DRAFT—DO NOT CITE OR QUOTE               EPA/63 5/R-10/001
                                           www.epa.gov/iris
         TOXICOLOGICAL REVIEW



                              OF



             INORGANIC ARSENIC

                       (CAS No. 7440-38-2)


          In Support of Summary Information on the
          Integrated Risk Information System (IRIS)


                         February 2010

                            NOTICE

This document is a final draft. This information is distributed solely for the purpose of
pre-dissemination peer review under applicable information quality guidelines. It has not
been formally disseminated by EPA. It does not represent and should not be construed to
represent any Agency determination or policy. It is being circulated for review of its
technical accuracy and science policy implications.
                  U.S. Environmental Protection Agency
                          Washington, DC

-------
                                    DISCLAIMER

       This document is a final draft for review purposes only.  This information is distributed
solely for the purpose of pre-dissemination peer review under applicable information quality
guidelines. It has not been formally disseminated by EPA.  It does not represent and should not
be construed to represent any Agency determination or policy. Mention of trade names or
commercial products does not constitute endorsement or recommendation for use.
                                          ii     DRAFT—DO NOT CITE OR QUOTE

-------
  TABLE OF CONTENTS—TOXICOLOGICAL REVIEW of INORGANIC ARSENIC
                              (CAS No. 7440-38-2)

LIST OF TABLES	VI

LIST OF FIGURES	VII

LIST OF ABBREVIATIONS	VIII

FOREWORD	XIII

AUTHORS, CONTRIBUTORS, AND REVIEWERS	XIV

1. INTRODUCTION	1

2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO   ASSESSMENTS 3
      2.1. PROPERTIES	3
      2.2. USES	3
      2.3. OCCURRENCE	4
      2.4. ENVIRONMENTAL FATE	5

3. TOXICOKINETICS	6
      3.1. ABSORPTION	7
      3.2. DISTRIBUTION	9
          3.2.1. Transport in Blood	9
          3.2.2. Tissue Distribution	10
          3.2.3. Cellular Uptake, Distribution, and Transport	14
      3.3. METABOLISM	15
          3.3.1. Reduction	19
          3.3.2. Arsenic Methylation	20
          3.3.3. Species Differences in the Methylation of Arsenic	23
          3.3.4. Thioarsenical Metabolites	24
      3.4. ELIMINATION	26
      3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS	27

4. HAZARD IDENTIFICATION	31
      4.1. STUDIES IN HUMANS	31
          4.1.1. Taiwan	32
          4.1.2. Japan	46
          4.1.3. South America	47
          4.1.4. North America (United States and Mexico)	52
          4.1.5. China	58
          4.1.6. Finland	59
          4.1.7. Denmark	61
          4.1.8. Australia	61
      4.2. PRECHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
          ANIMALS—ORAL	62
          4.2.1. Prechronic and Chronic Studies	62
          4.2.2. Cancer Bioassays	62
                4.2.2.1. Mice—Transplacental	62
                                      iii    DRAFT—DO NOT CITE OR QUOTE

-------
                4.2.2.2. Rat—Oral	65
                4.2.2.3. Other	66
      4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL	67
      4.4. OTHER STUDIES	67
          4.4.1. Possible Modes of Action and Key Events of Possible Importance	67
                4.4.1.1. In Vivo Human Studies	71
                4.4.1.2. In Vivo Experiments Using Laboratory Animals	75
                4.4.1.3. In Vitro Experiments	81
      4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS	93
      4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER
          CHARACTERIZATION	94
          4.6.1. Summary of Overall Weight-of-Evidence	94
          4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence	94
                4.6.2.1. Skin Cancer	96
                4.6.2.2. Lung Cancer	96
                4.6.2.3. Kidney, Bladder, and Liver Cancer	97
                4.6.2.4. In Utero Exposure	98
          4.6.3. Mode of Action Information	98
                4.6.3.1. General Comments onMOAs	98
                4.6.3.2. Low-Dose Extrapolation	102
      4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES	102
          4.7.1. Possible Childhood Susceptibility	102
          4.7.2. Possible Gender Differences	105
          4.7.3. Other	106
                4.7.3.1. Genetic Polymorphism	106
                4.7.3.2. Nutritional Status	109
                4.7.3.3. Cigarette  Smokers	Ill

5. DOSE-RESPONSE ASSESSMENTS	112
      5. LORAL REFERENCE DOSE (RfD)	112
      5.2. INHALATION REFERENCE CONCENTRATION (RfC)	112
      5.3. CANCER ASSESSMENT (ORAL EXPOSURE)	112
          5.3.1. Background: History of Cancer Risk Assessments for Arsenic	112
          5.3.2. Choice of Study/Data, Estimation Approach, and Input Assumptions	119
          5.3.3. Dose-Response Model Selection for Cancer Mortality in Taiwan	120
          5.3.4. Selection of Cancer Endpoints and Estimation of Risks for U.S. Populations
          	121
          5.3.5. Nonwater Arsenic  Intake and Drinking Water Consumption	123
          5.3.6. Dose-Response Data	125
          5.3.7. Risk Assessment Methodology	126
                5.3.7.1. Dose-Response Estimation Based on Taiwan Cancer Mortality Data
                       	127
                5.3.7.2. Estimation of Confidence Limits on Cancer Slope Parameters	128
                5.3.7.3. Estimation of LEDOT Values Using Relative Risk Models	129
                5.3.7.4. Estimation of Unit Risks	129
          5.3.8. Results	130
                5.3.8.1. Ingestion  Pathway Oral CSFs and Unit Risks	130
                5.3.8.2. Comparison to Previous Cancer Risk Estimates	132
                5.3.8.3. EDoi and  LEDM Estimates From Chen et al. (1988a, 1992), Ferreccio
                                        iv    DRAFT—DO NOT CITE OR QUOTE

-------
                     et al. (2000), and Chiou et al. (2001)	133
               5.3.8.4. Estimated Risk Associated With 10 ug/L Drinking Water Arsenic
                     From NRC (2001) and U.S. EPA (2005c)	135
               5.3.8.5. Sensitivity Analyses of Cancer Risk Estimates to Changes in
                     Parameter Values	137
               5.3.8.6. Sensitivity Analyses of Cancer Risk Estimates to Dose-Response
                     Model Form	142
               5.3.8.7. Significance of Cancer Risks at Low Arsenic Exposures	143
      5.4. CANCER ASSESSMENT (INHALATION EXPOSURE)	145

6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE-
RESPONSE	146
      6.1. HUMAN HAZARD POTENTIAL	146
      6.2. DOSE-RESPONSE	148
         6.2.1. Choice of Models	150
         6.2.2. Dose Metric	151
         6.2.3. Human Population Variability	151

7. REFERENCES	153

APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION	A-l

APPENDIX B. TABULAR DATA ON CANCER EPIDEMIOLOGY STUDIES	B-l

APPENDIX C. TABLES FOR STUDIES ON POSSIBLE MODE OF ACTION FOR
INORGANIC ARSENIC	C-l

APPENDIX D. IMMUNOTOXICITY	D-l

APPENDIX E. QUANTITATIVE ISSUES IN THE CANCER RISK ASSESSMENT FOR
INORGANIC ARSENIC	E-l

APPENDIX F. RISK ASSESSMENT FOR TOWNSHIPS AND LOW-EXPOSURE
TAIWANESE POPULATIONS	F-l
                                          DRAFT—DO NOT CITE OR QUOTE

-------
                                  LIST OF TABLES

Table 2-1. Chemical and Physical Properties of Arsenic and Selected Inorganic Arsenic
          Compounds (ATSDR, 2000; Merck Index, 1989)	4
Table 4-1. Summary of Number of Rows Derived From Peer-Reviewed Publications for
          Different Hypothesized Key Eventsa	71
Table 5-1. Historical Summary of Arsenic Risk Assessment Efforts	114
Table 5-2. Cancer Mortality Data Used in the Arsenic Risk Assessment	126
Table 5-3. Cancer Incidence Risk Estimates for Lung and Bladder Cancers in Males and
          Females21	131
Table 5-4. Combined Lung and Bladder Cancer Incidence Risk Estimate for the U.S. Population
          (Males and Females)	132
Table 5-5. Comparison of EDoi and LEDMa Estimates From Past Studies'3 With Those From the
          Current Analysis	134
Table 5-6. Comparison of cancer risk assessment results with estimates from NRC (2001) and
          U.S. EPA(2005c)	135
Table 5-7. Drinking water intake and body weight assumptions in females in recent arsenic risk
          assessments	136
Table 5-8. Theoretical maximum likelihood estimates of excess lifetime risk (incidence per
          10,000 people) of lung cancer and bladder cancer for US populations	137
Table 5-9. Arsenic oral CSFs (per mg/kg-d) for lung cancer and bladder cancer in US
          populations	137
Table 5-10. Sensitivity analysis of estimated cancer incidence risks associated with 10 ug/L to
          changes in modeling assumptions and inputs	139
Table 5-11. Proportional Changes in Cancer Risks at 10 ug/L Associated With Changes in
          Modeling Inputs and Assumptions	140
Table B-l. Taiwan Cancer Studies	B-4
Table B-2. Japan Cancer Studies	B-22
Table B-3. South America Cancer Studies	B-23
Table B-4. North America cancer studies	B-30
Table B-5. China cancer studies	B-38
Table B-6. Finland cancer studies	B-39
Table B-7. Denmark cancer studies	B-41
Table B-8. Australia Cancer Studies	B-42
Table C-l. In vivo human studies related to possible modes of action of arsenic in the
          development of cancer	C-l9
Table C-2. In vivo experiments on laboratory animals related to possible modes of action of
          arsenic in the development of cancer—only oral exposures	C-30
Table C-3. In vitro studies related to possible MOA of arsenic in the development of cancerC-57
Table D-l. Lymphocyte counts and labeling, mitotic, and replication indexes (mean ± se) in the
          peripheral blood lymphocytes in populations exposed to low (control) and high
          (exposed) levels of arsenic (Gonsebatt et al., 1994)	D-2
Table F-l. Coefficients from linear regressions of age-adjusted cancer risk versus arsenic doses
          for townships identified by Lamm et al. (2006)	F-5
Table F-2. Arsenic dose coefficients for study populations with median well water arsenic
          concentrations less than 127 ppb	F-6
                                          vi    DRAFT—DO NOT CITE OR QUOTE

-------
                                  LIST OF FIGURES

Figure 3-1. Traditional metabolic pathway for inorganic arsenic in humans	16
Figure 3-2. Alternative metabolic pathway for inorganic arsenic in humans proposed by
           Hayakawa et al. (2005)	17
Figure 3-3. Thioarsenical structures	25
Figure 4-1. Level of significant exposure of adult mice to sodium arsenite in drinking water in
           ppm As	76
Figure 5-1. Estimated oral CSFs for individual and combined cancer endpoints	132
Figure 5-2. Change in arsenic-related unit risk estimates associated with variations in input
           assumptions	140
Figure F-l. Lifetime crude total cancer risk (male + female) for the low- and high-exposure
           villages	F-4
                                          vii    DRAFT—DO NOT CITE OR QUOTE

-------
                  LIST OF ABBREVIATIONS
293 cells

2-AAAF
8-OHdG
AG06 cells
AGT
AIC
AMI
AP
APE
As
As111
Asv
AS3MT
AQP
ATG
ATO
ATSDR
B[or]P
BBDR
BCC
BER
BFD
BMI
BPDE

BrdU
BSO

BW or bw
CA
Caco-2
CAE
CASRN
CAT
CCA
CCRIS

cDNA
cen+
cen-
Chang cells

CHO
approximately (if before a listing of concentrations,
it applies to all)
a cell line derived from adenovirus-transformed
human embryonic kidney epithelial cells
2-acetoxyacetylaminofluorene
8-hydroxydeoxyguanosine
SV40-transformed human keratinocytes
average generation time
Akaike information criterion
acute myocardial infraction
activator protein or activating protein
apurinic/apyrimidinic endonuclease
arsenic
arsenite
arsenate
arsenic(+3 oxidation state) methyltransferase
aquaglycoporins
arsenic triglutathione
arsenic trioxide
Agency for Toxic Substances and Disease Registry
benzo[a]pyrene
biologically based quantitative dose-response
basal cell carcinoma
base excision repair
blackfoot disease
body mass index
benzo[a]pyrene diol epoxide, an active metabolite
ofB[or]P
bromodeoxyuridine
L-buthionine-S,R-sulphoximine (depletes GSH, y-
GCS inhibitor)
body weight
chromosome aberrations
a human intestinal cell line
cumulative arsenic exposure
Chemical Abstracts Service Registry Number
catalase (decomposes H2O2)
chromate copper arsenate
Chemical Carcinogenesis Research Information
System
complementary DNA
centromere positive
centromere negative
a human cell line thought to be derived from HeLa
cells
Chinese hamster ovary

             viii    DRAFT—DO NOT CITE OR QUOTE

-------
           Ill
     III
CI
c-Jun or c-jun
CL3 cells

COPD
CSF
DEB
DBS
dhfr gene
DHLP
DI-I or II or

dL
DMA

DMA
DMAV
DMAG
DMMTA111
DMMTAV
DMPS
DMSA

DNA
DNMT
DTT
DW
E. coli
ED
EGFR-ECD

EPA
ER-a
ERCC1
ERCC2
ERK
FAK
FPG

G6PDH
GAPDH
GI
GLM
GM04312C cells

GM-CSF
confidence interval
an AP-1 protein
human lung adenocarcinoma cells (established from
a non-small-cell lung carcinoma)
chronic obstructive pulmonary disease
cancer CSF
diepoxybutane (DNA crosslinking agent)
diethylstilbestrol
dihydrofolate reductase gene
dihydrolipoic acid
iodothyronine deiodinase-I or II orIn (are 3 forms of
this selenoenzyme)
deciliter
dimethyl arsenic (used when the oxidative state is
unknown or not specified)
dimethylarsenous acid
dimethylarsinic acid
dimethylarsinic glutathione
dimethylmonothioarsinic acid
dimethylmonothioarsonic acid
2,3-dimercaptopropane-l-sulfonic acid
dimercaptosuccinic acid or meso 2,3-
dimercaptosuccinic acid
deoxyribonucleic acid
DNA methyltransferase
dithiothreitol
drinking water
Escherichia coli
effective dose
extracellular domain of the epidermal growth factor
receptor
Environmental Protection Agency
estrogen receptor-alpha
excision repair cross-complement 1 component
excision repair cross-complementing rodent repair
deficiency, complementation group 2 (also known
as xeroderma pigmentosum group D or XPD)
extracellular signal-regulated kinase
focal adhesion kinase
formamidopyrimidine-DNA glycosylase (digestion
of DNA)
glucose-6-phosphate dehydrogenase
glyceraldehyde-3-phosphate dehydrogenase
gastrointestinal
generalized linear model
a SV-40 transformed XPA human fibroblast NER-
deficient cell line
granulocyte-macrophage colony-stimulating factor
                               ix    DRAFT—DO NOT CITE OR QUOTE

-------
GPx
GSH
GST
GSTO1
GSTP1-1
H69AR
H9c2 cells

HAC
HCC
HEALS
HELP cells
HepG2 cells

HGPRT
hGST-Ol
HMOX-1
hOGGl
HPBM
HSDB
HXT
IC50
IFN-y
IL
ILK
IRIS
IRR
iv
JAK
LED
LI
LOEC
LOEL
MADG
MAP
MCF-7 cells
M-CSF
MDA
mdm2
MEK
MI
MLE
MMA

MMAm
MMAV
MMS
MN
glutathione peroxidase
glutathione
glutathione-S-transferase
glutathione-^-transferase omega 1
glutathione-^-transferase Pl-1
a multi-drug resistant human cancer cell line
immortalized myoblast cell line derived from fetal
rat hearts
highest arsenic concentration
hepatocellular carcinoma
Health Effects of Arsenic Longitudinal Study
human embryo lung fibroblast cell line
human hepatocellular liver carcinoma cell line
(Caucasian)
hypoxanthine-guanine phosphoribosyltransferase
human glutathione-S-transferase omega 1
heme oxygenase 1
human 8-oxoguanine DNA glycosylase
human peripheral blood monocytes
Hazardous Substances Data Bank
hexose permease transporters
concentration that is needed to cause 50% inhibition
interferon-gamma
interleukin
integrin-linked kinase
Integrated Risk Information System
incidence rate ratio
intravenous
Janus kinase
lowest effective dose
labeling index
lowest observed effect concentration
lowest observed effect level
monomethylarsonic diglutathione
mitogen-activated protein
human breast carcinoma cell line
macrophage colony-stimulating factor
malondialdehyde
murine double minute 2 proto-oncogene
MAP/ERK kinase
mitotic index
maximum likelihood
monomethyl arsenic (used when oxidative state is
unknown or not specified)
monomethylarsonous acid
monomethyl arsonic acid
methyl methanesulfonate
micronuclei
                               x    DRAFT—DO NOT CITE OR QUOTE

-------
MNU
MOA
MPR2/cMOAT

mRNA
MRP
MTHFR
NAC
NAD
NADPH

NCHS
NCI
NER
NHEK cells
NK
NO
NRC
OATP-C
ODC
OGG1
OPP
OR
PARP
PBPK model
PBMC
PCNA
PCR
PGK
PHA
PMI
PNP
POD
ppb
ppm
PTEN
PYR
R15
RAGE
RBCs
RED
RfC
RfD
RI
RNS
ROS
RR
RT
N-methyl-N-nitrosourea
mode of action
multi-drug resistance associated protein 2
transporter
messenger ribonucleic acid
multidrug resistance protein
methylene trihydrofolate reductase
w-acetyl-cysteine
nicotinamide adenine dinucleotide
nicotinamide adenine dinucleotide phosphate-
oxidase
National Center for Health Statistics
National Cancer Institute
nucleotide excision repair
primary normal human epidermal keratinocytes
natural killer
nitric oxide
National Research Council
organic anion transporting polypeptide-C
ornithine decarboxylase
8-oxoguanine DNA glycosylase
Office of Pesticide Programs
odds ratio
poly(adenosine diphosphate-ribose) polymerase
physiologically based pharmacokinetic model
peripheral blood mononuclear cells
proliferating cell nuclear antigen
polymerase chain reaction
phosphoglyerate kinase
phytohemagglutinin
primary methylation indices
purine nucleoside phosphorylase
point of departure
parts per billion
parts per million
phosphatase and tensin homolog
person-years at risk
arsenic-resistent cells
receptor for advanced glycation end products
red blood cells
Reregi strati on Eligibility Decision
inhalation reference concentration
oral reference dose
replication index
reactive nitrogen species
reactive oxygen species
relative risk
real time
                               xi    DRAFT—DO NOT CITE OR QUOTE

-------
SAB
SAM
SBET
SCC
SCE
SCGE
Se
SEER
SHE cells
SIR
SMI
SMR
SOD
STAT
SV-HUC-1 cells
T3
T4
TAT
TCEP
Tg.AC
TGF-a
TMAm
TMAV
TMAO
TNF-a
TPA
Trx
TrxR
TWA
UCL
UROtsa
UV
V79 cells

VEGF
XRCC1
Science Advisory Board
S-adenosylmethionine
simplified bioaccessibility extraction test
squamous cell carcinoma
sister chromatid exchange
single cell gel electrophoresis
selenium
surveillance epidemiology and end result
Syrian hamster ovary cells
standardized incidence ratio
secondary methylation indices
standard mortality ratio
superoxide radical dismutase
signal transducer and activator of transcription
SV40 large T-transformed human urothelial cell line
thyroid hormone triiodothyronine
thyroid hormone thyroxine
tyrosine aminotransferase
tris(2-carboxylethyl)phospine
a strain of transgenic mice that contains the fetal
beta-globin promoter fused to the v-Ha-ras
structural gene (with mutations at codons 12 and 59)
and linked to a simian virus 40
polyadenylation/splice sequence
transforming growth factor alpha
trimethyl arsine
trimethylarsinic acid
trimethylarsine oxide
tumor necrosis factor alpha
12-O-tetradecanoyl phorbol-13-acetate
thioredoxin
thioredoxin reductase
time-weighted average
upper confidence limits
a SV40-immortalized human urothelium cell line
ultraviolet radiation
a cell line derived from lung fibroblasts of a male
Chinese hamster
vascular endothelial cell growth factor
X-ray repair cross-complimentary group 1
                               xii    DRAFT—DO NOT CITE OR QUOTE

-------
                                           FOREWORD
 1          The purpose of this Toxicological Review is to provide scientific support and rationale
 2   for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to inorganic
 3   arsenic.  It is not intended to be a comprehensive treatise on the chemical or toxicological nature
 4   of inorganic arsenic.
 5          The intent of Section 6, "Major Conclusions in the Characterization of Hazard and Dose
 6   Response," is to present the major conclusions reached in the derivation of the reference dose,
 7   reference concentration, and cancer assessment, where applicable, and to characterize the overall
 8   confidence in the quantitative and qualitative aspects of hazard and dose-response by addressing
 9   the quality of data and related uncertainties. The discussion is intended to convey the limitations
10   of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk
11   assessment process.
12          For other general information about this assessment or other questions relating to IRIS,
13   the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
14   hotline.iris@epa.gov (email address).
                                                xiii    DRAFT—DO NOT CITE OR QUOTE

-------
                 AUTHORS, CONTRIBUTORS, AND REVIEWERS

CHEMICAL MANAGER/AUTHOR

Santhini Ramasamy, Ph.D., MPH, DABT
Office of Science and Technology
Office of Water
U.S. Environmental Protection Agency
Washington, DC

OFFICE OF RESEARCH AND DEVELOPMENT CO-LEAD/AUTHOR

Reeder Sams, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC

AUTHORS

Robyn B. Blain, Ph.D.
Gregory M. Blumenthal, Ph.D.
William M. Mendez, Ph.D.
Welford C. Roberts, Ph.D.
ICF International
Fairfax, VA

Paul B  Selby, Ph.D., DABT
RiskMuTox
Oak Ridge, TN

Arthur W. Stange, Ph.D.
Oak Ridge Associated Universities
Arvada, CO 80005

Susan M. Wells, M.P.H.
Oak Ridge Associated Universities
Oak Ridge, TN 37831-0117

CONTRIBUTORS

Elizabeth Doyle, Ph.D.
Office of Science and Technology
Office of Water
U.S. Environmental Protection Agency
Washington, DC
                                        xiv    DRAFT—DO NOT CITE OR QUOTE

-------
Jonathan Chen, Ph.D.
Office of Pesticide Programs
U.S. Environmental Protection Agency
Washington, DC

Andrew Schulman, Ph.D.
Office of Enforcement and Compliance Assurance
U.S. Environmental Protection Agency
Washington, DC

Chao Chen, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Paul White, M.S.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Washington, DC

Irene Dooley
Office of Water
U.S. Environmental Protection Agency
Washington, DC

Brenda Foos, Ph.D.
Office of Children's Health Protection
U.S. Environmental Protection Agency
Washington, DC

Molly Rosett
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC
                                         xv    DRAFT—DO NOT CITE OR QUOTE

-------
REVIEWERS

       This document has been reviewed by EPA scientists, interagency reviewers from other
federal agencies, and the public, and peer reviewed by independent scientists external to EPA. A
summary and EPA's disposition of the comments received from the independent external peer
reviewers and from the public is included in Appendix A.

INTERNAL EPA REVIEWERS

Ila Cote, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency

Joyce Morrissey Donohue, Ph.D.
Office of Science and Technology
Office of Water
U.S. Environmental Protection Agency

Hisham El-Masri, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Nicole Hagan
ORISE, National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency

Elaina Kenyon, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Kirk Kitchin, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Andrew Kligerman, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Danelle Lobdell, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

                                         xvi    DRAFT—DO NOT CITE OR QUOTE

-------
Bob Hetes
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Stephen Nesnow, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Julian Preston, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

David Thomas, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

John Vandenberg, Ph.D.
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency

Tim Wade, Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency

Debra Walsh
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency

Doug Wolf, D.V.M., Ph.D.
National Health and Environmental Effects Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
                                        xvii   DRAFT—DO NOT CITE OR QUOTE

-------
EXTERNAL PEER REVIEWERS

Science Advisory Board Arsenic Review Panel

CHAIR

Genevieve Matanoski, M.D., Ph.D.
Johns Hopkins University

MEMBERS

H. Vasken Aposhian, Ph.D.
The University of Arizona

Aaron Barchowsky, Ph.D.
University of Pittsburgh

David Brusick, Ph.D.
Retired, Convance Labs

Kenneth P. Cantor, Ph.D.
National Cancer Institute

John (Jack) Colford, Ph.D.
University of California

Yvonne P. Dragan, Ph.D.
National Center for Toxicological Research,
Food and Drug Administration

Sidney Green, Ph.D.
Howard University

Sioban Harlow, Ph.D.
University of Michigan

Steven Heeringa, Ph.D.
University of Michigan

Claudia Maria Hopenhayn, Ph.D.
University of Kentucky

James E. Klaunig, Ph.D.
Indiana University

X. Chris Le, Ph.D.
University of Alberta
                                        xviii   DRAFT—DO NOT CITE OR QUOTE

-------
Michele Medinsky, Ph.D.
Toxcon

Kenneth Portier, Ph.D.
American Cancer Society
Atlanta, GA

Barry Rosen, Ph.D.
Wayne State University

Toby Rossman, Ph.D.
New York University

Miroslav Styblo, Ph.D.
University of North Carolina

Justin Teeguarden, Ph.D.
Pacific Northwest National Laboratory

Michael Waalkes, Ph.D.
National Institute of Environmental Health Science

Janice Yager, Ph.D.
Electric Power Research Institute
                                         xix    DRAFT—DO NOT CITE OR QUOTE

-------
                                       1. INTRODUCTION
 1          This document presents background information and justification for the Integrated Risk
 2    Information System (IRIS) Summary of the hazard and dose-response assessment of inorganic
 3    arsenic.  The IRIS Summary may include oral reference dose (RfD) and inhalation reference
 4    concentration (RfC) values for chronic and other exposure durations, as well as a carcinogenicity
 5    assessment.
 6          This document is based on EPA reviews of the reports Arsenic in Drinking Water and
 7    Arsenic in Drinking Water, 2001 Update published by the National Research Council (NRC) in
 8    1999 and 2001, respectively.  In writing those reports, the NRC arsenic committee considered
 9    presentations at the committee's public meetings, comments from the public, and the comments
10    made by technical experts on the draft NRC arsenic reports. The conclusions, recommendations,
11    and final content of the NRC (1999, 2001) reports rest entirely with the committee and the NRC.
12    This IRIS document—based on reviews of those reports—has undergone evaluation by EPA
13    health scientists  from several program offices and regional offices, interagency review, and
14    external peer review by the Science Advisory Board (SAB).
15          Compared to the draft Toxicological Review submitted to the SAB in 2005, this
16    assessment is expanded: it provides a detailed review of epidemiological studies and the mode of
17    action (MOA) studies, as well as revisions to the dose-response analysis to address the
18    recommendations of the SAB (SAB, 2007). Specifically, it includes additional sensitivity
19    analyses on the effects of modeling assumptions on estimated cancer risk.
20          The RfD and RfC, if derived, provide quantitative information for use in risk assessments
21    for health effects known or assumed to be produced through a non-linear (presumed threshold)
22    MO A. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with uncertainty
23    spanning perhaps an order of magnitude) of a daily exposure to the human population (including
24    sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a
25    lifetime. The inhalation RfC (expressed in units of mg/m3) is analogous to the oral RfD, but
26    provides a continuous inhalation exposure estimate. The  inhalation RfC considers both toxic
27    effects on the respiratory system (portal of entry) and toxic effects peripheral to the respiratory
28    system (extrarespiratory or systemic effects). Reference values are generally derived for chronic
29    exposures (up to a lifetime), but may also be derived for acute (< 24 hours), short-term (>24
30    hours to 30 days), and subchronic (>30 days to 10% of lifetime) exposure durations, all of which
31    are derived based on an assumption of continuous exposure throughout the duration specified.
32    Unless specified otherwise, the RfD and RfC are derived for chronic exposure duration.
33          The carcinogenicity assessment provides information on the carcinogenic hazard
34    potential of the substance in question and quantitative estimates of risk from oral and inhalation
35    exposures may be derived. The information includes a weight-of-evidence judgment of the
36    likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic
                                                 1

-------
 1   effects may be expressed.  Quantitative risk estimates may be derived from the application of a
 2   low-dose extrapolation procedure. If derived, the oral cancer CSF (CSF) is a plausible upper
 3   bound on the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk
 4   is a plausible upper bound on the estimate of risk per ug/m3 air breathed.
 5          Development of these hazard identification and dose-response assessments for inorganic
 6   arsenic has followed the general guidelines for risk assessment set forth by the National
 7   Research Council (NRC, 1983). EPA Guidelines and Risk Assessment Forum Technical Panel
 8   Reports that may have been used in the development of this assessment include the following:
 9   Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines
10   for Mutagenicity Risk Assessment (U. S. EPA, 1986b), Recommendations for and Documentation
11   of Biological Values for Use in Risk Assessment (U.S. EPA, 1988a), Guidelines for
12   Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Use of the Benchmark Dose
13   Approach in Health Risk Assessment (U.S. EPA, 1995), Guidelines for Reproductive Toxicity
14   Risk Assessment (U.S. EPA,  1996), Guidelines for Neurotoxicity Risk Assessment (U.S. EPA,
15   1998), Science Policy Council Handbook: Peer Review (U. S. EPA, 2000a), Science Policy
16   Council Handbook: Risk Characterization (U.S. EPA, 2000b), Benchmark Dose Technical
17   Guidance Document (U.S. EPA, 2000c), Supplementary Guidance for Conducting Health Risk
18   Assessment of Chemical Mixtures (U.S. EPA, 2000d), A Review  of the Reference Dose and
19   Reference Concentration Processes (U.S. EPA, 2002),  Guidelines for Carcinogen Risk
20   Assessment (U. S. EPA, 2005a), Supplemental Guidance for Assessing Susceptibility from Early-
21   Life Exposure to Carcinogens (U.S. EPA, 2005b), Science Policy Council Handbook: Peer
22   Review (U.S. EPA, 2006a), and A Framework for Assessing Health Risks of Environmental
23   Exposures to Children (U.S.  EPA, 2006b).
24          The literature search  strategy employed for this compound was based on the Chemical
25   Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
26   scientific information submitted by the public to the IRIS Submission Desk was also  considered
27   in the development of this document.  The relevant literature was reviewed through December,
28   2007; however, a few references from 2008 have also been included.
29

-------
        2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS

     2.1. PROPERTIES
 1          Arsenic (As) is a metalloid that can exist in the -3, 0, +3, and +5 oxidation states.l The
 2   arsenite (As111; +3) and arsenate (Asv ; +5) forms are the primary forms found in drinking water.
 3   The chemical and physical properties of arsenic are listed in Table 2-1.

     2.2. USES
 4          The metalloid, arsenic, is used for hardening copper and lead alloys (HSDB, 2005). It
 5   also is used in glass manufacturing as a decolorizing and refining agent, as a component of
 6   electrical devices, in the semiconductor industry, and as a catalyst in the production of ethylene
 7   oxide. Arsenic compounds are used as a mordant in the textile industry, for preserving hides, as
 8   medicinals, pesticides, pigments, and wood preservatives.  Production of chromate copper
 9   arsenate (CCA), a wood preservative whose production is currently being phased out, accounts
10   for about 90% of the domestic consumption of arsenic (ATSDR, 2007).
11
      1 Oxidation states for arsenic have been abbreviated differently by different organizations or authors. For example,
      arsenite can be abbreviated as either " As(m)'
      document uses the superscript abbreviation.
arsenite can be abbreviated as either "As(m)" or "As111"; both refer to trivalent inorganic arsenic compounds. This
                                                  3        DRAFT—DO NOT CITE OR QUOTE

-------
     Table 2-1. Chemical and Physical Properties of Arsenic and Selected Inorganic Arsenic
               Compounds (ATSDR, 2000; Merck Index, 1989)

CAS No.
Oxidation State
Molecular Weight
Synonyms
Physical State
(25°C)
Boiling Point (°C)
Melting Point (°C)
Density (g/cm3)
Vapor Pressure
(20°C)
Taste Threshold
Odor Threshold
Conversion Factor
Arsenic
7440-38-2
0
74.9
metallic arsenic,
gray arsenic
solid
613 (sublimes)
817@28atm
5.7
—
—
—
—
As2O3
1327-53-3
+3
197.8
arsenic trioxide,
arsenolite,
white arsenic
(+3)
solid
465
312
3.7
—
—
—
—
As2O5
1303-28-2
+5
229.8
arsenic
pentoxide,
arsenic acid
anhydride (+5)
solid
—
315
(decompose)
4.3
—
—
—
—
NaAsO2
7784-46-5
+3
129.9
sodium arsenite
(+3)
solid
—
—
1.8
—
—
—
—
Na2HAsO4
7778-43-0
+5
185.9
disodium
arsenate (+5)
solid
—
86.3
1.8
—
—
—
—
      — No data available

     2.3. OCCURRENCE
 1          Arsenic naturally makes up about 3.4 parts per million (ppm) of the Earth's crust, where
 2   it is the twentieth most abundant element (ATSDR, 2007; Merck Index, 1989). Arsenic leaches
 3   from natural weathering of soil and rock into water, and low concentrations of arsenic are found
 4   in water, food, soil, and air. However, industrial activities such as coal combustion and smelting
 5   operations release higher concentrations of arsenic to the environment (Adams et al., 1994). The
 6   highest background arsenic levels found in the environment are in soils, with concentrations
 7   ranging from 1 to 40 ppm (ATSDR, 2007). Food typically contains arsenic concentrations of 20
 8   to 140 parts per billion (ppb) (ATSDR, 2007). The majority of surface and ground waters
 9   contain less than 10 ppb (although levels of 1,000-3,400 ppb have been reported, especially in
10   areas of the western United States). Average arsenic content in drinking water in the United
11   States is 2 ppb; 12% of water supplies from surface water in the central United States and 12%
12   of ground water sources in the western United States exceed 20 ppb (ATSDR, 2007).  Mean
13   arsenic concentrations in ambient air have generally been found to range from 1 to 2,000 ng/m3
14   (ATSDR, 2007).
                                                        DRAFT—DO NOT CITE OR QUOTE

-------
     2.4. ENVIRONMENTAL FATE
 1          Arsenic as a free element (0 oxidation state) is rarely encountered in the environment
 2   (HSDB, 2005). Under normal conditions in water, arsenic is present as soluble inorganic Asv
 3   because it is more thermodynamically stable in water than As111. In soil there are many biotic
 4   and abiotic processes controlling arsenic's overall fate and environmental impact. Arsenic in
 5   soil exists in various oxidation states and chemical species, depending upon soil pH and
 6   oxidation-reduction potential (ATSDR, 2007). Arsenic is largely immobile in agricultural soils,
 7   and tends to remain in upper soil layers (ATSDR, 2007).  However, reducing conditions form
 8   soluble mobile forms of arsenic and leaching is greater in sandy soil than in clay loam (ATSDR,
 9   2007). The most influential parameter affecting arsenic mobility is the iron content of the soil.
10
                                                         DRAFT—DO NOT CITE OR QUOTE

-------
                                       3. TOXICOKINETICS
 1          This Toxicological Review discusses oral waterborne arsenic exposure.  It does not
 2    specifically address inhalation exposures, though they are also common. Dermal exposure and
 3    exposure from food consumption, however, can be significant and may be confounding variables
 4    in epidemiological studies. Therefore, this report's toxicokinetic information focuses on oral
 5    exposure from water sources, but absorption from dermal exposure and arsenic in food is also
 6    briefly addressed.
 7          The behavior of arsenic in the body is very complex.  After absorption, inorganic arsenic
 8    can undergo a complicated series of enzymatic and non-enzymatic oxidation, reduction, and
 9    conjugation reactions.  Although all these reactions may occur throughout the body, the rate at
10    which they occur varies greatly from organ to organ.  In addition, there are important differences
11    in arsenic metabolism across animal species, and these variations make it difficult to identify
12    suitable animal models for predicting human metabolic patterns.
13          Each metabolic transformation affects the subsequent biokinetic behavior (transport,
14    persistence, elimination) and toxicokinetics of the arsenic species.  Thus, absorption, transport,
15    and metabolic processes  are highly interdependent and cannot easily be discussed separately.
16    The general pattern described in this chapter involves the gastrointestinal (GI) absorption of
17    inorganic arsenic species, followed by a cascade of oxidation-reduction reactions and
18    methylation steps, resulting in the partial transformation of the inorganic species into mono- or
19    dimethylated species (collectively referred to as MMA and DMA, recognizing that there is often
20    ambiguity in characterizing the oxidation state of the methylarsenic compounds).  Conjugated
21    arsenic species, either methylated or not (e.g., glutathione conjugates or other sulfur-containing
22    derivatives), also may be produced.
23          As discussed in Section 3.3, several metabolic schemes have been proposed that describe
24    the general pathway that converts inorganic arsenic to its primary metabolites MMA and DMA.
25    These pathways involve  numerous enzymes and cofactors. Some of the proposed metabolic
26    pathways involve the cycling of arsenic  species back and forth between the +3 (trivalent) and +5
27    (pentavalent) oxidation states, and there is evidence that key metabolic processes may be
28    saturable, so that metabolic patterns differ with exposure levels.  MMA, DMA, and inorganic
29    arsenic levels in tissues, blood, and urine are the most easily and frequently measured
30    metabolites; the relative levels of these compounds in blood or urine are often the primary
31    evidence in support of one or another metabolic pathway. Genomic tools are being increasingly
32    employed to better characterize human arsenic metabolism and to identify individuals at higher
33    risk from arsenic exposures.
34
                                                          DRAFT—DO NOT CITE OR QUOTE

-------
      3.1. ABSORPTION
 1          Water-soluble forms of inorganic arsenic (both trivalent and pentavalent) are readily
 2    absorbed from the GI tract in experimental animal models (about 80-90% 0.62 mg/kg of sodium
 3    arsenate; Freeman et al., 1995) as well as humans (Pomroy et al., 1980, who recovered 62% of a
 4    0.06 ng dose of arsenic in seven days).  Monomethyl arsonic acid (MMAV) and dimethylarsinic
 5    acid (DMAV) also appear to be well absorbed (75-85%) in humans and experimental animals
 6    (Stevens et al., 1977; Buchet et al., 1981; Yamauchi and Yamamura, 1984; Hughes et al., 2005).
 7    Using an in vivo swine test, however, Juhasz et al. (2006) determined that MMA (oxidation
 8    state not specified) and DMA (oxidation state not specified) were poorly absorbed, with only
 9    16.7% and 33.3%, respectively, bioavailable.
10          Laparra et al. (2006) used a Caco-2 permeability model, which measured transport
11    through  a monolayer of human intestinal cells, to examine the intestinal permeability of As111.  A
12    decrease in the apical to basolateral permeability with increasing dose was found, indicating the
13    presence of a saturable intestinal transport system. The data also indicated that Caco-2 cells
14    have a secretory system for As111. In an earlier study, Laparra et al. (2005a) demonstrated that
15    the retention and transport of As111 in Caco-2 cells was more efficient than that of Asv. However,
16    this could have been due to the presence of phosphate in the culture medium, which would
17    compete with arsenate for transport across the membrane.
18          Gastrointestinal absorption  of low-solubility arsenic compounds such as arsenic
19    trisulfide, lead arsenate, arsenic selenide, gallium arsenide (Mappes, 1977; Webb et al., 1984;
20    Yamauchi et al., 1986), and arsenic-contaminated soil (Freeman et al., 1995) is much less
21    efficient than that of soluble inorganic arsenic compounds. The degree of absorption of arsenic
22    from soil was found to be dependent on the arsenic species present in the soil and on the type of
23    soil. Juhasz et al. (2007) performed in vivo bioavailability studies in swine and determined that
24    the bioavailability of total arsenic in soils was highly variable, with a range of 6.9% to 74.7%
25    depending on the soil type.  They also determined that a simplified bioaccessibility extraction
26    test (SBET; a rapid in vitro chemical extraction method) had results highly correlated with the in
27    vivo results.  Therefore, they concluded that the less expensive in vitro test was just as effective
28    for determining bioavailability.
29          There is little information concerning the bioavailability of inorganic arsenic from
30    various types of food (NRC, 1999,  2001). However, there have been recent studies examining
31    the bioaccessibility of arsenic from rice (Laparra et al., 2005b; Juhasz et al., 2006).  Laparra et
32    al. (2005b) determined that while cooking rice (they tested several types, but did not specify
33    them) in deionized water caused no change in arsenic content compared to the raw form, cooking
34    in water contaminated with 0.5 ug/mL of Asv increased the inorganic arsenic content 5- to 17-
35    fold over the raw rice.  Laparra et al. subjected the rice samples (10 grams) to an in vitro
36    simulated digestion process. They  measured levels of soluble arsenic to determine
                                                 7        DRAFT—DO NOT CITE OR QUOTE

-------
 1    bioaccessibility. The results demonstrated that large amounts of the arsenic (i.e., 63%-99%),
 2    mainly in the pentavalent form, were bioaccessible for intestinal absorption. Ackerman et al.
 3    (2005) also found 89%-105% bioaccessible arsenic in different samples of white and brown rice
 4    cooked in water containing Asv.
 5          Juhasz et al. (2006) examined the bioavailability of arsenic from rice (mainly white rice
 6    samples) using an in vivo swine assay.  Quest rice was grown in arsenic-contaminated water and
 7    cooked in arsenic-free water.  This caused the rice to contain arsenic, mainly in the form of
 8    DMA. Administration of the cooked rice to swine demonstrated a bioavailability similar to that
 9    observed after a single oral administration of DMA in water (i.e., 33.3%). Basmati white rice
10    cooked in water contaminated with 1,000 ppb of Asv,  which contained entirely inorganic arsenic
11    as a result of the arsenate in the cooking water, had a bioavailability of 89.4%.
12          Although there have been no studies performed on the rate of inorganic arsenic
13    absorption through intact human skin, systemic toxicity due to high dermal occupational
14    exposure to aqueous inorganic arsenic solutions indicates that the skin may be a significant
15    exposure route (Hostynek et al., 1993).  The systemic  absorption via the skin from less
16    concentrated solutions, however, appears to be low (NRC, 1999).  An in vivo study by Wester et
17    al. (1993) demonstrated that 2% to 6% of radiolabeled arsenate (as a water solution) was
18    absorbed by rhesus monkey skin over a 24-hour period.  Results demonstrated that the lower
19    dose (0.000024 ug/cm2) was absorbed at a greater rate (6%) than the higher arsenic exposure
20    (2.1 ug/cm2; 2%), but the difference did not reach statistical significance. Wester et al. (2004)
21    performed another in vivo dermal absorption study using female rhesus monkeys.  Using the
22    levels excreted in the urine and the applied dose, they  calculated that 0.6% to 4.4% was absorbed
23    in the three monkeys tested, which was  similar to their previous results. In vitro results on
24    human skin (from donors) demonstrated a 24-hour absorption of 1.9% (Wester et al., 1993).
25    Mouse dorsal skin was demonstrated to absorb 30% to 60% of applied arsenic (Rahman et al.,
26    1994) using similar in vitro testing, with 60% to 90%  of the absorbed arsenic being retained in
27    the skin. NRC (1999) suggests this indicates that inorganic arsenic binds significantly to skin
28    and hair.  Lowney et al. (2007) found that dermal absorption of arsenic from soils was negligible
29    in an in vivo study in rhesus monkeys.
30          Harrington et al. (1978) compared arsenic metabolite levels in the urine from a group of
31    people in Fairbanks, Alaska, who had arsenic-contaminated water (345 ppb) in their home, but
32    drank only bottled water, with the levels measured in a group of people who drank home water
33    containing less than 50 ppb. The results demonstrated that the group with high arsenic in their
34    water had close to the same average concentration of total  arsenic metabolites in their urine (i.e.,
35    43 ug/L) as the group who drank home water with less than 50 ppb arsenic (i.e., 38 ug/L in
36    urine), indicating possible dermal  absorption via bathing or other exposure sources. Levels of
                                                          DRAFT—DO NOT CITE OR QUOTE

-------
 1    arsenic in the bottled water, however, were not measured.  Possible exposure through using
 2    contaminated water for cooking also was not examined.

      3.2. DISTRIBUTION
 3           The retention and distribution patterns of arsenic species are strongly dependent on their
 4    chemical properties. While both As111 and Asv bind to sulfhydryl groups, As111 has approximately
 5    a 5- to 10-fold greater affinity for sulfhydryl groups than Asv (Jacobson-Kram and Montalbano,
 6    1985).  Cellular uptake rates and resulting tissue concentrations are substantially lower for the
 7    pentavalent than for the trivalent forms of arsenic. DMA (an important metabolite of inorganic
 8    arsenic) appears to be more readily excreted than MMA (NRC, 2001). Liu et al. (2002) found
 9    arsenite to be transported into cells by aquaglycoporins (AQP7 and AQP9), whose usual
10    substrates are water and glycerol. Liu et al. (2006a) also detected transport of
11    monomethylarsonous acid (MMA111) by AQP9. MMA111 was transported at a rate nearly 3 times
12    faster than As111. A hydrophobic residue at position 64 was required for the transport of both
13    species, suggesting that both species are transported by AQP9 using the same translocation
14    pathway. Asv, however, has been suggested to be transported by the phosphate transporter
15    (Huang and Lee, 1996). Retention of arsenic can vary not only with its form, but also with tissue
16    (Thomas et al., 2001).  Other factors that affect the retention and distribution of arsenic include
17    the chemical species, dose level, methylation capacity, valence state, and route of administration.
18
      3.2.1. Transport in Blood
19
20           Once arsenic is absorbed, it is transported in the blood throughout the body. In the blood,
21    inorganic arsenic species are generally bound to sulfhydryl groups of proteins and low-
22    molecular-weight compounds such as glutathione (GSH) and cysteine (NRC, 1999).  Binding of
23    As111 to GSH has been demonstrated by several investigators (Anundi et al., 1982; Scott et al.,
24    1993; Delnomdedieu et al., 1994a,b). Because of the different binding and transport
25    characteristics of various arsenic compounds, the persistence in the blood varies across species.
26    Inorganic arsenic elimination in humans has been observed to be triphasic, with first-order half-
27    lives for elimination of 1 hour, 30 hours, and 200 hours (Mealey et al., 1959, used As111; Pomroy
28    et al., 1980, used Asv). A single intravenous (iv) dose of 5.8 ug As/kg body weight (in the form
29    of 73 Asv) administered to two male chimpanzees had a half-life plasma elimination rate of
30    1.2 hours and a half-life elimination rate from red blood cells (RBCs) of about 5 hours (Vahter et
31    al., 1995a).
32          Rats retain arsenic in the blood considerably longer than other species because
33    dimethylarsenous acid (DMA111) and DMAV accumulate in RBCs, apparently bound to
34    hemoglobin (Odanaka et al., 1980; Lerman and Clarkson,  1983; Vahter, 1983; Vahter et al.,

                                                 9        DRAFT—DO NOT CITE OR QUOTE

-------
 1    1984). Naranmandura et al. (2007) found that 75% of an oral dose of arsenite accumulated in rat
 2    RBCs mainly in the form of DMA111; however, less than 0.8% of the same dose to hamsters was
 3    found in their RBCs. Rats maintained this level in their RBCs for at least 7 days whereas the
 4    treated hamsters had levels equivalent to those in controls by 3 days after the administered dose.
 5    Stevens et al. (1977) calculated an elimination half-life for inorganic arsenic of 90 days in rat
 6    whole blood after a single  oral dose of 200 mg/kg. Lanz et al.  (1950) also reported a high
 7    retention of arsenic in the blood of cats, although less than in the rat.  However, they did not
 8    determine if the retained arsenic was in the form of DMA.
 9          The relative concentration of arsenic in human plasma and RBCs apparently differs
10    depending on exposure levels and the health status of the exposed individuals.  Heydorn (1970)
11    reported that healthy people in Denmark with low arsenic exposures had similar arsenic
12    concentrations in their plasma and RBCs (2.4 ug/L and 2.7 ug/L, respectively; the RBC:plasma
13    ratio was 1.1).  However, normal healthy Taiwanese exposed to arsenic-contaminated water had
14    plasma levels of 15.4 ug/L and RBCs levels  of 32.7 ug/L (RBC:plasma ratio 2.1). Blackfoot
15    disease (BFD) patients and their unaffected family members had 38.1 ug/L and 93 ug/L of
16    arsenic species in their plasma and RBCs, respectively (RBC:plasma ratio 2.4). These results
17    indicate  a different distribution between the RBCs and the plasma depending on exposure levels.
18    However, examining the BFD patients and their families, who presumably have the same
19    exposure levels, demonstrates a different distribution, possibly due to disease state.  BFD
20    patients had a ratio of 3.3 (106 ug/L in RBCs and 32.3 ug/L in plasma) compared to 1.8 (81 ug/L
21    in RBCs and 45.2 ug/L in plasma) in family  members without BFD. This indicates that
22    accumulation of arsenic in the RBCs is greater as exposure increases and possibly even greater
23    when health is compromised.  The ratio between plasma and RBC arsenic concentrations may
24    also depend on the exposure form of arsenic (NRC, 1999).
25
      3.2.2. Tissue Distribution
26
27          Once arsenic compounds enter the blood, they are transported and taken up by other
28    tissues and organs, with a large proportion of ingested material being subject to "first pass"
29    processing through the liver.  Uptake varies with arsenic species, dose, and organ. The observed
30    uptake of inorganic arsenic (mainly As111) in  the skin, hair, oral mucosa, and esophagus is most
31    likely due to the binding of inorganic arsenic species with sulfhydryl groups of keratin in these
32    organs.  In studies using rabbits and mice, where the transfer of methyl groups from
33    S-adenosylmethionine (SAM; a proposed major reaction during arsenic metabolism;  see Section
34    3.3) was chemically inhibited, the concentration of arsenic in most tissues (especially the skin)
35    was found to be increased  (Marafante and Vahter, 1984).  The important role of chemical
                                                10       DRAFT—DO NOT CITE OR QUOTE

-------
 1    binding of arsenic species also is supported by the observed tissue distribution in the marmoset
 2    monkey, which does not methylate inorganic arsenic (Vahter et al., 1982).
 3          Human subjects also have demonstrated high concentrations of arsenic in tissues
 4    containing a high content of cysteine-containing proteins, including the hair, nails, skin, and
 5    lungs. Total arsenic concentrations in these tissues of human subjects exposed to background
 6    levels of arsenic ranged from 0.01 to  1.0 mg/kg of dry weight (Liebscher and Smith, 1968; Cross
 7    et al., 1979). Benign and malignant skin lesions from 14 patients, with a minimum of 4 years of
 8    exposure to inorganic arsenical medication, had higher arsenic  levels (0.8 to 8.9 ppm) than six
 9    subjects with no history of arsenic intake (0.4 to 1.0 ppm; Scott, 1958).  In West Bengal, India,
10    where the average arsenic concentration in the drinking water ranges from 193 to 737 ppb,
11    arsenic concentrations in the skin, hair,  and nails were 1.6-5.5, 3.6-9.6, and 6.1-22.9 mg/kg dry
12    weight, respectively (Das et al., 1995).  Mandal et al. (2004) measured different arsenic species
13    in the hair and fingernails of 41 subjects in West Bengal, India, who were drinking arsenic-
14    contaminated water and in blood from 25 individuals who had  stopped drinking contaminated
15    water 2 years earlier. Results were: fingernail contained As111 (62.4%), Asv (20.2%), MMAV
16    (5.7%), DMA111 (8.9%), and DMAV (2.8%); hair contained As111 (58.9%), Asv (34.8%), MMAV
17    (2.9%), and DMAV (3.4%); RBCs contained arsenobetaine (22.5%) and DMAV (77.5%); and
18    blood plasma contained arsenobetaine (16.7%), As111 (21.1%), MMAV (27.1%), and DMAV
19    (35.1).  However, the amount of arsenic in these tissues resulting from other exposure pathways
20    (e.g., dermal exposure) was not determined.
21          The  longest retention of inorganic arsenic in mammalian tissues during experimental
22    studies has been observed in the skin  (Marafante and Vahter, 1984), hair, squamous epithelium
23    of the upper GI tract (oral cavity, tongue, esophagus, and stomach wall), epididymis, thyroid,
24    skeleton, and the lens of the eye (Lindgren et al.,  1982). Although the study authors measured
25    radioactive arsenic (74As) in the various tissues, they did not differentiate between the different
26    species of arsenic and could not determine if accumulation was due to the originally
27    administered compound or metabolites.  Arsenic levels in all these tissues, with the exception of
28    the skeleton, were greater in mice administered As111 than in mice administered Asv.  This could
29    indicate that As111 is taken up more efficiently than Asv and that less was found in the tissues of
30    AsV-treated mice due to the initial reduction to As111. The calcified areas of the skeleton in mice
31    administered As  accumulated and retained more arsenic than mice administered As , most
32    likely due to the similarities between Asv and phosphate, causing a substitution of phosphate by
33    Asv in the apatite crystals in bone.  Marmoset monkeys were found not to accumulate arsenic in
34    the ocular lens or the thyroid (Vahter et al., 1982); however, intravenous administration of 74As-
35    labelled DMA to mice resulted in accumulation of DMA in the ocular lens and the thyroid.
36    Marmoset monkeys do not methylate arsenic and DMA was found to accumulate in the ocular
37    lens and thyroid; this suggests that only the methylated species are retained in these organs.

                                                11        DRAFT—DO NOT CITE OR QUOTE

-------
 1    Mouse tissues with the largest retention of DMA were the lens of the eyes, thyroid, lungs, and
 2    intestinal mucosa (Vahter et al., 1984). Methylated arsenic species (DMA), in general, have a
 3    shorter tissue retention time in mice than rats (i.e., more than 99% of the administered dose was
 4    eliminated in mice within 3 days as compared to 50% in rats due to accumulation in blood)
 5    (Vahter etal., 1984).
 6          Hughes et al. (2003) estimated that a steady-state, whole-body arsenic balance was
 7    established after nine repeated oral daily doses of 0.5 mg As/kg as radioactive Asv in adult
 8    female B6C3F1 mice.  Twenty-four hours after the last dose, the whole-body burden of arsenic
 9    was about twice that observed after a single dose.  The rate of elimination was slower following
10    repeated doses. Accumulation of radioactivity was highest in the bladder, kidney, and skin,
11    while the loss of radioactivity was greatest from the lungs and slowest from the skin. Atomic
12    absorption spectrometry was used to characterize the organ distribution of arsenic species.
13    MMA was detected in all tissues except the bladder. DMA was found at the highest levels in the
14    bladder and lung after a single oral exposure, with increases after repeated exposures. Inorganic
15    arsenic was predominantly found in the kidney.  After a single oral exposure of Asv (0.5 mg
16    As/kg), DMA was the predominant form of arsenic in the liver, but after nine repeat exposures,
17    the proportion of DMA decreased while the proportion of inorganic arsenic increased (this could
18    indicate metabolic saturation or GSH depletion; see Section 3.3 for more details). A
19    trimethylated form of arsenic also was detected in the liver.
20          Kenyon et al. (2005a) examined the  time course of tissue distribution of different arsenic
21    species after a single oral dose of 0, 10, or 100 umole As/kg as sodium arsenate to adult female
22    B6C3F1 mice.  The concentrations of all forms of arsenic were lower in the blood than in other
23    organs across all doses and time points. The concentration of inorganic arsenic measured in the
24    liver was similar to that measured in the kidney  at both dose levels, with peak concentrations
25    observed 1 hour after dosing. For the first 1 to 2 hours, inorganic arsenic was the predominant
26    form in both the liver and kidney, regardless of dose.  At the later times, DMA became the
27    predominant form. Kidney measurements 1 hour after dosing demonstrated that MMA levels
28    were 3 to 4 times higher than in other tissues.  DMA concentrations in the kidney reached their
29    peak 2 hours after dosing. DMA was the predominant form measured in the lungs at all time
30    points following exposure to 10 umole As/kg as Asv.  DMA concentrations in the lung were
31    greater than or equal to those of the other tissues beginning at four hours.  The study did not
32    distinguish the different valence states of the MMA or DMA compounds.
33          In a follow-up study by Kenyon et al. (2008), adult female C57B1/6 mice were
34    administered 0, 0.5, 2, 10, or 50 ppm of arsenic as sodium arsenate in the drinking water for
35    12 weeks.  The average daily intakes were estimated to be 0, 0.083, 0.35, 1.89, and 7.02 mg
36    As/kg/day, respectively.  After 12 weeks of exposure, the tissue distributions were as follows:
37    kidney > lung > urinary bladder > skin > blood > liver. In the kidney, MMA was the

                                                 12       DRAFT—DO NOT CITE OR QUOTE

-------
 1    predominant form measured, while DMA was more prominent in the lungs and blood.  The skin
 2    and urinary bladder had nearly equal levels of both inorganic arsenic and DMA and the liver had
 3    equal proportions of all three species.
 4          Naranmandura et al. (2007) characterized the tissue distribution in rats and hamsters
 5    administered a single oral dose of As111 (5.0 mg As/kg body weight, or BW).  In rats, the highest
 6    concentrations were found in RBCs. Because hamsters did not accumulate arsenic species in
 7    their RBCs, they exhibited a more uniform tissue distribution. While the quantity of arsenic in
 8    the liver and kidneys of the hamster were significantly greater than those observed in the rat,
 9    arsenic accumulated more and was retained longer in the kidneys than the liver in both species.
10    The hamster had greater levels of MMA111 bound to protein in the kidney than rats.
11          As111 and Asv, as well as methylated metabolites, cross the placenta at all stages of
12    gestation in mice, marmoset monkeys, and hamsters (Hanion and Perm,  1977; Lindgren et al.,
13    1984; Hood et al., 1987; Jin et al., 2006a), with tissue distribution of arsenic  similar between the
14    mother and the fetus in late gestation.  Jin et al. (2006a) found increased levels of inorganic
15    arsenic and DMA in the livers and brains of newborn mice from dams administered either As111
16    or Asv in their drinking water throughout gestation and lactation. The levels of total arsenic in
17    the mothers' livers increased in a dose-dependent manner and were greater than those observed
18    in the mothers' brains or in the newborns' brains or livers.  The levels of total arsenic in the
19    livers and brains of newborn mice, however, were greater than those  observed in  the mothers'
20    brains, suggesting easier passage through the placenta than through a mature blood-brain barrier.
21    Because the levels of inorganic arsenic in the newborn livers and brains were nearly identical, it
22    appears that there was no difficulty in  passing through an immature blood-brain barrier. In
23    addition, the nearly 2:1 ratio of DMA in the brains compared to the livers of newborns indicates
24    either a preferential distribution of DMA in the newborns' brains or an increased distribution of
25    inorganic arsenic to the brain that is subsequently metabolized.  The marmoset monkey (known
26    to not methylate arsenic) displayed somewhat less placental transfer after administration of As111
27    than was seen in mice (Lindgren et al., 1984).
28          The arsenic concentration in the cord blood (11 ug/L) was similar to that observed in
29    maternal blood (an average of 9 ug/L) in pregnant women living in a village  in northwestern
30    Argentina, where the arsenic concentration in the drinking water was approximately 200 ppb
31    (Concha et al., 1998a). Hall et al. (2007) also found a strong association between maternal (11.9
32    ug/L) and cord blood levels (15.7 ug/L)  in Matlab, Bangladesh (arsenic exposure ranged from
33    0.1 to 661 ppb in drinking water). They also measured arsenic metabolite levels and found that
34    the association also was observed for the metabolites MMA and DMA. Elevated arsenic
35    concentrations also were noted in pregnant women living in cities with low dust fall  (i.e., low
36    arsenic inhalation exposures), where an average of 3 ug/L was measured in the maternal blood
37    and 2 ug/L in cord blood (Kagey et al., 1977). Women living near smelters also have been

                                                 13       DRAFT—DO NOT CITE OR QUOTE

-------
 1    observed to have an increased concentration of placental arsenic (Tabacova et al., 1994).
 2    Although the human fetus is exposed to arsenic, it may be more in the form of DMA (at least in
 3    late gestation) because 90% or more of the arsenic in the urine and plasma of newborns and
 4    mothers (at time of delivery) was DMA.
      3.2.3. Cellular Uptake, Distribution, and Transport
 5          Cellular uptake of inorganic arsenic compounds also depends on oxidation state, with
 6    As111 generally being taken up at a much greater rate than arsenate (Cohen et al., 2006). In
 7    Chinese hamster ovary (CHO) cells, the rate of uptake was DMA111 > MMA111 > As111 (Dopp et
 8    al., 2004), with the pentavalent forms being taken up much more slowly than the trivalent forms.
 9    Delnomdedieu et al. (1995) demonstrated that As111 is taken up more readily than Asv, MMAV,
10    or DMAV by RBCs in rabbits. Drobna et al. (2005) found that MMA111 and DMA111 were taken
11    up by modified UROtsa cells expressing arsenic methyltransferase (this is a human urothelial
12    cell line that normally does not methylate inorganic arsenic) at an order of magnitude faster than
13    As111. Because arsenate uptake is inhibited in a dose-dependent manner by phosphate (Huang
14    and Lee, 1996), it has been suggested that a common transport system is responsible for the
15    cellular uptake for both compounds. As111 uptake, however, is not affected by phosphate;
16    therefore, Huang and Lee (1996) suggested that cellular uptake of As111 occurs through simple
17    diffusion. Liu et al. (2002, 2006a),  however, suggested that transport of As111 and MMA111 across
18    the cellular membrane may be mediated by AQP7 and AQP9 with MMA111  transported at a
19    higher rate. Lu et al. (2006) found that inorganic arsenic (both pentavalent and trivalent
20    oxidation states) can be transported by organic anion transporting polypeptide-C (OATP-C;
21    which was transfected into cells of a human embryonic kidney cell line), but not MMAV or
22    DMAV. In a cell line resistant to arsenic (R15), Lee et al. (2006a) found little AQP7 or AQP9
23    messenger RNA (mRNA) and only  half the AQP3 mRNA expression compared to the parental
24    cell line (CL3, a human lung adenocarcinoma cell line). Suppressing the AQP3 expression in
25    CL3 cells caused less arsenic to accumulate in these cells. Over-expression of AQP3 in a 293
26    cell line (a human embryonic kidney cell  line) resulted in an increase in arsenic accumulation in
27    the cells. Hexose permease transporters (HXT) also have been suggested as another influx
28    pathway for As111 (Thomas, 2007).
29          Shiobara et al.  (2001) demonstrated that the uptake of DMA in RBCs was dependent on
30    not only the chemical form (or oxidation state), but animal species. DMA111 and DMAV were
31    incubated with rat, hamster, mouse, and human RBCs. DMAV was only minimally absorbed by
32    RBCs, and the cellular uptake was very slow in all animal species tested. DMA111, on the other
33    hand, was efficiently taken up by the RBCs in the following order: rats > hamsters > humans.
34    Mouse RBCs were less efficient at the uptake of DMA111 than any of the other species. Rat
35    RBCs retained the DMA111 throughout the 4 hours of the experiment, but hamster RBCs were
36    found to excrete the arsenic absorbed as DMA111 in the form of DMAV. Human RBCs also
                                                14       DRAFT—DO NOT CITE OR QUOTE

-------
 1    excreted DMA111 as DMAV, though the rate of uptake of DMA111 and efflux of DMAV was much
 2    slower than in hamster RBCs.
 3          Cellular excretion of arsenic species also depends on oxidation state and the degree of
 4    methylation. Leslie et al. (2004), using membrane vesicles from a multi-drug resistant human
 5    lung cancer cell line (H69AR), found that a multi-drug resistance protein (MRP) called MRP1
 6    transports As111 in the presence of GSH but did not transport Asv under any conditions. This
 7    suggests that Asv must be reduced to As111 before being excreted from the cell.  Further, the
 8    MRP1 transport was more efficient with arsenic triglutathione (ATG) as the substrate.  This
 9    finding, along with the observation that As111 transport is more efficient at neutral or low pH
10    where ATG is more readily formed and more stable, suggests that ATG is formed prior to
11    transport. Leslie et al. (2004) also suggest that the formation of the conjugate is catalyzed by the
12    glutathione-S-transferase Pl-1 (GSTP1-1)  enzyme. MRP2 may also be involved in the efflux of
13    arsenic species from cells (Thomas, 2007). MRP2 expression was found to be five times higher
14    in arsenic-resistant (R15) cells compared to the parent cell line (CL3). However, expression
15    levels of MRP1 and MRP3 were similar to levels in parent cells (Lee et al., 2006a).  Suppressing
16    the multi-drug resistant transporters reduced  the efflux of arsenic from Rl 5 cells.
17          In a study of rabbits and mice exposed to radio-labeled arsenic (as As111), the majority of
18    the arsenic was found in  the nuclear and soluble fractions of liver, kidney, and lung cells
19    (Marafante et al., 1981; Marafante and Vahter, 1984).  The marmoset monkey had a different
20    intracellular distribution, with approximately 50% of the arsenic dose found in the microsomal
21    fraction in the liver (Vahter et al., 1982;  Vahter and Marafante, 1985). Chemical inhibition of
22    arsenic methylation in rabbits did not alter the intracellular distribution of arsenic (Marafante and
23    Vahter, 1984; Marafante et al., 1985).
24          Increases in tissue arsenic concentration (especially in the liver) have been found to be
25    associated with increased arsenic concentrations in the microsomal fraction of the liver in rabbits
26    fed diets containing low concentrations of methionine, choline, or proteins, which leads to
27    decreased arsenic methylation (Vahter and Marafante,  1987).  The levels of arsenic in the
28    microsomal fraction of the liver in these rabbits were similar to those observed in the marmoset
29    monkey (Vahter et al., 1982), indicating that nutritional factors may play a role in determining
30    the subcellular distribution of arsenic.

      3.3. METABOLISM
31          After entering the body, Asv can be reduced to As111, which can then proceed through a
32    series of methylation and conjugation reactions, some of which involve re-oxidation of arsenic to
33    Asv. The traditional metabolic pathways proposed for arsenic are shown in Figure 3-1. In this
34    metabolic scheme, less toxic species (i.e., Asv, MMAV, and DMAV) can be converted to more
35    toxic species (i.e.,  As111, MMAm, and DMA111). The trivalent species have been found to be more

                                                15       DRAFT—DO NOT CITE OR QUOTE

-------
 1    cytotoxic, genotoxic, and more potent inhibitors of enzyme activity (Thomas et al., 2001). While
 2    the final metabolite in humans is predominantly DMAV, as this is the form most highly excreted,
 3    some animal species further metabolize DMAV through DMA111 to trimethylarsine oxide
 4    (TMAO).
                    Arsenatev
                          \  <--
                         Arsenite111
Arsenate reductase, Glutathione S-transferase-(0,
Glyceraldehyde phosphate dehydrogenase
(GAPDH)?; Nonenzymatic pathway
                                  SAM
                          Monomethylarsonic acid (MMAV)
                                      X  •*	Glutathione S transferase-co
                              Monomethylarsonous acid (MMA111)
                                               VSAM
               	^___,                        SAH
               MethyltransferaseCASSMT)      Dimethylarsink add (DMAV)

                                                 \	J
                                           Dimethylarsenous acid (DMA111)
                                                         •SAM
                                               Trimethyl arsine oxide (TMAOV)
                                                            \  -
                                                       Trimethyl arsine (TMA111)
                   Source:  Sams et al. (2007).

                   Figure 3-1. Traditional metabolic pathway for inorganic arsenic in
                               humans.
 5          Hayakawa et al. (2005) suggested a possible alternate metabolic pathway for inorganic
 6    arsenic (Figure 3-2).  As in the previously described model, the first step involves reduction of
 7    Asv to As111.  A major difference, however,  is that Hayakawa et al. (2005) suggest that arsenic-
 8    glutathione complexes are important intermediates in the metabolism of arsenic and are the
 9    primary substrates for arsenic methyltransferases. The proposed model was based on the
10    observation that more DMAV is produced from As111 than from MMAV. This should not be the
11    case if the reactions depicted in Figure 3-1 are the primary arsenic metabolic pathways. Their
12    data suggest that arsenite, in the presence of GSH, non-enzymatically reacts to form ATG.  In
13    support of this mechanism, they observed a dose-dependent increase in concentration of ATG
14    with increasing doses of GSH, up to 4 mM.  Monomethyl and dimethyl arsenic species were
15    generated by the transfer of a methyl group from SAM in the presence of human recombinant
                                                 16
                  DRAFT—DO NOT CITE OR QUOTE

-------
1    arsenic (+3 oxidation state) methyltransferase (AS3MT), and only occurred when ATG or
2    monomethylarsonic diglutathione (MADG) was present. At concentrations of glutathione of
3    2.0 mM or greater, there was a dose-dependent increase in DMAV levels, accompanied by a
4    dose-dependent decrease in Asv.
                   o
                   ii
              OH - Asv - OH
                   I
                  OH
                Arsenate
                   O
                   II
              OH - Asv - OH
                   I
                  CH3
                MMA(V)
OH-As"1-OH
      I
     OH

   Arsenite
OH-As"1-OH
      I
    CH3

   MMA(III)
JGSH
JGSH
GS-As'"-SG
     I
    SG
                                                                     AsSMT
                                                                     SAM
       ATG
(Arsenic triglutathione)
GS-As1"
     I
    CH,
                  SG
                                                                      AsSMT
                                                                      SAM
        MADG

   (Monomethylarsinic
     diglutathione)
              OH-Asv-CH3
                   I
                  CH3
                 DMA(V)


      Hayakawa et al. 2005
OH-AsMI-CH3
      I
    CH3

   DMA(III)
         GS-Asm-CH3
               I
              CH3
                    DMAG

                (Dimethylarsinic
                 Glutathione)
  Arsenic 3 methyl transferase (As3MT); SAM -S-adenosyl methionine; GSH -Glutathione
                   Source: Hayakawa et al. (2005).
                   Figure 3-2. Alternative metabolic pathway for inorganic arsenic in
                               humans proposed by Hayakawa et al. (2005).
                                                  17
                      DRAFT—DO NOT CITE OR QUOTE

-------
 1          In summary, the proposed metabolic model of Hayakawa et al. (2005) suggests that Asv
 2    is first reduced to As111, which then reacts (non-enzymatically) with GSH (producing ATG). In
 3    the presence of AS3MT (specified as cyt!9 in the Hayakawa article),2 ATG is methylated to
 4    MADG if the GSH concentration is sufficient, which then comes to equilibrium with MMA111
 5    (GSH concentrations lower than 1 mM caused MADG to be unstable in solution and was readily
 6    hydrolyzed and oxidized to MMAV). While some of the MMA111 is oxidized to MMAV, some of
 7    the MADG is methylated by AS3MT to dimethylarsinic glutathione (DMAG), which, like
 8    MADG, is in equilibrium with its trivalent form and can be oxidized to its pentavalent form.
 9    This more recently proposed pathway leads to higher proportions of less toxic final species than
10    the original proposed metabolic pathway (Figure 3-1).
11          Results reported by Hughes et al. (2005) may provide support for the Hayakawa et al.
12    (2005) revised pathway.  B6C3F1 mice administered MMAV per os demonstrated its rapid
13    absorption,  distribution, and excretion,  with 80% of the dose eliminated within 8 hours.  Very
14    little of the  absorbed dose, however, was methylated to DMA and/or TMAO. Less than  10% of
15    the dose excreted in urine and 25% or less of the dose measured in the tissues were in the form
16    of DMA. In contrast, in MMA111-treated mice, more than 90% of the excreted dose and more
17    than 75% of the arsenic measured in the tissues was identified as DMA. This discrepancy
18    between the two forms of MMA is not  expected if the generally accepted metabolic pathway
19    (Figure 3-1) is followed. However, if MMA111 is the form methylated to DMA while MMAV is
20    an end product, as is suggested by Hayakawa et al.  (2005), then it would be expected that a
21    greater proportion of MMA111 would be methylated to DMA than MMAV.  There are, however,
22    factors that may limit the in vivo methylation of MMAV that are unrelated to the metabolic
23    pathway proposed by Hayakawa et al. (2005). First, MMAV does not appear to be taken up well
24    by the liver (Hughes et al., 2005), a major site of inorganic arsenic metabolism (Thomas  et al.,
25    2001). In fact, pentavalent species of arsenic are not taken up by cells as readily as trivalent
26    arsenicals (Dopp et al., 2004). In addition, in the generally accepted metabolic pathway  (Figure
27    3-1), MMAV needs to be reduced to MMA111 in order to be methylated.  Therefore, if very little is
28    taken up into cells, very little can be methylated.
29           Aposhian and Aposhian (2006) suggest that it is too early to accept AS3MT as the
30    primary methyltransferase responsible for arsenic methylation in humans because it has only
31    been observed in experiments involving deoxyribonucleic acid (DNA) recombinant technology
32    and because there is no indication that the enzyme is expressed in human liver. Although
33    AS3MT has been detected in human liver cell lines (Zakharyan et al., 1999), it has not been
      2 Arsenic (+3 oxidative state) methyltransferase (AS3MT) has been referred to by many investigators as cyt!9 in
      their references. According to Thomas et al. (2007), the Human Genome Nomenclature Committee
      (http://www.gene.ucl.ac.uk/cgi-bin/nomenclature/searchgenes.pl) recommends that this protein be systematically
      named AS3MT. In this document, references to cyt!9 it has been changed to AS3MT to avoid confusion and for
      uniform consistency.
                                                 18       DRAFT—DO NOT CITE OR QUOTE

-------
 1    isolated from surgically removed liver tissue.  Thomas et al. (2007) also states the evidence
 2    supports the conclusion that arsenic methylation catalyzed by AS3MT is not strictly dependent
 3    on the presence of GSH, which would suggest that other pathways may be involved in addition
 4    to those included in Hayakawa et al.'s (2005)  model.  GSH depletion would likely occur at high
 5    arsenic exposures under Hayakawa et al.'s proposed pathway.  Therefore, it is possible that both
 6    pathways work in conjunction, or one is predominant over the other depending on the
 7    concentration of arsenic. Hayakawa et al. (2005) found that levels of MMAV were not
 8    dependent on GSH level (from 2 to 5 mM), suggesting that this indicated possible further
 9    methylation to DMAV.  Since this is not part of the proposed Hayakawa et al. (2005) pathway, at
10    least some of the MMAV may be methylated through the classic pathway.
      3.3.1. Reduction
11          A substantial fraction of absorbed Asv is rapidly reduced to As111 in  most species studied;
12    in mice, rabbits, and marmoset monkeys, the reduction apparently occurs mainly in the blood
13    (Vahter and Envall, 1983; Vahter and Marafante, 1985; Marafante et al., 1985). Reduction also
14    may occur in the stomach or intestines prior to absorption, but quantitative  experimental  data are
15    not available to determine the importance of this GI reduction. In addition  to the reduction of
16    inorganic Asv, as shown in Figure 3-1,  methylated Asv species also may be reduced, apparently
17    by different enzymes.
18          GSH may play a role in the reduction of Asv, but apparently is not the only cofactor, as
19    cysteine and dithiothreitol (DTT) also have been found to reduce Asv to As111 in vitro (Zakharyan
20    et al., 1995; NRC, 1999; Nemeti and Gregus,  2002). Inorganic phosphate inhibits the formation
21    of As111 from Asv in intact RBCs (Nemeti and  Gregus, 2004), probably by competing with the
22    phosphate transporter for the uptake into cells.
23          Arsenate reductase enzymes have been detected in the human liver (Radabaugh and
24    Aposhian, 2000). At least one of these enzymes has been characterized as a purine nucleoside
25    phosphorylase (PNP) (Gregus and  Nemeti, 2002; Radabaugh et al., 2002).  This enzyme  requires
26    a thiol and a heat-stable cofactor for activation. According to Radabaugh et al. (2002),
27    dihydrolipoic acid (DHLP) is the most active  naturally occurring thiol in mammalian systems
28    and appears to be required for the enzymatic reduction of Asv to As111. PNP, however, did not
29    catalyze the reduction of MMAV to MMA111. An MMAV reductase  has been detected in rabbit
30    liver (Zakharyan and Aposhian, 1999),  hamster tissues (Sampayo-Reyes et  al., 2000), and human
31    liver (Zakharyan et al., 2001). In humans, this reductase is human glutathione-S-transferase co
32    (hGST-Ol), which is a member of the glutathione-S-transferase (GST) superfamily (Aposhian
33    and Aposhian, 2006).
34          Although PNP has been determined to reduce  Asv to As111, Nemeti et al. (2003) observed
35    this reduction only in vitro. PNP did not appear to be a major player in the  reduction of Asv to
36    As111 in either human erythrocytes or in rats in vivo. Nemeti and Gregus (2004, 2005) further
                                                19       DRAFT—DO NOT CITE OR QUOTE

-------
 1    demonstrated that human erythrocytes exhibit a PNP-independent Asv-reducing pathway that
 2    requires GSH, nicotinamide adenine dinucleotide (NAD), and a substrate for either one or both
 3    of the following enzymes: glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or
 4    phosphoglycerate kinase (PGK). This mechanism of reduction also was demonstrated in rat liver
 5    cytosol (Nemeti and Gregus, 2005). In addition, another unidentified enzyme in the liver cytosol
 6    had the capacity to reduce Asv. A further study (Gregus and Nemeti, 2005) demonstrated that
 7    GAPDH exhibited Asv reductase activity, but that PGK served as an auxiliary enzyme when
 8    3-phosphoglycerate was the glycolic substrate.
 9          The reduction of pentavalent arsenicals also has been observed to be catalyzed by
10    AS3MT (Waters et al., 2004a).  According to Waters et al. (2004b), AS3MT may possess both
11    As111 methyltransferase and Asv reductase activities. In the presence of an exogenous or
12    physiological reductant, AS3MT was found to catalyze the entire sequence converting arsenite to
13    all of its methylated metabolites through both methylation and reduction steps (Figure 3-1).
14    Thomas et al. (2007) also suggest that thioredoxin (Trx, isolated from E. coli) is necessary,
15    possibly reducing some critical cysteine residue in AS3MT as a step in the methyltransferase
16    reaction.  Cohen et al. (2006) suggest that Trx, thioredoxin reductase (TrxR), and nicotinamide
17    adenine dinucleotide phosphate-oxidase (NADPH) are the primary reducing agents involved in
18    the conversion of MMAV to DMAV, but they are orders of magnitude less effective than the
19    arsenic methyltransferase isolated from rabbit liver (i.e., AS3MT). Zakharyan and Aposhian
20    (1999) found that MMAv-reductase was the rate-limiting enzyme in arsenic biotransformation in
21    rabbit livers. Jin et al. (2006a) also suggest that Asv reduction is possibly a rate-limiting step in
22    arsenic metabolism at low concentrations.  At higher concentrations, saturation or methylation
23    inhibition may cause other reactions to become rate-limiting.
      3.3.2. Arsenic Methylation
24          Methylation is an important factor affecting arsenic tissue distribution and excretion.
25    Humans and most experimental animal models methylate inorganic arsenic to MMA and DMA,
26    with the amounts differing across species, as  determined by analysis of urinary metabolites. The
27    methylated metabolites in  and of themselves  have historically been considered less acutely toxic,
28    less reactive  with tissue constituents, less cytotoxic, and more readily excreted in  the urine than
29    inorganic arsenic (Vahter and Marafante, 1983; Vahter et al., 1984; Yamauchi and Yamamura,
30    1984; Marafante et al., 1987; Moore et al.,  1997a; Rasmussen and Menzel, 1997;  Hughes and
31    Kenyon, 1998; Sakurai et al., 1998). The trivalent species MMA111 and DMA111, however, have
32    been demonstrated to be more cytotoxic in a  human liver cell line called Chang cells (Petrick  et
33    al., 2000,  2001), CHO (Dopp et al., 2004),  and cultured primary rat hepatocytes (Styblo et al.,
34    1999a, 2000) than As111, Asv, MMAV, or DMAV.
35          Although the kinetics of arsenic methylation in vivo are not fully understood, it is
36    believed the  liver may be the primary site of  arsenic methylation.  However, the testes,  kidney,
                                                20        DRAFT—DO NOT CITE OR QUOTE

-------
 1    and lung also have been observed to have a high methylating capacity (Cohen et al., 2006).
 2    Marafante et al. (1985) found that DMA appeared in the liver prior to any other tissue in rabbits
 3    exposed to inorganic As. It also has been demonstrated oral administration of inorganic arsenic
 4    favors methylation more than either subcutaneous or intravenous administration (Charbonneau et
 5    al., 1979; Vahter,  1981; Buchet et al., 1984), presumably because the arsenic will pass through
 6    the liver first after oral administration.  However, liver disease (i.e., alcoholic, post-necrotic or
 7    biliary cirrhosis, chronic hepatitis, hemochromatosis, and steatosis) can be associated with
 8    increased ratios of DMA to MMA in the urine following a single injection of sodium arsenite
 9    (Buchet et al., 1984; Geubel et al., 1988). This appears to indicate that efficient methylation of
10    arsenic continues in the presence of liver damage, possibly indicating that a different organ is
11    responsible for methylation under these circumstances.  In addition, the site of methylation may
12    depend on the rate of reduction of Asv to As111. Isolated rat hepatocytes readily  absorbed and
13    methylated As111, but not Asv (Lerman et al., 1983).  Kidney slices, on the other hand, produced
14    five times more DMA from Asv than As111 (Lerman and Clarkson, 1983).  Therefore, it is likely
15    that any Asv not initially reduced can be efficiently methylated in the kidney for subsequent
16    urinary excretion.
17           Identifying the main organs responsible for methylation of arsenic in vivo has not been
18    straightforward because in vitro results do not necessarily reflect in vivo methylation patterns
19    (NRC,  1999). Buchet and Lauwerys (1985) identified the rat liver as the main organ for
20    methylation, with  the methylating capacities in the RBCs, brain, lung, intestine, and kidneys
21    being insignificant in comparison.  Assays of arsenite methyltransferases from mouse tissues
22    demonstrated the testes had the highest methylating  activity, followed by the kidney, lung, and
23    liver (Healy et al., 1998).  Aposhian (1997) determined that the amount of methyltransferases
24    vary in the liver of different animal species.  Arsenite bound to components of tissue can be
25    methylated and released (Marafante et al., 1981; Vahter and Marafante, 1983).  This may explain
26    the initial rapid phase (immediate methylation and excretion) followed by a slow elimination
27    phase (continuous release of bound arsenite through methylation) (NRC,  1999), as described in
28    Section 3.4.
29           It  has been demonstrated that inhibition of arsenic methylation results in increased tissue
30    concentrations of arsenic (Marafante and Vahter, 1984; Marafante et al., 1985).  Loffredo et  al.
31    (2003)  suggest that the second methylation step is inducible and that the inducibility is possibly
32    polymorphic (i.e., more than one enzyme or enzyme form may be involved, depending on the
33    individual). This suggestion is based on observations that human urinary DMA concentrations
34    in high-exposure groups were higher and more variable than urinary MMA levels, and because
35    urinary DMA levels appeared to have a bimodal distribution in a population from Mexico,
36    regardless of exposure status. Others have suggested that the second methylation step may be
37    saturable, which would be consistent with the decreasing excretion of DMA with increasing

                                                 21        DRAFT—DO NOT CITE OR QUOTE

-------
 1    arsenic exposures (Ahsan et al., 2007).  Cysteine, GSH, and DTT have been shown to increase
 2    the activity of arsenite methyltransferase and MMA methyltransferase (both later identified as
 3    AS3MT; Lin et al., 2002) in purified rabbit liver enzyme preparations (Zakharyan et al., 1995).
 4    Dithiols (e.g., reduced lipoic acid) have also been found to enhance arsenite methylation by
 5    MMA111 methyltransferase (Zakharyan et al., 1999). Glutathione-S-transferase omega 1
 6    (GSTO1) has also been associated with arsenic biotransformation (Meza et al., 2007). Although
 7    humans have been observed to methylate arsenic, no arsenic methyltransferase has yet been
 8    isolated from human tissues (Aposhian and Aposhian, 2006).
 9           In vitro studies using rat liver preparations indicate that the methylating activity is
10    localized in the cytosol, with SAM being the main methyl donor for As111 methylation (Marafante
11    and Vahter, 1984; Buchet and Lauwerys, 1985; Marafante et al., 1985; Styblo et  al., 1995,  1996;
12    Zakharyan et al., 1995).  AS3MT catalyzes the transfer of the methyl group from SAM to the
13    arsenic substrates (Lin et al., 2002; Thomas, 2007).  Expressing AS3MT in UROtsa (human
14    urothelial cells that do not normally methylate inorganic arsenic) caused the cells to effectively
15    methylate arsenite (Drobna et al., 2005). High concentrations of As111 or MMA111 in the culture
16    caused an inhibition in the formation of DMA, but had little effect on the formation of MMA.
17    The inhibition of DMA production resulted in MMA accumulation in cells.  Drobna et al. (2006)
18    demonstrated that AS3MT was the major enzyme for arsenic methylation in human
19    hepatocellular carcinoma (HepG2) cells, but reducing  it by 88% (protein levels) only accounted
20    for a 70% reduction in methylation capacity, suggesting that there is another methylation process
21    that is independent of AS3MT.
22           The addition of GSH has been found to increase the yield of mono- and dimethylated
23    arsenicals but suppressed the production of TMAO in  the presence of rat AS3MT (Waters et al.,
24    2004a), indicating that GSH suppresses the third methylation reaction but not the first two
25    (Thomas et al., 2007). Thomas et al. (2004) discovered a similar arsenic methyltransferase in the
26    rat liver, which they designated cyt!9 because an orthologous cyt!9 gene encodes an arsenic
27    methyltransferase in the mouse and human genome. It has subsequently been concluded that this
28    methyltransferase was the same as AS3MT.
29           GSH alone does not support recombinant rat AS3MT catalytic function, but when added
30    to a reaction mixture containing other reductants, the rate of arsenic methylation  increases
31    (Waters et al., 2004b). GSH alone (5mM) does not support the catalytic activity  of AS3MT, but
32    stimulates the methylation rate in the presence of the reductant tris(2-carboxylethyl)phosphine
33    (TCEP; 1  mM) (Thomas et al., 2007).  GSH (5 mM) did not have any effect on DTT (1 mM)-
34    induced arsenic methylation. Drobna et al. (2004) linked the genetic polymorphism of AS3MT
35    with other cellular factors and to the inter-individual variability in the capacity of primary human
36    hepatocytes to retain and metabolize As111 (see Section 4.7).
                                                22        DRAFT—DO NOT CITE OR QUOTE

-------
 1          The main products of arsenic methylation in humans are MMAV and DMAV, which are
 2    readily excreted in the urine (Marcus and Rispin, 1988).  MMA111 and DMA111 have recently been
 3    detected in human urine (NRC, 2001); however, most studies do not differentiate the valence
 4    state of mono- or dimethylated arsenic species detected in urine or tissue samples. Le et al.
 5    (2000a,b) and Del  Razo et al. (2001) noted that the concentration of trivalent metabolites in the
 6    urine may be underestimated because they are easily oxidized after collection. Le et al. (2000b)
 7    found 43 to 227 ug/L of MMA111 in the urine of populations from Inner Mongolia, China, who
 8    were exposed to 510-660 ppb (0.46 uM) of arsenic via the drinking water.
 9          A small percent of DMA111 may further be methylated to TMAO in mice and hamsters
10    (see Kenyon and Hughes, 2001, for a review). A single human volunteer ingesting DMA
11    excreted 3.5% of the dose as TMAO (Kenyon and Hughes, 2001).  TMAO can be detected in
12    urine following DMA exposure, but has not been detected in the blood or tissues of mice
13    exposed intravenously to DMA (Hughes et al., 2000) or in the urine of mammals orally exposed
14    to inorganic As. This may be due to rapid clearance of DMA and MMA from cells (Styblo et al.,
15    1999b); however, most analytical methods are not optimized for the detection of TMAO that
16    could have been present but not detected.
      3.3.3. Species Differences in the Methylation of Arsenic
17          There is considerable variation in the patterns of inorganic arsenic methylation among
18    mammalian species (NRC, 1999). Humans, rats, mice, dogs, rabbits, and hamsters have been
19    shown to efficiently methylate inorganic arsenic to MMA and/or DMA.  Rats and hamsters
20    appear to methylate administered DMA into TMAO more efficiently than other species (NRC,
21    1999; Yamauchi and Yamamura,  1984). About 40% of urinary arsenic was present as TMAO 1
22    week after exposure to DMA in the drinking water, while 24% was present as TMAO  after 7
23    months of exposure (100 mg/L) in male rats (Yoshida et al.,  1998).
24          Humans (mainly exposed  to background levels or exposed at work) have been  estimated
25    through a number of studies to excrete 10% to 30% of the arsenic in its inorganic form, 10% to
26    20% as MMA, and 55% to 75% as DMA (see Vahter,  1999a, for a review).  In contrast, a study
27    of urinary arsenic metabolites in a population from northern  Argentina exposed to arsenic via
28    drinking water demonstrated an average of only 2% MMA in the urine (Vahter et al., 1995b;
29    Concha et al., 1998b). This may indicate variations in methylation activity depending on the
30    route of exposure,  level of exposure, and possible nutritional or genetic factors.  Although
31    humans are considered efficient at arsenic methylation, they are less efficient than many animal
32    models, as indicated by the larger proportion of MMAV excreted in the urine (Vahter,  1999a).
33    This is important because it may explain why humans are more susceptible to cancer from
34    arsenic exposures, and why no adult animal model for  inorganic-arsenic-induced cancers has yet
35    been identified (Tseng et al., 2005).

                                               23        DRAFT—DO NOT CITE OR QUOTE

-------
 1          The rabbit (Marafante et al., 1981; Vahter and Marafante, 1983; Maiorino and Aposhian,
 2    1985) and hamster (Charbonneau et al., 1980; Yamauchi and Yamamura, 1984; Marafante and
 3    Vahter, 1987) appear to be more comparable to humans with respect to arsenic methylation than
 4    other experimental animals (NRC,  1999). However, rabbits and hamsters, in general, excrete
 5    more DMA and less MMA than humans.  In contrast, Flemish giant rabbits (De Kimpe et al.,
 6    1996) excrete MMA in amounts similar to humans. Mice and dogs, efficient methylators of
 7    arsenic, excrete more than 80% of a single arsenic dose administered as DMA within a few days
 8    (Charbonneau et al., 1979; Vahter,  1981). Guinea pigs (Healy et al., 1997), marmoset monkeys
 9    (Vahter et al., 1982; Vahter and Marafante, 1985), and  chimpanzees (Vahter et al., 1995a), on
10    the other hand, do not appear to appreciably methylate  inorganic arsenic. In addition, no
11    methyltransferase activity was detected in these species (Zakharyan et al., 1995, 1996; Healy et
12    al., 1997; Vahter, 1999a). Li et al.  (2005) identified a frameshift mutation in the chimpanzee
13    AS3MT gene that resulted in the production of an inactive truncated protein, possibly explaining
14    the lack of methylation activity in that species.
15          AS3MT homolog proteins with five fully conserved cysteine residues have been
16    observed in the genome of numerous species (Thomas et al., 2007). Chimpanzees were found to
17    differ from other  species  studied in that their AS3MT protein was shorter and lacked the 5th
18    cysteine (Thomas et al., 2007). Healy et al. (1999) identified marked variations in the activity of
19    methyltransferases, while Vahter (1999b) characterized differences in methylation efficiency
20    among different human populations.  The observed variations in methyltransferase activity and
21    methylation efficiency are probably the underlying reason for the cross-species variability in
22    methylation ability, as all the species had ample arsenate reductase  activity (Vahter, 1999a;
23    NRC, 2001).
24          Although  arsenic methylation is generally believed to take place in order to enhance
25    excretion, there are several species (guinea pigs, marmoset monkeys, and chimpanzees) that do
26    not methylate arsenic, but still  efficiently excrete it. In fact, these animals do not retain arsenic
27    any longer than species that methylate arsenic (Cohen et al., 2006), indicating that factors other
28    than methylation  also affect arsenic excretion rates. Supporting this is the fact that inorganic
29    arsenic is found in the urine of even the most efficient methylators (Vahter, 1994).
      3.3.4. Thioarsenical Metabolites
30          In 2004, Hansen et al. reported the detection of  unusual arsenic-containing metabolites in
31    the urine of sheep exposed to arsenic-contaminated vegetation. The metabolite  was tentatively
32    identified as dimethylmonothioarsinic acid (DMMTA111), a sulfur-containing derivative of
33    DMA111 as shown in Figure 3-3. Because the exposed sheep consumed algae known to contain
34    arsenosugars, some of which contain sulfur, the relevance of this finding to human exposures
35    was not initially clear.  Subsequently, Raml et al. (2006) detected the presence of DMMTA111 in

                                                 24        DRAFT—DO NOT CITE OR QUOTE

-------
 1   the urine of Japanese men, but again, consumption of arsenosugars was suspected as a source of
 2   the observed arsenic containing species.
                             SH-Asm-CH3        OH-Asv-CH3
                                     I                        I
                                    CH3                   CH3
                                DMMTA111               DMMTAV

                   Source:  Hansen et al. (2004).

                   Figure 3-3. Thioarsenical structures.
 3          In experiments addressing this issue, Adair et al. (2007) and Naramandura et al. (2007)
 4    found substantial concentrations of thioarsenical metabolites in arsenic-exposed experimental
 5    animals. Adair et al. (2007) administered drinking water containing 100 ppm Asv or up to 200
 6    ppm DMA111 to female Fisher 344 rats for 14 days. During analysis of the urine (collected during
 7    the last 24 hours of exposure) for metabolites, they found high levels of DMMTA111 and
 8    trimethylarsine sulfide (another sulfur-containing metabolite) in the urine of rats treated with
 9    DMA111.  Lower levels of the sulfur-containing metabolites were detected in the urine of
10    arsenate-treated animals.  They proposed a mechanism whereby the reaction of DMA111 and
11    DMAV with hydrogen sulfide resulted in the observed metabolites.
12          Naranmandura et al. (2007) administered single doses of 5.0 mg/kg As111 to Syrian
13    hamsters and Wistar rats by gavage and measured the levels of sulfur-containing arsenic
14    metabolites in urine. Both DMMTA111 and dimethylmonothioarsonic acid (DMMTAV) were
15    found at appreciable levels in urine from hamsters, but only the latter metabolite was found in rat
16    urine. A previously uncharacterized  metabolite, monomethylmonothioarsonic acid, was also
17    found in urine from both species.
18          These studies suggest that the generation of sulfur-containing arsenic metabolites does
19    not depend on exposures to arsenosugars, at least in rodents, but can occur during the
20    metabolism of inorganic arsenic compounds.  In 2007, Raml et al. presented evidence that this
21    pathway was also significant in humans. DMMTA111 was detected in the urine of 44% (33 of 75)
22    women exposed to inorganic arsenic-contaminated drinking water in Bangladesh.  The
23    metabolite was present in urine samples at concentrations between "trace" amounts and 24 ug/L,
24    with total arsenic concentrations ranging from 8 to 1034 ug/L. It was suggested that

                                               25       DRAFT—DO NOT CITE OR QUOTE

-------
 1    thioarsenical metabolites may have been present in urine from other epidemiological studies of
 2    arsenic-exposed populations, but may have not been detected due to analytical difficulties.

      3.4. ELIMINATION
 3          The major route of excretion for most arsenic compounds by humans is via the urine
 4    (Yamauchi and Yamamura 1979; Tarn et al., 1979; Pomroy et al., 1980; Buchet et al.,  1981). Six
 5    human subjects who ingested 0.01 ug of radio-labeled 74Asv excreted an average of 38% of the
 6    administered dose in the urine within 48  hours and 58% within 5 days (Tarn et al., 1979).
 7    Inorganic arsenic elimination in humans  has been observed to be triphasic, with first-order half-
 8    lives for elimination of 1  hour, 30 hours, and 200 hours (Mealey et al., 1959 used As111; Pomroy
 9    etal., 1980 used Asv).
10          As mentioned in the preceding section, MMA and DMA are important metabolites
11    generated after exposure to inorganic As. These methylated metabolites are excreted  in the
12    urine faster than the inorganic As. In humans orally exposed to MMA or DMA in aqueous
13    solution, about 78% of MMA and 75% of DMA were excreted in the urine within 4 days of
14    ingestion (Buchet et al., 1981).  In mice,  the half-time of MMA and DMA excretion was found
15    to be about 2 hours  following iv administration (Hughes and Kenyon, 1998).
16          Kenyon et al. (2008) administered 0, 0.5, 2, 10, or 50 ppm of arsenic as sodium arsenate
17    to adult C57B1/6 female mice in the drinking water for 12 weeks. The average daily intakes
18    were estimated to be 0, 0.083, 0.35, 1.89, and 7.02 mg As/kg/day, respectively. Levels of
19    MMA111, DMA111, DMAV, and TMAO in  the urine collected at the end of treatment increased in a
20    linear manner with dose, but Asv and MMAV did not.
21          Rats excrete DMA slowly compared to other species (Vahter et al., 1984), even though
22    they are efficient at methylating inorganic arsenic to DMA. The slow excretion is believed to be
23    associated with retention  of a significant portion of the DMA in erythrocytes (Odanaka et al.,
24    1980; Lerman and Clarkson, 1983; Vahter, 1983; Vahter et al., 1984).  The biliary excretion of
25    inorganic arsenic by rats is  about 800 times greater than observed in dogs and 37 times that of
26    rabbits, as proportion of administered dose.  Hughes et al. (2005) found that in mice the level of
27    MMAV excreted in  the urine compared to the bile was related to dose, with fecal excretion
28    increasing at higher doses.  Cui et al. (2004a) also found that rat biliary excretion rates varied
29    with dose, but found it was also related to route of administration and chemical form.  After oral
30    administration of inorganic arsenic (either form) to male Sprague-Dawley rats, MADG and
31    DMAV (likely present due to dissociation of DMAG) were the predominant forms in the bile.
32    MADG was found at a higher level after a higher (i.e., 100 ppm) dose, while DMAV was more
33    prevalent at the lower dose  (i.e., 10 ppm). Kala et al. (2000) found that the secretion of arsenic
34    into the bile of rats was dependent on the multi-drug resistance-associated protein 2 transporter
                                                26        DRAFT—DO NOT CITE OR QUOTE

-------
 1    (MPR2/cMOAT) and that GSH is necessary for the transport, as arsenic-glutathione complexes
 2    accounted for the majority of arsenic found in the bile.
 3          Although absorbed arsenic is removed from the body mainly via the urine, small amounts
 4    of arsenic are excreted through other routes (e.g., skin, sweat, hair, breast milk). While arsenic
 5    has been detected at low levels in the breast milk of women in northwestern Argentina (i.e., 2
 6    ug/kg), breastfeeding was associated with lower concentrations of arsenic in the urine of
 7    newborn children (Concha et al., 1998c) than formula feeding, owing to the use of arsenic
 8    contaminated water in formula preparation. Parr et al. (1991) measured arsenic (as well as other
 9    elements) in the breast milk from three groups of mothers from four countries (Guatemala,
10    Hungary, Nigeria, and the Philippines), and one to two groups from Sweden and Zaire.  The
11    breast milk was collected 3 months after birth. Levels of arsenic in the breast milk from women
12    in the Philippines were higher than other regions with levels about 19 ug/kg. Women from
13    Nigeria had levels similar to those observed by Concha et al. (1998c). Women from all the other
14    areas measured had levels of 0.24 to 0.55 ug/kg.
15          The  average concentration of arsenic in sweat induced in a hot and humid environment
16    was 1.5 ug/L, with an hourly loss rate of 2.1 ug (Vellar,  1969). Based on an average arsenic
17    concentration in the skin of 0.18 mg/kg, Molin and Wester (1976) estimated that the daily loss of
18    arsenic through desquamation was 0.1 to 0.2 ug in males with no known exposure to arsenic.

      3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
19          Physiologically based pharmacokinetic (PBPK) models for inorganic arsenic are
20    important for developing a biologically based dose-response (BBDR) model. The development
21    of useful BBDR models has  proved to be challenging because inorganic arsenic appears to
22    mediate its toxicity through a range of metabolites, and their roles with regard to specific adverse
23    effects are not clear (Clewell et al., 2007).
24          A PBPK model for exposure to inorganic arsenic (orally, intravenously, and
25    intratracheally) was developed in hamsters and rabbits by Mann et al. (1996a). The model
26    includes tissue compartments for lung (nasopharynx, tracheobronchial, pulmonary), plasma,
27    RBCs, liver, GI tract, skin, kidney, keratin, and combined other tissues.  Oral absorption of As111,
28    Asv, and DMA (pooled In and V oxidation states) was modeled as a first-order transport process
29    directly from the GI contents into the liver. Distribution to tissues was diffusion-limited, with
30    transfer rates estimated based upon literature values for capillary thickness and pore sizes for
31    each tissue.  Reductive metabolism of Asv to As111 was modeled as a first-order process occurring
32    in the plasma. Oxidative metabolism of As111 to Asv was modeled as first-order processes in the
33    plasma and  kidneys. Methylation of inorganic arsenic species to MMA (pooled In and V
34    oxidation states) and then to  DMA were modeled as saturable Michaelis-Menten processes
35    taking place in the liver. Urinary, biliary, and fecal excretion of As111, Asv, MMA, and DMA

                                                27       DRAFT—DO NOT CITE OR QUOTE

-------
 1    also are modeled as first-order processes. Parameters for absorption, tissue partition,
 2    metabolism, and biliary excretion were estimated by fitting the model to literature data on the
 3    urinary and fecal excretion of total arsenic from rabbits and hamsters administered various
 4    arsenic compounds by iv, oral gavage, or intratracheal instillation (Charbonneau et al., 1980;
 5    Yamauchi and Yamamura, 1984; Marafante et al., 1985, 1987). The model was found to
 6    accurately simulate the excretion of arsenic metabolites in the urine of rabbits and hamsters and
 7    to produce reasonable fits to liver, kidney, and skin concentrations in rabbits and hamsters
 8    (Yamauchi and Yamamura, 1984; Marafante et al., 1985; Marafante and Vahter, 1987).
 9           Mann et al. (1996b) extended their PBPK model for use in humans by adjusting
10    physiological parameters (organ weights, blood flows) and re-estimating absorption and
11    metabolic rate constants. The model was fit to literature data on the urinary excretion of total
12    arsenic following a single oral dose of As111 or Asv in human volunteers (Tarn et al., 1979;
13    Buchet et al., 1981).  The extended human model was further tested against empirical data on the
14    urinary excretion of the different metabolites of inorganic arsenic following oral intake of As111,
15    intake of inorganic arsenic via drinking water, and occupational exposure to arsenic trioxide
16    (ATO) (Harrington et al., 1978; Valentine et al., 1979; Buchet et al., 1981; Vahter et al., 1986).
17    The model predicted a slight decrease (about 10%) in the percentage of DMA in urine with
18    increasing single-dose exposure (highest dose of arsenic at 15 ug/kg of body weight), especially
19    following exposure to As111, and an almost corresponding increase in the percentage of MMA.
20    The model predicted that adults' drinking water containing 50 ppb would excrete more arsenic in
21    urine than an occupational  inhalation exposure of 10 ug/m3 (Mann et al., 1996b).
22           Yu (1999a,b) also developed a PBPK model for arsenic in humans that includes tissue
23    compartments for lung,  skin, fat, muscle, combined kidney and richly perfused tissues, liver,
24    intestine, GI and stomach contents, and bile.  Oral absorption of As111, Asv, and DMA (pooledIn
25    and v oxidation  states) was modeled as first-order transport from the GI contents into the
26    intestinal tissue. Distribution to tissues was modeled as perfusion-limited. Reductive
27    metabolism of Asv to As111  was modeled as a first-order, GSH-dependent process taking place in
28    the intestinal tissue, skin, liver, and kidney/rich tissues. Oxidative metabolism of As111 to Asv
29    was not modeled. Methylation of inorganic arsenic species to MMA (pooled In and v oxidation
30    states) and then  to DMA was modeled as saturable Michaelis-Menten processes occurring in the
31    liver and kidney. Urinary,  biliary, and fecal excretion of As111, Asv, MMA, and DMA were
32    modeled as first-order processes.  Parameters for absorption, tissue partition, metabolism, and
33    biliary excretion were estimated by fitting the model to literature data on tissue concentrations of
34    total arsenic from a fatal human poisoning (Saady et al., 1989), and blood, urine, and fecal
35    elimination of total arsenic following oral administration (Odanaka et al., 1980; Pomroy et al.,
36    1980).  The model was not tested further against external data, and fits to the data sets used for
37    parameter estimation were  not provided.

                                                28       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Gentry et al. (2004) adapted the model proposed by Mann et al. (1996a) to different
 2    mouse strains by adjusting physiological parameters (organ weights and perfusion rates). The
 3    absorption, partition, and metabolic rate constants were re-estimated by fitting the model to
 4    literature data on urinary excretion of various arsenic species following iv administration of
 5    MMA to B6C3F1 mice (Hughes and Kenyon, 1998) or single oral administration of As111 or Asv
 6    to mice (Kenyon et al., 1997; Hughes et al., 1999). Additionally, the description of methylation
 7    in the model was refined to include the uncompetitive inhibition of the conversion of MMA to
 8    DMA by As111.  The PBPK model was then validated using data from a single oral administration
 9    of Asv (Hughes et al., 1999) and a 26-week drinking water exposure of As111 to C57Black mice
10    (Moser et al., 2000). These data were found to adequately fit the model without further
11    parameter adjustment. Ng et al. (1999) had found arsenic-induced tumors in C57B1/6J mice,
12    while numerous other mouse strains (Swiss CR:Nffl[S], C57Bl/6p53 [+/-], C57Bl/6p53 [+/+], and
13    Swiss CD-I) had not experienced a significant increase in arsenic-induced tumors.  The Gentry
14    et al. (2004) model was unable to explain the different outcomes in the mouse bioassay on the
15    basis of predicted target organ doses.
16          The Mann et al. (1996a,b) and Gentry et al. (2004) models are well documented, were
17    validated against external data, and appear to capture the salient features of arsenic
18    toxicokinetics in rodents and humans.  The information provided by these models may help
19    explain the MO As involved in carcinogenesis along with possible reasons that humans are
20    apparently more susceptible to the carcinogenic effects of arsenic.
21          Clewell et al. (2007) noted that the then-available PBPK models did not incorporate the
22    most recent available information on arsenic methylation kinetics and suggested several steps for
23    improving the PBPK models.  El-Masri and Kenyon (2008) have developed a PBPK model
24    incorporating some of the improvements suggested by Clewell et al. (2007) (although not the
25    simulation of changes in gene expression).  The model predicts the levels of inorganic arsenic
26    and its metabolites in human tissues and urine following oral exposure of Asv, As111, and for oral
27    exposure to organoarsenical pesticides.  The model consists of interconnecting submodels for
28    inorganic arsenic (As111 and Asv), MMAV, and DMAV.  Reduction of MMAV and DMAV to their
29    trivalent forms  is also modeled. The submodels include the GI tract (lumen and tissue), lung,
30    liver, kidney, muscle, skin, heart, and brain, with reduction of MMAV and DMAV to their
31    trivalent forms  modeled as occurring in the lung, liver, and kidney.  The model also incorporates
32    the inhibitory effects of As111 on the methylation of MMA111 to DMA and MMA111 on the
33    methylation of As111 to MMA into consideration, modeled as noncompetitive inhibition. This
34    model differs from the other models described above because it provides an updated description
35    of metabolism using recent biochemical data on the mechanism of arsenic methylation.  In
36    addition, it uses in vitro studies to estimate most of the model parameters (statistically
37    optimizing those that are sensitive to urinary excretion levels to avoid problems with parameter

                                                29       DRAFT—DO NOT CITE OR QUOTE

-------
1   identifiability), and can predict the formation and excretion of trivalent methylated arsenicals.
2   The partition coefficients estimated in the model are comparable to those developed by Yu
3   (1999a). The performance of the model was tested against limited human data on urinary
4   excretion; the model needs to be evaluated for its ability to predict the tissue and urinary
5   concentrations of arsenicals in large numbers of subjects.  This model is  an improvement over
6   previous models because it can quantitatively assess impacts of parameter variability arising
7   from genetic polymorphism.
                                               30       DRAFT—DO NOT CITE OR QUOTE

-------
                                  4. HAZARD IDENTIFICATION

      4.1. STUDIES IN HUMANS
 1          Numerous epidemiologic investigations have examined the association between
 2    waterborne arsenic exposure and cancer outcome.  These epidemiologic investigations used
 3    many different study designs, each with their inherent limitations.  Regardless of the study type,
 4    the majority of these investigations found some level of association between arsenic exposure
 5    and cancer outcome. This association is not new, since arsenic exposure has been linked with
 6    cancer as far back as 1887 when Hutchinson reported an unusual number of skin tumors in
 7    patients treated with arsenicals. Since 1887, the association between skin cancer and arsenic has
 8    been reported in a number of studies (Tseng et al.,  1968; Tseng, 1977; Chen et al., 1985,
 9    1988a,b; Wu et al., 1989; Hinwood et al.,  1999; NRC, 1999; Tsai et al.,  1999; Karagas et al.,
10    2001; Knobeloch et al., 2006; Lamm  et al., 2007).
11          The SAB Arsenic Review Panel provided comments on key scientific issues associated
12    with arsenicals on cancer risk estimation in July 2007 (SAB, 2007). It was concluded that the
13    Taiwanese database is still the most appropriate source for estimating bladder and lung cancer
14    risk among humans (specifics provided in Section 5) because of: (1) the size and statistical
15    stability of the database relative to other studies; (2) the reliability of the population and
16    mortality counts; (3) the stability of residential patterns; and (4) the inclusion of long-term
17    exposures. However, SAB also noted considerable limitations within this data set (EPA-SAB-
18    07-008, http://www.epa.gov/sab). The Panel suggested that one way to mitigate the limitations
19    of the Taiwanese database would be to include other relevant epidemiological studies from
20    various countries. For example, SAB referenced other databases that contained studies of
21    populations also exposed to high levels of arsenic (e.g., Argentina and Chile), and recommended
22    that these alternate sources of data be used to compare the unit risks at the higher exposure levels
23    that have emerged from the Taiwan data.  SAB also suggested that, along with the Taiwan data,
24    published epidemiology studies from the United States and other countries where the population
25    is chronically  exposed to low levels of arsenic in drinking water (0.5 to 160 ppb) be critically
26    evaluated, using a uniform set of criteria presented in a narrative and tabular format. The
27    relative strengths and weaknesses of each study should be described in relation to each criterion.
28    The caveats and assumptions used should be presented so  that they are apparent to anyone who
29    uses these data.  The risk assessment background document should be a complete and transparent
30    treatment of variability within and among studies and how it affects risk estimates. Additionally,
31    SAB (2007) recommended considering the following issues when reviewing "low-level" and
32    "high-level" studies: (1) estimates of the level of exposure misclassification;  (2) temporal
33    variability  in assigning past arsenic levels from recent measurements; (3) the extent of reliance
34    on imputed exposure levels; (4) the number of persons exposed at various estimated levels of
                                                31       DRAFT—DO NOT CITE OR QUOTE

-------
 1    waterborne arsenic; (5) study response/participation rates; (6) estimates of exposure variability;
 2    (7) control selection methods in case-control studies; and (8) the resulting influence of these
 3    factors on the magnitude and statistical stability of cancer risk estimates.
 4          In order to address these issues, this Toxicological Review provides a comprehensive
 5    review of the significant epidemiologic investigations in the literature from  1968 to 2007 with
 6    the focus on the more recent publications.  The report includes data from all populations that
 7    have been examined in regards to cancer from arsenic exposure via drinking water. Earlier
 8    publications were reviewed and are included as needed to facilitate the understanding of results
 9    from certain  study populations. As recommended by SAB, studies were presented in both a
10    narrative (below) and tabular (Appendix B) format. Each publication was evaluated using a
11    uniform set of criteria, including the study type, the size of the study population and control
12    population, and the relative strengths and weaknesses of the study. While the information in the
13    tables  mirrors the information in the narrative, the narrative may provide additional important
14    information concerning the investigation.  The studies are presented by country of origin, then in
15    chronological order by publication year. In order to facilitate comparisons across the
16    epidemiological studies, the arsenic concentrations pertaining to water exposure levels have been
17    converted from milligrams (mg) per liter (or ppm) to parts per billion (ppb).  This was not
18    applied when discussing animal or in vitro MOA studies because a wide range of concentrations
19    was employed; converting the arsenic levels or doses into ppb would not be reader-friendly.
      4.1.1.  Taiwan
20          More than 80 years ago (between 1910 and 1920), parts of southwestern Taiwan began
21    using artesian (ground water) wells to increase water supplies and decrease  the salt content of
22    their drinking water.  Some of these artesian wells were discovered to be contaminated with
23    naturally occurring arsenic, thus resulting in widespread arsenic exposure.  As a result, the
24    Taiwanese population has been extensively studied. Due to the high arsenic content in the
25    artesian wells, water was piped into certain areas in Taiwan from the  reservoir of the Chia-Nan
26    irrigation system in 1956. This water was  reported to contain 10 ppb of arsenic (Tseng,  1977).
27    Almost 75% of the residences had tap water by the 1970s; however, a survey in  1988 noted that
28    artesian well water was still used for drinking, aquaculture, and agriculture in 1988, especially
29    during the dry season (Wu et al., 1989).
30           Tseng et al. (1968) conducted a general  survey using an ecological study design of
31    40,421 inhabitants (21,152 females, 19,269 males) from the southwest coast of Taiwan in order
32    to determine the potential relationship between skin cancer and chronic arsenicism. The arsenic
33    content was measured in 142 samples from 114 wells (110 deep artesian and 4 shallow) and
34    ranged from  10 to 1,820 ppb.  The authors noted, however, that the arsenic content varied
35    considerably over a 2-year period when measurements were taken. For example, in one well
36    measurements were 528 ppb in July, 1962; 530 ppb in June, 1963; and 1190 ppb in February,
                                                32       DRAFT—DO NOT CITE OR QUOTE

-------
 1    1964.  These variations made dose-response relationships difficult to determine.  Study subjects
 2    were categorized by arsenic exposure into three groups (low: 0-290 ppb, medium: 300-590 ppb,
 3    and high: 600 ppb or greater). The overall prevalence rate for skin cancer was 10.6 per 1,000.
 4    The male-to-female ratio was 2.9:1 for skin cancer. The prevalence rate increased steadily with
 5    age (recorded in 10-year increments), except for declining cancer prevalence rates for females
 6    older than 69 years. Age-specific (plotted in 20-year intervals) and sex-specific prevalence rates
 7    for skin cancer increased with arsenic concentration. The most common type of lesion was intra-
 8    epidermal carcinoma (51.7%), and the body areas most frequently involved were unexposed
 9    surfaces (74.5%). In addition, an extremely high percentage of cases with multiple skin cancer
10    (99.5%) was observed. The association between BFD and skin cancer was significantly higher
11    than expected.  Strengths of the Tseng et al. (1968) study include the large number of
12    participants and the inclusion of dose-response information.  Weaknesses include the lack of
13    individual exposure data (ecological  study design) and the potential for recall bias among study
14    participants in determining the age of cancer onset and the length of residence in the area.  In
15    addition, changes in water supply over time were not noted, information on smoking history was
16    not obtained, and the arsenic concentration from individual wells varied over time.
17          Tseng (1977) also used the general ecologic survey design discussed in Tseng et al.
18    (1968) to report skin  cancer incidence among the 40,421 individuals and to follow up on 1,108
19    patients with BFD (identified between 1958 and 1975).  By the end of the follow-up period, 528
20    of the BFD patients had died. Tseng (1977) identified 428 cases (prevalence of 10.6/1,000) of
21    skin cancer and 370 cases (prevalence of 9.0/1,000) of BFD, and analyzed the relationship
22    between the two.  Skin cancer and BFD occurred in 61 cases (1.51/1,000), but only 4 cases
23    (0.09/1,000) were expected.  The observed:expected ratio was 16.77. Tseng (1977) determined
24    that the patients with BFD consumed artesian water before the onset of the disease, and none of
25    the residents who had consumed only surface water or water from shallow wells developed BFD.
26    This finding illustrates that no cases were found among the inhabitants who were born after the
27    tap water supply was introduced, and supports the close association between the consumption of
28    arsenic contaminated water and the development of BFD. In addition, the study found that
29    patients with skin cancer or BFD had a greater incidence of death due to cancers of various sites
30    (28% and 19%, respectively) when compared to the general population of the endemic area
31    (13%) or to the entire population of Taiwan (8%).
32          Using similar arsenic exposure categories (low <300 ppb, medium 300-600 ppb, and
33    high >600 ppb) from the Tseng et al. (1968) investigation, the skin cancer and the BFD
34    prevalence rates showed an ascending gradient from low to high arsenic exposure for both sexes
35    (Tseng, 1977).  Skin cancer prevalence rates by age and arsenic exposure group were as follows:
36    20-39 years (high exposure:  11.5; medium exposure 2.2; low exposure: 1.3); 40-59 years (high:
37    72.0; medium: 32.6; low: 4.9); and 60+ years (high: 192.0; medium: 106.2; low: 27.1).  BFD

                                                33       DRAFT—DO NOT CITE OR QUOTE

-------
 1    prevalence rates by age and arsenic exposure group were as follows: 20-39 years (high: 14.2;
 2    medium: 13.2; low: 4.5); 40-59 years (high: 46.9; medium: 32.0; low: 10.5); 60+ years (high:
 3    61.4; medium: 32.2; low: 20.3). The common cause of death in the patients with skin cancer and
 4    BFD was carcinoma of various sites, including lung, bladder, liver, and kidney.  The Tseng
 5    (1977) investigation observed that the prevalence of skin cancer increased steadily with age.  It
 6    was difficult to obtain the age at onset of cancer from patient interviews, as most of the patients
 7    were unable to name a date.  Strengths and weaknesses of this study are the same as Tseng et al.
 8    (1968); however, this study also included adjusted analysis for age and gender.
 9          The objective of the Chen et al. (1985) ecological study was to evaluate the possible
10    association between exposure to elevated levels of arsenic from artesian well water and cancer in
11    the BFD-endemic area of southwestern Taiwan (i.e., Peimen, Hsuechia, Putai, and Ichu
12    townships). The population of the BFD-endemic area in 1982 was 120,607 and consisted
13    primarily of individuals engaged in farming, fishing, and salt production operations. The
14    educational and socioeconomic status of the BFD-endemic area was below average for Taiwan.
15    Chen et al. (1985) cited arsenic measurements from 83,565 wells across Taiwan taken by Lo et
16    al. (1977),  which showed that 29.1% of the wells in the study area had concentrations greater
17    than 50 ppb (with the highest concentration measuring 2500 ppb), while only 5.7% of wells in
18    other areas of Taiwan  exceeded 50 ppb. A previous study by Chen et al. (1962) demonstrated a
19    range of 350 to 1,140 ppb,  with a median of 780 ppb arsenic content in Taiwanese artesian wells
20    in BFD-endemic areas. As compared with the general population in Taiwan, both the
21    standardized mortality ratio (SMR) and cumulative mortality rate were significantly higher in
22    BFD-endemic areas.  SMRs for males were significant for bladder (11.00, 95% confidence
23    interval [CI]: 9.33-12.87),  kidney (7.72, 95% CI: 5.37-10.07), skin (5.34, 95% CI: 3.79-8.89),
24    lung (3.20, 95% CI: 2.86-3.54), liver (1.70, 95% CI: 1.51-1.89), and colon (1.80, 95% CI: 1.17-
25    2.03) cancers. SMRs for females also were significantly increased for bladder (20.09, 95% CI:
26    17.02-23.16), kidney (11.19, 95% CI: 8.38-14.00), skin (6.52, 95% CI: 4.69-8.35), lung (4.13,
27    3.60-4.66), liver (2.29, 95% CI: 1.92-2.66), and colon (1.68, 95% 1.26-2.10) cancers. Cancer
28    SMRs were greater in villages that used only artesian wells as the drinking water source, as
29    compared to villages that used both artesian and shallow wells.  Villages and townships using
30    only shallow wells generally had the lowest SMRs. Strengths of the investigation include the
31    use of general population of Taiwan and world population for determining SMRs and potential
32    confounders of age and gender were controlled for in the analysis. Weaknesses were that arsenic
33    measurements were not linked to cancer mortality, death certificates list the main cause of death
34    (Yang et al., 2005) rather than all causes, and SMRs were only presented by township and by
35    well type.
36          To  evaluate the association between high arsenic exposure from artesian well water and
37    cancer mortality in the BFD-endemic area of the southwest coast of Taiwan (i.e., the Peimen,

                                                34       DRAFT—DO NOT CITE OR QUOTE

-------
 1    Hsuechia, Putai, and Ichu townships), Chen et al. (1986) used a case-control study design to
 2    evaluate 69 bladder cancer, 76 lung cancer, and 65 liver cancer deceased cases and 368 alive
 3    community controls matched on age and gender.  The study area was the same one Chen et al.
 4    had used in 1985.  Cases were  selected from the Republic of China's National Health
 5    Department between January 1980 and December 1982. The age distribution for cases was
 6    significantly lower than the controls.  Similar gender distributions were observed for bladder and
 7    lung cancer cases and controls, though there was a slightly higher proportion of males in liver
 8    cancer cases than in controls. Other sociodemographic factors (marital status, education,
 9    occupation, and resident years) were comparable between cases and controls. Age and gender
10    differences were adjusted for in the analysis. The artesian well water arsenic content from the
11    BFD-endemic area ranged from 350 to 1,140 ppb (median 780 ppb), and the shallow well water
12    arsenic concentration ranged from below detection limits to 300 ppb (median 40 ppb).  A
13    positive dose-response relationship was observed between the exposure to artesian well water
14    and cancers of bladder, lung, and liver.  The age-gender-adjusted odds  ratios (ORs) of bladder,
15    lung, and liver cancers for those who had used artesian well water for 40 or more years were
16    3.90, 3.39, and 2.67, respectively, when compared with those who never used artesian well
17    water. Regression analyses examined the associations between exposure to artesian well water
18    and bladder, lung, and liver cancers after adjusting for other variables including age, gender, and
19    cigarette smoking. Results showed a statistically significant association between exposure to
20    artesian well water and bladder and lung cancers (p < 0.01) when other variables were
21    controlled, but the association between the exposure to artesian well water and liver cancer was
22    not statistically significant (p < 0.05). (The text of the article specifies that liver cancers are not
23    significantly associated with arsenic, but the table that the text refers to illustrates a significant
24    association.) Strengths of the Chen et al. (1986) study include that most cases were confirmed
25    using histology or cytology findings, cancer cases and controls were from the same BFD
26    community, and potential confounders were adjusted for in the analysis (i.e., age, gender,
27    smoking, tea consumption, vegetable consumption, and fermented bean consumption).
28    Weaknesses include  selection bias (control selection)  and not controlling recall bias for the
29    following confounders: lifestyle, diet, daily water consumption, and source of water.
30          In a cohort study conducted by Chen et al. (1988a), cancer mortality associated with BFD
31    was analyzed in area residents  (i.e., Peimen, Hsuechia, Putai, and Ichu  townships, Taiwan) from
32    1973 to  1986.  Arsenic levels in drinking water were measured between 1962 and 1964; these
33    levels were used to divide the study population into three groups: <300 ppb; 300-599 ppb; and
34    >600 ppb.  Sociodemographic  characteristics including lifestyle, diet, and living conditions were
35    comparable among study participants.  Between 1974 and 1976, water  from more than 83,000
36    wells in 313 villages throughout Taiwan was reanalyzed for arsenic content. The levels of
37    arsenic in the drinking water were consistent between the two measurement periods. Death

                                                3 5       DRAFT—DO  NOT CITE OR QUOTE

-------
 1    certificates (n = 1031) were obtained from Taiwanese health care registration offices.  Age-
 2    adjusted cancer mortality rates were calculated using the 1976 world population as the standard.
 3    Significantly elevated dose-response cancer mortality was observed among residents of the BFD
 4    area (<300 ppb, SMRfemale=118.8, male=154.0; 300-599 ppb SMRfemale=182.6,
 5    male=258.9; >600 ppb SMR female=369.1, male=434.7) as compared to the general population
 6    of Taiwan (SMR female=85.5, male=128.1).  For both genders, significantly elevated dose-
 7    response mortality also was observed for cancers of the liver, lung, skin, bladder, and kidney in
 8    comparison to the general population of Taiwan. A strength of the study is that data from
 9    arsenic monitoring conducted in 1962-64 and 1974-76 revealed similar results. A weakness of
10    the study is that arsenic exposure levels are not individualized.
11          The objective of the Chen et al. (1988b) cohort (nested case-control) study was to
12    examine multiple risk factors and their correlation to malignant neoplasms related to BFD. A
13    total of 241 BFD cases, including 169 with spontaneous or surgical amputations of affected
14    extremities and 759 age-sex-residence-matched healthy community controls, were identified and
15    studied in the Peimen, Hsuechia, Putai, and Ichu townships of southwest Taiwan. Multiple
16    logistic regression analysis showed that artesian well water consumption,  arsenic poisoning,
17    familial history of BFD, and undernourishment were significantly associated with the
18    development of BFD. In a nonconcurrent cohort, cancer mortality of 789 BFD patients followed
19    for 15 years also was examined using a life table.  Results showed a significantly higher
20    mortality from cancers of the bladder (SMR=38.80, p < 0.001), skin (SMR=28.46, p < 0.01),
21    lung (SMR=10.49, p < 0.001), liver (SMR=4.66, p < 0.001), and colon (SMR=3.81, p < 0.05) as
22    compared with the general population in Taiwan.  When non-BFD residents in the BFD-endemic
23    area were used as controls, significant differences in mortality rates were found for cancers of
24    the bladder (SMR=2.55, p < 0.01), skin (SMR=4.51, p < 0.05), lung (SMR=2.84, p < 0.01), and
25    liver (SMR=2.48, p < 0.01).  The results strongly suggest carcinogenic effects from the artesian
26    well water in the BFD-endemic area.  Study strengths include minimizing recall bias through
27    interview techniques, which identified the education, hours of occupational sunshine exposure,
28    artesian well use, family medical history, history of smoking and alcohol use, and frequency of
29    categories of food consumption. SMRs were calculated using both the national Taiwanese
30    population and the local endemic area population, and BFD cases were matched to healthy
31    community controls for age, sex, and residence. A weakness of the study was not providing the
32    individual arsenic dose levels.
33          Chiang et al. (1988) conducted a case-control prevalence study of bladder cancer in the
34    BFD-endemic and surrounding areas of the southwestern coast of Taiwan. Four groups (cases:
35    246 BFD patients; controls: 444 residents of the BFD-endemic area, 286 residents of the region
36    neighboring the endemic area, and 731 residents of the non-endemic area) were screened using a
37    detailed questionnaire and urinalysis. Three hundred and four subjects received urinary cytology

                                                36       DRAFT—DO NOT CITE OR QUOTE

-------
 1    examinations. The study revealed no difference in the prevalence of bladder cancer between the
 2    BFD patients and non-BFD controls in the BFD-endemic area, indicating that individuals in the
 3    BFD-endemic area were equally affected by a high prevalence of bladder cancer.  A high
 4    prevalence of bladder cancer in the BFD-endemic area was noted when compared with the
 5    neighboring region and residents of the non-endemic area. However, sporadic cases of bladder
 6    cancer were noted in the region neighboring the endemic area. This study also found that the
 7    non-BFD-endemic areas, which had a high arsenic content in the well water, did not have a high
 8    prevalence of bladder cancer, indicating other possible environmental factors.  The histological
 9    confirmation of bladder cancer diagnoses is a strength of the study; however, the lack of
10    individual arsenic exposure data is a limitation.
11          Wu et al. (1989) analyzed age-adjusted mortality rates using an ecological study design
12    to determine whether a dose-response relationship exists between ingested arsenic levels and the
13    risk of cancer among residents in the BFD endemic area.  The study population consisted of a
14    cohort of individuals from the southwestern coast of Taiwan (27 villages from the townships of
15    Peimen, Hsuechia, Putai, and Ichu and 15 villages from the townships of Yensui and Hsiaying).
16    The arsenic levels in well water for the 42 villages were determined from 1964 to 1966, while
17    mortality and population data were obtained for the years of 1973 to 1986 from the local
18    registration offices and from the Taiwan Provincial Department of Health. Age-adjusted
19    mortality rates from various cancers by gender were calculated using the 1976 world population
20    as the standard population.  A significant dose-response relationship was observed between
21    arsenic levels in well water and bladder, kidney, skin, and lung cancers in both males and
22    females.  A similar relationship was observed for prostate and liver cancers in males.  There was
23    no association for leukemia or cancers of the nasopharynx, esophagus, stomach, colon, and
24    uterine cervix.  Strengths of the study include the fact that adjustments were made for age and
25    gender, and that lifestyle, access to medical care, and socioeconomic status were similar among
26    the study groups.   The use of mortality data can be considered a weakness of the study, since
27    death certificates  may not list all cancers.  Additionally, associations observed at the local level
28    may not be accurate at the individual level.
29          The Chen and Wang (1990) ecological study was  carried out to examine correlations
30    between the arsenic level in well water and mortality from various malignant neoplasms in 314
31    precincts and townships of Taiwan.  The arsenic content of water from 83,656 wells was
32    available from measurements taken in 1974 through 1976. Mortality rates from 1972 to 1983
33    were derived from residents in study precincts and townships who displayed one or more of the
34    21  examined malignant neoplasms. Arsenic content in the water was available at the precinct or
35    township level. A statistically significant association with the arsenic level in well water was
36    observed for cancers of the liver, nasal cavity, lung, skin, bladder, and kidney in both males and
37    females, as well as for prostate cancer in males. These associations remained significant after

                                                37        DRAFT—DO NOT CITE OR QUOTE

-------
 1    adjusting for indices of urbanization and industrialization through multiple regression analyses.
 2    No significant association was identified for the other 14 cancers examined.  The multivariate-
 3    adjusted regression coefficient showed an increase in age-adjusted mortality for cancers in males
 4    and females for every 100 ppb increase in arsenic level in well water. Coefficients for males and
 5    females, respectively, were as follows: 6.8 and 2.0 (liver), 0.7 and 0.4 (nasal cavity), 5.3 and 5.3
 6    (lung), 0.9 and 1.0 (skin), 3.9 and 4.2 (bladder), and 1.1 and 1.7 (kidney).  Results were
 7    unchanged when 170 southwestern townships were included. Strengths of the study were that
 8    potential confounders (including socioeconomic differences, i.e., urbanization and
 9    industrialization) were controlled for, the study reported ecological correlations between arsenic
10    content in well water and mortality from various cancers,  and cancer rates in endemic BFD
11    townships were compared with cancer rates in non-endemic townships of Taiwan.  Potential
12    confounders not controlled for were gender,  other potential well water exposure contaminants,
13    and individual arsenic exposures that were not available.
14          Using an ecologic investigation, Chen et al. (1992) showed a comparable excess risk of
15    cancer of liver, lung, bladder, and kidney cancers induced by arsenic in drinking water. The
16    study area and population were previously described by Wu et al. (1989).  In order to compare
17    the risk of developing various cancers as the result of ingesting inorganic arsenic and to assess
18    the differences in risk between males and females, cancer potency indices were calculated with
19    the Armitage-Doll multistage model using mortality rates among residents of 42 villages in six
20    townships (Peimen, Hsuechia, Putai, Ichu, Yensui, and Hsiaying) located on the southwest coast
21    of Taiwan. Locations selected were considered to be chronic arsenicism endemic areas. Arsenic
22    exposure levels from drinking water in these villages were categorized into four groups: <100
23    ppb (13 villages), 100-299 ppb (8 villages), 300-599 ppb (15 villages), and 600 ppb or greater
24    (6 villages). Based on a total of 898,806 person-years during the study period from 1973
25    through 1986, a significant dose-response relationship was observed between the arsenic level in
26    drinking water and cancer mortality of the liver, lung, bladder, and kidney. The lifetime risk
27    (determined using the Armitage-Doll model) of developing cancer due to an intake of 10 ug/kg-
28    day of arsenic was estimated to be 4.3 x 10-3 (liver), 1.2 x 10-2 (lung), 1.2 x  10-2 (bladder), and
29    4.2 x 10-3 (kidney) for males and 3.6 x 10-3 (liver), 1.3 x 10-2 (lung), 1.7 x 10-2 (bladder), and
30    4.8 x 10-3 (kidney) for females. Strengths include that potential confounders including age,
31    gender, access to medical care,  socioeconomic status, and lifestyle were all controlled for during
32    the analysis, and that villages shared similar socioeconomic status, living environments,
33    lifestyles, dietary patterns, and medical facilities.  A weakness of the study is the assumption that
34    an individual's arsenic intake remained constant from birth to the end of the follow-up period;
35    this flaw possibly led to the underestimation of risk. Additional weaknesses include that the
36    Armitage-Doll model constrains risk estimates to be monotonically increasing function of age,
                                                 3 8        DRAFT—DO NOT CITE OR QUOTE

-------
 1    that dietary sources of arsenic were not quantified, and that age stratification was for under 30,
 2    over 70, and 20-year strata.
 3          To determine whether a dose-response relationship exists between ingested inorganic
 4    arsenic and cancer, Chiou et al. (1995) used a cohort study with a total of 263 BFD patients and
 5    2293 healthy residents in the arseniasis-endemic area of southwestern coast of Taiwan (Peimen,
 6    Hsuechia, Putai, and Ichu townships).  Participants were followed for an average of 4.97 years
 7    (range: 0.05-7.69 years). Data concerning the consumption of artesian well water containing
 8    high levels of arsenic, sociodemographic characteristics, lifestyle and dietary habits, and cancer
 9    histories were obtained through a standardized interview. Internal cancers were determined via
10    health examinations, personal interviews, household registration data checks, and Taiwan's
11    national death certification and cancer registry databases. Concentrations used in the assessment
12    were < 50 ppb, 50-70 ppb, 71+ ppb, and unknown.  Disregarding the unknown category, a dose-
13    response relationship was observed between the long-term arsenic exposure from drinking
14    artesian well water and the incidence of lung cancer, bladder cancer, and cancers of all sites
15    combined after adjusting for age, sex, and cigarette  smoking through a Cox's proportional
16    hazards regression analysis.  BFD patients had a significantly increased incidence of bladder
17    cancer and for all sites combined after adjusting for age, gender, smoking history, and
18    cumulative arsenic exposure (CAE). Strengths include that the analysis adjusted for BFD status,
19    age, gender, and smoking; incidence data were reported; and the results of the study showed a
20    significant dose-response relationship. A weakness of the study is that well water artesian
21    arsenic concentrations were unknown for some study subjects; consequently, this was a
22    significant confounder.
23          To further evaluate the association between arsenic exposure in drinking water and
24    urinary cancers of various cell types, Guo et al. (1997) conducted an ecological study
25    encompassing 243 townships using Taiwanese National  Cancer Registry data of patients
26    diagnosed with cancer between 1980 and 1987. Wells with known arsenic concentrations in
27    each township were used to separate people into the following exposures: <50 ppb, 50-80 ppb,
28    90-160 ppb, 170-320 ppb, 330-640 ppb, and >640  ppb. The effects of urbanization and
29    smoking were evaluated by an urbanization index based on 19 socioeconomic factors shown to
30    be good indicators of urbanization and the number of cigarettes sold per capita. For both
31    genders, Guo et al.  observed associations between high arsenic levels in drinking water and
32    transitional cell carcinomas (bladder, kidney, ureter, and all urethral cancers combined).
33    Positive associations between the proportion of wells with arsenic levels above 640 ppb and the
34    incidence of transitional cell carcinomas of the bladder, kidney,  ureter, and all urethral cancers
35    combined in both genders were identified after the model was adjusted for urbanization and age.
36    Arsenic exposure in males was associated with adenocarcinomas of the bladder, but not in
37    squamous cell carcinomas of the bladder or renal cell carcinomas or nephroblastomas of the

                                                39        DRAFT—DO NOT CITE OR QUOTE

-------
 1    kidney. Males also exhibited a positive association between the urbanization index and
 2    transitional cell carcinomas of the ureter.  The results support the case that the carcinogenicity of
 3    arsenic may be cell-type specific.  Analyses were adjusted for age, gender, urbanization, and
 4    smoking; however, the ecologic study design was a limitation.
 5          Tsai et al. (1999) conducted a cross-sectional study in BFD-endemic areas in the
 6    southwest coastal region of Taiwan (Peimen, Hsuechia, Putai, and Ichu townships) to analyze
 7    mortality from neglected cancers related to artesian well water containing high levels of arsenic.
 8    The median artesian well water arsenic content was 780 ppb (range: 250-1,140 ppb).  Local
 9    endemic area residents' daily ingestion of arsenic was estimated to be < 1 mg.  SMRs were
10    calculated for cancer diseases, by gender, during the period from 1971 to 1994.  These SMRs
11    were compared to the local reference group (Chiayi-Tainan County population) and a national
12    reference group (Taiwanese population). The comparisons revealed significant differences
13    between SMRs of the three groups. Mortality increases (p < 0.05) were found in males and
14    females, respectively, for all cancers (SMR=2.19, 95% CI: 2.11-2.28; SMR=2.40, 95% CI:
15    2.30-2.51) when compared to the local reference population. Additionally, the following other
16    cancers showed mortality increases in males and females, respectively, when compared to the
17    local reference population: bladder (SMR=8.92,  95% CI: 7.96-9.96; SMR=14.07, 95% CI:
18    12.51-15.78); kidney (SMR=6.76, 95% CI: 5.46-8.27; SMR=8.89, 95% CI: 7.42-10.57); skin,
19    lung, nasal-cavity, bone, and liver (SMR=1.83, 95% CI:  1.69-1.98; SMR=1.88, 95% CI: 1.64-
20    2.14);  and larynx, stomach, colon, intestine, rectum, lymphoma, and prostate cancer in males
21    only (SMR=2.52, 95% CI: 1.86-3.34). When compared to the national reference population,
22    significantly increased  (p < 0.05) mortality was found in males and females, respectively, for all
23    cancers (SMR=1.94, 95% CI: 1.87-2.01; SMR=2.05, 95% CI: 1.96-2.14) and for the other
24    following cancers: bladder (SMR=10.50, 95% CI: 9.37-11.73; SMR=17.65, 95% CI: 5.70-
25    19.79) and lung (SMR=2.64, 95% CI: 2.45-2.84; SMR=3.50, 95% CI: 3.19-3.84). The results
26    of the  Tsai et al. (1999) investigation indicate that the hazardous effect of arsenic may be
27    systemic. Key strengths of the study are that the exposed group and local reference group had
28    similar lifestyle factors; all cancers were pathologically confirmed; and the analysis controlled
29    for gender.  Weaknesses of the study are that death certificates indicated only one underlying
30    cause  of death (not multiple causes), resulting in possible distortion of association between
31    exposure and disease; individual exposure data were not provided; and certain potential
32    confounders were not controlled for (age,  smoking history, alcohol consumption, and
33    occupational exposures).
34          The Morales et  al. (2000) ecological investigation re-analyzed data originally reported by
35    Chen et al. (1988a, 1992) and Wu et al. (1989) from 42 villages in the arseniasis-endemic region
36    of southwestern Taiwan by considering the number of liver, lung, and bladder cancer deaths.
37    Morales et al. (2000) used a generalized linear model (i.e., Poisson distribution) and the

                                                40        DRAFT—DO NOT CITE OR QUOTE

-------
 1    multistage-Weibull models to determine lifetime cancer risk estimates. Liver, lung, and bladder
 2    cancer mortality data were collected from death certificates of residents in 42 villages during
 3    1973 through 1986.  Drinking water samples had been collected from wells in the 42 villages
 4    between 1964 and 1966.  SMRs were used to summarize the observed patterns of mortality in the
 5    collected data.  Morales et al. (2000) selected two comparison populations (the Taiwanese
 6    population as a whole and a population from a southwestern region of Taiwan) to account for
 7    urban versus non-urban populations differences. Although a non-significant trend was observed
 8    in the combined cancer analyses with respect to age, there was no observed tendency in liver,
 9    lung, or bladder SMRs with respect to age.  This suggests that there is no age dependency on the
10    risk ratio. Liver cancer mortality was higher than expected, although there was no strong
11    exposure-response relationship found. The Morales et al. (2000) investigation results showed
12    that exposure-response assessments were highly dependent on the choice of the analysis model
13    and whether or not a comparison population is used in the analysis. One possible explanation for
14    this observation is the inherent uncertainty associated with the limitations of an ecological study
15    design. Depending on the model used and the comparison population used in the analysis, the
16    effective dose at the 1% level (ED01) estimates ranged from 21 to 633 ppb for male bladder
17    cancer, and from 17 to 365 ppb for female bladder cancer. The lung cancer risk for males was
18    found to be slightly higher than the bladder cancer risk, with ED01  estimates ranging from 10 to
19    364 ppb. The risk for female cancer tended to be higher than that of males for each cancer type.
20    For lung cancer, female ED01 estimates ranged from 8  to 396 ppb.
21           In summary, the Morales et al. (2000) analysis of the Taiwan data suggests that excessive
22    cancer mortality may occur in many populations where the drinking water standard for arsenic is
23    set at 50 ppb, the drinking water standard for arsenic in the United States at the time of
24    publication.  A strength of the study was that person-years at risk (PYR) were stratified by 5-
25    year age groups, gender, and median arsenic level for each village.  Weaknesses  include the
26    ecological study design (i.e., there were no individual monitoring data and individual exposures
27    were not available) and the fact that potential confounders such as smoking,  dietary arsenic, and
28    the use of bottled water (U.S. population) were not controlled for in the analysis.
29           Between 1991 and 1994, Chiou et al. (2001) recruited  a cohort of 8,102 residents aged 40
30    years or older from four townships (18 villages) in northeastern Taiwan (4 villages in Chiaohsi, 7
31    in Chuangwei, 3 in Wuchih, and 4 in Tungshan) and followed it until the end of 1996. The study
32    examined the risk of transitional cell carcinoma in relation to ingested arsenic. The Chiou et al.
33    (2001) findings were consistent with previously reported findings from the arsenic-endemic area
34    of southwestern Taiwan.  Based on the arsenic concentration in well water, each  study subject's
35    individual exposure to inorganic arsenic was estimated. Information concerning  the duration of
36    consumption of the well water was obtained through standardized questionnaire interviews.
37    Urinary tract cancers were identified by follow-up interviews, community hospital records, the

                                                41        DRAFT—DO NOT CITE OR QUOTE

-------
 1    Taiwanese national death certification profile, and the cancer registry profile. A significantly
 2    increased incidence of urinary tract cancers for the study cohort was observed (standardized
 3    incidence ratio [SIR]=2.05; 95% CI:  1.22-3.24) when compared to the general population in
 4    Taiwan. In addition, a dose-response relationship was observed between the risk of cancers of
 5    the urinary organs, especially transitional cell carcinoma, and indices of arsenic exposure after
 6    adjusting for age, sex, and cigarette smoking. The relative risks (RR) of developing transitional
 7    cell carcinoma were 1.9, 8.2, and 15.3 for arsenic concentrations of 10.1-50.0 ppb, 50.1-
 8    100.0 ppb, and >100.0 ppb, respectively, compared with the referent level of < 10.0 ppb. No
 9    association was observed for the duration of well water  drinking (<40 years compared to
10    > 40 years).  The findings of this study suggest that arsenic ingestion may increase the risk of
11    urinary tract cancer at levels around 50 ppb. Strengths include adjustments for potential
12    confounders (age, gender, smoking history), individual arsenic exposure estimates, and  a dose-
13    response relationship even with the low levels of arsenic. Weaknesses include possible
14    diagnostic bias as the result of medical data collection from various community hospitals and
15    recall bias from self-reported information. The short duration of follow-up also is a limitation
16    because it impacted: (1) the number of person-years of observation; and (2) only a few cases
17    were recorded. This study also has an apparent supralinear curve, which is likely due to dose
18    misclassification in the low-dose individuals. If food arsenic concentrations (estimated in NRC,
19    2001, to be approximately 50 ug/day) were included, the curve might not be supralinear.
20           Guo et al. (2001) conducted an  ecological investigation of the 243 townships from their
21    1997 publication; however, this investigation focused on arsenic exposure through drinking
22    water and the potential association with skin cancers. Data regarding arsenic levels in drinking
23    water were available from the previous investigation, and cases of skin cancer were identified
24    using the Taiwanese National Cancer Registry. Data were analyzed with regression models
25    using multiple variables to describe exposures, including arsenic. To adjust for potential
26    confounding variables, an urbanization index based on 19 socioeconomic factors shown to be
27    good indicators of urbanization was developed.  A total  of 2,369 individuals with skin cancer
28    (954 females and 1,415 males) were registered with the  Cancer Registry between January 1980
29    and December 1989. After age and urbanization adjustment, arsenic levels above 640 ppb
30    showed a statistically significant (p < 0.01) association with the incidence of basal cell
31    carcinoma (BCC) in males.  Exposed females also exhibited an increased incidence in skin
32    cancer rates; however, this increase did not reach statistical significance (p = 0.20). For
33    squamous cell carcinomas (SCC), a significant (p < 0.01), positive association was found for
34    males exposed to 170-320 ppb and >640 ppb. However, a statistically significant (p < 0.01)
35    negative association was found for males exposed to 330-640 ppb. For females, a similar
36    statistically significant (p < 0.01) positive association was observed at >640 ppb, while  a
37    statistically significant (p < 0.05) negative association was observed in 330-640 ppb females.

                                                 42       DRAFT—DO NOT CITE OR QUOTE

-------
 1    For melanomas, no significant associations were identified in females or males at any exposure.
 2    The results of the investigation suggest that skin cancers are cell-type-specific, as previously was
 3    demonstrated for urinary tract cancers (Guo et al., 1997). Strengths of the study include that
 4    cases were identified from a government operated National  Cancer Registration Program,
 5    pathological  classifications were determined by board-certified pathologists, and potential
 6    confounders  (gender and age) were adjusted in the analysis. A limitation of the study is the
 7    ecological study design.
 8          Studies on cancers of the urinary system and skin showed that arsenic's carcinogenic
 9    effect was cell-type-specific (Guo et al.,  1997, 2001).  Guo (2003) conducted an ecological
10    investigation in 243 townships in Taiwan, previously used in the Guo et al. (1997, 2001)
11    investigations for urinary and skin cancers, to determine if a similar relationship could be
12    identified for liver cancer. Many previous epidemiologic studies did not provide data on
13    pathological  diagnoses; therefore, there was no information to support the hypothesis that
14    hepatocellular carcinoma (HCC) or cholangiocarcinoma of the liver were not associated with
15    arsenic ingestion.  Liver cancers were identified through the Taiwanese National Cancer
16    Registry. The distribution of cancer cell-types between an arseniasis-endemic area and a
17    township outside the arseniasis area were compared.  Between January 1980 and December
18    1999, 32,034 men and 8,798 women living in the study townships were diagnosed with liver
19    cancer. The  distribution of two cancer cell-types (HCC and cholangiocarcinoma) did not appear
20    to be different between the arseniasis-endemic and non-arseniasis-endemic areas, and an
21    association between HCC and arsenic ingestion was not observed. The remainder of the cell-
22    types did not have enough cases to provide stable estimates. Identified strengths of the study
23    include the following: cases were identified from the government-operated National Cancer
24    Registration Program; pathological classifications were determined by board-certified
25    pathologists; and analyses were adjusted for gender and age. Weaknesses include the limitations
26    of ecological study design (no monitoring data were presented).
27          A cohort investigation of residents from two arsenic endemic areas were followed for 8
28    years by Chen et al. (2004a) to investigate the dose-response relationship between arsenic
29    exposure and lung cancer, as well as how cigarette smoking influenced the relationship between
30    arsenic and lung cancer. Arsenic-endemic areas included the southwestern coast (Peimen,
31    Hsuechia, Putai, and Ichu; n = 2,503) and the northeastern coast (Tungshan, Chuangwei,
32    Chiaohsi, and Wuchieh; n = 8,088) of Taiwan. The amount of arsenic in well water from these
33    areas ranged from less than 0.15 ppb to more than 3,000 ppb.  The Taiwanese National Cancer
34    Registry was used to identify new cases of lung cancer diagnosed between January 1, 1985, and
35    December 31, 2000. For each participant, follow-up person-years were calculated using the time
36    from the initial interview date to the date of diagnosis, death, or December 31, 2000, whichever
37    came first. Arsenic concentration was arbitrarily divided into five categories: <10 ppb (referent),

                                                43        DRAFT—DO NOT CITE OR QUOTE

-------
 1    10-99.9 ppb, 100-299.9 ppb, 300-699.9 ppb, and >700 ppb.  Smoking histories were obtained
 2    from interviews. Cox proportional hazards regression models were used to estimate RR and
 3    95% CI. The final model was adjusted for age, gender, years of schooling, study cohort (BFD
 4    cases and matched controls of the southwestern coast, residents along the arseniasis-
 5    hyperendemic southwestern coast villages, and residents living in the northeastern coastal
 6    Lanyang Basin), smoking status, and alcohol consumption. During the study follow-up period,
 7    there were 139 lung cancers diagnosed, resulting in an incidence rate of 165.9 per 100,000
 8    person-years.  When the highest level of arsenic exposure was compared to the lowest,  the RR
 9    was 3.29 (95% CI: 1.60-6.78). The risk of lung cancer was four times higher for past and
10    current smokers compared to non-smokers. A synergistic effect of ingested arsenic and cigarette
11    smoking on lung cancer was noted, with synergy indices ranging from 1.62 to 2.52. Strengths of
12    the study include controlling for confounders (age, gender, education,  smoking history, and
13    alcohol consumption), having a long follow-up period, using a national computerized cancer
14    case registry, and pathologically confirming all lung cancer cases. Weaknesses include the lack
15    of historical monitoring data and possible misclassification bias (exposure measurements were
16    based on one survey).
17           Chiu et al. (2004), using a cohort study design, examined whether liver cancer mortality
18    rates were altered after the consumption of high-arsenic artesian well water ceased.  SMRs for
19    liver cancer were calculated for the BFD-endemic area of the southwest coast of Taiwan (i.e.,
20    Peimen, Hsuechia, Putai, and Ichu townships) for the years 1971 through 2000. Median well
21    water arsenic concentrations in the early 1960s were 780 ppb.  Temporal changes in the SMRs
22    were monitored using cumulative-sum techniques and were reported for 3-year intervals between
23    1971 and 2000.  Study results showed that female mortality from liver cancer started declining 9
24    years after consumption of high-arsenic artesian well water stopped. The SMR for liver cancer
25    in females was 2.041 during the 1983-1985 period (peak) and was 1.137 during 1998 through
26    2000.  Data in males, however, showed fluctuations in liver cancer mortality rates. The SMR for
27    liver cancer in males from 1989 to 1991 was 1.868 and 1.242 during 1998 to 2000. Based on
28    analyses by Chiu et al. (2004), it was determined that the relationship between arsenic exposure
29    and liver cancer mortality was possibly causal in females, but not in males.  Strengths of the
30    study are: (1) residents in the study area were similar in terms of socioeconomic status, living
31    environments, lifestyles, dietary patterns, and availability  of health service facilities;  and (2) the
32    study used an accurate death registration system. Weaknesses include the limitations of the
33    mortality data.
34           To obtain data on the potential dose-response relationship between lung cancer  and the
35    level of arsenic in drinking water, Guo (2004) conducted an ecological investigation in 10
36    townships (138 villages) in Taiwan.  Measurements of arsenic levels in drinking water were
37    available for the 138 villages from a census survey conducted by the Taiwanese government.

                                                44        DRAFT—DO NOT CITE OR QUOTE

-------
 1    Death certificates dated between January 1, 1971, and December 31, 1990, were reviewed, and
 2    673 males and 405 females were identified as dying from lung cancer. Multivariate regression
 3    models were applied to assess the relationship between arsenic levels in drinking water and lung
 4    cancer mortality. After adjusting for age, arsenic levels above 640 ppb were associated with a
 5    significant increase in lung cancer mortality for both genders; however, no significant effect was
 6    observed at lower arsenic exposure levels.  Regression analyses  and stratified analyses
 7    confirmed a dose-response relationship at >640 ppb.  Guo (2004) noted that the results of this
 8    investigation show a carcinogenic effect of high arsenic levels in drinking water on the lung.
 9    Guo (2004), however, recommended that further studies with exposure data on individuals were
10    warranted to confirm these findings. As a result of the study's ecologic design, the association
11    observed on an aggregate level may not necessarily represent the association that exists at an
12    individual level. In addition, the study design may have contributed to biases introduced by the
13    effects of population mobility. Strengths of the study include that analyses adjusted for gender
14    and age, and cases were ascertained using information from household registry offices in each
15    township. Weaknesses of the investigation include the inherent  limitations of ecological studies
16    and the fact that smoking was not controlled for in the analysis.
17          In a  cross-sectional study,  Yang et al. (2004)  examined whether kidney cancer mortality
18    decreased in the southwest coast of Taiwan (Peimen, Hsuechia, Puta,  and Ichu townships) after
19    the elimination of arsenic exposure in the 1970s. SMRs for kidney cancer were calculated for
20    the BFD-endemic area for the years 1971 through 2000. There were 308 kidney cancer deaths
21    (135 men and 173 women) in the BFD-endemic area between 1971 and 2000.  The means of the
22    3-year SMRs for female and male kidney cancer were significantly higher than for Taiwan as a
23    whole. Time series plots for male SMRs showed decreasing mortality rates. The estimated
24    slope for male SMRs (rate of decrease per year) in the linear time trend analysis was -15.13
25    (p < 0.01).  The time series plot for female SMRs also showed decreasing mortality rates.
26    Kidney cancer mortality rates among residents in the BFD-endemic area decreased after removal
27    of the arsenic source through tap water implementation. SMRs decreased each year, on average,
28    from 1971 to 2000 (p < 0.01). Study strengths include the adjustment of potential confounders
29    (gender and age); mandatory registering of all births, deaths, marriages, divorces, and migration
30    issues with the Household Registration Office in Taiwan, making it an accurate data source; and
31    a comparable study population (i.e., residents likely had similar  socioeconomic status, living
32    environments, lifestyles, dietary patterns; they worked in farming, fisheries, or salt production)
33    that had comparable access to medical care (i.e., all kidney cancer cases likely had similar access
34    to medical care). Weaknesses of the study include cross-sectional mortality limitations and not
35    adequately controlling for smoking histories.
36          Tsai et al. (2005) used a cross-sectional study to compare primary urethral carcinomas
37    from the BFD-endemic area of Taiwan with those in the United  States and explore the potential

                                                45        DRAFT—DO NOT CITE OR QUOTE

-------
 1    influence of chronic arsenic exposure.  Cases were identified by the only medical center near the
 2    BFD area.  There were 21 pathologically proven primary urethral carcinomas diagnosed (7
 3    females and 14 males) between 1988 and 2001.  Seven of 14 male patients had reported an
 4    average of 23 years of chronic arsenic exposure from drinking water. Tsai et al.  (2005)
 5    compared these cases to cases identified in three U.S. cancer centers (MD Anderson, Memorial
 6    Sloan-Kettering, and Barbara Ann Karmanos; n = 79 females, n = 80 males), and analyzed for a
 7    relationship with chronic arsenic exposure. In comparison to the three U.S.  cancer centers, there
 8    was a higher frequency of bulbomembranous adenocarcinoma (43% vs. 18%, 2%, and 0%,
 9    respectively, p < 0.0001). In those with chronic arsenic exposure, there was an even greater
10    association with bulbomembranous adenocarcinoma compared to those without chronic arsenic
11    exposure (73% vs. 14%, p=0.031).  Based on these results, Tsai et al. (2005) concluded that the
12    BFD-endemic area in Taiwan had a high frequency of bulbomembranous urethral
13    adenocarcinoma, which may be associated with chronic arsenic exposure. A strength of the
14    study is that cases were pathologically  confirmed. The small number of cases and the lack of
15    arsenic exposure information are study weaknesses.
16          The objective of the Yang et al. (2005) cross-sectional study was to determine whether
17    bladder cancer mortality decreased after the implementation of the tap  water system and the
18    subsequent elimination of arsenic exposure.  SMRs for bladder cancer  were  calculated for the
19    BFD-endemic area for the years 1971-2000. The study showed that bladder cancer mortality
20    decreased gradually after the instillation of the tap water system, thereby eliminating exposure to
21    arsenic through artesian well water, (1971, male SMR=10.25, female SMR=14.89; 2000,  male
22    SMR=2.15, female SMR=7.63). Strengths include similar access to medical care for bladder
23    cancer, the adjustment for age and gender, and the mandatory registering of all births, deaths,
24    marriages, divorces, and migration issues to the Household Registration Office in Taiwan,
25    making it an accurate data source.  Limitations of the study include the cross-sectional mortality
26    study design and smoking history confounding.
      4.1.2. Japan
27          Tsuda et al. (1995) used a cohort study to investigate the long-term effect of ingesting
28    arsenic in drinking water for an estimated  exposure period of 5 years (1955-1959).  Four
29    hundred and fifty-four residents identified in 1959 as living in an arsenic-polluted area of Niigata
30    Prefecture, Japan, were followed until 1992.  The mortality of these residents between October 1,
31    1959, and February 29, 1992, was examined using death certificates. These individuals used
32    arsenic-contaminated well water, and none worked at a nearby factory  that was the source of the
33    water contamination. Death certificates for the people who died between 1959 and 1992 were
34    examined and a total of 113 of the 454  residents were estimated to have consumed well water
35    containing a high concentration of arsenic (>1,000 ppb).  The SMRs of these 113 residents were
36    15.69 for lung cancer (95% CI: 7.38-31.02) and 31.18 for urinary tract cancer (95% CI: 8.62-
                                                46       DRAFT—DO NOT CITE OR QUOTE

-------
 1    91.75). Cox's proportional hazard analyses demonstrated that the hazard ratios of the highest
 2    exposure level group (> 1,000 ppb) versus the background exposure level group (1.0 ppb) were
 3    1.74 (95% CI: 1.10-2.74) for all deaths, 1972.16 (95% CI: 4.34-895,385.11) for lung cancer,
 4    and 4.82 (95% CI: 2.09-11.14) for all cancers.  The study also analyzed skin signs of chronic
 5    arsenicism, and results indicated that they were useful risk indicators for subsequent cancer
 6    development. These results indicate a relationship between well water arsenic exposure and lung
 7    and urinary tract cancers. The study also showed that arsenic-induced cancer could develop
 8    years following the end of arsenic exposure. For lung cancer, there was evidence of synergistic
 9    effects between arsenic exposure and smoking history. Strengths of this study include data on
10    smoking history, age, and gender, and an examination of the cohort by three arsenic exposure
11    categories. Weaknesses, however, include the lack of detailed arsenic intake information, a
12    small study population, as well as possible misclassification and recall bias pertaining to
13    smoking history.
      4.1.3.  South America
14          Hopenhayn-Rich et al. (1996a) used an ecological study design to investigate bladder
15    cancer mortality for the years 1986 through 1991 in the province of Cordoba, Argentina, using
16    rates for all of Argentina as the standard for comparison.  The study  compiled arsenic
17    measurements from a major water survey performed more than 50 years earlier.  Using these
18    earlier arsenic data, a crude estimate of exposure was made.  The data were matched to the
19    population listings from the national census bureau. This study grouped counties into three
20    defined arsenic exposure categories: low, medium, and high (groups were defined based on the
21    location of counties and the concentrations were only provided for the high group, which had a
22    mean arsenic level of 178 ppb). In the absence of smoking data for each county, mortality from
23    chronic obstructive pulmonary disease (COPD) was used as a surrogate.  SMRs for bladder
24    cancer were higher in counties with known elevated levels of arsenic exposure through drinking
25    water. The SMRs (95% CI) for corresponding arsenic exposure categories were 0.80 (0.66-
26    0.96),  1.42(1.14-1.74), and 2.14 (1.78-2.53) for males, and 1.21 (0.85-1.64),  1.58(1.01-2.35),
27    and 1.82 (1.19-2.64) for females, respectively.  Significant trends were noted in both males and
28    females.
29          Results of this study showed a dose-response relationship between arsenic exposure from
30    drinking water and bladder cancer in spite of the limitations inherent from the ecologic design.
31    Argentina has one of the world's highest rates of per capita beef consumption. The high-arsenic
32    region of Cordoba is an important agricultural and beef-producing area, and animal protein is
33    considered to be one of the basic foods of the population. This is important because protein
34    deficiency in the Taiwanese population has been suggested to diminish their capacity to detoxify
35    arsenic.  The similar findings between the two populations, regardless of genetic and dietary
36    differences, strengthens the link between arsenic exposure and bladder cancer.  Strengths of the
                                                47        DRAFT—DO NOT CITE OR QUOTE

-------
 1    study include the adjustment for age and gender, the use of stomach cancer as a non-arsenic-
 2    induced comparison, and that the analysis was restricted to rural counties to limit confounders.
 3    The lack of individual smoking history (mortality from COPD was used as a surrogate for
 4    smoking), the lack of arsenic measurements in low and medium groups, and the lack of
 5    individual arsenic exposure data (ecological study) are important potential weaknesses of this
 6    study.
 7           To investigate dose-response relationships between arsenic exposure from drinking water
 8    and cancer mortality, Hopenhayn-Rich et al. (1998) conducted an ecological  study in Cordoba,
 9    Argentina.  Cancer mortality from the lung, kidney, liver, and skin during the 1986-1991 period
10    in 26 counties of Cordoba were studied. This investigation expanded the analysis of the authors'
11    previous study (Hopenhayn-Rich et al.,  1996a),  which only examined bladder cancer in Cordoba.
12    Counties were grouped into low, medium, and high arsenic exposure categories based on arsenic
13    exposure data taken from Hopenhayn-Rich et al. (1996a).  In the absence of smoking data for
14    each county, mortality from COPD was used as a surrogate.  SMRs were calculated using all of
15    Argentina as the reference population.  Hopenhayn-Rich et al. (1998) found increasing trends for
16    kidney  and lung cancer mortality with increasing arsenic exposure (i.e., low,  medium, high) as
17    follows: male kidney cancer SMRs=0.87 (95% CI: 0.66-1.10), 1.33 (95% CI: 1.02-1.68), and
18    1.57 (95% CI: 1.17-2.04); female kidney cancer SMRs=1.00 (95% CI: 0.71-1.37), 1.36 (95% CI:
19    0.94-1.89), and  1.81 (95% CI: 1.19-2.64); male lung cancer SMRs=0.92 (95% CI: 0.85-0.98),
20    1.54 (95% CI: 1.44-1.64), and 1.77 (95% CI: 1.63-1.90); and female lung cancer SMRs=1.24
21    (95% CI: 1.06-1.42), 1.34 (95% CI:  1.12-1.58), and 2.16 (95% CI: 1.83-2.52), respectively
22    (p < 0.001 in trend test). These findings were similar to the previously reported bladder cancer
23    results.  Additionally, the Hopenhayn-Rich et al. (1998) study showed a weakly positive trend
24    for liver cancer,  with SMRs being significantly increased even in the lowest exposure category.
25    Skin cancer mortality was elevated only for females in the highest arsenic exposure group, while
26    males showed an increase in mortality only in the lowest exposure group. The results add to the
27    evidence that arsenic ingestion through drinking water increases the risk of lung and kidney
28    cancers. The association between arsenic and mortality from liver and skin cancers was not as
29    clear. Risk analyses were restricted to rural Cordoba counties to limit confounders and to
30    account for cancer diagnosis and detection bias.  Strengths and weaknesses are the same as those
31    observed for Hopenhayn-Rich et al. (1996a).
32           Smith et  al. (1998), using an ecological design, studied cancer mortality in a population
33    of approximately 400,000 people exposed to high arsenic levels in drinking water in past years in
34    Region II of northern Chile.  Arsenic concentrations in drinking water from 1950 to 1996 were
35    available. The population-weighted  average arsenic levels reached 570 ppb between 1955 and
36    1969, but decreased to less than 100  ppb by 1980.  SMRs were calculated for the years 1989 to
37    1993, and increased SMRs were identified for bladder, kidney, lung, and skin cancers. Bladder

                                                48       DRAFT—DO NOT CITE OR QUOTE

-------
 1    cancer mortality was the most elevated (female SMR=8.2, 95% CI: 6.3-10.5; male SMR=6.0,
 2    95% CI: 4.8-7.4).  Lung cancer mortality was likewise significantly elevated (female SMR=3.1,
 3    95% CI: 2.7-3.7; male SMR=3.8, 95% CI: 3.5-4.1).  Smoking survey data and mortality rates
 4    from COPD provided evidence that smoking did not contribute to the increased mortality from
 5    these cancers. These results provide additional evidence that ingestion of inorganic arsenic in
 6    drinking water can lead to increases in cancers of the bladder and lung.  Smith et al. (1998)
 7    estimated that approximately 7% of all deaths in individuals more than 30 years old might be
 8    attributable to arsenic exposure.  Strengths of the study are the large size of the study population,
 9    the adjustment of SMRs by age  and gender, and the use of Chilean national data for comparison.
10    Weaknesses include that arsenic levels were not available at the individual source level, dose-
11    response information was not provided, and only limited individual smoking history information
12    was available (i.e., participants were asked if they had smoked cigarettes over a 1-month period
13    in 1990).
14          In a case-control study, Ferreccio et al. (2000) investigated the association between lung
15    cancer and arsenic in drinking water by comparing patients diagnosed with lung cancer (1994-
16    1996; 152 cases) with frequency-matched hospital  controls (419 controls). Using a full-logistic
17    regression model, a clear trend in lung cancer ORs was observed with increasing concentration
18    of arsenic in drinking water: 10-29 ppb arsenic, OR:  1.6 (95% CI: 0.5-5.3), 30-49 ppb arsenic,
19    OR: 3.9 (95% CL1.2-12.3), 50-199 ppb arsenic, OR: 5.2 (95% CI: 2.3-11.7), and 200-400 ppb,
20    OR: 8.9 (95% CI: 4.0-19.6). Evidence of synergistic effects between arsenic in drinking water
21    and cigarette smoking history was much greater than expected, as the OR for lung cancer was
22    32.0 (95% CI: 7.2-198.0) among smokers exposed to more than 200 ppb.  In comparison, an  OR
23    of 8.0 was observed for those who never smoked but were in the highest arsenic category, and an
24    OR of 6.1 was observed for smokers in the lowest arsenic category. Based on these results, the
25    effect was considered synergistic because an OR of 13.1 was expected if the  effect was additive.
26    This study provided strong evidence that ingestion of inorganic arsenic through drinking water
27    is associated with lung cancer.  ORs for the full-analysis model were adjusted for age, gender,
28    cumulative lifetime cigarette smoking, working in copper smelting, and socioeconomic status;
29    this is considered a study strength. The fact that more controls were obtained from Antofagasta
30    than from the lower-exposure cities of Arica and Iquique, which could lead to an improper
31    (lower) estimation of risk, is considered a study limitation.
32          Bates et al. (2004) recognized that epidemiologic studies had found an association
33    between increased bladder cancer risk and high levels of arsenic in drinking water; however,
34    little information was found concerning cancer risks at lower concentrations.  It also was
35    recognized that ecologic studies in Argentina had found increased bladder cancer mortality in
36    Cordoba Province, where some wells were contaminated with moderate arsenic concentrations.
37    Therefore, Bates et al. (2004) decided to use a population-based bladder cancer case-control

                                                49        DRAFT—DO NOT CITE OR QUOTE

-------
 1    study during 1996-2000 in two Cordoba counties and recruited 114 case-control pairs, matched
 2    by age, sex, and county of residence over the past 40 years.  Three arsenic exposure metrics
 3    based on questionnaire and water sampling data were used: average arsenic concentration in
 4    domestic water, arsenic concentration adjusted to fluid intake, and reported years of well water
 5    consumption. Statistical analyses showed no evidence of an association of bladder cancer with
 6    arsenic exposure estimates based on arsenic concentrations in drinking water. Additional time-
 7    trend analyses, however, did suggest that the use of arsenic-contaminated well water at least 50
 8    years prior to the study was associated with increased bladder cancer risk. This positive
 9    association was limited to people who had ever smoked (OR=2.5, 95% CI:  1.1-5.5 for the time
10    period 51-70 years before the study interview).  Bates et al. (2004) suggested that it could not be
11    excluded that these associations were based on chance.
12           The results of this study suggest a decreased bladder cancer risk for arsenic exposure than
13    had been predicted from other studies. The results of the Bates et al. (2004) study did add to the
14    evidence that the latency for arsenic-induced bladder cancers may be longer than previously
15    thought and that increased lengths of follow-up for studies may be required to accurately
16    measure the induced risk.  Strengths include that potential confounders (age, gender, smoking
17    history, and residence county) were controlled for in the analysis. However, weaknesses related
18    to the lack of a cancer registry, arsenic exposure misclassification, and recall and selection bias
19    exist.  Selection bias may have occurred, as the controls had a significantly  lower rate of
20    participation than cases. Additional selection  bias may have occurred with the selection of cases
21    from the tumor registry. An additional weakness is that other harmful exposures (including
22    arsenic exposure through food)  were not measured.
23           Using a cohort study design, Smith et al. (2006) investigated lung cancer, bronchiectasis,
24    and COPD mortality rates in Antofagasta, Chile, from 1989 through 2000 and compared these
25    rates to the rest of Chile. Study subjects (30-49 years old at time of death) were selected
26    primarily from those born during or just prior  to the peak in the arsenic exposure period.  Results
27    show a lung cancer SMR of 7.0 (95% CI: 5.4-8.9, p < 0.001) for the cohort born just before the
28    peak exposure period (i.e., from 1950 through 1957), and, therefore, were exposed to arsenic
29    during their childhood.  For those cases born between 1958 and 1971 (i.e., the high-exposure
30    period), a lung cancer SMR of 6.1 (95% CI: 3.5-9.9, p < 0.001) was estimated; this group was
31    probable exposed to arsenic in utero and early childhood. These findings suggest that exposure
32    to arsenic in drinking water during early childhood or in utero has pronounced pulmonary effects
33    greatly increasing subsequent mortality in young adults from malignant lung disease.  The study
34    concluded that the observed effects are most probably due to arsenic in water, even though
35    possible effect-dilution occurred as the result of in-migration of those from  other regions of
36    Chile.  A strength of the study was the extensive documentation of drinking water arsenic levels
37    in the Antofagasta water system. Weaknesses include that place of residence was determined

                                                50        DRAFT—DO NOT CITE OR QUOTE

-------
 1    from the death certificates, which relates to residence at the time of death, and the reliance on
 2    death certificates (potential diagnostic bias).  Smoking, although considered unlikely by Smith et
 3    al. (2006), is a potential confounder for this study.
 4          Marshall et al. (2007) conducted an ecological study to investigate lung and bladder
 5    cancer mortality from 1950 to 2000 in a region of Chile where drinking water was contaminated
 6    with arsenic (Region II), and in another region of Chile where arsenic was not an issue (Region
 7    V).  Elevated arsenic exposure through  drinking water began in Region II in 1958 and continued
 8    into the early 1970s. Mortality data tapes and mortality data from death certificates for the two
 9    regions for 1950 to 1970 identified 307,541 deaths from the two regions for 1971 to 2000.
10    Poisson regression models were used to compare Region II with Region V by identifying time
11    trends in rate ratios of mortality from lung and bladder cancers. Lung and bladder cancer
12    mortality rate ratios for Region II compared with Region V began to increase approximately 10
13    years after high arsenic exposures commenced and continued  to rise, peaking between 1986 and
14    1997.  The peak lung cancer mortality rate ratios for women and men were 3.26 (95% CI: 2.50-
15    4.23)and3.61 (95% CI: 3.13-4.16), respectively.  The peak bladder cancer rate ratios for
16    women and men were 13.8 (95% CI:  7.74-24.5) and 6.10 (95% CI: 3.97-9.39), respectively.
17    Together, lung and bladder cancer mortality rates in Region II were highest from 1992 to 1994,
18    with mortality rates of 50/100,000 for women and 153/100,000 for men compared with
19    19/100,000 and 54/100,000, respectively, in Region V. The long latency for lung and bladder
20    cancer mortality continued to have a residual effect through the late 1990s, even though there
21    was a significant decrease in arsenic exposure through drinking water more than 25 years earlier.
22    Strengths of the investigation include the large study population, the availability of past
23    exposure data, and that potential  confounders of age, gender, and smoking history were
24    controlled for in the analysis. However, weaknesses include the inability to account for
25    migration, the ecologic design  (i.e., lack of individual exposure data) and lack of information
26    concerning occupation.
27          Yuan et al. (2007)  investigated mortality from 1950 to 2000 using an ecological study
28    design in the arsenic-exposed Region II of Chile and the unexposed population from Region V.
29    Before 1958, the drinking  water in Region II contained approximately 90 ppb of arsenic.  In
30    1958, it became necessary to supplement the  Region II water  supply using rivers that had an
31    average arsenic concentration of 870  ppb. After the installation of an improved water treatment
32    operation in the early  1970s, the arsenic concentrations in the  Region II water supply dropped
33    sharply (<10 ppb).  While  acute myocardial infarction (AMI)  mortality was the predominant
34    cause of excess deaths during and immediately after the high-exposure period, due to the longer
35    latency of cancer, excess deaths from lung and bladder cancer became predominated years later.
36    Yuan et al. (2007) concluded that after a 15- to 20-year lag period following initial exposure to
37    significantly  elevated levels  of arsenic from drinking water (1958-1970), mortality from bladder

                                                51        DRAFT—DO NOT CITE OR QUOTE

-------
 1    and lung cancer surpassed other causes of mortality. Strengths of the study included known
 2    arsenic concentrations and the large study population. In addition, to ensure appropriate
 3    selection of a control population, preliminary investigations were conducted to compare regional
 4    income, smoking history, and availability and quality of death certificate information. The major
 5    weakness of the study was its ecological study design (i.e., lack of individual arsenic exposure).
 6    In addition, potential confounders (i.e., smoking histories, diet, and exercise) were not examined
 7    on an individual basis, but were compared on a regional basis.
      4.1.4. North America (United States and Mexico)
 8           Bates et al. (1995), in a case-control study, used data obtained from Utah respondents for
 9    the 1978 National Bladder Cancer Study to examine the potential relationship between bladder
10    cancer in a U.S. population exposed to measurable levels of arsenic in drinking water. Arsenic
11    levels in drinking water were lower than those in Asian and South American studies.  A total of
12    117 cases and 266 controls were selected as participants for this study.  Restricting subjects to
13    those who had lived in study areas for at least half of their lives, the number of subjects still
14    eligible was 71 cases and 160 controls. Arsenic exposures ranged from 0.5 to 160 ppb (mean,
15    5.0 ppb).  Two measurements of arsenic exposure were used. One measure used was the total
16    CAE and the other was the arsenic concentration ingested adjusted for individual water
17    consumption. Bates et al. (1995) found no association between bladder cancer and either arsenic
18    exposure measure.  However, among smokers, positive trends in cancer risk were found for
19    arsenic exposures between 30 to 39 years prior to cancer diagnosis. The risk estimates were
20    stronger for the drinking water measure that estimated the ingested arsenic concentration than
21    the CAE.  The risk estimates obtained, however, were higher than predicted based on the results
22    of the Taiwanese studies, which raised concerns by Bates et al. (1995) regarding confounders,
23    bias,  and chance.
24           The data from this study raised the potential that smoking contributes to the increased
25    effect of arsenic on the risk of bladder cancer. Potential confounders included in the logistic
26    models were gender, age, smoking status, years of exposure to chlorinated water, history of
27    bladder infection, and the highest educational  level attained.  Strengths of the Bates et al. (1995)
28    investigation are that these confounders were controlled for; occupation, population size of
29    geographic area, and urbanization were addressed in the analysis; and cases were histologically
30    confirmed.  Potential weaknesses of the study are the small size of the study population, the fact
31    that the subjects were mostly male and the data on females were inadequate, and that arsenic
32    exposure levels were based on measurements close to the time that cases were diagnosed.  Due
33    to the low concentration in the water, the  lack of measurement of arsenic in the food was a
34    limitation of this study.  Although the purpose of the Bates et al.  (1995) study was to compare
35    low-level arsenic exposure and bladder cancer with the results from the Taiwanese population,

                                                52        DRAFT—DO NOT CITE OR QUOTE

-------
 1    the results cannot be interpreted without consideration of potential confounders and bias
 2    resulting from the retrospective study design.
 3          Employing a retrospective cohort mortality investigation of residents from Millard
 4    County, Utah, Lewis et al. (1999) examined the relationship between arsenic exposure from
 5    drinking water and mortality outcome. Median drinking water arsenic concentrations for
 6    selected study areas ranged from 14 to 166 ppb.  Drinking water samples were obtained from
 7    public and private sources and were collected and analyzed under supervision of the State of
 8    Utah Department of Environmental Quality, Division of Drinking Water. Cohort members were
 9    assembled using historical documents made available by the Church of Jesus Christ of Latter-
10    Day Saints.  Residential histories and median drinking water arsenic concentration were used to
11    construct a matrix for CAE.  Previous drinking water arsenic concentrations (from 1964 forward)
12    were obtained from historical records of arsenic measurements maintained by the state of Utah.
13    Without regard to specific exposure levels, statistically significant increases in mortality from
14    prostate cancer (SMR=1.45, 95% CI: 1.07-1.91) among cohort males was observed. Non-
15    significant increases in mortality for males were observed in cancer of the kidney (SMR=1.75,
16    95% CI: 0.80-3.32). There was no increased risk for cancer of the bladder  and other urinary
17    organs (SMR=0.42, 95% CI: 0.08-1.22) in males.  Among cohort females, no statistically
18    significant increase in mortality was observed. Females did, however, exhibit non-significant
19    increases in mortality from kidney cancer (SMR=1.60, 95% CI: 0.44-4.11) and melanoma of the
20    skin (SMR=1.82, 95% CI: 0.50-4.66). Female cancer of the bladder and other urinary organs
21    (SMR=0.81, 95% CI: 0.10-2.93) was not increased. Risk analysis using low-, medium-, and
22    high-arsenic exposure groups did not provide any clear indication of a dose-response for prostate
23    cancer. Confounding was not considered to be a significant concern by Lewis et al. (1999).
24    Exposure to other arsenic sources (food- or airborne), however, may have contributed to the total
25    exposure potential of this population. Strengths of the study included the cohort study design.
26    In this design type, the exposure precedes the effect being measured so a variety of effects from
27    a single type of exposure can be considered. The study population was mostly rural and
28    Mormon (low tobacco and alcohol use).  In addition, NRC (2001) and EPA (U.S. EPA, 2001)
29    identified that the Lewis et al. (1999) study was not powerful enough to estimate risk.
30          To address the association between skin cancer and arsenic exposure in drinking water,
31    Karagas et al.  (2001) used data collected on 587 basal cell and 284 squamous cell skin cancer
32    cases and 524 controls.  Cases and controls were interviewed as part of a case-control study
33    conducted in New Hampshire (and bordering regions) between 1993 and 1996. Arsenic
34    exposure levels were determined using toenail clippings. The ORs for SCC (range 0.93-1.10)
35    and BCC (range 0.72-1.06) were not significant and near unity (1.0) in all but the highest
36    category (0.345-0.81 ug/g).  For cases with significantly elevated toenail arsenic concentrations,
37    the adjusted ORs were 2.07 (95% CI 0.92-4.66) for SCC and 1.44 (95% CI: 0.74-2.81) for BCC,

                                                53       DRAFT—DO NOT  CITE OR QUOTE

-------
 1    compared with those with concentrations at or below the median.  Since the risks of SCC and/or
 2    BCC were not elevated in the range of toenail arsenic concentrations detected in most study
 3    subjects, the authors did not exclude the possibility of a dose-related increase at the highest
 4    levels of exposure.  Strengths include evaluating the effects of potential confounders such as age,
 5    gender, race, educational attainment, smoking status, skin reaction to first exposure to the sun,
 6    and history of radiotherapy.  Toenail arsenic concentrations can be considered a strength and a
 7    weakness.  They are a strength because they individualize the dose and could account for arsenic
 8    exposure from other sources (e.g., food), but they also could be considered a weakness because
 9    toenail arsenic is a biomarker of recent past exposure (covering a period of about one year
10    according to Cantor and Lubin, 2007).  Some confounding variables were not controlled for and
11    may have influenced the results.  The latency of arsenic-induced skin cancer is unknown and, as
12    a result, the follow-up period for this study may have been inadequate.
13          The identification of a potential leukemia cluster in Churchill County, Nevada, where
14    arsenic levels in water supplies are relatively high, prompted a study by Moore et al.  (2002).
15    Using an ecological study design, Moore et al. examined the incidence of childhood cancer
16    between 1979 and 1999 in all 17 Nevada counties. For analysis, arsenic exposures were grouped
17    into low (<10 ppb), medium (10-25 ppb), and high (35-90 ppb) population-weighted arsenic
18    levels based on the levels obtained from public drinking water.  SIRs for all childhood cancers
19    combined were 1.00 (95% CI:  0.94-1.06) for low-exposure, 0.72 (95% CI: 0.43-1.12) for
20    medium, and 1.25 (95% CI: 0.91-1.69) for high-exposure counties. Moore et al. (2002) found
21    no apparent relationship between the three arsenic levels and childhood leukemia with SIRs of
22    1.02 (95% CI: 0.90-1.15), 0.61 (95% CI: 0.12-1.79), and 0.86 (95% CI: 0.37-1.70) in the low,
23    medium, and high exposure categories, respectively.  No association was found for all childhood
24    cancers, excluding leukemia, with SIRs of 0.99 (95% CI: 0.92-1.07), 0.82 (95% CI: 0.47-1.33),
25    and 1.37 (95% CI: 0.96-1.91), respectively. There was, however, an excess for bone cancers in
26    5- to  9-year-olds and 10- to 14-year-olds and an excess in cancer (primarily lymphomas) in 15-
27    to 19-year-old young adults in the high-exposure group. The findings in this  study showed no
28    increase in leukemia risk at the concentrations of arsenic identified and categorized in the water.
29    Although the results did not eliminate the possibility for increased risks for non-leukemia
30    childhood cancers, there is no reason to suspect that the exposures to low levels of arsenic in the
31    small study group are responsible.  Strengths of the study are that the analysis of the data was
32    stratified by age, the study was a low-level arsenic exposure study, and the findings were
33    reported at different arsenic concentrations. Weaknesses of the study include the small study
34    size, the potential for exposure misclassification, and the limitations of the ecological study
35    design.
36          Steinmaus et al. (2003) used a case-control study to evaluate the effects of arsenic
37    ingestion on bladder cancer risk in  seven counties in the western United States.  These counties

                                                 54        DRAFT—DO NOT CITE OR QUOTE

-------
 1    contain the largest populations historically exposed to arsenic via drinking water at levels of
 2    approximately 100 ppb. These populations gave Steinmaus et al. the opportunity to critically
 3    evaluate the effects of relatively low-level arsenic exposure on bladder cancer incidence.
 4    Incident bladder cancer cases diagnosed between 1994 and 2000 were recruited based on
 5    information obtained from the Nevada Cancer Registry and the Cancer Registry of Central
 6    California. Arsenic measurements for community-supplied drinking water within the study were
 7    provided by the Nevada State Health Division and  the California Department of Health Services.
 8    Over 7000 arsenic measurements were obtained. Individuals' data on water sources, water
 9    consumption patterns, smoking history, and other sociodemographic factors were obtained for
10    181 bladder cancer cases and 328 matched controls. There was no observed increased risk for
11    bladder cancer associated with intakes greater than 80 ug/day (OR=0.94, 95% CI: 0.56-1.57;
12    linear trend, p=0.48).  This observed OR was below the risk predicted based on higher arsenic
13    concentrations in drinking water studies from Taiwan. However, when the analysis focused
14    solely on previous smokers who had arsenic exposures greater than 80 ug/day (median 177
15    ug/day) for more than 40 years, the risk was significantly increased (OR=3.67, 95% CI: 1.43-
16    9.42; linear trend, p< 0.01).  These data provide evidence that smoking and ingesting arsenic at
17    elevated concentrations (i.e., greater than 100  ug/day) may result in an increased risk of bladder
18    cancer. A strength of the Steinmaus et al. (2003) study is the use of individual exposure level
19    data to examine low-dose drinking water arsenic exposure; however, the lack of arsenic exposure
20    from food is a study weakness due to the low levels of exposure through drinking water. In
21    addition, the use of cancer registries allowed for improved case identification. Potential
22    confounders adjusted for in the analysis included gender, age, smoking history, education,
23    occupation associated with elevated rates of bladder cancer, and income. However, bias as the
24    result of next-of-kin interviews may have influenced the exposure assessment. Arsenic
25    exposures from outside the study area also may have influenced the exposure assessment. In the
26    arsenic-exposed areas, the percentage of non-participants was 5% higher among cases than
27    controls. This difference probably means that more exposed cases were missed in analyses  of
28    recent exposure, biasing the OR toward the null.
29          There has been little research investigating  the link between arsenic and cutaneous
30    melanoma, although arsenic has been associated with increased risk of non-melanoma skin
31    cancer. Beane-Freeman et al. (2004) performed a case-control study to examine the potential
32    relationship between melanoma and environmental arsenic exposure in a cohort from Iowa.
33    Study participants included 368 cutaneous melanoma cases  (selected from 645 eligible cases)
34    and 373 colorectal cancer controls (selected from 732  eligible controls) diagnosed in  1999 or
35    2000, frequency-matched on gender and age. Participants completed a mailed survey and
36    submitted toenail clippings (obtained from 355 cases and 353 controls) for analysis of arsenic
37    content.  The authors identified an increased risk of melanoma in study cases with elevated

                                                55        DRAFT—DO NOT CITE OR QUOTE

-------
 1    toenail arsenic concentrations (OR=2.1, 95% CI: 1.4-3.3; p-trend=0.001) and an increased risk
 2    of melanoma with previous diagnosis of skin cancer and elevated toenail arsenic concentrations
 3    (OR=6.6, 95% CI: 2.0-21.9). There was a greater association between the toenail arsenic and
 4    melanoma when subjects reported a previous diagnosis of melanoma. Strengths of this
 5    investigation include the fact that the potential confounders (age, gender, skin color/skin type,
 6    prior history of sunburn, education, and occupational exposure) were controlled for in the
 7    analysis.  Ascertainment of cases and controls was accomplished by using the Iowa Cancer
 8    Registry, a Surveillance, Epidemiology, and End Results Program registry. This allowed newly
 9    diagnosed melanoma cases to be identified for a specific period and assured a greater degree of
10    certainty regarding the accuracy of diagnosis. Another strength is that toenail arsenic
11    concentrations individualize the exposure and account for arsenic exposure from other sources.
12    A limitation of this study was that toenail samples were collected 2 to 3 years after diagnosis and
13    therefore do not measure arsenic concentrations prior to diagnosis, resulting in possible exposure
14    misclassification.
15          Karagas et al. (2004) used a case-control study design to examine the effects of low-level
16    arsenic exposure on the incidence of bladder cancer in New Hampshire  (and bordering regions),
17    where levels above 10 ppb are commonly found in private wells.  The authors studied  383 cases
18    of transitional cell carcinoma of the bladder, diagnosed between July 1,  1994, and June 30, 1998,
19    and 641 general population controls.  Individual exposure to arsenic was determined through the
20    use of toenail  clippings. Karagas et al. (2004) found arsenic concentrations ranged from 0.014 to
21    2.484 ug/g among bladder cancer cases and 0.009 to 1.077 ug/g among  controls. When stratified
22    by smoking history, toenail arsenic concentrations were not associated with the risk of bladder
23    cancer. However, among smokers in the uppermost category of arsenic exposure, an elevated
24    OR for bladder cancer was observed (OR: 2.17, 95% CI: 0.92-5.11 for >0.330 ug/g compared to
25    <0.06 ug/g). When Karagas et al. (2004) stratified their analysis by duration of current water
26    system usage (<15 years and >15 years), an increased bladder cancer OR for people who ever
27    smoked with the highest category of arsenic exposure with less than 15 years of use was
28    identified (<15 years, OR=3.09, 95% CI: 0.80-11.96; >15 years, OR=1.86, 95% CI: 0.57-6.03).
29    These data suggest that ingestion of low to moderate arsenic levels may affect bladder cancer
30    incidence and that cigarette smoking may act as a co-carcinogen. Strengths of the study include
31    its use of a stratified analysis to evaluate the potential that an extended latency period was
32    required for bladder cancer development and its minimizing of misclassification by using
33    biomarkers. The following potential confounders were adjusted for: age, gender, race,
34    educational attainment,  smoking status, family history of bladder cancer, study  period, and
35    average number of glasses of tap water consumed per day. Toenail clippings were used in an
36    attempt to minimize misclassification. This, however, is a limitation because it only measures
37    recent past exposures. Limitations of the study were that misclassification at the lower

                                                 56        DRAFT—DO NOT CITE OR QUOTE

-------
 1    exposures was possible and that lifetime exposure could not be calculated since data from
 2    previous residences could not be determined.  In addition, there was limited data at extreme ends
 3    of exposure.
 4          The Lamm et al. (2004) ecological study investigated the association between arsenic
 5    exposure from drinking water and bladder cancer mortality in 133 counties in the United States.
 6    Caucasian male county-specific bladder cancer mortality data between 1950 and 1979 and
 7    county-specific ground water arsenic concentration data were obtained for counties solely
 8    dependent on ground water for their public drinking water supply.  Arsenic exposure was based
 9    on measurements for at least 5 wells for each county.  No arsenic-related increase in bladder
10    cancer mortality (SMR=0.94, 95% CI: 0.90-0.98) was identified (arsenic exposure range: 3-60
11    ppb) using stratified analysis and regression analyses.  These findings are consistent with other
12    previously published U.S. studies.  Strengths of the study include the large nationwide study
13    population, which included more than 75 million person-years of observation.  Weaknesses,
14    however, are the lack of available individual exposure data, the assumption that study
15    participants consumed only local drinking water, the assumption that available data were
16    representative of actual arsenic content in the water, that arsenic contribution from food sources
17    were not analyzed, and that the analysis did not directly adjust for smoking, urbanization, or
18    industrialization.
19          The Wisconsin Division of Public Health, in July 2000 through January 2002, conducted
20    a cross-sectional study in 19 rural Wisconsin townships concerning private drinking-water wells
21    and arsenic exposure (Knobeloch et al., 2006). Residents in these townships were asked to
22    collect well-water samples and  complete  a questionnaire regarding residential history,
23    consumption of drinking water, and family health. In Wisconsin, skin cancer is not reportable;
24    therefore, no skin cancer registry data were available.  During the study, 2,233  private wells
25    were tested, and 6,669 residents provided information on water consumption and health. Water
26    arsenic levels ranged from less than 1.0 to 3,100 ppb.  The median  arsenic level was 2.0 ppb.
27    Eighty percent of the wells had arsenic levels below 10 ppb, but 11% had an arsenic level of
28    above 20 ppb. Age-, gender-, and smoking-adjusted ORs of residents 35 years of age and older
29    who had consumed water with arsenic levels greater than 1.0 ppb for at least 10 years showed a
30    significant increase in individuals who reported skin cancer compared to those whose water
31    arsenic levels were less than 1.0 ppb (arsenic 1.0-9.9 ppb OR=1.81, 95% CI: 1.10-3.14).
32    Similarly, adults whose well-water reportedly contained arsenic concentrations greater than 10
33    ppb were significantly more likely to report skin cancer than those  whose water arsenic  levels
34    were less than 1.0 ppb  (OR=1.92, 95% CI: 1.01-3.68). Tobacco use also was associated with
35    higher rates of skin cancer and may—synergistically with arsenic exposure—affect the
36    development of skin cancer. Strengths of the study include: the large sample size, a history of
37    individual tobacco use, arsenic well water analysis for each household, an exposure duration of

                                                 57        DRAFT—DO NOT CITE OR QUOTE

-------
 1    at least 10 years in participants who consumed water from the tested wells, and the fact that the
 2    analysis controlled for age, gender, and tobacco use. Weaknesses include the following: skin
 3    cancers were self-reported and not confirmed by a medical records review, few people could
 4    provide information about specific types of cancer, potential bias could have resulted from the
 5    participating families being concerned about arsenic exposure, sun exposure and occupation
 6    were not controlled for in the analysis, and food sources of arsenic were not considered.
 7
 8          Meliker et al. (2007) performed an ecological study in a contiguous six-county study area
 9    of southeastern Michigan to investigate the relationship between moderate arsenic levels (10-
10    100 ppb) and selected disease outcomes. This region of southeastern Michigan was chosen
11    because it had moderately high arsenic concentrations in the ground water and low rates of
12    migration. The six counties had a population-weighted mean arsenic concentration of 11.00 ppb
13    and a population-weighted median of 7.58 ppb. In comparison, the remainder of Michigan has a
14    population-weighted mean of 2.98 ppb with a median of 1.27 ppb.   SMRs for cancers were not
15    significantly different from the age- and race-adjusted expected values for males or females for
16    the state of Michigan (SMR skin melanoma female=0.97, 95% CI: 0.73-1.27, melanoma
17    male=0.99, 95% CI: 0.79-1.22; SMR bladder female=0.98, 95% CI: 0.80-1.19, bladder
18    male=0.94, 0.82-1.08; SMR kidney female=1.00, 95% CI: 0.80-1.20, kidney male=1.06, 95%
19    CI: 0.91-1.22; SMR trachea, lung, bronchus female=1.02, 95% CI: 0.96-1.07, trachea, lung,
20    bronchus  male=1.02, 95% CI: 0.98-1.06). The only exception was cancer of the female
21    reproductive organs (SMR=1.11, 95% CI: 1.03-1.19). The potential explanations for the lack of
22    significant cancer findings were the relatively low level of arsenic in the ground water of
23    southeastern Michigan, which may be below the threshold for cancer induction and other
24    moderating factors that were not considered by this study (i.e., food as a source of arsenic
25    exposure). Strengths include that mortality rates, which were gathered from Michigan Resident
26    Death Files for a 20-year period, were stratified by gender, age, and race. Weaknesses include
27    the following: the ecological study design did not provide individual arsenic exposure data and
28    may not permit the detection of significant risk, there may have been differences in reporting and
29    classification of underlying causes of death, case migration occurred, preferential sampling was
30    conducted based on home owners' request, arsenic contribution from food was not measured,
31    and there  was a lack of information concerning smoking history and obesity.
      4.1.5.  China
32          Using an ecological study design, Lamm et al. (2007) conducted dermatological
33    examinations for 3,179 of the 3,228 (98.5%) residents of three villages (Zhi Ji Liang, Tie Men
34    Geng, and Hei He) in Huhhot, Inner Mongolia, with well  water arsenic levels that ranged from
35    undetectable (<10 ppb) to 2,000 ppb.  Individual water consumption histories were obtained for
36    this population, and arsenic levels were measured for 184 wells.  Arsenic exposures were
                                                58        DRAFT—DO NOT CITE OR QUOTE

-------
 1    summarized as the highest arsenic concentration (HAC) and CAE. Thirty-five percent of the
 2    study population had HAC of less than 50 ppb, 86% had HAC less than 150 ppb, and only 1% of
 3    the participants had HAC greater than 500 ppb. The proportion of females to males was similar
 4    in each of the three villages (female range 49%-50% and male range 50%-51%), and almost all
 5    study subjects identified themselves as being of Chinese (99.8%) rather than Mongolian (0.2%)
 6    origin.  The median age for all participants was 29 years; however, participants from Hei He
 7    tended to be older than those from the other two villages (55.0% older than 30  in Hei He, 42.4%
 8    in Zhi Ji Liang and Tie Men Geng). Participants (female or male) who reported occupations
 9    listed "student" or "farmer." None of the examinations revealed any evidence of BFD. Analyses
10    included frequency-weighted, simple  linear regression,  and most likely estimate models. Eight
11    people were found to have skin cancer.  In addition to skin cancer, these eight cases also had
12    both hyperkeratoses and dyspigmentation.  Skin cancer cases were only identified in those
13    participants with HAC exposures >150 ppb or whose CAE was less than 1,000 ppb-years. The
14    models showed a threshold of 122-150 ppb. Lamm et al. (2007) identified a general exposure-
15    prevalence pattern (higher prevalence for HAC exposure group) for skin disorders
16    (hyperkeratosis, dyspigmentation, and skin cancers).  Duration of water usage (arsenic
17    exposure), age, latency, and misclassification did not appear to markedly affect the analysis.
18    Strengths of the study include the large study population, the fact that HAC and CAE were used
19    in the analyses, and the fact that arsenic concentrations  were measured in 184 wells.
20    Confounders that were controlled for  included age, differences in cumulative arsenic dose, and
21    duration of exposure.  A confounder not adjusted for in the analysis was sun exposure.
22    Additional weaknesses are the ecological study design and the potential for recall or
23    misclassification bias resulting from the collection of arsenic exposure histories through
24    interviews.
      4.1.6. Finland
25          In a case-cohort study, Kurttio et al. (1999) examined the levels of arsenic in Finnish
26    water wells and their relationship to the risk of bladder  and kidney cancers. The study
27    population consisted of 61 bladder cancer cases and 49  kidney cancer cases diagnosed between
28    1981 and 1995, and a randomly selected age- and gender-adjusted reference cohort of 275
29    subjects.  Arsenic exposure was estimated for cancer cases and for the reference cohort for two
30    periods. The first period was from the third to ninth calendar years (the shorter latency period)
31    prior to either the cancer diagnosis or the respective year for referent cohort, while the other was
32    from the tenth or earlier calendar years (the longer latency period). Water specimens were
33    obtained from the wells used by the study cohort from 1967 to  1980. The  arsenic concentrations
34    in the wells of the control population  were low, with approximately 1% exceeding 10 ppb.
35    Bladder cancer was associated with arsenic concentration and daily dose during the third to ninth
36    calendar years prior to the cancer diagnosis. The risk ratios for arsenic exposure concentration
                                                59       DRAFT—DO NOT CITE OR QUOTE

-------
 1    categories 0.1-0.5 and >0.5 ppb relative to the category with <0.1 ppb were 1.53 (95% CI: 0.75-
 2    3.09) and 2.44 (95% CI: 1.11-5.37), respectively.  In spite of low levels of arsenic exposure,
 3    Kurttio et al. (1999) found evidence of a relationship between exposure to arsenic at the higher
 4    exposure level and bladder cancer risk. No association, however, was observed between arsenic
 5    exposure level and kidney cancer risk.  Strengths include the following: Finnish Cancer Registry
 6    records were accessible; Statistics Finland's 1985 Population Census file was used to identify
 7    areas in which less than 10% of the population used the municipal water supply; and age, gender,
 8    and smoking histories were accounted for in the risk ratio calculations. Possible weaknesses
 9    include misclassification and recall bias resulting from the study choosing to use water
10    consumption from the  1970s. In addition, because of the low arsenic concentrations, arsenic
11    exposure from other sources (e.g., food) could bias the results.
12           Michaud et al. (2004) used  a cohort (nested case-control) study design to investigate the
13    relationship between arsenic levels in toenail and bladder cancer risk among Finnish male
14    smokers aged 50-69 years who were participating in the Alpha-Tocopherol, Beta-Carotene
15    Cancer Prevention Study. Data for 280 incident bladder cancer cases, identified between 1985
16    and 1988 as well as April  1999, were available for analysis.  Controls (n = 293) were matched to
17    each case on the basis of age, toenail collection date, intervention group, and duration of
18    smoking. Logistic regression analyses were performed to estimate ORs. Arsenic toenail
19    concentrations in this Finnish study (cases and controls) ranged between 0.01 and 2.11 ug/g,
20    with one control outlier at 17.5 ug/g. Arsenic toenail concentrations were similar to those
21    reported in the United States (range: 0.02-17.7 ug/g). Men were categorized into quartiles based
22    on the distribution of arsenic among the controls (O.050, 0.050-0.105, 0.106-0.161, and
23    >0.161).  The study observed no significant relationship between arsenic concentration and
24    bladder cancer risk (OR=1.13, 95% CI: 0.70-1.81  for the highest vs. lowest quartile). Strengths
25    of the Michaud et al. (2004) study were that the  authors excluded toenail samples with non-
26    detectable arsenic levels greater than 0.09 ug/g,  in an attempt to avoid potential misclassification
27    of samples with high detection limits, and that they controlled for potential confounders in the
28    analysis (i.e., smoking history, beverage intake,  place of residence, toenail weight, smoking
29    cessation, smoking inhalation,  educational level, beverage intake, and place of residence). Cases
30    and controls were matched according to age, toenail collection date, intervention group (alpha
31    tocopherol and beta carotene),  and  smoking duration. Toenail arsenic concentrations are a
32    strength because they individualize the dose and could account for arsenic exposure from other
33    sources, but they also could be considered a weakness because toenail arsenic is a biomarker of
34    recent past exposure (covering about 1 year according to Cantor and Lubin, 2007).  Another
35    weakness of the study was that water consumption was not included in the total beverage intake
36    variable.
                                                 60       DRAFT—DO NOT CITE OR QUOTE

-------
      4.1.7. Denmark
 1          The Baastrup et al. (2008) cohort study was designed to determine whether exposure to
 2    low levels of arsenic in drinking-water in Denmark is associated with an increased risk for
 3    cancer. The study population was selected from participants in the prospective Danish cohort
 4    Diet, Cancer, and Health.  A cohort of 56,378 people (39,378 from Copenhagen and 17,000 from
 5    Aarhus) accepted an invitation to participate in the study. Cancer cases were identified in the
 6    Danish Cancer Registry, and the Danish civil registration system was used to trace residential
 7    addresses of the cohort members.  The study used a geographic information system to link
 8    residential addresses with water supply  areas and using this information estimated arsenic
 9    exposure by addresses. The average arsenic exposure for the cohort ranged between 0.05 and
10    25.3 ppb (mean =1.2 ppb) and was based on 4,954 measurements reported between 1987 and
11    2004 (the majority between 2002 and 2004).  The exposure was generally higher among Aarhus
12    participants than those enrolled in the Copenhagen area (Aarhus mean = 2.3 ppb, min = 0.09  ppb
13    and max=25.3 ppb; Copenhagen mean = 0.7 ppb, min = 0.05 ppb, and max=15.8 ppb).
14    Regression models were used to analyze possible relationships between arsenic and cancer. The
15    study found no significant association between arsenic exposure and risk for cancers of the lung,
16    bladder, liver, kidney, prostate, colon, or melanoma skin cancer.  The incidence rate ratio (IRR)
17    for non-melanoma skin cancer (0.88, 95% CI: 0.84-0.94) decreased with per ppb increases in the
18    time-weighted average exposure to arsenic.  The study did identify a significant increased risk
19    for breast cancer in association with time-weighted average exposure to arsenic (IRR=1.05, 95%
20    CI:  1.01-1.10). Strengths of the study include the large study population, the
21    socioeconomic/demographic similarities of the cohort, and the adjustment for potential
22    confounders (smoking, alcohol consumption, education, body mass index [BMI], daily intake of
23    fruits/vegetables, red meat, fat and dietary fiber, skin reaction to the sun, hormone replacement
24    therapy use, reproduction, occupation, and enrollment area). Weaknesses of the study include
25    the low arsenic levels in Danish drinking water, the lack of information on other sources of
26    arsenic exposure, and the inability to assess arsenic exposures before 1970, all resulting in
27    possible misclassification bias.
      4.1.8. Australia
28          Hinwood et al.  (1999) conducted an ecological study that investigated areas of Victoria,
29    Australia, with elevated environmental arsenic concentrations, areas with arsenic concentrations
30    in the soil of more than 100 mg/kg and/or drinking water arsenic  concentrations greater than  10
31    ppb, and the relationship with cancer incidence.  SIRs for cancer  were generated for 22 areas
32    between 1982 and 1991 using cancer registry data. In addition, SIRs for combined areas
33    according to environmental exposure  (high soil and/or high water arsenic concentrations, etc.)
34    were generated.  The SIRs (females and males together) for the combined 22 areas were
35    significantly elevated for all cancers (1.06, 95% CI: 1.03-1.09), melanoma (1.36, 95% CI: 1.24-
                                                61        DRAFT—DO NOT CITE OR QUOTE

-------
 1    1.48), chronic myeloid leukemia (1.54, 95% CI 1.13-2.10), breast cancer in females (1.10, 95%
 2    CI: 1.03-1.18), and prostate cancer in males (1.14, 95% CI:  1.05-1.23).  The SIR for kidney
 3    cancer (females and males combined) was 1.16 (95% CI: 0.98-1.37), and although elevated was
 4    not statistically significant. When stratified by exposure category, the SIR for prostate cancer
 5    was significant at 1.20 (95% CI: 1.06-1.36) for the high soil/high water category only. This
 6    result was likely confounded by misclassification (level of population exposure) and limited by
 7    low statistical power.  There was no significant dose-response relationship observed between
 8    drinking water and any individual  cancer. Strengths of the study include that water and soil
 9    arsenic levels were provided and a large area was examined.  Hinwood et al. (1999) recognized
10    that the results of this study were potentially confounded by a number of factors, including the
11    ecological study design, socioeconomic status, race, occupation, and urban versus rural status.
12    Due to the low concentrations in the drinking water, the lack of arsenic exposure from food
13    could cause exposure misclassification.

      4.2. PRECHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN
      ANIMALS—ORAL

      4.2.1. Prechronic and Chronic Studies
14          Wei et al. (1999, 2002) demonstrated that 10-week-old male F344/DuCrj rats (36/group)
15    administered 12.5, 50, or 200 ppm DMAV (a major metabolite of inorganic arsenic) in their
16    drinking water for 104 weeks had no effect on the morbidity, mortality, body weights,
17    hematology, or serum biochemistry. Reductions in electrolyte concentrations in the urine were
18    related to an increase in urinary volume resulting from increased water consumption in the 50-
19    and 200-ppm groups.  There was no difference in the urinary pH between control and treated
20    rats.
      4.2.2. Cancer Bioassays
21          Cancer bioassays with inorganic arsenic have generally obtained negative results with
22    mice, rats, hamsters, rabbits, beagles, and cynomologus monkeys  (for review see Kitchin, 2001;
23    NRC, 1999). However, the following studies have observed increases in tumors in animals
24    exposed to arsenic species.
      4.2.2.1. Mice—Transplacental
25          Timed pregnant female C3H mice (10/group) were administered 0 (control), 42.5, or 85
26    ppm As111 in their drinking water ad libitum from day 8 to day 18 of gestation (Waalkes et al.,
27    2003). Strain and doses used in the experiment were determined through preliminary short-term
28    testing that determined C3H mice to be the most sensitive to arsenic toxicity of the three  strains
29    tested (i.e., C3H, C57BL/6NCr, and B6C3Fl/NCr), and the preliminary test indicated that a dose
30    of 100 ppm was unpalatable and resulted in approximately 10% reduced growth in the offspring.

                                                62       DRAFT—DO NOT CITE OR QUOTE

-------
 1    The doses used in this study did not affect maternal water consumption or body weight in the
 2    dams.  It was estimated that the pregnant females consumed 9.55 to 19.13 mg arsenic/kg-day, for
 3    a total dose of 95.6 to 191.3 mg arsenic/kg.
 4          Offspring were weaned at 4 weeks and received no additional exposure to arsenic. Male
 5    and female offspring (25/sex/group) were observed for the next 74 or 90 weeks, respectively.
 6    Males were sacrificed at 74 weeks due to high mortality in the high-dose group beginning at 52
 7    weeks. Both the 42.5- and 85-ppm males had a significant increase in the incidence of HCC
 8    (12.5% in the control group versus 38.1% in the 42.5-ppm group and 60.9% in the 85-ppm
 9    group) and adrenal cortical tumors (37.5% in the control group versus 66.6% in the 42.5-ppm
10    group  and 91.3% in the 85-ppm group), which followed a significant (p<0.001), dose-related
11    trend.  In addition, the 85-ppm group had a significant increase in the multiplicity (tumor/mouse)
12    for both HCC (0.13, 0.42, and 1.30, respectively) and adrenal tumors (0.71, 1.10, and 1.57,
13    respectively), which also had a significant (p<0.02), dose-related trend. Although there were no
14    differences in the incidence of hepatocellular adenomas in males, the multiplicity of
15    hepatocellular adenomas (0.71, 1.43, and 3.61, respectively) followed a significant (p < 0.0001),
16    dose-related trend.
17          Males and females had an increase in lung tumors (8.0%, 13.0%, and 25.0%,
18    respectively,  in females; 0%, 0%, and 13.0%, respectively, in males), which followed a
19    significant (p<0.03), dose-response trend.  In addition, females had increases in the incidence of
20    benign ovarian tumors, which reached statistical significance in the 85-ppm group. Although a
21    significant increase was  not observed in malignant ovarian tumors, the total incidence (benign
22    plus malignant) of ovarian tumors was significant in the 85-ppm group and followed a
23    significant (p=0.015), dose-related trend (8% in the control group versus 26% in the 42.5-ppm
24    group  and 37.5% in the 85-ppm group).  There was an increase in uterine tumors that was not
25    significant and did not follow a dose-response trend, but was accompanied by a significant
26    (p=0.0019), dose-related increase in hyperplasia occurring at both doses. Females also had  a
27    dose-related increase in hyperplasia of the oviduct. The number of both tumor-bearing and
28    malignant tumor-bearing males was significantly increased in both dose groups and followed a
29    significant (p=0.0006 and 0.0001, respectively), dose-related trend. Female animals had a slight
30    increase in the number of tumors, which did not reach statistical significance and did not appear
31    to be dose-related.  The number of females bearing malignant tumors was significantly  increased
32    for both dose groups, but not in a dose-dependent  manner.
33          Waalkes et al. (2004a) followed the same procedure (except that offspring were observed
34    for 104 weeks), but exposed 25 male and 25  female offspring from each exposure group (0, 42.5,
35    or 85 ppm in the drinking water from gestational days 8 to 18 with no additional exposure after
36    birth) to acetone or 12-O-tetradecanoyl phorbol-13-acetate (TPA; 2 ug/0.1 mL in acetone) twice
37    a week—via  a shaved area of dorsal skin—for 21 weeks after weaning in an attempt to  promote

                                                63       DRAFT—DO NOT CITE OR QUOTE

-------
 1    skin tumors.  However, very few skin lesions occurred and were not associated with arsenic
 2    exposure either in the absence or presence of TPA. As was noted in Waalkes et al. (2003), there
 3    was a dose-dependent increase in the incidence and/or multiplicity of hepatocellular adenomas
 4    and carcinomas in treated males, both in the absence and presence of TPA. In the absence of
 5    TPA, the incidence of adenomas was 41.7%, 52.2%, and 90.5% for the 0-, 42.5-, and 85-ppm
 6    exposure groups, respectively; the incidence of carcinomas was 12.5%, 34.8%, and 47.6%,
 7    respectively; total incidence was 50%, 60.9%, and 90.5%, respectively; and multiplicity was
 8    0.75, 1.87, and 2.14,  respectively.  In the presence of TPA, the incidence of adenomas was
 9    34.8%, 52.2%, and 76.2% for the 0-, 42.5-, and 85-ppm exposure groups, respectively; the
10    incidence of carcinomas was 8.7%, 26.0%, and 33.3%, respectively; total incidence was 39.1%,
11    65.2%, and 85.7%, respectively; and multiplicity was 0.61, 1.44, and 2.14, respectively.  A
12    statistically significant increase was noted at 85 ppm. Arsenic only caused a dose-dependent
13    increase in hepatocellular adenomas and carcinomas in the presence of TPA in females
14    (adenomas: 8.3%, 18.2%, and 28.6% for the 0-, 42.5-, and 85-ppm exposure groups with TPA
15    exposure, respectively; carcinomas: 4.2%, 9.1%,  and 19.0%, respectively; total incidence:  12.5,
16    27.3, and 38.1%, respectively; multiplicity: 0.13, 0.32, and 0.71, respectively), with a
17    statistically significant increase in total incidence and multiplicity for the  85-ppm group.
18          There also was an increase in ovarian adenomas in treated female  offspring regardless of
19    whether they were treated with TPA (0%, 22.7%, 19.0%, respectively) or acetone (0%, 17.4%,
20    and 19.0%, respectively). There was no effect on the incidence of ovarian carcinomas. This was
21    accompanied by increases in the incidence of uterine epithelial hyperplasia (cystic) and total
22    uterine proliferative lesions, which increased in severity with dose. There also was a dose-
23    dependent increase in oviduct hyperplasia.  Male offspring exposed to arsenic had an increase in
24    the incidence and multiplicity of cortical adenomas of the adrenal glands. The increases were
25    statistically significant for both arsenic exposure  groups, but were only related to dose in the
26    absence  of TPA (p=0.020).  Incidences were as follows: 37.5%, 65.2%, and 71.4% for the 0-,
27    42.5-, and 85-ppm dose groups, respectively, in the absence of TPA and 30.4%, 65.2%, and
28    57.1%, respectively,  with TPA treatment.  Multiplicities also were statistically significantly
29    increased in arsenic-exposed male offspring with a significant dose-dependent trend both in the
30    absence  (0.58, 2.13, and 2.19, respectively; p=0.0014) or presence (0.54,  1.65, and 1.62,
31    respectively; p=0.016) of TPA.
32          Lung adenomas were increased in a dose-dependent manner in females exposed to TPA
33    (4.2%, 9.1%, and 28.%, respectively; p=0.018), but not in the absence of TPA (4.2%, 8.7%, and
34    9.5%, respectively; not significant). Males only had a statistically significant increase (5-fold
35    increase) in lung adenomas  in the 42.5-ppm group exposed to TPA.
36          A statistically significant increase in the multiplicity of all tumors in males  (with or
37    without TPA) was observed after arsenic exposure, but was not dependent on dose. Although

                                                64       DRAFT—DO NOT CITE OR QUOTE

-------
 1    females also had an increase in the multiplicity of all tumors, the only statistically significant
 2    increase occurred in the 85-ppm group exposed to TPA. The increase in females exposed to
 3    TPA also appeared to be dose-dependent. The statistically significant increase observed in the
 4    multiplicity of malignant tumors in males was greater in the absence of TPA, but was dose-
 5    dependent in the presence of TPA. In females, there was also an increase in the multiplicity of
 6    malignant tumors in arsenic treated mice (regardless of TPA exposure), but the results did not
 7    reach statistical significance, nor were they dose-dependent.
 8          Waalkes et al. (2006a) used female CD1 mice, which have a low rate of spontaneous
 9    tumors. Thirty-five percent (12/34) of female offspring receiving 85 ppm of As111 via the dams'
10    drinking water on gestational days 8 to 18 developed urogenital tumors, with 9% being
11    malignant compared to 0% in the controls.
      4.2.2.2. Rat—Oral
12          Soffritti et al. (2006) administered male and female Sprague-Dawley rats 0, 50, 100, or
13    200 mg/L (i.e., ppm) of sodium arsenite via the drinking water for 104 weeks.  There was a
14    consistent dose-dependent decrease in water and food consumption accompanied by a dose-
15    related decrease in body weight (there was no difference in body weight in females administered
16    50 mg/L).  There was only a slight decrease in survival in male rats administered 100 or 200
17    mg/L beginning at 40 weeks of age. Females only had a decrease in survival rate after 104
18    weeks of age.  Males and females administered 100 mg/L had an increase in the number of
19    tumor-bearing animals and in the number of tumors. Although there is no dose-related trends in
20    tumors, there were sporadic benign and malignant tumors of the lung, kidney, and bladder
21    observed in treated rats that are extremely rare in the authors' extensive historical controls.
22    These tumors consisted of adenomas and carcinomas of the lung, adenomas and carcinomas of
23    the kidney, papillomas and one carcinoma of the renal pelvis transitional cell epithelium, and one
24    carcinoma of the bladder transitional cell  epithelium.
25          Wei et al. (1999 and 2002) demonstrated that 10-week-old male F344/DuCrj rats
26    (36/group) administered 50  or 200 ppm DMAV in their drinking water for 104 weeks developed
27    bladder tumors (mainly carcinomas) and papillary or nodular hyperplasia in a dose-dependent
28    manner.  Controls and rats administered 12.5 ppm did not develop any bladder tumors or
29    hyperplasia. There was a significant (p < 0.05) increase in bromodeoxyuridine (BrdU) labeling
30    of morphologically normal epithelium of the bladder in the 50- and 200-ppm groups (Wei et al.,
31    2002). There was no significant increase in any other tumor type related to DMAV treatment.
32    There appeared to be a dose-related increase in subcutis fibromas (i.e., 4% in controls, 12% in
33    the 12.5-ppm group, and 16% in both the 50- and 200-ppm groups). Data indicate that multiple
34    genes are involved in the stages of DMAv-induced urinary bladder tumors.  Wei et al. (2002)
35    further indicate that reactive oxygen species (ROS) may play an important role during the early
3 6    stages of DMA carcinogenesi s.
                                                65       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Shen et al. (2003) administered TMAO, an organic metabolite of inorganic As, to male
 2   F344 rats for 2 years via their drinking water at concentrations of 0, 50, or 200 ppm.  Total
 3   intakes were estimated to be 0,  638, and 2475 mg/kg, respectively.  From 87 weeks of treatment
 4   on, there was an increase in the incidence and multiplicity of hepatocellular adenomas in rats
 5   sacrificed or dead. Incidences of 14.3%, 23.8%, and 35.6%, respectively, were reported. The
 6   respective multiplicities were 0.21, 0.33, and 0.53.  The results were statistically significant in
 7   the 200-ppm dose group.
     4.2.2.3. Other
 8          Transgenic models also have been developed to examine arsenic carcinogenesis. Arsenic
 9   exposure (200 ppm sodium arsenite in drinking water for 4 weeks)  in Tg. AC transgenic mice
10   containing activated H-ras did not induce skin tumors alone; however, the group of mice that
11   were administered arsenic and a subsequent skin painting with TPA showed an increase in the
12   number of papillomas compared to mice treated with TPA alone. Thus, it was suggested that
13   arsenite may be a "tumor enhancer" in  skin carcinogenesis (Germolec et al., 1997; Luster et al.,
14   1995).
15          Ten ppm of either sodium arsenite or DMAV (cacodylic acid) administered for 5 months
16   in the drinking water of K6/ODC transgenic mice induced a small number of skin papillomas
17   (Chen et al., 2000a).  K6/ODC  transgenic mice have hair follicle keratinocytes (likely targets for
18   skin carcinogens), which  over express ornithine decarboxylase (ODC).  ODC is involved in
19   polyamine synthesis, which is needed in S phase. Over expression  of ODC is sufficient to
20   promote papilloma formation without administration of TPA, which has been demonstrated to
21   induce ODC (O'Brien et al., 1997).
22          Rossman et al. (2001) administered sodium arsenite (10 ppm) in the drinking water of
23   hairless Skh 1  mice for 26 weeks.  Mice were also administered 1.7 kJ/m2 solar ultraviolet
24   radiation (UV), which is considered a low, nonerythemic dose, 3 times weekly, either with or
25   without sodium arsenite exposure. Results demonstrated a 2.4-fold increase in the yield of skin
26   tumors for mice exposed to both sodium arsenite and UV than in mice administered UV alone.
27   A second experiment by the same group (Burns et al., 2004), demonstrated a 5-fold increase in
28   skin tumors using 5 mg/L As111  with 1 kJ/m2 solar UV, but also observed a significant increase
29   with  1.25 mg/L As111 with 1 kJ/m2 solar UV.  The skin tumors (mainly SCCs) occurred  earlier,
30   were larger, and were more invasive in mice administered As111.  Arsenite alone did not induce
31   skin tumors. Rossman (2003) concluded that this demonstrates that arsenite enhances the onset
32   and growth of malignant skin tumors induced by a genotoxic carcinogen in mice. Rossman
33   (2003) also suggested that the increased tumor incidence observed by Waalkes et al. (2003) may
34   be due to the same enhancement as C3H mice have a high background of spontaneous tumors
35   and suggests the need for examining the transgenic effects in another strain of mice with a lower
36   b ackground tumorgeni city.
                                                66       DRAFT—DO NOT CITE OR QUOTE

-------
 1          A critical review of the inhalation data was not conducted as part of the evaluation
 2    discussed in this report.

      4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL
 3          Not addressed in this document.

      4.4. OTHER STUDIES

      4.4.1. Possible Modes of Action and Key Events of Possible Importance
 4          As discussed in Section 3.3, the metabolism of inorganic arsenic in humans occurs
 5    through alternating steps of reduction and oxidative methylation mostly to DMAV. Many of the
 6    metabolites have been subjected to a variety of toxicological tests in vivo and in vitro, and they
 7    often differ considerably in their toxicological responses. The relative contributions of the many
 8    different forms of arsenic to the toxicity and carcinogenicity of inorganic arsenic are uncertain.
 9    Each of the arsenical metabolites exhibits its own pattern of toxicity, possibly via similar and/or
10    separate MO As that together are responsible for inorganic arsenic toxicity and tumor formation
11    (Kitchin, 2001).
12          The biotransformation and pharmacodynamics of inorganic arsenic are complex in
13    mammals, with inorganic arsenic being biotransformed through a complex cycle of reduction,
14    oxidation, and methylation steps to form the trimethylated TMAO metabolite,  and possibly its
15    reduced form, trimethylarsine, which may not be of consequence in humans. Arsenical forms of
16    greater instability (i.e., trivalent forms) are produced within each step, and those forms have
17    greater reactivity toward biological and biochemical intermediates and biological
18    macromolecules. The trivalent species MMA111 and DMA111 have been identified as the most
19    toxic and genotoxic forms in several assay systems (Thomas et al.,  2001). Each intermediate
20    arsenical form, however, has the potential to induce cancer or to affect the promotion and
21    progression of cancer, such as by disrupting signal transduction pathways and  gene expression.
22    Many of these forms have been detected in the urine of humans exposed to inorganic arsenic and
23    in rodents exposed to inorganic and organoarsenicals. Through the process of metabolizing
24    arsenic, cells and organs  are exposed to mixtures of these intermediates, which bring to the
25    forefront potential synergistic interactions between them that could enhance the tumorigenesis
26    process.
27          Inorganic arsenic has been demonstrated to cause tumors in humans at  multiple sites
28    (bladder, lung, skin, liver, and possibly kidney).  Rodents are generally much less sensitive to the
29    tumorigenic effects of inorganic arsenic, except for a few recent transplacental mouse studies in
30    which As111 caused liver, lung, ovarian, and/or adrenal cortical tumors (Waalkes et al., 2003,
31    2004a, and 2006a). Currently, there is insufficient information to fully explain the differences
32    between human and rodent sensitivity to arsenic carcinogenicity.
                                                67        DRAFT—DO NOT CITE OR QUOTE

-------
 1          Based on its extensive review of health consequences of inorganic arsenic in drinking
 2    water, NRC (1999) concluded that
 O
 4       •  "The mode of action for arsenic carcinogenicity has not been established.  Inorganic
 5          arsenic and its metabolites have been shown to induce deletion mutations and
 6          chromosomal alterations (aberrations, aneuploidy, and SCE [sister chromatid exchange]),
 7          but not point mutations. Other genotoxic responses that can be pertinent to the mode of
 8          action for arsenic carcinogenicity are co-mutagenicity, DNA methylation, oxidative
 9          stress, and cell proliferation; however,  data on those genotoxic responses are insufficient
10          to draw firm conclusions.  The most plausible and generalized mode of action for arsenic
11          carcinogenicity is that it induces structural and numerical chromosomal abnormalities
12          without acting directly with DNA."
13
14       •  "For arsenic carcinogenicity, the mode of action has not been established, but the several
15          modes of action that are considered most plausible (namely, indirect mechanisms of
16          mutagenicity) lead to a sublinear dose-response at some point below the level at which a
17          significant increase in tumors is observed. However, because a specific mode (or modes)
18          of action has not been identified at this time, it is prudent not to rule out the possibility of
19          a linear response."
20
21          Several of the report's other concluding statements drew attention to the possible
22    importance of ROS to several health effects caused by arsenic and suggested that "intracellular
23    production of ROS might play an initiating role in the carcinogenic process by producing DNA
24    damage" (NRC, 1999).  At the time of the NRC report, the prevailing view was that metabolism
25    of inorganic arsenic through several methylated forms represented a detoxification pathway.
26    One of the fundamental changes in thinking about the effects of inorganic arsenic since the NRC
27    report has been the growing awareness that some of those metabolites (specifically, MMAm and
28    DMA111) can have especially high levels of toxicity. Thus, metabolism also represents a
29    toxification pathway.  Regardless, when there  is a steady influx of inorganic arsenic into the
30    body as through continual exposure from drinking water, metabolism is essential to eliminate
31    that arsenic, including the highly reactive As111, from the body.
32          In 2001, NRC produced an update to its major review on inorganic arsenic in drinking
33    water. It summarized, in tabular format, the mechanistic studies completed since 1998 and
34    included a discussion of them.  It focused on experiments that appeared to induce biochemical
35    effects at moderate to relatively low concentrations of arsenic in vitro (e.g., less than 10  uM);
36    however, some studies that used higher concentrations were included for comparative purposes.
37    The focus was on moderate- to relatively low-dose studies because it was felt that studies that
38    required arsenic concentrations greater than 10 uM to produce a biological response in vitro
39    would be less likely to be relevant to the health effects related to chronic ingestion of arsenic in
40    drinking water. NRC (2001) concluded that "The mechanistic studies reviewed herein and those
41    reviewed previously in the 1999 NRC report suggest that trivalent arsenic species (primarily

                                                68        DRAFT—DO NOT CITE OR QUOTE

-------
 1    As111, MMAm, and, possibly, DMA111) are the forms of arsenic of greatest toxicological concern."
 2    They estimated concentrations of arsenic that could be expected in human urine from the known
 3    human experience and concluded that "Arsenite concentrations in excess of 10 uM generally
 4    exceed concentrations that can occur in the urine of individuals chronically exposed to arsenic in
 5    drinking water and have less direct relevance to understanding the modes of action responsible
 6    for human cancer induced by this route of exposure." They also stated that:
 7
 8       •  "Experiments in animals and in vitro have demonstrated that arsenic has many
 9          biochemical and cytotoxic effects at low doses and concentrations that are potentially
10          attainable in human tissues following ingestion of arsenic in drinking water. Those
11          effects include induction of oxidative damage to DNA; altered DNA methylation and
12          gene expression; changes in intracellular levels of murine double minute 2 proto-
13          oncogene (mdm2) protein and p53 protein; inhibition of thioredoxin reductase (TrxR;
14          MMA111 but not As111); inhibition of pyruvate dehydrogenase; altered colony-forming
15          efficiency; induction of protein-DNA cross-links; induction of apoptosis; altered
16          regulation of DNA-repair genes, thioredoxin, glutathione reductase, and other stress-
17          response pathways; stimulation or inhibition of normal human keratinocyte cell
18          proliferation, depending on the concentration; and altered function of the glucocorticoid
19          receptor."
20
21          Despite the extensive research on MO A up to that time, NRC stated that "the
22    experimental evidence does not allow confidence in distinguishing between various shapes
23    (sublinear, linear, or supralinear) of the dose-response curve for tumorigenesis at low doses."
24          The present review uses the terms "mode of action" and "key event" as they are
25    described in the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a). According to
26    EPA, '"mode of action' is defined as a sequence of key events and processes, starting with
27    interaction of an agent with a cell, proceeding through operational and anatomical changes, and
28    resulting in cancer formation. A 'key event' is an empirically observable precursor step that is
29    itself a necessary element of the mode of action or is a biologically based marker for such an
30    element. Mode of action is contrasted with 'mechanism of action', which implies a more
31    detailed understanding and description of events, often at the molecular level, than is meant by
32    mode of action.  The toxicokinetic processes that lead to formation or distribution of the active
33    agent to the target tissue are considered in estimating dose, but are not part of the mode of action
34    as the term is used here.  There  are many examples of hypothesized modes of carcinogenic
35    action, such as mutagenicity, mitogenesis, inhibition of cell death, cytotoxicity with reparative
36    cell proliferation, and immune suppression."
37          In this review, tables have been compiled in order to make a large amount of information
38    on the biological effects of inorganic arsenic readily available.  Appendix C contains tables that
39    deal with in vivo human studies (Table C-l), in vivo experiments on laboratory animals (Table
40    C-2), and in vitro studies (Table C-3). These tables include as many experiments published from

                                                 69        DRAFT—DO NOT CITE OR QUOTE

-------
 1    2005 through August 2007 as possible. Numerous earlier experiments have been included as
 2    well, based on various selection criteria: being mentioned in the SAB Arsenic Review Panel
 3    comments of July 2007 (SAB, 2007) or in NRC's update (NRC, 2001), or inclusion in an earlier
 4    draft that lacked tables (U.S. EPA, 2005c).  The tables provide information on: (1) the arsenic
 5    species tested; (2) the cell types, tissues, or species tested; (3) all concentrations or doses tested;
 6    (4) all durations of exposure; (5) estimates of the LOEC or LOEL (i.e.,  lowest observed effect
 7    concentration or level); (6) a summary of the most important results of each study; and (7) the
 8    citations. The 22 categories into which the hypothesized key events are grouped in those tables
 9    are listed in column 1 of Table 4-1, and the number of data rows under each category provide an
10    estimate of the amount of available data pertaining to each category topic. Data from a single
11    publication are sometimes entered under multiple event categories. For example, the results in
12    Wang et al. (1996) are summarized in rows under Apoptosis, Cytotoxicity, and Effects Related
13    to Oxidative Stress (ROS).
14          When judging the possible relevance of in vitro experiments or in vivo laboratory animal
15    experiments on human health, it is useful to keep in mind that the total concentration of As111 and
16    Asv in drinking water pumped from tube wells in Bangladesh (as an example of one country
17    with high exposures to inorganic arsenic in drinking water) ranges from 20 to over 2,000 ppb
18    arsenic (i.e., 0.3 to 27 uM).  In people exposed at those high levels, total blood arsenic levels
19    range from 0.5 to 1.2 uM (Snow et al., 2005), and total arsenic concentrations in urine would
20    probably not exceed 10 uM (NRC, 2001).
21
                                                70       DRAFT—DO NOT CITE OR QUOTE

-------
     Table 4-1. Summary of Number of Rows Derived From Peer-Reviewed Publications for
               Different Hypothesized Key Events3
Hypothesized Key Events
Aberrant Gene or Protein Expression13
Apoptosis
Cancer Promotion
Cell Cycle Arrest or Reduced Proliferation
Cell Proliferation Stimulation
Chromosomal Aberrations and/or Genetic Instability
Co-carcinogenesis
Co-mutagenesis
Cytotoxicity
DNA Damage
DNA Repair Inhibition or Stimulation
Effects Related to Oxidative Stress (ROS)
Enzyme Activity Inhibition
Gene Amplification
Gene Mutations
Hypermethylation of DNA
Hypomethylation of DNA
Immune System Response
Inhibition of Differentiation
Interference With Hormone Function
Malignant Rransformation or Morphological
Transformation
Signal Transduction
Number of Rows in Tables
In Vivo Human
Studies
(Table C-l)
6
1
0
0
0
13
0
0
0
5
2
2
0
0
1
2
1
1
0
0
0
1
In Vivo
Experiments
Using
Laboratory
Animals
(Table C-2)
32
6
3
1
18
3
2
1
2
6
0
30
0
0
2
1
2
0
0
1
0
2
In Vitro
Experiments
(Table C-3)
124
78
3
29
21
83
3
21
118
35
11
69
5
5
7
2
7
46
13
7
13
51
     a Details of the studies are presented in Appendix C.
     b Some hypothesized key events are shown in boldface to emphasize that in at least one of the tables they contain
     much more data than the other categories.

     4.4.1.1. In Vivo Human Studies
1          Table C-l summarizes in vivo human studies. Here and elsewhere in the consideration of
2    human studies there was particular interest in the subset of people who develop skin lesions
3    (usually keratoses, which are often considered premalignant, or hyperpigmentation) following
4    long-term exposure to inorganic arsenic in drinking water. Indeed, four of the six studies related
5    to Aberrant Gene or Protein Expression compared groups of people with and without arsenic-
6    related skin lesions following similar exposures to high levels of inorganic arsenic in drinking
7    water, and in three cases, they also compared them to groups of people with much lower
                                                71
DRAFT—DO NOT CITE OR QUOTE

-------
 1    inorganic arsenic exposure levels. The genomics study by Argos et al. (2006) showed that 312
 2    more genes were down-regulated in the group with skin lesions than in the inorganic arsenic-
 3    exposed group without such lesions. No genes were shown to be up-regulated. Other studies
 4    showed increased levels of the EGFR-ECD protein (i.e., extracellular domain of the epidermal
 5    growth factor receptor) in serum (Li et al., 2007), increased levels of transforming growth factor
 6    alpha (TGF-a) protein in bladder urothelial cells (Valenzuela et al., 2007), and decreased levels
 7    of three integrins in and around  skin lesions following exposures to inorganic arsenic in drinking
 8    water (Lee et al., 2006b). Integrins are important in the control of differentiation and
 9    proliferation of the epidermis. Many skin diseases, including arsenical keratosis, show altered
10    patterns of integrin distribution and expression.  In the first two instances, there were bigger
11    increases  in the group with skin lesions. The study on integrins only made comparisons to a
12    control group. One of the other studies showed a decrease in the concentration of the receptor
13    for advanced glycation end products (RAGE) protein in sputum when there was a higher
14    concentration of inorganic arsenic in the urine (Lantz et al., 2007). Changes  in that biomarker
15    are related to several chronic inflammatory diseases in the lung,  including lung cancer. The
16    remaining study showed that two oncogenes were up-regulated in tumor tissues in patients with
17    arsenic-related urothelial cancer, but not in those from patients with non-arsenic-related
18    urothelial cancer (Hour et al., 2006).
19          The Chromosomal Aberrations and/or Genetic Instability category has the most entries in
20    the table on human studies.  Although some of the studies found no effects (usually on SCE
21    induction) in people exposed to  inorganic As, most of the studies included in the table showed
22    clear increases of chromosomal  aberrations (C A) in lymphocytes, micronuclei (MN;  in various
23    cell types), or both CA and MN in people who had been exposed to high levels of inorganic
24    arsenic in drinking water or to Fowler's solution (i.e., a solution containing 1% arsenic that was
25    commonly used as a medicine in the 1800s and early  1900s). Arsenic was shown to increase the
26    incidence of MN specifically in bladder cells (Warner et al., 1994; Moore et  al., 1996, 1997b).
27    There also was suggestive evidence that some arsenic-induced MN (a minority of them) result
28    from aneuploidy (Moore et al., 1996).  There was some evidence for induction of SCE.  Three of
29    the papers showed that those persons with arsenic-induced skin lesions had higher frequencies of
30    induced chromosomal damage seen either as CA or MN than those without lesions (Gonsebatt et
31    al., 1997;  Ghosh et al., 2006; Banerjee et al., 2007).  It is intriguing that one of the studies
32    demonstrated an apparent predisposition to both skin  lesions and CA that was correlated with
33    (and was thus perhaps caused by) a single polymorphism of the ERCC2 (excision repair cross-
34    complementing rodent repair deficiency gene, complementation group 2) gene, which plays a
35    key role in the nucleotide excision repair (NER) pathway. The polymorphism resulted from an
36    A—>C mutation at codon 751 that caused a change from lysine to glutamine,  and the allele
37    conferring the higher predisposition in homozygotes had the remarkably high gene frequency of

                                                72        DRAFT—DO NOT CITE  OR QUOTE

-------
 1    0.40 in that population (Banerjee et al., 2007). Although only some of the homozygotes heavily
 2    exposed to inorganic arsenic in drinking water developed skin lesions or had chromosomal
 3    aberrations, those that were affected had both endpoints.
 4          Table C-l also provides data showing that oral inorganic arsenic exposure increases
 5    DNA damage. Two papers reported oxidative damage to DNA revealed by increases in the
 6    concentration of 8-hydroxydeoxyguanosine (8-OHdG) in the urine. Both studies were in Japan,
 7    with the first showing a positive correlation between urinary concentrations of arsenic and 8-
 8    OHdG after analyzing samples from 248 people in the general population (Kimura et al.,  2006).
 9    The other study (Yamauchi et al., 2004) involved clinical examination of 52 patients following
10    an incident in which 63 people (four of whom died within about 12 hours of being poisoned)
11    were poisoned by eating food  contaminated with ATO. Those 52 patients  were followed up for
12    various effects including levels of 8-OHdG in urine. Maximal levels of-150% compared to
13    normal Japanese levels were reached 30 days after the exposure, and by 180 days the levels had
14    returned to normal.  The same paper reported that people in Inner Mongolia, China, who  drank
15    water contaminated with about 130 ppb arsenic had a significant increase in urinary 8-OHdG,
16    which returned to normal after they drank "low-arsenic" water for one year.
17          Table C-l includes data that demonstrate DNA damage (i.e., single-strand breaks)
18    detected by the single cell gel  electrophoresis (SCGE) comet assay. One of those studies, in
19    which the high-exposure group drank water containing about 247 ppb As,  also included a comet
20    assay combined with formamidopyrimidine-DNA glycosylase (FPG) digestion and thereby
21    showed that arsenic also induced oxidative base damage. (Digestion with the FPG enzyme
22    breaks the DNA at the sites of oxidative damage so that those sites are seen in this modified
23    comet assay.) Besides looking at baseline DNA damage, the other comet study investigated the
24    capacity of the lymphocytes of subjects who used drinking water containing 13-93 ppb arsenic
25    to repair damage induced by an in vitro challenge with the mutagen 2-
26    acetoxyacetylaminofluorene (2-AAAF). Adducts formed following treatment with 2-AAAF are
27    primarily repaired through the NER pathway and lymphocytes from arsenic-exposed individuals
28    had more  adducts. The lymphocytes from the people with high-arsenic exposure had reduced
29    NER ability (Basu et al., 2005).  The remaining DNA damage study (Mo et al., 2006) used 8-
30    oxoguanine DNA glycosylase (OGG1) expression as an  indicator of oxidative-induced DNA
31    damage. The OGG1  gene codes for an enzyme involved in base excision repair (BER) of
32    residues that result from oxidative damage to DNA. OGG1 expression was found to be closely
33    linked to the levels of arsenic in drinking water and in toenails, thereby indicating a link between
34    ROS damage to DNA and inorganic arsenic exposure. An inverse relationship between OGG1
35    expression and selenium (Se) levels in toenails was found, which suggests a possible protective
36    effect of Se against arsenic-induced oxidative stress. As was often the case when populating the
37    MO A tables in Appendix C, some studies could equally well be placed into one or another

                                               73        DRAFT—DO NOT CITE OR QUOTE

-------
 1    hypothesized key event category, and clearly some studies listed under DNA Damage also relate
 2    to the hypothesized key events of DNA Repair Inhibition or Stimulation and Effects Related to
 3    Oxidative Stress (ROS).
 4          In another polymorphism study, homozygotes for two different alleles of the p53 gene
 5    were shown to be at higher risk (than those carrying other alleles) of developing arsenic-induced
 6    keratosis among individuals who used drinking water that contained roughly 180 ppb arsenic
 7    (De Chaudhuri et al., 2006).  Because that gene is so important in controlling apoptosis, that
 8    study was listed under Apoptosis. It is unclear, however, why mutations at that gene would
 9    predispose those who consume high levels of arsenic to develop skin lesions.  Two studies
10    described under DNA Repair Inhibition or Stimulation demonstrated reduced expression of three
11    nucleotide excision repair (NER) genes in a population that used drinking water that contained
12    10-75 ppb arsenic (Andrew et al., 2003, 2006). Still more evidence that arsenic causes Effects
13    Related to Oxidative Stress (ROS) comes from school children in Taiwan who showed a positive
14    correlation between urinary concentrations of arsenic and 8-OHdG; no information was provided
15    regarding the level of arsenic in their drinking water (Wong et al., 2005). Subjects with arsenic-
16    related skin lesions from a population in Inner Mongolia, China, that used drinking water with a
17    mean of 158 ppb arsenic showed a statistically significant positive correlation between 8-OHdG
18    adducts in their urine and individual urinary concentrations of inorganic As, MMA, and DMA.
19    In contrast, those without skin lesions showed no correlation (Fujino et al., 2005).
20          Evidence is presented under Hypermethylation of DNA that arsenic exposure causes
21    hypermethylation of the promoter sequence in the DNA for four tumor suppressor genes.  For
22    two of the genes, p53  and pi 6, there was a positive dose-response between arsenic
23    contamination of drinking water and the level of effect; however, this was only seen in
24    individuals with skin lesions (Chanda et al., 2006). For the other two genes, RASSF1A and
25    PRSS3, the association was demonstrated with regard to the level of arsenic consumption
26    estimated from toenail clippings (Marsit et al., 2006). Because the Marsit et al. (2006) study was
27    done on bladder cancer patients, it provides a potential link between arsenic exposure and
28    epigenetic alterations in patients with bladder cancer. The Chanda et al. (2006) study also
29    demonstrated hypomethylation in a few individuals, but it was found only in persons having
30    prolonged arsenic exposure at high doses.
31          Regarding the hypothesized key event category Immune System Response, there was
32    suggestive evidence of an association between changes in sensitive markers of lung
33    inflammation (i.e., metalloproteinase concentrations in induced sputum) and levels of only about
34    20 ppb of arsenic in drinking water. The initial comparison between the high- and low-level
35    exposure towns showed no difference with regard to these biomarkers, but a significant
36    association appeared when the analysis was adjusted for possible confounding factors (Josyula et
37    al., 2006).  Islam et al. (2007) found that IgG and IgE levels were significantly elevated in

                                                74        DRAFT—DO NOT CITE OR QUOTE

-------
 1    arsenic-exposed individual with skin lesions. More details about that experiment, including
 2    clinical findings possibly related to inflammatory reactions, are found in Appendix D.  Appendix
 3    D discusses several other studies (including in vitro experiments and experiments on laboratory
 4    animals) related to immunotoxicity, including some that are not included in any of the tables in
 5    Appendix C.
 6          The only study listed under Gene Mutations gave no more than a hint of an effect
 7    (Ostrosky-Wegman et al., 1991). Regarding Signal Transduction, a study in Taiwan showed that
 8    both the levels of plasma TGF-a and the proportion of individuals with TGF-a over-expression
 9    were significantly higher in the high CAE group than in the control group (Hsu et al., 2006).
10          Only limited information from the cited experiments has been included in this discussion.
11    Much more detail on these studies can be found in Table C-l of Appendix C as well as in Table
12    C-2 for in vivo experiments using  laboratory animals and Table C-3 for in vitro experiments.
13    Brief discussions of the information in Table C-2 and C-3 are found in Sections 4.4.1.2 and
14    4.4.1.3, respectively.
      4.4.1.2. In Vivo Experiments Using Laboratory Animals
15          Table C-2 summarizes in vivo experiments  using laboratory animals. All doses given in
16    this section are stated in terms of the amount of arsenic in the dose.  Twenty-four of the 112 rows
17    in Table C-2 involve studies of nine key event categories in mice that drank water containing
18    arsenic for several to many weeks. Results are of particular interest because they involved most
19    of the lowest dose levels tested,  and As111 is the most toxic oxidation state of inorganic As.
20    Figure 4-1 summarizes the results  according to key events by showing, for each endpoint, the
21    concentration of arsenic in the water that was the LOEL, the period of treatment, and the organ
22    or tissue in which the effect was seen. Because the result for gene mutations was a negative
23    finding, it is not shown in the figure.  Sometimes more than one entry in Table C-2 corresponds
24    to a single item in the figure, and sometimes a single entry in the table deals with separate groups
25    of animals.  Consequently, there may be multiple LOELs shown in the figure.  It should also be
26    kept in mind that sometimes only one dose was tested in an experiment, and, of course, if an
27    effect was found, that dose became the LOEL  (even though a much lower dose might have been
28    effective).  One benefit of the detailed descriptions found in Table C-2 is that all doses tested are
29    listed. As Figure 4-1  shows, roughly half the dose  levels used exceed 2,000 ppb and are thus
30    much higher than levels ever found in drinking water used for human consumption. While all of
31    the experiments summarized in Table C-2 are useful in terms of showing their effects in mice,
32    this discussion gives more attention to doses that overlap higher levels of exposure to humans
33    from drinking water.  A better understanding of the pharmacokinetic characteristics in different
34    species may aid in determining the relevance of the high-dose animal studies to human subjects
35    exposed to arsenic in drinking water at lower concentrations for a longer period.

                                                75       DRAFT—DO NOT CITE OR QUOTE

-------
ppm A
1000
100
10


1

0.1

0.01
0.001
0.0001
Cancer Promotion = CP Co-carcinogenesis = CC
Aberrant Gene or or Cell Cycle Arrest or Cell Proliferation or Effects Related to
Protein Expression Hypomethylation of DNA Reduced Proliferation Stimulation Co-mulagenesis Oxidative Stress (ROS)
• 4w, S CP«14w, S • 10w, S

• 48w,Li »48w.Li »4w'B •'6w'B • Sw.Bl.K.and
• 8w, Lu
.., „ •13w,Bl,Br,
• 16w,B an(] u
• 10 w, S
CC»29w, S
• 26 w, Li
• 20w,P •20w,VH CC»23w,S
• 5w, H
-
• 4w, Lu;10w, H;20w, H •Sw, V
• 9w, TS »8w, TS
• 5», V
_
• LOEL . B-bladder Br-brain It-kidney Lu-lung S-skin V-blood vessels
Effects seen in:
w-Week(s) Bl-blood H-heart Li-liver P-blood plasma T-tumor tissue

K,









                  Figure 4-1.  Level of significant exposure of adult mice to sodium
                              arsenite in drinking water in ppm As.
 2          The Aberrant Gene or Protein Expression effects seen at those lower levels included
 3   increases in levels of several proteins and in mRNA levels of a few genes that are important in
 4   angiogenesis and remodeling. For example, vascular endothelial cell growth factor [VEGF] and
 5   its receptors VEGFR1 and VEGFR2 were measured in hearts, and increases were sometimes
 6   restricted to areas around blood vessels (Kamat et al., 2005; Soucy et al., 2005). However,
 7   increases in dose (up to 0.5 ppm in drinking water) and duration (up to 20 weeks) actually
 8   caused decreases in the protein and mRNA levels for VEGFR1 and VEGFR2, suggesting that
 9   chronic exposure at these higher levels was toxic to the cardiac vasculature in mice.  Consistent
10   with the decreased mRNA levels seen for VEGFR1 and VEGFR2 following 20-week chronic
11   exposures to 0.5 ppm, the same treatment regimen produced evidence of reduced cell
12   proliferation, which was represented as a decrease in the density of microvessels of less than 12
                                               76
DRAFT—DO NOT CITE OR QUOTE

-------
 1    um in the heart (Soucy et al., 2005). These data thus provide an interesting example of the
 2    concentration and time-dependent effects of arsenic exposure that might be important in the
 3    etiology of some of the diseases that it causes. In contrast, stimulation of cell proliferation at
 4    low-dose levels involved increases in (1) blood vessel number in Matrigel implants (Soucy et al.,
 5    2005), (2) tumor growth rates after implantation of tumor cells (Kamat et al., 2005), and (3)
 6    number of metastases to the lungs after implantation of those tumor cells (Kamat et al., 2005).
 7          Proteomic analysis of bronchoalveolar lavage fluid from lungs of mice that drank 0.05
 8    ppm (i.e., 50 ppb) arsenic in water for 4 weeks showed an increase in peroxiredoxin-6 and
 9    enolase 1 levels and a decrease in GSTO1, RAGE, contraspin, and apolipoproteins A-I and A-IV
10    (Lantz et al., 2007). That same paper had demonstrated a decrease in the level of RAGE protein
11    in human sputum that was associated with  arsenic exposure.  Two microarray experiments at
12    much higher dose levels of 28.8 and 45 ppm showed changes in expression of dozens of genes
13    (Chen et al., 2004b; Lantz and Hays, 2006). In each experiment, the LOEL was the only dose
14    tested, which leaves open the possibility that such high doses might not have been necessary to
15    obtain these changes.
16          Mice that were exposed for 23 weeks to 0.7-5.8 ppm arsenic in drinking water developed
17    no skin tumors; however, when they were also exposed to UV thrice weekly for most of that
18    time, they showed a strong dose-related increase up through 2.9 ppm As, thus providing strong
19    evidence of co-carcinogenesis (Burns et al., 2004).  Another part of the same study  (reported in
20    Uddin et al., 2005) demonstrated that at 2.9 ppm there was oxidative DNA damage  caused by the
21    co-treatment. Effects Related to Oxidative Stress (ROS) following 26 weeks of exposure at 1.8
22    ppm included decreases in GSH content, and in the  activities of glucose-6-phosphate
23    dehydrogenase (G6PDH), glutathione peroxidase (GPx), and plasma membrane Na+/K+
24    ATPase. Additional changes suggestive of such damage, such as an increase in the
25    concentration of malondialdehyde (MDA), were apparent after  9, 12, or 15 months  at the same
26    dose level (Mazumder, 2005).
27          Eighteen of the 112 rows in Table C-2 involved rats that drank water containing sodium
28    arsenite for several to many weeks, but those studies are distributed among only two key event
29    categories and do not extend down to nearly as many effects at  low exposure levels. Most
30    experiments cited in the  18 rows involved drinking water containing 57.7 ppm arsenic for
31    several to many weeks and showed findings of numerous changes indicative of oxidative damage
32    in several organs. A few experiments show differing levels of oxidative damage in different
33    regions of the brain (Samuel et al., 2005; Shila et al., 2005a,b).  By far the lowest dose tested
34    among these experiments was 0.03 ppm As, and it was found to be effective in  decreasing the
35    GSH level and superoxide radical dismutase (SOD) activity in the liver.  The other two dose
36    levels tested, 1.4 and 2.9 ppm, caused bigger changes in these two variables, as well as
37    additional changes indicative of oxidative stress. It is of interest that the changes per unit dose

                                               77        DRAFT—DO NOT CITE OR QUOTE

-------
 1    were much higher for GSH and SOD at 0.03 ppb than they were at the two much higher doses
 2    tested (Bashir et al., 2006a).  In experiments using 5.8 ppm As, which rats drank for 4, 8, or 12
 3    weeks, activities of catalase (CAT) and SOD in kidney, liver, and RBCs were found to be
 4    elevated at 4 weeks, but they decreased to baseline levels or lower by 12 weeks; MDA levels
 5    were always elevated (Nandi et al., 2006). Consumption of water containing 1.4 ppm arsenic for
 6    60 days led to a demonstrable increase in apoptosis in liver cells (Bashir et al., 2006a).
 7          Twenty-six of the 112 rows in Table C-2 involve rats or mice that consumed pentavalent
 8    arsenicals (Asv, MMAV, DMAV, or TMAV) for several to many weeks, and in all but three rows
 9    they were delivered in drinking water instead of food.  As would be expected for these less
10    potent forms of arsenic, LOELs were typically high and usually above 50 ppm.  Only a few
11    results occurred at much lower concentrations, and are mentioned in this discussion.  After rats
12    were exposed for 28  days to 0.35 ppm arsenic in drinking water in the form of DMAV,
13    microarray analysis demonstrated significant effects on the expression of 503 genes (i.e., 11% of
14    the genes tested with that microarray) in urothelial cells. Even more genes were affected at the
15    three higher doses tested (i.e., 1.4, 14, and 35 ppm As).  Most of the effected genes related to the
16    functional categories of apoptosis, cell cycle regulation, adhesion, signal transduction, stress
17    response, or growth factor and hormone receptors. There was a change in the types of genes
18    affected at the different doses, particularly when comparing the higher two doses (both
19    cytotoxic) with the two non-cytotoxic doses (Sen et al., 2005). When rats were exposed to 0.24
20    ppm Asv for 1 or 4 months in drinking water, changes in signal transduction were increased
21    expression of integrin-linked kinase (ILK) and decreased expression of phosphatase and tensin
22    homolog (PTEN) in the liver. At higher doses, the expression of these genes and additional
23    cancer-related genes was affected (Cui et al., 2004b).
24          DNA damage (both fragmentation and oxidative) was demonstrated in peripheral blood
25    leukocytes of mice using the comet assay following exposure of 50, 200, or 500 ppb arsenic in
26    drinking water in the form of Asv for 3 months with and without a low-Se diet.   Arsenic caused
27    increased DNA fragmentation only in mice consuming the low-Se diet, and induced oxidative
28    damage only in mice consuming the normal-Se diet. Neither case showed a positive dose-
29    response (Palus et al., 2006). In lung adenocarcinomas from mice exposed for 18 months to
30    0.24, 2.4, or 24 ppm Asv in drinking water, there was an increase in the extent of
31    hypermethylation of promoter regions of tumor suppressor genes p!6INK4a and RASSF1A
32    (genes frequently found inactivated in many types of cancer including lung cancer), based on
33    methylation-specific polymerase chain reaction (PCR).  All doses had an effect, and there was a
34    positive  dose-response. Reduced expression or lack of expression of these two genes was
35    correlated with the extent of hypermethylation. Mice without tumors, whether control or
36    arsenic-treated, had normal (i.e., not reduced or eliminated) expression of these genes in their
                                                78       DRAFT—DO NOT CITE OR QUOTE

-------
 1    lungs. The authors concluded that epigenetic changes of tumor suppressor genes are involved in
 2    inorganic arsenic-induced lung carcinogenesis (Cui et al., 2006).
 3          Of the experiments described in Table C-2 in which arsenic exposure occurred through
 4    consumption of arsenic in drinking water or food, the only group not yet discussed consists of
 5    the series of experiments in which pregnant female mice drank water containing 42.5 or 85 ppm
 6    arsenic in the form of sodium arsenite for 10 days on gestation days 8 to 18.  These studies
 7    follow up on the interesting observation that arsenic seems to be a complete carcinogen in mice
 8    following such a treatment. The offspring were observed for effects (sometimes only after they
 9    had grown to be adults), and results are categorized in Table C-2 under Aberrant Gene or Protein
10    Expression, Cell Proliferation or Stimulation, Hypomethylation of DNA, and Signal
11    Transduction. Some of the more noteworthy findings were as follows. Numerous microchip
12    analyses were conducted, often with some of the findings confirmed by real-time (RT) PCR.
13    Microarrays containing from 588 to 22,000 genes were used. It was not unusual to find changes
14    in the expression of scores of genes (sometimes even of thousands) in the different studies.
15    Changes (often many-fold) included both increases and decreases of expression, occurring at
16    both dose levels.  Some of the many types of genes often altered included oncogenes, HCC
17    biomarkers, cell proliferation-related genes, stress proteins, insulin-like growth factors, estrogen-
18    linked genes, and genes involved in cell-cell communication.  Tissues in which gene expression
19    changes were found in offspring that had been exposed to arsenic in utero included: (1) arsenic-
20    induced HCC tumors that developed in adult males, (2) normal-appearing cells in livers of adult
21    males, (3) fetal livers of males right at the end of treatment, (4) livers of newborn males, (5) fetal
22    lungs of females right at the end of treatment, and (6) arsenic-induced adenomas and
23    adenocarcinomas that developed in lungs of adult females.
24          The expression of three estrogen-related genes was shown to increase synergistically in
25    the uteri  of females (at 11 days of age) that had been exposed in utero to arsenic and also
26    subcutaneously injected with diethylstilbestrol (DBS) on the first 5 days after birth. These and
27    other results showed that inorganic arsenic acts with estrogens to enhance production of
28    urogenital cancers in female mice (Waalkes et al., 2006a).  Females that had been exposed to
29    arsenic in utero and then received a 21-week post-weaning treatment with TPA showed changes
30    in gene expression that were similar to those seen in liver samples from males that had received
31    only the arsenic treatment in utero.  This is interesting because it parallels another situation in
32    which TPA-treated females showed a response similar to males without TPA treatment.
33    Specifically,  female mice exposed in utero to arsenic develop HCC only after TPA treatment
34    (Liu et al., 2006b); however, male mice exposed in utero to arsenic develop those tumors without
35    receiving any TPA treatment. Observed changes in estrogen-related genes sometimes seemed
36    especially important in the interpretation of results, and fetal lungs of females exposed to arsenic
37    in utero showed a large increase in estrogen receptor-alpha (ER-a), as well as several other

                                                79        DRAFT—DO NOT CITE OR QUOTE

-------
 1    estrogen-related genes and numerous other genes, including some associated with lung cancer.
 2    There also was a large increase in nuclear ER-a in adenomas and adenocarcinomas that
 3    developed in the lungs of adult females that had been exposed to arsenic in utero (Shen et al.,
 4    2007).
 5          Stimulation of cell proliferation during treatment of males while in utero at 85 ppm
 6    induced kidney cystic tubular hyperplasia in 23% of the animals, and although males did not
 7    develop bladder hyperplasia from the arsenic treatment alone, they often did if treated in
 8    conjunction with DBS or tamoxifen on the first 5 days after birth because of a synergistic
 9    interaction that occurred with those chemicals. Although females exposed while in utero showed
10    bladder hyperplasia similar to the males, arsenic exposure in utero alone caused no hyperplasia
11    in their kidneys (Waalkes et al., 2006a,b). Global hypomethylation of GC-rich  regions was
12    demonstrated in livers of newborn males that received 85 ppm in utero (Xie et al., 2007).
13          Almost all remaining experiments summarized in Table  C-2 involved treatments of mice
14    or rats by gavage, and those results are summarized under Aberrant Gene or Protein Expression,
15    Apoptosis, Chromosomal Aberrations and/or Genetic Instability, Effects Related to Oxidative
16    Stress (ROS), and Interference With Hormone Function. In all rows where As111 was
17    administered, it was usually as sodium arsenite, but sometimes as arsenic trioxide (ATO). One
18    study also included treatment with pentavalent arsenicals. By using gavage, the amount of the
19    arsenical administered to each animal was controlled precisely,  and it was given as a certain
20    weight of arsenic per animal, often with adjustment to the individual weight of  each animal  (i.e.,
21    ug/animal or mg/kg bw, respectively). Most treatments were administered repeatedly, with
22    treatment regimens in one case lasting an entire year.  As in all other  studies on experimental
23    animals, there was an attempt here to state all doses in terms of the amount of arsenic. Because
24    it was unclear from the reporting of a few experiments whether  doses were expressed as arsenic
25    compound or as As, Table C-2 always makes it clear whether or not such a correction was made.
26    In a gavage study with one of the smallest amounts of arsenic per dose (equivalent to 36
27    ug/mouse if a mouse weighed 25 g), Patra et al. (2005) found induction of chromosomal
28    aberrations in mice that received  1.44 mg As/kg bw given as sodium  arsenite by gavage once-
29    per-week for 4 weeks. Induction of chromosomal aberrations also was seen after 5 and 6
30    treatments; however, 7 and 8 treatments were lethal to the mice.  A 25 g mouse in that study
31    would have received the same amount of arsenic in that one day if it had drunk  water that
32    contained 6 ppm arsenic (assuming that it drank 6 mL of water,  which would be a reasonable
33    amount for a mouse).
34          In the only  gavage study with in utero treatments, 9 daily treatments of 4.35 mg As/kg
35    bw was shown to increase the activity of the selenoprotein iodothyronine deiodinase-II (DI-II) in
36    fetal brains and to decrease the activity of the selenoprotein  TrxR in fetal livers. In both cases,
37    these results were observed only if the mice were on a Se-deficient diet (Miyazaki et al., 2005).

                                                80       DRAFT—DO NOT CITE OR QUOTE

-------
 1    In a gavage study lasting a full year (Das et al., 2005), mice were administered 50, 100, or 150
 2    ug/mouse, 6 days a week for 3, 6, 9, or 12 months; it took 9 months before substantial increases
 3    were seen in the activities of tumor necrosis factor alpha (TNF-a) and interleukin (IL)-6 at any
 4    dose, but by then all doses had an effect and there was a positive dose-response. Three months
 5    later, both effects had increased substantially at all doses, still with a positive dose-response.  A
 6    similar response was seen for the concentration of total collagen, although increases were not as
 7    large in comparison to the control group. That same study examined six components of the
 8    antioxidant defense system and found numerous interesting changes over time. While  all of the
 9    affected components had a LOEL of 50 ug at the 3-, 9-, and 12-month test periods, all five
10    affected components had a LOEL of 100 ug at 6 months.  GSH levels and activities of GPx and
11    CAT increased by 3 months, but decreased by 9 and 12 months. In another experiment with
12    single, large doses of As111 or Asv given to mice by gavage, there were large increases in heme
13    oxygenase 1 (HMOX-1) activity within 6 hours in liver and kidney but not in the brain. The
14    effect was somewhat higher with As111, but DMAV had no effect.  This study also tested some
15    much smaller doses, and a dose as high as 2.25 mg/kg bw had no effect on this endpoint in
16    kidneys (Kenyon et al., 2005b).
17          Various biochemical indicators of apoptosis were seen in brain and liver 24 hours after
18    giving rats a single high dose of sodium arsenite by gavage (Bashir et al., 2006b).  The same
19    paper showed that single, large doses of sodium arsenite given to rats by gavage affected many
20    biochemical indicators of oxidative stress in liver and brain 24 hours after treatment. Some
21    studies on Effects Related to Oxidative Stress (ROS) included co-treatments with antioxidants
22    that were shown to reduce the level of effects seen (Modi et al., 2006; Sohini and Rana, 2007).
23    With regard to Interference With Hormone Function, rats given 30.3 mg Asin/kg bw as ATO by
24    gavage every other day for 30 days were shown to have a large increase in the levels of thyroid
25    hormones triiodothyronine (T3) and thyroxine (T4) in their blood serum (Rana and Allen, 2006).
      4.4.1.3. In Vitro Experiments
26          Table C-3 summarizes a large number of in vitro experiments; and some highlights are
27    discussed below. The potencies of many arsenicals, including both trivalent and pentavalent
28    forms, have been compared in several series of experiments, with the obvious conclusion that the
29    pentavalent forms  almost always have much higher LOECs (e.g., Moore et al., 1997a;  Sakurai et
30    al., 1998; Petrick et al., 2000; Drobna et al., 2002; Kligerman et al., 2003).  Consequently, the
31    discussion below does not focus on the studies that analyzed pentavalent arsenicals.
32          Three chemical properties of arsenic likely to account for its biological activity are:
33    (1) the soft acid/soft base principle (which is related to trivalent arsenicals and sulfhydryl
34    binding); (2) the nucleophilicity of trivalent arsenicals; and (3) the formation of free radicals,
35    ROS, or both by arsenicals (Kitchin et al., 2003).  As noted by Kitchin et al. (2003):

                                                81        DRAFT—DO NOT CITE OR QUOTE

-------
 1       •  "If trivalent arsenicals acting as soft acids are causally important, then the likely modes
 2          of action of arsenic carcinogenesis may include altered DNA repair, altered growth
 3          factors, cell proliferation, altered DNA methylation patterns and promotion of
 4          carcinogenesis."
 5
 6          Arsenic is readily absorbed from the GI tract in humans and is primarily transported in
 7    the blood bound to sulfhydryl groups in proteins and low-molecular-weight compounds, such as
 8    amino acids and peptides (NRC, 1999).  At any given time, about 99%  of absorbed As111 is bound
 9    to tissue sulfhydryls, mostly to monothiol sites (Kitchin and Wallace, 2006). Based on the
10    results of their peptide binding studies, Kitchin and Wallace (2006) suggested that dithiol- and
11    trithiol-binding sites would be "the most likely causal triggers of biological effects because of
12    their stronger affinity and because the bi- and tri-dentate complexes last so much longer than the
13    rapidly dissociating and reforming binding of arsenite to monothiol sites." While the As111
14    attachment to the monothiol-binding sites are short-lived, a substantial part of the total As111
15    attaches to those sites because of their great abundance in mammals. Because the  functional
16    group of the amino acid cysteine in a protein or peptide is a thiol group, any proteins that contain
17    cysteine are of importance  for interactions with As111. Although Table C-3 includes large
18    amounts of data under Effects Related to Oxidative Stress (ROS), arsenic's action  as a soft acid
19    and its nucleophilicity are not included as key events. It is  obvious, nonetheless, that those
20    chemical properties play important roles in the interactions of inorganic arsenic with organisms
21    at early stages in multiple key event(s) leading to tumor development.
22          Table  C-3 summarizes a great deal of data under Aberrant Gene or Protein Expression.
23    Abundant evidence is presented showing that changes can easily occur  at concentrations of As111
24    (as either sodium arsenite or arsenic trioxide) of less than 10 uM and often with durations of
25    exposure of 24 hours or less.  Results from 10 microarray analyses are found in this category,
26    and they all demonstrated changes in expression of large numbers of genes, often numbering in
27    the hundreds.  Two studies with longer exposures to especially low concentrations are of special
28    interest.  In one study, NB4 cells were exposed to 0.5 uM ATO for periods up to 72 hours for
29    transcriptome analysis and up to 48 hours for proteomic analysis.  The regulation of 487 genes
30    was affected at the transcriptome level; however, at the proteome level, 982 protein spots were
31    affected. The finding of more significant changes at the proteomic level, in comparison with the
32    relatively minor changes found at many of the corresponding genes at the transcriptome level,
33    suggests that ATO particularly enhances mechanisms of post-transcriptional/translational
34    modification (Zheng et al.,  2005). In the second experiment, which was a cDNA
35    (complementary DNA) microarray analysis of about 2,000 genes, the LOECs for SV40 large T-
36    transformed human urothelial cells (SV-HUC-1) exposed to As111, MMAm, or DMA111 for 25
37    passages (with subculturing twice weekly) were found to be 0.5, 0.05, and 0.2 uM, respectively.
38    DMA111 was shown to have a substantially different gene profile from the other two arsenicals.

                                                82       DRAFT—DO NOT CITE OR QUOTE

-------
 1    Most genes were down-regulated by these arsenicals, and evidence suggested that the
 2    suppression of two of these genes resulted from epigenetic hypermethylation (Su et al., 2006).
 3    Since each finding is presented only one time in Table C-3, subjectivity was often involved in
 4    the placement of data into the different key event categories. As a result, the densities of data in
 5    the different categories presented in Table 4-1 are only approximate estimates.  This situation
 6    was especially common for several key event categories that have large densities of data:
 7    Aberrant Gene or Protein Expression, Signal Transduction, and Effects Related to Oxidative
 8    Stress (ROS).
 9          Table C-3 also presents details on the genes and proteins affected and changes related to
10    dose and time. It also provides the possible significance of such changes, when available. A few
11    examples follow. When primary normal human epidermal keratinocytes (NHEK) cells were
12    exposed to 1  uM sodium arsenite for 24, 48, and 72 hours, there was an increase in focal
13    adhesion kinase (FAK) protein at 24 hours followed by a decrease to below the background level
14    at later times, with almost none being present at 72 hours (Lee et al., 2006b). The concentration
15    of some enzymes increased after exposures to 0.5 uM for 24 hours, but the concentrations
16    decreased at higher levels  of exposure up to 25 uM (Snow et al., 2001). DuMond and Singh
17    (2007) demonstrated the same relationship for proliferating cell nuclear antigen (PCNA) with
18    exposures to sodium arsenite lasting 70 days. The expression of PCNA increased at 0.008 uM,
19    but decreased at 0.77 and 7.7  uM.  Similar results have been observed for telomerase activity
20    (Zhang et al., 2003).  Numerous studies investigated effects of various modulators or inhibitors
21    or of different genetic conditions (e.g., knockout mutations or transfections).  Cell type can have
22    a major influence on the effect of arsenic on protein expression, as was shown for p53
23    expression, with some cells having no response to 50 uM sodium arsenite for 24 hours while
24    other cells showed an increase after exposure to only 1  uM sodium arsenite (Salazar et al., 1997).
25    Clearly, small levels of arsenic exposure can have large effects on many genes and proteins, and
26    the relationships involving time and dose can be complicated and subject to many influences.
27          Results found in the Apoptosis category show that ATO and sodium arsenite can often
28    induce apoptosis in cells with exposures to less than 10 uM (often much less) for a few days or
29    less.  Zhang et al. (2003) demonstrated a large difference in the sensitivity of cell lines to
30    arsenic-induced apoptosis. The authors found a positive association between telomerase activity
31    in cell lines and their susceptibility to induction of apoptosis by exposure to sodium arsenite.
32    Exposure to extremely low concentrations of sodium arsenite (i.e., 0.1-1 uM in HaCaT cells and
33    0.1-0.5 uM in HL-60 cells) for 5 days increased telomerase activity, maintained or elongated
34    telomere length, and promoted cell proliferation.  At higher concentrations, exposure of these
35    cell lines to sodium arsenite for 5 days decreased telomerase activity, decreased telomere length,
36    and induced apoptosis.  The positive association noted earlier means that cell lines that innately
37    have more telomerase activity are more likely to be affected by sodium arsenite in inducing

                                                 83       DRAFT—DO NOT CITE OR QUOTE

-------
 1    apoptosis.  Many experiments tested effects of modulators on the arsenic-induced apoptosis.  For
 2    example, Chen et al. (2006) demonstrated that co-treatment with L-buthionine-S,R-sulphoximine
 3    (BSO) markedly increased induction of apoptosis, presumably because of its effect in decreasing
 4    GSH levels. Other experiments looked at the effects of inhibitors of various proteins involved in
 5    signal transduction pathways. For example, Lunghi et al. (2005) showed that use of MAP/ERK
 6    kinase (MEK) 1  inhibitors greatly increased ATO-induced apoptosis. Other studies showed that
 7    different genetic conditions established using knockout mutations or transfections could
 8    markedly affect the extent of arsenic-induced apoptosis (e.g., Bustamante et al., 2005; Poonepalli
 9    et al., 2005; Ouyang et al., 2007). Many of the experiments  related to apoptosis were motivated
10    by the desire to improve methods for using ATO in cancer therapy, but in the process they have
11    provided much additional information about the complex pathways by which arsenic can affect
12    apoptosis.
13          In the hypothesized key event category Cancer Promotion, Tsuchiya et al. (2005) tested
14    sodium arsenite and three pentavalent arsenicals in a two-stage transformation assay in BALB/c
15    3T3 A31-1-1 cells.  Sodium arsenite caused cancer promotion at aLOEC of 0.5 uM when the
16    initiating treatment was exposure to 0.2 ug/mL 20-methylcholanthrene for 3 days before the 18-
17    day post-treatment with sodium arsenite.  Sodium arsenite caused promotion at a LOEC of 1  uM
18    when the initiating treatment was exposure to 10 uM sodium arsenite for 3 days before the 18-
19    day post treatment with sodium arsenite.  When Asv was tested in the same way with the same
20    initiating treatments, it was somewhat less potent, with LOECs of 1 and 5 uM respectively.  The
21    two methylated arsenicals had little or no effect. Paralleling their cancer promotion effects, the
22    same study demonstrated LOECs for As111 and Asv of 0.7 and 5 uM, respectively, for inhibition
23    of gap-junctional intercellular communication, which is a mechanism linked to many tumor
24    promoters.
25          The Cell  Cycle Arrest or Reduced Proliferation category includes many experiments that
26    showed that levels of exposure to ATO and sodium arsenite  of less than 10 uM (often much less)
27    for a few days or less can often increase the numbers of cells in mitosis and  otherwise disrupt
28    mitosis, so as to reduce cell proliferation. In the Drobna et al. (2002) experiment, the LOECs for
29    reduced cell proliferation were 1,1, and 5 uM for 24-hour exposures to As111, MMA111, and
30    DMA111, respectively; no effects were seen following exposures to the pentavalent forms of these
31    arsenicals at 200 uM.  By testing cells enriched in different phases of the cell cycle using
32    centrifugal elutriation, McCollum et al. (2005) showed that As111 slowed cell growth in every
33    phase of the cell cycle. Cell passage from any cell cycle phase to the next was inhibited by 5 uM
34    sodium arsenite. By looking at caspase activity, they showed that Asin-induced apoptosis
35    specifically in cell populations delayed in the G2/M phase.  Tests with knockout mutations
36    showed that poly(adenosine diphosphate-ribose) polymerase-1 (PARP-1) (Poonepalli et al.,
37    2005) and securin (Chao et al., 2006a) protect against arsenic-induced cell cycle disruption. Yih

                                                84       DRAFT—DO NOT CITE OR QUOTE

-------
 1    et al. (2005) provided evidence that 1 uM sodium arsenite appears to inhibit activation of the G2
 2    DNA damage checkpoint and thereby allows cells with damaged DNA to proceed from G2 into
 3    mitosis.
 4          Extremely small concentrations of As111 can stimulate cell proliferation.  For example,
 5    0.005 uM sodium arsenite exposure for 24 hours stimulated cell proliferation in NHEK;
 6    however, concentrations of 0.05 uM or higher inhibited it (Vega et al., 2001). In other studies,
 7    stimulation occurred at much higher concentrations. Mudipalli et al. (2005) exposed NHEK
 8    cells to many exposure levels of As111, MMAm, and DMA111 for 24 hours. The LOECs were 2,
 9    0.5, and 0.6 uM, respectively. There was increased stimulation of cell proliferation up to doses
10    of 6, 0.8, and 0.6 uM, respectively, and in all cases significant cytotoxicity was observed at
11    higher doses. Proliferation was often stimulated to a considerable extent. Yang et al. (2007)
12    showed that human embryo lung fibroblast (HELP) cells exposed to 0.5 uM sodium arsenite for
13    24 hours had 175% of the cell proliferation efficiency of control cells.  When the concentration
14    of As111 was increased to 5 uM, however, the cell proliferation efficiency decreased to 60% that
15    of the control. The increased proliferation rates can extend over long periods, as shown by
16    Bredfeldt et al. (2006), who exposed UROtsa cells to 0.05 uM MMA111 for  12, 24, or 52 weeks.
17    Cell population doubling times were 27, 25, and 21 hours, respectively, in comparison to the 42
18    hours observed in the control.
19          Mutations can play an important part in initiating carcinogenesis or in the development of
20    cancers, and they range from gene mutations that involve a single base-pair change to
21    chromosomal aberrations (CAs).  Much evidence is presented in Table C-3 under Chromosomal
22    Aberrations and/or Genetic Instability to show that inorganic arsenic can induce CAs, SCEs,
23    MN, multilocus deletions, and several other endpoints  such as changes in the length of
24    telomeres.  Arsenic appears to be ineffective in inducing  gene (point) mutations, but mutations at
25    some genes tend to be deletions that are so large that they extend over  several genes (termed
26    multilocus deletions). These multilocus deletions have been grouped with CA in Table C-3.
27    CD59 mutations (Liu et al., 2005) and gpt mutations (Klein et al., 2007) provide examples of
28    such mutations.  Numerous experiments are summarized in Table C-3  that show that CAs can be
29    induced by exposure to 10 uM or less of sodium arsenite for periods of 24 hours or less.
30    Following exposures of human primary peripheral blood lymphocytes  for 24 hours, LOECs for
31    As111, MMA111, and DMA111 were 2.5, 0.6, and 1.35 uM, respectively (Kligerman et al., 2003).
32    Examination of data shown in the table for the few  other experiments on MMA111 and DMA111 are
33    consistent with this experiment in  suggesting that both of those methylated arsenicals tend to be
34    more effective in inducing CAs than As111. The table includes estimates of about 15 LOECs for
35    induction of SCEs and about 20 LOECs for induction of MN following exposure to As111, and it
36    appears that CAs, SCEs, and MN are all induced to roughly the same extent by  As111. Some
37    experiments fail  to show a dose-response, which makes them difficult  to interpret.

                                                85       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Several of the experiments on CAs provided evidence of arsenic-induced changes in
 2    chromosome number (e.g., Barrett et al., 1989; Ochi et al., 2004). In the Ochi et al. (2004)
 3    experiment, DMA111 was much more potent than As111, and it induced mitotic spindle,
 4    centrosome, and microtubule elongation abnormalities.  Experiments on induction of MN were
 5    conducted in such a way as to distinguish between MN caused by aneuploidy and those caused
 6    by chromosomal breakage; these experiments provided evidence that both mechanisms may be
 7    important (e.g., Colognato et al., 2007; Ramirez et al., 2007). Chou et al. (2001) showed that
 8    exposure to 0.25 uM ATO for 4 weeks caused a decrease in telomere length.  Mouse embryo
 9    fibroblasts that are homozygous for the PARP knockout mutation were shown to be much more
10    sensitive to both arsenite-induced telomere attrition and induction of MN by As111 (Poonepalli et
11    al., 2005). Many experiments investigated the effects of various modulators on induction of
12    arsenic-induced chromosomal damage. For example, Jan et al. (2006) found that co-treatment
13    with low concentrations of dimercaptosuccinic acid, meso 2,3-dimercaptosuccinic acid (DMSA),
14    or 2,3-dimercaptopropane-l-sulfonic acid (DMPS) markedly increased the induction of MN by
15    sodium arsenite, ATO, MMA111, and DMA111, while co-treatment with high concentrations of the
16    same chemicals decreased the ability of arsenic to induce MN. Although the authors stated that
17    the reasons are obscure why these dithiol compounds effectively enhanced the toxic effects of
18    arsenic when they were at micromolar concentrations, they speculated that the observed  results
19    might be related to the influence of dithiols on retention of arsenite in cells, with low
20    concentrations of dithiols increasing arsenite levels and high concentrations of dithiol decreasing
21    them.  Ramirez et al. (2007) also showed that co-treatment with SAM blocked As111 induction of
22    centromere positive (cen+) MN without having any effect on its induction of centromere
23    negative (cen-) MN.  The authors suggested that the reason for this might be that SAM in some
24    way influences some components (probably microtubules) of the mitotic spindle.  As the main
25    methyl group donor,  SAM plays a major role  in chromatin methylation and condensation, and it
26    might stop the lagging of chromosomes by in some way correcting the cell's methylation status.
27    Alternatively, they suggested that SAM might interfere with the effects of ROS in causing
28    aneuploidy. Whatever SAM does to block induction of cen+ MN, it does not appear to affect
29    induction of double strand DNA breaks that would lead to cen- MN.
30          The results from the Co-Carcinogenesis category all relate to promotion of
31    benzo[a]pyrene (B[a]P)-mediated carcinogenesis via exposure to 1.5 uM sodium arsenite for 12
32    weeks.  Transformation (i.e., anchorage-independent growth in soft agar) of a rat lung epithelial
33    cell line occurred because of the arsenite treatment alone, and the transformed cells were shown
34    by proteomic analysis to have changes  in the amounts present of many proteins.  When the
35    arsenite treatment was preceded by exposure to 100 nM B[a]P for 24 hours, there was a
36    synergistic interaction. Results indicate that the transformation rate increased more than 500 and
37    200 times when compared to arsenite and B[a]P treatments alone, respectively. The findings in

                                                86       DRAFT—DO NOT CITE OR QUOTE

-------
 1    the proteomic analysis also showed synergistic interactions (Lau and Chiu, 2006). BPDE
 2    (benzo[a]pyrene diol epoxide) is an active metabolite of B[a]P. Shen et al. (2006) showed that a
 3    24-hour pretreatment of GM04312C cells, a SV-40 transformed XPA human fibroblast NER-
 4    deficient cell line, with 10 or 50 uM As111 markedly increased the cellular uptake of BPDE in a
 5    dose-dependent manner.
 6           The results found under Co-Mutagenesis showed that As111 affected the induction of
 7    mutations (using different assays) when there was also a treatment with UV, diepoxybutane
 8    (DEB), methyl methanesulfonate (MMS), X-radiation, gamma-radiation, or N-methyl-N-
 9    nitrosourea (MNU). Many of the types of mutations affected were gene mutations (i.e., point
10    mutations and numerous other changes in the DNA of single genes, such as small deficiencies),
11    which are not normally induced by arsenic alone. Arsenic treatment also caused co-mutagenesis
12    regarding CAs and MN.  Sometimes the timing of the As111 treatment relative to the treatment
13    with the other agent was of importance to the result observed. For example, a 24-hour
14    pretreatment with 10 uM sodium arsenite reduced the frequency of induction of hypoxanthine-
15    guanine phosphoribosyltransferase (HGPRT) mutations by MMS, but a 24-hour post-treatment
16    with the same concentration of sodium arsenite caused a synergistic interaction with MMS in
17    induction of HGPRT gene mutations (Lee et al., 1986).
18           The data found in Table  C-3 under Cytotoxicity are sometimes important to help
19    determine the possible relevance to human health of findings related to other key events. For
20    example, a large arsenic-induced increase in the expression of some protein that is important in
21    signal transduction is much more likely to have such relevance if it occurs at concentrations
22    having little or no cytotoxicity than if it occurs only when most cells are dying.  Table C-3 shows
23    that large differences in  LOECs for cytotoxicity can result from a change in any of the following
24    variables: species of arsenic, duration of treatment, cell line, and particular assay used.  As
25    another example, LOECs of As111 were 0.1 and 50 uM after 24-hour exposures in Jurkat cells and
26    HeLa cells, respectively (Salazar et al., 1997). Petrick et al. (2000) showed that three different
27    cytotoxicity assays yielded substantially different 24-hour LCSOs for each of five different
28    arsenic species. Sometimes the  different assays yield more similar results when treatments last
29    at least 48 hours (Komissarova et al., 2005). Overall it appears that in comparison to As111,
30    MMA111 has substantially higher cytotoxicity, DMA111 has higher cytotoxicity, and Asv has
31    sub stanti ally 1 ower cytotoxi city.
32           Effects of modulators  on arsenic-induced cytotoxicity were tested in many experiments.
33    Snow et al. (1999) showed that pretreatment with BSO, to decrease GSH levels, markedly
34    increased cytotoxicity of sodium arsenite following a 48-hour exposure. Jan et al. (2006) found
35    that co-treatment with low concentrations of DMSA or DMPS (dithiols that are currently used to
36    treat arsenic poisoning) markedly increased the cytotoxicity of ATO, while co-treatment with
37    high concentrations of DMSA or DMPS had the opposite effect. Probably the most important

                                                87        DRAFT—DO NOT CITE OR QUOTE

-------
 1    observation related to cytotoxicity from perusal of Table C-3 is that exposure of a large number
 2    of different cell lines to trivalent arsenicals results in significant cytotoxicity at molarities
 3    smaller than what would be found in urine, or even in the blood streams, of individuals exposed
 4    to high levels of inorganic arsenic in drinking water in places like Bangladesh. In some cell
 5    lines, even the pentavalent arsenicals destroyed more than 50% of the cells following a 7-day
 6    exposure with concentrations such as those observed in Bangladesh; As111 and MMA111 would do
 7    the same at concentrations far below such levels (Wang et al., 2007).  Also, from the numerous
 8    dose-response curves published in those papers, it is apparent that cytotoxicity generally has a
 9    threshold below which there is  no apparent effect.
10          DNA Damage is another key event category for which many experimental data are
11    summarized in Table C-3. Evidence showed induction of oxidative DNA damage, DNA single-
12    strand breaks,  and DNA-protein crosslinks by exposures at 10 uM (and often much less) of As111
13    for periods of often much less than one day. MMA111 is especially effective in inducing damage
14    detected by the comet assay (Gomez et al., 2005). Much more DNA damage was detected in the
15    comet assay by using enzyme treatments to reveal oxidative DNA adducts and DNA protein
16    crosslinks, and DNA damage was induced at levels of sodium arsenite that caused no
17    cytotoxicity in two different cell types (Wang et al., 2001). In a third cell type, no DNA damage
18    was observed up to the maximum concentration tested (2 uM), even though in each of the other
19    two cell types the LOEC was 0.25 uM. Jan et al. (2006) found that co-treatment with low
20    concentrations of DMSA or DMPS markedly  increased the DNA damage detected by the comet
21    assay following treatment with  ATO, while co-treatment with high  concentrations of DMSA or
22    DMPS had the opposite effect.  Several experiments looked at induction of 8-OHdG formation
23    as a measure of oxidative DNA damage.  In one such experiment, sodium arsenite was shown to
24    be effective. However, MMA111 was shown to be about 200 times more effective than As111 (with
25    an LOEC  of 0.05 uM) following a 1-hour treatment (Eblin et al., 2006). Pre-incubation with
26    SOD or catalase to reduce effects of ROS almost completely blocked induction of 8-OHdG
27    formation by a 24-hour treatment with sodium arsenite (Ding et al., 2005). Tests with a cell line
28    containing a knockout mutation of the PARP-1 gene showed that the PARP-1 protein protects
29    against arsenic-induced DNA damage detected by the comet assay at pH >13 in the version of
30    the assay that does not include further digestion to detect additional types of DNA damage
31    (Poonepalli et al., 2005).
32          The DNA Repair Inhibition or Stimulation category includes rather few experiments in
33    Table C-3. A microarray experiment that showed decreased expression of DNA repair genes
34    involved exposure to only 0.77  uM of sodium arsenite for 70 days (DuMond and Singh, 2007).
35    Arsenic does not always have the effect of decreasing repair. Snow et al. (2005) found that
36    W138 cells exposed to 0.1 uM  sodium arsenite for 24 hours showed increased DNA ligase
37    activity. Increasing the As111 concentration to  1 uM further increased the activity, but 5 uM

                                                88       DRAFT—DO NOT CITE OR QUOTE

-------
 1    decreased DNA ligase activity to below normal levels.  The same paper demonstrated a rather
 2    similar reversal-of-direction effect for DNA polymerase p. In another experiment, when CHO
 3    Kl cells were treated with MMS followed by 5 uM sodium arsenite for 6 hours, there was a
 4    decrease in repair of MMS-induced single-strand breaks in DNA (Lee-Chen et al., 1993).
 5    Andrew et al. (2006) demonstrated that in Jurkat cells the LOEC for sodium arsenite was 0.01
 6    uM for reduction of expression of NER gene ERCC1 (excision repair cross-complement 1
 7    component).  The decrease in expression was 45% at that concentration and 60% at
 8    concentrations of 0.1 and 1 uM. The functional effect of this decrease in expression was shown
 9    by reduced repair following a challenge with the mutagen 2-AAAF immediately after the sodium
10    arsenite treatment.   Clearly, exposure to inorganic arsenic at low concentrations can modify the
11    level of DNA repair.
12          The Effects Related to Oxidative Stress (ROS) category in Table C-3 includes many
13    experiments in which antioxidants or radical scavengers were used as modulators. When a
14    reduction in the effects was seen, it was taken as evidence that oxidative stress was the cause  of
15    the original effects observed, as, for example, in the study by Sasaki et al. (2007).  Results from a
16    series of experiments by Lynn et al. (2000) led to the conclusion that As111 activates NADH
17    oxidase to produce  superoxide, which then causes oxidative damage to DNA. Experiments by
18    Liu et al. (2005) dealt with the effects of various modulators on induction of CD59- mutations
19    and lead to the conclusion that peroxynitrites, which are formed as a result of ROS and reactive
20    nitrogen species, have an important role in the induction by As111 of such mutations.  Wang et al.
21    (2007) measured formation of oxidative damage to lipids, proteins, and DNA (comet assay) by
22    three trivalent arsenicals and three pentavalent arsenicals in two different cell lines.  For As111,
23    Asv, MMAm, and DMA111, the LOECs were all 0.2 uM for a 24-hour exposure for all three types
24    of damage. The order of effectiveness of the different arsenicals differed in the two cell lines
25    used and for the different types of damage. Consistent with these effects, increased levels of
26    nitric oxide, superoxide ions, hydrogen peroxide, and the cellular free iron pool were
27    consistently detected in both cell lines after treatments by all three trivalent arsenicals. A
28    microarray analysis in which genes were identified for which the response to ATO and hydrogen
29    peroxide was reversed by n-acetyl-cysteine (NAC) suggested that 26% of the genes significantly
30    responsive to ATO were directly altered by ROS (Chou et al., 2005). Further evidence that ROS
31    is likely involved in arsenite-induced DNA damage comes from comet assays done on splenic
32    lymphocytes from  SOD knockout mice (Kligerman and Tennant, 2007).  Results showed
33    homozygotes exhibiting a large decrease in splenic SOD levels and a large increase in arsenite-
34    induced DNA damage, while heterozygotes had intermediate changes in SOD levels and DNA
35    damage.
36          Table C-3 includes little information on Enzyme Activity Inhibition. Hu et al. (1998)  and
37    Snow et al. (1999) tested the effect of sodium arsenite on the activity of several purified enzymes

                                                89        DRAFT—DO NOT CITE OR QUOTE

-------
 1   in vitro, including enzymes required for DNA repair and some related to GSH metabolism.  The
 2   purpose of the study was to examine whether As111 binding to sulfhydryls caused protein
 3   denaturation and inhibited enzyme activity.  In almost all cases, the purified enzymes were not
 4   inhibited by physiologically relevant concentration of As111. The concentrations that are needed
 5   to cause 50% inhibition (ICSOs) for the rate of the reaction (over 6 minutes for many of those
 6   enzymes) ranged from 6.3 to 381 mM.  The one exception was purified pyruvate dehydrogenase
 7   for which the IC50 was 5.6 uM. Table C-3  also lists ICSOs for GSH peroxidase and ligase when
 8   tested in extracts of AG06 (SV40-transformed human keratinocyte) cells that were pretreated for
 9   24 hours with an unspecified concentration of sodium arsenite; these ICSOs were both low, i.e.,
10   2.0 and 14.5 uM, respectively.
11          Table C-3, under Gene Amplification, shows that As111 caused amplification of
12   dihydrofolate reductase (dhfr) genes in three different experiments with LOECs ranging from
13   0.0125 to 6 uM (Barrett et al., 1989; Rossman and Wolosin, 1992; Mure et al., 2003).  Takahashi
14   et al. (2002) showed that several neoplastic transformed cell lines produced by 48-hour
15   treatments with either < 8 uM As111 or < 150 uM Asv contained gene amplification of either the
16   c-Ha-ras or the c-myc oncogene. Almost all of the data in Table C-3 for Gene Mutations show
17   no induction of mutations by arsenic.
18          Hypermethylation of DNA was demonstrated in a number of specific DNA sequences in
19   two human kidney carcinoma cell lines and in one human lung carcinoma cell line.  In the lung
20   cell line, the LOEC for As111 was 0.08 uM for a 7-day exposure, and there was a positive dose-
21   response extending over the two higher doses tested (0.4 and 2.0 uM). Hypermethylation in this
22   cell line was demonstrated within a 341-base-pair fragment of the promoter region of p53 (Mass
23   and Wang, 1997; Zhong and Mass, 2001).
24          Hypomethylation of DNA has been demonstrated globally and for a number of specific
25   DNA sequences. In one instance, exposure of HaCaT cells to 0.2 uM  sodium arsenite for 10
26   serial passages in folic-acid depleted media caused genomic hypomethylation.  Sodium arsenite
27   repressed the expression of the DNA methyltransferase (DNMT) genes DNMT1 and DNMT3A
28   and caused depletion of SAM, the main cellular methyl donor. It is thought that long-term
29   exposure to sodium arsenite may have resulted in DNA hypomethylation as a consequence of
30   those two complementary mechanisms (Reichard et al., 2007). Singh and DuMond (2007)
31   demonstrated methylation changes in DNA  at 18 genetic loci in TM3 cells, with some showing
32   hypomethylation and others hypermethylation, following sodium arsenite exposures ranging
33   from 0.008-7.7 uM that lasted for either 25  or 75 days. The LOEC was the lowest dose.  Some
34   loci were affected only after 25 days of exposure, while others were affected after 75 days of
35   exposure.  In one of several other demonstrations of hypomethylation, a 19-week exposure of
36   TRL 1215 cells to 0.125 uM sodium arsenite was sufficient to cause global hypomethylation
37   (Zhao etal., 1997).

                                               90       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Under Immune System Response, Table C-3 describes a wide-range of effects on the
 2    immune system. This discussion provides highlights from that table and Appendix D, which is
 3    devoted entirely to the immunotoxicity of inorganic arsenic.  Appendix D discusses some aspects
 4    of the immunotoxicity of inorganic arsenic in much more detail, including more emphasis on
 5    human studies and in vivo experiments on laboratory animals, as well as on some older in vitro
 6    studies. It overlaps very little with data found in Table C-3.  Effects thought to be related to
 7    Immune System Response were grouped under that heading in Table C-3 even if they dealt
 8    mainly with other key events. For example, several findings related to Apoptosis, Cytotoxicity,
 9    or Signal Transduction are included in this section of Table C-3.
10          Exposures to low concentrations of As111 over 1-2 weeks inhibited maturation of human
11    peripheral blood monocytes (HPBMs) into the following types of cells: M-type and GM-type
12    macrophages, immature  dendritic cells, and multinucleated giant cells (Sakurai et al., 2006). The
13    ICSOs for this inhibition  ranged from 0.06 to  0.70 uM.  Lemarie et al. (2006a) showed that ATO
14    inhibited macrophage differentiation of peripheral blood mononuclear cells (PBMCs) and that
15    concentrations as low as 0.125 uM over 6 days induced apoptosis and necrosis in PBMCs co-
16    treated with granulocyte-macrophage colony-stimulating factor (GM-CSF) or macrophage
17    colony-stimulating factor (M-CSF). Differentiated macrophages developed from PBMCs treated
18    with GM-CSF for 6 days were exposed to 0.25 uM ATO for  6 days. The ATO treatment caused
19    major alterations in morphology, adhesion, and actin organization, giving the impression that the
20    ATO "de-differentiated" the macrophages back into monocytic cells (Lemarie et al., 2006b).
21    The same series of experiments showed that macrophages exposed to 1 uM ATO for 6 days also
22    caused a reduction in several surface markers, markedly decreased endocytosis and
23    phagocytosis, and increased the secretion of inflammatory cytokines in response to a co-
24    treatment with lipopolysaccharide.
25          Exposure of PBMCs that had been stimulated with phytohemagglutinin (PHA) after
26    exposure to 1-5 uM sodium arsenite for 120  hours caused a marked dose-related decrease in
27    both cell proliferation and the percentage  of divided cells (Tenorio and Saavedra, 2005).  Even at
28    the higher doses, most of the cells were viable but unable to divide. The treatments also
29    modified the expression  of CD4 and CDS molecules. Judging from evaluation of blast
30    transformation, CD4+ and CD8+ T cells appear to have  different sensitivities to As111. As the
31    concentration of the sodium arsenite increased from 1 to 5 uM in the 120-hour treatment, there
32    was an accumulation of resting  CD8+ cells with a positive dose-response, but there was not an
33    accumulation of CD4+ cells. The Janus kinase (JAK)-signal transducer and activator of
34    transcription (STAT) pathway is an essential  cascade for mediating normal functions of different
35    cytokines in the development of the hematopoietic and  immune systems.  Huang et al. (2007a)
36    showed that exposure of SV-HUC-1 cells to sodium arsenite for 48 hours caused changes in
37    levels of proteins that are part of that cascade, and the LOEC was 2 uM. Sometimes there was a

                                                91        DRAFT—DO NOT CITE OR QUOTE

-------
 1    dose-response, and sometimes the direction of the change reversed.  Cheng et al. (2004) showed
 2    that a 48-hour pretreatment of HepG2 cells with 4 uM sodium arsenite was sufficient to block
 3    induction of STATS activity by an IL-6 treatment. Other experiments showed that As111 acted
 4    directly on the JAK1 protein to cause JAK-STAT inactivation. Di Gioacchino et al. (2007)
 5    studied the effects of several arsenicals on PBMC proliferation and cytokine release. At a
 6    concentration of 100 uM, sodium arsenite was effective in decreasing PHA-induced cell
 7    proliferation and in reducing interferon-gamma (IFN-y) and TNF-a release.  However, at a
 8    concentration of 0.1 uM, As111 significantly increased cell proliferation.  More details about that
 9    experiment are found in Appendix D.
10          Regarding Inhibition of Differentiation, in experiments done on spontaneously
11    immortalized human keratinocytes and on normal human epidermal  cells derived from foreskin,
12    sodium arsenite was shown to delay differentiation and preserve the proliferative potential of
13    keratinocytes (Patterson et al., 2005; Patterson and Rice, 2007). A concentration of sodium
14    arsenite as low as 0.1 uM over 4 days had a noticeable effect, but most experiments were done
15    using 2 uM sodium arsenite over 4-14 days, which yielded  a much larger effect.  Treatment of
16    C3H 10T1/2 cells with 6 uM sodium arsenite for 8 weeks completely inhibited their
17    differentiation into adipocytes following dexamethasone/insulin treatment, and treatment with
18    3 uM sodium arsenite for only 48 hours was the LOEC for that effect (Trouba et al., 2000).
19          Interference With Hormone Function was demonstrated in experiments by Bodwell et al.
20    (2004, 2006).  Some effects were observed at approximately 0.09  uM of sodium arsenite;
21    however, the increases  found in glucocorticoid-receptor-mediated gene transcription of reporter
22    genes that contained tyrosine aminotransferase (TAT) response elements were highly dependent
23    on, and inversely related to, the amount of activated  steroid receptor within cells. More detailed
24    information on interference with hormone function can be found in Table C-3.
25          Under Malignant Transformation or Morphological  Transformation, Table C-3 shows
26    that concentrations of less than 1 uM of As111, MMA111, or DMA111 are capable of causing
27    transformation. HaCaT cells exposed to 0.5 uM As111 for 20 passages caused the cells to become
28    tumorigenic, as shown by production of tumors 2 months after injection into Balb/c nude mice
29    (Chien et al., 2004). Zhao et al. (1997) found similar results with another cell line after 18 weeks
30    of exposure to 0.25 uM As111. UROtsa cells exposed to 0.05 uM MMA111 for 52 weeks caused
31    anchorage-independent growth as detected by colony formation in soft agar, and cells from those
32    colonies  showed enhanced tumorigenicity in SCID mouse xenographs (Bredfeldt et al., 2006).
33    After 26  weeks, this experiment showed much anchorage-independent growth but not yet
34    enhanced tumorigenicity. Syrian hamster ovary (SHE) cells exposed to DMA111 for 48 hours
35    showed morphological  transformation at a concentration of only 0.1 uM, and at the highest dose
36    tested of 1.0 uM, 3.35% of the surviving colonies had become transformed (Ochi et al., 2004).
                                                92       DRAFT—DO NOT CITE OR QUOTE

-------
 1    In contrast, at a dose of 10 uM after the same exposure duration of 48 hours, As111 had only
 2    transformed 0.48% of the surviving cells.
 3          Table C-3 summarizes many findings related to the Signal Transduction category, even
 4    though considerable data found under Aberrant Gene or Protein Expression could have been
 5    placed into this category. Most of the data in this category are for sodium arsenite or ATO.  In
 6    addition, there are numerous LOECs smaller than 10 uM (often much less), and they are often
 7    for treatments that lasted much less than one day. Drobna et al. (2002) evaluated
 8    phosphorylation of extracellular signal-regulated kinase (ERK)-2, activator protein (AP)-l
 9    binding activity, and phosphorylation of c-Jun (an AP-1 protein) by six arsenicals in treatments
10    lasting up to 2 hours.  Asv, MMAV, and DMAV were all tested  at concentrations up to 100 uM
11    and had no effect.  As111, MMA111, and DMA111 each had an LOEC of 0.1 for at least one endpoint.
12    Details presented  in Table C-3 show that the responses of those three arsenicals were different
13    and that, in some cases, the direction of the response reversed as the concentration increased. In
14    some cases a reduction from an increase was observed, which is interesting because various
15    responses for some endpoints described above showed a reversal in which the lowest doses
16    caused a bigger effect. Another experiment showing a reversal in response (from a decrease to
17    an increase) was for phosphorylation of Akt Thr308 in JB6 C141  cells (P+ mouse epidermal cell
18    line) (Ouyang et al., 2006).  Following 1-hour exposures to sodium arsenite, there was slight
19    decrease at 0.1 uM, a larger decrease at 0.5 uM, increases above the control level at 1 and 5 uM,
20    and a much larger  increase at 10 uM. Additionally, several experiments in this category related
21    to different ways in which arsenic affects signal transduction to either increase or decrease
22    apoptosis.  For example, MCF-7 cells exposed to 2 uM ATO for 1 hour activated the pro-
23    survival MEK/ERK pathway (Ye et al., 2005).  By decreasing apoptosis,  such an effect might
24    permit the survival of cells containing damage that could eventually lead  to a cancer.  Yancy et
25    al. (2005) did a series of experiments on H9c2 cells (an immortalized myoblast cell line derived
26    from fetal rat hearts) and concluded that sodium arsenite exposure decreases cell migration
27    through an effect on focal adhesions and by  disrupting cell interactions with the extra-cellular
28    matrix. Focal adhesions are involved in integrin signaling. Florea et al. (2007) showed that
29    ATO triggered three different kinds of Ca2+  signals (i.e., steady state increases, transient
30    elevations, and calcium spikes). The Ca2+ concentration in cells was substantially increased (and
31    by rather similar amounts) by exposure to either 0.1 or 1  uM ATO for about 1 hour in two
32    different cell lines (i.e., the  human neuroblastoma cell line SY-5Y and the human embryonic
33    kidney cell line HEK 293).

      4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER  EFFECTS
34          Not addressed in this document.
                                                93        DRAFT—DO NOT CITE OR QUOTE

-------
      4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER CHARACTERIZATION

      4.6.1. Summary of Overall Weight-of-Evidence
 1          Based upon the EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a)
 2    inorganic arsenic is categorized as "carcinogenic to humans" due to convincing epidemiological
 3    evidence of a causal relationship between oral exposure of humans to inorganic arsenic and
 4    cancer. Arsenic is a multisite carcinogen, with numerous studies finding an association between
 5    arsenic and increased incidences of a number of different types of cancers. The carcinogenic
 6    effect of arsenic has been reported for populations in many different countries. While the studies
 7    detailed in this document provide evidence for cancer after oral exposure to arsenic, arsenic also
 8    has been associated with cancer after inhalation exposure (U.S. EPA, 1994).
      4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence
 9          Numerous epidemiologic investigations,  each conducted differently and containing its
10    own biases (e.g., lack of confounding variables, possible recall bias), provide support for an
11    association between oral exposure to inorganic arsenic and cancer including skin, bladder,
12    kidney, lung, liver, and prostate. The most extensively studied population is from southwest
13    Taiwan.  This is because between 1910 and 1920, water supplies were changed from shallow
14    surface water wells to artesian wells, which were subsequently found to contain high levels of
15    arsenic in various regions.  Studies in these arsenic-endemic regions of Taiwan have found
16    increases in all of the aforementioned cancer types.  The link between these cancers and arsenic
17    exposure in drinking water also have been observed in other parts of the world, including Japan,
18    Chile, and Argentina. Therefore, it is unlikely that any single environmental factor (e.g.,
19    nutritional habits) associated with a single population is entirely responsible for the increased
20    cancer rates. Although many studies did not account for confounding variables (e.g., cigarette
21    smoking in association with lung cancer), the positive associations between arsenic intake and
22    cancer risk were still observed in studies that did account for confounding variables (e.g.,
23    lifestyle habits, age, and socioeconomic status).
24          Most of the epidemiology studies examining the relationship between arsenic exposure
25    from drinking water and cancers are ecological in nature and are therefore subject to the
26    limitations inherent in such studies (e.g., lack of measured individual exposure).  For a number
27    of reasons, the southwest Taiwanese database remains the most appropriate source for estimating
28    bladder and lung cancer risk among humans (NRC, 1999, 2001; SAB, 2000, 2007), despite
29    lacking individual water consumption and nonwater arsenic intake. Strengths of the data include
30    the size of the population, the reliability of the population and mortality counts, the stability of
31    residential patterns, the homogenous lifestyle as  confirmed by surveys, the long-term exposures,
32    the extensive follow-up (almost 900,000 person-years), the large number of exposed villages
33    (42), and the large number of cancer deaths (1152 recorded from 1973 to 1986).  Population

                                                94       DRAFT—DO NOT CITE OR QUOTE

-------
 1    records in Taiwan have been well kept since 1905, and death certificates include all primary
 2    cancers. In addition, cancer cases were pathologically confirmed in some of the Taiwanese
 3    studies.
 4          Although dose-response relationships have been observed for the majority of cancers
 5    noted in areas with high levels of arsenic in their drinking water, results for low-level arsenic
 6    epidemiologic investigations (primarily from the United States and Europe) have been equivocal
 7    with regard to the relationship between these cancers and arsenic exposure. This could be due to
 8    the fact that none of the studies accounted for arsenic exposure through food sources. Kile et al.
 9    (2007) found that as the level of arsenic in the water decreased for women in Bangladesh, the
10    contribution of arsenic from dietary sources became of greater importance.  Uchino et al. (2006)
11    found that with concentrations of 50 ppb or less of arsenic in the drinking water in a population
12    in West Bengal, India, the contribution of arsenic from food was the main source of arsenic
13    exposure (i.e., contribution from water with less than 50 ppb was less than 27% of the total
14    arsenic consumed). Therefore, as the exposure of arsenic from drinking water decreases and the
15    relative contribution from food increases, misclassification of exposure groups can become
16    significant.  The average estimate of inorganic arsenic consumption in food ranges from  1.34
17    ug/day in infants to 18 ug/day in adults, for a total arsenic average of 62 ug/day for people in the
18    United States  (NRC, 1999). At the lower concentrations, dietary intake could easily create total
19    arsenic intake levels to be similar between the referent group and what is considered the
20    exposure group.
21          Cantor and Lubin (2007) also conclude that misclassification  occurs because exposure is
22    not necessarily assessed during disease-relevant exposure periods.  In regards  to cancer, there is
23    a long latency period, which appears to vary depending on the type of cancer and exposure.  This
24    means that exposure to arsenic sources during the decades prior to cancer outcome is necessary.
25    Therefore, studies with low levels of exposure that are ecological in nature (no individual
26    exposure) are  more prone to misclassification, which means they are  biased toward the null
27    hypothesis.  In addition, studies that attempted to individualize exposure by examining toenail
28    arsenic levels  are looking at only the prior year of exposure (Cantor and Lubin, 2007) and may
29    miss the important exposure period. Despite all these numerous limitations in low-level
30    exposure studies, significant associations have been observed for cancers of the prostate
31    (Hinwood et al., 1999; Lewis et al., 1999), skin (Hinwood et al., 1999; Karagas et al., 2001;
32    Beane-Freeman et al., 2004; Knobeloch et al., 2006), and bladder (Kurttio et al., 1999;
33    Steinmaus et al., 2003; Karagas et al., 2004). In most cases, however, there is no dose-response
34    with increases observed at the highest concentrations only and in many cases significant  results
35    occurred in smokers only.
36          There are very few animal data demonstrating the carcinogenic potential of arsenic. This
37    is likely due to the fact that rodents, which are the most likely animal model, are better

                                                 95        DRAFT—DO NOT CITE OR QUOTE

-------
 1    methylators of arsenic than humans (Vahter, 1999a). Since it has been noted that humans who
 2    are better methylators are at lower risk (Yu et al., 2000; Chen et al., 2005a; Steinmaus et al.,
 3    2005; Valenzuela et al., 2005; Ahsan et al., 2007; Huang et al., 2007b; McCarthy et al., 2007a),
 4    it is not surprising that animals that are better methylators are at even lower risk.  As stated
 5    before, arsenic has been associated with cancers of the skin, lung, kidney, bladder, and liver.
 6    Below is a summary these different types of cancers and their association with arsenic exposure
 7    in drinking water.
      4.6.2.1. Skin Cancer
 8          Epidemiologic investigations of populations in the arseniasis-endemic areas of Taiwan
 9    have shown that exposure to arsenic from drinking water is associated with skin cancer (Tseng et
10    al., 1968; Tseng, 1977; Chen  et al., 1985, 1988a,b; Wu et al., 1989; Chen and Wang, 1990; Tsai
11    et al., 1999). The prevalence rate for skin cancer showed an increasing gradient according to the
12    arsenic content of the well water.  Guo et al. (2001) found significant increases in SCCs at the
13    highest dose only (>640 ppb) with results at lower doses variable, suggesting that skin cancers
14    may be cell-type specific. Contrastingly, Karagas et al. (2001) found increases in both SCC and
15    BCC in the highest toenail arsenic concentration in a population in  the United States. Beane-
16    Freeman et al. (2004) also found an increase in the risk of melanoma with elevated toenail
17    arsenic concentrations.  Therefore, these results demonstrate that skin cancers may not be cell-
18    type-specific.  Although Taiwan has been the area most associated with skin  cancers in relation
19    to arsenic exposure, the association has been made in other populations as well.  Arsenic has also
20    been associated with skin cancers in Argentina, where signs of arsenicism also have been
21    observed (Smith et al., 1998). Hopenhayn-Rich et al. (1998),  however, found a significant
22    association in women in the highest category and surprisingly in males in the lowest category
23    only. Skin cancer has also been found in China with drinking water concentrations of 150 ppb or
24    greater (Lamm et al., 2007). Skin cancer was not found associated  with arsenic in Denmark
25    (Baastrup et al., 2008) or in the United States (Meliker et al., 2007), but these studies were at
26    lower concentrations of arsenic.
27          Skin tumors have only been induced in transgenic mice or with subsequent TPA or UV
28    exposure (indicating co-carcinogenesis) in mice.  Because co-carcinogenesis has been
29    demonstrated in animal models, it is possible that the same  occurs in humans. Sun exposure
30    would likely be high and the use of sunblock is less likely in the areas where skin cancer has
31    been noted (i.e., Taiwan and Argentina).  Therefore, a possible co-carcinogenic effect also may
32    be contributing to the association.
      4.6.2.2. Lung Cancer
33          Lung cancer has been associated with arsenic in populations that were exposed to
34    exceedingly high arsenic levels in Taiwan, Chile, and Argentina. Studies of populations with

                                                 96       DRAFT—DO NOT CITE OR QUOTE

-------
 1    lower arsenic exposure, especially <50 ppb, have not conclusively found an association between
 2    arsenic and lung cancer.  Lung cancer was not associated with arsenic exposure in the United
 3    States (Lewis et al., 1999 and Meliker et al., 2007), Denmark (Baastrup et al., 2008), or Australia
 4    (Hinwood et al., 1999).  Yang et al. (2004) found that lung cancer incidence in endemic areas of
 5    Taiwan remained elevated even after the use of the arsenic-containing well water ceased. Yuan
 6    et al. (2007) also found that mortality from lung cancers exceeded that observed in regions with
 7    consistently low arsenic exposure even after a 10- to 20-year lag period after removal of the
 8    arsenic source. These were likely due to the long latency for cancer.  Many of the studies have
 9    not controlled for smoking  history, which is a potential confounder for lung cancer.
      4.6.2.3. Kidney, Bladder, and Liver Cancer
10           Significant increases in mortality rates for cancers of the kidney, bladder, and liver have
11    been identified in populations from Taiwan, Argentina, and Chile.  These three regions all have
12    elevated levels of arsenic exposure through drinking water.  Yang et al. (2004) found that arsenic
13    was associated with kidney cancers in Taiwan. Unlike lung  cancer, the mortality associated with
14    kidney cancer decreased after reducing arsenic exposure. Yang et al. (2005) also found a
15    reduction in bladder cancer after removal of arsenic exposure (through tap water instillation), but
16    the decline was gradual.  In Chile, supplementation of drinking water with water from rivers
17    caused exposure to high levels of arsenic, but after the installation of improved water treatment
18    in the early 1970s, arsenic exposure dropped dramatically. Yuan et al. (2007), however, found
19    that even after a 10- to 20-year lag period after removal of the arsenic source, mortality from
20    bladder cancers still exceeded that observed in regions with consistently low arsenic exposure.
21          While high levels of arsenic have been found to be related to bladder, kidney, and liver
22    cancers, low-dose exposures from the United States, Europe, and Australia have been less clear.
23    Lewis et al. (1999) observed increased SMRs in kidney cancer for both males (SMR=1.75) and
24    females (SMR=1.60), but the results were not significant.  Because the highest concentration in
25    this population was 166 ppb, the results are still noteworthy. Kurttio et al. (1999) found that
26    despite the low levels of arsenic (median = 0.1 ppb; max=64 ppb) there was evidence of a
27    relationship between exposure to arsenic at levels above 0.5  ppb and bladder cancer risk. No
28    association was observed for kidney cancer risk. Hinwood et al. (1999), Meliker et al. (2007),
29    and Baastrup et al. (2008) did not find associations between  these cancers and the low levels of
30    exposure in Australia, the United States, and Denmark.
31          Although inorganic arsenic exposure in rodents has not been observed to cause increases
32    in cancer, long-term (104 weeks) exposure to DMAV in rats  has been found to increase bladder
33    tumors with doses of 50 ppm or greater.  These concentrations are quite high in comparison to
34    the amount of inorganic arsenic exposure in humans.
                                                97       DRAFT—DO NOT CITE OR QUOTE

-------
      4.6.2.4. In Utero Exposure
 1          There is no adult animal model available to study the relationship between arsenic
 2    exposure via drinking water and cancer outcome; however, lung and liver tumors have been
 3    induced by inorganic arsenic in mice when exposed during gestation. Pregnant dams were
 4    exposed for 10 days during gestation only; this increases the evidence that lung and liver cancers
 5    are associated with oral exposure to inorganic arsenic. Reproductive and adrenal tumors also
 6    have been observed with transplacental exposure in mice.
 7          There is very little epidemiology information specifically linking in utero arsenic
 8    exposure to cancer outcome. Although the available epidemiological studies conducted in
 9    Taiwan and other countries included women of reproductive age, the cancer outcomes from adult
10    exposures were not differentiated from in utero exposures. Recently, Smith et al. (2006)
11    examined lung cancer rates (and other respiratory diseases) in cohorts born just before the peak
12    exposure period in Antofagasta, Chile (meaning that they were not exposed in utero to high
13    levels of arsenic, but were exposed during childhood) and cohorts born during the high-exposure
14    period (indicating likely in utero exposure). Results demonstrated that exposure during either
15    period of development caused increased risk of lung cancer; however, the results from early
16    childhood exposures and/or in utero exposures were not compared to exposures during adulthood
17    to determine the possible cancer sensitivity effects in humans.
18          Because both in utero studies in mice and a study in humans by Smith et al. (2006)
19    indicate that lung cancer development may be associated with transplacental arsenic exposure,
20    there is an opportunity to examine the similarities in mechanistic effects mediating lung cancers
21    between the two species.  Several PBPK models exist for humans (Yu, 1999a,b; El-Masri and
22    Kenyon, 2008) and mice (Gentry et al., 2004). However, these studies are inadequate in
23    interpreting the findings from the in utero studies in mice and relating them to human exposure
24    concentrations.
      4.6.3. Mode of Action Information
      4.6.3.1. General Comments on MOAs
25          The carcinogenic MOA for inorganic arsenic is unknown. Multiple MOAs for inorganic
26    As seem likely in view of the numerous ways in which arsenic acts upon living organisms and
27    the several metabolites produced before it is excreted from the body. While this review focuses
28    on inorganic As, the methylated species produced during its metabolism, especially the highly
29    reactive MMA111 and DMA111, probably play an important role in the carcinogenesis of inorganic
30    arsenic consumed in drinking water. Each successive product in the metabolic pathway has its
31    own toxicity and carcinogenic potential, with possible differential transport into and out of
32    different organs.  In comparison to laboratory animals, humans excrete more MMA in urine and
33    are more prone to arsenic-induced carcinogenesis.  These findings suggest that MMA (probably

                                                98        DRAFT—DO NOT CITE OR QUOTE

-------
 1    in the trivalent form) may be of special importance to arsenic-induced carcinogenesis in humans.
 2    The finding of numerous different tumor types associated with arsenic exposure both in humans
 3    and transplacental animal models also supports the view that multiple MOAs are likely. Due to
 4    the complexities of the available data related to MO A, including the range of possible toxicities
 5    of the different arsenic species, the different levels of each arsenic compound in target tissues,
 6    multiple hypothesized key events, and multiple tissue tumor effects in humans, there is a need
 7    for improved PBPK models to assist in understanding the MO A. Although there are several
 8    PBPK models available (see Section 3.5), none have sufficiently addressed the complex nature
 9    of the kinetics associated with arsenic carcinogenesis; therefore, this is an ongoing effort along
10    with BBDR modeling.
11           It seems useful to describe a few MOAs for cancer to use as a frame of reference when
12    considering arsenic specifically. Although inorganic arsenic and its metabolites have not been
13    found to induce gene (point) mutations, the key events involved in mutagenesis—i.e., (1)
14    exposure of target or stem cells; (2) reaction with DNA to produce DNA damage; (3)
15    misreplication  of a damaged DNA template or misrepair of DNA damage leading to a mutation
16    in a critical gene in the replicating target cell; (4) replication forming a clone of mutated cells;
17    (5) DNA replication, possibly leading to additional mutations in critical genes; (6) unbalanced
18    and uncontrolled clonal growth of mutant cells, possibly leading to pre-neoplastic lesions; (7)
19    progression of pre-neoplastic cells in those lesions, resulting in emergence of overt neoplasms,
20    solid tumors (which require neoangiogenesis), or leukemia; (8) additional mutations in critical
21    genes occurring as a result of uncontrolled cell division; and (9) cancer occurring due to
22    malignant behavior (adapted from Preston and Williams, 2005)—may contribute to one or more
23    arsenic-mediated MOA(s) for carcinogenesis.  A mutagen with the above MOA would likely be
24    thought to have a linear dose-response.  It is unclear what the shape of the dose-response curve is
25    for any specific key event that might be involved in the MOA for arsenic and its metabolites.
26    Therefore, a linear dose-response is the prudent choice unless the dose-response of the identified
27    key events mediating the carcinogenesis is fully understood.
28           A second example of a MOA is the one hypothesized for arsenical-induced urinary
29    bladder carcinogenesis as follows: after the requisite arsenical ingestion, absorption, and
30    metabolism, (1) DMA111 is excreted into urine above a critical concentration, (2) it reacts with
31    urothelial critical sulfhydryl groups,  (3) urothelial cytotoxicity and necrosis results, (4) urothelial
32    regenerative cell proliferation (hyperplasia) results, and (5) urothelial cancer develops; oxidative
33    damage might possibly stimulate both steps 3 and 4 (adapted from Cohen et al., 2007).
34    Obviously this MOA directly relates to the topic of this review, and any combination of factors
35    in which consumption of inorganic arsenic would lead to more than the critical (threshold)
36    concentration of DMA111 for a particular individual for a sufficient time could result in bladder
37    cancer.

                                                 99        DRAFT—DO NOT CITE OR QUOTE

-------
 1           Section 4.4.1 provided abundant evidence that many potential key events can occur at
 2    levels of exposure that would be encountered in populations exposed to high levels of inorganic
 3    arsenic in drinking water.  It seems possible that those key events could fit together in many
 4    ways to result in a MOA for carcinogenesis. For example, some known mutagen and/or
 5    carcinogen commonly encountered in the environment might cause the initiation step, and then
 6    various arsenic-induced key events would provide the later steps necessary to result in a cancer.
 7    Alternatively, oxidative damage to DNA (or other types of DNA damage caused by arsenic)
 8    would make the DNA more prone to be acted upon by some other agent to produce a mutation
 9    that fulfills the initiation step. Although arsenic exposure does not induce gene mutations,
10    evidence from all three tables in Appendix C shows that chromosomal aberrations  can be
11    induced, and if a chromosome happened to break, for example in a tumor suppressor gene, that
12    mutation might provide an important step in a MOA. After the steps in a MOA resulted in cell
13    proliferation and genomic instability, cancer would result when changes occurred that provided
14    evasion of apoptosis, self-sufficiency of growth signals and insensitivity to anti-growth signals,
15    and limitless replicative potential (Hanahan and Weinberg, 2000). Vascularization would also
16    be needed to help the tumors grow larger.
17           Many detailed reviews in the past decade have discussed possible MO As for arsenic
18    carcinogenesis. Numerous ideas expressed in these reviews agree that exposure to inorganic
19    arsenic may be able to cause cancer by many alternative MOAs.  For example, Kitchin (2001)
20    discussed nine possible MOAs for arsenic carcinogenesis, suggesting that the three with the most
21    positive evidence in both animals and human cells are chromosomal abnormalities, oxidative
22    stress, and a continuum of altered growth factors leading to increased cell proliferation and then
23    the promotion of carcinogenesis. Florea et al. (2005) suggested that genomic damage, apoptosis,
24    and changes in gene expression associated with arsenic exposure are related to arsenic-induced
25    intracellular calcium disruption. Rossman (2003), Huang et al. (2004), and Simeonova and
26    Luster (2000) also provided noteworthy reviews related to MOAs of arsenic carcinogenesis.
27    Snow et al. (2005) reviewed effects  of arsenic at low concentrations and suggested that hormesis
28    (i.e., a biphasic response)  occurs in regard to cell proliferation and/or viability, base excision
29    DNA repair, and telomerase activity. While some low-dose effects (e.g., increased DNA repair)
30    may be protective of carcinogenesis, other effects (e.g., cell proliferation or telomerase
31    activation) may be protective and thus permit mutant cells to survive by preventing cellular
32    senescence and death and  may thereby be involved in arsenic's cancer-promoting capacity.
33           Kitchin and Ahmad (2003) provided an in-depth review on oxidative stress. They did not
34    reach a definitive conclusion on the role of oxidative stress in arsenic carcinogenesis, but rather
35    stated,
36
                                                100       DRAFT—DO NOT CITE OR QUOTE

-------
 1       •  " .. .it may eventually be found that many arsenic species act through several modes of
 2          carcinogenic action at many stages of multistage carcinogenesis and that the concept of a
 3          single cause of arsenic carcinogenesis simply does not fit the existing facts."
 4
 5          Oxidative stress seems particularly attractive as an important early step for some of the
 6    following reasons.  Some ROS can interconvert between themselves or react with nitric oxide
 7    (NO) to become reactive nitrogen species (RNS). RNS have their own spectra of biological
 8    reactivity.  High-energy ROS can convert to lower-energy forms and in the process can damage
 9    biological molecules.  ROS and related species can be inactivated by cellular defenses.
10    Extended, high-level exposure to reactive arsenic species might result in the depletion of
11    generalized cellular defense mechanisms against oxidative damage. ROS have been postulated
12    to be involved in both the initiation and promotional stages of carcinogenesis (Zhong et al.,
13    1997; Bolton et al., 1998, 2000; Shackelford et al., 2000; Chen et al., 2000b). Low levels of
14    ROS can modulate gene expression by acting as  a secondary messenger, while high doses of
15    ROS can cause oxidative injury leading to cell death (Perkins et al., 2000).  It has also been
16    demonstrated or suggested that ROS can (or does) damage cells by the following mechanisms:
17    lipid peroxidation; DNA and protein-modification; structural alterations in DNA including base-
18    pair mutations, rearrangements, deletions, insertions, and sequence amplifications (but not point
19    mutations); involvement in the signaling of the cell transformation response; affecting
20    cytoplasmic and nuclear signal transduction pathways that regulate gene expression; and
21    increasing the expression of certain genes (e.g., MDM2 protein, a key regulator of the tumor
22    suppression gene p53) (Li et al., 1998; Sen and Parker, 1996; Lander, 1997). Activation of
23    signal transduction pathways that enhance cell proliferation, reduce antiproliferative signaling,
24    and override checkpoints controlling cell division after genotoxic insult also have been
25    considered as possible mechanisms of arsenic's co-carcinogenic properties (Rossman, 2003).
26    Luster and Simeonova (2004) cited the results of in vitro studies suggesting that arsenic
27    stimulates cell proliferation through specific signal transduction pathways that are similar to
28    other classic tumor promotors. There has been much research in the last few years on the
29    effectiveness of As111,  especially ATO, on apoptosis, with much of it aimed at improving cancer
30    therapy.  Those results reveal the extreme complexity of the signal transduction cascades
31    involved in controlling apoptosis. Regarding causation of cancer, any  effects that inorganic
32    arsenic ingestion might have on signal transduction pathways that inhibit apoptosis could result
33    in proliferation of damaged cells and thereby lead to cancer.
34           The few animal studies (Waalkes et al., 2006a, 2006b, 2004a, 2004b, 2003a, 2003b) that
35    suggest inorganic arsenic is a complete carcinogen are those of Waalkes and his group that
36    involved treatments in utero.  Doses received by the pregnant dams were large compared to
37    human exposures, but tissue levels in the fetuses were reported as being comparable to levels
38    sometimes seen in humans. Almost all  of the categories of key events  discussed in this

                                                101       DRAFT—DO NOT CITE OR QUOTE

-------
 1    document can be caused by inorganic arsenic at exposure levels comparable to, or lower than,
 2    those that would be present in large population groups presently. The experiments also indicate
 3    that typically when a treatment is extended over a longer period of time, the concentration of
 4    inorganic arsenic necessary to cause an effect decreases.  This indicates that the impact in
 5    humans suggested by the in vitro findings might be substantially greater than might be expected
 6    by just comparing the concentrations found in humans and in those used in experiments. Due to
 7    the complexities of the possible MO As of inorganic-arsenic-mediated carcinogenesis, various
 8    scientific tools (e.g., genomic tools, human pharmacokinetic and biologically based dose-
 9    response models) may be needed in order to interpret the data for the hypothesized key events
10    qualitatively and quantitatively in a meaningful way.
      4.6.3.2. Low-Dose Extrapolation
11           According to the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), a
12    linear extrapolation to low doses is to be used either when there are MOA data to indicate that
13    the dose-response curve is expected to have a linear component below the point of departure
14    (e.g., DNA-reactivity or direct mutagenic activity) or when the available data are insufficient to
15    establish the MOA for a tumor site. Since the MOA of inorganic arsenic is unknown, a linear
16    low-dose extrapolation was applied as a default option.

      4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
17           Several studies (Yu et al., 2000; Chen et al., 2005a; Steinmaus et al., 2005; Valenzuela et
18    al., 2005; Ahsan et al., 2007; Huang et al., 2007b; McCarthy et al., 2007a) have observed a
19    correlation between increased disease risk and low urinary DMA and/or high urinary MMA,
20    indicating a slower secondary methylation.  Valenzuela et al.  (2005) measured the levels of
21    MMA111 in the urine of the residents of the Zimapan region of central Mexico. They found that
22    individuals exposed chronically to arsenic who also had arsenic-related skin lesions had
23    significantly greater concentrations and proportions of MMA111 in their urine than exposed
24    individuals without skin lesions. These findings support the hypothesis that any factor (e.g.,
25    genetic variability in metabolic enzymes) associated with reduced secondary methylation (i.e.,
26    the conversion of MMA to DMA) may also be correlated with increase susceptibility to arsenic-
27    induced disease. In the  following sections, factors affecting DMA and/or MMA ratios and level
28    in the urine or secondary methylation will be evaluated with regard to how they may affect
29    individual susceptibility.
      4.7.1. Possible Childhood Susceptibility
30           Although children are exposed to arsenic through generally the same sources as adults
31    (i.e., air, water, food, and soil), their behaviors  and physiology may result in them receiving
32    higher absorbed doses in relation to their body weight than adults for a given set of exposure

                                                102       DRAFT—DO NOT CITE OR QUOTE

-------
 1    conditions. Because children tend to eat less varied foods than adults, exposure to contaminated
 2    food, juice, or infant formula prepared with contaminated water may result in higher doses than
 3    adults.  In addition, children are more likely to ingest arsenic-contaminated soil, either
 4    intentionally or by putting dirty hands in their mouths.
 5           There are few data on the relative efficiency of absorption of arsenic from the
 6    gastrointestinal tract of children compared to adults, but measurement of urinary arsenic levels in
 7    children indicate that absorption does occur. ATSDR (2007) suggests that there is some
 8    evidence that children may be less efficient at methylating arsenic.  A decreased methylation
 9    capacity could lead to different tissue distribution and longer retention times that might possibly
10    increase their susceptibility relative to adults. Adults have been demonstrated to excrete 40% to
11    60% of the arsenic as DMA, 20% to 25% as inorganic As, and 15% to 25% as MMA.  Concha et
12    al. (1998b), however, determined that children ingesting 200 ppb (ug/L) arsenic in their drinking
13    water excreted about 49% as inorganic arsenic and 47% as DMA.  Women in the same study
14    were found to excrete 66% of the arsenic as DMA and 32% as inorganic arsenic.  In contrast,
15    others (Chowdhury et al., 2003; Meza et al., 2005, 2007; Sun et al., 2007) have found that
16    children have a higher urinary DMA:MMA ratio than adults, suggesting increased capacity for
17    secondary methylation.  Lindberg et al. (2008) also concluded that children and adolescents (i.e.,
18    <20 years of age) are more efficient methylators than adults (i.e., >20 years of age).  Studying a
19    population in Bangladesh exposed to high levels of arsenic in drinking water, Sun et al. (2007)
20    found increased secondary methylation indices (SMI) in children exposed to 90 or 160 ppb of
21    arsenic in drinking water, but not in controls. Chowdhury et al. (2003) also found that the
22    increased methylation in children was only observed in exposed individuals (average
23    concentration in drinking water 382 ppb) and not in the controls (<3 ppb in drinking water).
24    This could indicate a lower saturation point for secondary methylation in adults than in children.
25    Primary  methylation indices (PMI) were not age-dependent in any case.
26           Epidemiological studies provide only limited data on whether childhood exposures to
27    arsenic may result in increased cancer risk later in life.  Because a significant dose-response
28    relationship has been found between cancer mortality and increased years of exposure to the
29    high-arsenic artesian well water of southwestern Taiwan (Chen et al., 1986), it is important to
30    consider the extent to which childhood exposures contributed to lifetime arsenic intake.  The
31    analysis of cancer risks in the same population (Chen et al.,  1992) included "only residents who
32    had lived in the study area after birth," and assumed that the arsenic intake of each person
33    continued from birth to the end of the follow-up period  (1973  to 1986)3.  No information was
34    provided on the exposure of pregnant women in this population to the artesian well water.
       The artesian wells were introduced in 1910 to 1920; prior sources of fresh water included ponds, streams, and
      rainwater (Tseng, 1968).

                                                 103       DRAFT—DO NOT CITE OR QUOTE

-------
 1    Arsenic has been found to pass through the placenta (Hanion and Perm, 1977; Lindgren et al.,
 2    1984; Hood et al., 1987; Concha et al., 1998a; Jin et al., 2006a).
 3          Chen et al. (1992) stated that their cancer study results may somewhat underestimate
 4    arsenic-related risks in this population because tap water with lower arsenic concentrations was
 5    introduced into the study area in  1956 and was available to almost 75% of the residents in the
 6    1970s. Thus, the actual lifetime arsenic ingestion may be lower than estimated as residents
 7    switched from the high-arsenic artesian wells to alternate water sources.  Also, because this
 8    study is based on mortality records (1973 to 1986) from the study region, it would not capture
 9    cancer incidence among individuals exposed during childhood and early adulthood who then
10    migrated from the region.  Chen et al. (1986) reported that the 1982 migration rate for this area
11    was 27%, with primarily the youths and young adults leaving the area to move to cities and  those
12    45+ years old emigrating at a rate less than 6%. There is limited migration into this region,  and
13    it has been reported that more than 90% of the local residents lived in the study area all their
14    lives (Wuetal., 1989).
15          There is very little  epidemiology information specifically linking in utero arsenic
16    exposure to cancer outcome. Although the available epidemiological studies conducted in
17    Taiwan and other countries included women of reproductive age, the cancer outcomes from adult
18    exposures were not differentiated from in utero exposures. Recently, Smith et al. (2006),
19    examined lung cancer rates (and other respiratory diseases) in cohorts born just before the peak
20    exposure period in Antofagasta, Chile (meaning that they were not exposed in utero to high
21    levels of arsenic, but were exposed during childhood) and cohorts born during the high-exposure
22    period (indicating likely in utero exposure). Results demonstrated that exposure during either
23    period of development caused increased risk of lung cancer; however, the results from early
24    childhood exposures and/or in utero exposures were not compared to exposures during
25    adulthood to determine the possible cancer sensitivity effects in humans.
26          Although there is no adult animal model available for arsenic carcinogenesis,
27    administering inorganic arsenic to mice for 10 days during gestation has been found to increase
28    the incidence of lung, liver, reproductive, and adrenal tumors (Waalkes et al., 2003, 2004a,
29    2006a). This demonstrates that, at least in animals, embryos are more sensitive to the
30    carcinogenic effects of arsenic.
31          The Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
32    Carcinogens (U.S. EPA, 2005b) indicates that age-dependent adjustment factors  should be
33    applied to the CSF and combined with early-life exposure estimates when estimating cancer
34    risks from exposures to carcinogens with a mutagenic MOA. A mutagenic MOA for inorganic
35    arsenic has not been determined; therefore, the application of age-dependent adjustment factors
36    is not recommended.
                                                104       DRAFT—DO NOT CITE OR QUOTE

-------
      4.7.2. Possible Gender Differences
 1          Differences in methylation patterns have been noted between men and women in a
 2    number of studies. Higher MMA:DMA ratios have been observed in men than in women in a
 3    variety of populations tested, including in the United States (Hopenhayn-Rich et al., 1996b;
 4    Steinmaus et al., 2005, 2006, 2007), Taiwan (Tseng et al., 2005), and Bangladesh (Ahsan et al.,
 5    2007). In contrast, Loffredo et al. (2003) found that gender differences in arsenic methylation
 6    varied across populations studied in Mexico, China, and Chile, sometimes by exposure level.
 7    Based on mean urinary metabolite levels, they found no difference in the MMA:DMA ratio
 8    between males and females in China in the group with the highest arsenic levels in their drinking
 9    water (i.e., 405 ppb).  Low-exposure Chinese males (i.e., those exposed to 18 ppb in drinking
10    water) had MMA:DMA ratios similar to both the high-dose males and females (0.31 to 0.32), but
11    low-dose females had a much lower (i.e., 0.22) MMA:DMA ratio. In Mexico, there was a
12    difference between the sexes at high concentrations (408 ppb in the drinking water) of arsenic
13    (i.e., the MMA:DMA ratio was 0.23 in males vs. 0.18 in females), but there was no differences
14    in the MMA:DMA ratio (0.11) at low concentrations (i.e., 30 ppb in the drinking water). In
15    Chile, a completely different pattern was observed, with females  exposed to high concentrations
16    (600 ppb in the drinking water) demonstrating a higher MMA:DMA ratio (0.27) than males
17    (0.20), while the opposite pattern was seen at low concentrations (30 ppb in the drinking water;
18    0.18 in males vs. 0.13 in females).  Studying a population in Bangladesh  exposed to high levels
19    of arsenic in drinking water, Heck et al. (2007) found a higher percentage of urinary MMA in
20    men and a higher proportion of urinary DMA in women.
21          Age and reproductive status also may affect the male-female differences in arsenic
22    methylation patterns.  Concha et al. (1998a) demonstrated that pregnant women in their third
23    trimester excrete approximately 90% of arsenic as DMA. Engstrom et al. (2007) also found
24    pregnant women to have an increased proportion of DMA in their urine compared to non-
25    pregnant women in the same population, with increases occurring with gestational age.  This
26    indicates possible hormonal effects on arsenic methylation.  Lindberg et al. (2007) also found
27    possible hormonal effect on arsenic methylation, noting that females younger than 60 (i.e., likely
28    pre-menopausal) generally had a more efficient methylation than men of the same age, while the
29    difference narrowed considerably in males and females over 60.  Lindberg et al. (2008) found
30    that although females of all ages generally were better at methylating arsenic than males, the
31    greatest disparity between the sexes occurred between the ages of 20 and 55  (childbearing age in
32    women). Lindberg et al. (2007) also found that selenium, BMI, and AS3MT polymorphism
33    affected the observed proportions of methylated urinary arsenic metabolites in males only.  The
34    pattern of arsenic methylation was also altered in males with mutations in one allele of the
35    methylenetetrahydrofolate reductase (MTHFR) gene, but in females variants in both alleles were
36    required.

                                                105      DRAFT—DO NOT CITE OR QUOTE

-------
 1          Brenton et al. (2006) used a case-control study with 900 case-control pairs to examine the
 2    effect of hemoglobin levels on skin lesion prevalence in Pabna, Bangladesh. A 1.0 g/dL increase
 3    in hemoglobin was found to be associated with a 21% decrease in the odds for having skin
 4    lesions even after adjusting for toenail arsenic levels, BMI, education, biri or cigarette smoking,
 5    chewing tobacco, and betel nut chewing. However, when the data was examined further, it was
 6    discovered that the hemoglobin levels were correlated with decreased skin lesion prevalence
 7    only in males  (40% reduction), but not in females. Females, however, were more likely to have
 8    anemia than males (18.2% vs.  8.2%; p < 0.0001). A subsequent cohort study (Brenton  et al.,
 9    2006) found that hemoglobin levels were not associated with changes in urinary arsenic levels or
10    MMA/DMA ratios.
      4.7.3. Other
      4.7.3.1. Genetic Polymorphism
11          Despite the observed differences in methylation related to age and sex, data from
12    Bangladesh analyzed by Lindberg et al. (2008) suggest that genetic polymorphism is the most
13    important factor affecting the methylation of inorganic arsenic, with only 30% of variation in
14    methylation patterns attributable to level of arsenic exposure, gender, and age. Most humans
15    excrete 10% to 30% of absorbed inorganic arsenic as unchanged in urine, 10% to 20% as MMA,
16    and 60% to 80% as DMA.  Excretion patterns vary across populations, however. A study of
17    urinary arsenic in a population in northern Argentina exposed to arsenic via drinking water
18    demonstrated  an average of only 2% MMA in the urine (Vahter et al., 1995b; Concha et al.,
19    1998b).  Studies on populations in San Pedro and Toconao in northern Chile demonstrated
20    differences  in the ratio of MMA:DMA excretion between the two populations (Hopenhayn-Rich
21    et al., 1996b). Chiou et al. (1997) found that in a population in northeastern Taiwan, 27% of the
22    arsenic consumed was excreted as MMA. Although these variations have not been
23    unequivocally linked with genetic factors, as opposed to environmental or nutritional factors,
24    human genetic polymorphism has been reported for methyltransferases believed to be involved
25    in arsenic metabolism (e.g., thiopurine S-methyltransferase; Yates et al.,  1997).
26          Chung et al. (2002) studied the association of familial relationships with urinary arsenic
27    methylation patterns in 11 families (father, mother, and two children studied from each family)
28    from Chile where drinking water concentrations were 735-762 ppb. Their results indicate that
29    13-52% of the variation in methylation patterns could be explained by being a member of a
30    specific family. There was a high and significant correlation in the methylation patterns between
31    siblings and a much lower correlation between parent and child, which could be attributed to
32    inherent differences in methylation patterns between children and adults. Adjusting for
33    nutritional factors (blood levels of methionine, homocysteine, folate, vitamin Be, selenium, and
34    vitamin 812) did not notably alter the correlation. As might be expected, the correlation between

                                                106       DRAFT—DO NOT CITE OR QUOTE

-------
 1    father and mother was relatively low, even when adjusted for age and gender.  However, the
 2    correlation became stronger when adjusting for homocysteine levels as well.
 3          Meza et al. (2005) found a strong association between the variations in the DNA
 4    sequence of AS3MT and urinary DMA:MMA ratios in native populations in Yaqui Valley in
 5    Sonora, Mexico.  Three polymorphic sites were found to be associated with increased
 6    DMA:MMA levels in the study population, but site 30585 was most strongly associated with
 7    urinary arsenic metabolite patterns.  Using a stepwise linear regression model with DMA:MMA
 8    as the dependent variable and 30585 genotype, age, sex, and log-converted daily arsenic dose as
 9    independent variables, only the 30585 genotype and age were found to have a highly significant
10    association with DMA:MMA levels. Further investigation determined that there was no
11    significant genetic association observed in adults, but there was a highly significant effect in
12    children aged 7 to 11 years.  There was no difference in the allele frequencies at the 23 sites
13    examined between the adults and children.
14          Engstrom et al. (2007)  also found a strong association between the presence of three
15    intronic single nucleotide polymorphisms in AS3MT (i.e., G12390C, C14215T, and A35991G)
16    and increased DMA levels. The study population consisted of adult women living in San
17    Antonio de los Cobres (a village in the northern Argentinean Andes) who were exposed to
18    approximately 200 ppb of arsenic in their drinking water.  This group provided a rather uniform
19    genetic background against which to examine the impact of polymorphism alone as a variant.
20    Subjects who were homozygous for one or more of the variant alleles had lower MMA and
21    higher DMA levels than heterozygotes, who in turn had lower MMA:DMA ratios than
22    individuals lacking the alleles. Because the proportion of ingested inorganic arsenic that was
23    excreted was relatively constant across the groups, the effects of the variants were attributed
24    primarily to increased secondary methylation. Individuals homogenous for all three variant
25    alleles were found to have the  lowest proportions of urinary MMA and the highest proportions of
26    DMA among all the groups studied.
27          A case-referent study in Bangladesh evaluated arsenic metabolite patterns in 594
28    individuals with arsenic-related skin lesions compared to 1,041 controls (Ahsan et al., 2007).  A
29    correlation was found between increased arsenic concentrations in the drinking water, increased
30    proportions of MMA in the urine, and the risk of skin lesions, suggesting that variations in
31    secondary methylation could increase the risk of developing such lesions. Individuals with
32    variants in MTHFR (677TT/1298AA and 677CT/1298AA diplotypes) also had slightly increased
33    skin lesion risk (OR 1.66 and 1.77, respectively). However, the risk for developing skin lesions
34    in relation to all at-risk alleles  for the GSTO1 diplotype was 3.91.  Additivity of effect was
35    observed when the genotypes were analyzed jointly with water arsenic concentrations and
36    proportion of urinary MMA.
                                                107       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Steinmaus et al. (2007) examined the association between genetic polymorphisms in
 2   MTHFR and GST and urinary arsenic metabolites in 170 subjects from Argentina. Subjects with
 3   the TT/AA variant of MTHFR 677/1298 were found to have higher urinary proportions of
 4   inorganic arsenic and MMA (not statistically significant) and lower levels of DMA, with the
 5   results being more pronounced in males. A null genotype of GSTM1 in women was
 6   significantly associated with lower proportions of urinary MMA and higher proportions  of
 7   urinary DMA compared to women with the active genotype. While the same trend was observed
 8   in males, it was weaker and did not achieve statistical significance. Polymorphism in the GSTT1
 9   gene was not associated with differences in arsenic methylation. Lindberg et al. (2007) also
10   found that carriers of the variant allele of the M287T (C—>T) polymorphism of the AS3MT gene
11   or the A222V (C—>T) polymorphism in the MTHFR gene had higher proportions  of urinary
12   MMA.
13          McCarthy et al. (2007a,b) examined the effect of GST polymorphisms on  skin lesion risk
14   in a case-control (600 pairs) study in Pabna, Bangladesh. In one study (2007a), they found that a
15   10-fold increase in MMA/inorganic arsenic ratio was associated with a 1.5-fold increase in risk
16   of skin lesions.  There was a significant interactive effect between GSTT1 wild-type and
17   secondary methylation on skin lesions, but no interactive effects with the GSTM1 or GSTP1
18   genotypes or any of the genotypes with primary methylation.  In their second study (2007b),
19   however, they found a greater risk for skin lesions in GSTT1 wild-type (OR=1.56, 95% CI 1.10-
20   2.19) compared to GSTT1 null status (referent group). The presence of the GSTP1 GG genotype
21   was associated with a 1.86-fold increase (95% CI: 1.15-3.00)  in risk of skin lesions over the AA
22   genotype. However, none of the polymorphisms examined (i.e., GSTT1, GSTM1, and GSTP1)
23   were found to modify the association between arsenic exposure and skin lesion risk.
24          Banerjee et al. (2007) also found a significant correlation between genetic polymorphism
25   and skin lesions in a population in West Bengal, India. This population was selected because
26   even though over 6 million people are exposed to high arsenic levels, only 15% to 20%
27   developed skin lesions. Polymorphisms in ERCC2, which is a NER pathway gene, was
28   examined.  Specifically, the relationship between the ERCC2 codon 751 A—>C polymorphism
29   (lysine to glutamine) and skin lesion risk.  Subjects exposed to arsenic-contaminated drinking
30   water with hyperkeratosis (n = 165) were compared to those without skin lesions (n = 153).
31   Occurrence  of hyperkeratosis was strongly associated with the Lys/Lys genotype  in the ERCC2
32   codon 751, with an OR of 4.77 (95% CI: 2.75-8.23). A significant increase in chromosomal
33   aberrations in individuals with the AA genotype compared to either the AC or CC genotypes
34   combined was also observed.
35          Brenton et al. (2007a) observed a positive association between total urinary arsenic and
36   oxidative stress (as measured by 8-OHdG) in healthy women (only females were studied) from
37   Pabna, Bangladesh, with the GSTM1 null genotype.  No such  association was found in GSTM1

                                               108       DRAFT—DO NOT CITE OR QUOTE

-------
 1    positive women. APE1 (apurinic/apyrimidinic endonuclease) was found to be a predictor of 8-
 2    OHdG levels with the variant allele associated with a decrease in 8-OHdG. Other factors that
 3    also were predictive of 8-OHdG levels included creatinine, betel nut chewing, presence of
 4    environmental tobacco smoke in the home (even though none of the women reportedly smoked
 5    themselves), and education.
 6          In a case-control study with 792 pairs with and without skin lesions in Pabna,
 7    Bangladesh, Brenton et al. (2007b) studied the association between genetic polymorphisms in
 8    the base excision DNA repair pathway and arsenic-induced skin lesions. Four common base
 9    excision repair (BER) genetic polymorphisms (X-ray repair cross-complimentary group 1
10    [XRCC1] Arg399Gln, XRCC1 Argl94Trp, human 8-oxoguanine DNA glycosylase [hOGGl]
11    Ser326Cys, and APE1 Aspl48Glu) were examined.  APE1 148 Glu/Glu individuals were twice
12    as likely to have skin lesions as APE1 148 Asp/Asp individuals, even after adjusting for toenail
13    arsenic concentration, BMI, education, smoking, and betel nut use. Presence of the Glu/Glu
14    variant of APE1 Aspl48 Glu was associated with a 2- to 2.5-fold increased OR for skin lesions
15    compared to the Asp/Asp variant, in the low and middle tertiles, but no increase was observed in
16    risk at the highest tertile of exposure.  XRCC1 Argl94 Trp genotypes, however, were not
17    associated with skin lesion risk in the low and middle tertiles, but were associated with a 3-fold
18    difference in the highest exposure tertile (i.e., OR of 2.9 for Trp/Trp compared to 8.4 for
19    Arg/Arg where Arg/Arg at the lowest tertile is the referent group).  No association was observed
20    between skin lesions and genetic polymorphisms in XRCC1 Arg399Gln or hOGGl Ser326Cys
21    alleles.
      4.7.3.2. Nutritional Status
22          In many of the epidemiological studies discussed above (e.g., southwestern Taiwan and
23    Bangladesh), the study subjects were relatively poor and had poor nutritional status. Mazumder
24    et al. (1998) demonstrated that people in and around West Bengal who had body weights below
25    80% for their age and sex had an increased RR (2.1 for females and 1.5 for males) in the
26    prevalence of arsenic-associated keratosis. Lindberg et al. (2008), however, found that women
27    in Bangladesh were better at methylating arsenic than men even though they were less likely to
28    eat nutritious food (e.g.,  meat and fresh vegetables) than men, indicating that gender was a better
29    predictor  of methylation capacity than nutritional status in this group.
30          Selenium has been demonstrated to reduce the teratogenic, clastogenic, and cytogenic
31    effects of arsenic (ATSDR, 1993).  Chen et al. (2007) found that individuals in the Health
32    Effects of Arsenic Longitudinal Study (HEALS; population from Araihazar, Bangladesh) with
33    low selenium intake were at a greater risk for developing pre-malignant skin lesions than those
34    with adequate intake.  In 93 pregnant women from Antofagasta, Christian et al. (2006) found that
35    increases  in urinary selenium levels were associated  with increased urinary arsenic excretion,
36    and with a greater percent excreted as DMA and less excreted as inorganic arsenic. The
                                               109       DRAFT—DO NOT CITE OR QUOTE

-------
 1    proportion of urinary MMA was fairly consistent in the study population. Using four quartiles of
 2    increasing urinary selenium levels, results showed that the total arsenic excretion increased
 3    steadily across quartiles of selenium intake.  The proportion of DMA excreted increased, and the
 4    proportion of inorganic arsenic excreted decreased with increasing selenium intake, but only in
 5    the first two quartiles. Although different gestational stages of pregnancy have been associated
 6    with differences in urinary arsenic excretion patterns, this was controlled for in the analysis.
 7           Gamble et al. (2005) suggest that adequate folate is necessary for both primary and
 8    secondary arsenic methylation and that adequate folate intake is associated with increased
 9    urinary DMA.  Gamble et al. (2006) found that providing folate supplements to individuals from
10    Araihazar, Bangladesh, with a diet low in folate significantly increased the proportion of arsenic
11    excreted as DMA in the urine.  Heck et al. (2007), however, found that levels of folate
12    consumption (measured by levels in the food) were directly related to percentages of urinary
13    MMA, but not to changes in urinary DMA in a population from Bangladesh (participants of the
14    HEALS study) exposed to arsenic in drinking water. Heck et al. found no correlation between
15    intake of folate-related nutrients and urinary DMA levels, but found that increases in methionine,
16    vitamin B12, calcium, protein, and riboflavin were associated with decreases in the proportion of
17    urinary inorganic arsenic and increases in the percent of urinary MMA. Niacin and choline were
18    found to be the better predictors of secondary methylation (as measured by DMA/MMA).
19    Although high levels of plasma homocysteine were not associated with urinary MMA levels,
20    they were associated with a decrease in DMA levels (Gamble et al., 2005).
21           Mitra et al. (2004) studied whether nutritional deficiencies increased the susceptibility of
22    individuals to arsenic-related health effects as measured by skin lesions.  In West Bengal, India,
23    where exposures were <500 ppb, nutritional assessments were based on a 24-hour recall for
24    major dietary constituents and a 1-week recall  for less common constituents. Increases in risk
25    were associated with low intake of animal protein (OR=1.94, 95% CI: 1.05-3.59), calcium
26    (OR=1.89, 95% CI:  1.04-3.43), fiber (OR=2.20, 95% CI: 1.15-4.21), and folate (OR=1.67, 95%
27    CI: 0.87-3.2).  Nutrient intake was not related  to arsenic exposure. The authors concluded that
28    the potential protective effects of these nutrients were small in comparison to eliminating the
29    exposure to arsenic.
30           Steinmaus et al. (2005) found an association between low dietary protein, iron, zinc, and
31    niacin, and decreased production of urinary DMA accompanied by increased levels of urinary
32    MMA in arsenic-exposed individuals from a U.S. population. An associations between arsenic
33    methylation patterns and dietary intake of thiamine,  vitamin B6, lutein, and a-carotene were
34    found, but the links were not as clear when adjusted for confounding variables (i.e., age, sex,
35    smoking, and total urinary arsenic levels). The authors suggest, however, that the effect of
36    specific nutrient intake levels on methylation patterns was small in comparison with the known
                                                 110       DRAFT—DO NOT CITE OR QUOTE

-------
 1   magnitude of inter-individual variability associated with genetic polymorphisms. Kreppel et al.
 2   (1994) found that dietary zinc protects mice against acute arsenic toxicity.
     4.7.3.3. Cigarette Smokers
 3          Cigarette smokers (current or former) were found to have a decreased secondary
 4   methylation capacity, resulting in increased urinary MMA and decreased DMA concentrations
 5   (Huang et al., 2007b). Tseng et al. (2005) reported a decrease in secondary metabolism in
 6   cigarette smokers exposed to arsenic-contaminated drinking water, resulting in a significant
 7   increase in the secreted MMA as a fraction of total metabolites. Steinmaus et al. (2005) found
 8   that current smokers in a U. S. population had lower proportion of arsenic excreted as DMA than
 9   either former or never-smokers (although the difference was not statistically significant).
10   Steinmaus et al. (2006) found that in a population in Argentina the  proportion of excreted MMA
11   was associated with bladder cancer risk in former smokers, but not in individuals who had never
12   smoked. Subjects who had ever smoked and had proportions  of MMA in the upper tertile had a
13   2-fold elevated risk of bladder cancer compared to subjects with proportions of MMA in the
14   lower two tertiles. Therefore, it was concluded that individuals who smoke had  an increased
15   susceptibility to arsenic toxicity. Steinmaus et al. (2006) also studied a population in the United
16   States.  Although the results indicated increased MMA was associated with increased cancer
17   risk, the number of cases was too small to estimate separate ORs for never-smokers and ever-
18   smokers.
19
                                                111       DRAFT—DO NOT CITE OR QUOTE

-------
                              5. DOSE-RESPONSE ASSESSMENTS

      5.1. ORAL REFERENCE DOSE (RfD)
 1                 An RfD was developed for inorganic arsenic and posted on the IRIS database in
 2    1991.  An oral noncancer dose-response estimation is not addressed in this document.  However,
 3    the Agency is currently reviewing the literature and will develop an updated RfD at a later date.

      5.2. INHALATION REFERENCE CONCENTRATION (RfC)
 4          An inhalation noncancer dose-response estimation is not addressed in this document. An
 5    RfC is not developed for inorganic arsenic, nor does a current value exist on the IRIS database.

      5.3. CANCER ASSESSMENT (ORAL EXPOSURE)

      5.3.1. Background: History of Cancer Risk Assessments for Arsenic
 6          This assessment is unusual in that it builds on a long history of previous efforts by EPA
 7    and others to evaluate potential risks from oral exposure to arsenic via drinking water. Table 5-1
 8    summarizes previous assessments and expert reviews of arsenic carcinogenicity.
 9           The table starts (chronologically) with EPA's  1988 risk assessment for skin cancer (U.S.
10    EPA, 1988b).  The scope of the 1988 assessment was to review the applicability of EPA's 1984
11    assessment (U.S. EPA, 1984) on skin cancer risk from the Taiwanese population to the U.S.
12    population. The skin cancer risk from oral exposure was estimated based on two studies (Tseng
13    et al., 1968; Tseng, 1977) of age-specific prevalence rates for skin cancer in a large cohort of
14    Taiwanese (40,241 subjects in 37 villages) in an "arseniasis-endemic" area, where arsenic
15    concentrations in water supply wells ranged from less than 10 ug/L (ppb) to 1,820 ug/L. The
16    occurrence of skin cancer was estimated in a survey lasting approximately 2 years (U.S. EPA,
17    1988b). Preliminary data from the same cohort suggested that risks of internal cancers (lung,
18    liver, and bladder) were also elevated, but U.S. EPA (1988b) concluded that insufficient data
19    were available to support a dose-response assessment for these effects.
20          The second entry in the table is the National Research Council's 1999 review (NRC,
21    1999) of EPA's 1988 risk assessment. EPA commissioned NRC to review the U.S. EPA (1988b)
22    assessment and also the qualitative and quantitative evidence on arsenic and health effects for
23    reassessment of human health risks from arsenic in drinking water.  One of the major
24    recommendations of NRC's 1999 review was that studies from the arsenic-endemic area of
25    Taiwan (Wu et al.,  1989; Chen et al., 1988a, 1992) provide the best available empirical human
26    data for assessing the risks of arsenic-induced cancer.  The report explored quantitative modeling
27    approaches for the male bladder cancer data, but did not provide a formal risk assessment;
                                               112      DRAFT—DO NOT CITE OR QUOTE

-------
 1    additional modeling analyses were recommended. NRC 1999 applied absolute Taiwan risks to
 2    the U.S. populations.
 3          NRC (1999) published the arsenic concentration in village wells, person-years of males
 4    and females by village and the village-specific lung, bladder, and liver deaths for the Wu et al.
 5    (1989) and Chen et al. (1992) studies. Additional raw data were obtained from study authors by
 6    Morales and Ryan during reanalysis and these data were subsequently provided to EPA
 7    (personal communications). All of the succeeding assessments summarized in Table 5-1 derive
 8    dose-response estimates based on the internal cancer data.
 9          In the first of these efforts, Morales et al. (2000) gathered data on lung, bladder, and liver
10    cancer, as well as detailed exposure data (well arsenic concentrations) from the three
11    epidemiological studies (Wu 1989; Chen et al., 1988a, 1992), and evaluated a range of statistical
12    models for estimating potential arsenic-related cancer risks in the Taiwanese population and for
13    extrapolating these risks to the U.S. population. In promulgating the Primary Drinking Water
14    Standard for Arsenic, U.S. EPA (2001) adopted one of Morales et al.'s models, with adjustments
15    of some exposure assumptions, for estimating the health benefits of regulatory alternatives. The
16    Office of Pesticide Programs (OPP) also recently applied oral CSFs based on the U.S. EPA
17    (2001) assessment in their Reregi strati on Eligibility Decision (RED) Documents for organic
18    arsenic pesticides (U.S. EPA, 2006c) and for Inorganic Arsenicals and/or Chromium Based
19    Wood Preservatives (U.S. EPA, 2008).
20          In response to continued public concern over arsenic-related cancer risks, EPA asked
21    NRC to update its 1999 recommendations in light of new scientific evidence, and to review the
22    risk assessment in support of the 2001 drinking water standard. NRC (2001) reviewed the
23    methodology used in EPA's arsenic risk assessment (U.S. EPA, 2001) and provided a systematic
24    analysis of and recommendations for applying the Taiwanese epidemiological data for assessing
25    cancer risks from arsenic exposure in U.S. populations.  Recommendations included the
26    inclusion of a reference population in the dose-response assessment, the form of the dose-
27    response model, exposure assumptions, and approaches for extrapolating risks to the U.S.
28    population.  As the committee noted, the cancer risk estimates that it developed were higher than
29    those reported by U.S. EPA (2001), and reasons for those differences were reviewed.  EPA
30    examined and applied the NRC (2001) statistical methodology and submitted its revised analysis
31    (U.S. EPA, 2005c) to SAB for review and comment.  SAB (2007) provided additional discussion
32    related to the treatment of arsenic exposure, and recommended expanded sensitivity analyses of
33    other exposure-related assumptions. EPA adopted these recommendations, along with responses
34    to comments from interagency reviewers, into the current assessment.  The current quantitative
35    risk assessment can thus be described as EPA's reimplementation of the technical cancer risk
36    modeling recommendations in NRC (2001), with additional  examination of arsenic exposure
                                                113       DRAFT—DO NOT CITE OR QUOTE

-------
1   assumptions and taking into account SAB's (2007) advice for the expansion of sensitivity
2   analyses of modeling methods and choices.
    Table 5-1. Historical Summary of Arsenic Risk Assessment Efforts
Assumption/
Method
Goals/Scope of
Assessment







Critical Study






Critical Study
Endpoint(s)

Dose-Response
Model






Reference
Population


Arsenic
Concentration





U.S. EPA
(1988b)
Revise EPA's
1984 risk
assessment
for skin
cancer,
evaluate
evidence of
arsenic
essentiality
Taiwan skin
cancer
prevalence
studies
(Tseng etal.,
1968; Tseng,
1977)
Skin cancer
incidence

Linear
multistage






Taiwanese
outside
arseniasis-
endemic area
Stratified: 0-
300, 300-
600, 600-900
ug/L in well
water,
unknown
exposure
NRC (1999)
Review EPA's
1988b risk
assessment,
suggest
alternative
approaches;
was "not a
risk
assessment"
Taiwan
epidemiologic
al studies (Wu
etal., 1989;
Chen et al.,
1988a, 1992)

Bladder
cancer
mortality
Weibull,
Poisson
regression





With and
without all-
Taiwan

Median well
arsenic
concentrations




Morales et
al. (2000)
Test dose-
response
models,
modeling
assumptions




U.S. EPA
(2001)
Estimate U.S.
cancer risks in
support of
drinking water
standard




NRC (2001)
Review
EPA's 2001
methods and
results





U.S. EPA
(2005c)
Incorporate
NRC (2001)
recommenda-
tions for SAB
Review




Taiwan epidemiological studies (Wu et al., 1989; Chen et al., 1988a,
1992)





Bladder, lung,
liver cancer
mortality
Nine Poisson
forms with
varying age,
dose
representation
s; one
multistage
Weibull
None,
southwest
Taiwan, all-
Taiwan
Median well
arsenic
concentration
s



Bladder, lung cancer mortality


Morales et al.
"Model 1"
(multiplicative
linear dose,
quadratic age)



None



Median well
arsenic
concentrations




Additive
Poisson,
linear dose,
quadratic age




All-Taiwan,
southwest
Taiwan

Median;
sensitivity
analysis of
other values



Additive
Poisson, linear
dose, quadratic
age; UCLs on
dose coefficients
estimated by
Bayesian
simulation
Southwest
Taiwan


Median well
arsenic
concentrations




                                             114
DRAFT—DO NOT CITE OR QUOTE

-------
Assumption/
Method
Taiwanese
Water Intake







Taiwanese
Body Weight
Nonwater
arsenic Intake











Risk Model
for U.S.
Population
U.S. Incidence,
Mortality Data
U.S. Water
Intake






U.S. Body
Weight
U.S. EPA
(1988b)
3.5 L/day
(M),
2.0 L/day (F)






55 kg (M), 50
kg(F)
None (0
ug/day)











Simple life
table

Not specified

2.0 L/day
(approximate
90th
percentile
value)



70 kg (M and
F)
NRC (1999)
3.5 L/day (M),
2.0 L/day (F)







55 kg (M), 50
kg(F)
Not explored












Simple life
table

NCHS 1994
mortality data
2.0 L/day
(approximate
90th percentile
value)




70 kg (M and
F)
Morales et
al. (2000)
Water intakes
not specified







Body weights
not specified
None (0
ug/day)











Life table, 5-
year age strata

U.S. EPA
(2001)
3. 5 L/day (M),
2.0 L/day (F) +
1.0 L/day
cooking





55 kg (M),
50 kg (F)
50 ug/day
(exposed
population)










Life table,
5 -year age
strata
NRC (2001)
Recommenda
tions based
on approx. 2
L/day;
sensitivity
analysis of
U.S./Taiwan
intake ratios
is presented
55 kg (M),
50 kg (F)
None (0
ug/day)
in baseline
assessment;
sensitivity
analysis
showed little
effect of
adding 30 or
50 ug/day to
study village
exposure
estimates
U.S. EPA
(2005c)
2.0 L/day








50 kg (M and F)

30 ug/day
exposed
population only,
sensitivity
analyses of
0-50 ug/day







BEIR IV survival model (relative
risk)

NCHS 1996 mortality

Average U.S.
water intake






Average U.S.
body weights
1.0-1. 2 L/day
used as central
tendency
values;
2. 1-2.3 L for
90th percentile
risk in Monte
Carlo model
1.0 L/day
with
sensitivity
analyses




1.0 L/day







70 kg (M and F)

115
DRAFT—DO NOT CITE OR QUOTE

-------
Assumption/
Method
Endpoints
Calculated



























U.S. EPA
(1988b)
Unit risk =
3xlQ-5per
Hg/L
(females),
7xlO'5per
ug/L (males);
CSFs = 1 to 2
per mg/kg-
day
(incidence)



















NRC (1999)
Lifetime
bladder cancer
risk at 10 ug/L
= 3xl(T3
(males),
9xlO"3
(females);
EDM = 404-
443 ug/L,
LEDoi =
323-407 ug/L


















Morales et
al. (2000)
"Model 1," no
reference pop.

Males (ug/L)
EDoi
LEDoi
Lung 364
294
Bladder 3 95
326

Females
(iig/L)
EDoi
LEDoi
Lung 258
213
Bladder 252
211
Many other
results
presented







U.S. EPA
(2001)
CSFs derived
from Morales
et al. (2000)
EDoi,
LED 01 values

Unit risk, per
ug/L:

Male bladder=
2.5xlO'5
(MLE),
3.1xlO'5(UCL)

Male lung =
2.8xlO"5
(MLE),
3.4xlO'5
(UCL)
Female bladder
= 4.0xlO"5
(MLE),
4.7xlO"5
(UCL)
Female lung=
3.9xlO'5,
(MLE),
4.7xlO"5
(UCL)
NRC (2001)
Lifetime
cancer risk
incidence
from
10 ug/L:

Male
lung =
l.SxlO"3
bladder =
2.3 xlO'3

Female
lung =
1.4xlO"3
bladder =
1.2xlO"3












U.S. EPA
(2005c)
Female lung +
bladder
incidence:
unit risk =
1.6xlO'4per
ug/L

Incidence at 10
ug/L in drinking
water = 1.6xlO"3

Drinking water
concentration
for 10'4
incidence risk =
0.63 ug/L













 1          The techniques and assumptions used in arsenic risk assessment have evolved and
 2   changed over time, and it is not possible to do justice to all of the changes and innovations in
 3   each assessment in this chapter.  Table 5-1 provides a general summary of the  important data
 4   sources, techniques, and assumptions employed in each assessment. Where cells in the table are
 5   merged across the columns, it indicates that the same assumptions were used in more than one
 6   assessment, implying a solidification of a technical consensus.  The major issues addressed in
 7   each study include:
 8
 9      •   Scope and goals. Some of the efforts in Table 5-1 (the NRC studies most importantly)
10          were not intended to be comprehensive risk assessment, but to provide recommendations
11          for EPA and other agencies. Some were pure modeling studies (Morales et al., 2000),
12          and some were employed to derive quantitative risk estimates for regulatory support
13          purposes (U.S. EPA, 2001) or for health criteria development (U.S. EPA, 2005c).
14
15      •   Selection of critical studies for use in the risk assessment. As noted above, the U.S.
16          EPA (1988b) assessment was based on skin cancer prevalence data (Tseng et al., 1968;
17          Tseng, 1977).  All of the subsequent assessments in the table use data from later
                                              116
DRAFT—DO NOT CITE OR QUOTE

-------
 1          epidemiological studies (Wu et al., 1989; Chen et al., 1988a, 1992), which provide
 2          information on PYR and cancer mortality in narrowly defined age strata, and exposure
 3          concentrations from individual water supply wells.
 4
 5       •  Critical study endpoints.  Over time, assessments have moved from evaluating skin
 6          cancer (U.S. EPA, 1984, 1988b) to internal cancers (lung and bladder).  As discussed
 7          below, the change in endpoint is the major reason that the cancer potency estimated in the
 8          current assessment is so different from that derived in 1988.  Wu et al. (1989) and Chen
 9          et al. (1988a, 1992) also reported data on liver cancer, but in response to concerns related
10          to a high incidence of viral hepatitis in Taiwan (U.S. EPA, 2001), liver cancer has not
11          been included as an endpoint in recent assessments.
12
13       •  Dose-response models.  The form of the dose-response models used to assess risks in the
14          Taiwanese population has evolved over time as different investigators explored the
15          performance of various models under a wide range of exposure assumptions.  In the early
16          models, linear regression and multistage models were used for dose-response  assessment
17          in the Taiwanese population. In the more recent analyses, Poisson regression with linear
18          dose terms and quadratic age terms have been employed, as recommended by NRC
19          (2001), to derive primary risk estimates.  In addition, sensitivity analyses of other Poisson
20          models (different transformations of dose) have been conducted, as recommended by
21          SAB (2007).  Changes in the modeling approaches, like changes in the endpoints
22          modeled, have resulted in changes in estimated cancer potency.
23
24       •  Inclusion/exclusion of a reference population. EPA's 2001 risk assessment was based
25          on a dose-response model for the Taiwanese population that did not include a reference
26          population (i.e., a group with similar characteristics not exposed to arsenic in  drinking
27          water). In keeping with NRC (2001) and SAB (2007) comments, the primary estimates
28          in this chapter are derived based on the inclusion of a reference population from
29          southwest Taiwan; sensitivity analyses are provided for risk estimates with the reference
30          population excluded and with a reference population from all regions of Taiwan (i.e.,
31          "all-Taiwan").
32
33       •  Arsenic concentration used in the dose-response model.  The available exposure data
34          (Wu et al., 1989; Chen et al., 1992) consist of measurements from 155 village drinking
35          water wells taken between  1964  and 1966 for 42 exposed villages. Most of the
36          assessments in Table 5-1 employed the median exposure concentrations for each group.
37          That approach also is followed in this assessment; however, following SAB (2007)
38          recommendations, a sensitivity analyses  on the impacts of using minimum and maximum
39          village arsenic concentrations in the risk assessment has been conducted.
40
41       •  Water intake and body weight of the exposed population. As discussed in Section
42          5.3.5, there are few precise data available concerning the distribution of daily drinking
43          water intake volumes in the exposed populations. As shown in Table 5-1, past
44          assessments have employed a range of assumptions; the basic consensus is that
45          Taiwanese men appear to consume more water than men in the U.S. owing to the hotter
46          climate, and because a large proportion of them engage in vigorous outdoor activity as
47          part of their livelihood. Consistent with the limited information, the current analysis has
                                               117       DRAFT—DO NOT CITE OR QUOTE

-------
 1          followed this consensus. Following other analyses, this assessment assumes an average
 2          body weight of 50 kg for both Taiwanese men and women.
 3
 4       •  Nonwater arsenic intake. Because the risk modeling for the Taiwanese population is
 5          based on estimated daily arsenic dosage, it is important to include reasonable
 6          assumptions about the amount of arsenic intake coming from non-drinking water sources.
 7           This is an area where there is relatively little data, and considerable confusion about, for
 8          example, whether and how to include a contribution from cooking water, reasonable
 9          estimates of arsenic concentrations in food, and whether the arsenic-exposed and
10          reference populations should be assumed to receive the same nonwater arsenic intake.
11          The various assumptions used in previous analyses are summarized in Table 5-1, and the
12          basis for nonwater arsenic intake  estimates used in this assessment is discussed in Section
13          5.3.5.  As is the case for many other assumptions, the approach to dealing with
14          uncertainty in nonwater arsenic intake is to conduct sensitivity analyses based on a
15          reasonable range of values.
16
17       •  Risk model for the U.S. population. The outputs of the dose-response modeling for the
18          Taiwanese population were arsenic dose-response coefficients that described the
19          relationship between estimated arsenic intake in the Taiwanese population and
20          proportional increases in age-specific lung and bladder cancer mortality risk. Consistent
21          with NRC (2001) recommendations, lifetime cancer incidence in U.S. populations was
22          then estimated by using a modified version of the "BEIRIV" relative risk model, as
23          described in Appendix E. A key assumption underlying this model is that the risk of
24          arsenic-related cancer mortality or incidence for the U.S. population is a constant
25          multiplicative function of the current "background" age profile of cancer risks in the
26          same U.S. population.
27
28       •  U.S. mortality and cancer incidence data.  Models for extrapolating cancer risks for the
29          U.S population require data on overall mortality, and the BEIR IV model requires non-
30          arsenic related cancer incidence data for the U.S. population. One source of variation in
31          the cancer risk estimates over time has been the use of more recent mortality and cancer
32          incidence data in the most recent  assessments.
33
34       •  U.S. water intake and body weight.  Estimates of the drinking water intake and typical
35          body weight of the exposed population are also needed to predict cancer risks in the U.S.
36          population. All of the recent assessments assume body weight of 70  kg for males and
37          females. For the primary risk estimates, the  current assessment assumes a water intake of
38          2.0 L/day, as  discussed in Section 5.3.5, with sensitivity analyses of other values. Adult
39          water intake of 2.0 L/day is used as a  standard factor in EPA IRIS assessments, and
40          represents approximately the 90th percentile of intake of community water in the U. S.
41          population. Other intake assumptions (e.g., mean versus upper percentile) can be used in
42          risk assessments, depending on target population characteristics and assessment needs.
43
44       •  Endpoints calculated. As can be seen in Table 5-1, different assessments have
45          calculated a range of risk endpoints, including EDoiS, LEDMs, lifetime cancer risks, CSFs,
46          and drinking water concentrations corresponding to various cancer risk levels. As
47          discussed in Section 5.3.8.2, this can create some difficulty in comparing the results
48          across assessments,  since converting from one measure to another can require

                                                118      DRAFT—DO NOT CITE OR QUOTE

-------
 1          assumptions related to exposure that may not have been clearly specified.  Where they
 2          have been calculated, the most commonly used and easily comparable endpoints are
 3          provided, including drinking water unit risks (lifetime cancer incidence associated with
 4          1 ug/L exposure), estimated cancer risk at 10 ug/L in drinking water, and the drinking
 5          water concentration associated with a lifetime cancer risk of 10"4.
 6
 7          Given the many features of the risk assessment for arsenic that have changed over time, it
 8    is not surprising that the magnitude of the risk estimates has also varied from assessment to
 9    assessment. As discussed above, the CSF from U.S. EPA's (1988b) assessment, which is
10    derived based on skin cancer prevalence, is not directly comparable to CSFs derived from
11    internal cancer data in the later assessments.  Section 5.3.8.2 discusses modeling methods and
12    assumptions used in the current assessment, describing precisely how they differ from previous
13    analyses.
      5.3.2. Choice of Study/Data, Estimation Approach, and Input Assumptions
14          As discussed in Section 4.2, the  few  animal carcinogenicity bioassays that have been
15    conducted on inorganic arsenic compounds do not provide data of high enough quality to use in
16    human dose-response modeling (NRC, 2001; SAB, 2000, 2007). There are, however, several
17    epidemiologic studies that relate human exposures to arsenic in drinking water to cancer risk.
18    NRC (2001) and SAB (2007) concluded that the epidemiological studies by Chen et al. (1988a,
19    1992) and Wu et al. (1989) that use the  southwestern Taiwanese population provide the best
20    available data for conducting a quantitative risk assessment for exposure to arsenic in drinking
21    water. SAB (2007) cited the important  strengths of the data, including the large population,
22    extensive follow-up (almost 900,000 person-years), large number of exposed villages (42), large
23    number of lung and bladder cancer deaths (441), reliability of the population and mortality
24    counts, and stability of residential patterns, stating that:
25       •  ".. .in view of the size and statistical  stability of the database relative to other studies, the
26          reliability of the population and  mortality counts, the stability of residential patterns, and
27          the inclusion of long-term exposures, it is the Panel's view that this [the Taiwanese]
28          database remains, at this time, the most appropriate choice for estimating cancer risk
29          among humans.  Supporting this view is the  fact that the datasets from Taiwan have been
30          subjected to many years of peer  review as part of published studies."

31          In keeping with SAB's recommendations, epidemiological studies by Smith et al. (1998)
32    and Ferreccio et al. (2000) on arsenic-related lung cancer in Chile, as well as studies by Chiou et
33    al. (2001) and Chen et al. (2004a), were evaluated (see Section 4.1 and Appendix B); however,
34    these studies were not considered to be  of comparable quality to the Taiwanese data set for use
35    in the quantitative assessment. The dose-response estimation discussed below, like previous
36    analyses, is based on the southwest Taiwanese data and incorporates the NRC and SAB
37    recommendations for modeling approaches and sensitivity analyses.
                                                119       DRAFT—DO NOT CITE OR QUOTE

-------
      5.3.3. Dose-Response Model Selection for Cancer Mortality in Taiwan
 1          Despite the high quality of the data set, estimation of dose-response relationships based
 2    on the Taiwanese data is challenging for a number of reasons. First, owing to the "ecological"
 3    nature of the study, drinking water exposure information is not available for individual study
 4    subjects. Instead,  drinking water arsenic exposure must be estimated based on measured arsenic
 5    concentrations in wells serving the 42 population groups ("villages") that constitute the study
 6    population. For 20 of the 42 villages, water was supplied by a single well at the time of
 7    sampling.  For another 10 villages, water was  supplied by two wells; the remaining villages used
 8    more than two wells.  Data provided are related to all the arsenic measurements for each well in
 9    each village, but no information is available concerning the time variability of arsenic levels in
10    individual wells.
11          In addition to villages where drinking water arsenic concentrations were measured, the
12    epidemiological data used in this assessment include information on the cancer mortality in two
13    reference populations (southwest Taiwan and  all of Taiwan) for the same period covered by the
14    Chen et al. (1988a, 1992) studies.  Drinking water concentrations for the reference populations
15    were not measured, but are assumed to be lower than those seen in the 42 arsenic-exposed
16    villages (zero drinking water arsenic intake was assumed for the reference populations). As
17    discussed below, the data on the nonwater arsenic intakes available for both the exposed and
18    reference populations are very limited (Schoof et al., 1998), so the impacts of different
19    assumptions are explored through a  sensitivity analysis.
20          It is clear that cancer mortality in the reference population and in the arsenic-exposed
21    villages is strongly age-dependent, with the older study subjects generally exhibiting higher
22    mortality.  The age-dependence does not appear to be monotonic, however, but rather peaks
23    around age 60 and declines thereafter. This non-linear age-dependence complicates the
24    estimation of dose-response relationships because it requires the estimation of models using non-
25    standard methods.
26          Chen et al.  (1992) used an Armitage-Doll time-to-tumor model to estimate cancer risks as
27    a function of dose  in this population for 20-year age strata, but the model they used assumed
28    monotonically increasing cancer risk with age. As discussed below, using narrower age strata (5
29    years), the non-monotonic dependence of cancer risk on age becomes more apparent.  Morales et
30    al. (2000) used a variety of non-linear models  to fit dose-response functions to data derived from
31    the Chen et al. (1988a, 1992) and Wu et al. (1989) studies. They derived cancer slope estimates
32    for arsenic-associated cancers of the bladder, liver, and lung by using Poisson regression with a
33    number of different methods for expressing the dependence of risks on age and arsenic intake.
34    When no reference population was included in the data, the best-fitting model included a
35    quadratic function of age and a linear exponential term for dose.  When the southwest Taiwan
36    reference population was included in the risk modeling, the best-fitting model again included a

                                                120       DRAFT—DO NOT CITE OR QUOTE

-------
 1    quadratic age model, but an exponential function of log-transformed dose.  A number of other
 2    models with different age and dose terms were found to fit nearly as well as judged by the
 3    Akaike Information criterion (AIC).  Many of the models also were very sensitive to changes in
 4    input assumptions.
 5          NRC (2001) reviewed the U.S. EPA (2001) cancer assessment including application of
 6    the model from the Morales et al. (2000) study and conducted independent analyses of the data
 7    in order to systematically evaluate the effects of different modeling approaches, assumptions
 8    related to background cancer rates, and individual variability in exposures. As noted above, they
 9    recommended two specific changes to EPA's modeling approach; the inclusion of a reference
10    population, and the use of an additive (rather than multiplicative) linear dose term in the Poisson
11    regression. SAB (2007) also reviewed EPA's modeling procedures. Given the NRC
12    recommendations and results of the SAB review, the current model (see Section 5.3.7) employs
13    the following approaches:
14
15       •  Poisson regression (of cancer mortality against age and dose) fit by maximum likelihood
16          estimation (MLE).
17
18       •  A quadratic age model.
19
20       •  Additive linear dose term.
21
22       •  Confidence limits on the dose terms estimated by profile likelihood.
23
24       •  Primary risk estimates derived for the data set that includes the southwest Taiwan
25          reference population.
26
27          As recommended by SAB, sensitivity analyses  were conducted to evaluate the impacts of
28    different modeling assumptions (nonwater arsenic intake, daily water intake, and reference
29    population) on risk estimates. Several different model forms (quadratic, exponential linear, and
30    exponential quadratic dose transformations) also were  evaluated (see Section 5.3.8.4 for further
31    detail).
      5.3.4.  Selection of Cancer Endpoints and Estimation of Risks for U.S. Populations
32          Lung and bladder cancer mortality in the Taiwanese population have been chosen as
33    endpoints in the dose-response modeling because they  are the internal cancers most consistently
34    observed and best characterized in epidemiological studies of arsenic exposure (U.S. EPA, 2001;
35    NRC,  2001). Oral CSFs and other risk metrics were calculated  separately for each endpoint and
36    gender.
37          Although liver cancer risks also were examined by Morales et al. (2000), they were not
38    included in the quantitative risk assessment because the observed liver cancer mortality in the

                                                121       DRAFT—DO NOT CITE OR QUOTE

-------
 1    southwest Taiwanese population was thought to be affected by a high incidence of viral
 2    hepatitis, which made attribution of risks to arsenic problematic. As noted in Section 4.1,
 3    arsenic-related skin cancer also has been noted in the Taiwanese population (and in other
 4    arsenic-exposed groups), but this endpoint was not included in the cancer risk assessment for
 5    several reasons. The high mortality rates for internal cancers, compared to skin cancers which
 6    are rarely fatal, makes the internal cancers an appropriate critical health endpoints for the cancer
 7    risk assessment. In addition, the internal cancers were identified as the critical endpoints
 8    because the estimated cancer potency of arsenic for lung and bladder cancers was much greater
 9    than the potency estimated for skin cancers (see Section 5.3.8.1). The development of pre-
10    cancerous skin lesions (as reported by Ahsan et al., 2006) is being addressed separately in EPA's
11    noncancer risk assessment.
12           The current risk model includes multiplicative terms for age and dose.  Therefore, the
13    risk calculated for a target population (e.g., a U.S.  population exposed to arsenic in drinking
14    water) depends on the "background" cancer risk, i.e., the expected age-specific cancer risk in the
15    U.S. population in the absence of arsenic exposure. Morales et al. (2000) calculated lifetime
16    arsenic-related mortality risks for the U.S.  population exposed to different drinking water
17    concentrations by applying age-specific hazard functions (derived from the dose-response
18    models estimated for the Taiwanese population) to a "life table" of age-specific probabilities of
19    death for the U.S. population.  These calculations were based on data from 1996.
20           In response to comments from NRC and SAB, a slightly different approach to estimate
21    cancer risks for U.S. populations is being used.  In the following analysis, arsenic concentrations
22    corresponding to an additional 1% lifetime cancer incidence (effective dose; ED01 values) above
23    "background" are derived for each endpoint. Also derived are lowest effective dose (LED01)
24    values, which represent the lower confidence limits on the dose corresponding to  a one percent
25    lifetime incidence risk in the U.S. population. Consistent with EPA's Guidelines for Carcinogen
26    Risk Assessment (U.S. EPA, 2005a) and the NRC (2001) cancer assessment, risk estimates are
27    derived based on a linear extrapolation  from the points of departure (LEDOTs for lung, bladder,
28    and combined cancers) because the MOA for inorganic arsenic is unknown.
29           The ED01 and LEDM values are  estimated using a variation on the "BEIRIV" model
30    derived for use in estimating population cancer risks for radionuclide exposures (NRC, 2001).
31    This method, which is described further in Section 5.3.7.3 and Appendix E.2, includes the
32    application of relative cancer risk estimate derived from the Taiwanese dose-response
33    assessment multiplicatively to age-specific cancer risks for the United States.  In this model, the
34    background hazard consists  of age-specific cancer incidence data for bladder and lung cancer
35    from the United States for the years 2000 to 2003 (NCI, 2006).  The ratios of cancer mortality to
36    incidence for arsenic-related cancers are assumed to be the same in the U.S. and Taiwanese
37    populations.

                                                122       DRAFT—DO NOT CITE OR QUOTE

-------
      5.3.5. Nonwater Arsenic Intake and Drinking Water Consumption
 1          It is important to clarify that the nonwater arsenic intake value corresponds to the arsenic
 2    amount from dietary sources (rice and yams, the dietary staples for the Taiwanese population in
 3    the endemic area) only.  It does not include the arsenic intake value from water used for cooking
 4    rice or produce, which was addressed separately via sensitivity analysis modeling with higher
 5    water intake values.
 6          For the baseline risk calculations, the nonwater arsenic intake was assumed to be 10
 7    ug/day for the reference and exposed populations. Although the data supporting this value are
 8    scarce, it appears to be a reasonable intake estimate for the reference populations based on the
 9    available information. U.S. EPA (1989) estimated the arsenic intake based on soil arsenic level
10    and rice consumption in Taiwan to be between 2 and 16 ug/day.  The higher value was presumed
11    to result from possible soil contamination by organic arsenical herbicides applications. U.S.
12    EPA (1989) found no reliable data to estimate arsenic intake from sweet potato (yam)
13    consumption by the southwest Taiwanese population. In a separate study, Schoof et al. (1998)
14    estimated that the total inorganic arsenic intake from food sources in the endemic area in Taiwan
15    ranged between  15 and 211 ug/day, with the average intake value as 50 ug/day.  This arsenic
16    intake value is based on analysis of limited rice and yam samples collected in the endemic area
17    of Taiwan during 1993  and 1995 (Schoof et al., 1998). It is likely that the arsenic intake in the
18    non-endemic area (background arsenic intake value for reference population) is lower than that
19    reported in the endemic area.
20          EPA also examined the arsenic intake value from food sources in countries where the
21    arsenic exposures are much lower than in Taiwan. The average nonwater inorganic arsenic
22    intake from food consumption is reported to range from 8.3 to 14 ug/day in the United States and
23    from 4.8 to 12.7 ug/day in Canada, with variation across age groups (Yost et al., 1998). Based
24    on the available  information, EPA selected 10 ug/day as the best estimate for nonwater arsenic
25    intake (food sources) in baseline calculations. Alternate values of nonwater arsenic intake were
26    also explored in the sensitivity analysis (Section 5.3.8.3).
27          NRC (1999) reported the background arsenic intake of 50 ug/day in endemic areas based
28    on the Schoof et al. (1998) findings.  It is not clear if this value was ever used for dose-response
29    modeling in estimating bladder cancer risk. However, NRC (2001) included the background
30    intake of 30 ug/day in the dose-response modeling; the basis for the latter value is not clear.
31    NRC (2001) also reported that there was no difference in the lung and bladder cancer risk
32    estimates when 30 or 50 ug/day were used as the nonwater intake values in the exposed
33    populations.  It is not clear if NRC (2001) assumed any nonwater arsenic intake value for the
34    reference populations.  In the draft Toxicological Review submitted to SAB in 2005 (U.S. EPA,
35    2005c), nonwater arsenic intake values of 0, 30, and 50 ug/day were assumed for the exposed
36    populations only, and the background inorganic arsenic intake was assumed to be zero for the

                                                123       DRAFT—DO NOT CITE OR QUOTE

-------
 1    reference populations.  SAB (2007) recommended that the background arsenic intake for
 2    reference (control) populations should not be assumed to be zero.  However, SAB did not specify
 3    a nonwater inorganic arsenic intake value for the reference population.
 4          Given the state of the available data and the recommendations from SAB, EPA has
 5    assumed 10 ug/day nonwater arsenic intake for the current assessment for both reference and
 6    exposed populations in the baseline risk calculations.  EPA also evaluated 0, 30, and 50 ug/day
 7    for dietary arsenic intake assumption for reference populations, and up to 200 ug/day for
 8    exposed populations. The high-end background arsenic intake value was recommended by SAB
 9    in 2007 (i.e., the background arsenic intake value in the exposed populations as high as 200
10    ug/day should be included to assess the impact in lung and bladder cancer risk estimates)
11    (Section 5.3.8.3).
12          In the current assessment, the drinking water consumptions for Taiwanese males and
13    females are assumed to be 3.5 L/day and 2.0 L/day, respectively, in the baseline risk
14    calculations. These values are consistent with the assumptions applied by U.S. EPA (1988b),
15    Chen et al. (1992), and NRC (1999 and 2001) for cancer risk estimations. There is conflicting
16    information concerning the  extent to which these values include both direct drinking water
17    consumption and water used for cooking. To examine the impact of additional water
18    consumption in cancer risk estimations, NRC (2001) also examined different ratios of water
19    intake-rates between Taiwanese and U.S. populations (up to ratio of 3.0).
20          In the U.S. EPA (1989) report, the arsenic workgroup estimated that the total water
21    consumption for the Taiwanese men, including the water used for cooking rice and yams (the
22    dietary staples in the southwest Taiwanese population), was 4.5 L/day since Taiwanese workers
23    could drink 3.0 to 4.0 L/day of water and the 3.5 L/day seemed to be a reasonable estimate for
24    direct water consumption.  Indirect water consumption from cooking rice and yams was
25    estimated to be 1.0 L/day.  The basis for the derivation of the drinking water values in the U.S.
26    EPA (1989) report is approximate and gathered from very limited populations (three or four
27    residents were surveyed). In the Arsenic Rule (U.S. EPA, 2001), the total water Taiwanese
28    consumption rates (including water used for cooking) were assumed to be 4.5 L/day for males
29    and 3.5 L/day for females.
30          SAB  (2007) did not recommend  specific water intake values to be used for cancer risk
31    modeling in the Taiwanese populations.  Therefore, in the current assessment, the baseline water
32    intake values modeled are 3.5 L/day for  males and 2.0 L/day for females, to be consistent with
33    NRC (1999) recommendations.  In addition, a range of water consumption values (up to 5.1
34    L/day in males and 4.1  L/day in females) were evaluated in the sensitivity analysis to study the
35    impact of alternate water consumption in the cancer risk estimates. The water consumption
36    values modeled in the baseline calculations for Taiwanese populations are also close to the
37    average estimates provided  for populations in West Bengal, India (Chowdhury et al., 2001),

                                                124       DRAFT—DO  NOT CITE OR QUOTE

-------
 1   where the climate is close to Taiwan.  The average drinking water intake values for children,
 2   adult females, and adult males were reported as 2.0, 3.0, and 4.0 L/day, respectively.
 3          The drinking water consumption for the U.S. reference population is estimated to be 2.0
 4   L/day for both men and women. This is approximately equal to the 90th percentile estimate
 5   (2.014 L/day) from the 1994-1996 and 1998 data gathered as part of the Continuing Survey of
 6   Food Intake by Individuals (U.S. EPA, 2004), and is consistent with upper percentile estimates
 7   from previous surveys. Alternative assumptions about U.S. drinking water consumption result in
 8   simple reciprocal adjustments to CSF estimates (discussed further in Section 5.3.8.3). Within
 9   the range analyzed, changes in the assumptions about Taiwanese drinking water consumption
10   also result in nearly linear effects on estimated dose-response slope estimates.
     5.3.6. Dose-Response Data
11          Table 5-2 summarizes the cancer mortality data from the Morales et al. (2000) study. For
12   this assessment, the original data set containing age-specific PYR, mortality statistics, and
13   village water concentration data was obtained from Dr. Morales (Morales et al., 2000).
14          Water arsenic concentration data were provided for each village. Single concentration
15   measurements were provided for each well. For 20 of the 42 villages only data for one well was
16   reported. However, for the remaining 22 villages, multiple well concentrations were available
17   (range between 2 and 47 measurements) (NRC, 1999).  For dose-response estimation, models
18   were fit to the median well concentration for each village.  As part of the sensitivity analysis, the
19   reported maximum or minimum well arsenic concentrations were also applied to the models.
                                                125      DRAFT—DO NOT CITE OR QUOTE

-------
  Table 5-2. Cancer Mortality Data Used in the Arsenic Risk Assessment
Gender
Male
Female
Village Water
Concentration,
jig/L
<100
100-299
300-599
>600
Total
<100
100-299
300-599
>600
Total
Age
PYRa
Deaths'3
PYR
Deaths
PYR
Deaths
PYR
Deaths
PYR
Deaths
PYR
Deaths
PYR
Deaths
PYR
Deaths
PYR
Deaths
PYR
Deaths
20-30
35,818
(0, 0, 0)
18,578
(0, 0, 0)
27,556
(0, 3, 0)
16,609
(0, 0, 1)
98,561
(0, 3, 1)
27,901
(0, 0, 0)
13,381
(0, 0, 0)
19,831
(0, 0, 0)
12,988
(0, 0, 0)
74,101
(0, 0, 1)
30-49
34,196
(1, 10, 2)
16,301
(0, 4, 3)
25,544
(5, 7, 9)
15,773
(4, 12, 3)
91,814
(10, 33, 17)
32,471
(3, 1, 5)
15,514
(0, 3, 4)
24,343
(0, 5, 6)
15,540
(0, 4, 6)
87,868
(3, 13,21)
50-69
21,040
(6, 17, 12)
10,223
(7, 15, 14)
15,747
(15,23,30)
8,573
(15, 15,23)
55,583
(43, 70, 79)
21,556
(9, 6, 18)
11,357
(9, 6, 10)
16,881
(19, 6, 20)
9,084
(21,7, 28)
58,878
(58, 25, 76)
>70
4,401
(10, 4, 14)
2,166
(2, 4, 13)
3,221
(12, 6, 14)
1,224
(8, 2, 6)
11,012
(32, 16, 47)
5,047
(9, 5, 5)
2,960
(2, 5, 5)
3,848
(11,2, 10)
1,257
(7, 1, 4)
13,112
(29, 13, 24)
Total
95,455
(17,31,28)
47,268
(9, 23, 30)
72,068
(32, 39, 53)
42,179
(27, 29, 33)
256,970
(85, 122, 144)
86,975
(21, 12, 29)
43,212
(11, 14, 19)
64,903
(30, 13, 36)
38,869
(28, 12, 38)
233,959
(90, 51, 122)
1 PYR = person-years at risk
' Numbers in parentheses = number of cancer deaths due to bladder, liver, and lung cancer, respectively.
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
  5.3.7. Risk Assessment Methodology
        The cancer risk assessment for U.S. population exposure to arsenic in drinking water was
  conducted in four steps:

     •  Models were fit to the data using mg/kg-day intake metrics calculated from the estimated
        water consumption values for the Taiwanese population and village water arsenic
        concentrations, assuming a 10 ug/day nonwater dietary intake in the baseline analysis.
        Dose-response models were fit to the Morales et al. (2000) data for bladder and lung
        cancer in both genders using maximum likelihood methods (see Section 5.3.7.1).

     •  Upper confidence limits (UCLs) on the dose coefficients from the fitted models were
        estimated using the profile likelihood method (see Section 5.3.7.2).
             values for U.S. populations were calculated for each endpoint and gender based on
        the dose coefficient UCLs calculated for the Taiwanese populations in the previous step.
        Using the "BEIR IV" methodology, U.S. bladder and lung cancer incidence data for the
                                           126
                                                         DRAFT—DO NOT CITE OR QUOTE

-------
 1          years 2000 to 2003 (NCI, 2006) were used as the reference values for calculating U.S.
 2          lifetime cancer risks.  Thus, the LEDM values are expressed in terms of lifetime cancer
 3          incidence for the U.S. population (see Section 5.3.7.3).
 4
 5          The LEDoi values were used to calculate ingestion drinking water unit risks for lung and
 6    bladder cancer for arsenic-exposed men  and women in the United States. This step involved
 7    linear extrapolation from the LED01 values to zero dose and risk, yielding estimates of low-dose
 8    CSFs. Unit risk and CSF  calculations were adjusted for differences between body weights and
 9    drinking water ingestion rates in Taiwan and the United States. Other risk metrics (estimated
10    lifetime incidence risk per mg/kg-day arsenic intake and corresponding to specific drinking
11    water concentrations) were calculated for each endpoint from the LEDM values (see Section
12    5.3.7.4).
      5.3.7.1. Dose-Response Estimation Based on Taiwan Cancer Mortality Data
13          A "Poisson model" was used to fit the cancer mortality data for the Taiwanese
14    population. The general form of the Poisson model is:
15
16                                       h(x,t) = h0(t) x g(x)                      (Equation 5-1)
17
18    where: h(x,t)  = cancer mortality risk at dose "x" and age "t"
19    h0(t)  = cancer mortality risk in the reference population at age "t"
20    g(x)  = risk attributable to arsenic exposure at dose "x" (mg/kg-day)
21
22          Taiwanese cancer mortality and PYR data were available for 5-year ranges for ages 20 to
23    84.  Cancer mortality data for the southwest Taiwan reference groups also were included in the
24    preferred version of the model;  estimates were derived without the reference population and with
25    cancer mortality statistics  from  all regions of Taiwan.  In the Poisson model, which is widely
26    applied in the analysis of epidemiology data, cancer deaths are assumed to be "rare" events and
27    Poisson-distributed within each age-dose group. When hO(t) and/or g(x) are non-linear
28    functions, as is the case for arsenic, the model cannot be fit using conventional least-squares
29    regression methods or general linear models (GLM). Based on recommendations from NRC
30    (2001) and after testing a number of different models, the following model form was selected for
31    primary risk estimates based on goodness-of-fit and parsimony criteria:4
32
33                       h(x,t) = exp(ai + a2 x age + as x age2) x (1 + b x dose)      (Equation 5-2)
34
35    where:  ai, a2, as = age coefficients;  b = dose coefficient
36
       Results obtained using alternative model forms are discussed in Section 5.3.8.4.
                                                127       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Specifically, the model parameters in h(x,t) in Equation 5-2 were obtained by assuming
 2    that the number of cases in each exposure-age category has a Poisson distribution with parameter
 3    ^(x,t), Cases ~ Poisson (Py x X(x,t)), where Py is person-years, and X is the intensity of Poisson
 4    parameter at the exposure-age,(x,t), category. Because data are given in 5-year age intervals, the
 5    parameter X is related to hazard rate h which is equal to X/5.
 6          In this model, the exponential term represents "hO(t)"in Equation 5-1, the age-dependent
 7    risk of cancer at the "background" doses of arsenic (zero from drinking water and 10 ug/day
 8    from  diet in the preferred model).  The last term in the equation captures the dependency of risk
 9    on the daily ingestion dose of arsenic.
10          Cancer mortality data were stratified across 13 5-year age groups and 43 villages (42
11    exposed villages plus the reference population).  This stratification yielded 559 data points per
12    cancer endpoint for model fitting.  Mid-range values for the age ranges were standardized to
13    their mean values and treated as nuisance parameters.
14          The unit of dose used in the modeling was mg/kg-day. In the primary (baseline) risk
15    model, the estimated nonwater arsenic intake was 10 ug/day for both the exposed and reference
16    populations.  The total arsenic dose received by the population of any village was estimated as
17    the sum of the nonwater dietary intake plus the median arsenic well water concentration for the
18    village (baseline model), multiplied by the estimated water Taiwanese consumption rates (3.5
19    L/day for men, 2.0 L/day for women) and divided by estimated average body weights for
20    Taiwanese men and women (50 kg for both genders; Chen et al., 1992). The southwest
21    Taiwanese population outside of the arseniasis-endemic area (Morales et al., 2000)  served as the
22    reference population in the baseline model.
23          Values for the coefficients al, a2, a3, and b were fit using MLE methods. Likelihood
24    maximization was performed using the Solver add-in of Excel®. The MLE fits for the baseline
25    model were replicated using the Non-Linear Estimation module of Statistica®. Replicated
26    results (estimated age and dose coefficients) were identical to Solver estimates to the third
27    decimal place for all endpoints.
      5.3.7.2. Estimation of Confidence Limits on Cancer Slope Parameters
28          The LEDoi values were derived based on estimated upper confidence limits on the
29    estimated dose coefficients ("b") for each endpoint and gender. The confidence limits were
30    calculated using the likelihood profile method (Venson and Moolgavkar, 1988). In this
31    approach, the value of the dose parameter, b, was varied from its estimated mean value. The
32    ratio of the log likelihood for the best-fit model to the log likelihood for other values of "b" is
33    known to follow an approximate chi-squared  distribution with one degree of freedom.  Thus, the
34    5th and 95th confidence limits on the dose coefficient "b" correspond to the values where the
35    likelihood ratio is equal to 1.92.  Upper and lower confidence limits were calculated using
36    Solver®. The fact that the profile likelihood method ignores the likelihood impact of the age
                                                128       DRAFT—DO NOT CITE OR QUOTE

-------
 1    "nuisance parameters" implies that the calculated confidence limits are only approximate.
 2    Confidence limit calculations using other methods (empirical Bayesian simulations and
 3    "bootstrap-t") gave comparable results (within a few percent of the values estimated by profile
 4    likelihood).
      5.3.7.3. Estimation ofLED01  Values Using Relative Risk Models
 5           Once the dose coefficients were calculated, they were used to estimate arsenic-associated
 6    lifetime risks in the U.S. population.  In this analysis, LEDM values for the U.S. population were
 7    calculated using a variant of the "BEIRIV" relative risk model recommended by NRC (2001).
 8    The method applied the relative risk estimated as (1 + bUCL x dose) to the age profile of cancer
 9    incidence for the reference (U.S. male or female) population, where bUCL is the 95% upper
10    confidence limit on "b" (the arsenic coefficient from the dose-response model for the Taiwanese
11    population, estimated as explained in Section 5.3.7.2).  The BEIR IV model also takes into
12    account the effect of noncancer mortality, cancer mortality, and previous cancer incidence on the
13    number of individuals in the exposed population who survive to the start of each 5-year age
14    stratum.  To estimate cancer risks  in the U.S. population, incidence risks are calculated for each
15    5-year age stratum and summed to give an estimate of lifetime incidence.  The dose is then
16    adjusted until the estimated extra incidence risk from arsenic-associated cancer risk equals 0.01
17    (1%) for the U.S. reference population.  The dose (in mg/kg-day) that fulfills this condition  is the
18    LEDoi, which becomes the point of departure (POD) for estimating the CSF.
19           The BEIR IV model takes  as its input age-specific mortality data and lung and bladder
20    cancer incidence for the U.S. reference population.6 U.S.  cancer incidence was estimated in this
21    analysis based on mortality data for the year 2000 (NCHS, 2000).  Lung and bladder incidence
22    data for the years 2000 to 2003 were obtained from the National Cancer Institute's  SEER
23    (surveillance epidemiology and end result) program (NCI, 2006). Arsenic intakes resulting in
24    10"4 lifetime risks were estimated using Solver®. Details  of the relative risk methodology are
25    provided  in Appendix E.2.
      5.3.7.4. Estimation of Unit Risks
26           For each endpoint and gender, the slope of a line from the LED0i  dose through the
27    intercept  (water-related arsenic dose = 0, water-related arsenic risk = 0)  was calculated. The
28    slopes of these lines represent the  oral CSF for the endpoint:
29
       The empirical Bayes modeling involved taking random samples within the neighborhoods of the MLE coefficient
      values, calculating the log likelihood, and after many iterations, building up an estimate of the posterior distribution
      of the "b" coefficient (mean and standard error). Confidence limits were then estimated assuming the posterior
      probability of b was normally distributed.

       Note that the age dependence estimated for the Taiwanese population—represented by the parameters a1; a2, and
      a3—is specific to that population, and is not carried over to the United States.
                                                 129       DRAFT—DO NOT CITE OR QUOTE

-------
 1                              oral CSF (per mg/kg-day) = 0.01/LEDM              (Equation 5-3)
 2
 3          Linear low-dose extrapolation was employed consistent with EPA's finding that
 4    insufficient mode of action data are available to justify the use of non-linear, low-dose models
 5    (Section 4.6.3.2). Unit risks (cancer risk per ug/L arsenic in drinking water) also were
 6    estimated:
 7
 8    unit risk (per ug/L) = CSF (per mg/kg-day) x 0.001  x DW/BW                  (Equation 5-4)
 9
10    where: 0.001  = conversion from milligrams to micrograms
11          BW   = body weight for exposed population in kilograms (U.S. male and female)
12          DW   = daily drinking water consumption for exposed population in liters (U.S. male
13                 and female)
14          As discussed previously, the estimated drinking water consumption for the U.S. adult
15    population is 2.0 L/day for both males and females.  U.S. male and female body weights are
16    estimated to be 70 kg. The 2.0 L/day is a standard factor used in EPA IRIS assessments, and
17    represents approximately the 90th percentile of intake of community water in the U.S.
18    population. Other intake assumptions (e.g., mean versus upper percentile) can be used in risk
19    assessments, depending on target population characteristics and assessment needs.
      5.3.8. Results
      5.3.8.1. Ingestion Pathway Oral CSFs and Unit Risks
20          Table 5-3 presents the estimated risk metrics for lung and bladder cancers in males and
21    females under baseline assumptions (see Footnote "a" to the table for baseline modeling
22    assumptions).
23          The estimated oral CSF for female lung cancer (16.6 per mg/kg-day) is higher than that
24    for males (6.7 per mg/kg-day), but the bladder cancer oral CSFs for males and females are
25    comparable (11.2 and 10.5 per mg/kg-day, respectively). Drinking water unit risks for lung
26    cancer are 1.9 x 10"4 and 4.8 x 10"4 per ug/L, respectively, for males and females while the
27    drinking water unit risks for bladder cancer are 3.2 x 10"4 and 3.0 x 10"4 per ug/L, respectively.
28    Estimated lifetime incidence risks corresponding to  10 ug/L arsenic in drinking water follow
29    similar patterns for the various endpoints. Estimated drinking water concentrations associated
30    with 10"4 lifetime incidence range from 0.21 ug/L (female lung  cancer) to 0.52 ug/L (male lung
31    cancer).
                                                130       DRAFT—DO NOT CITE OR QUOTE

-------
Table 5-3.  Cancer Incidence Risk Estimates for Lung and Bladder Cancers in Males and
           Females3
Metric
Lung Cancer
Bladder Cancer
Males
ED0i, mg/kg-day
LEDoi, mg/kg-day
Oral CSF, per mg/kg-day
Unit risk, per ug/L drinking water
Lifetime incidence risk at 10 ug/L in drinking water
Water concentration for 10~4 risk, ug/L
1.9E-03
1.5E-03
6.7
1.9E-04
1.9E-03
0.52
1.1E-03
8.9E-04
11.2
3.2E-04
3.2E-03
0.31
Females
ED0i, mg/kg-day
LEDoi, mg/kg-day
Oral CSF, per mg/kg-day
Unit risk, per ug/L drinking water
Lifetime incidence risk at 10 ug/L in drinking water
Water concentration for 10"4 risk, ug/L
7.5E-04
6.0E-04
16.6
4.8E-04
4.8E-03
0.21
1.2E-03
9.5E-04
10.5
3.0E-04
3.0E-03
0.33
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
 aBaseline assumptions: reference population = southwest Taiwan; Taiwanese male and female body weight = 50
 kg, Taiwanese male water intake = 3.5 L/day, Taiwanese female water intake = 2.0 L/day; reference and exposed
 population nonwater arsenic intake = 10 ug/day. Male and female U.S. body weights are assumed to be 70 kg,
 and U.S. water intake for both males and females is assumed to be 2.0 L/day.

       Arsenic-related cancer risks also are calculated for the population as a whole, that is, for
combined bladder and lung cancer incidence in a population composed of both men and women.
 In this analysis, total cancer risk (lung plus bladder) for males and females is calculated by
combining the risk for the individual tumor types. Upper confidence limits on the combined
cancer risks can be calculated based in the assumption that the uncertainties in the two CSFs are
both normally distributed. If this is the case, the 95% upper bound, U, for the combined cancer
potency can be calculated as:
U =
                                                   +(u2 -m2)2
                                                                           (Equation 5-5)
where mi and ui, i = 1,2, are respectively mean and 95% upper bound cancer potency for the two
tumor types. The results of these calculations are summarized in Table 5-4. Using this
approach, the combined cancer potency factor estimate for males is 16.9 per mg/kg-day for
males and 25.7 per mg/kg-day for females.  The estimated drinking water unit risk for combined
male lung and bladder cancer is 4.8 x 10"4 per ug/L; for females, the estimated value is 7.3 x 10"4
per ug/L. The drinking water concentrations corresponding to 10-4 combined cancer risks for
males and females are 0.21 and 0.14 ug/L, respectively.
                                           131
                           DRAFT—DO NOT CITE OR QUOTE

-------
    Table 5-4. Combined Lung and Bladder Cancer Incidence Risk Estimate for the U.S.
              Population (Males and Females)
Metric
Oral CSF, per mg/kg-day
Unit risk, per ug/L drinking water
Lifetime incidence risk at 10 ug/L
in drinking water
Water concentration for 10~4 risk,
Mg/L
Male Combined
Lung+Bladder
16.9
4.8E-04
4.8E-03
0.21
Female Combined
Lung+Bladder
25.7
7.3E-04
7.3E-03
0.14
1          Figure 5-1 shows the estimated oral CSFs for each of the endpoints separately, along
2   with oral CSF estimates for the combined cancers in males and females.  In keeping with EPA
3   policy, the combined oral CSF for women (25.7 per mg/kg-day) is appropriate for use in
4   establishing health criteria, since, based on the available data, women appear to be the
5   more sensitive group.
           OJ
           CJ
           C
           «
           U
Male Lung    Male
           Bladder
  Male      Female     Female    Female
Combined    Lung      Bladder   Combined
                 Figure 5-1. Estimated oral CSFs for individual and combined cancer
                            endpoints.

    5.3.8.2. Comparison to Previous Cancer Risk Estimates
6         As discussed in Section 5.3.1, a number of risk assessments have been conducted by EPA
7   and others. Results of the present dose-response assessment were compared to cancer risk
8   estimates derived from the same and other data sets in previous studies (NRC, 2001; U.S. EPA,
                                            132       DRAFT—DO NOT CITE OR QUOTE

-------
 1    2005c). Note that the results of the U.S. EPA (1988b) analysis, which estimated a CSF of 1.0-
 2    2.0 per mg/kg-day, are not comparable to the results of the current assessment (CSF 25.7 per
 3    mg/kg-day), because the former was based on skin cancer, while all of the more recent analyses
 4    estimate risks of internal (lung and bladder) cancers.  Thus, the detailed comparisons in this
 5    section are limited to assessments that also address lung and bladder cancer.  The drinking water
 6    standard (U.S. EPA, 2001) also provides numerical risk estimates for exposures to arsenic in
 7    drinking water. However, Tables III.D-2(a)  and (b) of the rule (U.S. EPA, 2001) display ranges
 8    of cancer risks for populations exposed to distributions of arsenic concentrations in drinking
 9    water at and above the proposed MCL  options. Thus, the numerical risk results of that analysis
10    are also not directly comparable to the NRC  (2001), U.S. EPA (2005c), and current assessments,
11    which apply to populations exposed to single concentrations. In the analyses that follow, some
12    of the risk comparisons are based on mortality estimates that have been converted to incidence
13    using recent U.S. incidence-mortality ratios.  This conversion introduces additional uncertainty
14    into the comparisons; different results would have been obtained had the incidence been
15    modeled directly rather than estimated after the fact.
      5.3.8.3. EDoi andLED01 Estimates From Chen et aL (1988a, 1992), Ferreccio et al (2000),
            and Chiou et al. (2001)
16          Consistent with SAB (2007) recommendations, Table 5-5  presents risk estimates from
17    previous studies and compares them to estimates derived in this analysis.  The estimates in Table
18    5-5 come from Table 5-3 of NRC (2001), and include EDOI and LEDM estimates (expressed as
19    ug/L arsenic  in drinking water) from a number of studies of arsenic-related cancer risks in Chile
20    (Ferreccio et al., 2000) and Taiwan (Chiou et al., 2001; Chen et al., 1988a, 1992).
21          NRC  calculated EDOI and LEDM values for lung and bladder cancer mortality from the
22    same Taiwanese cohort used in the current assessment, based on the results presented in Chen et
23    al. (1988a,  1992), but without a reference population.  In addition, these values do not account
24    for differences in drinking water consumption between the U.S. and Taiwanese populations, and
25    did not apply life-table adjustments.
                                                133       DRAFT—DO NOT CITE OR QUOTE

-------
      Table 5-5. Comparison of ED0i and LED,/ Estimates From Past Studies'1 With Those From
                the Current Analysis
Study
Chenetal. (1988a,
1992), Taiwan
Ferreccio et al.
(2000), Chile
Chiouetal. (2001),
Taiwan
Current analysis
Male Lung
EDoi
38-84
5-17
—
66
LED01
37-72
3-14
—
52
Female Lung
EDoi
33-
94
7-27
—
26
LED01
31-84
5-21
—
21
Male Bladder
EDoi
102-317
—
129-500+
40
LED01
94-286
—
42-
500+
31
Female Bladder
EDoi
138-443
—
231-500+
41
LED01
125-406
—
88-500+
33
        a Units = ug/L arsenic in drinking water
        b Source of estimates: NRC (2001)
 1          NRC also estimated EDOI and LEDM values based on data from the Ferreccio et al. (2000)
 2    case-control study of male and female lung cancer data from a Chilean population that included
 3    151 lung cancer cases and 419 controls. The EDOI and LED01 derived by NRC were obtained by
 4    linear regression of mortality odds ratio estimates on exposures, with the intercept forced to a
 5    value of 1.0 at zero exposure.  These estimates are shown in the second row of Table 5-5.
 6    Multiplicative linear dose and log dose models were used to derive EDOI and LED0i estimates
 7    from the study by Chiou et al. (2001) of urinary tract cancer incidence over a 4-year period in
 8    8,000 Taiwanese exposed to arsenic in drinking water. These results are presented in the third
 9    row of Table 5-5. Where ranges are given in the table, the minimum and maximum values
10    represent the lowest and highest EDOI  or LEDM estimates that were derived when different
11    models were used.
12          The bottom row of the table shows the EDOI  and LEDM values for cancer incidence
13    derived in this analysis using the Poisson regression and BEIRIV models.  The EDOI and LEDM
14    values for lung cancer derived in the current assessment fall within, or are close to, the ranges
15    estimated from the Chen et al. (1988a,  1992) data. This finding is not surprising because the
16    results are estimated for the same cohort in both cases, and because the case mortality for lung
17    cancer is so high (nearly 100%). The EDOI and LED0i values derived in the current assessment
18    are, however, higher than those estimated by Ferreccio et al. (2000). One possible  explanation
19    involves differences in modeling methods; to estimate EDOI and LED0i values from the Ferreccio
20    study, NRC applied linear regression to the odds ratio estimates, forcing  the intercept through
21    1.0 at zero dose. Thus, these values must be considered highly uncertain. The differences also
22    may be due to differences in exposure conditions (e.g., NRC did not account for differences in
23    drinking water intake between the Chilean and U.S. populations) or other covariates (e.g.,
24    smoking) between the two studies.
                                               134
DRAFT—DO NOT CITE OR QUOTE

-------
 1          For bladder cancer, the ED01 and LEDM values estimated in this analysis are lower (2.5-
 2    to 10-fold) than those derived from the Chen et al. studies (1988a, 1992). In addition to the
 3    differences in modeling approaches outlined above, another possible reason for this difference is
 4    that the Chen et al.  (1988a, 1992) studies are based on bladder cancer mortality, while the ED01
 5    and LEDM values in this analysis are for bladder cancer incidence. Adjustment for bladder cancer
 6    case mortality (in the order of 16-20%) would make EPA's current results much more similar to
 7    those of Chen et al. (1988a, 1992).
 8          Finally, the  ED01 and LEDOT values from the current analysis are below the lower end of
 9    the ranges estimated by Chiou et al. (2001). Reasons for this finding are not entirely clear. The
10    sensitivity of the Chiou et al. study may have been limited by the short follow-up period (NRC,
11    2001), and only 18  total urinary tract cancers were identified in the study.  Only four exposure
12    categories were analyzed (less than 10 ug/L, 10-50, 50-100, and more than 100  ug/L in water;
13    nonwater exposures were not evaluated).  The low sensitivity could have caused the ED01 and
14    LEDoi estimates derived by Chiou et al. (2001) to be biased upward from what would have been
15    seen with a more extended follow-up period.
      5.3.8.4. Estimated Risk Associated With 10 fig/L Drinking Water Arsenic From NRC
            (2001)
16          Table 5-6 provides an additional set of comparisons between the current risk estimates
17    and the results from a previous  analysis by NRC (2001).  Lifetime incidence risks are presented
18    for a hypothetical U.S. population exposed to 10 ug/L arsenic in drinking water. NRC (2001)
19    estimated arsenic-associated risks using an "additive Poisson model with dose entered as a linear
20    term and using the BEIRIV formula" (p.  201).
      Table 5-6. Comparison of cancer risk assessment results with estimates from NRC (2001)
Source of Estimate
NRC (2001), Taiwan
Current analysis
Estimated Cancer Incidence at 10 ug/L Arsenic in
Drinking Water (per 10,000 Exposed Population)
Bladder
Male
23
32
Female
12
30
Lung
Male
14
19
Female
18
48
              a The original mortality risk estimates from U.S. EPA (2005c) were multiplied by incidence-
              mortality ratios for the various endpoints to obtain incidence estimates. For the Taiwanese
              populations, case mortality for lung cancer was assumed to be 100% and mortality for bladder
              cancer was assumed to be 80% (NRC, 2001).

21          The incidence risks derived in the current analysis, however, are reasonably close, but not
22    identical, to the NRC (2001) estimates. Differences in the calculated cancer potency relate to
23    several factors. Changes in the assumed drinking water intake in females in the current
                                                13 5       DRAFT—DO NOT CITE OR QUOTE

-------
 1    assessment compared to the NRC (2001) and U.S. EPA (2005c) analyses are summarized in
 2    Table 5-7. In particular, the change in the assumed ratios of Taiwanese/U.S. female water intake
 3    from 2.8 in the earlier assessments to 1.4 in the current analysis are relevant to the differences in
 4    risk shown in Table 5-6.  The lower ratio in the current analysis translates into a slightly greater
 5    than 2-fold greater estimated risk for females in the current assessment than in the NRC (2001)
 6    and current analyses.
      Table 5-7. Drinking water intake and body weight assumptions in females in recent arsenic
                risk assessments
Assessment
NRC (2001)
U.S. EPA (2005c)
Current analysis
Body Weight, kg
Taiwan
50
50
50
U.S.
70
70
70
Water Intake, L/day
Taiwan
2
2
2
U.S.
1
1
2
Ratio of Taiwan/U.S.
Drinking Water Intake
2.8
2.8
1.4
 7          In addition, the NRC (2001) risk estimates are based on maximum likelihood estimates
 8    (MLE) of the arsenic slope parameters in the Poisson regression, while U.S. EPA (2005c) and
 9    the current assessment derive risks based on the statistical upper confidence bounds on these
10    parameters. As shown in Table 5-3, the difference between the MLE estimates (ED01 values)
11    compared to the upper confidence limit (LEDM) is on the order of 20%.  This would translate into
12    approximately 20% greater risks calculated based on the upper confidence limit values compared
13    to the MLE estimates.
14           The use of more recent cancer incidence and mortality data in the BEIRIV model than
15    in the previous risk assessments also probably contributes to the differences in risks in Table 5-6.
16    Also,  the current assessment includes a modification to the BEIR IV model suggested by Gail et
17    al. (1999) for obtaining more accurate estimates of incidence within multi-year age strata. The
18    modifications to the model are described in detail in Appendix E.2.
19          Changes in the assumptions related to nonwater arsenic intake also would be expected to
20    have small to moderate effects on the results within the range in question. In this assessment,
21    both the reference and exposed populations are assumed to receive 10 ug/day  nonwater arsenic
22    intake (see Section 5.3.5). Section 5.3.8.3 presents the results of uncertainty analyses that
23    explore the effects of changes in selected modeling assumptions, including nonwater arsenic
24    intake, on the risk estimates.
25          The cancer risk estimates presented in Table 5-8 for consumption of drinking water with
26    specified arsenic concentrations provide information that is scientifically equivalent to estimates
27    of CSFs. The NRC's (200l)recommended risk models provide estimates that consumption of
28    drinking water containing 10 ug/L arsenic is associated with the site specific cancer risks below.
                                                136
DRAFT—DO NOT CITE OR QUOTE

-------
 1    Note that the same CSF values, other than small differences due to rounding error, would be
 2    obtained starting with any of the water concentrations presented in the NRC (2001) Table S-l.
 3
 4
 5
 6
 7
 8
 9
10
11
12
      Table 5-8. Theoretical maximum likelihood estimates of excess lifetime risk (incidence per
                10,000 people) of lung cancer and bladder cancer for US populations
Arsenic
concentration
(HS/L)
10
Bladder
Male
23
Female
12
Lung
Male
14
Female
18
The equivalent CSFs can be calculated as follows:
Using the exposure factors for US populations applied in NRC (2001), consumption of
10 ug/L arsenic in drinking water results in a daily exposure of (10 ug/L) x (1 L/d) x
(1 mg/1,000 ug) x (1/70 kg) = 0.000143 mg/kg-d of inorganic arsenic. As the NRC risk
estimates are linear (proportional to dose) for these exposures, equivalent CSF values
come from the equation:
Risk = CSF (per mg/kg-d) x dose (mg/kg-d)
As an example, applying this equation to bladder cancers in females:
12 x 10'4 = CSF x 0.000143 mg/kg-d, or CSF = 8.4 per mg/kg-d
Thus the CSF estimates resulting from Table 5-8 are shown below in Table 5-9.
      Table 5-9. Arsenic oral CSFs (per mg/kg-d) for lung cancer and bladder cancer in US
                populations
Bladder
Male
16
Female
8
Lung
Male
10
Female
13
13          As these are maximum likelihood estimates, it is appropriate to add risks across the two
14    sites resulting in combined CSFs for lung and bladder cancer of 21 and 26 per mg/kg-d in
15    females and males respectively.
      5.3.8.5. Sensitivity Analyses of Cancer Risk Estimates to Changes in Parameter Values
16          NRC (2001) and SAB (2007) recommended that the impacts of different modeling
17    assumptions and input parameter values be investigated in the risk assessment for arsenic in
18    drinking water. EPA, therefore, examined several aspects of the cancer risk modeling through
19    single-value sensitivity analysis. The Agency felt that the currently available data were
20    insufficient to support detailed probabilistic uncertainty and variability estimation. In response
21    to SAB comments, EPA evaluated the impacts of:
                                               137
                                             DRAFT—DO NOT CITE OR QUOTE

-------
 1       •  Varying the assumed daily nonwater arsenic intake of the exposed and reference
 2          populations.  Sensitivity cases were run in which the nonwater arsenic intake in the
 3          exposed populations was varied from its default value of 10 ug/day to 0, 100, and 200
 4          ug/day. An additional case was run in which both the exposed and reference populations
 5          were assumed to receive 0, 30, and 50 ug/day nonwater arsenic exposure. Because the
 6          Poisson risk model for female bladder cancer is particularly sensitive to changes in
 7          assumptions related to nonwater arsenic intakes (see below), nonwater arsenic intake was
 8          limited to below 50 ug/day in reference populations.
 9
10       •  Varying assumptions related to drinking water intake by the exposed Taiwanese
11          population.  Cases were run in which the male drinking water consumption was varied
12          from its baseline value of 3.5 L/day to 5.1  L/day, 3.0 L/day, and 2.75 L/day. Female
13          drinking water intake in the Taiwanese population was varied from its baseline value of
14          2.0 L/day to 2.75 and 4.1 L/day.
15
16       •  Varying the arsenic well concentrations used to fit the dose-response model for the
17          Taiwanese population.  The baseline risk model used the median village arsenic
18          concentrations as the exposure metric.  In the sensitivity analysis, cases also were run
19          using the minimum and maximum well concentrations in each village.
20
21       •  Including different Taiwanese reference populations in the  dose-response assessment.
22          The baseline (southwest Taiwan) reference population was replaced by data from all
23          Taiwan.  The model also was run without any distinct reference population.
24
25       Tables 5-10 and 5-11 summarize the results of the sensitivity analysis runs.  Table 5-10
26    shows the estimated (incidence) risks associated with a drinking water concentration of 10 ug/L
27    for the U.S. population estimated when calculated using the assumptions specified in the left-
28    hand column of the table.  Table 5-11 shows the proportional changes in estimated risks in
29    relations to the baseline estimate. Figure 5-2 summarizes the impact of alternative modeling
30    assumptions, showing the ratios of estimated cancer risks to the base case estimates for changes
31    in input variables having a substantial (>20%) effect on the risk estimates.
                                                13 8       DRAFT—DO NOT CITE OR QUOTE

-------
Table 5-10. Sensitivity analysis of estimated cancer incidence risks associated with 10 ug/L
           to changes in modeling assumptions and inputs
Estimated Cancer Risk at 10 |j,g/L
Baseline (all default values)3
Nonwater arsenic intake = 0 ug/day (reference and
exposed populations)
Nonwater arsenic intake = 30 ug/day (reference and
exposed populations)
Nonwater arsenic intake = 50 ug/day (reference and
exposed populations)
Nonwater arsenic intake (exposed population) = 0
ug/day
Nonwater arsenic intake (exposed population) = 100
ug/day
Nonwater arsenic intake (exposed population) = 200
ug/day
Taiwan water consumption =3.0 L/day (M), 2.0 L/day
(F)
Taiwan water consumption =5.1 L/day (M), 4.1 L/day
(F)
Taiwan water consumption = 2.75 L/day (M, F)
Village water arsenic concentrations = minimum values
Village water arsenic concentrations = maximum values
Reference population = none
Reference population = all Taiwan
Male
Lung
1.9E-03
1.9E-03
2.0E-03
2.0E-03
1.9E-03
1.8E-03
1.7E-03
2.3E-03
1.3E-03
2.5E-03
2.5E-03
1.4E-03
1.2E-03
2.4E-03
Female
Lung
4.8E-03
4.6E-03
5.1E-03
5.5E-03
4.8E-03
4.4E-03
3.9E-03
4.8E-03
2.3E-03
3.4E-03
5.7E-03
3.5E-03
1.5E-03
3.9E-03
Male
Bladder
3.2E-03
3.0E-03
3.5E-03
3.9E-03
3.2E-03
3.0E-03
2.8E-03
3.8E-03
2.2E-03
4.1E-03
4.0E-03
2.3E-03
8.3E-04
4.8E-03
Female
Bladder
3.0E-03
2.6E-03
4.5E-03
1.1E-02
3.0E-03
2.8E-03
2.4E-03
3.0E-03
1.4E-03
2.1E-03
4.0E-03
2.1E-03
3.5E-04
6.2E-03
 aBaseline inputs: reference population = southwest Taiwan; male and female body weight = 50 kg, male water
 intake = 3.5 L/day, female water intake = 2.0 L/day, reference and exposed population nonwater arsenic intake
 = 10 |ag/day. U.S. population male and female body weights = 70 kg, male and female water consumption =
 2.0 L/day.
                                             139
DRAFT—DO NOT CITE OR QUOTE

-------
Table 5-11. Proportional Changes in Cancer Risks at 10 ug/L Associated With Changes in
           Modeling Inputs and Assumptions
Modeling Assumptions/Input Values
Baseline (all default values)3
Nonwater arsenic intake = 0 ug/day (reference and exposed
populations)
Nonwater arsenic intake = 30 ug/day (reference and exposed
populations)
Nonwater arsenic intake = 50 ug/day (reference and exposed
populations)
Nonwater arsenic intake (exposed population) = 0 ug/day
Nonwater arsenic intake (exposed population) = 100 ug/day
Nonwater arsenic intake (exposed population) = 200 ug/day
Taiwan water consumption =3.0 L/day (M), 2.0 L/day (F)
Taiwan water consumption =5.1 L/day (M), 4. 1 L/day (F)
Taiwan water consumption = 2.75 L/day (M, F)
Village water arsenic concentrations = minimum values
Village water arsenic concentrations = maximum values
Reference population = none
Reference population = all Taiwan
Male Lung
0%
0%
5%
5%
0%
-5%
-11%
21%
-32%
32%
32%
-26%
-37%
26%
Female
Lung
0%
-4%
6%
15%
0%
-8%
-19%
0%
-52%
-29%
19%
-27%
-69%
-19%
Male
Bladder
0%
-6%
9%
22%
0%
-6%
-13%
19%
-31%
28%
25%
-28%
-74%
50%
Female
Bladder
0%
-13%
50%
267%
0%
-7%
-20%
0%
-53%
-30%
33%
-30%
-88%
107%
 a Baseline inputs as described in footnote to Table 5-8.
4.0 -i

Q) ^ n
.1
0)

ro ? n
CQ ^u
o
0)
^ -in
_> 1 .U
"ro
0)
°^ 0.0

























^^


1
| |

(/)












c JS









in
\lon-water A


-
























55
0)
(0
c








-



M
§
1





















1-1



-













|f









ru~u
n

M " •— >
i >•» ^*
c -§ "o













1






















—



-









-i
















^
>s
ro ^^
R
i- CD -- i- in - i- ,-g








rv
n
n

in
llage H20 A
= minimum
>







DMale


Lung
• Female Lung
D Male
Bladder
D Female Bladder











(/)
<
$ S
3 ^C
(0 (D
>









= maximum




















rt

(D
C °
"5 Q.
IV O
Q.






—



0)
Referenc



-
1












TO C
Q- ro


             Figure 5-2. Change in arsenic-related unit risk estimates associated
                        with variations in input assumptions.
                                         140
DRAFT—DO NOT CITE OR QUOTE

-------
 1          These results indicate that varying most of the risk modeling inputs within the tested
 2    ranges have a small or moderate effect on risk estimates for most endpoints. For all of the
 3    endpoints except female bladder cancer, changing assumptions related to nonwater arsenic intake
 4    for the reference and/or exposed populations results in small changes (<25%) in the estimated
 5    oral CSF and cancer risks at 10 ug/L in drinking water. Risk estimates for female bladder
 6    cancer, in contrast, are  quite sensitive to changes in nonwater arsenic intake in the range from 0
 7    to 50 ug/day. When nonwater arsenic intake is assumed to be 30 ug/day (rather than 10 ug/day
 8    in the baseline estimate), estimated female bladder cancer risks are approximately 50% higher
 9    than under baseline assumptions. When nonwater arsenic intake increases to 50 ug/day, female
10    bladder cancer risk increases by 267% compared to baseline. The sensitivity of the risk
11    estimates is greater for  changes in reference population arsenic intake; when nonwater intake
12    increases to 100 and 200 ug/day for the exposed populations alone, the impacts on female
13    bladder cancer risks are much less (7% and 20%, respectively).
14          As  expected, the risk estimates obtained when making different assumptions concerning
15    Taiwanese drinking water consumption are very nearly inversely proportional to the assumed
16    water intake. For example, when male drinking water consumption is assumed to be 5.1 L/day,
17    rather than 3.5 L/day in the baseline case, estimated cancer risks for male lung and bladder
18    cancer are both approximately 0.69 (= 3.5/5.1) times the values derived using baseline
19    assumptions. Similar results are seen for the other endpoints.
20          Using different  exposure concentration metrics also shows relatively limited impacts on
21    the estimated cancer risks. When the village minimum water concentrations are used as inputs to
22    the Poisson risk model, the estimated cancer risks increase slightly (32%, 19%, 25%, and 33%
23    over baseline) for male and female lung and male and female bladder cancer, respectively.
24    When village maximum water concentrations are used as model inputs, the estimated cancer
25    incidence risks  decrease between 26 and 30% relative to baseline. These changes are roughly
26    reciprocal to the changes in average exposure concentrations, as expected.
27          The final two rows of Tables 5-8 and 5-9 illustrate the impact of alternative assumptions
28    about which reference populations are included in the Taiwanese risk assessment model.  When
29    no reference population is included (the Poisson model is fit only to the data from the 42
30    exposed villages), the estimated risks for all four endpoints are considerably lower than under
31    the baseline case, which included the  southwest Taiwan population. This finding is not
32    unexpected, because the addition of the relatively large reference population serves to "anchor"
33    the low-exposure end of the model and decrease the impact of the high variability ("noise") in
34    the exposed population data. When the reference population is excluded from the assessment,
35    estimated cancer risks are reduced between 37% (male lung) and 88% (female bladder cancer)
36    compared to the baseline model that included the southwest Taiwan reference populations. All
37    of the exposure-response "b" parameters  retain statistical significance, however, even when the

                                               141       DRAFT—DO NOT CITE OR QUOTE

-------
 1    reference population is excluded. Finally, including the "all Taiwan" reference population,
 2    rather than southwest Taiwan, has smaller and variable effects on the risk estimates. Predicted
 3    risks for male lung and bladder cancer are increased (decreased) by approximately 26% and
 4    19%, respectively, while risks for female lung and bladder cancer are increased by 50% and
 5    107%, respectively, compared to baseline.
 6          Based on these outcomes, it appears that the risk model results are relatively stable and
 7    react predictably to reasonable changes in exposure assumptions. The exception is female
 8    bladder cancer, for which the dose-response parameter estimated in the Poisson model is very
 9    sensitive to the assumed nonwater arsenic intake by the reference population in the range
10    between 0 and 50 ug/day. In addition, risk estimates for all endpoints are strongly affected by
11    the inclusion or exclusion of a low-dose reference population in the Poisson risk model.
      5.3.8.6. Sensitivity Analyses of Cancer Risk Estimates to Dose-Response Model Form
12          In the course of this analysis, EPA has investigated the impact of alternative model forms
13    on the cancer risks estimated for the Taiwanese and U.S. populations for individual endpoints
14    (lung and bladder cancer). Based on the past  experience of Morales et al. (2000) and modeling
15    results presented by NRC (2001), this effort was limited to exploring alternative forms for the
16    dose dependence of risks. Equation 5-5 shows EPA's baseline model, which is "linear Poisson"
17    with the form:
18
19                       h(x,t) = exp(ai + a2 x age + a3 x age2) x (l + b x dose)      (Equation 5-5)
20
21    In addition to the linear model, three other models were evaluated.  First, the quadratic form of
22    dose dependence:
23
24              h(x,t) = exp(ai + a2 x age + as x age2) x (l + bl x dose +b2 x dose2)   (Equation 5-6)
25
26    Next, two models in which the dose dependence was exponential, one linear and one quadratic:
27
28                    h(x,t) = exp(ai + a2 x age + as x age2) x Exp(bO + bi x dose)    (Equation 5-7)
29
30           h(x,t) = exp(ai + a2 x age + a3 x age2) x Exp(bO + bi x dose + b2 x dose2) (Equation 5-8)
31
32          The last model (Equation 5-8) was specifically recommended by SAB (2007) for
33    evaluation.  In the discussion that follows, these four models are referred to, respectively, as the
34    "linear"  (baseline) model (Equation 5-5), quadratic model (Equation 5-6), linear exponential
35    model (Equation 5-7), and quadratic exponential model (Equation 5-8).7
       "Absolute risk" models (models in which arsenic exposure was assumed to result in additive, rather than
      multiplicative, increments in risks) were found to fit the data much less well than the multiplicative forms shown in
      Equations 5-6 to 5-8 and are not discussed further.
                                                142       DRAFT—DO NOT CITE OR QUOTE

-------
 1          All four models were fit to lung cancer data from the Taiwanese population, using the
 2    baseline exposure parameter values and including the southwest Taiwanese reference population.
 3    Models were fit using the Non-Linear Estimation module of Statistica®. For males, the
 4    quadratic and quadratic exponential models curve sharply downward at high doses, whereas the
 5    linear exponential model curves sharply upward. Over the dose range from 0 to 0.05 mg/kg-day
 6    in males, which corresponds to an arsenic drinking water concentration range of 0 to 710 ug/L
 7    (which covers approximately 95% of the exposed population years at risk), predictions from the
 8    non-linear models are never more than 22% higher or 24% lower than the predictions from the
 9    linear (baseline) model.  As noted previously, these differences are relatively small compared to
10    the degree of statistical uncertainty in the estimates of the dose-response coefficients.
11          For females, two of the models (quadratic and quadratic exponential) predict lung cancer
12    risks for 60- to 65-year-olds that are very close to those predicted by the linear model.  The
13    linear exponential model, however, curves strongly upward at high doses. Over the dose range
14    from 0 to 0.03 mg/kg-day in females (corresponding  to 0 to 750 ug/L arsenic in drinking water,
15    about 95% of the exposed population years at risk), the cancer risks predicted by the non-linear
16    models are never more than 9% above or 37% below the risks predicted by the linear (baseline)
17    model.
18          These analyses indicate that, within the range of exposures covered by the
19    epidemiological data, the alternative model forms predict very similar risks (i.e., variations in
20    risk estimates across models are well within the estimated statistical uncertainty of the models).
21    The behavior to the various models at the extremes of the data (high and low exposures) depends
22    to a large extent on the model specification; models with non-linear dose specifications will
23    predict risks that increase more or less rapidly in the  extremes than the linear additive Poisson
24    regression, depending on the form of the dose term.  As discussed in Section 4.6.3, given the
25    limitations in data related to mode of action, there is  no compelling reason to prefer non-linear
26    models,  and the additive Poisson model is the simplest, best-fitting, and most parsimonious
27    model currently available for establishing a point of departure for establishing health criteria.
      5.3.8.7. Significance of Cancer Risks at Low Arsenic Exposures
28          Several recently published studies have called into question the strength and significance
29    of the exposure-response relationship for arsenic in the Taiwanese population studied by Chen et
30    al. (1988a, 1992) and Wu et al. (1989) that have been used by EPA for estimating cancer risk.
31    Based on "graphical  and regression analysis," Lamm et al. (2003)  found no significant dose-
32    response relationship for arsenic-related bladder cancer in the subset of the Taiwanese
33    population with median drinking water well concentrations less than 400 ug/L.  Kayajanian
34    (2003) found that combined male and female lung, bladder, and liver cancers were relatively
35    elevated at low arsenic exposures, then decreased to  minimums for villages with water arsenic
36    concentrations in the range between 42 and 60 ug/L,  and then again increased with increasing
                                                143       DRAFT—DO NOT CITE OR QUOTE

-------
 1    arsenic exposure.  In a more recent analysis, Lamm et al. (2006) found that (1) dummy variables
 2    related to "township" location were significant (along with arsenic well concentration) when all
 3    the townships were included in the analysis and (2) the dose-response parameter for arsenic
 4    exposure became insignificant for arsenic well concentrations less than 151 ug/L when only a
 5    subset of the data was included in the regression.
 6           The studies by Lamm et al. (2003, 2006) and Kayajanian (2003) have severe limitations.
 7    In evaluating the findings of these studies, it is important to recognize the complexity and
 8    limitations of the Taiwanese data set. Cancer mortality and person-years at risk observations are
 9    provided for a large  number (n =  559) of relatively small age- and village-stratified populations
10    (median person-years at risk ~ 340 for both males and females). Most population groups have
11    zero cancer deaths, and the data are very "noisy." Cancer mortality is strongly age-dependent,
12    and simultaneously evaluating the age- and dose-dependence of cancer mortality based on a data
13    set in which cancer deaths are "rare events" requires appropriately structured models. All of
14    these features of the data drove the selection of the Poisson regression  methods described in
15    Section 5, and the use of simpler models (linear regression, for example) can (and did) produce
16    misleading results.
17           With regard to the Lamm  et al. (2003) paper, it is likely that the use of linear regression
18    and the failure to correctly account for the age-dependency of bladder cancer risks combined to
19    make it impossible to detect a significant exposure-response relationship in villages with water
20    arsenic levels less than 400 ug/L.  U.S. EPA (2005d) evaluated this study and noted the
21    following weaknesses:
22       •   Classification of wells as artesian or shallow was based solely on arsenic concentration.
23
24       •   Age was not included as a variable in the regression analysis, despite the clear strong
25           dependence of cancer risks on age.
26
27       •   Previous studies have found little evidence for the presence of other potential carcinogens
28           in the sampled wells.
29
30           The major limitation of Kayanjaian's (2003) analysis of the Taiwanese data is that it
31    breaks the data into  strata that are too small to be used to calculate reliable mortality risks, and
32    that it is very sensitive to the specific way that the data are stratified. The observed trend in
33    cancer mortality versus arsenic dose would be very different if only few cancer deaths were
34    misclassified, or if the pattern of cancer deaths had been slightly different by  chance.  Lamm et
35    al.'s (2006) failure to find a significant exposure-response relationship in villages with arsenic
36    water concentrations below 151 ug/L can also be explained by (1) the use of linear regression
37    without age-adjustment; and (2) the omission of data from three of the six townships from the
38    regression.


                                                144      DRAFT—DO NOT CITE OR QUOTE

-------
1          Appendix F provides additional analyses supporting the significance and robustness of
2   the dose-response relationship for arsenic at low doses and in the defined subsets of the
3   population studied by Lamm et al. (2006).

    5.4. CANCER ASSESSMENT (INHALATION EXPOSURE)
4           An inhalation unit risk was developed for inorganic arsenic and posted on the IRIS
5   database in 1988. This document does not present a re-assessment of the cancer dose-response
6   estimation for inhalation exposure to inorganic arsenic.
                                              145      DRAFT—DO NOT CITE OR QUOTE

-------
      6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE-
                                               RESPONSE
      6.1. HUMAN HAZARD POTENTIAL
 1          Arsenic is readily absorbed from the GI tract, either from drinking water or food sources.
 2    Although dermal absorption is not significant compared to absorption from oral exposure, it
 3    may have contributed to the total arsenic exposures and health effects reported in many
 4    epidemiological studies in the literature. There appears, however, to be little if any dermal
 5    absorption (NRC, 1999) except at high occupational exposures (Hostynek et al., 1993).
 6    Inhalation is not being addressed in this document.
 7          After absorption, inorganic arsenic can undergo a complicated series of enzymatic and
 8    non-enzymatic reduction, enzymatic oxidative methylation, and conjugation reactions.  Although
 9    these reactions occur throughout the body, the rate at which they occur varies greatly from organ
10    to organ, with major metabolism occurring in the liver.  While there are two proposed pathways
11    (Figures 3-1 and 3-2)  for arsenic metabolism—with each pathway likely to occur depending on
12    exposure level and/or individual—the main urinary excretion products in humans are MMA and
13    DMA and the parent compound.  Arsenic metabolism (mainly methylation) varies greatly across
14    different species (Vahter, 1994, 1999a), which may explain why there has been no adult animal
15    model for the carcinogenic potential of arsenic. Although a few animal bioassays have been
16    conducted, they have  all been negative. Arsenic-induced cancers have been observed with
17    transplacental exposure in mice.  Transplacental exposure to arsenic in mice has found increases
18    in the development of lung, liver, reproductive, and adrenal tumors. Skin tumors in animals have
19    only been induced in transgenic models or in co-carcinogenesis studies.
20          Despite the lack of a good animal model for arsenic carcinogenesis, numerous
21    epidemiological studies have examined the carcinogenic potential of inorganic arsenic via oral
22    exposure. Although each of the investigations has its  own inherent strengths and weaknesses,
23    the combination of all the study results supports an association between oral exposure to
24    inorganic arsenic and  cancer including bladder, kidney, skin, lung, liver, and prostate. Because
25    the association between arsenic and these cancers has  been found in different populations, it is
26    unlikely that any single attribute (e.g., nutritional habits) associated with a single population is
27    responsible for the increased cancer rates. However, genetic polymorphisms have been found to
28    be an important factor in the methylation of arsenic. Evidence suggests that people who have a
29    greater capacity to methylate arsenic completely to DMA are at a lower risk for developing
30    arsenic-related cancers. Nutritional and personal habits including smoking also affect the
31    methylation rate. Therefore, genetic, nutritional, and lifestyle factors contribute to the inter-
32    individual variations.
                                                146       DRAFT—DO NOT CITE OR QUOTE

-------
 1          Although dose-response relationships have been observed for the majority of cancers
 2    noted in areas with high levels of arsenic in their drinking water, results for low-level arsenic
 3    epidemiologic investigations (primarily from the United States and Europe) have been equivocal
 4    in the relationship  between these cancers and arsenic exposure.  This could be due to the fact that
 5    none of the studies accounted for arsenic exposure through food sources, which would be a
 6    significant source as the levels in the drinking water decreased (Uchino et al., 2006; Kile et al.,
 7    2007). Because cancer has a long latency period, misclassification also occurs due to lack of
 8    data on disease-relevant exposures (Cantor and Lubin, 2007), which would be more significant
 9    in studies examining lower exposures. Therefore, studies with low levels of exposure that are
10    ecological in nature (no individual exposure) are more prone to exposure misclassification,
11    which means they  are biased toward the null hypothesis.  Despite all these numerous limitations
12    in low-level exposure studies, positive associations have been observed for cancers of the
13    prostate  (Hinwood et al., 1999; Lewis et al., 1999), skin (Hinwood et al., 1999; Karagas et al.,
14    2001; Beane-Freeman et al., 2004; Knobeloch et al., 2006), and bladder (Kurttio et al., 1999;
15    Steinmaus et al., 2003; Karagas et al., 2004).  In most cases, however, there is no dose-response
16    with increases observed at the highest concentrations only and in many cases significant results
17    occurred in smokers only.
18          Based upon current EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a),
19    inorganic arsenic is determined to be "carcinogenic to humans" due to convincing
20    epidemiological evidence of a causal relationship between oral exposure of humans to inorganic
21    arsenic and cancer.
22          The available evidence is inadequate to establish a MOA by which arsenic induces
23    tumors.  The genotoxicity data for arsenic are equivocal.  Chromosomal aberrations have been
24    observed in humans and animals exposed to arsenic, but arsenic has been generally negative in
25    bacterial mutagenicity tests and has only been observed to be a weak mutagen at the hprt locus in
26    Chinese hamster V79 cells at toxic concentrations (Li and Rossman, 1989a).  In addition, even
27    though it appears genotoxic in animal models, it does not generally induce tumors in animal
28    models.  Arsenic does not appear to cause point mutations in standard assays, but instead causes
29    large deletion mutations (Rossman, 1998).  These large deletions can cause lethality when
30    closely linked to essential genes. Therefore, the mutations are not easily observed in standard
31    bacterial and mammalian cell mutation assays.  However, even in transgenic cell lines, which
32    were tolerant of large deletions, arsenic was still only weakly mutagenic at doses causing overt
33    cytotoxicity (Rossman, 2003).  It has been suggested that arsenic acts as an aneugen (affects the
34    number of chromosomes) at low doses, but as a clastogen (causes chromosomal breaks) at high
35    doses (Rossman, 2003). However, arsenic has also been demonstrated to affect other processes
36    possibly involved with carcinogenesis, including aberrant gene/protein expression, ROS, DNA
37    repair inhibition, signal transduction, and cancer promotion. Therefore, it is likely that arsenic

                                                147       DRAFT—DO NOT CITE OR QUOTE

-------
 1    acts via multiple MO As, which would explain the number of different internal cancers associated
 2    with arsenic.

      6.2. DOSE-RESPONSE
 3          Only the oral cancer assessment is addressed in this document. Lung and bladder cancer
 4    mortality in the Taiwanese population were selected as endpoints in the dose-response modeling
 5    because they are the internal cancers with the most consistent results and are best characterized
 6    in epidemiology studies of arsenic exposure (NRC, 1999, 2001; SAB 2000, 2007). Dose-
 7    response models were estimated for the Taiwanese population using additive Poisson regression
 8    with linear dose terms and quadratic age terms.
 9          ED0i values were derived from the MLE dose-response parameter estimates.  LED01
10    estimates were derived from the 95% upper confidence limits on the dose-response parameters,
11    as described in Appendix E.  The analysis was done in two phases. The first phase consisted of
12    the derivation and fitting of dose-response models using the Taiwanese epidemiology data from
13    Chen et al. (1988a,  1992) and Wu et al. (1989).  The outputs of this phase of the analysis were
14    arsenic dose-response coefficients that described the relationship between estimated arsenic
15    intake in the Taiwanese population and proportional increases in age-specific lung and bladder
16    cancer mortality risk.  Lifetime cancer incidence in U.S. populations was then estimated by using
17    a modified version of the "BEIRIV" relative risk model.  A key assumption underlying this
18    model is that the risk of arsenic-related cancer is a constant multiplicative function of the
19    "background" age profile of cancer risks in the target U.S. population. Estimates of arsenic-
20    related cancer risks in a (hypothetical) U.S. population exposed to arsenic at varying levels in
21    drinking water were then derived.
22          The oral CSFs for lung and bladder cancers in U.S. males  and females were derived using
23    the following assumptions: nonwater arsenic intake for the reference and exposed populations
24    was 10 ug/day; drinking water consumption was 3.5 and 2.0 L/day in Taiwanese men and
25    women, respectively; 50 kg was the average Taiwanese body weight; and a 70 kg individual in
26    the United States consumes 2.0 L/day of water (Section 5.3.5).  The oral CSF is dependent on
27    assumptions related to the volume of contaminated water consumed over the course of a day and
28    the amount of arsenic consumed through the diet. Changes in these assumptions would result in
29    different cancer potency estimates (as discussed in Section 5.3.8.3), and corresponding changes
30    in the  other risk criteria (drinking water unit risk, drinking water concentration associated with
31    lOLEDoi lifetime cancer risk, etc.). Sensitivity analyses were performed to test the effects of
32    differences in drinking water intake assumptions, nonwater arsenic intake assumptions, using
33    median well water values compared to minimum and maximum values, and including different
34    Taiwanese reference populations on the estimates (Section 5.3.8.3). Based on the results of the
35    sensitivity  analyses, the risk model results, with the exception of female bladder cancer, appear

                                               148       DRAFT—DO NOT CITE OR QUOTE

-------
 1    to be relatively stable and react predictably to reasonable changes in exposure assumptions.
 2    Female bladder cancer estimates were particularly sensitive to variations in nonwater arsenic
 3    intake.
 4          Estimated cancer potency factors for lifetime U.S. male lung and bladder cancer
 5    incidence were 6.7 and 11.2 per mg/kg-day, respectively.  The corresponding values for females
 6    were 16.6 and 10.5 per mg/kg-day (Table 5-3).  Cancer potency for combined lung and bladder
 7    cancer risks were estimated for males and females, as described in Section 5.3.8.1. The
 8    estimated cancer potency factors for combined (lung plus bladder) cancer incidence were 16.9
 9    and 25.7 per mg/kg-day, respectively. The potency factor estimate for women (25.7 per
10    mg/kg-day) was identified as the recommended point of departure for derivation of health
11    criteria, with women being the more sensitive population.
12          The cancer potency estimates derived in this analysis are not directly comparable to those
13    estimated in EPA's 1988 assessment (U.S. EPA, 1988b). That analysis derived a much lower
14    potency factor estimate (1.0-2.0 per mg/kg-day) based on an analysis of skin cancer incidence in
15    the Taiwanese population studied by Tseng et al. (1968; Tseng,  1977).  Since the exposure-
16    response data on internal cancers has become available, all the subsequent assessments
17    (including this one) have been based on internal (bladder and/or lung) cancer (see Section 5.3.1).
18    The difference in endpoints (skin versus internal cancers) is the main reason for the relatively
19    large difference in estimated cancer potency in the more recent assessment compared to the 1988
20    assessment.
21          As discussed  in Section 5.3.8.2, the lifetime risk estimates for male and female lung and
22    bladder cancer  calculated in this assessment are generally consistent with the risk estimates from
23    previous analyses that used the internal cancers (NRC, 2001). The bulk of the difference
24    between the cancer potency estimates in this assessment and those from previous analyses can be
25    explained by differences in dose-response models, changes in the  assumptions related to the
26    relative drinking water consumption by women in Taiwan and the United States, and the use of
27    more recent data on U.S. population mortality and cancer incidence in the BEIRIV relative risk
28    model.
29          The Supplemental Guidance for Assessing Susceptibility From Early-Life Exposure to
30    Carcinogens (U.S. EPA, 2005b) indicates that age-dependent adjustment factors should be
31    applied to the CSF and combined with early-life exposure estimates when estimating cancer
32    risks from exposures to carcinogens with a mutagenic MO A.  As discussed in Section 4.6.3,
33    insufficient data are available to adequately demonstrate a mutagenic mode of action for
34    inorganic arsenic.  Therefore, the application of age-dependent adjustment factors is not
35    recommended.
36          The overall level of confidence in the data is high. The data used in the dose-response
37    assessment come from human epidemiology rather than animal bioassays.  The Taiwanese

                                                149      DRAFT—DO NOT CITE OR QUOTE

-------
 1    studies characterize the cancer risks of an extremely large, well-characterized population with a
 2    wide range of exposure concentrations. Reliability and accuracy of mortality records,
 3    verification of endpoints with histological examinations, several decades of exposure to arsenic
 4    in drinking water to detect internal cancer outcomes, apparent similarities in lifestyle habits
 5    (similar urbanization in the endemic area versus the rest of southwestern Taiwan) between
 6    exposed and reference populations, and the residential stability of the population (i.e., little
 7    migration or emigration) are high.  The data demonstrate a statistically significant dose-related
 8    effect in humans, across the entire range of exposures (i.e., 10-934 ppb median levels) evaluated.
 9    The currently used BEIRIV model is an improvement over previous models because it contains
10    a quadratic age model, an additive linear dose term, and a reference population, and adjusts for
11    differences between the exposed and target (i.e., U.S.) populations.
12           Despite all their strengths, the Chen et al. (1988a, 1992) and Wu et al. (1989) studies are
13    "ecological"; data on individual exposure (which are a function of both water consumption rates
14    and concentrations) are not available.  In addition, smoking information was  not provided in the
15    critical studies (however, it appears comparable—40% vs. 32% in endemic area vs. the rest of
16    Taiwan according to Chen et al., 1985).  Lacking this information introduces an unquantifiable
17    degree of uncertainty into the risk estimates.  In EP A's judgment, these factors are equally likely
18    to have resulted in overestimates or underestimates of risks.
      6.2.1. Choice of Models
19           As  discussed in Section 5.3.1, the Taiwanese data have been used as the basis for
20    quantitative risk assessment by a number of investigators.  In this current analysis, EPA is
21    building on the experience of previous efforts by itself and others, and has incorporated
22    comments  and recommendations by NRC (2001) and SAB (SAB, 2007) in the selection of
23    statistical methods for use in the risk assessment. As discussed in Section 5.3.7.1, the current
24    assessment employs a Poisson regression model with additive linear dose terms and quadratic
25    age terms for dose-response model fitting in the Taiwanese population. This model was found to
26    be the simplest, best-fitting model among a number of alternatives tested. Sensitivity analyses of
27    other models (quadratic, exponential linear, and exponential quadratic dose transformation) were
28    also conducted (see Section 5.3.8.4 for further details).
29           To  extrapolate arsenic-related cancer  risks to the U.S. population, the current assessment
30    employs a  variant of the "BEIR IV" relative risk model (Section 5.3.7.3). This model takes as its
31    inputs the dose-response coefficients from the Poisson regressions and "background" cancer
32    incidence and population mortality data from the target (U.S.) population. Population mortality
33    data for the year 2000 (NCHS, 2000) and background lung and bladder cancer incidence for
34    2000-2003 (NCI, 2006) were used as inputs to the BEIR IV model.
                                                150       DRAFT—DO NOT CITE OR QUOTE

-------
      6.2.2. Dose Metric
 1          Inorganic arsenic is metabolized in vivo, with some of the known metabolites being more
 2    toxic than the parent compound. However, it is not known whether it is a metabolite, the parent
 3    compound, or a combination of the two that is responsible for the observed carcinogenic
 4    potential. An increase in MMA or decreased DMA in the urine has been associated with an
 5    increase in disease risk (Yu et al., 2000; Chen et al., 2005a; Steinmaus et al., 2005; Valenzuela et
 6    al., 2005; Ahsan et al., 2007; Huang et al., 2007b; McCarthy et al., 2007a); therefore, the actual
 7    carcinogenic moiety may not be proportional to administered exposure  and use of administered
 8    exposure may produce a bias in the model. However, the exposure assessment for the model is
 9    ecological in nature and produces its own inherent bias. Detailed arsenic speciation data are not
10    available for the Taiwanese population used  in the risk assessment. Therefore, estimated total
11    daily arsenic dose (water + other dietary) has been used as the dose metric in the risk assessment.
12    Arsenic dose is estimated based on well water concentration data, and it is assumed that the
13    arsenic concentrations have been constant over the period of exposure.  Since there are no data
14    related to the temporal variability in the well water concentrations, this  introduces uncertainty
15    into the dose estimates for the 43 villages. Sensitivity analyses were  conducted to investigate the
16    impact of using alternative exposure indices, as discussed in Section 5.3.8.3.
      6.2.3. Human Population Variability
17          Although the extent of inter-individual variability in arsenic metabolism has not been
18    adequately characterized, genetic polymorphism, nutritional status, and personal habits (e.g.,
19    smoking) have all been associated with differences in arsenic methylation. Data exploring
20    whether there is a differential sensitivity to arsenic carcinogenicity across life stages is limited.
21    Data by Waalkes et al. (2003, 2004a) indicate that transplacental exposure in mice is a sensitive
22    stage for carcinogenic potential. These are the only studies in which  inorganic arsenic exposure
23    has been associated with cancer in rodents. Lung, liver, reproductive, and adrenal tumors were
24    associated with arsenic administration during gestation (10 days only).  A single epidemiological
25    study by Smith et al. (2006) examined lung cancer rates (and other respiratory diseases) in
26    cohorts exposed during childhood and cohorts likely exposed in utero to arsenic concentrations
27    of 860 ppb that subsequently dropped to 100 ppb.  Results demonstrated that exposure during
28    either period of development caused increased risk of lung cancer in females aged 40 to 49 born
29    between 1950 and 1957 and in males aged 30 to 49 born between 1950  and  1970. However, the
30    risks associated with early childhood exposures and/or in utero exposures were not compared to
31    risks from exposures during adulthood. Thus, the available data do not allow for a quantitative
32    assessment of the relative sensitivity to arsenic exposures between the Taiwanese population
33    used in the dose-response assessment and U.S. populations exposed to arsenic in drinking water.
34          SAB (2007) acknowledged "the possible issue of compromised  nutrition among
35    segments of the exposed population" in the Taiwanese study population, along with the lack of
                                                151       DRAFT—DO NOT CITE OR QUOTE

-------
1    data related to smoking history. However, data are not available that would allow quantitative
2    evaluation of these factors. Therefore, this risk assessment assumes that the observed
3    carcinogenic potency in the Taiwanese population, with suitable corrections for differences in
4    drinking water intake and background cancer incidence, is an appropriate predictor of the
5    potential for human cancer risk in the U.S. population.
                                               152       DRAFT—DO NOT CITE OR QUOTE

-------
                                         7. REFERENCES

Achanzar, WE; Brambila, EM; Diwan, BA; et al. (2002) Inorganic arsenite induced malignant transformation of
human prostate epithelial cells. JNatl Cancer Inst 94:1888-1891.

Ackerman, AH; Creed, PA; Parks, AN; et al. (2005) Comparison of a chemical and enzymatic extraction of arsenic
from rice and an assessment of the arsenic absorption from contaminated water by cooked rice. Environ Sci Technol
39:5241-5246.

Adair, BM; Moore, T; Conklin, SD; et al. (2007) Tissue distribution and urinary excretion of dimethylated arsenic
and its metabolites in dimethyl arsenic acid- or arsenate-treated rats. Toxicol Appl Pharmacol 222:235-242.

Adams, MA; Bolger, PM; Gunderson, EL. (1994) Dietary intake and hazards of arsenic. In: Chappell, WR;
Abernathy, CO; and Cothern, CR; eds. Arsenic: exposure and health. Northwood, UK: Sci Tech Lett; pp. 41-49.

Ahsan, H; Chen, Y; Parvez, F; et al. (2006) Arsenic exposure from drinking water and risk of premalignant skin
lesions in Bangladesh: baseline results from the health effects of arsenic longitudinal study. Am J Epidemiol
163(12):1138-1148.

Ahsan, H; Chen, Y; Kibriya, MG; et al. (2007) Arsenic metabolism, genetic susceptibility, and risk of premalignant
skin lesions in Bangladesh. Cancer Epidemiol Biomarkers Prev 16(6): 1270-1278.

An, Y; Kato, K; Nakano, M; et al. (2005) Specific induction of oxidative stress in terminal bronchiolar Clara cells
during dimethylarsenic-induced lung tumor promoting process in mice. Cancer Lett 230:57-64.

Andersen, O. (1983) Effects of coal combustion products and metal compounds on sister chromatid exchange (SCE)
in a macrophage like cell line. Environ Health Perspect 47:329-253.

Andrew, AS; Karagas, MR; Hamilton, JW. (2003) Decreased DNA repair gene expression among individuals
exposed to arsenic in United States drinking water. Int J Cancer 104:263-268.

Andrew, AS; Burgess, JL; Meza, MM; et al. (2006) Arsenic exposure is associated with decreased DNA repair in
vitro and in individuals exposed to drinking water arsenic. Environ Health Perspect 114:1193-1198.

Andrewes, P; Kitchin, KT; Wallace, K. (2003) Dimethylarsine and trimethylarsine are potent genotoxins in vitro.
ChemRes Toxicol 16:994-1003.

Anundi, I; Hogberg, J; Vahter, M. (1982) GSH release in bile as influenced by arsenite. FEES Lett 145:285-288.

Aposhian, HV. (1997) Enzymatic methylation of arsenic species and other new approaches to arsenic toxicity. Annu
Rev Pharmacol Toxicol 37:397-419.

Aposhian, HV; Aposhian, MM. (2006) Arsenic  toxicology: five questions. ChemRes Toxicol 19(1): 1-15.

Argos, M; Kibriya, MG; Parvez, F; et al. (2006) Gene expression profiles in peripheral lymphocytes by arsenic
exposure and skin lesion status in a Bangladeshi population. Cancer Epidemiol Biomarkers Prev 15:1367-1375.

Arkusz, J; Stariczyk, M; Lewiriska, D; St^pnik, M. (2005) Modulation of murine peritoneal macrophage function by
chronic exposure to arsenate in drinking water. Immunopharmacol Immunotoxic 27:315-330.

Askar, N; Cirpan, T; Toprak, E; et al.  (2006) Arsenic trioxide exposure to ovarian carcinoma cells leads to decreased
level of topoisomerase II and cytotoxicity. Int J  Gynecol Cancer 16:1552-1556.

ATSDR (Agency for Toxic Substances and Disease Registry).  (1993) Toxicological profile for arsenic (update).
Public Health Service, U.S. Department of Health and Human  Services, Atlanta, GA.
                                                  153        DRAFT—DO NOT CITE OR QUOTE

-------
ATSDR (Agency for Toxic Substances and Disease Registry). (2000) Toxicological profile for arsenic (update).
Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.

ATSDR (Agency for Toxic Substances and Disease Registry). (2007) Toxicological profile for arsenic (update).
Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.

Avani, G; Rao, MG. (2007) Genotoxic effects in human lymphocytes exposed to arsenic and vitamin A. Toxicol in
Vitro 21:626-631.

Baastrup, R; Sorensen, M; Balstrom, T;  et al. (2008) Arsenic in drinking-water and risk for cancer in Denmark.
Environ Health Prospect 116:231-237.

Banerjee, M; Sarkar, J; Das, JK; et al. (2007) Polymorphism in the ERCC2 codon 751 is associated with arsenic-
induced premalignant hyperkeratosis and significant chromosome aberrations. Carcinogenesis 28(3):672-676.

Barchowsky, A; Dudek, EJ; Treadwell, MD; Wetterhahn, KE. (1996) Arsenic induces oxidant stress and NF-KB
activation in cultured aortic endothelial cells. Free Radic Biol Med 21:783-790.

Barchowsky, A; Roussel, RR; Klei, LR; et al. (1999a) Low levels of arsenic trioxide stimulate proliferative signals
in primary vascular cells without activating stress effector pathways. Toxicol Appl Pharmacol 159:65-75.

Barchowsky, A; Klei, LR; Dudek, EJ; et al. (1999b)  Stimulation of reaction oxygen, but not reactive nitrogen
species,  in vascular endothelial cells exposed to low  levels of arsenite. Free Radic Biol Med 27:1405-1412.

Barrett, JC; Lamb, PW; Wang, TC; Lee, TC (1989) Mechanism of arsenic-induced cell transformation. Biol Trace
Element Res 21:421^29.

Bashir, S; Sharma, Y; Irshad, M; et al. (2006a). Arsenic induced apoptosis in rat liver following repeated 60 days
exposure. Toxicol 217:63-70.

Bashir, S; Sharma, Y; Irshad, M; et al. (2006b) Arsenic-induced cell death in liver and brain of experimental rats.
Basic ClinPharmacol Toxicol 98:38-43.

Basu, A; Mahata, J; Roy, AK; et al. (2002) Enhanced frequency of micronuclei in individuals exposed to arsenic
through  drinking water in West Bengal,  India. Mutat Res 516:29^0.

Basu, A; Som, A; Ghoshal, S; et al. (2005) Assessment of DNA damage in peripheral blood lymphocytes of
individuals susceptible to arsenic induced toxicity in West Bengal, India. Toxicol Lett 159:100-112.

Bates, MN; Smith, AH; Cantor, KP. (1995) Case-control study of bladder cancer and arsenic in drinking water. Am J
Epidemiol 141:523-530.

Bates, MN; Rey, OA; Biggs, ML;  et al.  (2004) Case-control study of bladder cancer and exposure to arsenic in
Argentina. Am J Epidemiol 159(4):381-389.

Beane-Freeman, LE; Dennis, LK; Lynch, CF; et al. (2004) Toenail arsenic content and cutaneous melanoma in Iowa.
Am J Epidemiol 160(7):679-687.

Beckman, G; Beckman, L; Nordenson, I. (1977) Chromosome aberrations in workers exposed to arsenic. Environ
Health Perspect 19:145-146.

Benbrahim-Tallaa, L;  Waterland, RA; Styblo, M; et al. (2005) Molecular events associated with arsenic-induced
malignant transformation of human prostatic epithelial cells: aberrant genomic DNA methylation and K-ras
oncogene activation. Toxicol Appl Pharmacol 206:288-298.

Bodwell, JE; Kingsley, LA; Hamilton, JW. (2004) Arsenic at very low concentrations alters glucocorticoid receptor
(GR)-mediated gene activation but not GR-mediated gene repression: complex dose-response effects are closely
                                                  154         DRAFT—DO NOT CITE OR QUOTE

-------
correlated with levels of activated GR and require a functional GR DNA binding domain. Chem Res Toxicol
17:1064-1076.

Bodwell, JE; Gosse, JA; Nomikos, AP; Hamilton, JW. (2006) Arsenic disruption of steroid receptor gene activation:
complex dose-response effects are snared by several steroid receptors. Chem Res Toxicol 19(12): 1619-1629.

Bolton, JL; Pisha, E; Zhang, F; Qiu, S. (1998) Role of quinoids in estrogen carcinogenesis. Chem Res Toxicol
11:1113-1127.

Bolton, JL; Trash, MA; Penning, TM; et al. (2000) Role of quinones in toxicology. Chem Res Toxicol 13:135-160.

Bredfeldt, TG; Kopplin, MJ; Gandolfi, AJ. (2004) Effects of arsenite onUROtsa cells: low-level arsenite causes
accumulation of ubiquitinated proteins that is enhanced by reduction in cellular glutathione levels. Toxicol Appl
Pharmacol 198:412^118.

Bredfeldt, TG; Jagadish, B; Eblin, KE; et al. (2006) Monomethylarsonous acid induces transformation of human
bladder cells.  Toxicol Appl Pharmacol 216:69-79.

Brenton, CV; Houseman, EA; Kile, ML; et al. (2006) Gender-specific protective effects of hemoglobin on arsenic-
induced skin lesions. Cancer Epidemiol Biomarkers Prev 15(5):902-907.

Brenton, CV; Kile, ML; Catalano, PJ; et al. (2007a) GSTM1 and APE1 genotypes affect arsenic-induced oxidative
stress: a repeated measures study. Environ Health 6:39. Available online at http://www.ehjournal.net/content/6/l/39.

Brenton, CV; Zhou, W; Kile, ML; et al. (2007b) Susceptibility to arsenic-induced skin lesions from polymorphisms
in base pair excision repair genes. Carcinogenesis 28(7): 1520-1525.

Buchet,  JP; Lauwerys, R; Roels, H. (1981) Comparison of the urinary  excretion of arsenic metabolites after a single
dose of sodium arsenite, monomethylarsonate or dimethylarsinate in man. Int Arch Occup Environ Health 48:71-79.

Buchet,  JP; Geubel, A; Pauwels, S; et al. (1984) The influence of liver disease on the methylation of arsenite in
humans. Arch Toxicol 55:151-154.

Buchet,  JP; Lauwerys, R. (1985) Study of inorganic arsenic methylation by rat liver in vitro: relevance for the
interpretation of observations in man. Arch Toxicol 57:125-129.

Bunderson, M; Pereira, F; Schneider, MC; et al. (2006) Manganese enhances peroxynitrite and leukotriene E4
formation in bovine aortic endothelial cells exposed to arsenic. Cardiovasc Toxicol 6:15-23.

Burgdorf, W; Kurvink, K; Cervenka, J. (1977) Elevated sister chromatid exchange rate in lymphocytes of subjects
treated with arsenic. Hum Genet 36(l):69-72.

Bums, FJ; Uddin, AN; Wu, F; et al.  (2004) Arsenic-induced enhancement of ultraviolet radiation carcinogenesis in
mouse skin: a dose-response study. Environ Health Perspect 112(5):599-603.

Bustamante, J; Nutt, L; Orrenius, S;  Gogvadze,  V. (2005) Arsenic stimulates release of cytochrome c from isolated
mitochondria via induction of mitochondrial permeability transition. Toxicol Appl Pharmacol 207(2 Suppl): S110-
S116.

Cantoni, O; Hussain, S; Guidarelli, A; et al. (1994) Cross-resistance to heavy metals in hydrogen peroxide-resistant
CHO cell variants. Mutat Res 324:1-6.

Cantor, KP; Lubin, JH. (2007) Arsenic, internal cancers, and issues in inference from studies of low-level exposures
in human populations. Toxicol Appl Pharmacol 222:252-257.

Chanda, S; Dasgupta, UB; Guha Mazumder, D; et al. (2006) DNA hypermethylation of promoter of gene p53 and
p!6 in arsenic-exposed people with and without malignancy. Toxicol Sci 89:431-437.
                                                  155         DRAFT—DO NOT CITE  OR QUOTE

-------
Chao, J-I; Hsu, SH; Tsou, T-C. (2006a) Depletion of securin increases arsenite-induced chromosome instability and
apoptosis via a p53-independent pathway. Toxicol Sci 90:73-86.

Chao, HR; Tsou, TC; Li, LA; et al. (2006b) Arsenic inhibits induction of cytochrome P450 1A1 by 2,3,7,8-
tetrachlorodibenzo-p-dioxin in human hepatoma cells. J Hazard Mater 137:716-722.

Charbonneau, SM; Tarn, GKH; Bryce, F; et al. (1979) Metabolism of orally administered inorganic arsenic in the
dog. Toxicol Lett 3:107-114.

Charbonneau, SM; Hollins, JG; Tarn, GKH; et al. (1980) Whole-body retention, excretion, and metabolism of
[  As]arsenic acid in the hamster. Toxicol Lett 5:175-182.

Chen, CJ; Wang, CJ. (1990) Ecological correlation between arsenic level in well water and age-adjusted mortality
from malignant neoplasms.  Cancer Res 50:5470-5474.

Chen, KP; Wu, HY; Wu, TC. (1962) Epidemiologic studies onblackfoot disease in Taiwan. 3. Physicochemical
characteristics of drinking water in endemic blackfoot disease areas. In: Memoirs, College of Medicine, National
Taiwan University, Vol. 8, pp. 115-129. Taipei: National Taiwan University College of Medicine.

Chen, C-J; Chuang, Y-C; Lin, TM; Wu, HY. (1985) Malignant neoplasms among residents of a blackfoot disease-
endemic area in Taiwan: high-arsenic artesian well water and cancers. Cancer Res 45:5895-5899.

Chen, C-J; Chuang, Y-C; You, S-L; et al. (1986) A retrospective study on malignant neoplasms of bladder, lung, and
liver in blackfoot disease endemic area in Taiwan. Br J Cancer 53:399-405.

Chen, C-J; Kuo, T-L; Wu, M-M. (1988a) Arsenic and cancers. Lancet: letter to the editor. February 20, 1988.

Chen, C-J; Wu, M-M; Lee,  S-S; et al. (1988b) Atherogenicity and carcinogenicity of high-arsenic artesian well
water. Arteriosclerosis 8:452^60.

Chen, C-J; Chen, C-W; Wu, M-M; Kuo, T-L. (1992) Cancer potential in liver, lung, bladder, and kidney due to
ingested inorganic arsenic in drinking water. Br J Cancer 66(5):888-892.

Chen, Y; Megosh, LC; Gilmour,  SK; et al. (2000a) K6/ODC transgenic mice as a sensitive model for carcinogenic
information. Toxicol Lett 116:27-35.

Chen, Y; Liu, X; Pisha, E; et al. (2000b) A metabolite of equine estrogens, 4-hydroxyequilenin, induces DNA
damage and apoptosis in breast cancer cell lines. ChemRes Toxicol 13:342-350.

Chen, H; Liu, J; Merrick, BA; Waalkes, MP. (2001) Genetic events associated with arsenic-induced malignant
transformation: applications of cDNA microarray technology. Mol Carcinog 30:79-87.

Chen, CL; Hsu LI; Chiou HY; et al. (2004a) Ingested arsenic, cigarette smoking, and lung cancer risk:  a follow-up
study in arseniasis-endemic areas in Taiwan. JAMA 292:2984-90.

Chen, H; Li, S; Liu, J; et al. (2004b) Chronic inorganic arsenic exposure induces hepatic global  and individual gene
hypomethylation: implications for arsenic hepatocarcinogenesis. Carcinogenesis 25(9): 1779-1786.

Chen, Y-C; Su, H-JJ; Guo, Y-LL; et al. (2005a) Interaction between environmental tobacco smoke and arsenic
methylation ability on the risk of bladder cancer. Cancer Causes Control 16:75-81.

Chen, P-H; Lan, C-CE; Chiou, M-H; et al. (2005b) Effects of arsenic  and UVB on normal human cultured
keratinocytes: impact on apoptosis and implication on photocarcinogenesis. Chem Res Toxicol 18:139-144.
                                                  156        DRAFT—DO NOT CITE OR QUOTE

-------
Chen, D; Chan, R; Waxman, S; Jing, Y. (2006) Buthionine sulfoximine enhancement of arsenic trioxide-induced
apoptosis in leukemia and lymphoma cells is mediated via activation of c-Jun NH2-terminal kinase and up-
regulation of death receptors. Cancer Res 66:11416-11423.

Chen, Y; Hall, M; Graziano, JH; et al. (2007) A prospective study of blood selenium levels and the risk of arsenic-
related premalignant skin lesions. Cancer Epidemiol Biomarkers Prev 16(2):207-213.

Cheng, HY; Li, P; David, M; et al. (2004) Arsenic inhibition of the JAK-STAT pathway. Oncogene 23:3603-3612.

Cheng, Y; Chang, LW; Tsou, TC. (2006) Mitogen-activated protein kinases mediate arsenic-induced down-
regulation of survivin in human lung adenocarcinoma cells. Arch Toxicol 80:310-318.

Cheng, TJ; Wang, YJ; Kao, WW; et al. (2007) Protection against arsenic trioxide-induced autophagic cell death in
U118 human glioma cells by use of lipoic acid. Food Chem Toxicol 45:1027-1038.

Chiang, HS; Hong, CL; Guo, H-R; et al. (1988) Comparative study on the high prevalence of bladder cancer in the
blackfoot disease endemic area in Taiwan. J Formosan Med Assoc 83(11): 1074-1080.

Chien, C-W; Chiang, M-C; Ho, I-C; Lee, T-C. (2004) Association of chromosomal alterations with arsenite-induced
tumorigenicity of human HaCaT keratinocytes in nude mice. Environ Health Perspect 112(17): 1704-1710.

Chiou, HY; Hsueh, YM; Liaw, KF; et al. (1995) Incidence of internal cancers and ingested inorganic arsenic: a
seven-year follow-up study in Taiwan. Cancer Res 55:1296-1300.

Chiou, HY; Hsueh, YM; Hsieh, LL; et al. (1997) Arsenic methylation capacity, body retention, and null genotypes
of glutathione S-transferase Ml and Tl among current arsenic-exposed residents in Taiwan. Mutat Res 386:197-
207.

Chiou, H-Y; Chiou, S-T; Hsu, Y-H; et al. (2001) Incidence of transitional cell carcinoma and arsenic in drinking
water: a follow-up study of 8,102 residents in an arseniasis-endemic area in northeastern Taiwan. Am J Epidemiol
153:411-418.

Chiu, HF; Ho, S-C; Wang, L-Y; et al. (2004) Does arsenic exposure increase the risk for liver cancer? J Toxicol
Environ Health A 67:1491-1500.

Chou, W-C; Hawkins, AL; Barrett, JF; et al.  (2001) Arsenic inhibition of telomerase transcription leads to genetic
instability. J Clinlnves 108:1541-1547.

Chou, WC; Chen, HY; Yu, SL; et al. (2005) Arsenic suppresses gene expression in promyelocytic leukemia cells
partly through Spl oxidation. Blood  106:304-310.

Chowdhury, UK; Rahman, MM; Mondal, BK; et al. (2001) Groundwater arsenic contamination and human suffering
in West Bengal - India and Bangladesh. Environ Sci 8(5):393^U5.

Chowdhury, UK; Rahman, MM; Sengupta, MK; et al. (2003) Pattern of excretion of arsenic compounds [arsenite,
arsenate, MMA(V), DMA (V)] in urine of children compared to adults from an arsenic exposed area in Bangladesh.
J Environ Sci Health (Part A) A38(l):87-113.

Christian, WJ; Hopenhayn, C; Centeno,  JA; Todorov, T. (2006) Distribution of urinary selenium and arsenic among
pregnant women exposed to arsenic in drinking water. Environ Res 100:115-122.

Chung, JS; Kalman, DA; Moore, LE; et al. (2002) Family correlations of arsenic methylation patterns in children
and parents exposed to high concentrations of arsenic in drinking water. Environ Health Perspect 110:729-733.

Clewell, HJ; Thomas, RS; Gentry, PR; et al. (2007) Research toward the development of a biologically based dose
response assessment for inorganic arsenic carcinogenicity: a progress report.  Toxicol Appl Pharmacol 222:388-398.


                                                 157        DRAFT—DO NOT CITE OR QUOTE

-------
Cohen, SM; Arnold, LL; Uzvolgyi, E; et al. (2002) Possible role of dimethylarsinous acid in dimethylarsinic acid-
induced urothelial toxicity and regeneration in the rat. Chem Res Toxicol 15(9): 1150-1157.

Cohen, SM; Arnold, LL; Eldan, M; et al. (2006) Methylated arsenicals: the implications of metabolism and
carcinogenicity studies in rodents to human risk assessment. Crit Rev Toxicol 36:99-133.

Cohen, SM; Ohnishi, T; Arnold, LL; Lee, XC. (2007) Arsenic-induced bladder cancer in an animal model. Toxicol
Appl Pharmacol 222:258-263.

Colognato R; Coppede, F; Ponti, J; et al. (2007) Genotoxicity induced by arsenic compounds in peripheral human
lymphocytes analysed by cytokinesis-block micronucleus assay. Mutagenesis 22:255-261.

Concha, G; Vogler, G; Lezeano, D; et al. (1998a) Exposure to inorganic arsenic metabolites during early human
development. Toxicol Sci 44:185-190.

Concha, G; Nermell, B; Vahter, M. (1998b) Metabolism of inorganic arsenic in children with chronic high arsenic
exposure in northern Argentina. Environ Health Perspect 106:355-359.

Concha, G; Vogler, G; Nermell, B; Vahter, M. (1998c) Low-level arsenic excretion in breast milk of native Andean
women exposed to high levels of arsenic in the drinking water. Int Arch Occup Environ Health 71:42-46.

Cross, JD; Dale, IM; Leslie, ACD; Smith, H. (1979) Industrial exposure to arsenic. J Radioanal Chem 48:197-208.

Crossen, PE. (1983) Arsenic and SCE in human lymphocytes. Mut Res 119:415-419.

Cui, X; Kobayashi, Y; Hayakawa, T; Hirano, S. (2004a)  Arsenic speciation in bile and urine following oral and
intravenous exposure to inorganic and organic arsenics in rats. Toxicol Sci 82:478-487.

Cui, X; Li, S; Shraim, A; et al. (2004b) Subchronic exposure to arsenic through drinking water alters expression of
cancer-related genes in rat liver. Toxicol Pathol 32:64-72.

Cui, X; Wakai, T; Shirai, Y. (2006) Chronic oral exposure to inorganic arsenate interferes with methylation status of
p!6INK4a and RASSF1A and induces lung cancer in A/J mice. Toxicol Sci 91:372-381.

Das, D;  Chatterjee, A; Mandal, BK; et al. (1995) Arsenic in ground water in six districts of West Bengal, India: the
biggest arsenic calamity in the world. Part 2. Arsenic concentration in drinking water, hair, nails, urine, skin-scale,
and liver tissue (biopsy) of the affected people.  Analyst 120:917-924.

Das, S; Santra, A; Lahiri, S; Guha Mazumder, DN. (2005) Implications of oxidative stress and hepatic cytokine
(TNF-a and IL-6) response in the pathogenesis of hepatic collagenesis in chronic arsenic toxicity. Toxicol Appl
Pharmacol 204:18-26.

Datta, S; Talukder, G; Sharma, A. (1986) Cytotoxic effects of arsenic in dietary oil primed rats. Sci Culture 52:196-
198.

Deaglio, S; Canella, D; Baj, G; et al. (2001) Evidence of an immunologic mechanism behind the therapeutic effects
of arsenic trioxide (As2O3) on myeloma cells. Leuk Res  25:227-235.

De Chaudhuri, S; Mahata, J; Das, JK; et al. (2006) Association of specific p53 polymorphisms with keratosis in
individuals exposed to arsenic through drinking water in West Bengal, India. Mutat Res 601:102-112.

De Kimpe, J; Cornells, R; Mees, L; Vanholder, R. (1996) Basal metabolism of intraperitoneally injected carrier-free
  As-labeled arsenate in rabbits. Fundam Appl Toxicol 34:240-248.

De La Fuente, H; Portales-Perez, D; Baranda, L; et al. (2002) Effect of arsenic, cadmium, and lead on the induction
of apoptosis of normal human mononuclear cells. Clin Exp Immunol 129:69-77.


                                                  158        DRAFT—DO NOT CITE OR QUOTE

-------
Del Razo, LM; Styblo, M; Cullen, WR; Thomas DJ. (2001) Determination of trivalent methylated arsenicals in
biological matrices. Toxicol Appl Pharmacol 174:283-293.

Delnomdedieu, M; Basti, MM; Otvos, JD; Thomas, DJ. (1994a) Reduction and binding of arsenate and
dimethylarsinate by glutathione: a magnetic resonance study. Chern Biol Interact 90:139-155.

Delnomdedieu, M; Basti, MM; Styblo, M; et al. (1994b) Complexation of arsenic species in rabbit erythrocytes.
ChemRes Toxicol 7:621-627.

Delnomdedieu, M; Styblo, M; Thomas, DJ. (1995) Time dependence of accumulation and binding of inorganic and
organic arsenic species in rabbit erythrocytes. ChemBiol Interact 98:69-83.

Diaz, Z; Colombo, M; Mann, KK; et al. (2005) Trolox selectively enhances arsenic-mediated oxidative stress and
apoptosis in APL and other malignant cell lines. Blood 105:1237-1245.

Di Gioacchino, M; Verna, N; Di Giampaolo, L; et al. (2007) Immunotoxicity and sensitizing capacity of metal
compounds depends on speciation. Int J Immunopathol Pharmacol 20(2): 15-22.

Ding, W; Hudson, LG; Liu, KJ. (2005) Inorganic arsenic  compounds cause oxidative damage to DNA and protein by
inducing ROS and RNS generation in human keratinocytes. Mol Cell Biochem 279:105-112.

DiPaolo, JA; Casto, BC. (1979) Quantitative studies of in vitro morphological transformation of Syrian hamster cells
by inorganic metal salts. Cancer Res 39:1008-1013.

Dong, J-T; Luo, X-M. (1993) Arsenic-induced DNA-strand breaks associated with DNA-protein crosslinks in
human fetal lung fibroblasts. Mutat Res 302(2):97-102.

Dopp, E; Hartmann, LM; Florea, A-M; et al. (2004) Uptake of inorganic and organic derivatives of arsenic
associated with induced cytotoxic  and genotoxic effects in Chinese hamster ovary (CHO) cells. Toxicol Appl
Pharmacol 201:156-165.

Drobna, Z; Jaspers, I; Thomas, DJ; Styblo, M. (2002) Differential activation of AP-1 in human bladder epithelial
cells by inorganic and methylated arsenicals. FASEB J 17(l):67-69.

Drobna, Z; Waters,  SB; Walton, FS; et al.  (2004) Interindividual variation in the metabolism of arsenic in cultured
primary human hepatocytes. Toxicol Appl Pharm 201(2): 166-177.

Drobna, Z; Waters,  SB; Devesa, V; et al. (2005) Metabolism and toxicity of arsenic in human urothelial cells
expressing rat arsenic (+3 oxidation state)-methyltransferase. Toxicol Appl Pharmacol 207:147-159.

Drobna, Z; Xing, W; Thomas, DJ; Styblo, M. (2006) shRNA silencing of AS3MT expression minimizes arsenic
methylation capacity of HepG2 cells. ChemRes Toxicol  19:894-898.

DuMond, JW, Jr; Singh, KP. (2007) Gene expression changes and induction of cell proliferation by chronic
exposure to arsenic  of mouse testicular Leydig cells. J Toxicol Environ Health A 70:1150-1154.

Eblin, KE; Bowen, ME; Cromey, DW; et al. (2006) Arsenite and monomethylarsonous acid generate oxidative stress
response in human bladder cell culture. Toxicol Appl Pharmacol 217:7-14.

Eblin, KE; Bredfeldt, TG; Buffmgton, S; Gandolfi,  AJ. (2007) Mitogenic signal transduction caused by
monomethylarsonous acid in human bladder cells: role in arsenic-induced carcinogenesis. Toxicol Sci 95:321-330.

El-Masri, H; Kenyon, EM. (2008) Development of a human physiologically based pharmacokinetic (PBPK) model
for inorganic arsenic and  its mono- and di-methylated metabolites. J Pharmacokinet Pharmacodyn 35:31-68.

Engstrom, KS; Broberg, K; Concha, G; et  al. (2007) Genetic polymorphism influencing arsenic metabolism:
evidence from Argentina. Environ Health Perspect 115(4):599-605.
                                                 159        DRAFT—DO NOT CITE OR QUOTE

-------
Fan, S-R; Ho, I-C; Yeoh, FL-F; et al. (1996) Squalene inhibits sodium arsenite-induced sister chromatid exchanges
and micronuclei in Chinese hamster ovary-Kl cells. Mutat Res 368:165-169.

Felix, K; Manna, SK; Wise, K; Barr, J; Ramesh, GT.  (2005) Low levels of arsenite activates nuclear factor-KB and
activator protein-1 in immortalized mesencephalic cells. J Biochem Mol Toxicol 19:67-77.

Ferreccio, C; Gonzalez, C; Milosavjlevic; et al.  (2000) Lung cancer and arsenic concentrations in drinking water in
Chile. Epidem 1:673-679.

Fischer, JM; Robbins, SB; Al-Zoughool, M; et al. (2005) Co-mutagenic activity of arsenic andbenzo[a]pyrene in
mouse skin. MutRes 588:35-46.

Fischer, JM; Robbins, SB; Kannamkumarath, SS; et al. (2006) Exposure of mice to arsenic and/or benzo[a]pyrene
does not increase the frequency of Aprt-deficient cells recovered from explanted skin of Aprt heterozygous mice.
Environ Mol Mutagen 47:334-344.

Flora, SJ. (1999) Arsenic-induced oxidative stress and its reversibility following combined administration of N-
acetylcysteine and meso 2,3-dimercaptosuccinic acid  in rats. ClinExp Pharmacol Physiol 26:865-869.

Flora, SJS; Bhadauria, S; Pant, SC; Dhaked, RK. (2005) Arsenic induced blood and brain oxidative stress and its
response to some thiol chelators in rats. Life Sci 77:2324-2337.

Florea, A-M; Yamoah, EN; Dopp, E. (2005) Intracellular calcium disturbances induced by arsenic and its methylated
derivatives in relation to genomic damage and apoptosis induction. Environ Health Perspect 113(6):659-664.

Florea, A-M; Splettstoesser, F; Btisselberg, D. (2007) Arsenic trioxide (As2O3) induced calcium signals and
cytotoxicity in two human cell lines: SY-5Y neuroblastoma and 293 embryonic kidney (HEK). Toxicol Appl
Pharmacol 220:292-301.

Freeman, GB; Schoof, RA; Ruby, MV; et al. (1995) Bioavailability of arsenic in soil and house dust impacted by
smelter activities following oral administration in Cynomologus monkeys. Fund Appl Toxicol 28:215-222.

Fujino, Y; Guo, X; Liu, J; et al.  (2005) Chronic arsenic exposure and urinary 8-hydroxy-2'-deoxyguanosine in an
arsenic-affected area in  Inner Mongolia, China.  JExpo Anal Environ Epidemiol 15:147-152.

Gail MH; Kessler M; Midthune D;  Scoppa  S. (1999)  Two approaches for estimating disease prevalence from
population-based registries of incidence and total mortality. Biometrics 55:1137-1144.

Galicia, G; Leyva, R; Tenorio, EP;  et al. (2003) Sodium arsenite retards proliferation of PHA-activated T cells by
delaying the production and secretion of IL-2. Int Immunopharmacol 3:671-682.

Gamble, MV; Liu, X; Ahsan, H; et al. (2005) Folate,  homocysteine, and arsenic metabolism in arsenic-exposed
individuals in Bangladesh. Environ Health Perspect 113(12): 1683-1688.

Gamble, MV; Liu, X; Ahsan, H; et al. (2006) Folate and arsenic metabolism: a double-blind, placebo-controlled
folic acid-supplementation trial in Bangladesh. Am J  Clin Nutr 84:1093-1101.

Ganyc, D; Talbot, S; Konate, F; et al. (2007) Impact of trivalent arsenicals on selenoprotein synthesis. Environ
Health Perspect 115:346-353.

Gentry, PR; Covington, TR; Mann, S; et al. (2004)  Physiologically based pharmacokinetic modeling of arsenic in the
mouse. J Toxicol Environ Health A 67(1):43-71.

Germolec, DR; Spalding, J; Boorman, GA; et al. (1997) Arsenic can mediate skin neoplasia by chronic stimulation
of keratinocyte-derived  growth factors. Mutat Res 386:209-218.


                                                  160         DRAFT—DO NOT CITE  OR QUOTE

-------
Germolec, DR; Spalding, J; Yu, HS; et al. (1998) Arsenic enhancement of skin neoplasia by chronic stimulation of
growth factors. AmJPathol 153:1775-1785.

Geubel, AP; Mairlot, MC; Buchet, JP; Lauwerys, R. (1988) Abnormal methylation capacity in human liver cirrhosis.
Int J ClinPharmacol Res 8:117-122.

Ghosh, P; Basu, A; Mahata, J; et al. (2006) Cytogenetic damage and genetic variants in the individuals susceptibile
to arsenic-induced cancer through drinking water. Int J Cancer 118:2470-2478.

Gomez, SE; del Razo, LM; Sanchez, JLM. (2005) Induction of DNA damage by free radicals generated either by
organic or inorganic arsenic (As  ,MMA  , and DMA  ) in cultures of B and T lymphocytes. Bio Trace Elem Res
108:115-126.

Gonsebatt, ME; Vega, L; Montero, R; et al. (1994) Lymphocyte replicating ability in individuals exposed to arsenic
via drinking water. Mutat Res 313:293-299.

Gonsebatt, ME; Vega, L; Salazar, AM. (1997). Cytogenic effects in human exposure to arsenic. Mutat Res 386:219-
228.

Gonzalez-Rangel, Y; Portales-Perez, DP; Galicia-Cruz, O; Escudero-Lourdes, C. (2005) Chronic exposure to arsenic
sensitizes CD3+ and CD56+ human cells to sodium arsenite-mediated apoptosis. Proc West Pharmacol Soc 48:89-
91.

Gottschalg, E; Moore, NE; Ryan, AC; et al. (2006) Phenotypic anchoring of arsenic and cadmium toxicity in three
hepatic-related cell systems reveals compound- and cell-specific selective up-regulation of stress protein expression:
implications for fingerprint profiling of cytotoxicity. Chem Biol Interact 161:251-261.

Graham-Evans, B; Cohly, HH; Yu, H; Tchounwou, PB. (2004) Arsenic-induced genotoxic and cytotoxic effects in
human keratinocytes, melanocytes and dendritic cells. Int J Environ Res Public Health 1:83-89.

Gregus, Z; Nemeti, B. (2002) Purine nucleoside phosphorylase as a cytosolic arsenate reductase. Toxicol Sci 70:13-
19.

Gregus, Z; Nemeti, B. (2005) The glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase works as an
arsenate reductase in human red blood cells and rat liver cytosol. Toxicol Sci 85(2):859-869.

Griffin, RJ; Monzen, H; Williams, BW; et al. (2003) Arsenic trioxide induces selective tumor vascular damage via
oxidative stress and increases thermosensitivity of tumours. Int J Hypertherm 19:575-589.

Guo, HR; Chiang, HS; Hu, H; et al. (1997) Arsenic in drinking water and incidence of urinary cancers. Epidem
8:545-550.

Guo, HR; Yu, H-S; Hu, H; Monson, RR. (2001) Arsenic in drinking water and skin cancers: cell-type specificity
(Taiwan, R.O.C.). Cancer Causes Control 12:909-916.

Guo, HR. (2003) The lack of a specific association between arsenic in drinking water and hepatocellular carcinoma.
JHepatol 39:383-388.

Guo, HR. (2004) Arsenic level in drinking water and mortality of lung cancer (Taiwan). Cancer Causes Control
15(2):171-177.

Gupta, R; Flora, SJ. (2005) Therapeutic value of Hippophae  rhamnoides L. against subchronic arsenic toxicity in
mice. J Med Food 8:353-361.

Gupta, R; Flora, SJ. (2006) Effect of Centella asiatica on arsenic induced oxidative stress and metal distribution in
rats. J Appl Toxicol 26:213-222.


                                                  161        DRAFT—DO NOT CITE OR QUOTE

-------
Guyton, KZ; Xu, QB; Holbrook, NJ. (1996) Induction of the mammalian stress response gene GADD153 by
oxidative stress: role of AP-1 element. Biochem J 314:547-554.

Hagiwara, M; Watanabe, E; Barrett, JC; Tsutsui, T. (2006) Assessment of genotoxicity of 14 chemical agents used
in dental practice: ability to induce chromosome aberrations in Syrian hamster embryo cells. Mutat Res 603:111-
120.

Hall, M; Gamble, M; Slakovich, V; et al. (2007) Determinants of arsenic metabolism: blood arsenic metabolites,
plasma folate, cobalamine, and homocysteine concentrations in maternal-newborn pairs. Environ Health Perspect
115:1503-1509.

Hamadeh, HK; Vargas, M; Lee, M; et al. (1999) Arsenic disrupts cellular levels of p53 and mdm2: a potential
mechanism of carcinogenesis. Biochem Biophys Res Commun 263:446-449.

Han, SG; Castranova, V; Vallyathan, V. (2005) Heat shock protein 70 as an indicator of early lung injury caused by
exposure to arsenic. Mol Cell Biochem 277:153-164.

Hanahan, D; Weinberg, RA. (2000) The hallmarks of cancer. Cell  100:57-70.

Hanlon, DP; Perm, VH. (1977) Placenta! permeability of arsenate ion during early embryogenesis in the hamster.
Experientia 33(9):1221-1222.

Hansen, HR- Raab, A; Jaspars, M; et al. (2004) Sulfur-containing arsenicals mistaken for dimethylarsinous
acid[DMA(  )] and identified as a natural metabolite in urine: major implication for studies on arsenic metabolism
andtoxicity. Chem Res Toxicol 17:1086-1091.

Hansen, JM; Zhang, H; Jones, DP. (2006) Differential oxidation of thioredoxin-1, thioredoxin-2, and glutathione by
metal ions. Free Radic Biol Med 40:138-145.

Harrington, JM; Middaugh, JP; Morse, DL; Housworth, J. (1978) A survey of a population exposed to high
concentrations of arsenic in well water in Fairbanks, Alaska. Am J Epidemiol 108:377-385.

Harrison, MT; McCoy, KL. (2001) Immunosuppression by arsenic: a comparison of cathepsin L inhibition and
apoptosis. Int Immunopharmacol 1:647-656.

Hasgekar, N; Beck, JP; Dunkelberg, H; et al. (2006) Influence of antimonite, selenite, and mercury on the toxicity of
arsenite in primary rat hepatocytes. Biol Trace Elem Res 111:167-183.

Hayakawa, T; Kobayashi, Y; Cui, X; Hirano, S.  (2005) A new metabolic pathway of arsenite: arsenic-glutathione
complexes are substrates for human arsenic methyltransferase Cytl9. Arch Toxicol 79:183-191.

He, X; Chen, MG; Lin, GX; Ma, Q. (2006) Arsenic induces NAD(P)H-quinone oxidoreductase I by disrupting the
Nrf2 x Keapl x Cul3 complex and recruiting Nrf2 x Maf to the antioxidant response element enhancer. J Biol Chem
281:23620-23631

Healy, SM; Zakharyan, RA; Aposhian, HV. (1997) Enzymatic methylation of arsenic compounds: IV. In vitro and in
vivo deficiency of methylation of arsenite and monomethylarsonic acid in the guinea pig. Mutat Res 386:229-239.

Healy, SM; Casarez, EA; Ayala-Fierro, F; Aposhian, HV. (1998) Enzymatic methylation of arsenic compounds: V.
Arsenite methyltransferase activity in tissues of mice. Toxicol Appl Pharmacol 148:65-70.

Healy, SM; Wildfang, E; Zakharyan, RA; Aposhian, HV. (1999) Diversity of inorganic arsenite biotransformation.
Biol Trace Elem Res 68(3):249-266.

Heck, JE; Gamble, MV; Chen, Y; et al. (2007) Consumption of folate-related nutrients and metabolism of arsenic in
Bangladesh. Am J Clin Nutr 85:1367-1374.


                                                  162        DRAFT—DO NOT CITE OR QUOTE

-------
Hei, TK; Liu, SX; Waldren, C. (1998) Mutagenicity of arsenic in mammalian cells: role of reactive oxygen species.
Proc Natl Acad Sci USA 95:8103-8107.

Hernandez-Zavala, A; Cordova, E; Del Razo, LM; et al. (2005) Effects of arsenite on cell cycle progression in a
human bladder cancer cell line. Toxicology 207:49-57.

Heydorn, K. (1970) Environmental variation of arsenic levels in human blood determined by neutron activation
analysis. Clin Chem Acta 28:349-357.

Hinwood, AL; Jolley, DJ; Sim, MR. (1999) Cancer incidence and high environmental arsenic concentrations in rural
populations: results of an ecological study. Int J Environ Health Res 9(2):131-141.

Hirano,  S; Kobayashi, Y; Cui, X; et al. (2004) The accumulation and toxicity of methylated arsenicals in endothelial
cells: important roles of thiol compounds. Toxicol Appl Pharmacol 198:458^67.

Hood, RD; Vedel-Macrander, GC; Zaworotko, MJ; et al. (1987) Distribution, metabolism, and fetal uptake of
pentavalent arsenic in pregnant mice following oral or intraperitoneal administration. Teratology 35:19-25.

Hopenhayn-Rich, C; Biggs, ML; Fuchs, A; et al.  (1996a) Bladder cancer mortality associated with arsenic in
drinking water in Argentina. Epidemiology 7:117-124.

Hopenhayn-Rich, C; Biggs, ML; Kalman, DA; et al. (1996b) Methylation study of a population environmentally
exposed to arsenic in drinking water. Environ Health Perspect 104:620-628.

Hopenhayn-Rich, C; Biggs, ML; Smith, AH. (1998) Lung and kidney cancer mortality associated with arsenic in
drinking water in Cordoba, Argentina. Int J Epidemiol 27:561-569.

Hornhardt, S; Gomolka, M; Walsh, L; Jung, T. (2006) Comparative investigations of sodium arsenite, arsenic
trioxide and cadmium sulphate in combination with gamma-radiation on apoptosis, micronuclei induction and DNA
damage in a human lymphoblastoid cell line. Mutat Res 600:165-176.

Hostynek, JJ; Hinz, RS; Lorence, CR; et al. (1993) Metals and the skin. Crit Rev Toxicol 23:171-235.

Hou, YC; Hsu, CS; Yeh,  CL;  et al. (2005) Effects of glutamine on adhesion molecule expression and leukocyte
transmigration in endothelial cells exposed to arsenic. J Nutr Biochem 16:700-704.

Hour, T-C; Pu, Y-C; Lin, C-C; et al. (2006) Differential expression of molecular markers in arsenic- and non-
arsenic-related urothelial cancer. Anticancer Res  26:375-378.

HSDB (Hazardous Substance Data Bank). (2005) National Library of Medicine. Records for element arsenic and
arsenic compounds. Bethesda, MD: NLM, HSDB.

Hsu, S-H; Tsou, T-C; Chiu, S-J; Chao, J-I. (2005) Inhibition of a7-nicotinic acetylcholine receptor expression by
arsenite in the vascular endothelial cells. Toxicol Lett 159:47-59.

Hsu, K-H; Brandt-Rauf, P; Lin, T-M; et al. (2006) Plasma-transforming growth factor-alpha expression in residents
of an arseniasis area in Taiwan. Biomarkers 11:538-546.

Hu, Y; Su, L; Snow, ET.  (1998) Arsenic toxicity is enzyme specific and its affects on ligation are not caused by the
direct inhibition of DNA repair enzymes. Mutat Res 408:203-218.

Huang, RN; Lee, TC. (1996) Cellular uptake of trivalent arsenite and pentavalent arsenate in KB cells cultured in
phosphate-free medium. Toxicol Appl Pharmacol 136(2):243-249.

Huang,  SC; Lee, TC. (1998) Arsenite inhibits mitotic division and perturbs spindle dynamics in HeLa S3 cells.
Carcinogenesis 19:889-896.


                                                  163        DRAFT—DO NOT CITE OR QUOTE

-------
Huang, H; Huang, CF; Wu, DR; et al. (1993) Glutathione as a cellular defense against arsenite toxicity in cultured
Chinese hamster ovary cells. Toxicology 79:195-204.

Huang, C; Ma, WY; Li, J; et al. (1999a) Requirement of Erk, but not INK, for arsenite-induced cell transformation. J
Biol Chem 274:14595-14601.

Huang, C; Ma, WY; Li, J; et al. (1999b) Arsenic induces apoptosis through a c-Jun NH2-terminal kinase-dependent,
p-53-independent pathway.  Cancer Res 59:3053-3058.

Huang, C; Ke, Q; Costa, M; Shi, X. (2004) Molecular mechanisms of arsenic carcinogenesis. Mol Cell Biochem
255:57-66.

Huang, HS; Liu, ZM;  Ding, L; et al. (2006) Opposite effect of ERK1/2 and JNK on p53-independent
p21 WAF1/CIP1 activation involved in the arsenic trioxide-induced human epidermoid carcinoma A431 cellular
cytotoxicity. J Biomed Sci 13:113-125.

Huang, Y-C; Hung, W-C; Kang, W.Y; et al. (2007a) Expression of STATS and Bcl-6 oncoprotein in sodium
arsenite-treated SV-40 immortalized human uroepithelial cells. Toxicol Lett 173:57-65.

Huang, Y-K; Tseng, C-H; Huang, Y-L; et al. (2007b) Arsenic methylation capability and hypertension risk in
subjects living in arseniasis-hyperendemic areas in southwestern Taiwan. Toxicol Appl Pharmacol 218:135-142.

Hughes, MF; Kenyon, EM.  (1998) Dose-dependent effects on the disposition of monomethylarsonic acid and
dimethylarsinic acid in the mouse after intravenous administration. J Toxicol Environ Health 53:95-l 12.

Hughes, M; Kenyon, E; Edwards, B; et al.  (1999) Strain-dependent disposition of inorganic arsenic in the mouse.
Toxicology 137:95-108.

Hughes, MF; Del Razo, LM; Kenyon, EM. (2000) Dose-dependent effects on tissue distribution and metabolism of
dimethylarsinic acid in the mouse after intravenous administration. Toxicology 143(2): 155-166.

Hughes, MF; Kenyon, EM;  Edwards, BC; et al. (2003) Accumulation and metabolism of arsenic in mice after
repeated oral administration of arsenate. Toxicol Appl Pharmacol 191(3):202-210.

Hughes, MF; Devesa,  V;  Adair, BM; et al. (2005) Tissue dosimetry, metabolism and excretion of pentavalent and
trivalent monomethylated arsenic in mice after oral administration. Toxicol Appl Pharmacol 208:186-197.

Hutchinson, J. (1887)  Arsenic cancer. BrMed J 2:1280-1281.

Hwang, BJ; Utti, C; Steinberg, M. (2006) Induction of cyclinDl by submicromolar concentrations of arsenite in
human epidermal keratinocytes. Toxicol Appl Pharmacol 217:161-167.

Islam, LN; Nabi, AHM; Rahman, MM; Zahid, MSH. (2007) Association of respiratory complications and elevated
serum immunoglobulins with drinking water arsenic toxicity in humans. J Environ Sci Health A 42:1807-1814.

Ivanov, VN; Hei, TK.  (2006) Sodium arsenite accelerates TRAIL-mediated apoptosis in melanoma cells through
upregulation of TRAIL-R1/R2 surface levels and downregulation of cFLIP expression. Exp Cell Res 312:4120-
4138.

Jacobson-Kram, D; Montalbano, D. (1985) The reproductive effects assessment group's report on the mutagenicity
of inorganic arsenic. Environ Mutagen 7:787-804.

Jan, KY; Wang, TC; Ramanathan, B; Gurr, JR. (2006) Dithiol compounds at low concentrations increase arsenite
toxicity. Toxicol Sci 90:432-439.

Jha, AN; Noditi, M; Nilsson, R; Natarajan, AT. (1992) Genotoxic effects of sodium arsenite on human cells. Mutat
Res284(2):215-221.

                                                 164       DRAFT—DO NOT CITE OR QUOTE

-------
Jiang, HY; Jiang, L; Wek, RC. (2007) The eukaryotic initiation factor-2 kinase pathway facilitates differential
GADD45a expression in response to environmental stress. J Biol Chem 282:3755-3765.

Jin, Y; Xi, S; Li, X; et al. (2006a) Arsenic speciation transported through the placenta from mother mice to their
newborn pups. Environ Res 101:349-355.

Jin, HO; Yoon, SI; Seo, SK; et al. (2006b)  Synergistic induction of apoptosis by sulindac and arsenic trioxide in
human lung cancer A549 cells via reactive oxygen species-dependent down-regulation of survivin. Biochem
Pharmacol 72:1228-1236.

Jing, Y; Dai, J; Chalmers-Redman, RM; et al. (1999) Arsenic trioxide selectively induces acute promyelocytic
leukemia cell apoptosis via a hydrogen peroxide-dependent pathway. Blood 94:2101-2 111.

Josyula, AB; Poplin, GS; Kurzius-Spencer, M; et al. (2006) Environmental arsenic exposure and sputum
metalloproteinase concentrations. Environ Res 102:283-290.

Juhasz, AL; Smith, E; Weber, J; et al.  (2006) In vivo assessment of arsenic bioavailability rice and its significance
for human health risk assessment. Environ Health Perspect 114:1826-1831.

Juhasz, AL; Smith, E; Weber, J; et al.  (2007) Comparison of in vivo and in vitro methodologies for the assessment of
arsenic bioavailability in contaminated soils. Chemosphere 69:961-966.

Kagey, BY; Bumgarner, JE; Creason,  JP. (1977) Arsenic levels in maternal-fetal tissue sets. In: Hemphill, DD; ed.
Trace substances in environmental health-XI, proceedings of the University of Missouri's llth Annual Conference
on Trace Substances in Environmental Health. Columbia, MO: University of Missouri Press. Pp. 252-256.

Kala, SV; Neely, MW; Kala, G; et al.  (2000) The MRP2/cMOAT transporter and arsenic-glutathione complex
formation are required for biliary excretion of arsenic. J Biol Chem 275(43):33404-33408.

Kalia, K; Narula, GI; Kannan, GM; Flora, SJ. (2007) Effects of combined administration of captopril and DMSA on
arsenite induced oxidative stress and blood and tissue arsenic concentration in rats. Comp Biochem Physiol C
Toxicol Pharmacol 144:372-379.

Kaltreider, RC; Pesce, CA; Ihnat, MA; et al. (1999) Differential effects of arsenic( ) and chromium(VI) on nuclear
transcription factor binding. Mol Carcinog 25:219-229.

Kaltreider, RC; Davis, AM; Lariviere, JP; Hamilton, JW. (2001) Arsenic alters the function of the glucocorticoid
receptor as a transcription factor. Environ Health Perspect 109:245-251.

Kamat, CD; Green, DE; Curilla, S; et al. (2005) Role of HIF signaling on tumorigenesis in response to chronic low-
dose arsenic administration. Toxicolog Sci 86(2):248-257.

Kann, S; Huang, MY; Estes, C; et al. (2005a) Arsenite-induced aryl hydrocarbon receptor nuclear translocation
results in additive induction of phase I genes and synergistic induction of phase II genes. Mol Pharmacol 68:336-
346.

Kann, S; Estes, C; Reichard, JF; et al.  (2005b) Butylhydroquinone protects cells genetically deficient in glutathione
biosynthesis from arsenite-induced apoptosis without significantly changing their prooxidant status.  Toxicol Sci
87:365-384.

Karagas, MR; Stukel, TA; Moriss,  JS; et al. (2001) Skin cancer risk in relation to toenail arsenic concentrations in a
U.S. population-based case-control study. Am JEpidemiol  153:559-565.

Karagas, MR; Tosteson, TD; Morris, JS; et al. (2004) Incidence of transitional cell carcinoma of the bladder an
arsenic exposure in New Hampshire. Cancer Causes Control 15:465-472.


                                                   165        DRAFT—DO NOT CITE OR QUOTE

-------
Karasawas, N; Carcamo, JM; Stratis, G; Golde, DW. (2005) Vitamin C protects HL60 and U266 cells from arsenic
toxicity. Blood 105:4004-4012.

Kawaguchi, K; Oku, N; Rin, K; et al. (1996) Dimethylarsenics reveal DNA damage induced by superoxide anion
radicals. Biol Pharm Bull 19:551-553.

Kawata, K, Yokoo, H, Shimazaki, R and Okabe, S. (2007) Classification of heavy-metal toxicity by human DNA
microarray analysis. Environ Sci Technol 41:3769-3774.

Kayajanian, G. (2003) Arsenic, cancer, and thoughtless policy. EcotoxEnv Safety 55:139-142.

Kenyon, E; Hughes, M; Levander, O. (1997) Influence of dietary selenium on the disposition of arsenate in female
B6C3F1 mouse. J Toxicol Environ Health 51:279-299.

Kenyon, EM; Hughes, MF. (2001) A concise review of the toxicity and carcinogenicity of dimethylarsinic acid.
Toxicology 160(l-3):227-236.

Kenyon, EM; Del Razo, LM; Hughes, MF. (2005a) Tissue distribution and urinary excretion of inorganic arsenic
and its methylated metabolites in mice following acute oral administration of arsenate. Toxicol Sci 85(1):468^175.

Kenyon, EM; Del Razo, LM; Hughes, MF; Kitchin, KT. (2005b) An integrated pharmacokinetic and
pharmacodynamic study of arsenite action. 2. Heme oxygenase induction in mice.  Toxicology 206:389-401.

Kenyon, EM; Huges, MF; Adair, BM; et al. (2008) Tissue distribution and urinary excretion of inorganic arsenic and
its methylated metabolites in C57BL6 mice following subchronic exposure to arsenate in drinking water. Toxicol
Appl Pharmacol 232:448-455.

Kessel, M; Liu, SX; Xu, A; et al. (2002) Arsenic induces oxidative DNA damage in mammalian cells. Mol Cell
Biochem 234-235:301-308.

Khalil, S; Luciano, J; Chen, W; Liu, AY. (2006) Dynamic regulation and involvement of the heat shock
transcriptional response in arsenic carcinogenesis. J Cell Physiol 207:562-569.

Kile, ML;  Houseman, EA; Brenton, CV; et al. (2007) Dietary arsenic exposure in Bangladesh.  Environ Health
Perspect 115:889-893.

Kimura, S; Yamauchi, H; Hibino, Y; et al. (2006) Evaluation of urinary 8-hydroxydeoxyguanine in healthy Japanese
people. Basic Clin Pharmacol Toxicol 98:496-502.

Kinoshita, A; Wanibuchi, H; Wei, M; et al. (2007a) Elevation of 8-hydroxydeoxyguanosine and cell proliferation via
generation of oxidative stress by organic arsenicals contributes to their carcinogenicity in the rat liver and bladder.
Toxicol Appl Pharmacol 221:295-305.

Kinoshita, A; Wanibuchi, H; Morimura, K; et al. (2007b) Carcinogenicity of dimethylarsinic acid in Oggl-deficient
mice.  Cancer Sci 98:803-814.

Kitchin, KT. (2001) Review. Recent advances in arsenic carcinogenesis:  modes of action, animal model systems,
and methylated arsenic metabolites. Toxicol Appl Pharmacol 172:249-261.

Kitchin, KT; Ahmad, S. (2003) Oxidative stress as a possible mode of action for arsenic carcinogenesis. Toxicol Lett
137:3-13.

Kitchin, KT; Wallace, R. (2006) Dissociation of arsenite-peptide complexes: triphasic nature, rate constants, half-
lives,  and biological importance. J Biochem Mol Toxicol 20(l):48-56.
                                                  166        DRAFT—DO NOT CITE OR QUOTE

-------
Kitchin, KT; Wallace, K; Andrewes, P. (2003) Some chemical properties underlying arsenic's biological activity. In:
Chappell, WR; Abernathy, CO; Calderon, RL; Thomas, DL; eds. Chapter 26 in arsenic exposure and health effects.
Elserier Science B.V. Publishers.

Klein, CB; Leszczynska, J; Hickey, C; Rossman, TG. (2007) Further evidence against a direct genotoxic mode of
action for arsenic-induced cancer. Toxicol Appl Pharmacol 222:289-297.

Kligerman, AD; Doerr, CL; Tennant, AH; et al. (2003) Methylated trivalent arsenicals as candidate ultimate
genotoxic forms of arsenic: induction of chromosomal mutations but not gene mutations. Environ Mol Mutagen
42:192-205.

Kligerman, AD; Doerr, CL; Tennant, AH. (2005) Oxidation and methylation status determine the effects of arsenic
on the mitotic apparatus. Mol Cell Biochem 279:113-121.

Kligerman, AD; Tennant, AH. (2007) Insights into the carcinogenic mode of action of arsenic. Toxicol Appl
Pharmacol 222(3):281-288.

Knobeloch, LM; Zierold, KM; Anderson, HA. (2006) Association of arsenic-contaminated drinking-water with
prevalence of skin cancer in Wisconsin's Fox River Valley. J Health Popul  Nutr 24(2):206-213.

Kochhar, TS; Howard, W; Hoffman, S; Brammer-Carleton, L. (1996)  Effect of trivalent and pentavalent arsenic in
causing chromosome alterations in cultured Chinese hamster ovary (CHO) cells. Toxicol Lett 84:37-42.

Kojima, C; Qu, W; Waalkes, MP; et al. (2006) Chronic exposure to methylated arsenicals stimulates arsenic
excretion pathways and  induces arsenic tolerance in rat liver cells. Toxicol Sci 91:70-81.

Kokilavani, V; Devi, MA; Sivarajan, K; Panneerselvam, C. (2005) Combined efficacies of DL-a-lipoic acid and
meso 2,3 dimercaptosuccinic acid against arsenic induced toxicity in antioxidant systems of rats. Toxicol Lett 160:1-
7.

Komissarova, ET; Saha, SK; Rossman, TG. (2005) Dead or dying: the importance of time in cytotoxicity assays
using arsenite as an example. Toxicol Appl Pharmacol 202:99-107.

Konig, H; Hartel, N; Schultheis, B; et al. (2007) Enhanced Bcr-Abl-specific antileukemic activity of arsenic trioxide
(Trisenox) through glutathione-depletion in imatinib-resistant cells. Haematologica. 92:838-841.

Kreppel, H; Liu, J; Liu,  Y; et al. (1994) Zinc-induced arsenite tolerance in mice. Fundam Appl Toxicol 23:32-37.

Kurttio, P; Pukkala, E; Kahelin, H; et al. (1999). Arsenic concentration in well water and risk of bladder and kidney
cancer in Finland. Environ Health Perspect  107:705-710.

Lamm, SH; Byrd, DM; Kruse, MB; et al. (2003) Bladder cancer and arsenic exposure: Differences in the two
populations enrolled in a study in southwest Taiwan. Biomed Environ Sci 16:355-368.

Lamm, SH; Engel, A; Kruse, MB; et al.  (2004) Arsenic in drinking water and bladder cancer mortality in the U.S.:
an analysis based on 133 U.S. counties and thirty years of observation. J Occupation Environ Med 46:298-306

Lamm, SH; Engel, A; Perm, CA;  et al. (2006) Arsenic cancer risk confounder in southwest Taiwan data set. Env
Health Persp 114(7): 1077-1082.

Lamm, S; Luo, Z-D; Bo, F-B; et al. (2007) An epidemiologic study of arsenic-related skin disorders and skin cancer
and the consumption of  arsenic contaminated well waters in Huhhot, Inner Mongolia, China. Hum Ecol Risk Assess
13:713-746.

Lander, HM. (1997)  An essential role for free radicals and derived species in signal transduction. FASEB J 11:118-
124.


                                                  167        DRAFT—DO NOT CITE  OR QUOTE

-------
Lantz, RC; Hays, AM. (2006) Role of oxidative stress in arsenic-induced toxicity. Drug Metab Rev 38:791-804.

Lantz, RC; Lynch, BJ; Boitano, S; et al. (2007) Pulmonary biomarkers based on alterations in protein expression
following exposure to arsenic. Environ Health Perspect 115:586-591.

Lanz, H Jr; Wallace, PC; Hamilton, JG. (1950) The metabolism of arsenic in laboratory animals using As74 as a
tracer. Univ Calif Publ Pharmacol 2:263-282.

Laparra, JM, Velez, D, Barbera, R, et al. (2005a) An approach to As( ) and As(V) bioavailability studies with
Caco-2 cells. Toxicol In Vitro 19:1071-1078.

Laparra, JM; Velez, D; Barbera, R; et al. (2005b) Bioavailability of inorganic arsenic in cooked rice: practical
aspects for human health risk assessments. J Agric Food Chem 53:8829-8833.

Laparra, JM; Velez, D; Barbera, R; et al. (2006) Cytotoxic effect of As( ) in Caco-2 cells and evaluation of its
human intestinal permeability. Toxicol iln Vitro 20:658-663.

Larramendy, ML; Popescu, NC; DiPaolo, JA. (1981) Induction by inorganic metal salts of sister chromatid
exchanges and chromosome aberrations in human and Syrian hamster strains. Environ Mutagen 3:597-606.

Lau, ATY; Chiu, J-F. (2006) Proteomic and biochemical analyses of in vitro carcinogen-induced lung cell
transformation: synergism between arsenic andbenzo[a]pyrene. Proteomics 6:1619-1630.

Le, XC; Ma, M; Cullen WR; et al. (2000a) Determination of monomethylarsonous acid, a key arsenic methylation
intermediate, in human urine. Environ Health Perspect 108(11): 1015-1018.

Le, XC; Lu, X; Ma, M; et al. (2000b) Speciation of key arsenic metabolic intermediates in human urine. Anal Chem
72:5172-5177.

Lee, T-C; Ho, I-C. (1995) Modulation of cellular antioxidant defense activities by sodium arsenite in human
fibroblasts. Arch Toxicol 69:498-504.

Lee, T-C; Huang, RY; Jan, KY. (1985) Sodium arsenite enhances the cytotoxicity clastogenicity and 6-thioguanine-
resistant mutagenicity of ultraviolet light in Chinese hamster ovary cells. Mutat Res 148:83-89.

Lee, T-C; Wang-Wuu, S; Huang, RY; et al. (1986) Differential effects of the pre- post-treatment of sodium arsenite
on the genotoxicity of methylmethanesulfonate in Chinese hamster ovary cells.  Cancer Res 46:1854-1857.

Lee, C; Lee, YM; Rice, RH. (2005) Human epidermal cell protein responses to  arsenite treatment in culture. Chem-
Bio Interactions 155:43-54.

Lee, T-C; Ho, I-C; Lu, W-J; Huang, J-D. (2006a) Enhanced expression of multidrug resistance-associated protein 2
and reduced expression of aquaglyceroporin 3 in an arsenic-resistant human cell line. J Biologic Chem
281(27): 18401-1840.

Lee, C-H; Chen, J-S; Sun, Y-L; et al. (2006b) Defective pl-integrins expression in arsenical keratosis and arsenic-
treated cultured human keratinocytes. J Cutan Pathol 33:129-138.

Lee, M; Hwang, JT; Yun, H; et al. (2006c) Critical roles of AMP-activated protein kinase in the carcinogenic metal-
induced expression of VEGF and HIF-1 proteins inDU145 prostate carcinoma. Biochem Pharmacol 72:91-103.

Lee-Chen, SF; Gurr, JR; Lin, IB; Jan, KY. (1993) Arsenite exchanges DNA double-strand breaks and cell killing of
methyl methanesulfonate-treated cells by inhibiting the excision of alkali-labile sites. Mutat Res 294(l):21-28.

Lemarie, A; Morzadec, C; Merino, D; et al. (2006a). Arsenic trioxide induces apoptosis of human monocytes during
macrophagic differentiation through nuclear factor-KB-related survival pathway down-regulation. J Pharmacol Exp
Ther 316:304-314.

                                                  168        DRAFT—DO NOT CITE OR QUOTE

-------
Lemarie, A; Morzadec, C; Bourdonnay, E; et al. (2006b). Human macrophages constitute targets for immunotoxic
inorganic arsenic. J Immunol 177:3019-3027.

Lerman, SA; Clarkson, TW. (1983) The metabolism of arsenite and arsenate by the rat. Fundam Appl Toxicol
3(4):309-314.

Lerman, SA; Clarkson, TW; Gerson, RJ. (1983) Arsenic uptake and metabolism by liver cells is dependent on
arsenic oxidation state. ChemBiol Interact 45:401-406.

Leslie, EM; Haimeur, A; Waalkes, MP. (2004) Arsenic transport by the human multidrug resistance protein 1
(MRP1/ABCC1). JBiol Chem 279(31):32700-32708.

Lewis, DR; Southwick, JW; Ouellet-Hellstrom, R; et al. (1999) Drinking water arsenic in Utah: a cohort-mortality
study. Environ Health Perspect 107:359-365.

Li, J-H; Rossman, TG. (1989a) Mechanism of comutagenesis of sodium arsenite with N-methyl-N-nitrosourea. Biol
Trace Element Res 21:373-381.

Li, J-H; Rossman, TG. (1989b) Inhibition of DNA ligase activity by arsenite: a possible mechanism of its
comutagenesis. Mol Toxicol 2:1-9.

Li, J-H; Rossman, TG. (1991) Comutagenesis of sodium arsenite with ultraviolet radiation in Chinese hamster V79
cells. Biol Met 4:197-200.

Li, W; Chou, I-N. (1992) Effects of sodium arsenite on the cytoskeleton and cellular glutathione levels in cultured
cells. Toxicol Appl Pharmacol 114:132-139.

Li, YM; Broome, JD. (1999) Arsenic targets tubulins to induce apoptosis in myeloid leukemia cells. Cancer Res
59:776-780.

Li, JJ; Oberley, LW; Fan, M; et al. (1998) Inhibition of AP-1 andNF-kappaB by  manganese-containing superoxide
dismutase in human breast cancer cells. FASEB J 12:1713-1723.

Li, D; Morimoto, K; Takeshita, T; Lu, Y. (2001) Formamidopyrimidine-DNA glycosylase enhances arsenic-induced
DNA strand breaks in PHA-stimulated and unstimulated human lymphocytes. Environ Health Perspect 109:523-
526.

Li, J; Waters, SB; Drobna, Z; et al. (2005) Arsenic (+3oxidation state) methyltransferase and the inorganic arsenic
methylation phenotype. Toxicol Appl Pharmacol 204(2): 164-169.

Li, J-P; Lin, J-C; Yang, J-L. (2006a) ERK activation in arsenite-treated Gl-enriched CL3 cells contributes to
survival, DNA repair inhibition, and micronucleus formation. Toxicol Sci 89:164-172.

Li, Q; Chen, H; Huang, X; Costa, M. (2006b) Effects of 12 metal ions on iron regulatory protein 1 (IRP-1) and
hypoxia-inducible factor-1 alpha (HIF-la) and HIF-regulated genes. Toxicol Appl Pharmacol 213:245-255.

Li, Y; Chen, Y; Slavkovic, V; et al. (2007) Serum levels of the extracellular domain of the epidermal growth factor
receptor in individuals exposed to arsenic in drinking water in Bangladesh. Biomarkers 12(3):256-265.

Liao, WT; Chang, KL; Yu, CL; et al. (2004) Arsenic induces human keratinocyte apoptosis by the FAS/FAS ligand
pathway, which correlates with alterations in nuclear factor-KB and activator protein-1 activity. J Invest Dermatol
122:125-129.

Liebscher, K; Smith, H. (1968) Essential and nonessential trace elements. A  method of determining whether an
element is essential or nonessential in human tissue. Arch Environ Health 17:881-890.


                                                  169        DRAFT—DO NOT CITE OR QUOTE

-------
Lin, S; Cullen, WR; Thomas, DJ. (1999) Methylarsenicals and arsinothiols are potent inhibitors of mouse liver
thioredoxin reductase. ChemRes Toxicol 12:924-930.

Lin, S; Del Razo, LM; Styblo, M; et al. (2001) Arsenicals inhibit thioredoxin reductase in cultured rat hepatocytes.
ChemRes Toxicol 14(3):305-311.

Lin, S; Shi, Q; Brent, F; et al. (2002) A novel S-adenosyl-L-methionine: arsenic(  ) methyltransferase from rat liver
cytosol. J Biol Chem 277(13): 10795-10803.

Lin, YL; Ho, 1C; Su, PF; Lee, TC. (2006) Arsenite pretreatment enhances the cytotoxicity of mitomycin C in human
cancer cell lines via increased NAD(P)H quinone oxidoreductase 1 expression. Toxicol Appl Pharmacol 214:309-
317.

Lindberg, A-L; Kumar, R; Goessler, W; et al. (2007) Metabolism of low-dose inorganic arsenic in a central
European population: influence of sex and genetic polymorphism. Environ Health Perspect 115:1081-1086.

Lindberg, A-L; Ekstrom, E-C; Nerwell, B; et al. (2008) Gender and age differences in the metabolism of inorganic
arsenic in a highly exposed population in Bangladesh. Environ Res 106:110-120.

Lindgren, A; Vahter, M; Dencker, L. (1982) Autoradiographic studies on the distribution of arsenic in mice and
hamster administered  As-arsenite or -arsenate. Acta Pharmacol Toxicol 51:253-365.

Lindgren, A; Danielsson, BRG; Dencker, L; Vahter, M. (1984) Embryotoxicity of arsenite and arsenate: distribution
in pregnant mice and monkeys and effects on embryonic cells in vitro. Acta Pharmacol Toxicol 54:311-320.

Lindskog,  M; Gleissman, H; Ponthan, F; et al. (2006) Neuroblastoma cell death in response to docosahexaenoic
acid: sensitization to chemotherapy and arsenic-induced oxidative stress. Int J Cancer 118:2584-2593.

Ling, Y-H; Jiang, J-D; Holland, JF; Perez-Soler, R. (2002) Arsenic trioxide produces polymerization of microtubules
and mitotic arrest before apoptosis in human tumor cell lines. Mol Pharmacol 62:529-538.

Liou, SH; Lung, J-C; Chen, YH; et al. (1999) Increased chromosome-type chromosome aberration frequencies as
biomarkers of cancer risk in a blackfoot endemic area. Cancer Res 59:1481-1484.

Liu, SX; Athar, M; Lippai,  I; etal. (2001) Induction of oxyradicals by arsenic: Implication for mechanism of
genotoxicity. Proc Natl Acad Sci USA 98:1643-1648.

Liu, J; Xie, Y; Ward, JM; et al. (2004) Toxicogenomic analysis of aberrant gene expression  in liver tumors and
nontumorous livers of adult mice exposed inutero to inorganic arsenic. Toxicol Sci 77:249-257.

Liu, SX; Davidson, MM; Tang, X; et al. (2005) Mitochondrial damage mediates genotoxicity of arsenic in
mammalian cells. Cancer Res 65:3236-3242.

Liu, Z; Styblo, M; Rosen, BP. (2006a) Methylarsonous acid transport by aquaglyceroporins. Environ Health Persp
114(4):527-531.

Liu, J; Xie, Y; Merrick, BA; et al. (2006b). Transplacental arsenic plus postnatal 12-O-teradecanoyl phorbol-13-
acetate exposures associated with hepatocarcinogenesis induce similar aberrant gene expression patterns in male and
female mouse liver. Toxicol Appl Pharmacol 213:216-223.

Liu, J; Xie, Y; Ducharme, DMK; et al. (2006c) Global gene expression associated with hepatocarcinogenesis in
adult male mice induced by inutero arsenic exposure. Environ Health Persp 114(3):404-411.

Liu, J; Benbrahim-Tallaa, L; Qian, X; et al. (2006d) Further studies on aberrant gene expression associated with
arsenic-induced malignant transformation in rat liver TRL1215 cells. Toxicol Appl Pharmacol 216:407-415.
                                                  170        DRAFT—DO NOT CITE OR QUOTE

-------
Liu, B; Pan, S; Dong, X; et al. (2006e) Opposing effects of arsenic trioxide on hepatocellular carcinomas in mice.
Cancer Sci 97:675-681.

Liu, J; Xie, Y; Cooper, R; et al. (2007a) Transplacental exposure to inorganic arsenic at a hepatocarcinogenic dose
induces fetal gene expression changes in mice indicative of aberrant estrogen signaling and disrupted steroid
metabolism. Toxicol Appl Pharmacol 220:284-291.

Liu, QZ; Jiang, GF; He, Y;  et al. (2007b) Arsenite-induced alterations in Ku70-deficient cells: a model to study
genotoxic effects. J Toxicol Environ Health A 70:938-946.

Lo, MC; Hsen, YC; Lin, BK. (1977) The second report on the investigation of arsenic content in underground water
in Taiwan Providence, Talchung. Provincial Institute of Environmental Sanitation.

Loffredo, CA; Aposhian, HV; Cebrian, ME; et al. (2003) Variability in human metabolism of arsenic. Environ Res
92:85-91.

Lowney, YW; Wester, RC; Schoof, RA; et al. (2007) Dermal absorption of arsenic from soils as measured in the
rhesus monkey. Toxicol Sci 100(2):381-392.

Lu, WJ; Tamai, I; Nezu, J; et al. (2006) Organic anion transporting polypeptide-C mediates arsenic uptake in HEK-
293 cells. J Biomed Sci 13(4):525-533.

Lunghi, P; Tabilio, A; Lo-Coco, F; et al. (2005) Arsenic trioxide (ATO) and MEK1 inhibition synergize to induce
apoptosis in acute promyelocytic leukemia cells. Leukemia 19:234-244.

Lunghi, P; Costanzo, A; Salvatore, L; et al. (2006) MEK1 inhibition sensitizes primary acute myelogenous leukemia
to arsenic trioxide-induced apoptosis. Blood 107:4549-4553.

Luster, MI; Simeonova, PP. (2004) Review. Arsenic and urinary bladder cell proliferation. Toxicol Appl Pharmacol
198:419-423.

Luster, MI; Wilmer, JL; Germolec, DR; et al. (1995) Role of keratinocyte-derived cytokines in chemical toxicology.
Toxicol Lett 82-83:471-476.

Lynn, S; Gurr, JR; Lai HT;  Jan, KY. (2000) NADH oxidase activation is involved in arsenite-induced oxidative
DNA damage in human vascular smooth muscle cells. Circ Res 86:514-519.

Mahata, J; Ghosh, P; Sarkar, JN; et al.  (2004) Effect of sodium arsenite on peripheral lymphocytes in vitro:
individual susceptibility among a population exposed to arsenic through the drinking water. Mutagenesis 19(3):223-
229.

Maier, A; Dalton TP; Puga, A.  (2000) Disruption of dioxin-inducible phase I and phase II gene expression patterns
by cadmium, chromium, and arsenic. Mol Carcinog 28:225-235.

Maiorino, RM; Aposhian, HV.  (1985) Dimercaptan metal-binding agents influence the biotransformation of arsenite
in the rabbit. Toxicol Appl Pharmacol 77:240-250.

Mandal, BK; Ogra, Y; Anzai, K; Suzuki, KT. (2004) Speciation of arsenic in biological samples. Toxicol Appl
Pharmacol 198:307-318.

Mann, S; Droz, PO; Vahter, M. (1996a) A physiologically based pharmacokinetic model for arsenic exposure: I.
development in hamster and rabbits. Toxicol Appl Pharmacol 137:8-22.

Mann, S; Droz, PO; Vahter, M. (1996b) A physiologically based pharmacokinetic model for arsenic exposure: II.
validation and application in humans. Toxicol Appl Pharmacol 140:471^86.

Mappes, R. (1977) Experiments on the excretion of arsenic in urine. Int Arch Occup Environ Health 40:267-272.

                                                  171        DRAFT—DO NOT CITE OR QUOTE

-------
Marafante, E; Vahter, M. (1984) The effect of methyltransferase inhibition on the metabolism of (  As)arsenite in
mice and rabbits. Chem-Biol Interact 50:49-57.

Marafante, E; Vahter, M. (1987) Solubility, retention, and metabolism of intratracheally and orally administered
inorganic arsenic compounds in the hamster. Environ Res 42:72-82.

Marafante, E; Rade, J; Sabbioni, E; et al. (1981) Intracellular interaction and metabolic fate of arsenite in the rabbit.
ClinToxicol 18:1335-1341.

Marafante, E; Vahter, M; Envall, J. (1985) The role of methylation in the detoxication of arsenate in the rabbit.
Chem-Biol Interact 56:225-238.

Marafante, E; Vahter, M; Norin, H; et al. (1987). Biotransformation of dimethylarsinic acid in mouse, hamster, and
man. J Appl Toxicol 7:111-117.

Marcus, WL; Rispin, AS. (1988) Threshold carcinogenicity using arsenic as an example. In: Cothern, CR; Mehlman,
MA; eds. Advances in modern environmental toxicology. Vol. XV. Risk assessment and risk management of
industrial and environmental chemicals. Princeton, NJ: Princeton Scientific Publishing Co., Inc.; pp. 133-158.

Marshall, G; Ferreccio, C; Yuan, Y; et al. (2007) Fifty-year study of lung and bladder cancer mortality in Chile
related to arsenic in drinking water. J Natl Can Inst 99(12):920-928.

Marsit, CJ; Karagas, MR; Danaee, H; et al.  (2006) Carcinogen exposure and gene promoter hypermethylation in
bladder cancer. Carcinogenesis 27:112-116.

Mass, MJ; Wang, L. (1997) Arsenic alters cytosine methylation patterns of the promoter of the tumor suppressor
gene p53 in human lung cells: a model for mechanism of carcinogenesis. Mutat Res 386:263-277.

Mass, MJ; Tennant, A; Roop, BC; et al. (2001) Methylated trivalent arsenic species are genotoxic. Chem Res
Toxicol  14:355-361.

Massrieh, W; Derjuga, A; Blank, V. (2006) Induction of endogenous Nrf2/small Maf heterodimers by arsenic-
mediated stress in placental choriocarcinoma cells. Antioxid Redox Signal 8:53-59.

Mazumder, DNG; Haque, R; Ghosh, N; et al. (1998) Arsenic levels in drinking water and the prevalence of skin
lesions in West Bengal, India. Int J Epidemiol 27:871-877.

Mazumder, DNG. (2005) Effect of chronic intake of arsenic-contaminated water on liver. Toxicol Appl Pharmacol
206:169-175.

McCarthy, KM; Chen, Y-C; Quamruzzaman, Q; et al. (2007a) Arsenic methylation, GSSTT1, GSTM1, GSTP1
polymorphisms, and skin lesions. Environ Health Perspect 115:341-345.

McCarthy, KM; Ryan, L; Houseman, EA; et al. (2007b) A case-control  study of GST polymorphisms and arsenic
related skin lesions. Environ Health 6:5.

McCollum, G; Keng, PC; States, JC; McCabe, MC Jr. (2005) Arsenite delays progression through each cell cycle
phase and induces apoptosis following G2/M arrest in U937 myeloid leukemia cells. J Pharmacol Exp Ther
313:877-887.

McNeely, SC; Xu, X; Taylor, BF; et al. (2006) Exit from arsenite-induced mitotic arrest is p53 dependent. Environ
Health Perspect 114:1401-1406.

Mealey, J; Brownwell, GL; Sweet, WH. (1959) Radioarsenic in plasma, urine, normal tissues, and intracranial
neoplasms. Distribution and turnover after intravenous injection in man. AMA Arch Neurol Psychiatry 81(3):310-
320.

                                                  172        DRAFT—DO NOT CITE OR QUOTE

-------
Meliker, JR; Wahl, RL; Cameron, LL; Nriagu, J. (2007) Arsenic in drinking water and cerebrovascular disease,
diabetes mellitus, and kidney disease in Michigan: a standardized mortality ratio analysis. Environ Health 6:4. Page
number not for citation purposes.

Meng, Z; Hsie, AW. (1996) Polymerase chain reaction-based deletion analysis of spontaneous and arsenite-
enhanced gpt mutants in CHO-AS52 cells. Mutat Res 356:255-259.

Merck Index (1989) Budavari, S., M.J. O'Neil, A. Smith and P.E. Heckelman, eds. The Merck Index—an
encyclopedia of chemicals, drugs, and biologicals, llth Ed. Rahway, NJ: Merck & Co., Inc.; pp.126-127, 8515.

Meza, MM; Yu, L; Rodriguez, YY; et al. (2005) Developmentally restricted genetic determinants of human arsenic
metabolism: Association between urinary methylated arsenic and CYT19 polymorphisms in children. Environ
Health Persp 113(6):775-781.

Meza, M; Gandolfi, AJ; Klimecki, WT. (2007) Developmental and genetic modulation of arsenic biotransformation.
A gene by environment interaction?  Toxicol Appl Pharmacol 222:381-387.

Michaud, DS; Wright, ME; Cantor, KP; et al. (2004) Arsenic concentrations in prediagnostic toenails and the risk of
bladder cancer in a cohort study of male smokers. Am J Epidemiol 160(9):853-859.

Mitra, SR; Guha Mazumder, DN; Basu, A; et al. (2004) Nutritional factors and susceptibility to arsenic-caused skin
lesions in West Bengal, India. Environ Health Perspect  112:1104-1109.

Mittal, M; Flora, SJS. (2006) Effects of individual and combined exposure to sodium arsenite and sodium fluoride
on tissue oxidative stress, arsenic and fluoride levels in male mice. Chem Biol Interact 162:128-139.

Miyazaki, K; Watanabe, C; Mori, K; et al. (2005) The effects of gestational arsenic exposure and dietary selenium
deficiency on selenium and selenoenzymes in maternal and fetal tissues in mice. Toxicology 208:357-365.

Mizoi, M; Takabayashi, F; Nakano, M; et al. (2005) The role of trivalent dimethylated arsenic in dimethylarsinic
acid-promoted skin and lung tumorigenesis in mice: tumor-promoting action through the induction of oxidative
stress. Toxicol Lett 158:87-94.

Mo, J; Xia, Y; Wade, TJ; et al. (2006) Chronic arsenic exposure and oxidative stress: OGG1 expression and arsenic
exposure, nail selenium, and skin hyperkeratosis in inner Mongolia. Environ Health Persp 114(6):835-841.

Modi, M; Kaul, RK; Kannan, GM; Flora, SJ. (2006) Co-administration of zinc and n-acetylcysteine prevents
arsenic-induced tissue oxidative stress in male rats. J Trace Elem Med Biol 20:197-204.

Molin, L; Wester, PO. (1976) The estimated daily loss of trace elements from normal skin by desquamation. Scand J
ClinLab Invest 36:679-682.

Mondal, TK; Li, D; Swami, K; et al. (2005) Mercury impairment of mouse thymocyte survival in vitro: involvement
of cellular thiols. J Toxicol Environ Health A 68:535-556.

Moore, LE; Warner, ML; Smith, AH; et al. (1996).  Use of the fluorescent micronucleus assay to detect the genotoxic
effects of radiation and arsenic exposure in exfoliated human epithelial cells.  Environ Mol Mutagen 27:176-184.

Moore, MM; Harrington-Brock, K; Doerr, CL. (1997a) Relative genotoxic potency of arsenic and its methylated
metabolites. Mutat Res 386:279-290.

Moore, LE; Smith, AH; Hopenhayn-Rich, C; et al. (1997b) Micronuclei in exfoliated bladder cells among
individuals chronically exposed to arsenic in drinking water. Cancer Epidemiol Biomarkers Prev 6(1):31-36.

Moore, LE; Lu, M; Smith, AH. (2002) Childhood cancer incidence and arsenic exposure in drinking water in
Nevada. Arch Environ Health 57(3):201-206.

                                                  173       DRAFT—DO NOT CITE OR QUOTE

-------
Morales, KH; Ryan, L; Kuo, T-L; et al. (2000) Risk of internal cancers from arsenic in drinking water. Environ
Health Perspect 108:655-661.

Moser, G; Goldsworthy, T; Tice, R. (2000). Sodium arsenite studies inp53 +/- mice. Denver, CO; AWWA Research
Foundation and the American Water Works Association.

Motiwale, L; Ingle, AD; Rao, KVK. (2005) Mouse skin tumor promotion by sodium arsenate is associated with
enhanced PCNA expression. Cancer Lett 223:27-35.

Mouron, SA; Grille, CA; Dulout, FN; Golijow, CD. (2005) DNA-protein cross-links and sister chromatid exchanges
induced by dimethylarsinic acid in human fibroblasts cells. MutatRes 581:83-90.

Mouron, SA; Grillo, CA; Dulout, FN; Golijow, CD. (2006) Induction of DNA strand breaks, DNA-protein
crosslinks and sister chromatid exchanges by arsenite in a human lung cell line. Toxicol In Vitro 20:279-285.

Mudipalli, A; Owen, RD; Preston, RJ. (2005) The effect of arsenicals on ultraviolet-radiation-induced growth arrest
related signaling events in human keratinocytes. Int J Oncol 27:769-778.

Mure, K; Uddin, A; Lopez, LC; et al. (2003) Arsenite induces delayed mutagenesis and transformation in human
osteosarcoma cells at extremely low concentrations. Environ MolMutagen 41:322-331.

Namgung, U; Xia, Z. (2001) Arsenic induces apoptosis in rat cerebellar neurons via activation of JNK3 and p38
MAP kinases. Toxicol Appl Pharmacol 174:130-138.

Nandi, D; Patra, RC; Swamp, D. (2005) Effect of cysteine, methionine, ascorbic acid, and thiamine on arsenic-
induced oxidative stress and biochemical alterations in rats. Toxicology 211:26-3 5.

Nandi, D; Patra, RC; Swamp, D. (2006) Oxidative stress indices and plasma biochemical parameters during oral
exposure to arsenic in rats. Food Chem Toxicol 44:1579-1584.

Naranmandura, H; Suzuki, N; Iwata, K; et al. (2007) Arsenic metabolism and thioarsenicals in hamsters and rats.
Chem Res Toxicol 20:616-624.

NCI (National Cancer Institute). (2006) U.S. population data 2000-2003. Surveillance, epidemiology, and end
results (SEER) program populations (1969-2003) DCCPS, Surveillance Research Program, Cancer Statistics
Branch, released April 2006. Available online at www.seer.cancer.gov/popdata.

NCHS (National Center for Health Statistics). (2000) Vital statistics of the United States.

NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process.
National Academy Press, Washington, DC.  Available online at
http://www.nap.edu:80/openbook.php?isbn=0309033497.

NRC (National Research Council). (1999) Arsenic in drinking water. National Academy Press, Washington, DC.
Available online at http://www.nap.edu/openbook/0309063337/html/Rl.html.

NRC (National Research Council). (2001) Arsenic in drinking water (2001 update). National Academy Press,
Washington, DC. Available online at http://www.nap.edu/openbook/0309076293/html/Rl.html.

Nayak, AS; Lage, CR; Kim, CH. (2007) Effects of low concentrations of arsenic on the innate immune system of the
zebrafish (Danio rerio). Toxicol Sci 98(1): 118-124.

Nemeti, B; Gregus, Z. (2002) Reduction of arsenate to arsenite in hepatic cytosol. Toxicol Sci 70:4-12.

Nemeti, B; Gregus, Z. (2004) Glutathione-dependent reduction of arsenate in human erythrocytes: a process
independent of purine nucleoside phosphorylase. Toxicol Sci 82(2):419-428.
                                                  174        DRAFT—DO NOT CITE OR QUOTE

-------
Nemeti, B; Gregus, Z. (2005) Reduction of arsenate to arsenite by human erythrocyte lysate and rat liver cytosol:
characterization of a glutathione- and NAD-dependent arsenate reduction linked to glycolysis. Toxicol Sci
85(2):847-858.

Nemeti, B; Csanaky, I; Gregus, Z. (2003) Arsenate reduction in human erythrocytes and rats: testing the role of
purine nucleoside phosphorylase. Toxicol Sci 74:22-31.

Nesnow, S; Roop, BC; Lambert, G; et al. (2002) DNA damage induced by methylated trivalent arsenicals is
mediated by reactive oxygen species. ChemRes Toxicol 15:1627-1634.

Ng, J; Seawright, S; Wi, L; et al. (1999) Tumours in mice induced by exposure to sodium arsenate in drinking water.
In: Chappell, WR; Abernathy, CO; Cothern, CR; eds. Arsenic: exposure  and health. Northwood, UK:  Sci Tech Lett;
pp. 217-223.

Nordenson, I; Salmonsson, S; Bran, E; et al. (1979) Chromosome aberrations inpsoriatic patients treated with
arsenic. Hum Genet 48:1-6.

Nordenson, I; Beckman, L. (1991) Is the genotoxic effect of arsenic mediated by oxygen free radicals? Hum Hered
41:71-73.

Nutt, LK; Gogvadze, V; Uthaisang, W; et al. (2005) Indirect effects of Bax and Bak initiate the mitochondria!
alterations that lead to cytochrome c release during arsenic trioxide-induced apoptosis. Cancer Biol Ther 4:459-467.

O'Brien, TG; Megosh, LC; Gillard, G; Soler, AP. (1997) Ornithine decarboxylase over expression is a sufficient
condition for tumor promotion in mouse skin. Cancer Res 57:2630-2637.

Ochi, T; Kaise, T; Oya-Ohta, Y. (1994) Glutathione plays different roles in the induction of the cytotoxic effects of
inorganic and organic arsenic compounds in cultured BALB/c 3T3  cells. Experientia 50:115-120.

Ochi, T; Nakajima, F; Nasui, M. (1999a) Distribution of gamma-tubulin  in multipolar spindles and multinucleated
cells induced by dimethylarsinic acid, a methylated derivative of inorganic arsenics, in Chinese hamster V79 cells.
Toxicology 136:79-88.

Ochi, T; Nakajima, F; Shimizu, A; Harada, M. (1999b) Induction of multinucleated cells in V79 Chinese hamster
cells exposed to dimethylarsenic acid, a methylated derivative of inorganic arsenics: Mechanism associated with
aberrant mitotic spindles. Toxicol In Vitro 13:11-25.

Ochi, T; Suzuki, T; Barrett, JC; Tsutsui, T.  (2004) A trivalent dimethylarsenic compound, dimethylarsine iodide,
induces cellular transformation, aneuploidy, centrosome abnormality, and multipolar spindle formation in Syrian
hamster embryo cells. Toxicol 203:155-163.

Odanaka, Y; Matano, O; Goto, S. (1980) Biomethylation of inorganic arsenic by the rat and some laboratory
animals. Bull Environ Contam Toxicol 24:452-459

O'Hara, KA; Nemec, AA; Alam, J; et al. (2006) Chromium (VI) inhibits  heme oxygenase-1 expression in vivo and in
arsenic-exposed human airway epithelial cells. J Cell Physiol 209:113-121.

Ohno, H; Hanaoka, F; Yamada, M. (1982) Inducibility of sister chromatid exchanges by heavy-metal ions. Mutat
Res 104(1-3): 141-146.

Ostrosky-Wegman, P; Gonsebatt, ME; Montero, R; et al. (1991) Lymphocyte proliferation kinetics and genotoxic
findings in a pilot study on individuals chronically exposed to arsenic in  Mexico. Mutat Res 250:477-482.

Othumpangat,  S; Kashon, M; Joseph, P. (2005) Sodium arsenite-induced inhibition of eukaryotic translation
initiation factor 4E (eIF4E) results in cytotoxicity and cell death. Mol Cell Biochem 279:123-131.


                                                  175        DRAFT—DO NOT CITE  OR QUOTE

-------
Ouyang, W; Ma, Q; Li, J; et al. (2005) Cyclin Dl induction through M3 Kinase P/Nuclear Factor-KB pathway is
responsible for arsenite-induced increased cell cycle Gl-S phase transition in human keratinocytes. Cancer Res
65(20):9287-9293.

Ouyang, W; Li, J; Ma, Q; Huang, C. (2006) Essential roles of PI-SK/Akt/IKKp/NFicB pathway in cyclinDl
induction by arsenite in JB6 CI41 cells. Carcinogenesis 27:864-873.

Ouyang, W; Zhang, D; Ma, Q; et al. (2007) Cyclooxygenase-2 induction by arsenite through the IKK(3/NFKB
pathway exerts an antiapoptotic effect in mouse epidermal C141 cells. Environ Health Perspect 115:513-518.

Oya-Ohta, Y; Kaise, T; Ochi, T. (1996) Induction of chromosomal aberrations in cultured human fibroblasts by
inorganic and organic arsenic compounds and the different roles of glutathione in such induction. Mutat Res
375:123-129.

Palus, J; Lewinska, D; Dziubaltowska, E; et al. (2006) Genotoxic effects in C57B1/6J mice chronically exposed to
arsenate in drinking water and modulation of the effects by low-selenium diet. J Toxicol Environ Health A 69:1843-
1860.

Park, WH; Seol, JG; Kim, ES; et al. (2000) Arsenic trioxide-mediated growth inhibition in MC/CAR myeloma cells
via cell cycle arrest in association with induction of cyclin-dependent kinase inhibitor, p21, and apoptosis. Cancer
Res 60:3065-3071.

Parr, RM; De Maeyer, EM; lyengar, VG; et al. (1991) Minor and trace elements in human milk from Guatemala,
Hungary, Nigeria, Philippines, Sweden, and Zaire: results from a WHO/IAEA joint project. Biol Trace Element Res
29:51-75.

Partridge, MA; Huang, SX; Hernandez-Rosa, E; et al. (2007) Arsenic induced mitochondrial DNA damage and
altered mitochondrial oxidative function: implications for genotoxic  mechanisms in mammalian cells. Cancer Res
67:5239-5247.

Patra, M; Haider, A; Bhowmik, N; De, M. (2005) Use of black tea in modulating clastogenic effects of arsenic in
mice in vivo. J Environ Pathol Toxicol Oncol 24:201-210.

Patterson, TJ; Rice, RH. (2007) Arsenite and insulin exhibit opposing effects on epidermal growth factor receptor
and keratinocyte proliferative potential. Toxicol Appl Pharmacol 221:119-128.

Patterson, TJ; Reznikova, TV; Phillips, MA; Rice, RH. (2005) Arsenite maintains germinative state in cultured
human epidermal cells. Toxicol Appl Pharmacol 207(l):69-77.

Peraza, MA; Cromey, DW; Carolus, B; et al. (2006) Morphological  and functional alterations in human proximal
tubular cell line induced by low level inorganic arsenic: evidence for targeting of mitochondria and initiated
apoptosis. J Appl Toxicol 26:356-367.

Perez-Pasten, R; Martinez-Galero, E; Garduno-Siciliano, L; et al. (2006) Effects of dimethylsulphoxide on mice
arsenite-induced dysmorphogenesis in embryo culture and cytotoxicity in embryo cells. Toxicol Lett 161:167-173.

Perkins, C; Kim, CN; Fang,  G; Bhalla, KN. (2000) Arsenic induces  apoptosis of multi-drug-resistant human myeloid
leukemia cells that express Ber-Abl or over express MDR, Bel-2, or Bel-x(L). Blood 95:1014-1022.

Petrick, JS; Ayala-Fierro, F; Cullen, WR; et al. (2000) Monomethylarsonous acid (MMA  ) is more toxic than
arsenite in Chang human hepatocytes. Toxicol Appl  Pharmacol 163:203-207.

Petrick, JS; Jagadish, B; Mash, EA; Aposhian, HV. (2001). Monomethylarsonous acid (MMA  ) and arsenite: LD50
in hamsters and in vitro inhibition of pyruvate dehydrogenase. Chem Res Toxicol 14:651-656.
                                                  176        DRAFT—DO NOT CITE OR QUOTE

-------
Pi, J; He, Y; Bornter, C; et al. (2005) Low level, long- term inorganic arsenite exposure causes generalized
resistance to apoptosis in cultured human keratinocytes: potential role in skin co-carcinogenesis. Int J Cancer
116:20-26.

Piga, R; Saito, Y; Yoshida, Y; Niki, E. (2007) Cytotoxic effects of various stressors onPC12 cells: involvement of
oxidative stress and effect of antioxidants. Neurotoxicology 28:67-75.

Pomroy, C.; Charbonneau, SM; McCullough, RS; Tarn, GKH (1980) Human retention studies with 74As. Toxicol
Appl Pharmacol 53:550-556.

Poonepalli, A; Balakrishnan, L; Khaw, AK; et al. (2005) Lack of Poly(ADP-Ribose) Polymerase-1 gene product
enhances cellular sensitivity to arsenite. Cancer Res 65:10977-10983.

Preston, PJ; Williams, GM. (2005) DNA-reactive carcinogens: mode of action and human cancer hazard. Crit Rev
Toxicol 35:673-683.

Pu, YS; Jan, KY; Wang, TC; et al. (2007) 8-Oxoguanine DNA glycosylase and MutY homolog are involved in the
incision of arsenite-induced DNA adducts. Toxicol Sci 95:376-382.

Qian, Y; Liu, KJ; Chen, Y; et al. (2005) Cdc42 regulates arsenic-induced NADPH oxidase activation and cell
migration through actin filament reorganization. J Biol Chem 280:3875-3884.

Radabaugh, TR; Aposhian, HV. (2000) Enzymatic reduction of arsenic compounds in mammalian systems:
reduction of arsenate to arsenite by human liver arsenate reductase. Chem Res Toxicol 13(1)26-30.

Radabaugh, TR; Sampayo-Reyes, A; Zakharyan, RA; Aposhian, HV. (2002) Arsenate reductase II. Purine
nucleoside phosphorylase in the presence of dihydrolipoic acid as a route for reduction of arsenate to arsenite in
mammalian systems. Chem Res Toxicol 15(5):692-698.

Rahman, MS; Hall, LL; Hughes, MF. (1994) In vitro percutaneous absorption of sodium arsenate in B6C3F1 mice.
Toxicol In Vitro 8:441-448.

Ramanathan, K; Anusuyadevi, M; Shila, S; Panneerselvam, C. (2005) Ascorbic acid and a-tocopherol as potent
modulators of apoptosis on arsenic induced toxicity in rats. Toxicol Lett 156:297-306.

Ramirez, P; Eastmond, DA; Laclette, JP;  Ostrosky-Wegman, P. (1997) Disruption of microtubule assembly and
spindle formation for the  induction of aneuploid cells by sodium arsenite and vanadium pentoxide. Mutat Res
386:291-298.

Ramirez, P; Del Razo, LM; Gutierrez-Ruiz, MC; Gonsebatt, ME. (2000) Arsenite induces DNA-protein crosslinks
and cytokeratin expression in the  WRL-68 human hepatic cell line. Carcinogenesis 21:701-706.

Ramirez, T; Stopper, H; Hock, R; Herrera, LA. (2007) Prevention of aneuploidy by  S-adenosyl-methionine  in human
cells treated with sodium arsenite. Mutat Res 617:16-22.

Raml, A; Goessler, W; Francesconi, KA.  (2006) Improved chromatographic separation of thio-arsenic compounds
by reverse-phase high performance liquid chromatography-inductively coupled plasma mass spectrometry (HPLC-
ICPMS). J Chromatogr A 1128:164-170.

Raml, R; Rumpler, A; Goessler, W; et al. (2007) Thio-dimethylarsenite is a common metabolite in urine samples
from arsenic-exposed women in Bangladesh. Toxicol Appl Pharmacol 222:374-380.

Ramos, AM; Fernandez, C; Amran, D; et al. (2005) Pharmacologic inhibitors of PI3K/Akt potentiate the apoptotic
action of the antileukemic drug arsenic trioxide via glutathione depletion and increased peroxide accumulation in
myeloid leukemia cells. Blood 105:4013-4020.
                                                  177        DRAFT—DO NOT CITE OR QUOTE

-------
Rana, SVS; Allen, T. (2006) Influence of thyroxine and n-propylthiouracil on nephro-toxicity of inorganic arsenic in
rat. Toxicol Ind Health 22:137-145.

Rasmussen, RE; Menzel, DB. (1997) Variation in arsenic-induced sister chromatid exchange in human lymphocytes
and lymphoblastoid cell lines. Mutat Res 386:299-306.

Reichard, JF; Schnekenburger, M; Puga, A. (2007) Long-term low-dose arsenic exposure induces loss of DNA
methylation. Biochem Biophys Res Commun 352:188-192.

Rizki, M; Kossatz, E; Velazquez, A; et al. (2006) Metabolism of arsenic in Drosophila melanogaster and the
genotoxicity of dimethylarsinic acid in the Drosophila wing spot test. Environ Mol Mutagen 47:162-168.

Rossman, TG. (1981) Enhancement of UV-mutagenesis by low concentrations of arsenite inE. coli. Mutat Res
91:207-211.

Rossman, TG. (1998) Molecular and genetic toxicology of arsenic. In: Rose, J, ed. Environmental toxicology:
current developments. Amsterdam, NL: Gorden and Breach Science Publishers; pp. 171-187.

Rossman, TG. (2003) Review: Mechanism of arsenic carcinogenesis: an integrated approach. Mutat Res 533:37-65.

Rossman, TG; Wolosin, D. (1992) Differential susceptibility to carcinogen-induced amplification of SV40 and dhfr
sequences in SV40-transformed humankeratinocytes. Mol Carcinog 6:203-213.

Rossman, TG; Stone, D; Molina, M; Troll, W. (1980) Absence of arsenite mutagenicity inE. coli and Chinese
hamster cells. Environ Mutagen 2:371-379.

Rossman, TG; Molina, M; Meyer, LW. (1984) The genetic toxicology of metal compounds: I. Induction of I
prophage inE.coli WP2s (I). Environ Mutagen 6:59-69.

Rossman, TG; Uddin, AN; Burns, FJ; et al. (2001) Arsenic is a co-carcinogen with solar ultraviolet radiation for
mouse skin: an animal model for arsenic carcinogenesis. Toxicol Appl Pharmacol 176:64-71.

Rousselot, P; Labaume, S; Marolleau, J; et al. (1999) Arsenic trioxide and melarsoprol induce apoptosis in plasma
cell lines and in plasma cells from myeloma patients. Cancer Res 59:1041-1048.

Saady, JJ; Blanke, RV; Poklis, A. (1989) Estimation of the body burden of arsenic in a child fatally poisoned by
arsenite weedkiller. J Anal Toxicol 13:310-312.

SAB (Science Advisory Board). (2000) Arsenic proposed drinking water regulation: a U.S. EPA Science Advisory
Board review of certain elements  of the proposal: a report by the U.S. EPA Science Advisory Board. December
2000. EPA/SAB/DWC/01/001.

SAB (Science Advisory Board). (2007) Advisory onEPA's asssessments of carcinogenic effects of organic and
inorganic arsenic: a report of the U.S. EPA Science Advisory Board. June 2007. Available online at
http://yosemite.epa.gov/sab/SABPRODUCT.NSF/EAD ABBF40DED2A0885257308006741EF/$File/sab-07-
008.pdf.

Sahu, GR; Jena, RK. (2005) Significance of intracellular arsenic trioxide for therapeutic response in acute
promyelocytic leukemia. Am J Hematol 78:113-116.

Sakurai, T; Kaise, T; Matsubara, C. (1998) Inorganic and methylated arsenic compounds induce cell death in murine
macrophages via different mechanisms. Chem Res Toxicol 11:273-283.

Sakurai, T; Ochiai, M; Kojima, C; et al. (2005a). Preventive mechanism of cellular glutathione in
monomethylarsonic acid-induced cytolethality.  Toxicol Appl Pharmacol 206:54-65.
                                                 178        DRAFT—DO NOT CITE OR QUOTE

-------
Sakurai, T; Ohta, T; Fujiwara, K. (2005b) Inorganic arsenite alters macrophage generation from human peripheral
blood monocytes. Toxicol Appl Pharmacol 203:145-153.

Sakurai, T; Ohta, T; Tomita, N; et al. (2006) Evaluation of immunotoxic and immunodisruptive effects of inorganic
arsenite on human monocytes/macrophages.  Int Immunopharmacol 6:304-315.

Salazar, AM; Ostrosky, WP; Menendez, D; et al. (1997) Induction of p53 protein expression by sodium arsenite.
Mutat Res 381:259-265.

Sampayo-Reyes, A; Zakharyan, RA; Healy,  SM; Aposhian, HV. (2000) Monomethylarsonic acid reductase and
monomethylarsonic acid in hamster tissue. Chem Res Toxicol 13(11): 1181-1186.

Sams, RII; Wolf, DC; Ramasamy S; Ohanian E; Chen J; Lowit A. (2007) Workshop overview: Arsenic research and
risk assessment. Toxicol Appl Pharmacol 222: 245-251.

Samuel, S; Kathirvel, R; Jayavelu, T; Chinnakkannu, P. (2005) Protein oxidative damage in arsenic induced rat
brain: influence of DL-a-lipoic acid. Toxicol Lett 155:27-34.

Sasaki, A; Oshima, Y; Fujimura, A. (2007) An approach to elucidate potential mechanism of renal toxicity of arsenic
trioxide. Exp Hematol 35:252-262.

Schoof, RA; Yost, LJ; Crecelius, C; et al. (1998) Dietary arsenic intake in Taiwanese districts with elevated arsenic
in drinking water. Hum Ecol Risk Assess 4:117-135.

Sciandrello,  G; Barbara, R; Caradonna, F; et al. (2002) Early induction of genetic instability and apoptosis by
arsenic in cultured Chinese hamster cells. Mutagenesis 17:99-103.

Sciandrello,  G; Caradonna, F; Mauro, M; Barbata, G. (2004) Arsenite-induced DNA hypomethylation affects
chromosomal instability in mammalian cells. Carcinogenesis 25(3):413^U7.

Scott, N;  Hatlelid, KM; MacKenzie, NE; Carter,  DE. (1993) Reactions of arsenic(  ) and arsenic(V) species with
glutathione.  Chem Res Toxicol 6:102-106.

Scott, A.  (1958) The retention of arsenic  in the late cutaneous complications of its administration. Br J Dermatol
70:195-200.

Sen, CK; Parker, L. (1996) Antioxidant and redox regulation of gene transcription. FASEB J 10:709-720.

Sen, B; Wang, A; Hester, SD; et al. (2005) Gene expression profiling of responses to dimethylarsinic acid in female
F344 rat urothelium. Toxicology 215:214-226.

Seol, JG; Park, WH; Kim, ES; et al. (1999) Effect of arsenic trioxide on cell cycle arrest in head and neck cancer cell
line PCI-1. Biochem Biophys Res Commun 265:400-404.

Shackelford, RE; Kaufmann, WK; Paules, RS. (2000) Oxidative stress and cell cycle checkpoint function. Free
Radic Biol Med 28:1387-1404.

Shen, J; Wanibuchi, H; Salim, El; et al. (2003) Liver tumorigenicity of trimethylarsine oxide in male Fischer 344
rats-association with oxidative DNA damage and enhanced cell proliferation. Carcinogen 24(11): 1827-1835.

Shen, S; Lee, J;  Sun, X; et al. (2006) Elevation of cellular BPDE uptake by human cells: a possible factor
contributing to co-carcinogenicity by arsenite. Environ Health Perspect 114:1832-1837.

Shen, J; Liu, J; Xie, Y; et al. (2007) Fetal onset of aberrant gene expression relevant to pulmonary carcinogenesis in
lung adenocarcinoma development induced by inutero arsenic exposure. Toxicol Sci 95:313-320.
                                                  179         DRAFT—DO NOT CITE OR QUOTE

-------
Shila, S; Subathra, M; Devi, MA; Panneerselvam, C. (2005a) Arsenic intoxication-induced reduction of glutathione
level and of the activity of related enzymes in rat brain regions: reversal by DL-a-lipoic acid. Arch Toxicol 79:140-
146.

Shila, S; Kokilavani, V; Subathra, M; Panneerselvam, C. (2005b). Brain regional responses in antioxidant system to
a-lipoic acid in arsenic intoxicated rat. Toxicol 210:25-36.

Shinkai, Y; Sumi, D; Fukami, I; et al. (2006) Sulforaphane, an activator of Nrf2, suppresses cellular accumulation of
arsenic and its cytotoxicity in primary mouse hepatocytes. FEES Lett 580:1771-1774.

Shiobara Y; Ogra, Y; Suzuki, KT. (2001) Animal species difference in the uptake of dimethylarsinous acid
(DMA( )) by red blood cells. ChemRes Toxicol 14(10): 1446-1452.

Simeonova, PP; Wang, S; Toriuma, W; et al. (2000) Arsenic mediates cell proliferation and gene expression in the
bladder epithelium: association with activating protein-1 transactivation. Cancer Res 60:3445-3453.

Simeonova, PP; Luster MI. (2000) Mechanisms of arsenic carcinogenicity: genetic orepigenetic mechanisms?
Environ Pathol Toxicol Oncol 19:281-286.

Singh, KP; DuMond, JW, Jr. (2007) Genetic and epigenetic changes induced by chronic low  dose exposure to
arsenic of mouse testicular Leydig cells. January 2007. Int J Oncol 30(1):253-260.

Sinha, D; Bhattacharya, RK; Siddiqi, M; Roy, M. (2005a) Amelioration of sodium arsenite-induced clastogenicity
by tea extracts in Chinese hamster V79 cells. J Environ Pathol Toxicol Oncol 24:129-140.

Sinha, D; Roy, M; Siddiqi, M; Bhattacharya, RK. (2005b) Arsenic-induced micronuclei formation in mammalian
cells and its counteraction by tea. J Environ Pathol Toxicol Oncol 24:45-56.

Smith, AH; Goycolea, M; Haque, R; Biggs, ML. (1998) Marked increase in bladder and lung cancer mortality in a
region of Northern Chile due to arsenic in drinking water. Am J Epidemiol 147(7):660-669.

Smith, KR; Klei, LR; Barchowsky, A.  (2001) Arsenite stimulates plasma membrane NADPH oxidase in vascular
endothelial cells. Am J Physiol Lung Cell Mol Physiol 280:L442-L449.

Smith, AH; Marshall, G; Yuan, T; et al. (2006) Increased mortality from lung cancer and bronchiectasis in young
adults following exposure to arsenic in utero and early childhood. Environ Health Perspect 114(8): 1-4.

Snow, ET; Hu, Y; Yan, CC; Chouchane, S. (1999) Modulation of DNA repair and glutathione levels in human
keratinocytes by micromolar arsenite. In: Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic exposure and
health effects. Oxford, UK: Elsevier; pp. 243-251.

Snow, ET; Schuliga, M; Chouchane, S; Hu, Y. (2001) Sub-toxic arsenite induces a multi-component protective
response against oxidative stress in human cells. Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic
exposure and health effects.  Amsterdam, NL: Elsevier; pp. 265-276.

Snow, ET; Sykora, P; Durham, TR; Klein, CB. (2005) Arsenic, mode of action at biologically plausible low doses:
what are the implications for low dose cancer risk? Toxicol Appl Pharmacol 207:8557-8564.

Soffritti, M; Belpoggi, F; Degli, D; et al. (2006) Results of a long-term carcinogenicity bioassay on Sprague-Dawley
rats exposed to sodium arsenite administered in drinking water. Ann NY Acad Sci 1076:578-591.

Sohini; Rana, SV. (2007) Protective effect of ascorbic acid against oxidative  stress induced by inorganic arsenic in
liver and kidney of rat. Indian J Exp Biol 45:371-375.

Soto-Reyes, E; del Razo, LM; Valverde, M; Rojas, E. (2005) Role of the alkali labile sites, reactive oxygen species
and antioxidants in DNA damage induced by methylated trivalent metabolites of inorganic arsenic. BioMetals
18:493-506.

                                                  180        DRAFT—DO NOT CITE OR QUOTE

-------
Soucy, NV; Ihnat, MA; Kamat, CD; et al. (2003) Arsenic stimulates angiogenesis and tumorigenesis in vivo. Toxicol
Sci 76:271-279.

Soucy, NV; Mayka, D; Klei, LR; et al. (2005) Neovascularization and angiogenic gene expression following chronic
arsenic exposure in mice. Cardiovasc Toxicol 5:29-41.

Steinmaus, C; Yuan, Y; Bates, MN; Smith AH. (2003) Case-control study of bladder cancer and drinking water
arsenic in the western United States. Am J Epidemiol 158(12): 1193-1201.

Steinmaus, C; Carrigan, K; Kalman, D; et al. (2005) Dietary intake and arsenic methylation in a U.S. population.
Environ Health Persp 113(9): 1153-1159.

Steinmaus, C; Bates, MN; Yuan, Y; et al.  (2006) Arsenic methylation and bladder cancer risk in case-control studies
in Argentina and the United States. J Occup Environ Med 48:478-488.

Steinmaus, C; Moore, LE; Shipp, M; et al. (2007) Genetic polymorphisms inMTHFR 677 and 1298, GSTM1 and
Tl, and metabolism of arsenic. J Toxicol Environ Health A 70:159-170.

Stepnik, M; Staficzyk, M; Arkusz, J; Lewifiska, D. (2005) Assessment of apoptosis inthymocytes and splenocytes
from mice exposed to arsenate in drinking water: cytotoxic effects of arsenate on the cells in vitro. J Environ Sci
Health A40:369-384.

Stevens, JT; Hall, LL; Farmer, JD; et al. (1977) Disposition of 14C and/or   As-cacodylic acid in rats after
intravenous, intratracheal orperoral administration. Environ Health Perspect 19:151-157.

Styblo, M; Yamauchi, H; Thomas, DJ. (1995) Comparative in vitro methylation of trivalent and pentavalent
arsenicals. Toxicol Appl Pharmacol 135:172-178.

Styblo, M; Delnomdedieu, M; Thomas, DJ. (1996) Mono- and dimethylation of arsenic in rat liver cytosol in vitro.
Chem Biol Interact 99:147-164.

Styblo, M; Vega, L; Germolec, DR; et al.  (1999a) Metabolism and toxicity of arsenicals in cultured cells. In:
Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic exposure and health effects. Oxford, UK: Elsevier; pp.
311-323.

Styblo, M; Del Razo, LM; LeCluyse, EL; et al.  (1999b) Metabolism of arsenic in primary cultures of human and rat
hepatocytes.  Chem Res Toxicol 12(7):560-565.

Styblo, M; Del Razo, LM; Vega, L; et al. (2000) Comparative toxicity of trivalent and pentavalent inorganic and
methylated arsenicals in rat and human cells. Arch Toxicol 74(6):289-299

Su, P-F; Hu, Y-J; Ho, I-C; et al. (2006) Distinct gene expression profiles in immortalized human urothelial cells
exposed to inorganic arsenite and its methylated trivalent metabolites. Environ Health Perspect 114:394-403.

Sumi, D; Taguchi, K ; Sun, Y ; et al. (2005) Monomethylarsonous acid inhibits endothelial nitric oxide synthase
activity. J Health Science 51:728-730.

Sun, G; Xu, Y; Li, X; et al. (2007) Urinary arsenic metabolites in children and adults exposed to arsenic in drinking
water in Inner Mongolia, China. Environ Health Persp 115(4):648-652.

Szymczyk, KH; Kerr, BA; Freeman, TA; et al. (2006) Involvement of hydrogen peroxide in the differentiation and
apoptosis of preosteoclastic cells exposed to arsenite. Biochem Pharmacol 72:761-769.

Tabacova, S; Baird, DD; Balabaeva, L; et al. (1994) Placental arsenic and cadmium in relation to lipid peroxides and
glutathione levels in maternal-infant pairs from a copper smelter area. Placenta 15:873-881.


                                                   181        DRAFT—DO NOT CITE OR QUOTE

-------
Takahashi, M; Barrett, JC; Tsutsui, T. (2002) Transformation by inorganic arsenic compounds of normal Syrian
hamster embryo cells into a neoplastic state in which they become anchorage-independent and cause tumors in
newborn hamsters. Int J Cancer 99:629-634.

Tarn, GKH; Charbonneau, SM; Bryce, F; et al. (1979) Metabolism of inorganic arsenic (  As) in humans following
oral ingestion. Toxicol Appl Pharmacol 50:319-322.

Tang, F; Liu,  G; He, Z; et al. (2006) Arsenite inhibits p53 phosphorylation, DNA binding activity, and p53 target
gene p21 expression in mouse epidermal JB6 cells. Mol Carcinog 45:861-870.

Tenorio, EP; Saavedra, R. (2005) Differential effect of sodium arsenite during the activation of human CD4+ and
CD8+ T lymphocytes. Int Immunopharmacol 5:1853-1869.

Terek, MC; Karabulut, B; Selvi, N; et al. (2006) Arsenic trioxide-loaded, microemulsion-enhanced cytotoxicity on
MDAH 2774  ovarian carcinoma cell line. Int J Gynecol Cancer 16(2):532-537.

Tezuka, M; Hanioka, K; Yamanaka, K; Okada, S. (1993) Gene damage induced in human alveolar type II (L-132)
cells by exposure to dimethylarsenic acid. Biochem Biophys Res Commun 191:1178-1183.

Thomas, DJ. (2007) Molecular processes in cellular arsenic metabolism. Toxicol Appl Pharmacol 222:365-373.

Thomas, DJ; Styblo, M; Lin, S. (2001) The cellular metabolism and systemic toxicity of arsenic. Toxicol Appl
Pharmacol 176:127-144.

Thomas, DJ; Waters, SB; Styblo, M. (2004) Elucidating the pathway for arsenic methylation. Toxicol Appl
Pharmacol 198(3):319-326.

Thomas, DJ; Li, J; Waters, SB; et al. (2007) Arsenic(+3 oxidative state) methyltransferase and methylation of
arsenicals. Exp Biol Med 232:3-13.

Trouba, KJ; Wauson, EM; Vorce, RL. (2000) Sodium arsenite inhibits terminal differentiation of murine C3H
10T1/2 preadipocytes. Toxicol Appl Pharmacol 168:25-35.

Trouba, KJ; Germolec, DR. (2004) Micromolar concentrations of sodium arsenite induce cyclooxygenase-2
expression and stimulate p42/44 mitogen-activated protein kinase phosphorylation in normal human epidermal
keratinocytes. Toxicol Sci 79:248-257.

Tsai, S-M; Wang, T-N; Ko, Y-C. (1999) Mortality for certain diseases in areas with high levels of arsenic. Arch
Environ Health 54(3): 186-193.

Tsai, YS; Yang, W-H; Tong, T-C; et al. (2005) Experience with primary urethral  carcinoma from the blackfoot
disease-endemic area of south Taiwan: increased frequency of bulbomembranous adenocarcinoma?  Urol Int 74:229-
234.

Tseng, WP. (1977) Effects and dose-response relationships of skin cancer and blackfoot disease with arsenic.
Environ Health Perspect 19:109-119.

Tseng, WP; Chu, HM; How, SW; et al. (1968) Prevalence of skin cancer in an endemic area of chronic arsenicism in
Taiwan. J Natl Cancer Inst 40(3):453-463.

Tseng, C-H; Huang, Y-K; Huang, Y-L; et al. (2005) Arsenic exposure, urinary arsenic speciation, and peripheral
vascular disease in blackfoot-disease-hyperendemic villages in Taiwan. Toxicol Appl Pharmacol 206:299-308.

Tseng, YS; Tzeng, CC; Huang, CY; et al. (2006) Aurora-A overexpression associates with Ha-ras codon-12
mutation and blackfoot disease endemic area in bladder cancer. Cancer Lett 241:93-101.
                                                  182        DRAFT—DO NOT CITE OR QUOTE

-------
Tsuchiya, T; Tanaka-Kagawa, T; Jinno, H; et al. (2005) Inorganic arsenic compounds and methylated metabolites
induce morphological transformation in two-stage B ALB/c 3T3 cell assay and inhibit metabolic cooperation in V79
cell assay. Toxicolog Sci 84:344-351.

Tsuda, T; Babazono, A; Yamamoto, E; et al. (1995) Ingested arsenic and internal cancer: a historical cohort study
followed for 33 years. Am J Epidemiol 141:198-209.

Uchino, T; Roychowdhury, T; Ando, M;  Tokunaga, H. (2006) Intake of arsenic from water, food composites and
excretion through urine, hair from a studied population in West Bengal, India. Food Chem Toxicol 44(4):455-61.

Uddin, AN; Burns, FJ; Rossman, TG. (2005) Vitamin E and organoselenium prevent the cocarcinogenic activity of
arsenite  with solar UVR in mouse skin. Carcinogenesis 26:2179-2186.

Upreti, RK; Kannan, A; Pant, AB. (2007) Experimental exposure of arsenic in cultured rat intestinal epithelial cells
and cell line: lexicological consequences. Toxicol In Vitro 21:32^0.

U.S. EPA (Environmental Protection Agency). (1984) Health assessment document for inorganic arsenic. Final
report. Office of Health and Environmental Assessment; EPA/600/8-83-021.

U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical
mixtures. FedRegist 51(185):34014-34025.

U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk assessment. Fed Regist
51(185):34006-34012.

U.S. EPA (Environmental Protection Agency). (1988a) Recommendations for and documentation of biological
values for use in risk assessment. EPA/600/6-87/008. Available from the National Technical Information Service,
Springfield, VA; PB88-179874/AS, February 1988.

U.S. EPA (Environmental Protection Agency). (1988b) Special report on ingested inorganic arsenic: skin cancer;
nutritional essentiality. Risk Assessment Forum, Washington, DC; EPA/625/3-87/013.

U.S. EPA (1989) Report on Arsenic (As) Work Group meetings. Memorandum. From: Abernathy, CO; Marcus, W;
Office of Drinking Water; and Chen, C; Gibb, H; White,  P; Office of Research and Development; to Cook, P; Office
of Drinking Water; and Preuss, P; Office  of Regulatory Support and Scientific Management. February 23.

U.S. EPA (Environmental Protection Agency). (1991) Guidelines for developmental toxicity risk assessment. Fed
Regist 56(234):63798-63826. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.

U.S. EPA (Environmental Protection Agency). (1994) Integrated Risk Information System (IRIS) on inorganic
arsenic.  Verification date February 3, 1994.Cincinnati, OH. Accessed on June 26, 2008.

U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose approach in health risk
assessment. EPA/630/R-94/007.

U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive toxicity risk assessment. Fed
Regist 61(212):56274-56322. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.

U.S. EPA (Environmental Protection Agency). (1998) Guidelines for neurotoxicity risk assessment. Fed Regist
63(93):26926-26954. Available online at http://www.epa.gov/ncea/raf/rafguid.htm.

U.S. EPA (Environmental Protection Agency). (2000a) Science policy council handbook: peer review. 2nd ed.
Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-00/001.

U.S. EPA (Environmental Protection Agency). (2000b) Science policy council handbook: risk characterization.
Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-00/002.


                                                 183        DRAFT—DO NOT CITE OR QUOTE

-------
U.S. EPA (Environmental Protection Agency). (2000c) Benchmark dose technical support document. External
review draft. Office of Research and Development, Risk Assessment Forum, Washington, DC; EPA/630/R-00/001,
October 2000.

U.S. EPA (Environmental Protection Agency). (2000d) Supplementary guidance for conducting health risk
assessment of chemical mixtures. Office of Research and Development, Risk Assessment Forum, Washington, DC;
EPA /630/R-00/002.

U.S. EPA (Environmental Protection Agency). (2001) National primary drinking water regulations: arsenic and
clarifications to compliance and new source contaminants monitoring. Fed Regist 66(14):6975-7022, January 2001.

U.S. EPA (Environmental Protection Agency). (2002) A review of the reference dose concentration and reference
concentration processes. Risk Assessment Forum, Washington, DC; EPA/630/P-02/002F. Available online at
http ://www. epa. gov/ncea/raf.

U.S. EPA (Environmental Protection Agency). (2004) Estimated per capita water ingestion and body weight in the
United States—an update based on data collected by the United States Department of Agriculture's 1994-1996 and
1998 continuing survey of food intakes by individuals. EPA Office of Water; EPA/822/R-00/001, October 2004.

U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Office of
Research and Development, Risk Assessment Forum, Washington, DC; EPA/630/P-03/001F.

U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from
early-life exposure to carcinogens. Risk Assessment Forum; EPA/630/R-03/003F, March 2005.

U.S. EPA (Environmental Protection Agency). (2005c) Toxicological review of inorganic arsenic internal review
draft. Health Effects Criteria Division. Available online  at http://epa.gov/waterscience/criteria/arsenic/sab.

U.S. EPA (Environmental Protection Agency). (2005d). Issue paper: inorganic arsenic cancer CSF. Arsenic Cancer
CSF Workgroup. (Submitted to SAB)

U.S. EPA (Environmental Protection Agency). (2006a) Science policy council handbook: peer review. 3rd ed.  Office
of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-06/002. Available online at
http ://www. epa. gov/OS A/spc/2peerrev. htm.

U.S. EPA (Environmental Protection Agency). (2006b) A framework for assessing health risk of environmental
exposures to children. National Center for Environmental Assessment, Washington, DC; EPA/600/R-05/093F.
Available online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363.

U.S. EPA (Environmental Protection Agency). (2006c).  Revised re-registration eligibility decision document for
MSMA, DSMA, CAMA, and cacodylic acid. Office of Pesticides, Prevention, and Toxic Substances; EPA/738-R-
06-021.

U.S. EPA (Environmental Protection Agency). (2008). Human health risk assessment and ecological risk assessment
for the reregistration eligibility. Decision document for inorganic arsenicals and chromium-based wood
preservatives. Office of Pesticides, Prevention, and Toxic Substances; EPA/739/R-08/006.

Vahter, M. (1981) Biotransformation of trivalent and pentavalent inorganic arsenic in mice and rats. Environ Res
25:286-293.

Vahter, M. (1983) Metabolism of inorganic arsenic in relation to chemical form and animal species. Doctoral Thesis.
Karolinska Institutet, Stockholm.

Vahter, M. (1994) Species differences in  the metabolism of arsenic compounds. Appl Organomet Chem 8:175-182.

Vahter, M. (1999a) Methylation of inorganic arsenic in different mammalian species and population groups. Sci
Prog82(Pt.l):69-88.

                                                  184        DRAFT—DO NOT  CITE OR QUOTE

-------
Vahter, M. (1999b) Variation in human metabolism of arsenic. In: Chappell, WR; Abernathy, CO; Calderon, RL;
eds. Arsenic exposure and health effects. Oxford, UK: Elsevier; pp. 267-279.

Vahter, M; Envall, J. (1983) In vivo reduction of arsenate in mice and rabbits. Environ Res 32:14-24.

Vahter, M; Marafante, E. (1983) Intracellular interaction and metabolic fate of arsenite and arsenate in mice and
rabbits. ChemBiol Interact 47:29^4.

Vahter, M; Marafante, E. (1985) Reduction and binding of arsenate in marmoset monkeys. Arch Toxicol 57:119-
124.

Vahter, M; Marafante, E. (1987) Effects of low dietary intake of methionine, choline or proteins on the
biotransformation of arsenite in the rabbit. Toxicol Lett 37:41^6.

Vahter, M; Marafante, E; Lindgren, A; Dencker, L. (1982) Tissue distribution and subcellular binding of arsenic in
marmoset monkeys after injection of   As-arsenite. Arch Toxicol 51:65-77.

Vahter, M; Marafante, E; Dencker, L. (1984) Tissue distribution and retention of  As-dimethylarsinic acid in mice
and rats. Arch Environ Contam Toxicol 13:259-264.

Vahter, M; Friberg L; Rahnster B; et al. (1986) Airborne arsenic and urinary excretion of metabolites of inorganic
arsenic among smelter workers. Int Arch Occup Environ Health 57:79-91.

Vahter, M; Couch, R; Nermell, B; Nilsson, R. (1995a) Lack of methylation of inorganic arsenic in the chimpanzee.
Toxicol Appl Pharmacol 133:262-268.

Vahter, M; Concha, G; Nermell, B; et al. (1995b) A unique metabolism of inorganic arsenic in native Andean
women. Eur J Pharmacol 293:455-462.

Valentine, JL; Kang, HK; Spivey, G. (1979) Arsenic levels in human blood, urine, and hair in response to exposure
via drinking water. Environ Res 20:24-32.

Valenzuela, OL; Borja-Aburto, VH; Garcia-Vargas, GG; et al. (2005) Urinary trivalent methylated arsenic species in
a population chronically exposed to inorganic arsenic. Environ Health Persp 113(3):250-254.

Valenzuela, OL; Germolec, DR; Borja-Aburto, VH; et al. (2007) Chronic arsenic exposure increases TGF-alpha
concentration in bladder urothelial cells of Mexican populations environmentally exposed to inorganic arsenic.
Toxicol Appl Pharmacol 222:264-270.

Vega, L;  Gonsebatt, ME; Ostrosky-Wegman, P. (1995) Aneugenic effect of sodium arsenite on human lymphocytes
in vitro: an individual susceptibility effect detected. Mutat Res 334:365-373.

Vega, L;  Styblo, M; Patterson, R; et al. (2001) Differential effects of trivalent and pentavalent arsenicals on cell
proliferation and cytokine secretion in normal human epidermal keratinocytes. Toxicol Appl Pharmacol 172:225-
232.

Vellar, OD. (1969) Nutrient losses through sweating. Thesis. Oslo University, Oslo, Norway.

Venson DJ; Moolgavkar AH. (1988). A method for computing profile-likelihood based confidence intervals. Appl
Stat 37:87-94.

Vig, BK; Figueroa, ML; Cornforth, MN; Jenkins, SH. (1984) Chromosome studies in human subjects chronically
exposed to arsenic in drinking water. Am J Ind Med 6(5):325-338.

Vogt, BL; Rossman, TG. (2001) Effects of arsenite onp53, p21 and cyclinD expression in normal human
fibroblasts—a possible mechanism for arsenite's comutagenicity. Mutat Res 478:159-168.
                                                  185        DRAFT—DO NOT CITE OR QUOTE

-------
Waalkes, MP; Ward, JM; Liu, J; Diwan, BA. (2003) Transplacental carcinogenicity of inorganic arsenic in the
drinking water: induction of hepatic, ovarian, pulmonary, and adrenal tumors in mice. Toxicol Appl Pharmacol
186:7-17.

Waalkes, MP; Ward, JM; Diwan, BA. (2004a) Induction of tumors of the liver, lung, ovary and adrenal in adult mice
after brief maternal gestational exposure to inorganic arsenic: promotional effects of postnatal phorbol ester exposure
on hepatic and pulmonary, but not dermal cancers. Carcinogenesis 25(1): 133-141.

Waalkes, MP; Liu, J; Ward, JM; Diwan, BA. (2004b) Animal models for carcinogenesis: inorganic arsenic is a
transplacental carcinogen in mice. Toxicol Appl Pharmacol 198:377-384.

Waalkes, MP; Liu, J; Ward, J; et al. (2006a) Urogenital carcinogenesis in female GDI mice induced by in utero
arsenic exposure is exacerbated by postnatal diethylstilbestrol treatment. Cancer Res 66:1337-1345.

Waalkes, MP; Liu, J; Ward, JM; Diwan, BA. (2006b). Enhanced urinary bladder and liver carcinogenesis in male
GDI mice exposed to transplacental inorganic arsenic and postnatal diethylstilbestrol or tamoxifen. Toxicol Appl
Pharmacol 215:295-305.

Walter, I; Schwerdtle, T; Thuy,  C; et al. (2007) Impact of arsenite and its methylated metabolites on PARP-1
activity, PARP-1 gene expression and poly(ADP-ribosyl)ation in cultured human cells. DNA Repair (Amst) 6:61-
70.

Wan, B; Christian, RT; Soukup, SW. (1982) Studies of cytogenetic effects of sodium arsenicals on mammalian cells
in vitro. Environ Mutag 4:493-498.

Wang, TS; Huang, H. (1994) Active oxygen species are involved in the induction of micronuclei by arsenite in XRS-
5 cells. Mutagenesis 9:253-257.

Wang, TS; Kuo, CF; Jan, KY; Huang, H. (1996) Arsenite induces apoptosis in Chinese hamster ovary cells by
generation of reactive oxygen species. J Cell Physiol 169:256-268.

Wang, TS; Shu, YF; Liu, YC; et al. (1997) Glutathione peroxidase and catalase modulate the genotoxicity of
arsenite. Toxicol 121:229-237.

Wang, TS; Hsu, TY; Chung, CH; et al. (2001) Arsenite induces oxidative DNA adducts and DNA-protein crosslinks
in mammalian cells. Free Radical Biol Med 31:321-330.

Wang, XJ; Yang, J; Gang, H;  et al. (2005) Gene expression alteration during redox-dependent enhancement of
arsenic cytotoxicity by emodin inHeLa cells. Cell Res 15:511-522.

Wang, TC; Jan, KY; Wang, AS; Gurr, JR. (2007) Trivalent arsenicals induce lipid peroxidation, protein
carbonylation, and oxidative DNA damage in human urothelial cells. Mutat Res 615:75-86.

Warner, ML; Moore, LE; Smith, MT; et al. (1994) Increased micronuclei in exfoliated bladder cells of individuals
who chronically ingest arsenic-contaminated water in Nevada. Cancer Epidemiol Biomarkers Prev 3:583-590.

Waters, SB; Devesa, V; Fricke,  MW; et al. (2004a) Glutathione modulates recombinant rat arsenic (+3 oxidation
state) methyltransferase-catalyzed formation of trimethylarsine oxide and trimethylarsine. Chem Res Toxicol
17(12):1621-1629.

Waters, SB; Devesa, V; Del Razo, LM;  et al. (2004b) Endogenous reductants support the catalytic function of
recombinant rat cyt!9, an arsenic methyltransferase. Chem Res Toxicol 17:404-409.

Webb, DR; Sipes,  IG; Carter, DE (1984) In vitro solubility and in vivo toxicity of gallium arsenide. Toxicol Appl
Pharmacol 76:96-104.


                                                  186         DRAFT—DO NOT CITE OR QUOTE

-------
Wei, M; Wanibuchi, H; Morimura, K; et al. (1999) Urinary bladder carcinogenicity of dimethylarsinic acid in male
F344 rats. Carcinogenesis 20(9): 1873-1876.

Wei, M; Wanibuchi, H; Morimura, K; et al. (2002) Carcinogenicity of dimethylarsinic acid in male F344 rats and
genetic alterations in induced urinary bladder tumors. Carcinogenesis 23(8): 1387-1397.

Wei, M; Arnold, L; Cano, M; Cohen, SM. (2005) Effects of co-administration of antioxidants and arsenicals on the
rat urinary bladder epithelium. Toxicol Sci 83:237-245.

Wester, RC; Maibach, HI; Sedik, L; et al. (1993) In vitro and in vivo percutaneous absorption and skin
decontamination of arsenic from water and soil. Fundam Appl Toxicol 20:336-340.

Wester, RC; Hui, X; Barbadillo, S; et al. (2004) In vivo percutaneous absorption of arsenic from water and CCA-
treated wood residue. Toxicol Sci 79:287-295.

Wetzler, M; Earp, JC; Brady, MT; et al. (2007) Synergism between arsenic trioxide and heat shock protein 90
inhibitors on signal transducer and activator of transcription protein 3 activity—pharmacodynamic drug-drug
interaction modeling. Clin Cancer Res 13:2261-2270.

Wiencke, JK; Yager, JW. (1992)  Specificity of arsenite in potentiating cytogenic damage induced by the DNA
crosslinking agent diepoxybutane. Environ Mol Mutagen 19:195-200.

Wong, R-H; Kuo, C-Y; Hsu, M-L; et al. (2005) Increased levels of 8-hydroxy-2'-deoxyguanosine attributable to
carcinogenic metal exposure among schoolchildren. Environ Health Perspect 113:1386-1390.

Wu, MM; Kuo, TL; Hwang, YH; Chen CJ. (1989) Dose-response relation between arsenic concentration in well
water and mortality from cancers and vascular diseases. Am J Epidemiol 130:1123-1132.

Wu, F; Burns, FJ; Zhang, R; et al. (2005) Arsenite-induced alterations of DNA photodamage repair and apoptosis
after solar-simulation UVR in mouse keratinocytes in vitro. Environ Health Perspect 113:983-986.

Xie, Y; Liu, J; Benbrahim-Tallaa, L; et al. (2007) Aberrant DNA methylation and gene expression in livers of
newborn mice transplacentally exposed to a hepatocarcinogenic dose of inorganic arsenic. Toxicology 236:7-15.

Yamanaka,  K; Hayashi, H; Tachikawa, M; et al. (1997) Metabolic methylation is a possible genotoxicity-enhancing
process of inorganic arsenics. MutatRes 394(13):95-101.

Yamanaka,  K; Mizoi, M; Tachikawa, M; et al. (2003) Oxidative DNA damage following exposure to
dimethylarsinous iodide: the formation of cis-thymine glycol. Toxicol Lett 143:145-153.

Yamauchi, H; Yamamura, Y. (1979) Dynamic change of inorganic arsenic and methylarsenic compounds in human
urine after oral intake as arsenic trioxide. Ind Health 17:79-84.

Yamauchi, H; Yamamura, Y. (1984) Metabolism and excretion of orally administered dimethylarsinic acid in the
hamster. Toxicol  Appl Pharmacol 74:134-140.

Yamauchi, H; Takahashi, K; Yamamura, Y. (1986) Metabolism and excretion of orally and intraperitoneally
administered gallium arsenide in the hamster. Toxicology 40:237-246.

Yamauchi, H; Aminaka, Y; Yoshida, K; et al. (2004) Evaluation of DNA damage in patients with arsenic poisoning:
urinary 8-hydroxydeoxyguanine.  Toxicol Appl Pharmacol 198:291-296.

Yancy, SL;  Shelden, EA; Gilmont, RR; Welsh, MJ. (2005) Sodium arsenite exposure alters cell migration, focal
adhesion localization and decreases tyrosine phosphorylation of focal adhesion kinase in H9C2 myoblasts. Toxicol
Sci 84:278-286.
                                                  187        DRAFT—DO NOT CITE OR QUOTE

-------
Yang, JL; Chen, MF; Wu, CW; Lee, TC. (1992) Posttreatment with sodium arsenite alters the mutation spectrum
induced by ultraviolet light irradiation in Chinese hamster ovary cells. Environ Mol Mutagen 20:156-164.

Yang, CH; Kuo, ML; Chen, JC; Chen, YC. (1999) Arsenic trioxide sensitivity is associated with low level of
glutathione in cancer cells. Br J Cancer 81:796-799.

Yang, CY; Chiu, H-F; Wu, T-N; et al. (2004) Reduction in kidney cancer mortality following installation of a tap
water supply system in an arsenic-endemic area of Taiwan. Arch Environ Health 59(9):484-488.

Yang, CY; Chiu, HF; Chang, C-C; et al. (2005) Bladder cancer mortality reduction after installation of a tap-water
supply system in an arsenious-endemic area in southwestern Taiwan. Environ Res 98:127-132.

Yang, P;  He, XQ; Peng, L; et al.  (2007) The role of oxidative stress in hormesis induced by sodium arsenite in
human embryo lung fibroblast (HELP) cellular proliferation model. J Toxicol Environ Health A 70:976-983.

Yates, CR; Krynetski, EY; Loennechen, T; et al. (1997) Molecular diagnosis of thiopurine S-methyltransferase
deficiency: genetic basis for azathioprine and mercaptopurine intolerance. Ann Intern Med 126:608-614.

Ye, J; Li, A; Liu, Q; et al. (2005) Inhibition of mitogen-activated protein kinase kinase enhances apoptosis induced
by arsenic trioxide in human breast cancer MCF-7 cells. Clin Exp Pharmacol Physiol 32:1042-1048.

Yedjou, CG; Tchounwou, PB. (2007) In-vitro cytotoxic and genotoxic effects of arsenic trioxide on human leukemia
(HL-60) cells using the MTT and alkaline single cell gel electrophoresis (comet) assays. Mol Cell Biochem
301:123-130.

Yi, J; Yang, J; He, R; et al. (2004) Emodin enhances arsenic trioxide-induced apoptosis via generation of reactive
oxygen species and inhibition of survival signaling. Cancer Res 64:108-116.

Yih, J-H; Lee, T-C. (1999) Effects of exposure protocols on induction of kinetochore-plus and -minus micronuclei
by arsenite in diploid human fibroblasts. Mutat Res 440:75-82.

Yih, LH; Hsueh, SW; Luu, WS; et al. (2005) Arsenite induces prominent mitotic arrest via inhibition of G2
checkpoint activation in CGL-2 cells. Carcinogenesis 26:53-63.

Yoshida, K; Inoue, Y; Kuroda, K; et al. (1998) Urinary excretion of arsenic metabolites after long-term oral
administration of various arsenic compounds to rats. J Toxicol Environ Health 54(3): 179-192.

Yost, LJ; Schoof, RA; Aucoin, R. (1998) Intake of inorganic arsenic in the North American diet. Hum Ecol Risk
Assess 4:137-152.

Yu, D. (1999a) A physiologically based pharmacokinetic model of inorganic arsenic. Regul Toxicol Pharmacol 29(2
Ptl):128-141.

Yu, D. (1999b) A pharmacokinetic modeling of inorganic arsenic: a short-term oral exposure model for humans.
Chemosphere 39(15):2737-2747.

Yu, H-S; Chang, K-L; Yu, C-L; et al. (1998) Defective IL-2 receptor expression in lymphocytes of patients with
arsenic-induced Bowen's disease. ArchDermatol Res 290:681-687.

Yu, RC; Hsu, K-H; Chen, C-J; Froines, JR. (2000) Arsenic methylation capacity and skin cancer. Cancer Epidemiol,
Biomarkers Prev 9:1259-1262.

Yuan, Y; Marshall, G; Ferreccio, C; et al. (2007) Acute myocardial infarction mortality in comparison with lung and
bladder cancer mortality in arsenic-exposed Region II of Chile from 1950 to 2000.  Am J Epidemiol Advance Access
1-11.
                                                  188        DRAFT—DO NOT CITE OR QUOTE

-------
Zakharyan, RA; Aposhian, HV. (1999) Enzymatic reduction of arsenic compounds in mammalian systems: the rate-
limitin
1283.
limiting enzyme of rabbit liver arsenic biotransformation is MMA  reductase. Chem Res Toxicol 12(12): 1278-
Zakharyan, RA; Wu, Y; Bogdan, GM; Aposhian, HV. (1995) Enzymatic methylation of arsenic compounds: assay,
partial purification, and properties of arsenite methyltransferase and monomethylarsonic acid methyltransferase of
rabbit liver. Chem Res Toxicol 8:1029-1038.

Zakharyan, RA; Wildfang, E; Aposhian, HV. (1996) Enzymatic methylation of arsenic compounds:  . The
marmoset and tamarin, but not the rhesus, monkeys are deficient in methyltransferases that methylate inorganic
arsenic. Toxicol Appl Pharmacol 140:77-84.

Zakharyan, RA; Ayala-Fierro, F; Cullen, WR; et al. (1999) Enzymatic methylation of arsenic compounds. VII.
Monomethylarsonous acid (MMA   ) is the substrate for MMA methyltransferase of rabbit liver and human
hepatocytes. Toxicol Appl Pharmacol 158(1):9-15.

Zakharyan, RA; Sampayo-Reyes, A; Healy, SM; et al. (2001) Human monomethylarsonic acid (MMA(V)) reductase
is a member of the glutathione-S-transferase superfamily. Chem Res Toxicol 14(8): 1051-1057.

Zhang, T-C; Cao, EH; Li, JF; et al. (1999) Induction of apoptosis and inhibition of human gastric cancer MGC-803
cell growth by arsenic trioxide. Eur J Cancer 35:1258-1263.

Zhang, T-C; Schmitt, MT; Mumford, JL. (2003) Effects of arsenic on telomerase and telomeres in relation to cell
proliferation and apoptosis in human keratinocytes and leukemia cells in vitro. Carcinogenesis 24(11): 1811-1817.

Zhang, Y; Lu, Y; Ding, M; et al. (2005) Deficiency in Ikk(3 gene enhances arsenic-induced gadd45a expression. Mol
Cell Biochem 279:163-168.

Zhang, Y; Bhatia, D; Xia, H; et al.  (2006) Nucleolin links to arsenic-induced stabilization of GADD45a mRNA.
Nucleic Acids Res 34:485^95.

Zhao, CQ; Young, MR; Diwan, BA; et al. (1997) Association of arsenic-induced malignant transformation with
DNA hypomethylation and aberrant gene expression. Proc Natl Acad Sci USA 94:10907-10912.

Zheng, PZ ; Wang, K-K; Zhang, Q-Y; et al. (2005) Systems analysis of transcriptome and proteome in retinoic acid /
arsenic trioxide-induced cell differentiation / apoptosis of promyelocytic leukemia. Proc Natl Acad Sci USA
102:7653-7658.

Zhong, W; Oberley, LW; Oberley, TD; Clair, DK.  (1997) Suppression of the malignant phenotype of human glioma
cells by overexpression of manganese superoxide dismutase. Oncogene 14:481^190.

Zhong, CX; Mass, MJ. (2001) Both hypomethylation and hypermethylation of DNA associated with arsenite
exposure in cultures of human cells identified by methylation-sensitive arbitrary-primed PCR. Toxicol Lett 122:223-
234.
                                                 189        DRAFT—DO NOT CITE OR QUOTE

-------
            APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
                                COMMENTS AND DISPOSITION
 1          The Toxicological Review of Inorganic Arsenic has been formally reviewed by scientists
 2    outside EPA—i.e., the SAB Arsenic Review Panel—in accordance with EPA guidance on peer
 3    review (U.S. EPA, 2000a).  The reviewers on the Panel were tasked with providing written
 4    answers to general questions on the overall assessment and on chemical-specific charge
 5    questions, addressing key scientific issues of the assessment. While the Panel was supplied with
 6    questions regarding both DMAV and inorganic arsenic, this appendix addresses only questions
 7    and responses pertaining to inorganic arsenic.  Charge question B3 asked SAB to comment on
 8    EPA's hypothesis that inorganic arsenic acts via different modes of action for carcinogenicity.
 9    SAB agreed with EPA's conclusion, but during a discussion on the mode of action of DMAV, a
10    member of the Panel stated that the description for inorganic arsenic's mode of action could be
11    strengthened. In addition to strengthening the mode of action discussion, studies on the mode of
12    action for inorganic arsenic have been placed in a table in Appendix C.  Section 4.4.1 provides a
13    summary of the specifics in the tables instead of detailed write-ups for all the studies.  A
14    summary of significant comments made by the external reviewers and EPA's responses to these
15    comments arranged by charge question follow. Public comments were submitted to SAB and
16    were taken into consideration by the Panel during their review. The summary of significant
17    comments and responses below is inclusive  of the major issues raised by public commenters
18    which specifically focused on the choice of  study for cancer quantitation and the nature of the
19    dose-response.  Editorial comments were considered and incorporated into the document as
20    appropriate and are not discussed further.
21
22    Charge Question B3
23
24          EPA concluded that inorganic arsenic mostly likely causes human cancer by many
25    different modes of action.  This is based on the observed findings that inorganic arsenic
26    undergoes successive methylation steps in humans and results in the production of a number of
27    intermediate metabolic products and that each has its own toxicity. EPA asked SAB to comment
28    on the soundness of its conclusion.
29
30    SAB Comments
31
32          The Panel concluded that:

33          1)     Multiple modes of action may operate in carcinogenesis induced by inorganic
34                 arsenic because there is simultaneous exposure to multiple metabolic products

                                               A-1       DRAFT—DO NOT CITE OR QUOTE

-------
 1                 as well as multiple target organs and the composition of metabolites can differ
 2                 in different organs.
 3          2)     Each arsenic metabolite has its own cytotoxic and genotoxic capability.
 4          3)     Inorganic arsenic (iAs111) and its metabolites are not direct genotoxicants because
 5                 these compounds do not directly react with DNA. However, iAs111 and some of its
 6                 metabolites can exhibit indirect genotoxicity, induce aneuploidy, cause changes in
 7                 DNA methylation, and alter signaling and hormone action.  In addition, inorganic
 8                 arsenic can act as a transplacental carcinogen and a cocarcinogen.
 9          4)     Studies of indirect genotoxicity strongly suggest the possibility of a threshold for
10                 arsenic carcinogenicity. However, the studies discussed herein do not show
11                 where such a threshold might be, nor do they show the shape of the dose-response
12                 curve at these  low levels. In addition, a threshold has not been confirmed by
13                 epidemiological studies.  This issue is an extremely important area for research
14                 attention, and  it is an issue that should be evaluated in EPA's continuing risk
15                 assessment for inorganic arsenic.
16          5)     Arsenic essentiality and the possibility of hormetic effects are in need of
17                 additional research to determine how they would influence the determination of
18                 a threshold for specific arsenic-associated health endpoints.

19    EPA Response
20
21          EPA agrees that the available data potentially support multiple modes of action for
22    inorganic arsenic. The Agency believes that, at this point, the data concerning  mode of action
23    are not well-enough understood to support their use in quantitative risk assessment.
24
25    Charge Question C2
26
27          EPA reviewed the available epidemiologic studies, including those published since the
28    NRC 2001 review, for U.S. populations exposed to inorganic arsenic via drinking water.  EPA
29    concluded that the Taiwanese data set remains the most appropriate choice for  estimating cancer
30    risk in humans.  SAB was asked to comment on the soundness of this conclusion and also on
31    whether these data provide adequate characterization of the impact of childhood exposure to
32    inorganic arsenic.
33
34    SAB Comments
35
36          The Panel concluded that:
37
38          1)     Because  of various factors (e.g., the size and statistical stability  of the
39                 Taiwanese database relative to other studies, the reliability of the population
40                 and mortality counts, the stability of residential patterns, and the inclusion of
41                 long-term exposures), this database remains, at this time, the most appropriate
42                 choice for estimating bladder cancer risk among humans, though the data  have
43                 considerable limitations that should be described qualitatively or quantitatively
44                 to help inform risk managers about the strength of the conclusions.
                                                A-2      DRAFT—DO NOT CITE OR QUOTE

-------
 1           2)     There are other epidemiologic databases from studies of populations also
 2                 exposed at high levels of arsenic, and the panel recommends that these be used
 3                 to compare the unit risks at the higher exposure levels that have emerged from
 4                 the Taiwan data.
 5           3)     The panel also suggests that published epidemiology studies of U.S. and other
 6                 populations chronically exposed from 0.5 to 160 ug/L inorganic  arsenic in
 7                 drinking water be critically evaluated, using a uniform set of criteria, and that the
 8                 results from these evaluations be transparently documented in EPA's assessment
 9                 documents. If, after this evaluation, one or more of these studies are shown to be
10                 of potential utility, the low-level studies and Taiwan data may be compared for
11                 concordance.  Comparative analyses could lead to further insights into the
12                 possible influence of these differences on population responses to arsenic in
13                 drinking water.
14           4)     Regarding childhood exposure to inorganic arsenic, it was the Panel's view that,
15                 based on available data, it is not clear whether children differ from adults with
16                 regard to their sensitivity to the carcinogenic effects of arsenic in drinking water.
17                  However, the possibility of a different response in degree or kind should not be
18                 ignored and needs to be investigated.
19
20    EPA Response
21
22           After considering additional studies, EPA agreed with SAB that the Taiwanese data were
23    the best available for quantitative analysis.  Studies assessed, but not used in the analysis, are
24    summarized in Section 4.1 of the document. The studies were systematically evaluated for  their
25    suitability in risk assessment based on a uniform set of criteria including the study type, the size
26    of the study population and control population, and the relative strengths and weaknesses of the
27    study based on SAB-recommended criteria (i.e., estimates of the level of exposure
28    misclassification; temporal variability in assigning past arsenic levels from recent measurements;
29    the extent of reliance on imputed exposure levels; the number of persons exposed at various
30    estimated levels of waterborne arsenic; study response/participation rates; estimates of exposure
31    variability; control selection methods in case-control studies; and the resulting influence of these
32    factors on the magnitude and statistical stability of cancer risk estimates).  Study summaries are
33    also provided in tabular form in Appendix B for ease of comparison.  Studies are arranged
34    geographically and include other areas of high arsenic exposure (e.g., South America) as well as
35    areas of low exposure (e.g., U.S. and Europe). Studies examining children were evaluated and
36    are discussed in Section 4.7.1 of the document, but EPA believes that the available data do not
37    yet allow a definitive conclusion on children's differential susceptibility to arsenic exposure.
38    EPA notes that recent animal studies demonstrating the potential for cancer after in utero arsenic
39    exposures give rise to additional concerns regarding exposures early in development.
40
                                                A-3       DRAFT—DO NOT CITE OR QUOTE

-------
 1    Charge Question D2
 2
 3          EPA determined that the most prudent approach for modeling cancer risk from inorganic
 4    arsenic is to use a linear model because of the remaining uncertainties regarding the ultimate
 5    carcinogenic metabolites and whether mixtures of toxic metabolites interact at the site(s) of
 6    action. EPA asked SAB if it concurred with the selection of a linear model following the
 7    recommendations of the NRC (2001) to estimate cancer risk in light of the multiple modes of
 8    carcinogenic action for inorganic arsenic.
 9    SAB Comments
10
11          The Panel concluded that:

12          1)     Inorganic arsenic has the potential for a highly complex mode of action.
13          2)     Until more is learned about the complex PK and PD properties of inorganic
14                 arsenic and its metabolites, there is not sufficient justification for the choice of a
15                 specific nonlinear form of the dose-response  relationship.
16          3)     The NRC (2001) recommendation to base risk assessments on a linear dose-
17                 response model that includes the southwestern Taiwan population as a
18                 comparison group seems the most appropriate approach.
19          4)     The Panel also recommends that EPA perform a sensitivity analysis of the
20                 Taiwanese data with different exposure metrics, with the subgroup of villages
21                 with more than one well measurement, and using a multiplicative model that
22                 includes a quadratic term for dose.

23    EPA Response
24
25          As  discussed in Section 5.3, EPA investigated a range of model forms for use in the risk
26    assessment, building on previous efforts,  including U.S. EPA (2001) and Morales et  al. (2000).
27    The model used in the derivation of the preferred risk assessments (see Section 5.3.3) employs:
28
29       •  Poisson regression (of cancer mortality against age and dose) fit by maximum likelihood
30          estimation (MLE).
31       •  A quadratic age model.
32       •  A linear multiplicative dose term.
33       •  Confidence limits on the dose term estimated by profile likelihood.
34       •  Estimates derived for the data set that includes the southwest Taiwan reference
35          population.
36
37          A range of alternative model forms were investigated, as discussed in Section 5.3.8.4,
38    and the impacts of alternative assumptions about nonwater arsenic intake, drinking water
39    consumption, and other exposure factors were investigated through  sensitivity analyses, as
40    described in Section 5.3.8.3. EPA also investigated the properties of the dose-response
41    relationship in the low-dose range of the Taiwanese data, and found that arsenic slope
42    coefficients were positive and statistically significant even when high-exposure groups were

                                               A-4       DRAFT—DO NOT CITE OR QUOTE

-------
 1    excluded from the analysis.  EPA's dose-response modeling found no indication of the existence
 2    of a threshold arsenic exposure below which cancer risks are not elevated. As discussed in
 3    Section 4.6.3, EPA believes that the available mode of action data do not justify the use of non-
 4    linear low-dose extrapolation from the point of departure (POD).
 5
 6    Charge Question D3
 1
 8          EPA re-implemented the model presented in the NRC (2001) in the language R as well as
 9    in an Excel spreadsheet format. In addition, extensive testing of the resulting code was
10    conducted. Please comment upon precision and accuracy of the re-implementation of the model.
11    SAB Comments
12
13          The Panel concluded:

14          1)     That the EPA program conformed to the NRC (2001) recommendation for
15                 modeling cancer hazard as a function of age and the average daily dose of
16                 exposure to arsenic through drinking water sources.
17          2)     The Panel did, however, identify and report to the EPA on two potential
18                 discrepancies in the data inputs and one computational error in the portion of the
19                 program that employs the BEIR-IV formula to evaluate excess lifetime cancer
20                 risk from  arsenic exposure.
21          3)     The Panel made several suggestions for improvements in the model's
22                 programming and documentation conventions, as well as recommendations for
23                 specific sensitivity analyses designed to test the robustness of the model to
24                 alternative formulations of the hazard function and aggregate population data
25                 inputs.

26    EPA Response
27
28          EPA made a number of changes to the model implementation in response to the SAB
29    comments. As in the previous analyses, the linear Poisson dose-response models were estimated
30    using maximum likelihood methods; models were implemented in Excel® and replicated using
31    Statistica®.  In the latest analyses, confidence limits on the arsenic dose-response coefficients
32    were estimated using profile likelihood, rather than Bayesian simulation. The confidence limit
33    estimates derived using profile likelihood were very similar to those obtained using Bayesian
34    simulation and estimates derived by "bootstrap" methods.
35          In this latest analysis, the BEIRIV formula for estimating lifetime cancer incidence risks
36    was modified in response to SAB and internal EPA comments.  The revised model estimates
37    lifetime cancer incidence data based on "background" cancer incidence and mortality data from
38    the NCI SEER program (see Section 5.3.7.3).  The revised approach is discussed in detail in
39    Appendix E.2.
                                               A-5       DRAFT—DO NOT CITE OR QUOTE

-------
 1          As discussed in the previous response, EPA conducted sensitivity analyses on a number
 2    of model parameters.  These analyses are described in Section 5.3.8.3.
 O

 4    Charge Question D4
 5
 6          In calculating estimated cancer risk to the U.S. general population from drinking water
 7    exposure to inorganic arsenic, the EPA used epidemiologic data from Taiwan. EPA followed the
 8    NRC (2001) recommendations to account for the differences in the drinking water consumption
 9    rates for the Taiwanese population and U.S. populations. On the basis of more recent data
10    (noted in U.S. EPA, 2005b), EPA used water intake adjustments for 2 to 3.5 liters/day. EPA
11    asked SAB to recommend a drinking water value.
12
13    SAB Comments
14
15          The Panel agreed that water consumption (via drinking as water, in beverages, or in
16    cooking water) assumptions have a substantial impact on the assessment of arsenic's risk.
17    However, the Panel did not recommend specific values for EPA to use in evaluating dose-
18    response in the Taiwanese study nor for levels of exposure in the U.S. population risk estimates.
19    It did recommend that uncertainty in this parameter be evaluated for both the Taiwanese study
20    population and the U.S. populations at risk. The Panel recommended that EPA should:
21
22          1)     Evaluate the impact of drinking water consumption rates associated with more
23                 highly exposed population groups with differing exposures and susceptibilities
24                 (e-g-, children, pregnant women).
25          2)     Incorporate variability parameters for individual water consumption into their
26                 analysis for dose-response in the Taiwanese population, as they have done for
27                 the U.S. population.
28          3)     Conduct sensitivity analyses of the impact of using a range of
29                 consumption values for the Taiwanese population.
30          4)     Provide a better justification for assuming different consumption levels by gender
31                 or, in the absence of such a justification, conduct additional sensitivity analyses to
32                 examine the impact of equalizing the gender-specific consumption level.
33          5)     More fully articulate and document how different sources of water intake, as
34                 well as variability, are incorporated into the risk model (e.g., data for intake from
35                 beverages and cooking water).

36    EPA Response
37
38          Data are not available regarding individual water consumption rates and background
39    (nonwater) arsenic intake in the Taiwanese study populations. EPA, therefore, conducted a
40    series of sensitivity analyses involving ranges of drinking water consumption and "background"
41    (nonwater) arsenic consumption that the Agency believes spans a reasonable range of values for
42    these parameters. Arsenic dose-response models were fit assuming nonwater arsenic intakes of

                                               A-6       DRAFT—DO NOT CITE OR QUOTE

-------
 1    0, 10, 30, 50, 100, and 200 ug/day in the exposed populations, nonwater arsenic intake of 0, 30,
 2    and 50 ug/day in the reference population, and daily water consumption ranging from 2.75 to 5.1
 3    L/day for (Taiwanese) males and water consumption ranging from 2.0 to 4.1 L/day for females.
 4    Risk models also were fit using three different sets of village arsenic drinking water
 5    concentrations (median, minimum, and maximum), and three sets of assumptions related to
 6    reference (unexposed) populations (southwest Taiwan, all Taiwan, and none).  The results of
 7    these analyses are summarized in Tables 5-8 and 5-9. Overall, EPA found that cancer slope
 8    estimates for male and female lung cancer and male bladder cancer were relatively insensitive to
 9    assumptions related to nonwater arsenic intake and varied more or less inversely with the
10    assumed daily water consumption, and with drinking water arsenic concentration estimates.
11    When alternative reference populations were assumed (all Taiwan or none), cancer slope
12    coefficients were lower than when the southwest Taiwan comparison group was included in the
13    analysis.  The cancer slope estimates for female bladder cancer were generally more sensitive to
14    changes in exposure assumptions than the other endpoints.
15
16    Charge Question D5
17
18          As recommended by NRC (2001), EPA considered the background dietary intake of
19    inorganic arsenic and incorporated adjustment values of 0, 10, 30, and 50 ug per day into the
20    cancer modeling based on available new data.  SAB was asked to recommend a value for the
21    background dietary  intake of inorganic arsenic for both the control population and study
22    population of southwestern Taiwan.
23
24    SAB Comments
25
26          The Panel agreed that arsenic levels in food are important considerations for EPA's
27    assessment of lung and bladder cancer risk associated with exposures to arsenic in drinking
28    water.  However, the Panel did not recommend a specific value for EPA to use  in its base risk
29    assessment. It did recommend a range of values for consideration by EPA in its sensitivity
30    analysis and the Panel offered suggestions to EPA for additional analytical steps to clarify the
31    impact of food levels of arsenic on dose-response and exposure as it revises its  risk estimates.
32    These Panel recommendations include that EPA should:
33
34          1)     Conduct sensitivity analyses using a range of total arsenic food intake values
35                 from at least 50 to 100 ug /day to perhaps as high as 200 ug/day to assess the
36                 impact of this range of dietary intakes on risk of lung and bladder cancer from
37                 exposure via drinking water in the Taiwan cohort.
38          2)     Not assume that the control population has an intake value of zero arsenic from
39                 food.
40          3)     Apply greater rigor in their discussions of data used in these assessments (e.g.,
41                 sources, methodological and analytical issues, bioavailability).
                                               A-7       DRAFT—DO NOT CITE OR QUOTE

-------
 1          4)     Give immediate research attention to the issue of arsenic bioavailability.
 2
 3   EPA Response

 4                 As discussed in the previous response, EPA conducted sensitivity analyses that
 5   assumed nonwater arsenic intakes (doses) for the exposed populations ranging from 0 to 200
 6   ug/day and ranging from 0 to 50 ug/day in the reference population. EPA did not specifically
 7   conduct sensitivity analyses related to arsenic bioavailability. The Agency notes, however, that
 8   the range of absorbed dose that was evaluated implicitly addresses potential bioavailability
 9   differences. For example, assuming 50 ug arsenic intake absorbed dose is equivalent to
10   assuming 50% of absorption of a 100 ug/day dose, etc. The Agency believes that the range of
11   arsenic intake that was considered covers the plausible ranges of nonwater dietary arsenic and
12   bioavailability thereof.
                                                A-8       DRAFT—DO NOT CITE OR QUOTE

-------
            APPENDIX B. TABULAR DATA ON CANCER EPIDEMIOLOGY STUDIES
 1          The SAB Arsenic Review Panel provided comments on key scientific issues associated
 2    with arsenicals on cancer risk estimation in July 2007 (SAB, 2007). It was concluded that the
 3    Taiwanese database is still the most appropriate source for estimating bladder and lung cancer
 4    risk among humans (specifics provided in Section 5) because of: (1) the size and statistical
 5    stability of the database relative to other studies; (2) the reliability of the population and
 6    mortality counts; (3) the stability of residential patterns; and (4) the inclusion of long-term
 7    exposures. However, SAB also noted considerable limitations within this data  set (SAB, 2007).
 8    The Panel suggested that one way to mitigate the limitations of the Taiwanese database would be
 9    to include other relevant epidemiological studies from various countries. For example, SAB
10    referenced other databases that contained studies of populations also  exposed to high levels of
11    arsenic (e.g., Argentina and Chile), and recommended that these alternate sources of data be used
12    to compare the unit risks at the higher exposure levels that have emerged from the Taiwan data.
13    SAB also suggested that, along with the Taiwan data, published epidemiology studies from the
14    United States and other countries where the population is chronically exposed to low levels of
15    arsenic in drinking water (0.5 to 160 ppb) be critically evaluated, using a uniform set of criteria
16    presented in a narrative and tabular format.  The relative strengths and weaknesses of each study
17    should be described in relation to each criterion. Additionally, SAB (2007) recommended
18    considering the following issues when reviewing "low-level" and "high-level"  studies: (1)
19    estimates of the level of exposure misclassification, (2) temporal variability in assigning past
20    arsenic levels from recent measurements, (3) the extent of reliance  on imputed  exposure levels,
21    (4) the number of persons exposed at various estimated levels of waterborne arsenic, (5) study
22    response/participation rates, (6) estimates of exposure variability, (7) control selection methods
23    in case-control studies, and (8) the resulting influence of these factors on the magnitude and
24    statistical stability of cancer risk estimates.
25          In light of the SAB recommendations, epidemiological studies in the literature from 1968
26    to 2007 have been reviewed.  The report includes data from all populations that have been
27    examined in regard to cancer from arsenic exposure via drinking water. Earlier publications
28    were reviewed and are included as needed to facilitate the understanding of results from certain
29    study populations.  As recommended by SAB, studies were presented in both a narrative
30    (Section 4.1) and tabular (below) format.  Each publication was evaluated using a uniform set of
31    criteria, including the study type, the size of the  study population and control population, and the
32    relative strengths and weaknesses of the study, focusing on the major strengths and weaknesses.
33    While the information in the tables mirrors the information in the narrative, the narrative may
34    provide additional important information concerning the investigation.  The studies are presented
                                                B-1    DRAFT—DO NOT CITE OR QUOTE

-------
 1    by country of origin, then in chronological order by publication year.  Below also are definitions
 2    of terms that are used in the tables (and the narratives in Section 4.1).
 3           Cross-sectional studies have inherent limitations including: (1) difficulty in making
 4    causal inference; (2) the fact that data are collected for only one point in time, so that different
 5    results may be found if another time-frame had been chosen; and (3) prevalence-incidence bias
 6    (also called Neyman bias), which is especially prevalent for longer-lasting diseases, where any
 7    risk factor that results in death will be under-represented among those with the disease.
 8          Ecological studies provide low cost, convenience, simplicity of analysis, and ease of
 9    exposure measurement at population or group level rather than at the individual level; therefore,
10    a wider range of exposures can often be obtained.  Concerns about the methodological weakness
11    of ecological studies arise from three facts: estimates of effect do not equate to estimates of
12    biological effect obtained from individual level analysis, exposure data from this design cannot
13    be used to obtain direct estimates of the rate of injury in exposed and unexposed populations,
14    existing data sources are often flawed, and it is difficult to control confounding.
15           Cohort studies are research studies in which the medical records of groups of individuals,
16    who are alike in many ways, but differ by a certain characteristic (for example, individuals who
17    smoke and those who do not smoke) are compared for a particular outcome (such as lung
18    cancer). Cohort studies are generally used to follow large groups over a long period to study rare
19    or long-latency diseases.
20           A case-control study is a retrospective  study that compares two groups of people: those
21    with the disease or condition under study (cases) and a very similar group of people (matched
22    controls) who do not have the disease or condition. Researchers study the medical and lifestyle
23    histories of the people in each group to determine  which factors may be associated with the
24    disease or condition under investigation. An example is where one group may have been
25    exposed to a particular substance that the other was not.
26           In a nested case-control study, cases of a disease that occur in a defined cohort are
27    identified and, for each, a specified number of matched controls is selected from among those in
28    the cohort who have not developed the disease by the time of disease occurrence in the case.
29    The nested case-control design can potentially offer a lower cost and effort for data collection
30    and analysis compared with the full cohort approach, with relatively minor loss in statistical
31    efficiency.  The nested case-control design is particularly advantageous for studies of biologic
32    precursors of disease.
33          Recall bias is a type of systematic bias  that occurs when the way a survey  respondent
34    answers a question is affected not just by the correct answer, but also by the respondent's
35    memory.
36          Selection bias is the error of distorting a statistical analysis due to the methodology of
37    how the samples were collected. As an example, sample selection may involve pre- or post-

                                                 B-2   DRAFT—DO NOT CITE OR QUOTE

-------
 1    selecting the samples that may preferentially include or exclude certain kinds of results.
 2    Selection bias is possible whenever the group of people being studied has any form of control
 3    over whether to participate making the participants a non-representative sample. Selection bias
 4    may also occur when investigators preferentially select individuals to be included as cases or
 5    controls based on prior knowledge of study hypotheses or outcomes.  Selection bias in
 6    epidemiology is a distortion of data that arises from the way that the data have been collected. If
 7    the selection bias is not taken into account, conclusions drawn from the results obtained may be
 8    wrong. Self-selection bias is when individuals who make up the study population have any
 9    control over whether or not they are allowed to participate.  An individual's decision to
10    participate in a study may be associated with other factors that affect the study, which results in
11    the participants being a non-representative sample.
12           The standardized mortality ratio (SMR) in epidemiology is the ratio of observed deaths
13    to expected deaths in a population for a specific health outcome. The SMR also serves as an
14    indirect means for adjusting a rate. The number of observed deaths is obtained for a particular
15    sample of a population that is under investigation, and the number of expected deaths reflects the
16    number of deaths for a larger population from  which the study sample has been taken.  The
17    calculation used to determine the SMR is simply the number of observed deaths divided by the
18    number of expected deaths. The SMR may be displayed as either a ratio or sometimes as a
19    percentage. If the  SMR is shown as a ratio and is equal to  1.0, this means the number of
20    observed deaths equals that of expected cases.  If the SMR is greater than 1.0 there is a higher
21    number of deaths than expected, and if the SMR is less than 1.0 there is a lower number of
22    observed than expected deaths.
23           The standardized incidence ratio (SIR) is a common tool for monitoring disease rates.
24    Incidence is the number of newly diagnosed cases in a given location during a given time period.
25    An SIR compares the actual number of cases for a given place and time to  the number that
26    would be expected based on disease rates in some comparison area.
27           In statistics and epidemiology, relative risk (RR) is the risk of an event (or of developing
28    a disease) relative to exposure.  Relative risk is a ratio of the probability of the event occurring in
29    the exposed group versus the control (non-exposed) group.
30           Time-weighted average (TWA) is the average exposure to a contaminant or condition
31    (such as noise) to which workers are exposed over a period, such as in an 8-hour work day.
                                                B-3   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-l. Taiwan Cancer Studies
Study
Period
Not
indicated


















































Subjects/
Controls
19,269 males
2 1,1 52 females
(40,421 total)

















































Exposure
Assessment
Arsenic
concentration
in well water
(ppb):
low (L) = 0-
290
mid (M) =
300-590
high(H) =
>600
undetermined
(U)








































Study Outcome
Age-/gender-specific
skin cancer prevalence
rate (1/1000) by arsenic
concentration (L, M, H,
U):
Males, 20-39 yrs.—
L=1.5,M = 4.3,H =
22.4, U= 1.7
Males, 40-59 yrs.—
L = 6.5, M = 47.7, H =
98.3, U = 51.7
Males, 60 yrs. and
over —
L = 48. 1,M= 163.4, H
= 255. 3, U= 148.2
Total all males
combined —
L = 4.0, M= 14.4, H =
31.0, U= 16.5
Females, 20-3 9 yrs.—
L = 0.1,M = 0.7,H =
3.5, U = 0.9
Females, 40-59 yrs. —
L = 3.6, M= 19.7, H =
48.0, U = 9.2
Females, 60 yrs. and
over —
L = 9.1, M = 62.0, H =
110.1,U = 62.9
Total all females
combined —
L= 1.3,M = 6.3,H =
12.1, U = 4.7
Both genders, 20-39
yrs. —
L=1.3,M = 2.2,H =
11.5,U=1.2
Both genders, 40-59
yrs.—
L = 4.9, M= 32.6, H =
72.0, U = 28.3
Both genders, 60 yrs.
and over —
L = 27. 1,M= 106.2, H
= 192.0,
U= 107.9
Total both genders
combined —
L = 2.6, M= 10.1, H =
21.4, U= 10.4
Observed rate/1000:
hyperpigmentation =
Strengths/
Weaknesses
Strengths:
-Large number of
participants.
-Dose-response
information provided.

Weaknesses:
-No individual
exposure data.
-Possible recall bias
among study
participants in
determining the age
of cancer onset and
length of residence in
the study area.
-Water supply
changes over time
were not collected,
nor was information
on smoking histories;
the arsenic
concentration from
individual wells
varied over time.



























Reference/
Type of
Study
Tseng et al.,
1968
Ecological

















































                                    B-4   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period





1958-
1975





























Subjects/
Controls





40,421
individuals





























Exposure
Assessment





Arsenic
concentration
in well water
(ppb):
<300 = low
(L)
300-600 =
mid(M)
>600 = high
(H)






















Study Outcome
183.52
keratosis = 70.95
skin cancer = 10.59
blackfoot disease =
8.91
Age-specific prevalence
(per 1000):
Skin cancer
20-3 9 y ears—
L=1.3,M = 2.2,H =
11.5
40-59 years—
L = 4.9, M= 32.6, H =
72.0
60+ years —
L = 27. 1,M= 106.2, H
= 192.0
Blackfoot disease
20-3 9 years—
L = 4.5, M= 13.2, H =
14.2
40-59 years—
L = 10.5, M = 32.0, H =
46.9
60+ years —
L = 20.3, M = 32.2, H =
61.4
Skin cancer and BFD
combined:
observed — 61 cases,
1.51/1000
expected — 4 cases,
0.09/1000
observed to expected
ratio = 16.77

Strengths/
Weaknesses





Strengths:
-Large study
population.
-Adjusted for age and
gender.

Weaknesses:
-No individual
monitoring data.
-Possible recall bias
among study
participants
(interviews and
mailed surveys) in
determining the age
of cancer onset and
the length of
residence in the area.












Reference/
Type of
Study





Tseng, 1977
Ecological




























B-5   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1968-
1982
























Subjects/
Controls
Subjects from
BFD-endemic
area























Exposure
Assessment
Median
arsenic
concentration
(ppb):
artesian well
water — 780
(range: 350-
1140)
shallow well
water — 40
(range: 0-300)















Study Outcome
Cancer SMRs
(95% CI,p value
O.05):
Males —
bladder =11.00 (9.33-
12.67)
kidney = 7.72 (5.37-
10.07)
skin = 5. 34 (3.79-8.89)
lung =3.20 (2.86-3. 54)
liver = 1.70(1.51-1.89)
colon =1.60 (1.17-
2.03)
Females —
bladder = 20.09 (17.02-
23.16)
kidney =11. 19 (8.38-
14.00)
skin = 6.52 (4.69-8.35)
lung = 4.13 (3.60-4.66)
liver =2.29 (1.92-2.66)
colon =1.68 (1.26-
2.10)



Strengths/
Weaknesses
Strengths:
-The SMRs for the
study cohort taken
from BFD endemic
area in Taiwan were
determined using the
general population of
Taiwan and world
population.
-Controlled for the
potential confounders
age and gender.

Weakness:
-Arsenic
measurements not
linked to cancer
mortality.
- Death certificates
list the main cause of
death rather than all
causes
- SMRs were only
presented by
township and
villages.
Reference/
Type of
Study
Chenetal.,
1985
ecological























B-6   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
January
1980-
December
1982


























Subjects/
Controls
Deceased
cancer cases:
69 bladder
76 lung
59 liver

Controls: 368
(community
matched)





















Exposure
Assessment
Median
arsenic
concentration:
artesian well
water — 780
ppb (range:
350-1140)
shallow well
water — 40 ppb
(range: 0-300)




















Study Outcome
Age-/sex-adjusted odds
ratios, well water use
>40 years:
bladder cancer =3.90
lung cancer = 3.39
liver cancer = 2.67

Mantel-Haenszel x2:
bladder cancer = 13.74*
lung cancer = 8.49*
liver cancer = 9.01*
*p<0.01

Multivariate logistic
regression:
improvement x2
value —
bladder cancer = 11.45*
lung cancer = 9.04*
liver cancer = 6.34*
*p<0.01









Strengths/
Weaknesses
Strengths:
-Cases confirmed
using histology or
cytology findings.
-Cancer cases and
controls were from
the same BFD
community.
-Potential
confounders adjusted
for in the analysis
included age, gender,
smoking, tea
drinking, vegetable
consumption, and
fermented bean
consumption.

Weaknesses:
-Confounders not
controlled for
included recall bias
from case and control
interviews regarding
lifestyle, diet, and
daily water
consumption and
source of water.
-Selection bias
(control selection).
Reference/
Type of
Study
Chenetal.,
1986
Case-
control


























B-7   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1973-
1986




















Subjects/
Controls
Blackfoot-
endemic area
residents
Population of
Taiwan as
reference
population
World
population as
reference
population

















Exposure
Assessment
Three
exposure
categories
(ppb):
<300
300-590
>600




















Study Outcome
Age-standardized
mortality per 100,000
for various cancers:
World population:
<300 ppb
Males — all sites =
154.0, liver = 32.6, lung
= 35.1, skin =1.6,
prostate = 0.5, bladder
= 15.7, kidney = 5.4
Females — all sites =
118.8, liver = 14.2, lung
= 26.5, skin =1.6,
bladder = 16.7, kidney
= 3.6
300-590 ppb
Males — all sites =
258.9, liver =42.7, lung
= 64.7, skin =10.7,
prostate = 5.8, bladder
= 37.8, kidney =13.1
Females — all sites =
182.6, liver= 18.8, lung
= 40.9, skin = 10.0,
bladder =35.1,
kidney = 12.5
>600 ppb
Males — all sites =
434.7, liver =68.8, lung
= 87.9, skin =28.0,
prostate = 8.4, bladder
= 89.1, kidney = 2 1.6
Females — all sites =
369.4, liver =3 1.8, lung
= 83.8, skin =15.1,
bladder =91. 5,
kidney = 35.3
Taiwan:
Males — all sites =
128.1, liver =28.0, lung
= 19.4, skin =0.8,
prostate = 1.5, bladder
= 3.1, kidney = 1.1
Females — all sites =
85.5, liver =8.9,
lung = 9.5, skin = 0.8,
bladder = 1.4, kidney =
0.9
Strengths/
Weaknesses
Strengths:
-Data from arsenic
monitoring conducted
in 1962-64 and
1974-76 found
similar results.
Weaknesses:
-Individual arsenic
exposure levels were
not presented.


















Reference/
Type of
Study
Chenetal.,
1988a
Cohort




















B-8   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
January
1968-
December
1983
















August
1983-
February
1987




















Subjects/
Controls
241 cases

759 controls

General
population of
Taiwan

Local endemic
area population










246 BFD
bladder cancer
cases

444 BFD-
endemic area
residents

286 residents
neighboring the
endemic area

731 non-
endemic
area residents









Exposure
Assessment
Arsenic
concentration
(ppb): artesian
well water —
median = 780
range = 350-
1140 shallow
well water —
median = 40
range = 0-300










Percent of area
well water
with arsenic
content of

>50 ppb:
Pei-men= 81
Hsueh-Chia =
27
Pu-Tai = 58
Jinag-Jium =
24
Tai-Pao = 45
Pao-Chung =
54
>350 ppb:
Pei-men = 62
Hsueh-Chia =
7
Pu-Tai = 8
Jinag-Jium = 0
Tai-Pao = 6
Pao-Chung = 0


Study Outcome
Significant SMRs (p
values) (compared to
population of Taiwan):
Cancers —
bladder =3 8. 80
(0.001)
skin = 28.46 (O.01)
lung= 10.49(0.001)
liver =4.66(0.001)
colon =3.81(0.05)

Significant SMRs (p
values) (compared to
population of BFD-
endemic area):
Cancers —
bladder = 2.55 (0.01)
skin = 4.51 (O.05)
lung = 2.84(0.01)
liver =2.48(0.01)
Positive cytology
(bladder cancer/atypia)
prevalence rate (%):
BFD cases = 4.5
endemic area = 2.5
neighboring area = 0.7
non-endemic area =
0.13
















Strengths/
Weaknesses
Strengths:
-Cases consisted of
blackfoot disease
cases, matched to
healthy community
controls for age, sex,
and residence.
-Recall bias was
minimized through
interview techniques.
-SMRs were
determined using both
the national
Taiwanese population
and the local endemic
area population.

Weakness:
-Arsenic dose levels
were not provided.
Strengths:
-Histological
confirmation of
bladder cancer
diagnoses.

Weaknesses:
-Lack of individual
exposure data.














Reference/
Type of
Study
Chenetal.,
1988b
Cohort/
nested case-
control















Chiang et
al., 1988
Case-
control



















B-9   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
1973-
1986
















1972-
1983





























Subjects/
Controls
Residents of 42
villages

1976 world
population used
as comparison












Arsenic-
exposed
subjects from
314 townships
and precincts


























Exposure
Assessment
Three
exposure
categories
(ppb):
<300
300-590
>600











Total wells
tested =
83,656,
>50 ppb in
15,649 wells
(18.7%),
> 350 ppb in
2,224 wells
(2.7%)

Concentrations
in the
remainder of
the wells were
not given

















Study Outcome
Trend test of the
extension of the
Mantel-Haenszel Chi
square test:
Cancers —
Both genders:
bladder, skin, lung —
p< 0.001
Males only:
kidney, liver,
prostate— p < 0.05
Females only:
kidney— p< 0.001





Multivariate adjusted
regression coefficient
for cancers (SE):
Males —
liver =6.8 (1.3), nasal
cavity = 0.7(0.2), lung
= 5.3(0.9), skin = 0.9
(0.2), bladder =3. 9
(0.5), kidney = 1.1
(0.2), prostate = 0.5
(0.2)
Females —
liver = 2.0 (0.5), nasal
cavity = 0.4 (0.1), lung
= 5.3(0.7), skin = 1.0
(0.2), bladder = 4.2
(0.5), kidney = 1.7(0.2)
No p values indicated.













Strengths/
Weaknesses
Strengths:
-Adjustments made
for age and gender.
-Lifestyle, access to
medical care, and
socioeconomic status
were similar among
the study groups.

Weaknesses:
-Limitations of
mortality data.
-Associations
observed at the local
level may not be
accurate at the
individual level
(ecological fallacy).
Strengths:
-Potential
confounders
controlled for
included
socioeconomic
differences, i.e.,
urbanization and
industrialization.
-Cancer rates in
endemic BFD
townships were
compared with cancer
rates in non-endemic
townships of Taiwan.
-Ecological
correlations reported
between arsenic
content in well water
and mortality from
various cancers.
Weaknesses:
-Potential
confounders not
controlled for were
gender and other
potential well water
contaminants.
-No individual arsenic
exposures.
Reference/
Type of
Study
Wu et al.,
1989
Ecological















Chen and
Wang,
1990
Ecological


























B-10  DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
1973-
1986













































Followed
up for
0.05-7.69
years
(4.97
±1.72
[SD]

Subjects/
Controls
Arsenic-
exposed
subjects from
42 villages











































263 BFD cases
2,293 healthy
residents





Exposure
Assessment
Well water
arsenic
exposure
categories
(ppb):
<100
100-290
300-590
>600
Overall range:
10-1,752




































Artesian well
water median
arsenic level =
780 ppb

Shallow well
water median


Study Outcome

Cancer development
potency index (daily
arsenic intake of 10
|ig/kg):
Males —
liver = 4.3 x 10'3
lung= 1.2 x 10'2
bladder =1. 2 x io~2
kidney = 4.2 x 10'3
Females —
liver =3. 6 x 10'3
lung= 1.3 x 10'2
bladder =1. 7 x io~2
kidney = 4.8 x 10'3
































Multivariate adjusted
RR (95% CI), cancer:
All sites —
Age: every -1-yr
increment = 1.05
(1.03-1.06)*
Sex: men = 1.00,

Strengths/
Weaknesses
Strengths:
-Potential
confounders included
age, gender, access to
medical care,
socioeconomic status,
and lifestyle and were
all controlled for in
the analysis.
-Villages share
similar
socioeconomic status,
living environments,
lifestyles, dietary
patterns, and even
medical facilities.
Weaknesses:
-Armitage-Doll model
constrains risk
estimates to be
monotonically
increasing function of
age.
-Age stratification
only available for 20-
year strata.
-Possible
underestimation of
risk because it was
assumed that an
individual's arsenic
intake remained
constant from birth to
the end of the follow-
up period.
-Assumption that an
individual's arsenic
intake remained
constant from birth to
the end of the follow-
up period and the
possible
underestimation of
risk because other
sources of arsenic
exposure were not
considered.
Strengths:
-Showed a significant
dose-response
relationship with
increasing
concentrations of
arsenic.
Reference/
Type of
Study
Chenetal.,
1992
Ecological












































Chiou et al.,
1995
Cohort




B-11  DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
years)
until
January
1993



















































Subjects/
Controls























































Exposure
Assessment
arsenic level =
40ppb





















































Study Outcome
women = 0.72
(0.43-1.18)*
Cigarette smoking:
no= 1.00, yes = 1.52
(1.00-2.48)*
Status of blackfoot
disease:
no= 1.00, yes = 2.69
(1.80-4.01)*
Cumulative arsenic
exposure (mg/liter x
yr):
0= 1.00
0.1-19.9= 1.39(0.82-
2.37)
20+ =1.76 (1.01-
3.06)*
unknown = 0.72 (0.42-
1.22)
Lung-
Age: every -1-yr
increment = 1.06(1.02-
1.10)*
Sex: men= 1.00,
women = 1.79(0.44-
7.32)*
Cigarette smoking:
no= 1.00, yes = 4.31
(1.08-17.20)*
Status of blackfoot
disease:
no= 1.00, yes = 2.45
(1.07-0.57)*
Cumulative arsenic
exposure (mg/liter x
yr):
0=1.00
0.1-9.9 = 2.74(0.69-
11.0)
20+ = 4.01 (1.00-
16.12)*
unknown =2.01 (0.55-
7.36)
Bladder-
Age: every 1-yr
increment = 1.04 (1.05-
1.08)*
Sex: men= 1.00,
women =0.45 (0.18-
1.16)
Cigarette smoking:
no= 1.00, yes = 1.00
(0.37-2.31)
Status of blackfoot
disease:
Strengths/
Weaknesses
-Analysis adjusted for
BFD status, age, sex,
and smoking.
-Reported incidence
data.

Weaknesses:
-Artesian well water
arsenic concentration
was unknown for
some study subjects.












































Reference/
Type of
Study























































B-12   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period













January
1980-
December
1987




















Subjects/
Controls













2,915 urinary
cancer cases






















Exposure
Assessment













6 categories of
arsenic
exposure
(ppb):
<50
50-80
90-160
170-320
330-640
>640















Study Outcome
no= 1.00, yes = 4.41
(2.06-9.45)*
Cumulative arsenic
exposure (mg/liter x
yr):
0= 1.00
0.1-19.9=1.57(0.44-
5.55)
20+ = 3.58 (1.05-
12.19)*
unknown = 1.25(0.38-
4.12)
*p < 0.05
Rate differences (SE)*
with positive
associations:
Males —
Bladder cancer:
transitional cell
>640 ppb = 0.57(0.07),
adenocarinoma
>640 ppb =
0.027(0.008)
Kidney cancer:
transitional cell
330-640 ppb =
0.05(0.02)
Females —
Urethral cancer, all cell
types combined
>640 ppb =
0.027(0.007)
*Estimates for 1 unit
increase (1%) in
predictor (exposure
category)

Strengths/
Weaknesses













Strengths:
-Adjusted for age,
gender, urbanization,
and smoking.

Weaknesses:
- Limitations of
ecological study
design.














Reference/
Type of
Study













Guo et al.,
1997
Ecological




















B-13  DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1971-
1994





























Subjects/
Controls
11,193
mortalities
from all causes
of disease

Local reference
population

National
reference
population




















Exposure
Assessment
Median
artesian wells
water arsenic
content: 780
ppb
(range = 250-
1140 ppb)

Individual
exposure data
not available




















Study Outcome
Males —
BFD area compared to
local reference — SMR
(95% CI):
all cancers = 2.19
(2.11-2.28)
BFD area compared to
national reference —
SMR (95% CI):
all cancers =1.94
(1.87-2.01)
Females —
BFD area compared to
local reference — SMR
(95% CI):
all cancers = 2.40
(2.30-2.51)
BFD area compared to
national reference —
SMR (95% CI):
all cancers = 2.05
(1.96-2.14)
p<0.05








Strengths/
Weaknesses
Strengths:
-Exposed group and
local reference group
had similar lifestyle
factors.
-All cancers were
pathologically
confirmed.
-Controlled for
gender, a potential
confounder.

Weaknesses:
-Only one underlying
cause of death (not
multiple causes) was
indicated on death
certificate, resulting
in possible distortion
of association
between exposure and
disease.
-Lack of individual
exposure data.
-Potential
confounders not
controlled for were
age, smoking, alcohol
consumption, and
occupational
exposures.
Reference/
Type of
Study
Tsaietal.,
1999
Cross-
sectional



























B-14   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1973-
1986










































Subjects/
Controls
42 arseniasis-
endemic
villages

Population of
Taiwan






































Exposure
Assessment
Arsenic
exposure
categories
(ppb) =
0-50
50-100
100-200
200-300
300^00
400-500
500-600
600+
































Study Outcome
SMRs (male and female
combined.)
Bladder cancer SMRs:*
0-50 ppb = 10.02
50-100 ppb = 4. 15
100-200 ppb = 10.47
200-300 ppb = 7.66
300-400 ppb = 7.44
400-500 ppb = 29.68
500-600 ppb = 14.90
600+ ppb = 32.71

Lung cancer SMRs:*
0-50 ppb = 1.56
50-100 ppb = 1.43
100-200 ppb = 2.43
200-300 ppb = 3. 08
300-400 ppb = 1.97
400-500 ppb = 3. 65
500-600 ppb = 3. 32
600+ ppb = 5. 14

Liver cancer SMRs:*
0-50 ppb = 1.18
50-100 ppb = 0.65
100-200 ppb = 1.74
200-300 ppb = 1.44
300-400 ppb = 0.77
400-500 ppb =1.60
500-600 ppb =1.59
600+ ppb = 2. 17
Bladder, lung, and liver
combined cancer
SMRs:*
0-50 ppb = 1.83
50-100 ppb =1.16
100-200 ppb = 2.51
200-300 ppb = 2.47
300-400 ppb = 1.63
400-500 ppb = 3. 93
500-600 ppb = 3. 06
600+ ppb = 4.86
*No significance levels
presented.
Strengths/
Weaknesses
Strengths:
-Person-years at risk
stratified by age,
gender, and arsenic
level.
-Individual well
concentrations were
available for each
village.

Weaknesses:
-Ecological study
design (no individual
monitoring data,
individual exposures
not available).
-Potential
confounding by
smoking, use of
bottled water, and
dietary intake, since
this information was
not available.





















Reference/
Type of
Study
Morales et
al., 2000
Ecological









































B-15  DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
October
1991-
September
1994 with
follow-up
through
the end of
1996





























Subjects/
Controls
8, 102 residents
(4,056 men and
4,046 women)

General
population of
Taiwan used as
comparison





























Exposure
Assessment
Exposure
categories
(ppb) :
<10.00
10.1-50.0
50.1-100.0
>100.0






























Study Outcome
Standardized incidence
ratio (95% CI):
urinary cancer = 2.05
(1.22, 3.24)
bladder =1.96 (0.94-
3.61)
kidney = 2.82 (1.29-
5.36)
p<0.05

Multivariate adjusted
RR (95% CI):
Well water arsenic
concentration (ppb):
Urinary organs —
10.1-50.0=1.5(0.3-
8.0)
50.1-100.0 = 2.2(0.4-
13.7)
>100.0 = 4.8 (1.2-19.4)
TCC
10.1-50.0=1.9(0.1-
32.5)
50.1-100.0 = 8.2(0.7-
99.1)
>100.0= 15.3(1.7-
139.9)










Strengths/
Weaknesses

Strengths:
- Showed a
significant dose-
response relationship
with increasing
concentrations of
arsenic.
-Potential
confounders adjusted
for included age,
gender, and smoking.
-Individual exposure
estimates were
available.

Weaknesses:
-Possible diagnosis
bias, since data were
collected from
various community
hospitals.
-Possible recall bias
resulting from self-
reported information.
- Short duration of
follow-up, which
limited the number of
person-years of
observation.
-Possible
misclassification,
especially in the low-
dose region due to
lack of arsenic
exposure information
in the food.
Reference/
Type of
Study
Chiou et al.,
2001
Cohort


































B-16  DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
January
1980-
December
1989





















January
1980-
December
1999
















Subjects/
Controls
2,369 skin
cancer cases
(1,415 men and
954 women)





















40,832 liver
cancer patients
(32,034 men
and 8,798
women)















Exposure
Assessment
6 categories of
arsenic
exposure
(ppb):
<50,
50-80,
90-160,
170-320,
330-640,
>640















BFD area
average
arsenic
concentration
= 220 ppb

Non-BFD area
average
arsenic
concentration
= 20 ppb










Study Outcome
Statistically significant
rate differences per
100,000 person-years
(SE):*
Males —
Basal cell carcinoma
>640 ppb =
0.128(0.025)**
Squamous cell
carcinoma
170-320 ppb =
0.073(0.024)**
330-640 ppb= -
0.10(0.031)**
>640 ppb =
0.155(0.028)**
Females —
Squamous cell
carcinoma
330-640 ppb = -
0.064(0.027)*
>640 ppb =
0.212(0.024)**
*p < 0.05
**p<0.01
No statistically
significant (P > 0.05)
differences were noted
for cell types of liver
cancer between the
BFD area and the other
areas.













Strengths/
Weaknesses
Strengths:
-Cases were identified
from government
operated National
Cancer Registration
Program.
-Pathological
classifications
determined by board-
certified pathologists.
-Potential
confounders adjusted
for in the analysis
included gender and
age.

Weaknesses:
-Limitations of
ecological study
design. (No
monitoring data were
presented.)



Strengths:
-Cases identified from
government operated
National Cancer
Registration Program.
-Pathological
classifications were
determined by board-
certified pathologists.
-Potential
confounders adjusted
for included gender
and age.
Weaknesses:
-Limitations of
ecological study
design. (No
monitoring data were
presented).
Reference/
Type of
Study
Guo et al.,
2001
Ecological






















Guo, 2003
Ecological

















B-17  DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
January
1985-
December
2000;
average
follow-up
of 8 years
















1971-
2000













Subjects/
Controls
2,503 residents
in southwestern
area

8,088 residents
in northeastern
area
















Residents of 4
BFD-endemic
townships












Exposure
Assessment
Southwestern
area average
arsenic
exposure
categories
(ppb):
<10
10-99.9
100-299.9
300-699.9
>700
Unknown











Median well
water arsenic
level, early
1960s = 780
ppb











Study Outcome
Multivariate-adjusted
RR of lung cancer for
average arsenic level in
well water (ppb):
<10= 1.00 (referent)
10-99.9=1.09(0.63-
1.91)
100-299.9 =2.28 (1.22-
4.27)
300-699.9 = 3.03
(1.62-5.69)
>700 = 3.29(1.60-
6.78)
Unknown = 1.10(0.60-
2.03)








SMR liver cancer:
Males —
1989-1991 = 1.868
1998-2000=1.242

Females —
1983-1985 = 2.041
1998-2000= 1.137







Strengths/
Weaknesses
Strengths:
-Confounders
controlled for were
age, gender,
education, and
alcohol consumption.
-Long follow-up
period and the use of
a national
computerized cancer
case registry.
-All lung cancer cases
were pathologically
confirmed.

Weaknesses:
-Historical monitoring
data not available.
-Possible
misclassification bias
because exposure
measurements were
based on one survey.
Strengths:
-Residents in the
study area were
similar in terms of
socioeconomic status,
living environments,
lifestyles, dietary
patterns, and health
service facilities.
-Accurate death
registration system.
Weaknesses:
-Limitations of
mortality data.
Reference/
Type of
Study
Chenetal.,
2004a
Cohort




















Chiuetal.,
2004
Cohort











B-18  DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
January
1971-
December
1990














Subjects/
Controls
1,078 lung
cancer
mortality cases















Exposure
Assessment
Arsenic
exposure
levels (ppb):
<050
50-80
90-160
170-320
330-640
>640









Study Outcome
Lung cancer mortality
increase with 1,000 ppb
increase in mean
arsenic level (p=0.01):
Men —
27.45/100,000 person-
years
Women —
18.93/100,00 person-
years








Strengths/
Weaknesses
Strengths:
-Adjusted for gender
and age.
-Cases were
ascertained using
information from
household registry
offices in each
township. Taiwanese
law requires timely
reporting of deaths to
these offices.
Weaknesses:
-Limitations of
ecological studies.
-Smoking was not
controlled for in the
analysis.
Reference/
Type of
Study
Quo, 2004
Ecological
















B-19  DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
1971-
2000
















1988-
2001




Subjects/
Controls
Residents of 4
BFD-endemic
townships
















7 females
14 males




Exposure
Assessment
Median
arsenic level
(ppb), early
1960s = 780
(range: 350-
1140)















No exposure
data




Study Outcome
Kidney cancer SMR
(observed vs.
expected):
1971—
Men= 19.04 (4 vs.
0.21)
Women = 23. 52 (8 vs.
0.34)
2000—
Men = 4.46 (8 vs. 1.79)
Women = 6. 52 (9 vs.
1.38)














Chi square (Taiwan
case series compared to
3 U.S. case series
studies):
Males —
urethra!
adenocarcinoma:
p< 0.0001

Strengths/
Weaknesses
Strengths:
-Adjusted for gender
and age.
-Mandatory
registering of all
births, deaths,
marriages, divorces,
and migration to the
Household
Registration Office in
Taiwan, making it an
accurate data source.
-Most residents had
similar
socioeconomic status,
living environments,
lifestyles, dietary
patterns, and health
service facilities and
worked in farming,
fisheries, or salt
production.
-All kidney cancer
cases in the area
probably had similar
access to medical
care.
Weaknesses:
-Mortality data
limitations.
-Cross-sectional study
limitations.
-Smoking may
possibly have been a
confounder not
adequately controlled
for.
Strengths:
-Cases were
pathologically
confirmed.
Weaknesses:
-Limited number of
cases.
-No exposure
information.
Reference/
Type of
Study
Yangetal.,
2004
Cross-
sectional
















Tsaietal.,
2005
Cross-
sectional



B-20   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1971-
2000






















Subjects/
Controls
Residents in 4
BFD-endemic
area townships





















Exposure
Assessment
Median
arsenic level,
early 1960s =
780 ppb




















Study Outcome
Bladder cancer SMRs
(observed vs.
expected):
1971—
Males = 10.25 (8 vs.
0.78)
Females = 14.89 (7 vs.
0.47)
2000—
Males = 2. 15 (5 vs.
2.32)
Females = 7.63 (10 vs.
1.31)











Strengths/
Weaknesses
Strengths:
-All bladder cancer
cases in the area
probably had similar
access to medical
care.
-Adjusted for age and
gender.
-Mandatory
registering of all
births, deaths,
marriages, divorces,
and migration to the
Household
Registration Office in
Taiwan, making it an
accurate data source.
Weaknesses:
-Limitations of a
cross-sectional
mortality study.
-Smoking may
possibly have been a
confounder.
Reference/
Type of
Study
Yangetal.,
2005
Cross-
sectional




















B-21   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-2. Japan Cancer Studies
Study
Period
1959-
1992





































Subjects/
Controls
454
residents





































Exposure
Assessment
Well arsenic
concentration
(ppb):
<50
50-990
>1000

































Study Outcome
>1000 ppb
SMRs (95% CI):
Males —
all deaths =1.88
(1.17-2.96)
all cancers = 4.19
(2.20-7.56)
lung cancer = 19.08
(8.88-38.76)
urinary cancer = 33.16
(5.92-121.58)
all cancers except lung
=
2.22 (0.87-5.22)
Females —
all deaths = 1.31
(0.76-2.18)
all cancers = 3.00
(1.40-6.13)
lung cancer = 7.15
(0.36-41.11)
urinary cancer = 27.85
(1.42-159.89)
all cancers except lung
2.73 (1.19-6.04)
Cox's proportional
hazard analysis (95%
CI), highest group vs.
background:
concentration
categories (ppb)
>1 000 vs. 1
all deaths = 1.74
(1.10-2.74)
all cancers = 4.82
(2.09-11.14)
lung cancer =
1,972.16(4.34-
895,385.11)
Strengths/
Weaknesses
Strengths:
-Cohort examined by 3
exposure categories.
-Included information on
smoking, age and gender.

Weaknesses:
-Lacking detailed arsenic
intake information.
-Small study population.
-Possible misclassification
bias.
-Recall bias (smoking
history)..

























Reference/
Type of
Study
Tsuda et al.,
1995
Cohort




































                                   B-22   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-3.  South America Cancer Studies
Study
Period
1986-
1991























Subjects/
Controls
Bladder
cancer
deaths in
26
Cordoba
counties

Population
of
Argentina















Exposure
Assessment
Exposure
categories:
low
medium
high
(crude average
estimate of 178
ppb)

Two counties in
high-exposure
group













Study Outcome
Bladder cancer SMR
(95% CI) by exposure
category:
Men —
low = 0.80 (0.66-
0.96)
medium = 1.42 (1.14-
1.74)
high =2.14 (1.78-
2.53)
test for trend:
p=0.001
Women —
low = 1.21 (0.85-
1.64)
medium = 1.58(1.01-
2.35)
high= 1.82(1.19-
2.64)
test for trend: p=0.04





Strengths/
Weaknesses
Strengths:
-Adjusted for age and
gender.
-Analysis restricted to
rural counties to limit
confounders.
-To account for cancer
diagnosis and detection
bias, stomach cancer,
which is known not to be
related to arsenic
exposure, was used as a
comparison cancer.

Weaknesses:
-Limitations of ecological
studies.
-Lack of comprehensive,
systematic monitoring
data.
-No arsenic exposure
levels in low and medium
groups reported.
-Lack of individual
smoking history.
Reference/
Type of
Study
Hopenhayn-
Richetal.,
1996a
Ecological





















                                    B-23  DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1986-
1991































Subjects/
Controls
Population
from 26
counties in
Cordoba

Population
of
Argentina

























Exposure
Assessment
Exposure
categories:
low
medium
high
(crude average
estimate of 178
ppb)

























Study Outcome
SMRs (95% CI) by
exposure categories:
Kidney cancer —
Men
low = 0.87 (0.66-
1.10)
medium = 1.33 (1.02-
1.68)
high =1.57 (1.17-
2.05)
Women
low= 1.00(0.71-
1.37)
medium = 1.36(0.94-
1.89)
high =1.81 (1.19-
2.64)
Lung cancer
Men
low = 0.92 (0.85-
0.98)
medium = 1.54 (1.44-
1.64)
high =1.77 (1.63-
1.90)
Women
low = 1.24 (1.06-
1.42)
medium = 1.34(1.12-
1.58)
high =2.16 (1.83-
2.52)
p< 0.001 in trend test
Strengths/
Weaknesses
Strengths:
-Adjusted for age and
gender.
-Analysis restricted to
rural counties to limit
confounders.
-To account for cancer
diagnosis and detection
bias, stomach cancer, that
is known not to be related
to arsenic exposure, as a
comparison cancer.

Weaknesses:
-Limitations of ecological
studies.
-Lack of comprehensive,
systematic monitoring
data.
-No arsenic exposure
levels in low and medium
groups reported.
-Lack of individual
smoking history.









Reference/
Type of
Study
Hopenhayn-
Richetal.,
1998
Ecological





























B-24   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1989-
1993
























Subjects/
Controls
390,340
residents
national
mortality
data from
1991

Population
of Chile
used as
reference
group














Exposure
Assessment
Region II average
water arsenic level
(ppb):
1950-1954 = 123
1955-1959 = 569
1960-1964 = 568
1965-1969 = 568
1970-1974 = 272
1975-1979 = 176
1980-1984 = 94
1985-1989 = 71
1990-1994 = 43














Study Outcome
SMRs (95% CI, p
value)
>30 years old:
Men —
bladder =6.0 (4.8-
7.4, 0.001)
kidney =1.6 (1.1-2.1,
0.012)
liver =1.1 (0.8-1. 5,
0.392)
lung =3.8 (3. 5-4.1,
O.001)
skin = 7.7 (4.7-11. 9,
0.001)

Women —
bladder =8.2 (6.3-
10.5,0.001)
kidney = 2.7 (1.9-3. 8,
O.001)
liver =1.1 (0.8-1. 5,
0.377)
lung =3. 1(2.7-3. 7,
0.001)
skin =3.2 (1.3-6.6,
0.016)
Strengths/
Weaknesses
Strengths:
-Large study size.
-Used national data for
comparison. No other
major populations in Chile
were exposed to arsenic in
drinking water.
-SMRs adjusted for age
and gender.

Weaknesses:
-Arsenic levels in drinking
water available only by
city or town.
-Deaths were not linked to
town so individual
exposure is not known.
-Limited smoking data.
-No dose-response
information provided.






Reference/
Type of
Study
Smith etal.,
1998
Ecological























B-25   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1994-
1996



















































Subjects/
Controls
152 lung
cancer
cases

419
controls















































Exposure
Assessment
Average water
arsenic
concentration (ppb)
during peak
exposure years:
0-10
10-29
30-59
60-89
90-199
200-399
400-699
700-999








































Study Outcome
Lung cancer odds
ratio (95% CI):
Age/gender
adjusted —
0-10 ppb = 1
(referent)
10-29 ppb = 0.4 (0.1-
0.5)
30-59 ppb = 0.0 (0.6-
7.2)
60-89 ppb = 0.1 (1.8-
9.2)
90-199 ppb = 0.8
(1.1-7.0)
200-399 ppb = 0.4
(2.0-10.0)
400-699 ppb = 0.9
(2.4-19.8)
700-999 ppb = 0.3
(3.1-12.8)
Male vs. female = 0.7
(1.1-2.7)
Full model (95% CI)
(included smoking
and copper smelting):
0-10 ppb = 1
(referent)
10-29 ppb = 0.3 (0.1-
1.2)
30-59 ppb = 1.8(0.5-
6.9)
60-89 ppb = 4.1 (1.8-
9.6)
90-199 ppb = 2.7
(1.0-7.1)
200-399 ppb = 4.7
(2.0-11.0)
400-699 ppb = 5.7
(1.9-16.9)
700-999 ppb = 7.1
(3.4-14.8)
Male vs. female =1.1
(0.6-1.8)
Ever vs. never
smoked =
4.3 (2.6-7.3)
SES medium vs. low
= 1.3 (0.7-2.5)
SES high vs. low =
2.3(0.5-12.1)
Copper smelting
(ever/never) =1.7
(0.7-4.4)
Strengths/
Weaknesses
Strengths:
-Odds ratios adjusted for
age, gender, cumulative
lifetime cigarette smoking,
working in copper
smelting, and
socioeconomic status.
-Because the control
group selection was
complex, several validity
checks were completed.

Weaknesses:
-Relatively more controls
were chosen from the
highly exposed city of
Antofagasta than from the
lower exposure cities of
Arica and Iquique
resulting in possible
underestimation of risk.
































Reference/
Type of
Study
Ferreccio et
al., 2000
Case-
control

















































B-26   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
1996-
2000





















1989-
2000
















Subjects/
Controls
114
bladder
cancer
cases

114
individuals
without
bladder
cancer













-200,000
residents
















Exposure
Assessment
Average arsenic
concentration (ppb)
of 5 years of
highest exposure
during the period
of 6-40 years prior
to interview:
0-50
51-100
101-200
>200
(mean: 164 ppb)











Water arsenic
levels:
prior to 1958, -90
ppb; in the late
1950s, water
supplementation
from a nearby river
where arsenic
levels approached
1000 ppb was
added to the
existing city water
supply






Study Outcome
Bladder cancer
Odds ratio (95%
CI) — ever smokers by
time before interview:
5 1-60 years earlier =
2.65 (1.2-5.8)
61-70 years earlier =
2.54 (1.0-6.4)
periods combined =
2.5(1.1-5.5)













SMRs (95% CI):
1950-1957 birth
cohort (early
childhood exposure):
lung cancer = 7.0
(5.4-8.9, p< 0.001)

High exposure period
(1958-1971) with
probable exposure in
utero and early
childhood:
lung cancer = 6.1
(3.5-9.9, p< 0.001)




Strengths/
Weaknesses
Strength:
-Potential confounders
controlled included age,
gender, smoking, and
county of residence.

Weaknesses:
-Lack of a cancer registry,
arsenic exposure
misclassification (use of
current water source
arsenic measurements
possibly causing
underestimation of
exposure), and recall bias.
-Possible selection bias
since controls had a
significantly reduced rate
of participation than cases
and cases were selected
from the tumor registry.
-Other harmful exposures
not measured.
Strengths:
-Extensive documentation
of arsenic in drinking
water in the Antofagasta
water system.

Weaknesses:
-Residence was
determined from death
certificates and relates to
residence at the time at
death.
-Reliance on death
certificates resulting in
potential diagnostic bias.
-Information bias
(smoking history).
Reference/
Type of
Study
Bates et al.,
2004
Case-
control



















Smith etal.,
2006
Cohort














B-27   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1950-
2000




























Subjects/
Controls
Region II
residents

Region V
residents as
comparison
group

Population
of Chile




















Exposure
Assessment
Average arsenic
concentration
(ppb):
Region II
1950-1954 = 123
1955-1959 = 569
1960-1964 = 568
1965-1969 = 568
1970-1974 = 272
1975-1979 = 176
1980-1984 = 94
1985-1989 = 71
1990-1994 = 43
Region V
unexposed















Study Outcome
Peak rate ratios (95%
CI) compared to
Region V and Chile:
Lung Cancer
1992-1994
Men
3.61 (3.13-4.16)
(Region V)
4.20 (3.76-4.70)
(Chile)

1989-1991
Women
3.26 (2.50-4.23)
(Region V)
3.41 (2.76-4.22)
(Chile)
Bladder Cancer
1986-1988
Men
6.10(3.97-9.39)
(Region V)
5.99(4.41-8.14)
(Chile)
1992-1994
Women
13.8 (7.74-24.5)
(Region V)
9.32 (6.67-13.0)
(Chile)
Strengths/
Weaknesses
Strengths:
-Large population size.
-Accurate past exposure
data.
-Known exposure pattern.
-Controlled for potential
confounding by age,
gender, and smoking.

Weaknesses:
-Could not account for
migration.
-No individual exposure
data or data on other risk
factors (smoking and
occupation).














Reference/
Type of
Study
Marshall et
al., 2007
Ecological



























B-28   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1950-
2000



























Subjects/
Controls
314,807
exposed

1,230,498
unexposed
























Exposure
Assessment
Average water
concentration (ppb)
in Region II:
Before arsenic
removal plant —
1950-1957 = 90
1958-1970 = 870
After arsenic
removal plant —
1971-1985=110
1986-2000 = 40
Present = 10

















Study Outcome
Excess deaths as
percentage of total
deaths (%) due to
acute myocardial
infarction, lung
cancer, and bladder
cancer combined:
Males —
1950-1957=1.00
1958-1964 = 4.19
1965-1970 = 6.03
1971-1979* = 6.48
1980-1985 = 8.94
1986-1990 = 10.07
1991-1995 = 10.87
1996-2000 = 7.92
Total = 6.93
Females —
1950-1957 = 0.48
1958-1964= 1.59
1965-1970 = 3.11
1971-1979* = 3.78
1980-1985 = 2.75
1986-1990 = 3.85
1991-1995 = 4.00
1996-2000 = 3.36
Total = 2.94
*No data available for
1976
Strengths/
Weaknesses
Strengths:
-Almost all drinking water
came from a few
municipal water sources,
which had known arsenic
concentrations.
-The study involved a
large population that
experienced a rapid
increase in arsenic
exposure followed by a
rapid decrease in arsenic
exposure.
-To ensure that an
appropriate comparison
population was chosen,
preliminary investigations
were conducted to
compare income,
smoking, and quality of
death certificate
information.

Weaknesses:
-Possible biases resulting
from a lack of individual
exposure data and
confounders.

Reference/
Type of
Study
Yuan et al.,
2007
Ecological


























B-29   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-4. North America cancer studies
Study
Period
3 9 years
(endpoint-
1978
diagnosis)






























Subjects/
Controls
71
National
Bladder
Cancer
Study
participants

160
National
Bladder
Cancer
Study
participants
without
bladder
cancer


















Exposure
Assessment
Mean arsenic
level (ppb) = 5.0
(range = 0.5-160 )

Exposure indices:

Index 1 —
cumulative dose
(<19, 19to<33,
33 to <53, >53
mg)

Index 2 — intake
concentration
adjusted to fluid
intake (<33, 33 to
<53, 53 to <74,
>74 mg- years)
















Study Outcome
Odds ratio for
bladder cancer and
arsenic exposure: no
association of
bladder cancer with
Index 1 or Index 2.
Among smokers,
positive trend in 10
year intervals.

























Strengths/
Weaknesses
Strengths:
-Age, gender, smoking
status, years of
chlorinated surface water
exposure, history of
bladder infection,
education, occupation,
population size of
geographic area, and
urbanization were
addressed.
-Cases were
histologically confirmed.

Weaknesses:
-Small size of study
population.
-Absence of historical
monitoring data and data
on arsenic levels in
public water supplies
were collected in 1978-
1979.
-The subjects were
mostly males and the
data on females were
inadequate.
-Arsenic exposure levels
were based on
measurements close to
the time that cases were
diagnosed.
-Arsenic from food was
not considered.
Reference/
Type of
Study
Bates et al.,
1995
Case-
control






























                                    B-30   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
1996






















1993-
1996






















Subjects/
Controls
2,203
deceased
individuals
from
Millard
County

General
Utah
population
used as
comparison











587 BCC
cases
284 SCC
cases
524
controls


















Exposure
Assessment
Arsenic exposure
index (ppb-years):
low = <1000
medium = 1000-
4999
high = >5000

















Toenail arsenic
level (ug/g):
BCC cases =
0.01-2.03
SCC cases =
0.01-2.57
controls = 0.01-
0.81

















Study Outcome
Cancer SMRs (95%
CI):
kidney —
males = 1.75(0.80-
3.32)
females =1.60
(0.44-4.11)
bladder and other
urinary organs —
males = 0.42 (0.08-
1.22)
females = 0.81
(0.10-2.93)
melanoma of the
skin —
females = 1.82
(0.50-4.66)
prostate = 1.45*
(1.07-1.91)

*p<0.05


OR (95% CI),
toenail arsenic
concentrations above
the 97th percentile:
SCC = 2.07 (0.92-
4.66)
BCC = 1.44 (0.74-
2.81)
















Strengths/
Weaknesses
Strengths:
-A major strength of the
study is that it measured
the effects of chronic
arsenic exposure in U.S.
population.
-Advantages of cohort
design include the fact
that the exposure
precedes the effect being
measured and that the
cohort design has the
ability to measure a
variety of effects from a
single type of exposure.

Weaknesses:
-Exposure assessment.
-Study power.
-Exposure to atmospheric
arsenic and arsenic from
food were potential
confounder.
Strengths:
-Evaluated the effects of
age, gender, race,
educational attainment,
smoking status, skin
reaction to first exposure
to the sun, history of
radiotherapy (potential
confounders).
-Toenail concentrations
individualize exposure
and account for arsenic
from other sources.
Weaknesses:
-Latency of arsenic-
induced skin cancer
unknown, follow-up
period may have been
inadequate.
-Toenail arsenic
measurements only
account for recent past
exposure.
Reference/
Type of
Study
Lewis etal.,
1999
Cohort




















Karagas et
al., 2001
Case-
control



















B-31   DRAFT—DO NOT CITE OR QUOTE

-------

Study
Period
1979-
1999





















1994-
2000






























Subjects/
Controls
Not
applicable





















181 cases
328
controls





























Exposure
Assessment
Arsenic exposure
categories (ppb):
low = <10
medium = 10-25
high = 35-90


















Exposure
categories (ppb):
0-19
20-79
80-120
>120

Arsenic exposure
indices:
(1) highest
average daily
arsenic intake for
any one year, (2)
highest average
daily arsenic
intake averaged
over any
contiguous 5
years, (3) highest
average daily
arsenic intake
averaged over any
contiguous 20
years, and (4) total
lifetime
cumulative
exposure






Study Outcome
SIR (95% CI),
childhood leukemia
and all childhood
cancers excluding
leukemia:
Low-exposure
group-
leukemia = 1.02
(0.90-1.15)
all cancers = 0.99
(0.92-1.07)
Medium-exposure
group:
leukemia = 0.61
(0.12-1.79)
all cancers = 0.82
(0.47-1.33)
High-exposure
group:
leukemia = 0.86
(0.37-1.70)
all cancers = 1.37
(0.96-1.91)
Bladder cancer OR
(95% CI):

>80 ug/day = 0.94
(0.56-1.57)
linear trend, p = 0.48

>80 jig/day, >40
years ago — smokers
= 3.67 (1.43-9.42)
linear trend, p <0.01





















Strengths/
Weaknesses
Strengths:
-The analysis was
stratified by age.
-Low arsenic exposure
study.
-Findings were reported
for different
concentration ranges.

Weaknesses:
-Small study size.
-Limitations of
ecological study design.
-Arsenic from food was
not measured, leading to
possible exposure
misclassification.






Strengths:
-Potential confounders
adjusted included gender,
age, smoking history,
education, occupation
associated with elevated
rates of bladder cancer,
and income.
-Use of cancer registry.
-Individual exposure
levels.

Weaknesses:
-Information bias (next-
of-kin interviews).
-Arsenic exposures
outside the study area
were not incorporated.
-In the arsenic-exposed
areas, the percentage of
nonparticipants was 5%
higher among cases than
controls. This difference
would probably mean
that more exposed cases
were missed in analyses
of recent exposure,
biasing the odds ratio
toward the null.
-Arsenic exposure from
food was not considered.
Reference/
Type of
Study
Moore et
al, 2002
Ecological




















Steinmaus
et al., 2003
Case-
control



























B-32   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1999-
2000
































Subjects/
Controls
368
cutaneous
melanoma
cases

373
colorectal
cancer
controls

























Exposure
Assessment
Median toenail
arsenic
concentration:
cases = 0.06 ug/g,
controls = 0.04
ug/g




























Study Outcome
OR = 2.1(95%CI =
1.4-3.3,
p-trend = 0.001) for
increased risk of
melanoma with
elevated toenail
arsenic
concentrations

OR = 6.6 (CI = 2.0-
2 1.9) for increased
risk of melanoma
with previous
diagnosis of skin
cancer and elevated
toenail arsenic
concentrations

















Strengths/
Weaknesses
Strengths:
-Potential confounders
controlled for were age,
gender, skin color/skin
type, prior history of
sunburn, education, and
occupational exposure(s).
-Ascertainment of cases
and controls was
accomplished by using
the Iowa Cancer
Registry, a Surveillance,
Epidemiology, and End
Results Program registry.
This allowed newly
diagnosed melanoma
cases to be identified for
a specific period and
ensured a greater degree
of certainty regarding the
accuracy of diagnosis.
-Toenail arsenic
measurements
individualize exposure
and account for arsenic
exposure from other
sources.
Weaknesses:
-A limitation was that
toenail samples were
collected 2-3 years after
diagnosis, resulting in
possible exposure
misclassification.
Reference/
Type of
Study
Beane-
Freeman et
al., 2004
Case-
control





























B-33   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
July 1,
1994 and
June 30,
1998































Subjects/
Controls
383
transitional
cell
bladder
cancer
cases

641
controls


























Exposure
Assessment
Toenail arsenic
level (ug/g):
cases = 0.014-
2.484
controls = 0.009-
1.077





























Study Outcome
Odds ratio (95%
CI)-
bladder cancer
among smokers:
>0.330ug/g = 2.17
(0.92-5.11)





























Strengths/
Weaknesses
Strengths:
-Evaluated the following
potential confounders:
age, gender, race,
educational attainment,
smoking status, family
history of bladder cancer,
study period and average
number of glasses of tap
water consumed per day.
-Conducted stratified
analyses according to
how long subjects used
their current water
system (<15 years, >15
years) to evaluate the
possibility that an
extended latency period
is required for bladder
cancer development.
-Attempted to minimize
misclassification by
using biomarker
(toenails).
Weaknesses:
-Possible
misclassification at lower
end of exposure range.
-Limited data at extreme
ends of exposure.
-Lifetime exposure could
not be calculated since
data from previous
residences could not be
determined.
Reference/
Type of
Study
Karagas et
al., 2004
Case-
control































B-34   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1950-
1979





















Subjects/
Controls
2,498,185
white
males

1970 U.S.
standard
population
















Exposure
Assessment
Median water
arsenic
concentration
(ppb):
3.0-3.9
4.0-4.9
5.0-7.4
7.5-9.9
10.0-19.9
20.0-49.9
50.0-59.9












Study Outcome
Bladder cancer
SMRs (95% CI),
white males by
median arsenic
concentration in
ground water (ppb):
3.0-3.9 = 0.95
(0.89-1.01)
4.0-4.9 = 0.95 (
0.88-1.02)
5.0-7.4 = 0.97
(0.85-1.12)
7.5-9.9 = 0.89
(0.75-1.06)
10.0-19.9 = 0.90
(0.78-1.04)
20.0-49.9 = 0.80
(0.54-1.17)
50.0-59.9 = 0.73 (
0.41-1.27)
All levels combined
=
0.94 (0.90-0.98)
Strengths/
Weaknesses
Strengths:
-Large study population.
-Study was nationwide.
-Included over 75 million
person-years of
observation.

Weaknesses:
-No individual exposure
data.
-Assumed that study
participants consumed
local drinking water.
-Available data assumed
to represent actual
arsenic content of water.
-Analysis did not directly
adjust for smoking,
urbanization, and
industrialization.
-Arsenic contribution
from food was not
measured.
Reference/
Type of
Study
Lammet
al., 2004
Ecological




















B-3 5   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
July
2000-
January
2002

































Subjects/
Controls
6,669
residents



































Exposure
Assessment
Three arsenic
exposure
categories (ppb):
<1.0
1.0-9.0
>10































Study Outcome
Skin cancer
adjusted odds ratio
(95% CI):
Arsenic level
(ppb)—
<1.0 = referent
1-9.9=1.81(1.10-
3.41)
> 10 =1.92 (1.10-
3.68)
Age (years) —
35-64 = referent
> 65 = 4.53 (2.79-
7.38)
Gender —
female = referent
males = 2.25 (1.33-
3.79)
Cigarette use —
no = referent
yes= 1.37(0.84-
2.24)















Strengths/
Weaknesses
Strengths:
-Large sample size.
-History of individual
tobacco use.
-Arsenic well water
analysis for each
household.
-Participants consumed
water from the tested
wells for at least 10
years.
-Analysis controlled for
age, gender, and tobacco
use.

Weaknesses:
-Skin cancers were serf-
reported and not
confirmed by a medical
records review.
-Few people could
provide information
about specific types of
cancer.
-Families that
participated may have
been especially
concerned about arsenic
exposure or family
members may have had
existing health
conditions.
-Not controlled for sun
exposure or occupation.
-Arsenic contribution
from food was not
measured.
Reference/
Type of
Study
Knobeloch
et al., 2006
Cross-
sectional

































B-36   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1979-
1997


































Subjects/
Controls
Residents
of six
Michigan
counties

Remainder
of
Michigan
population
as
comparison

























Exposure
Assessment
Population-
weighted mean
arsenic
concentration
(ppb):
exposed counties
= 11.00
remainder of
Michigan = 2.98



























Study Outcome
Elevated cancer
SMRs (95% CI):
Males —
liver/biliary = 0.85
(0.72-1.00)
trachea, bronchus,
lung= 1.02
(0.98-1.06)
melanoma = 0.99
(0.79-1.22)
other skin cancer =
1.24 (0.86-1.72)
bladder =0.94
(0.82-1.08)
kidney /urinary =
1.06 (0.91-1.22)

Females —
liver/biliary =1.04
(0.89-1.20)
trachea, bronchus,
lung =
1.02 (0.96-1.07)
melanoma = 0.97
(0.73-1.27)
other skin cancer =
1.06 (0.60-1.72)
female reproductive
organs = 1.11*
(1.03-1.19)
bladder =0.98
(0.80-1.19)
kidney /urinary
organs =
1.00 (0.82-1.20)
*p<0.01
Strengths/
Weaknesses
Strengths:
-Mortality data gathered
from Michigan Resident
Death Files for 20-year
period.
-Mortality rates stratified
by gender, age, and race.

Weaknesses:
-Possible differences in
reporting and
classification of
underlying causes of
death.
-No assessment of
individual exposures and
case migration.
-Smoking and obesity,
possible confounders,
were not included in the
analysis.
-Preferential sampling
based on home owners'
request.
-Arsenic contribution
from food was not
measured.









Reference/
Type of
Study
Meliker et
al., 2007
Ecological

































B-37   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-5. China cancer studies
Study
Period
1990





















Subjects/
Controls
3,179
residents




















Exposure
Assessment
HAC (ppb):
<10
10-
30-
50-
60-
100-
150-
500+
CAE (ppb-year):
<10
10-
32-
100-
316-
1000-
3162-
10000+




Study Outcome
Grade and (age-
adjusted) skin cancer
prevalence rates by
HAC:
<10 = 0.0 (0.0)
10- = 0.0 (0.0)
50- = 0.0 (0.0)
150-= 1.2(1.0)
500+ = 7. 1(5.9)

Grade and (age-
adjusted) skin cancer
rates by CAE:
<10 = 0.0 (0.0)
10- = 0.0 (0.0)
32- = 0.0 (0.0)
100- = 0.0 (0.0)
3 16- = 0.0 (0.0)
1000- = 0.4 (0.3)
3162- =0.8(0.2)
10000+ = 2.7 (2.0)

Strengths/
Weaknesses
Strengths:
-Large study population.
-Used both HAC and CAE
in the analyses.
-Arsenic concentrations
measured in 184 wells.
-Controlled for age and
differences in cumulative
arsenic exposure dose and
duration of exposure.

Weaknesses:
-Possible recall and
misclassification bias
resulting from the
collection of exposure
histories through
interviews.
-Inherent limitations of
ecological study design.
-Did not control for sun
exposure.
Reference/
Type of
Study
Lammet
al., 2007
Ecological



















                                    B-3 8  DRAFT—DO NOT CITE OR QUOTE

-------
Table B-6.  Finland cancer studies
Study
Period
1981-
1995





































Subjects/
Controls
61
bladder
cancer
cases and
49
kidney
cancer
cases

275
referents




























Exposure
Assessment
Water arsenic
concentration
(ppb):
<0.1
0.1-0.5
>0.5

Arsenic daily dose
(ug/day):
0.2
0.2-1.0
>1.0

Cumulative dose
(V&):
<500
500-2000
>2000





















Study Outcome
Bladder cancer risk
ratios (95% CI):
Shorter latency —
Water arsenic
concentration (ppb):
0.1-0.5= 1.53(0.75-
3.09)
>0.5 = 2.44 (1.11-
5.37)
Daily arsenic dose
(ug/day):
0.2-1.0= 1.34(0.66-
2.69)
>1.0= 1.84(0.84-
4.03)
Cumulative dose (ug):
500-2000=1.61
(0.74-3.54)
>2000= 1.50(0.71-
3.15)

Longer latency —
Water arsenic
concentration (ppb):
0.1-0.5 = 0.81(0.41-
1.63)
>0.5= 1.51(0.67-
3.38)
Daily arsenic dose
(ug/day):
0.2-1.0 = 0.76(0.38-
1.52)
>1.0= 1.07 (0.48-
2.38)
Cumulative dose (ug):
500-2000 = 0.81
(0.39-1.69)
>2000 = 0.53 (0.25-
1.10)
Strengths/
Weaknesses
Strengths:
-Cases were identified
through the Finnish Cancer
Registry.
-The 1985 Population
Census file of Statistics
Finland was used to
identify areas in which less
than 10% of the population
used the municipal water
supply.
-Risk ratios adjusted for
age, gender, and smoking.

Weaknesses:
-Possible misclassification
and possible recall bias
resulting from the study
choosing to use water
consumption from the
1970.
-Lacks other sources of
arsenic exposure.
















Reference/
Type of
Study
Kurttio etal.,
1999
Case-cohort




































                                    B-39   DRAFT—DO NOT CITE OR QUOTE

-------
Study
Period
1985-
1988
and
April
1999



























Subjects/
Controls
280
incident
bladder
cancer
cases

293
controls
























Exposure
Assessment
Arsenic exposure
quartiles (ug/g)—
1:0.050
2:0.050-0.105
3:0.106-0.161
4:>0.161


























Study Outcome
Bladder cancer odds
ratio (95% CI):
highest vs. lowest
quartile of toenail
arsenic = 1.13, (0.70,
1.81)
p trend = 0.65 for the
highest vs. lowest
quartile)























Strengths/
Weaknesses
Strengths:
-Study used toenail arsenic
as biomarkers of exposure.
-Cases and controls
matched according to age,
toenail collection date,
intervention group (alpha
tocopherol and beta
carotene), and smoking
duration.
-Study adjusted for
matching factors, smoking,
educational level, beverage
intake, and place of
residence.
-Cut point of >0.09 ug/g
used to avoid sample
misclassification.
-Potential confounders,
including smoking
cessation, smoking
inhalation, educational
level, beverage intake, and
place of residence, were
controlled for in the study
analysis.
Weaknesses:
-Water intake was not
included in the total
beverage variable.
-Toenail arsenic measures
recent past exposures.
Reference/
Type of
Study
Michaud et
al., 2004
Cohort/nested
case-control




























B-40   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-7. Denmark cancer studies
Study
Period
1970-
2003



































Subjects/
Controls
39,378
Copenhagen
residents

17,000
Aarhus
residents






























Exposure
Assessment
TWA arsenic
exposure (ppb)
from 41 years old
to date of
enrollment:
Copenhagen:
min = 0.05
max= 15.8

Aarhus:
min = 0.09
max = 25.3

Entire cohort:
min = 0.05
max = 25.3





















Study Outcome
Cancer incidence
rate ratios
(95% CI):
Time-weighted
average exposure:
Copenhagen —
melanoma = 0.73
(0.46-1.14)
non-melanoma =
1.09 (0.95-1.24)
breast = 1.04(0.88-
1.22)

Aarhus —
melanoma = 0.85
(0.61-1.20)
non-melanoma =
0.97 (0.90-1.05)
breast =1.06 (1.01-
1.11)

Cumulative
exposure:
Copenhagen —
melanoma = 0.94
(0.81-1.08)
non-melanoma =
1.01 (0.97-1.06)
breast =1.0 1(0.95-
1.06)
Aarhus —
melanoma = 0.97
(0.90-1.05)
non-melanoma =
0.98 (0.95-1.01)
breast =1.01 (0.99-
1.03)
Strengths/
Weaknesses
Strengths:
-Large study population.
-
Socioeconomic/demographic
similarities of the cohorts.
-Potential confounders
adjusted were smoking,
alcohol consumption,
education, body mass index,
daily intake of
fruits/vegetables, red meat,
fat and dietary fiber, skin
reaction to the sun, hormone
replacement therapy use,
reproduction, occupation,
and enrollment area.

Weaknesses:
-Possible misclassification
bias.
-Overall low arsenic
concentration in drinking
water in Denmark.
-Lack of data regarding
other sources of arsenic.












Reference/
Type of
Study
Baastrup et
al., 2008
Cohort


































                                   B-41   DRAFT—DO NOT CITE OR QUOTE

-------
Table B-8. Australia Cancer Studies
Study
Period
1982-
1991
















Subjects/Control
Victoria Cancer
Registry cancer
data

Australian Bureau
of Statistics
denominator data











Exposure
Assessment
Water/soil
exposure
groups:
High
water/high
soil —
>10ppb/>100
mg/kg
High water/low
soil —
>10ppb/<100
mg/kg
High soil/low
water —
<10ppb/>100
mg/kg


Study Outcome
Cancer SIRs (95%
CI):
Males and
females —
all cancers = 1.06
(1.03-1.09)
prostate =1.14
(1.05-1.23)
kidney =1.16
(0.98-1.37)
melanoma = 1.36
(1.24-1.48)
chronic myeloid
leukemia = 1.54
(1.13-2.10)
Females —
breast = 1.10
(1.03-1.18)
Strengths/
Weaknesses
Strengths:
-Study included both
water and soil in
exposure categories.
-Twenty-two areas
included in the study.

Weaknesses:
-Socioeconomic status,
race, occupation
and living in a rural area
were possible
confounders.
-Possible exposure
misclassification.
-Ecological study
limitations.

Reference/
Type of
Study
Hinwood et
al., 1999
Ecological















                                    B-42   DRAFT—DO NOT CITE OR QUOTE

-------
         APPENDIX C. TABLES FOR STUDIES ON POSSIBLE MODE OF ACTION FOR
                                     INORGANIC ARSENIC
 1          This appendix contains three tables that deal with possible MOAs of arsenic in the
 2    development of cancer based on in vivo human studies (Table C-l), in vivo experiments on
 3    laboratory animals (Table C-2), and in vitro studies (Table C-3). They describe numerous
 4    experiments published from 2005 through August 2007, as well as earlier experiments that were
 5    mentioned in the Science Advisory  Board Arsenic Review Panel comments of July 2007 (SAB,
 6    2007), 2001 NRC document on arsenic (NRC, 2001), or a detailed early draft of this document
 7    that lacked MOA tables. The data from these studies are distributed among 22 key-event
 8    categories, with the data from different experiments from a single publication often being
 9    summarized under different key-event categories.  For example, the results in Wang et al. (1996)
10    are summarized by rows under Apoptosis, Cytotoxicity, and Effects Related to Oxidative Stress
11    (ROS). The advantage of distributing the data in this way is that it helped to focus on a
12    particular key event for each set of data.  The disadvantage of using this approach is that it
13    spatially separated the different parts of each experiment.  An exception to this procedure is the
14    category Immune System Response, in which results from different parts of each experiment are
15    presented in successive rows.
16          A brief discussion of the approaches and conventions used in preparing the tables is
17    included here. Abbreviations are used liberally in an attempt to reduce the size of the table. An
18    attempt was made to provide a summary of the main findings of each experiment, with the
19    expectation that any reader wanting more detail would read the publication. A search for any
20    specific citation should make it easy to pull together the information from the numerous parts of
21    some studies that related to different categories. Although, for example, cytotoxicity data are
22    generally summarized in the  Cytotoxicity category, exceptions sometimes were made in an
23    attempt to decrease the size of the table. For example, if data presented on apoptosis contained
24    only slight, but interesting, data on cytotoxicity, a brief summary of those cytotoxicity findings
25    was sometimes added at the end of the results column in the row that described the results on
26    apoptosis. When an experiment that tested only one concentration yielded interesting results, the
27    results column is sometimes merged with one or more columns to its left in that same row so the
28    long description of results did not drastically increase the height of the table.  In such a case, the
29    only dose tested was obviously the LOEC or LOEL.
30          In vivo experiments on laboratory animals were almost always restricted to experiments
31    in which the route of exposure was  oral.  In most cases this meant that the arsenical was
32    administered in drinking water or was given by gavage.  A few experiments had the arsenical in
33    the feed.  Two experiments on chicken embryos had a solution (with concentration in uM)  put
34    onto the embryo, and one genetic assay done on Drosophila melanogaster had the concentration

                                                C-1    DRAFT—DO NOT  CITE OR QUOTE

-------
 1    (given in mM) reported for the media. All other in vivo experiments were done on mice or rats.
 2    Numerous studies were excluded on other non-mammalian species, including, for example, fish,
 3    nematodes, and algae.
 4          Tables C-2 and C-3 list all doses or concentrations tested as well as the duration of
 5    testing. It was often necessary to estimate the concentrations or doses tested from figures. For
 6    brevity, the control dose of 0 is not listed as a concentration tested.  In the rare instances in
 7    which there was no zero-dose control group, this omission is mentioned in the results section.  In
 8    many cases the papers themselves did not specify the LOECs or LOELs, and those values were
 9    estimated from tables or figures.  Because of the large variation in the way that papers presented
10    data and variability in their findings, and because of the rather common failure to clearly define
11    the error bars around data points in figures, there was often subjectivity involved in selecting the
12    LOEC or LOEL. There was no strict requirement that the LOEC or LOEL declared for each
13    experiment had to be shown to be statistically significantly higher than the control, although it
14    was not uncommon for that to be the case.  The wording in the results column often helps to
15    clarify this situation. If six concentrations were tested, for example, and if the second from the
16    lowest concentration had error bars that did not overlap those of the control, and if the third from
17    the lowest concentration was identified as being statistically significantly higher then the control,
18    then the second from the lowest concentration tested would have been declared the LOEC. The
19    LOEC, for example, should be viewed as the lowest concentration that was "quite likely" to have
20    caused an effect—without any specific statistical interpretation being attached to it. As long as
21    this was made clear, it was felt that this approach would be most useful to readers who want to
22    know the lowest concentration level at which a particular effect would  probably occur.
23          Arrows are used to indicate changes that were increases or decreases from the control.  If
24    the change was relative to some other group, it was clearly indicated as such.  In most cases, the
25    changes in magnitude of effects relative to the control were described as, for example, "2.34x" or
26    "0.46x"—2.34 times higher than the control or  only 46% as high as the control. When those
27    ratios were based on estimates made from a graph, they are generally preceded by a "~" mark; if
28    they were calculated from tabulated values,  they are generally presented without that  mark.
29          In Table  C-2 the doses are presented in terms of the amount of arsenic. When doses were
30    reported in mg arsenic/L or in ppm As, it was assumed that the doses included adjustment to
31    determine the amount of arsenic administered.  In a few publications it was unclear if the
32    reported doses were for the compound or for the amount of arsenic administered.  Partly because
33    of this uncertainty, all doses shown in the table that were  corrected to the amount of arsenic from
34    values that were clearly reported as concentrations of some arsenical compound (or for which
35    that was assumed to be the case) are preceded by an asterisk. Species of arsenic are shown in
36    Tables C-2 and C-3, and Asv is almost always sodium arsenate.
37

                                                C-2   DRAFT—DO NOT CITE OR QUOTE

-------
             Abbreviations for Tables in Appendix C
ft
1RB3AN27 cells

1T1 cells
293 cells

2-AAAF
2BS cells
3-NT
4HNE
4NQO
5-aza-dC
6-4 PPs
7-AAD
8-OHdG

8-oxoG
A2780 cells
A431 cells
A5/SG assays
A549 cells

AA
AB assay
ABTS
AC
ADM
ADSB
AFP
AG06 cells
AGT
Ahr+l+ MEFs
Aktl

ALAD
ALAS
increase
decrease
approximately (if before a listing of concentrations,
it applies to all)
approximately equal
an immortalized dopamine-producing rat
mesencephalic cell line
a human epithelial cell line
a cell line derived from adenovirus-transformed
human embryonic kidney epithelial cells
2-acetoxyacetylaminofluorene
human fetal lung fibroblasts
3-nitrotyrosine
4-hydroxy-2-nonenal
4-nitroquinoline 1-oxide
5-aza-deoxycytidine, a demethylating agent
6-4 photoproducts (UV-induced DNA photoproduct)
7-aminoactinomycin D
8-hydroxy-2'-deoxyguanosine or 8-
hydroxydeoxyguanosine (synonym)
7,8-dihydro-8-oxoguanine
human ovarian carcinoma cell line
human epidermoid carcinoma cell line
A5 (Annexin V-Alexa568) and SG (a green
fluorescent DNA dye) staining assays; A5+/SG- cells
are apoptotic
human non-small cell lung cancer (NSCLC) cell line
(alveolar basal epithelial cell line)
ascorbic acid (vitamin C)
AlamarBlue assay
2,2'-azinobis(3-ethylbenzothiazoline)-6-sulfonic acid
arsenic chloride
adriamycin
apparent DNA strand break
a-fetoprotein
SV40-transformed human keratinocytes
average generation time
mouse embryo fibroblasts of genotype Ahr+ + from
C57BL/6J mice, which are cells known to respond to
a B(a)P or TCCD challenge by activation of the AhR
V-akt murine thymoma viral oncogene homolog 1  (a
human gene)
5-aminolevulinic acid dehydratase
5-aminolevulinic acid synthetase
                              C-3    DRAFT—DO NOT CITE OR QUOTE

-------
AL hybrid cells
AMs
AML
AMPK
AO
APE/Ref-1
AP-PCR
Aprt
AP sites
AR230 cells

AR230-r cells

AR230-S cells

ARE
AS52 cells
   in
As
As:
Asv
ASK1
ATO
B0653

B16-F10 cells
BAEC
BALF
B[or]P
BCS
BEAS-2B cells
BER
BFTC905  cells
BFU
BHMT
BHT
Bid
BPDE
BrdU
BSO

BUG
a cell line that contains structural set of CHO-K1
chromosomes and one copy of human chromosome
11
alveolar macrophages
acute myelogenous leukemia
adenosine monophosphate-activated protein kinase
acridine orange
apurinic/apyrimidinic endonuclease (hAPEl)
arbitrarily primed polymerase chain reaction
adenosine phosphoribosyl transferase
sites of base loss (apurinic/apyrimidinic [AP] sites)
a CML cell line that expresses large amounts of Bcr-
Abl
AR230 cells that are resistant to the Bcr-Abl
inhibitor imatinib mesylate
AR230 cells that are sensitive to the Bcr-Abl
inhibitor imatinib mesylate
antioxidant response element
a pSV2 gpt-transformed Chinese hamster ovary cell
line; cells in this line carry a single copy  of a
transfected E. coli gpt gene
arsenic
arsenite
arsenate
apoptosis signal-regulating kinase 1
arsenic trioxide
2,3-dihydro-5-hydroxy-2,2-dipentyl-4,6-di-tert-
butylbenzofuran
mouse melanoma cells
bovine aortic endothelial cells
bronchoalveolar lavage fluid
benzo[a]pyrene
bathocuproinedisulphonic acid
human bronchial (pulmonary) epithelial cell line
base excision repair
a human urothelial carcinoma cell line
burst-forming units
betaine-homocysteine methyltransferase
butylated hydroxytoluene
a BH3 domain-containing proapoptotic Bcl2 family
member that is a specific proximal substrate of
CaspS in the Fas apoptotic signaling pathway
benzo[a]pyrene diol epoxide
bromodeoxyuridine
L-buthionine-S,R- sulphoximine (depletes GSH, y-
GCS inhibitor)
bladder urothelial cells
                              C-4   DRAFT—DO NOT CITE OR QUOTE

-------
C-33A cells

CAM
CAM assay
CAs
CAT
Cdc
Cdc42

cen+
cen-
CFE
c-Fos
CFSE
CFU
CGL-2 cells
cGpx
Chang cells

ChAT
CHO
CI
c-Jun or c-jun
CK8
CL3 cells

CL3R15 cells

c-met

c-Mos
CM-H2DCFDA
CML
Cone
Contraspin
COS-7 cells
CoTr
COX

COX-2
CPDs

Cpp32
CREEP
a transformed human non-differentiated carcinoma
cell line
cell adhesion molecule
chorioallantoic membrane assay of angiogenesis
chromosome aberrations
catalase (decomposes H^C^)
cell division cycle
a small GTPase in the Rho/Rac subfamily of Ras-
like GTPases
centromere positive (micronuclei)
centromere negative (micronuclei)
colony-forming efficiency
an AP-1 protein
5,6-carboxyfluorescein diacetate succinimidyl ester
colony-forming units
a cell line derived from a hybrid (ESH5) of the HeLa
variant, D98/AH2, and a normal human fibroblast
strain, GM77
cellular glutathione peroxidase
a human cell line thought to be derived from HeLa
cells
choline acetyltransferase
Chinese hamster ovary
confidence interval
an AP-1 protein
cytokeratin 8
human lung adenocarcinoma cells (established from
a non-small-cell lung carcinoma)
cell line derived from CL3 cells that were maintained
in 4 uM arsenic SA
the oncogene that encodes HGF (hepatocyte growth
factor) receptor
proto-oncogene
5-(and-6)-carboxy-2',7'-dichlorofluorescein diacetate
chronic myeloid leukemia
concentration
a serine—or cysteine—proteinase inhibitor isoform
African green monkey kidney fibroblast cell line
containing 10,000 glucocorticoid receptors per cell
that are transcriptionally inactive
co-treatment
cytochrome c oxidase; its activity is a measure of
mitochondrial function
cyclooxygenase-2
cyclobutane pyrimidine dimers (UV-induced DNA
photoproduct)
caspase-3
cAMP response element binding protein
                              C-5    DRAFT—DO NOT CITE OR QUOTE

-------
CRL1675 cells
CRL-1609 cells
cRNA

CSTP
Cul3
CV assay

CYP1A1
CYP7B1

DA
DAP
DCF assay
DCFH-DA
DCHA
DEB
DENA
DES
Dex
DHA
dhfr gene
DHR123
DIG
DI-I or II or
           m
DKO
dL
DMA
     in
DMA

DMAmI
DMBA
DMN
DMNQ
DMPO
DMPS
DMSA

DMSO
DNA
DNA-PK

D-NMMA

DNMT
a human melanocyte cell line
chimpanzee transformed skin fibroblast cells
RNA derived from complimentary DNA through
standard RNA synthesis
clonal survival treat and plate
Cullin 3, an Nrf2-fmding protein
crystal violet assay; it measures cellular protein,
which is related to cell number
cytochrome P450 1 Al
cytochrome P450 family 7, subfamily b polypeptide
1
disodium arsenate
2,6-diaminopurine
dichlorofluorescein assay
2',7'-dichlorofluorescein diacetate
docosahexaenoic acid, a co-3 polyunsaturated fatty
acid vital for the developing nervous system
diepoxybutane (DNA crosslinking agent)
diethylnitrosamine
diethylstilbestrol
dexamethasone (synthetic glucocorticoid)
dehydroascorbic acid
dihydrofolate reductase gene
dihydrorhodamine 123
dicumarol,  andNqol inhibitor
iodothyronine deiodinase-I or II or m (are 3 forms of
this selenoenzyme)
double knock out
deciliter
dimethylarsenous acid
dimethylarsinic acid
dimethyl arsenic (used when the oxidative state is
unknown or not specified)
dimethylarsinous iodide
dimethylbenzanthracene
dimethylnitrosamine
2,3-dimethoxy-l,4-naphthoquinone
5,5'-dimethyl-1-pyrroline TV-oxide (a spin-trap agent)
2,3-dimercaptopropane-l-sulfonic acid
dimercaptosuccinic acid or meso 2,3-
dimercaptosuccinic acid
dimethyl sulfoxide
deoxyribonucleic acid
DNA-dependent protein kinase, which has 3
subunits, of which the Ku70 protein is one
NG-methyl-D-arginine, the inactive enantiomer of a
nitric oxide synthase inhibitor
DNA methyltransferase
                               C-6   DRAFT—DO NOT CITE OR QUOTE

-------
DPC
DPI
DPIC

DR
DRE-CALUX

DSB
DTNB
DTT
DU145 cells
DW
E2N
E7 cells
EA
EB
E. coli
EDR3 cells

EGCG
EOF
EGFR
EGFRECD

EGR
elF
eIF4E
ELISA
Emodin
EMSA
Enm
eNOS
ER-a
ERCC1
ERCC2
Erk or ERK
EROD
ESR
ETU
FACS
FADD
FAK
FBS
DNA protein crosslinks
diphenyleneiodonium
diphenylene iodonium chloride, an NADPH-oxidase
inhibitor
death receptor
dioxin-responsive element (DRE)-mediated
Chemical Activated LUciferase expression
double strand break (in DNA)
5,5'-dithiobis(2-nitrobenzoicacid)
dithiothreitol
a human prostate carcinoma cell line
drinking water
ubiquitin-conjugating enzyme
an immortalized human bladder cell line
ethacrynic acid (a GST inhibitor)
ethidium bromide
Escherichia coli
a rat hepatoma cell line (glucocorticoid receptor
negative, with neither protein nor mRNA detectable)
(-)-epigallocatechin gallate
epidermal growth factor
epidermal growth factor receptor
extracellular domain of the epidermal growth factor
receptor
early growth response
eukaryotic initiation factor
eukaryotic translation initiation factor 4E, which is
the mRNA cap binding and rate-limiting factor
required for translation
enzyme-linked immunosorbent assay
(l,3,8-trihydroxy-6-methylanthraquinone)
electrophoretic mobility shift assays
endonuclease m
endothelial  nitric acid synthase
estrogen receptor-a
excision repair cross-complement 1 component
excision repair cross-complementing rodent repair
deficiency,  complementation group 2 (also known as
xeroderma pigmentosum group D or XPD)
extracellular signal-regulated kinase
ethoxyresorufin-O-deethylase
electron spin resonance
S-ethylisothiourea, a NOS inhibitor
fluorescence-activated cell sorting
Fas-associated death domain protein
focal adhesion kinase
fetal bovine serum
                               C-7   DRAFT—DO NOT CITE OR QUOTE

-------
FeTMPyP        5,10,15,20-tetrakis (jV-methyl-4'-pyridyl) porphinato
                 iron(in) chloride (ONOCT decomposition catalyst)
FGC4 cells       rat hepatoma cells
FGF-2           fibroblast growth factor -2
FGFR1           fibroblast growth factor receptor 1
FISH            fluorescent in situ hybridization
FITC            fluorescein isothiocyanate
FLIP            FLICE-inhibitory protein, an antiapoptotic protein
                 controlled by NF-KB
FLIPL            long-splice variant of FLIP
Fox O3a         an oxidative stress inducible forkhead transcription
                 factor
FPG             formamidopyrimidine-DNA glycosylase (digestion
                 ofDNA)
G12  cells         a pSV2gpt-transformed Chinese hamster V79 (hprf)
                 cell line
G6PDH          glucose-6-phosphate dehydrogenase
G-6-P            glucose-6-phosphatase; the paper that presented data
                 on this chemical called it G-6-PD in the discussion
GADD           growth arrest and DNA damage-inducible
GCLM           glutamate cysteine ligase modifier, GCLM knockout
                 mice (-/-) have only 9%-16% of GSH level of wt
                 littermates
GCR            glucocorticoid receptor
GFP             green fluorescent protein (GFP expressing tumor
                 cells)
GLN            glutamine
GlycoA          glycophorin A
GM043 1 2C       a S V-40 transformed XP A human fibroblast NER-
cells             deficient cell line
GM847  cells      a SV-40-transformed human lung fibroblast cell line
GM-CSF         granulocyte-macrophage  colony-stimulating factor
GM-Mp          GM-type macrophage
gpt              guanine phosphoribosyltransferase
GPx             glutathione peroxidase
GR              glutathione reductase
GRE            glucocorticoid response elements
GSH            glutathione
GSSG           glutathione disulfide
GST             glutathione-S-transferase
GTP             guanosine-5'-triphosphate
Gy              gray (unit of ionizing radiation)
HI 3 55 cells       a human lung adenocarcinoma cell line
H2C>2            hydrogen peroxide
H22  cells         a hepatocellular carcinoma cell line
H41 IE cells      a rat hepatoma cell line
H460 cells        a human non-small-cell lung cancer cell line (also
                 called human lung large cell carcinoma cells)
                               C-8   DRAFT— DO NOT CITE OR QUOTE

-------
H9c2 cells

HaCaT cells
Hb
HCC
HCT116 cells

HCT15 cells
HEC
HEK 293 cells
HEK293T cells
Hepa-lclc?
cells
HepG2 cells

HeLa cells
HeLa S3 cells
HELP cells
HEL cells
hEp cells
HFF cells
HFW cells
HGF
HGPRT
HIF
HK-2 cells
HL-60 cells
HLA
HLA-DR

HLF cells
HLFC cells

HLFK cells
HMEC-1 cells
HMOX-1
HO-
HOS cells
an immortalized myoblast cell line derived from fetal
rat hearts
a human epidermal keratinocyte cell line
hemoglobin
hepatocellular carcinoma
a human colorectal cancer cell line (available in
securin-wild-type and securin-null forms)
a human colon adenocarcinoma cell line
hamster embryo cells
an adenovirus-transformed human embryonic kidney
epithelial cell line (non-tumor cells), also called
HEK293 cells
human embryonic kidney cells
a mouse hepatoma cell line known to respond to a
B[a]P or TCCD challenge by activation of the AhR
a human hepatocellular liver carcinoma cell line
(Caucasian)
a human cervical adenocarcinoma cell line
a human cervical carcinoma cell line, derived from
the parent HeLa cell line; adapted to grow in
suspension (spinner) culture and has the same virus
susceptibility as the parent line
a human embryo lung fibroblast cell line
an AML cell line that is a cytokine-independent
human erythroleukemia cell line that has constitutive
STAT3 activity
normal human epidermal cells derived from foreskin
a human foreskin fibroblasts cell line
a diploid human fibroblast cell line
hepatocyte growth factor
hypoxanthine-guanine phosphoribosyltransferase
hypoxia inducible factor
a human proximal tubular cell line
human promyelocytic leukemia cells
human leukocyte antigen
human leukocyte antigen DR, which is a major
histocompatibility  complex class-II antigen
human embryo lung fibroblasts
an HLF subline that is not Ku70 deficient; it has the
null pEGFP-Cl vector transferred into it
an HLF subline that is Ku70 deficient; it has a
recombinant plasmid of Ku70 gene antisense RNA
transferred into it; it had 38% as much Ku70 protein
content as the HLFC cell line
human microvascular endothelial cells
heme oxygenase 1
hydroxyl radicals
a human osteogenic sarcoma cell line
                              C-9   DRAFT—DO NOT CITE OR QUOTE

-------
Hpall or HP All

HPBM
HPLC
HPRT
HRE

Hr
HSF1
HSP
HT1080 cells
hTER
hTERT
HT1197 cells

HU
Huh? cells
HuR
HUVEC cells

IAP
iAs
icAA

ICAM-1
ICE
IC50
ID1
IEC cells
IEC-6 cells
IGF
IGFBP-1
IKKP
IL
ILK
Imatinib
IM9 cells
IRE
IRP-1
J82 cells
JAK
JAR cells
JB6C141 cells

JB6C141PG13
cells
Haemophilusparainfluenzae (restriction
endonucleases)
human peripheral blood monocytes
high-performance liquid chromatography
hypoxanthine phosphoribosyl transferase
hypoxia response element, the DNA binding element
of HIF-mediated transactivation
hour(s)
heat shock transcription factor 1
heat shock protein
a human sarcoma cell line
RNA component of telomerase
human telomerase reverse transcriptase
a human (Caucasian) epithelial bladder cancer cell
line
hydroxyurea
a human hepatoma cell line
RNA binding protein
a human umbilical vein endothelial cell line (or
HUVECs)
inhibitor of apoptosis protein family
inorganic arsenic
intracellular ascorbic acid, which is accumulated at
up to high concentrations by culturing cells in DHA
inter-cellular adhesion molecule-1
interleukin-lp-converting enzyme
concentration that causes 50% inhibition of activity
inhibitor of DNA binding-1
a primary culture of rat intestinal epithelial cells
a rat intestinal epithelial cell line
insulin growth factor (system)
insulin-like growth factor binding protein  1
inhibitor of kappa light polypeptide gene enhancer in
B-cells, kinase beta; also called IkappaB kinase beta
subunit
interleukin
integrin-linked kinase
imatinib mesylate
a human multiple myeloma cell line
iron responsive element
iron regulatory protein 1
human bladder tumor cells
Janus kinase
a human placental choriocarcinoma cell line
a P+ mouse epidermal cell line (sometimes called
JB6 Cl 41 cells)
stable p53 luciferase reporter plasmid transfectant of
cell line JB6C141
                              C-10   DRAFT—DO NOT CITE OR QUOTE

-------
JB6C141P+1-1
cells
JC-1
JNK
Jurkat cells

K1735-SW1
cells
K562 cells
KCL22 cells
KCL22-r cells

KCL22-S cells

kDa
Keapl
KMS12BM
cells
Ku70

L-132 cells
LAK cells
LCL-EBV cells
LDH
LD50
LI
LOEC
LOEL
LOH
LPO
Luc
LU1205 cells
Z-NAME

L-NMMA

LPS
LTE4
Lys
Maf
stable activator protein-1 (AP-1) transfectant of cell
lineJB6C141
voltage-sensitive lipophilic cationic fluorescence
probe 5,5',6,6'-tetrachloro-l,l',3,3'-
tetraethylbenzimidazolcarbocyanine iodide
c-Jun N-terminal kinase
a transformed human T-lymphocyte cell line (also
called lymphoblast cells)
a mouse melanoma cell line

a human immortalized myelogenous leukemia cell
line that is a bcrabl positive erythroleukemia line
derived from a 53-year-old female CML patient in
blast crisis
a Bcr-Abl positive CML cell line
KCL22 cells that are resistant to the Bcr-Abl
inhibitor imatinib mesylate
KCL22 cells that are sensitive to the Bcr-Abl
inhibitor imatinib mesylate
kilodalton, a unit of mass
the cytoplasmic Nrf2-binding protein
a human multiple myeloma cell line

one of the three subunits of DNA-dependent protein
kinase
human alveolar type II cells
lymphokine activated killers (effector cells)
mononuclear cells obtained from healthy donors and
transformed by Epstein-Barr virus
50% lethal concentration
lactate dehydrogenase
50% lethal dose
labeling index
lowest observed effect concentration
lowest observed effect level
loss of heterozygosity
lipid peroxidation
the PEPCK-luciferase construct
a human melanoma cell line
7Va>-nitro-L-arginine methyl ester (an inhibitor of
NOS)
NG-methyl-L-arginine, the active enantiomer of a
nitric oxide synthase inhibitor
lipopolysaccharide
leukotriene, a proinflammatory mediator
lysine
musculoaponeurotic fibrosarcoma (transcription
factor)
                              C-11   DRAFT—DO NOT CITE OR QUOTE

-------
MAP
MAPK
MCA
MC/CAR cells
MCF-7 cells
MCR
M-CSF
MDA
MDAH 2774
cells
MDA-MB-231
cells
MDA-MB-435
mdm2
MDR
MED
MEF
MEF cells
MEK
MGC-803 cells
MI
MiADMSA
min
MK-571
MKP-1
MMA

MMA111
MMAinO
MMAV
MMC
MMP
MMP-2
MMP-9
MMP-13
MMS
MN
MNNG
MnTMPyP

MNU
MRC-5 cells
mRNA
MRP
mitogen-activated protein
mitogen-activated protein kinase
20-methylcholanthrene
a human multiple myeloma cell line
human breast carcinoma cell line
mineralocorticoid receptor
macrophage colony-stimulating factor
malondialdehyde (the thiobarbituric acid-reactive
substance in the brain that reflects extensive lipid
peroxidation)
human ovarian carcinoma cells

a human breast cancer cell line (an invasive estrogen
unresponsive cell line)
a human metastatic breast cancer cell line
murine double minute 2 proto-oncogene
multidrug resistance gene
minimal erythemic dose
mouse embryo fibroblasts
a mouse embryonic fibroblast cell line
MAP/ERK kinase (also, a family of related serine-
threonine protein kinases that regulate mitogen-
activated protein kinase)
a human gastric cancer cell line
mitotic index
monoisoamyl meso 2,3- dimercaptosuccinic acid
minutes(s)
MRP antagonist
MAP kinase phosphatase  1
monomethyl arsenic (used when oxidative state is
unknown or not specified)
monomethylarsonous acid
methylarsine oxide
monomethyl arsonic acid
mitomycin  C
mitochondrial membrane potential
matrix metalloproteinase-2
matrix metalloproteinase-9
matrix metalloproteinase-13
methyl methanesulfonate
micronuclei
1 -methyl-3 -nitro-1 -nitrosoguanidine
Mn(m)tetrakis(l-methyl-4-pyridyl)porphyrin
pentachloride (a cell permeable SOD mimic)
N-methyl-N-nitrosourea
a human lung fibroblast cell line
messenger ribonucleic acid
multidrug resistance-associated protein
                              C-12   DRAFT—DO NOT CITE OR QUOTE

-------
Mrps
MS
MT
mtDNA
MTOC
MTS assay
MTT

MIX
MT-1
MT2A
MW
MYH
MYP3 cells
N-18 cells
NAC
NADH
NADPH

Namalwa cells
NB4 cells
NB4-AsR

NB4-M-AsR2
cells

NCE
NCI cells
NE
NER
NF-KB
NHEK cells
NIH3T3 cells
NO'
NOS
Nqol
NR
Nrf2
efflux transporters encoded by MRP genes
mass spectrometer or mass spectrometry
metallothionein
mitochondrial DNA
microtubule-organizing center
3-(4,5-dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-
tetrazolium, inner salt assay; in Yi et al. (2004) study
this was referred to as the CellTiter 96 AQueous
Non-Radioactive Cell Proliferation Assay (MTS) Kit
(Promega, Madison, WI)
3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium
bromide
methotrexate
metallothionein-1
gene symbol for metallothionein 2A
molecular weight
MutY homolog, an endonuclease
rat epithelial cells line (urinary bladder cells)
a mouse neuroblastoma cell line
w-acetyl-cysteine (precursor of GSH; it elevates
cellular GSH levels, also an antioxidant), also N-
acetyl-Z-cysteine
reduced form of nicotinamide adenine dinucleotide
nicotinamide adenine dinucleotide phosphate-
oxidase
a human Burkitt' s lymphoma cell line
a human acute promyelocytic leukemia cell line
an arsenic-resistant subline of NB4 that was made by
culturing and maintaining cells in luM As2C>3
an arsenic-resistant human acute promyelocytic
leukemia cell line, which is routinely grown in RPMI
1640 media containing 2 uM As2C>3
normochromatic erythrocytes
a human myeloma cell line
nuclear extract
nucleotide excision repair (pathway)
nuclear factor-kappa B
primary normal human epidermal keratinocytes
a mouse fibroblast cell line
nitric oxide
nitric oxide synthase
nicotinamide adenine dinucleotide phosphate-
quinone oxidoreductase (or NAD(P)H-quinone
oxidoreductase)
neutral red
cap 'n' collar basic leucine zipper transcription
factor (nuclear factor erythroid 2-related factor 2)
                              C-13   DRAFT—DO NOT CITE OR QUOTE

-------
NSAID
NSE
NTUB1 cells
NuF
OATP-C
ODA
OGG1
OM431 cells
ONOO
OR
p21
PAEC cells
PAI-1
PARP
PBMC
PC
PC12 cells

PCE
PCI-1 cells

PCNA
PCR
PDH
PDT
PD98059

PEG
PEPCK

pEpREpgeo

PGE2
P-gP

PHA
PHEN
PI
PI3K
PK
FLAP
PLC/PR/5 cells
PMA
PMN
non-steroidal anti-inflammatory drug
no significant effect (often not based on a statistical
test but on whether an effect appears likely to be real
based on examination of graphs)
a human urothelial carcinoma cell line
nuclear fragmentation
organic anion transporting polypeptide-C
oxidative DNA adducts
8-oxoguanine DNA glycosylase
a human melanoma cell line
peroxynitrite
odds ratio
a cyclin-dependent kinase inhibitor
porcine aortic endothelial cells
plasminogen activator inhibitor-1
poly(adenosine diphosphate-ribose) polymerase
peripheral blood mononuclear cell (human)
protein carbonyl (form of protein oxidation)
a rat sympathetic (neuronal) pheochromocytoma cell
line
polychromatic erythrocyte
a human head and neck squamous cell carcinoma
cell line
proliferating cell nuclear antigen
polymerase chain reaction
pyruvate dehydrogenase
population doubling time
inhibitor of MEK1/2, which are ERK upstream
kinases (structurally unrelated to U0126)
monomethoxypolyethylene glycol (covalent
attachment of PEG to CAT or SOD extends their
plasma half-lives)
phosphoenolpyruvate carboxykinase gene (a
hormone-inducible gene)
p-galactosidase-neomycin-resistance reporter
plasmid
prostaglandin E2
P-glycoprotein, the efflux transporter encoded by
MDR
phytohemagglutinin
o-phenanthroline (an iron chelator)
propidium iodide
phosphatidylinositol 3-kinase
proteinase K
placental alkaline phosphatase
a human hepatocellular carcinoma cell line
phorbol 12-myristate 13-acetate
polymorphonuclear neutrophils (or PMNs)
                              C-14   DRAFT—DO NOT CITE OR QUOTE

-------
PMs
PNA
ppb
P-PKB
ppm
PQ
PR
PRCC
PSH
P-STAT3
pt
PTEN

p-XSC
R-3T3 cells

Rac
RACs
Raf
RAGE
RANKL
RAPD-PCR

Ras
RAW264.7
cells
RBC
RFU
RHMVE cells
RI
RKO cells

ROCK
RNA
RNS
ROS
RPMI-8226
cells
RT-PCR
RWPE-1 cells
SA
SACs
peritoneal macrophages
peptide nucleic acid
parts per billion
phosphorylated protein kinase B
parts per million
paraquat (a generator of 02")
progesterone receptor
primary renal cortical cell
protein thiol
phosphorylated-STAT3
pretreatment
phosphatase and tensin homolog (mutated in
multiple advanced cancers 1)
l,4-phenylenebis(methylene)selenocyanate
Ras-transformed NIH 3T3 cells, a mouse fibroblast
cell line
a subfamily of the Rho family of GTPases, which are
small (-21 kDa) signaling G proteins (more
specifically GTPases).
rapidly adhering cells; epidermal cells with the
highest proliferative potential and with properties of
stem cells
a proto-oncogene
receptor for advanced glycation end products
receptor activator of NFicB ligand
random(ly) amplified polymorphic DNA polymerase
chain reaction
a name of a proto-oncogene
a mouse macrophage cell line (another source
described it as mouse macrophage-like cells)
red blood cell, erythrocyte
relative fluorescence units (units of ROS)
rat heart microvessel endothelial cells
replicative index
a human colorectal  carcinoma cell line that expresses
wild-type p53 proteins
Rho/kinase, and effector molecule of RhoA
ribonucleic acid
reactive nitrogen species
reactive oxygen species
a human myeloma cell line

reverse transcription-polymerase chain reaction
human prostate epithelial cell line
sodium arsenite
slowly adhering cells; epidermal cell fraction that
contains cells undergoing terminal differentiation,
with little ability to form colonies
                              C-15   DRAFT—DO NOT CITE OR QUOTE

-------
SAH
SAM
SCC
SCE
SCGE
Se
SE
SEM
Se-Met
Ser
SF
SFN
SHE cells
SIK cells

siRNA
SLC30A1

SMART
SMC cells
socs
SOD
SP
SRB assay
Src

SSB
STAT
StRE site
SU5416
SVEC4-10 cells
SV-HUC-1 cells
SV-40
SW13 cells
SW480 cells
SY-5Y cells
T3
T4
T47D cells
TAM
S-adenosylhomocysteine
S-adenosylmethionine
squamous cell carcinoma
sister chromatid exchange
single cell gel electrophoresis (assay)
selenium
standard error of the mean
scanning electron microscopy
selenomethionine
serine, an amino acid
sodium formate, an 'OH radical scavenger
sulforaphanem, an activator of transcription factor
Nrf2, which plays a critical role in metabolism and
excretion of xenobiotics
Syrian hamster ovary  cells
spontaneously immortalized human keratinocytes (or
epidermal cells)
small interfering RNA (ribonucleic acid)
gene symbol for the zinc transporter, solute carrier
family 30, member 1
somatic mutation and recombination test
human bladder smooth muscle cells
suppressors of cytokine  signaling
superoxide dismutase (an antioxidant to (V")
shock protein
sulforhodamine B colorimetric assay
first oncogene discovered, the transforming protein
of the chicken retrovirus, Rous sarcoma virus
single strand break (in DNA)
signal transducer and  activator of transcription
stress response element recognition site
inhibitor of VEGF receptor-2 kinase
a C3H/HeN mouse vascular endothelium cell line
(also called immortalized mouse endothelial cell
line)
an SV40 large T-transformed human urothelial cell
line (non-tumor cells, derived from urethra,
immortalized)
simian virus 40
a human adrenal carcinoma cell line
a colorectal adenocarcinoma cell line derived from a
Caucasian male that has two base-pair substitution
mutations in the p53 gene
a human neuroblastoma cell line
thyroid hormone triiodothyronine
thyroid hormone thyroxine
a human mammary adenocarcinoma cell line
tamoxifen
                              C-16   DRAFT—DO NOT CITE OR QUOTE

-------
TAT
TEARS

tBHQ
TCDD
TF
Tg.AC
TGF
THP-1 +
A23187 cells

TIG-112 cells
TIMP-1
Tiron

TK6 cells
TM
TMAVO
TM3 cells

TNF-a
TPA
TR9-7 cells
TRAIL
TRAIL-R
TRAP
TRF
TRL 1215 cells
Trolox

Trx
TrxR
TrxRl
Trxl
Trx2
TUNEL assay
tyrosine aminotransferase
thiobarbituric acid reactive substances (a measure of
tissue lipid peroxidation)
^-butylhydroquinone
2,3,7,8-tetrachlorodibenzo-p-dioxin
theaflavin
strain of transgenic mice that contains the fetal beta-
globin promoter fused to the v-Ha-ras structural gene
(with mutations at codons 12 and 59) and linked to a
simian virus 40 polyadenylation/splice sequence
transforming growth factor
a human dendritic cell line; THP-1 cells acquire the
characteristics of dendritic cells in the presence of
the calcium ionophore A23187
human normal skin diploid cells
tissue inhibitor of metalloproteinase-1
4,5-dihydroxy-w-benzenedisulfonic acid, disodium
salt
human lymphoblastoid cells
tail moment
trimethylarsine oxide
immortalized Ley dig cells derived from normal
mouse testis
tumor necrosis factor a (an inflammatory cytokine)
12-O-tetradecanoylphorbol-13-acetate
a spontaneously immortalized human fibroblast cell
line, derived from a Li-Fraumeni patient, and
subsequently stably transfected with a tetracycline-
regulated p53 expression vector
TNF-related apoptosis-inducing ligand
TRAIL receptor
tartrate resistant acid phosphatase (RAW264.7 cells
can undergo osteoclast differentiation, which is
accompanied by an increase in the number of
multinucleate cells expressing TRAP)
terminal restriction fragment
nontumorigenic adhesive rat epithelial liver cells
originally derived from the liver of 10-day-old Fisher
F344 rats
6-hydroxy-2,5,7,8- tetramethylchroman-2-carboxylic
acid
thioredoxin
thioredoxin reductase
cytosolic thioredoxin reductase
cytoplasmic thioredoxin-1
mitochondrial thioredoxin-2
terminal deoxynucleotidyl transferase-mediated
deoxyuridine nick-end labeling assay
                              C-17   DRAFT—DO NOT CITE OR QUOTE

-------
U0126

U118MG cells

U266 cells
U937 cells


U-937 cells

U-2OS cells
Ub
UROtsa cells
UV
UVA
UVB
UVC
V79 cells

VEGF

VEGFR1

VEGFR2

V-FITC
VH16
vs.
VSMC
W138
wk
wt
WM9 cells
WRL-68
WT-1
XIAP

XPA (B or F)

XRS
XTT

YC-1
yptlocus
Z-DEVD-FMK

ZPP
inhibitor of MEK1/2, which are ERK upstream
kinases (structurally unrelated to PD98059)
a human glioblastoma cell line, also called Ul 18MG
(ATCCHTB-lS)cells
a human multiple myeloma cell line
a human leukemic monocyte lymphoma cell line
(also described as a human promonocytic cell line or
as a human myeloid leukemia cell line)
human diffuse histiocytic lymphoma cells, perhaps
the same as U937 cells
a human osteogenic sarcoma cell  line
ubiquitin
an SV40-immortalized human urothelium cell line
ultraviolet radiation
ultraviolet radiation A
ultraviolet radiation B
ultraviolet radiation C
a cell line derived from lung fibroblasts of a male
Chinese hamster
vascular endothelial growth factor or vascular
endothelial cell growth factor
a vascular endothelial cell growth factor receptor
(flt-1)
a vascular endothelial cell growth factor receptor
(Flk-1, KDR)
V-fluorescein isothiocyanate
human primary fibroblasts
versus
vascular smooth muscle cells
a human diploid lung fibroblast cell line
week(s)
wild-type
a human melanoma cell line
a human hepatic cell line
Wilm's tumor protein-1
X-linked inhibitor of apoptosis protein, an
antiapoptotic protein controlled by NF-KB
xeroderma pigmentosum, complementation group A
(B or F)
X-ray sensitive
2,3-bis[2-methyloxy-4-nitro-5-sulfophenyl]-2H-
tetrazolium-5-carboxanilide
a small molecule inhibitor of HIF signaling
xanthine-guanine phosphoribosyltransferase locus
benzyloxycarbonyl-L-Asp-Glu-Val-Asp-
fluoromethyl ketone, a  caspase 3 inhibitor
zinc protoporphyrin
                              C-18   DRAFT—DO NOT CITE OR QUOTE

-------
           Z-VAD-FMK

           a7-nAChR
           a-Toc
           yGCS
           yH2A.X

           p° cells
Z-Val-Ala-DL-Asp-fluoromethylketone, a general
caspase inhibitor
a7-nicotinic acetylcholine receptor
a-tocopherol, an antioxidant
y-glutamylcysteine synthetase
phosphorylated histone variant H2A.X that is
indicative of DNA double strand breaks
AL hybrid cells made highly deficient in
mitochondrial DNA by long-term treatment with
ditercalinium
Table C-l.  In vivo human studies related to possible modes of action of arsenic in the
           development of cancer
Topic(s)
Population
Sampled
Information on Exposure
Levels and Durations and
on Biomarkers
Results
Reference
Aberrant Gene or Protein Expression
Effect of
inorganic arsenic
exposure from
DWon
concentration of
RAGE protein in
sputum
Effect of
inorganic arsenic
exposure from
DW on serum
levels of
extracellular
domain of EGFR
(i.e., EGFR
BCD)
Effect of
inorganic arsenic
exposure from
DW on levels of
TGF-a in
bladder
urothelial cells
(BUC)
People in Ajo
(high dose)
and Tucson
(low dose),
Arizona,
USA
Araihazar
area of
Bangladesh
3 towns in
central
Mexico
Compared subjects from
Ajo (-20 ppb of arsenic in
DW) with subjects from
Tucson (~5 ppb of arsenic
inDW). They also
determined total inorganic
arsenic concentrations in
urine in individuals.
Estimates of inorganic
arsenic exposure level were
based on well water arsenic
(ranged from 0. 1 to 768
ppb), urinary arsenic, and
cumulative arsenic index.
Such estimates and EGFR
BCD protein levels were
compared in 574 people.
Estimates of inorganic
arsenic exposure level were
based on levels of different
metabolites of arsenic in
urine from 72 women who
used drinking water that
contained 2-378 ppb As.
No difference was seen in concentration of RAGE
protein in sputum between cities. Since there was
much overlap of total inorganic arsenic
concentrations in urine in individuals in those cities,
a comparison was also made using inorganic arsenic
levels in urine. The regression analysis yielded a
significant negative association between urinary
total inorganic arsenic concentrations and RAGE
concentrations in sputum. Thus inorganic arsenic
exposure caused U in RAGE level as was seen in
mice.
Found significant positive correlation between
EGFR BCD protein levels in serum and all of these
measures of inorganic arsenic exposure, with the
association being strongest among individuals with
As-induced skin lesions.
Found significant positive correlation between TGF-
a protein levels in exfoliated BUC and each of 6
arsenic species present in urine. Women from areas
with high arsenic exposures had significantly higher
TGF-a protein levels in BUC than those from areas
of low arsenic exposure. BUC cells from people
with As-induced skin lesions contained significantly
more TGF-a.
Lantz et al.,
2007
Li et al.,
2007
Valenzuela
etal.,2007
                                        C-19   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
Microarray-
based gene
expression study
comparing
groups with and
without arsenical
skin lesions,
both of which
were exposed to
inorganic arsenic
in DW but to
different extents
Comparison of
expression of
several genes
between patients
with As-related
urothelial cancer
and non- As-
related urothelial
cancer


Comparison of
expression of
several integrins
between people
with arsenic-
related keratosis
and people with
normal skin





Population
Sampled
Bangladesh











Taiwan,
patients with
urothelial
cancer







Taiwan,
patients with
arsenical
keratosis








Information on Exposure
Levels and Durations and
on Biomarkers
Compared subjects with
cutaneous signs of
arsenicism (mean of
343+258 ppb of arsenic in
DW) with asymptomatic
individuals (mean of 40+50
ppb of arsenic in D W in one
set, and 95+91 ppb in
another).



All 33 patients with arsenic-
related urothelial cancer had
been living in the arseniasis-
endemic area of southwest
Taiwan, where people had
drunk the As-contaminated
artesian well water for at
least 10 years. They were
compared with 25 patients
who had nonarsenic -related
urothelial cancer.
All 25 arsenical keratosis
patients were from
arseniasis-endemic areas of
southwest Taiwan, where
water is contaminated by
high concentrations of
inorganic arsenic. Control
specimens were obtained
from the non-sun-exposed
skin of 8 age-comparable
patients who did not live in
the endemic areas.


Results
Looked at expression of -22,000 transcripts in RNA
from peripheral blood lymphocytes. When the
comparison was restricted to female never-smokers,
3 12 differentially expressed genes were identified
between those with and without As-induced skin
lesions, with all of them being down-regulated in the
skin-lesion group. Signal transduction through the
IL-1 receptor was identified as a significant pathway
of differentially expressed genes between the
arsenical skin lesion (n = 11) and nonlesion (n = 2)
groups. It discriminated between the 2 groups.

Comparisons were made of protein expression of
GST-Ti, p53, Bcl-2, and c-Fos by Western blotting oi
tumor tissues. A significantly higher proportion of
the patients with the arsenic-induced cancer had the
proteins present for Bcl-2 (33/33 vs. 19/25) and for
c-Fos (30/33 vs 16/25), suggesting that up-
regulation of these 2 oncoproteins may play
important roles in arsenic -mediated urothelial
carcinogenesis. Cellular GSH content was down-
regulated in both types of tumors, but to a greater
extent in the arsenic -induced ones.
Immunohistochemical staining patterns of integrin
Pi, a2pi, and a3pi were observed. The various
patterns of staining among the patients in
comparison to the controls showed decreased
expression of all 3 integrins in both arsenical
keratosis and in perilesional skin. None showed the
normal expression pattern of all 3 integrins.
However, there was no association with the
occurrence of basal cell carcinoma or squamous cell
carcinoma and the expression pattern of any of the 3
integrins.



Reference



Argos etal.,
2006










Hour etal.,
2006









Lee et al.,
2006b







Apoptosis
Possible
association of
specific p53
polymorphisms
with arsenic-
related keratosis
in individuals
exposed to
arsenic in DW





West Bengal,
India












Compared 177 arsenic-
exposed subjects with
keratosis (mean of 177 ppb
of arsenic inDW) with 189
arsenic -exposed subjects
without such skin lesions
(mean of 180 ppb of arsenic
in DW), and looked for
association of keratosis with
3 specific p53
polymorphisms. Used
arsenic concentration
comparisons in DW, urine,
nails, and hair.
Homozygotes for alleles at 2 of the polymorphisms
were significantly over represented in the
individuals with keratosis. Results suggest that
individuals carrying the arginine homozygous
genotype at codon 72 and/or the no duplication
homozygous genotype at intron 3 are at higher risk
for the development of arsenic -induced keratosis. In
both cases the OR was 2.086 and the 95% CI did not
overlap 1. Urinary excretion of arsenic was slightly
lower (NSE) in the group with keratosis suggesting
higher retention of arsenic in the body, which was
reflected in significantly higher arsenic content in
nails and hair.




De
Chaudhuri
etal., 2006








C-20   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)

Population
Sampled
Information on Exposure
Levels and Durations and
on Biomarkers


Results


Reference
Chromosomal Aberrations and/or Genetic Instability
Nested case-
control study/
CAs and/or
SCEs as
biomarkers for
the prediction of
cancer
development

Induction of MN








Induction of MN
and CAs
(relationship to
presence of
arsenicism and
GST
polymorphisms)











Blackfoot-
endemic area
in Taiwan






West Bengal,
India







West Bengal,
India
















Looked at CAs and SCEs in
lymphocytes from venous
blood samples






Compared subjects with
cutaneous signs of
arsenicism (368 ppb of
arsenic in DW) with
asymptomatic individuals
(5.5 ppb of arsenic inDW).
Also used arsenic
concentration comparisons
in urine, nails, and hair.
Compared arsenic-exposed
subjects with cutaneous
signs of arsenicism (mean
of 242 ppb of arsenic in
DW), arsenic -exposed
subjects without cutaneous
signs of arsenicism (mean
of 202 ppb of arsenic in
DW), and arsenic-
unexposed subjects (mean
of 7.2 ppb of arsenic in
DW), and looked for
association of effects with
different GSTT1 and
GSTM1 genotypes. Used
arsenic concentration
comparisons in DW, urine,
nails, and hair.
Chromosome-type CAs, but not chromatid-type CAs
or SCEs, were significantly higher in the cases than
in the controls. The cancer risk OR for subjects
with >0 chromosome-type breaks was 5.0 (95% CI =
1 .09-22.82). The OR became even higher with
more refinements. Thus chromosome-type CAs (but
not chromatid-type CAs or SCEs) can serve as
useful biomarkers for prediction of cancer
development.
In the exposed group, the frequencies of MN per
1,000 cells were highly elevated over those of the
control group (# per 1000 cells): 5.15 vs 0.77 in the
oral mucosa, 5.74 vs 0.56 in urothelial cells, and
6.39 vs. 0.53 in peripheral lymphocytes,
respectively.



arsenic-exposed groups showed ft in MN in the
lymphocytes, oral mucosa, and urothelial cells and ft
in frequencies of CAs in lymphocytes. The
symptomatic (i.e., with cutaneous signs of
arsenicism) exposed group had more of all types of
cytogenetic damage than the asymptomatic exposed
group, and the asymptomatic exposed group had
more of all types of cytogenetic damage than the
unexposed group. Asymptomatic and symptomatic
exposed groups demonstrated rather similar
concentrations in the urine, nails, and hair.
Individuals carrying at least one GSTM1 -positive
allele had a significantly higher risk of developing
cutaneous signs of arsenicism.






Liou et al.,
1999







Basuetal.,
2002









Ghosh et al.,
2006












C-21  DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
Association
between a
polymorphism in
ERCC2 codon
751 that
probably
improves NER
and (1) the
incidence of
CAs and (2) the
presence of
inorganic
arsenic-induced
hyperkeratosis








Induction of MN
(bladder cells)








Induction of
SCEs
(Fowler's
solution,
(lymphocytes)




Induction of CAs
and SCEs
(Fowler's
solution,
lymphocytes)










Population
Sampled
West Bengal,
India




















Chile, men









6 patients
treated with
Fowler's
solution who
developed
arsenicism
and biopsy-
proven skin
cancers
8 psoriasis
patients
treated with
Fowler's
solution were
compared
with 8
psoriasis
patients not
treated with
inorganic
arsenic (7
men in each
group)
Information on Exposure
Levels and Durations and
on Biomarkers
Comparisons were made
between people with
hyperkeratosis and
individuals with no skin
lesions who were drinking
similar inorganic arsenic-
contaminated water.
Groups with and without
hyperkeratosis had means ol
195 and 185 ppb arsenic in
DW, respectively, with
large standard deviations.










Compared subjects having
high (average 600 ppb of
arsenic in DW) and low
(average 15 ppb of arsenic
in DW) exposures.





Nothing is known about
doses; duration of treatment
with inorganic arsenic
ranged from 4 months to 27
years, and in most cases
treatment ceased decades
before this cytogenetic
analysis.

The total doses of inorganic
arsenic were from 300 to
1 200 mg for the 7 with
known doses. Inorganic
arsenic treatments ceased
many years before this
study. Comparisons were
also made to 30 apparently
healthy untreated males.







Results
The polymorphism resulted from a base pair change
from A to C at codon 75 1 that resulted in an amino
acid substitution from lysine to glutamine. The A/A
(i.e., Lys/Lys) genotype was compared with the A/C
and C/C genotypes combined. In the study
population, the allele frequencies of A and C were
0.4 and 0.6, respectively. A/A individuals were
shown to be at significantly higher risk of having
hyperkeratosis and also to have a higher frequency
of CAs in their lymphocytes, as follows: A/A
individuals were over-represented among
individuals with inorganic arsenic-induced
hyperkeratosis (OR = 4.77, 95% CI = 2.75-8.23).
There was a higher percentage of cells with CAs in
A/A individuals than in (A/C and C/C) individuals:
43% more in those exposed to inorganic arsenic but
not having hyperkeratosis, 18% more in those
exposed to inorganic arsenic and having
hyperkeratosis, and 3 1% in both groups combined.
Also, CAs were significantly more frequent in
inorganic arsenic-exposed people with
hyperkeratosis.
Used a fluorescent version of exfoliated bladder cell
MN assay to identify presence or absence of whole
chromosomes within MN. Significant ft in
induction of MN by arsenic was found, and
chromosome breakage appeared to be its major
cause. 4th highest quintile of exposure groups gave
the highest response, but there was a significant ft in
each of quintiles 2-4. Highest (5th) quintile (729-
1894 ppb) returned to baseline MN level, perhaps
because of cytostasis or cytotoxicity.
Patients treated with Fowler's solution had mean of
14.0 SCE/mitotic cell, while 44 normal controls had
mean of 5.8 SCEs/mitotic cell. They saw no
difference in chromosome breakage between the
groups.




ft in frequency of chromosomal breaks (i.e.,
chromatid and chromosome aberrations together) in
psoriasis patients with inorganic arsenic treatment
and an even bigger ft in comparison to healthy
untreated males. Inorganic arsenic treatment had
NSE on SCE frequency.










Reference





Banerjee et
al., 2007


















Moore et
al., 1997b







Burgdorf et
al., 1977








Nordenson
etal., 1979









C-22   DRAFT—DO NOT CITE OR QUOTE

-------
Topic(s)
Induction of CAs
(mostly airborne
inorganic
arsenic,
lymphocytes)
Induction of CAs
and SCEs
(lymphocytes)
Induction of CAs
andMN
(lymphocytes for
CAs)
Induction of MN
Induction of MN
(chromosome
breakage
and/or
aneuploidy)
Population
Sampled
9 workers
exposed to
inorganic
arsenic at
smelter in
northern
Sweden
People in
Fallen
(exposed)
and Reno
(control),
Nevada, USA
People in
Santa Ana
(high dose)
and Nazareno
(control),
Mexico
People in
Nevada,
USA, with
either very
high or low
exposure to
inorganic
arsenic in
DW
People in
Nevada,
USA, with
either very
high or low
exposure to
inorganic
arsenic in
DW
Information on Exposure
Levels and Durations and
on Biomarkers
Little information was
presented except to say that
there was no obvious
relationship between
exposure and CA
frequencies.
The exposed sample of 104
used DW containing >50
ppb arsenic (mostly >100
ppb As) for at least 5 years
and the control sample of 86
used DW containing <50
ppb arsenic (and often much
less) for the same period.
The high-dose group used
DW containing a mean of
408 ppb As, and the control
(i.e., low dose) group used
DW containing a mean of
30 ppb As. They also
considered arsenic
concentrations in urine and
blood and concentrations of
arsenic metabolites.
The high-dose group of 18
used DW containing a mean
of 13 12 ppb As, and the
individually matched
control (i.e., low-dose)
group used DW containing
a mean of 16 ppb As. They
also considered the
concentration of inorganic
arsenic and methylated
metabolites in urine.
The high dose group of 18
used DW containing a mean
of 1,3 12 ppb As, and the
individually matched
control (i.e., low-dose)
group used DW containing
a mean of 16 ppb As. They
also considered the
concentration of inorganic
arsenic and methylated
metabolites in urine.
Results
87 C As/8 19 mitotic cells among smelter workers
and 13 CAs/1012 mitotic cells in controls. Person
with highest CA frequency had also been exposed to
lead and selenium.
No hint of any effect of inorganic arsenic on CA or
SCE frequencies was seen, even though there was
an approximately 9-fold difference in the mean
inorganic arsenic concentrations in DW between the
2 groups.
inorganic arsenic caused ft in CA (chromatid and
isochromatid deletions) frequency in lymphocytes
and an ft in MN frequency in exfoliated epithelial
cells obtained from the oral mucosa and from urine
samples. MN frequencies were higher in people
with skin lesions, by a factor of 2.3 in oral mucosa
and 4.3 in urothelial cells. There was also much
more induction of MN in males than in females for
both cell types.
inorganic arsenic caused ft in MN frequency in
exfoliated bladder cells to 1.8x (90% CI, 1.06x to
2.99x). The MN frequency was positively
correlated with the urinary concentration of
inorganic arsenic plus methylated metabolites. In
contrast, inorganic arsenic had no effect on the MN
frequency in epithelial cells obtained from the
buccal mucosa.
The exfoliated cell MN assay using FISH with a
centromeric probe was applied: frequencies of MN
containing acentric fragments (MN-) and those
containing whole chromosomes (MN+) both showed
ft, to 1.65x (statistically significant) and 1.37x
(p = 0.15), respectively, suggesting that arsenic has
clastogenic and possibly even aneuploidogenic
properties. Effect was stronger in males than in
females. Thus, in males the increases were 2.06x (p
= 0.07) and 1.86x (p = 0.08), respectively. The
frequencies of MN- and MN+ were both positively
correlated with urinary arsenic and its metabolites.
Reference
Beckman et
al., 1977
Vig et al.,
1984
Gonsebatt et
al., 1997
Warner et
al., 1994
Moore et
al., 1996
C-23   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)


Induction of CAs
and SCEs
(lymphocytes for
CAs)











Population
Sampled
People in
Santa Ana
(high dose)
and Nuevo
Leon (low-
exposure
group),
Mexico








Information on Exposure
Levels and Durations and
on Biomarkers
The high-exposure group of
1 1 used DW containing a
mean of 390 ppb arsenic
(98% as Asv), and the low-
exposure group of 13 used
DW that ranged from 19 to
60 ppb As. They also
considered arsenic
concentrations in urine.









Results
Examined the levels of CAs and SCEs in peripheral
blood lymphocytes. There were no skin lesions in
the control subjects, but 4 of the 11 exposed subjects
had cutaneous signs of arsenicism The percentages
of total CAs and SCEs were similar in the two
groups; however, the finding of a higher point
estimate of the frequency of complex CAs (i.e.,
dicentrics, rings, and translocations) in the high-
exposure group was considered suggestive of a
possible effect of inorganic arsenic. Average
generation times (AGT) of lymphocytes were 19.02
hr in the laboratory control, 19.90 hr in the low-
exposure group, and 28.70 hr in the high-exposure
group, with this difference being statistically
significant. It was suggested that this effect might
suggest an impairment of the immune response.


Reference
Ostrosky-
Wegman et
al., 1991













DNA Damage
DNA damage
detected using
SCGE (comet)
assay
(lymphocytes)














New
Hampshire,
USA
















Low-exposure (control)
group had < 0.7 ppb arsenic
in DW and high-exposure
group had >13 (nd up to 93)
ppb arsenic in DW.














Using the SCGE (comet) assay, baseline DNA
damage as well as the capacity of the lymphocytes
from these subjects to repair damage induced by an
in vitro challenge with 2-AAAF were assessed. 2-
AAAF was used because its adducts are primarily
repaired through the NER pathway. High-exposure
group had ft in baseline damage (i.e., damage
resulting from inorganic arsenic exposure only) to
~1.8x. Two hours after identical in vitro 2-AAAF
treatments to cells from both high- and low-
inorganic arsenic -exposure groups, cells from both
groups showed big ft in DNA damage, with
inorganic arsenic-high-exposure group showing
-15% more DNA damage than control (NSE). Aftei
4-hr repair period, significantly more DNA damage
remained in lymphocytes from individuals in high-
exposure group (~1.54x), and essentially all 2-
AAAF-induced DNA damage had been repaired in
the control cells.




Andrew et
al., 2006













C-24   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
Oxidative DNA
damage




















Correlation of
urinary
8-OHdGwith
urinary metal
elements and
many other
substances






Levels of urinary
8-OHdG
following acute
arsenic
poisoning
incident









Population
Sampled
Residents of
Bayingnorme
n
(Ba Men),
Inner
Mongolia,
China, with
exposures to
a wide range
of
concentration
s of inorganic
arsenic in
DW








6 regions of
Japan











Wakayama,
Japan












Information on Exposure
Levels and Durations and
on Biomarkers
Concentrations of inorganic
arsenic in DW were
determined for individuals;
-70% of subjects used DW
containing nondetectable
arsenic through 200 ppb As,
with the rest using DW
containing up to -830 ppb
As, with all exposures
lasting at least 5 years.
They also determined
arsenic levels in toenail
clippings as a biomarker of
exposure.








128 men and 120 women
from Japan who did not live
within several kilometers of
large chemical factories or
garbage incinerator facilities








63 people were poisoned by
eating food contaminated
with arsenic trioxide, with 4
dying about 12 hours after
eating. Doses in individuals
were poorly known.










Results
OGG1 expression was used as an indicator of
oxidative stress. OGG1 was selected because it
codes for the enzyme 8-oxoguanine DNA
glycosylase, which is involved in base excision
repair of 8-oxoguanine residues that result from
oxidative damage to DNA. The study found that
OGG1 expression was closely linked to the levels of
arsenic in the drinking water and toenails of the
individuals examined, indicating a link between
ROS damage to DNA and arsenic exposure in
humans. There were no significant differences in
arsenic-induced expression due to gender, smoking,
or age. OGG1 expression was also associated with
skin hyperkeratosis in males, and there was a hint of
the same in females. There was an inverse
relationship between OGG1 expression and Se
levels in toenails, indicating possible protective
effects of Se against arsenic -induced oxidative
stress. The maximal OGG1 response appeared to be
at a water arsenic concentration of 149 ppb, after
which its expression leveled off and was gradually
down-regulated.
The association was investigated between urinary
concentrations of 8-OHdG and urinary
concentrations of As, Al, Cr, Ni, Hg, Zn, Cu, Pb (in
ng of element/mg creatinine) as well as with 5
antioxidants and several other substances.
Statistically significant positive correlations were
found with As, Cr, and Ni and not with any other
substances. (The correlation coefficient for arsenic
was 0.25.) It thus appears that exposure of healthy
people to these 3 metals under normal conditions
may increase oxidative DNA damage. Urinary
arsenic levels ranged from -0 to -230 ng As/mg
creatinine.
Some interesting observations were made among the
52 poisoned individuals who were tested for 8-
OHdG levels in urine following acute poisoning.
After 30 days, urinary 8-OHdG levels were
maximal, with a mean for all patients of ~1.5xthe
normal level in Japanese people. By 180 days after
the poisoning, levels returned to normal. About
37% of the patients never showed any increase in
the concentration of 8-OHdG in urine. The same
paper documented a significant increase in urinary
8-OHdG in people from Outer Mongolia, China,
who drank water contaminated with about 130 ppb
As. The increase in urinary 8-OHdG disappeared
after they drank "low-arsenic" water for 1 year.


Reference






Mo et al.,
2006

















Kimura et
al., 2006











Yamauchi et
al., 2004









C-25   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
DNA damage in
peripheral blood
lymphocytes
detected by
alkaline comet
assay



Population
Sampled
West Bengal,
India






Information on Exposure
Levels and Durations and
on Biomarkers
Low-exposure (control)
group had 7.7+0.5 ppb
arsenic in DW. High-
exposure group had 247+19
ppb arsenic in DW. They
also considered arsenic
levels in nails, hair, and
urine.


Results
Used SCGE (comet) assay with DNA denaturation
at pH >13. High-exposure group had significantly
more DNA damage in lymphocytes. Assay was also
combined with FPG enzyme digestion to
demonstrate that arsenic induced oxidative base
damage.




Reference
Basuetal.,
2005






DNA Repair Inhibition or Stimulation
Decreased DNA
repair
(lymphocytes)





















Decreased DNA
repair
(lymphocytes)







New
Hampshire,
USA, and the
towns of
Esperanza
and Colonia
Allende,
Mexico
















New
Hampshire,
USA







Subjects from New
Hampshire were from an
ongoing epidemiological
study of bladder cancer.
Low-exposure (control)
group had 0.007-5.3 ppb
(average of 0.7) arsenic in
DW. High-exposure group
had 10.4-74.7 ppb (average
of 32) arsenic in DW.
Subjects from Colonia
Allende had 5.5 ± 0.20 ppb
arsenic in DW, and those
from Esperanza had 43.3 ±
8.4 ppb arsenic in DW.
Comparisons between the
low (i.e., control) and high
exposure groups used either
5 (for protein analysis) or 6
ppb (for mRNA analysis) as
the dividing line between
low and high. They also
considered arsenic levels in
urine and toenails.
Subjects from New
Hampshire were from an
ongoing epidemiological
study of bladder cancer.
They compared levels of
expression of 5 NER genes
in 6 cases and 10 controls
with the inorganic arsenic
levels in their DW and in
their toenails.
Earlier work suggested that inorganic arsenic
exposure was correlated with decreased expression
of the nucleotide excision repair genes ERCC1,
XPB, and XPF. This study focused on ERCC1 and,
besides considering gene expression, it looked at
both the protein and DNA repair functional levels
(for latter, see part of study described in DNA
damage part of this table). Inorganic arsenic
exposure was associated with U in expression of
ERCC1 in isolated lymphocytes both at the mRNA
and protein levels. In combined data, there was a U
to -0.71 x, with a significant effect in New
Hampshire alone and in the total data. Estimate of
effect in Mexico was U to ~0.84x (NSE). U in
ERCC1 protein level to ~0.28x was also
demonstrated in high-exposure group in New
Hampshire.







Toenail and DW arsenic levels were inversely
correlated with expression of ERCC1, XPB, and
XPF. The arsenic levels in toenails were more
strongly negatively correlated with the changes in
gene expression that the arsenic concentrations in
DW. In these comparisons, expression levels were
compared between high and low levels of arsenic
exposure. By definition a high level in DW was
anything >2 ppb arsenic and a high level in toenails
was anything >2 ppm As.







Andrew et
al., 2006















Andrew et
al., 2003








Effects Related to Oxidative Stress (ROS)
Evidence of
oxidative
damage to DNA
caused by As,
but not
necessarily from
inorganic arsenic
inDW


Taichung
County,
Taiwan







School children ages 10-12,
with attention being given
to possibility of oxidative
stress to DNA from
exposure to environmental
pollutants As, Cr, and Ni.
No information given on
concentrations of inorganic
arsenic in DW.

When oxidative damage occurs in DNA, the excised
8-OHdG adduct is excreted into urine and is a
biomarker of oxidative stress. In this cross-sectional
study, subjects with higher urinary arsenic tended to
have more (19% more, p = 0.09) urinary 8-OHdG
than those with lower urinary As. Cr was also on
the borderline of showing a significant ft; when both
arsenic and Cr were at a higher level in urine, there
was a highly significant ft of 39% in urinary 8-
OHdG.



Wongetal.,
2005





C-26   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
Evidence of
oxidative
damage to DNA
caused by
inorganic arsenic
in DW, and the
relationship of
that DNA
damage to
arsenic -related
skin lesions












Population
Sampled
2 villages in
Wuyuan
prefecture in
Hetao Plain,
Inner
Mongolia,
China















Information on Exposure
Levels and Durations and
on Biomarkers
Adults from low-arsenic-
exposure village (mean of
5.3 ppb arsenic inDW) and
from high-arsenic-exposure
village (mean of 158.3 ppb
arsenic in DW). They also
measured levels of MMA
and DMA in the urine, and
the levels of those
metabolites in the urine in
the high-arsenic-exposure
village were at least 17x
higher than they were in the
low-arsenic-exposure
village.









Results
When oxidative damage occurs in DNA, the excised
8-OHdG adduct is excreted into urine and is a
biomarker of oxidative stress. For subjects without
arsenic-related skin lesions in the high-arsenic-
exposure village, there was no statistically
significant correlation found between inorganic
arsenic, MMA, or DMA and 8-OHdG adducts in the
urine. However, for subjects with arsenic -related
skin lesions in the high-arsenic-exposure village,
there was a significant positive correlation in urine
between levels of each those 3 types of arsenic and
the level of 8-OHdG adducts. There was so much
individual variability that overall there was no
excess of 8-OHdG adducts in urine in the high- As
village compared to the low-As village, even if
restricted to only those with arsenic -related lesions.
An overall comparison did, however, show an
excess of 8-OHdG adducts in urine in the high-
arsenic village among those who had been drinking
well water for more than 12 years when compared to
those who had been drinking it for less than 12
years, regardless of whether they had skin lesions.


Reference
Fujino et al.,
2005




















Gene Mutations
Induction of
HGPRT
mutations
(isolated
mononuclear
cells)



People in
Santa Ana
(high dose)
and Nuevo
Leon (low-
exposure
group),
Mexico

The high-exposure group of
1 1 used DW containing a
mean of 390 ppb arsenic
(98% as Asv), and the low-
exposure group of 13 used
DW that ranged from 19 to
60 ppb As. They also
considered arsenic
concentrations in urine.
The frequency of monocytes resistant to
thioguanine (i.e., mutants) was twice as high in the
high-exposure group, but this suggestion of an ft
was not statistically significant.





Ostrosky-
Wegman et
al., 1991






Hypermethylation of DNA
Extent of
methylation of
the promoters of
tumor suppressor
genes p53 and
p!6 (relationship
to arsenicosis)








West Bengal,
India













Criteria for diagnosis of
arsenicosis included a
history of using DW
containing > 50 ppb arsenic
for more than 6 months and
presence of skin lesions
characteristic of chronic
arsenic toxicity.
Comparisons were made to
individuals without skin
lesions or those who live in
non-arsenic affected areas.



Methylation of the p53 promoter region of DNA
obtained from blood samples was studied using
methyl-sensitive restriction endonuclease HP AIL
Methylation of p!6 was studied using bisulfite
modification of the DNA followed by methyl
sensitive PCR. Hypermethylation of the promoter
region of both genes was observed in people with
arsenicosis, and there was a positive dose-response
for this hypermethylation. There was a strong
suggestion that the promoter region of p53 is
hypermethylated in individuals with arsenic -induced
skin cancer in comparison to those with skin cancer
unrelated to inorganic arsenic exposure, but this
comparison did not reach statistical significance (p <
0.2)




Chanda et
al., 2006









C-27   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
Relationship
between
epigenetic
silencing of 3
tumor suppressor
genes and
exposure to
arsenic in
patients with
bladder cancer

Population
Sampled
New
Hampshire
patients with
bladder
cancer





Information on Exposure
Levels and Durations and
on Biomarkers
Estimated internal dose of
arsenic exposure from
toenail measurements. 18
patients with bladder cancer
had >0.26 ppm arsenic in
their toenails, and 318 had
O.26 ppm arsenic in their
toenails. 0.26 ppm was the
95th percentile of arsenic
exposure in this population.


Results
They applied methylation-specific PCR. A
significant relationship was identified between
arsenic exposure and promoter methylation of
RASSFlAandPRSSSbutnotpie™5^. The
promoter hypermethylation was associated with
advanced tumor state. Thus the data provide a
potential link between arsenic exposure and
epigenetic alterations in patients with bladder
cancer.



Reference


Marsit et al.,
2006






Hypomethylation of DNA
Extent of
methylation of
the promoters of
tumor suppressor
genes p53 and
p!6
(relationship to
arsenicosis)






West Bengal,
India












Criteria for diagnosis of
arsenicosis included a
history of using DW
containing >50 ppb arsenic
for more than 6 months and
presence of skin lesions
characteristic of chronic
arsenic toxicity.
Comparisons were made to
individuals without skin
lesions or those who live in
non-arsenic-affected areas.


Methylation of the p53 promoter region of DNA
obtained from blood samples was studied using
methyl-sensitive restriction endonuclease HP AIL
Methylation of p!6 was studied using bisulfite
modification of the DNA followed by methyl
sensitive PCR. In the study described in the row
above, a small number of people with high arsenic
exposure showed hypomethylation.
Hypomethylation occurs only after prolonged
arsenic exposure at higher doses. The authors noted
that cases of both hyper- and hypomethylation
leading to silencing of tumor suppressor genes and
activation of oncogenes have been documented in
different types of cancers.
Chanda et
al., 2006












Immune System Response
Association
between
biomarkers of
lung
inflammation
and level of
inorganic arsenic
exposure from
DW








Ajo and
Tucson,
Arizona,
USA













40 subjects were from the
high-arsenic-exposure town
of Ajo (20.3 +3. 7 ppb
arsenic in DW), and 33
were from the low-arsenic-
exposure town of Tucson
(4.0 + 2.3 ppb arsenic in
DW). They also measured
inorganic arsenic levels in
urine, with the mean in Ajo
being 2.6 times higher than
that in Tucson.





Proteolytic enzymes including MMP-2 and MMP-9
are continually secreted in the airways, and their
activities are regulated mainly by TIMP-1. The log-
normalized concentrations of these 3 substances in
induced sputum were not significantly different
between these towns. However, after adjusting for
town, asthma, diabetes, urinary MMA/inorganic
arsenic, and smoking history, total urinary arsenic
was negatively associated with MMP-2 and TIMP-1
levels and positively associated with the ratio of
MMP-2/TIMP-landMMP-9/TIMP-l. This
suggests an association between changes in sensitive
markers of lung inflammation and levels of
inorganic arsenic of only -20 ppb in DW. It appears
that inorganic arsenic levels in DW and the extent of
arsenic methylation may be important predictors of
lung metalloproteinase concentrations.




Josyula et
al., 2006











Signal Transduction
C-28   DRAFT—DO NOT CITE OR QUOTE

-------


Topic(s)
Association
between TGF-a
and/or EGFR
and cumulative
inorganic arsenic
exposure from
DW



















Population
Sampled
Taiwan
























Information on Exposure
Levels and Durations and
on Biomarkers
150 persons were selected
from the arseniasis-endemic
area in Ilan county in
northeast Taiwan, with 30
each coming from those
having residential well
water in the following
ranges (all inppb of As): 0-
50, >50-100, >100-300,
>300-600, and >600. Of
them, the 66 who agreed to
participate in medical
surveillance were compared
to 35 healthy individuals
with no known arsenic
exposure. Those with
arsenic exposure were
further divided on the basis
of cumulative arsenic dose
(i.e., total DW inorganic
arsenic levels x years of
exposure) into the following
2 groups: 32 with <6 ppm-
years and 34 with >6 ppm-
years.


Results
Blood plasma was collected and tested for TGF-a
and EGFR levels using immunoassays. No
relationship between arsenic exposure and EGFR
protein levels was found. However, both levels of
plasma TGF-a and the proportion of individuals
with TGF-a overexpression were significantly
higher in the high cumulative arsenic exposure
group than in the control group. After adjusting for
age and sex, there was also a significant linear trend
between cumulative arsenic exposure and the
prevalence of plasma TGF-a overexpression.
















Reference






Hsu et al.,
2006


















C-29   DRAFT—DO NOT CITE OR QUOTE

-------
Table C-2. In vivo experiments on laboratory animals related to possible modes of action
          of arsenic in the development of cancer—only oral exposures

Tissue or Cell
Type/Species

Arsenic
Species
Dose in
Elemental Asa
(in Units Stated)
Duration
of
Treatment

LOELb


Results



Reference
Aberrant Gene or Protein Expression
Lung/mouse
(C57BL/6)













Lung/mouse
(C57B16, male,
21 days of age at
start of exposure)



Urothelial
cells/rat
(F344, female)





















AsmSA














AsmSA





DMAV
sodium
cacodyl
ate-
trihydrat
e

















* 5.8, 28.8 ppm
(DW)













10, 50 ppb (DW)
Note in ppb!





*0.35, 1.4, 14,35
ppm (DW)





















8wk














4wk






28 days






















28.8 ppm














50 ppb






0.35 ppm






















mRNA levels were determined in a
microchip analysis and validated using
real-time PCR: 29 genes were up-
regulated and 42 down-regulated. 15%
of affected genes were associated with
inflammation, including HSP27 and
HSP90 (both up-regulated). Numerous
extracellular matrix genes were affected,
as reflected in phenotypic lung changes
related to the organization of elastin and
collagen. Protein levels were
determined by a Western blot assay: ft
for 4 genes, U for 14. No correlation
was found between altered genes and
altered proteins.
Protein levels in BALF determined by
proteomic analysis: it is unclear if
samples from 10 ppb were examined, ft
after dose of 50 ppb: peroxiredoxin-6
and enolase 1. U after dose of 50: GST-
omega- 1, RAGE, contraspin, and
apolipoproteins A-I and A-IV.
Microarray analysis using chip for 4395
genes: gene trees generated by
hierarchical clustering of the 510
responsive genes showed marked
changes at every dose in comparison to
the dose (or dose of 0) below it. Of the
510 genes, 38% were up-regulated and
9% down-regulated by >3-fold. Most
affected genes related to the functional
categories of apoptosis, cell cycle
regulation, adhesion, signal transduction,
stress response, or growth factor and
hormone receptors. There was a change
in the types of genes affected at the
different doses, particularly when
comparing the higher 2 doses (both
cytotoxic) with the 2 non-cytotoxic
doses. The dose with most genes
affected was 14 ppm. At the lowest
dose, 503 genes (11%) were
significantly affected, of which 41%
were up-regulated and 6% down-
regulated by >3-fold.




Lantz and
Hays, 2006










Lantz et al.,
2007









Senetal.,
2005
















                                      C-30   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Liver cells/mouse
(129/SvJ)








Brain, liver,
placenta/mouse
(only pregnant
ICR females
drank the water)





Fetal brain, fetal
liver/mouse
(only pregnant
ICR females
drank the water)




Liver/mouse
(BALB/c, male)








Liver/mouse
(BALB/c, male)







Arsenic
Species

AsmSA








AsmSA









AsmSA








AsmSA









AsmSA







Dose in
Elemental Asa
(in Units Stated)
45 ppm (DW)









* 4.35 mg/kg
(gavage)








* 4.35 mg/kg
(gavage)







50, 100, 150
ug/mouse/day for
6 days/week
(gavage)






50, 100, 150
ug/mouse/day for
6 days/week
(gavage)




Duration
of
Treatment
48 wk









1 time only
on each of
9 days,
gestation
days 7 to 16





1 time only
on each of
9 days,
gestation
days 7 to 16




3, 6, 9, 12
months








3, 6, 9, 12
months







LOELb

45 ppm









4.35 mg/kg









4.35 mg/kg









50
at 9 and 12
months
only






50
at 9 and 12
months





Results

Microarray analysis, RT-PCR, and
immunochemistry : big ft in ER-a and
cyclin Dl mRNA and protein levels. Of
588 genes tested in microarray analysis,
30 showed aberrant expression,
including steroid-related genes,
cytokines, apoptosis-related genes, cell
cycle-related genes, and genes encoding
for growth factors and hormone
receptors.
Activities of selenoenzymes GPx, TrxR,
DI-I, DI-II, and DI-m in maternal tissues
when examined on gestation day 17 of
their litter:
liver: U of DI-I to ~0.61x when Se-
adequate diet;
liver: U of DI-I to ~0.30x when Se-
deficient diet;
all other comparisons were either slight
or NSE.
Activities of selenoenzymes GPx, TrxR,
DI-I, DI-II, and DI-m in fetal tissues
when examined on gestation day 17:
brain: ft of DI-II to ~4.1x when Se-
deficient diet;
liver: U of TrxR to ~0.78x when Se-
deficient diet;
all other comparisons were either slight
or NSE.
Levels of TNF-a and IL-6:
NSE on either one at any dose in first 6
months.
At 9 months:
TNF-a: 50, ~1.2x; 100, ~1.2x; 150,
IL-6: 50, ~2.0x; 100, ~2.5x; 150, ~2.7x.
At 12 months:
TNF-a: 50, ~1.9x; 100, ~2.3x; 150,
~3.0x;
IL-6: 50, ~2.8x; 100, ~5.7x; 150, ~9.5x.
Concentration of total collagen:
At 3 months: NSE at all doses, but hint
of ft at 100 (~1.2x) and 150 (~1.3x).
At 6 months: NSE at all doses, but hint
of ft at 100 (~1.3x) and 150 (~1.4x).
At 9 months: 50, ~1.3x; 100, ~1.4x; 150,
At 12 months: 50, ~1.5x; 100, ~1.9x;
150, ~2.1x.


Reference


Chen et al.,
2004b








Miyazaki et
al., 2005








Miyazaki et
al., 2005








Das etal.,
2005








Das etal.,
2005



C-31  DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Liver, kidney,
and lung/mouse
(B6C3F1,
female)







Liver and
kidney/mouse
(B6C3F1,
female)





HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)






HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)





HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)





Arsenic
Species
AsmSA


Asv
sodium
arsenate

DMAV



AsmSA








AsmSA










AsmSA









AsmSA








Dose in
Elemental Asa
(in Units Stated)
In all cases,
dissolved in water
and administered
once by gavage:
* 9.58 mg/kg for
all


* 9.58 mg/kg for
all
*391 mg/kg for all
In all cases,
dissolved in water
and administered
once by gavage:
* 0.0749, 0.749,
2.25, 7.49 mg/kg
for both


42.5, 85 ppm
(DW);
report did not state
if molecular
analysis was done
on one or both of
these doses
combined



42.5, 85 ppm
(DW);
report did not state
if molecular
analysis was done
on one or both of
these doses
combined


42.5, 85 ppm
(DW);
report did not state
if molecular
analysis was done
on one or both of
these doses
combined

Duration
of
Treatment
One dose
for all









One dose








10 days,
gestation
days
8 to 18







10 days,
gestation
days
8 to 18






10 days,
gestation
days
8 to 18






LOELb

9.58 mg/kg


9.58 mg/kg


None




2.25 mg/kg
in liver

7.49 mg/kg
in kidney




42.5 or 85
ppm









42.5 or 85
ppm








42.5 or 85
ppm








Results

HMOX-1 activity 6 hr after the single
oral dose was administered by gavage:
Liver: Asm, ~7.5x; Asv, ~5.1x, DMAV,
~0.96x (NSE).
Kidney: Asm, ~7.6x; Asv, ~3.2x,
DMAV, ~1.03x(NSE).
Lung: none of the arsenicals induced
HMOX-1 activity.



HMOX-1 activity in liver 6 hr after the
single oral dose was administered by
gavage:
at 2 lower doses, NSE; 2.25, ~2.5x;
7.49, ~7.5x.
HMOX-1 activity in kidney 4 hr after
the single oral dose was administered by
gavage:
at 3 lower doses, NSE; 7.49, ~3.5x.
Comparisons of gene expression based
on microarray analysis, with
comparisons being made between HCC
tumors from offspring of exposed dams
and normal liver tissue from offspring of
unexposed dams: ft of AFP to -18. 5x; U
ofIGF-lto0.78x;
ft of IGFBP-1 to ~8.8x; ft of CK8 to
~2.4x;
ft of CK18 to ~8.8x; U of BHMT to
~0.33x.
Comparisons of gene expression based
on microarray analysis, with
comparisons being made between HCC
tumors of offspring of exposed dams and
spontaneous liver tumors of offspring of
unexposed dams:
ft of AFP to ~6.2x; NSE for IGF-1;
ft of IGFBP-1 to ~1.7x; ft of CK8 to
ft of CK18 to ~5.8x; U of BHMT to
~0.36x.
Comparisons of gene expression based
on microarray analysis, with
comparisons being made between HCC
tumors and normal-appearing liver cells
of offspring of exposed dams:
ft of AFP to ~7.4x; U of IGF-1 to
~0.68x; ft of IGFBP-1 to ~3.7x;
ft of CK8 to ~1.3x; ft of CK18 to -7.0 x;
UofBHMTto~0.32x.


Reference
Kenyon et
al., 2005b









Kenyon et
al., 2005b









Waalkes et
al., 2004b









Waalkes et
al., 2004b








Waalkes et
al., 2004b





C-32   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Liver cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)




HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)



Uterus/mouse
(only pregnant
CD1 females
drank the water,
female offspring
only)













HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)












Arsenic
Species
AsmSA








AsmSA







AsmSA


















AsmSA















Dose in
Elemental Asa
(in Units Stated)
42.5, 85 ppm
(DW);
report did not state
if molecular
analysis was done
on one or both of
these doses
combined

42.5, 85 ppm
(DW);
report did not state
if molecular
analysis was done
on one or both of
these doses
combined
85 ppm (DW)


















42.5, 85 ppm
(DW)














Duration
of
Treatment
10 days,
gestation
days 8 to 18






10 days,
gestation
days 8 to 18





10 days,
gestation
days 8 to 18
















10 days,
gestation
days 8 to 18














LOELb

42.5 or 85
ppm







42.5 or 85
ppm






85 ppm if
also treated
with DBS
orTAM















42.5 ppm
















Results

Comparisons of gene expression based
on microarray analysis, with
comparisons being made between
normal-appearing liver cells in both
offspring of exposed dams and
unexposed dams:
ft of AFP to ~2.5x; ft of IGF- 1 to -l.lx;
ft of IGFBP-1 to ~2.4x; ft of CK8 to
NSEforCKlSorBHMT.
In general, the results in the 4 previous
rows were confirmed by real-time RT-
PCR analysis. Aberrant gene expression
was also noted in the microarray
analysis for numerous other genes
including those related to cell
proliferation, oncogenes, stress, and
metabolism.
Expression (by real-time RT-PCR) of
various estrogen-related genes in uteri at
1 1 days of age: ft in ER-a to 1.56x.
Some female offspring were also
exposed by subcutaneous injection to
DBS on the first 5 days after birth. DBS
alone or (inorganic arsenic + DBS) did
not significantly increase ER-a
expression. Inorganic arsenic alone did
not ft expression of pS2, CYP2A4, or
lactoferrin. However, DBS alone caused
large ft in expression of all 3 of these
genes, and (inorganic arsenic + DBS)
caused a further ft to 3.0 times, 7.8
times, and 1.47 times that of DBS alone,
respectively. These and other results
showed that inorganic arsenic acts with
estrogens to enhance production of
urogenital cancers in female mice.
Comparisons of gene expression based
on microarray analysis of RNA, with
comparisons being made between HCC
tumors from offspring of exposed dams
and normal (i.e., non-tumorous) liver
tissue from offspring of unexposed
dams: 13.7% of 600 genes were
significantly up-regulated or down-
regulated. Only 7.7% of those 600
genes were similarly affected in
spontaneous tumors in liver tissue from
offspring of unexposed dams. The 600
genes studied included oncogenes and
genes associated with cell proliferation,
differentiation, or otherwise related to
cancer outcome.


Reference


Waalkes et
al., 2004b







Waalkes et
al., 2004b








Waalkes et
al., 2006a

















Liu et al.,
2004










C-33   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)










Liver cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)






HCC cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)











Liver cells/mouse
(only pregnant
C3H females
drank the water,
male offspring)




Arsenic
Species
AsmSA














AsmSA










AsmSA















AsmSA







Dose in
Elemental Asa
(in Units Stated)
42.5, 85 ppm
(DW)













42.5, 85 ppm
(DW)









85 ppm (DW)















85 ppm (DW)







Duration
of
Treatment
10 days,
gestation
days
8 to 18











10 days,
gestation
days
8 to 18







10 days,
gestation
days
8 to 18












10 days,
gestation
days
8 to 18





LOELb

42.5 ppm














42.5 ppm










85 ppm















85 ppm








Results

Comparisons of gene expression based
on microarray analysis of RNA (see row
above): up-regulated genes included
oncogene/tumor suppressor genes and
genes related to cell proliferation,
hormone receptors, metabolism, stress,
apoptosis, growth arrest, and DNA
damage. A wide array of different types
of genes was also down-regulated.
Real-time RT-PCR analysis largely
confirmed the findings of microarray
analysis. The higher dose tended to
yield more significant differences, but a
positive dose-response was not always
evident.
Comparisons of gene expression based
on microarray analysis of RNA, with
comparisons being made between non-
tumorous liver cells in both offspring of
exposed dams and unexposed dams:
-10% of 600 genes were significantly
up-regulated or down-regulated. The
600 genes studied included oncogenes
and genes associated with cell
proliferation, differentiation, or
otherwise related to cancer outcome.
Comparisons of gene expression based
on microarray analysis of RNA, with
comparisons being made between HCC
tumors from offspring of exposed dams
and normal liver tissue from offspring of
unexposed dams: statistically significant
alterations in expression were seen for
2,540 genes. Real-time RT-PCR and
Western blot analyses of selected genes
or proteins showed >90% concordance.
Affected gene expression included
oncogenes, HCC biomarkers, cell
proliferation-related genes, stress
proteins, insulin-like growth factors,
estrogen-linked genes, and genes
involved in cell-cell communication.
Comparisons of gene expression based
on microarray analysis of RNA, with
comparisons being made between non-
tumorous liver cells in both offspring of
exposed dams and unexposed dams:
statistically significant alterations in
expression were seen for 2010 genes.
See row above for results in HCC cells.


Reference



Liu et al.,
2004













Liu et al.,
2004










Liu et al.,
2006c












Liu et al.,
2006c




C-34   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Fetal
livers/mouse
(only pregnant
C3H females
drank the water,
male offspring)






Livers of
newborn
males/mouse
(only pregnant
C3H females
drank the water)






Liver and liver
tumors/mouse
(only pregnant
C3H females
drank the water)


















Arsenic
Species
AsmSA











AsmSA











AsmSA





















Dose in
Elemental Asa
(in Units Stated)
85 ppm (DW)











85 ppm (DW)











85 ppm (DW)





















Duration
of
Treatment
10 days,
gestation
days
8 to 18








10 days,
gestation
days
8 to 18








10 days,
gestation
days
8 to 18



















LOELb

85 ppm











85 ppm












Results

Comparisons of gene expression based
on microarray analysis of RNA from
fetal livers just after treatment ended,
with confirmation by real-time RT-PCR:
alteration of expression of 187 genes (of
22,000 in array) was demonstrated, with
-25% of them being related to either
estrogen signaling or steroid
metabolism — some with dramatic (here
meaning »100x) up-regulation.
Expression of some genes important in
methionine metabolism was suppressed.
Comparisons of gene expression based
on microarray analysis of RNA from
livers of newborn males, with
confirmation by real-time RT-PCR:
among 600 genes examined, marked
alteration of expression of 40 genes was
demonstrated. Affected genes included
genes related to stress (several in the
glutathione system), metabolism (several
cytochrome P450 genes), growth factors
(several insulin-like growth factor
genes), and hormone metabolism.
Samples from adults of both sexes were tested.
Some had had a post-weaning 21-wk dermal
treatment with TPA. Comparisons with the TPA-
treatment-only control were made regarding gene
expression based on microarray analysis of RNA,
with confirmation by real-time RT-PCR. Alteration
of expression of ~70 genes (of 588 in array) was
demonstrated. There were generally similar gene
alteration patterns in both sexes both in inorganic
arsenic/TPA exposed non-tumorous livers and in
inorganic arsenic/TPA-induced tumors. The tumors
themselves generally had more pronounced
alterations in gene expression than the normal tissue
around them. In general, the inorganic arsenic/TPA-
induced gene expression alterations were similar to
those seen in liver samples from male mice exposed
only to inorganic arsenic in utero. It should be noted
that while in utero inorganic arsenic -exposed males
developed hepatocellular carcinoma without the TPA
treatment, in utero inorganic arsenic-exposed
females only developed those tumors after TPA
treatment.


Reference



Liuetal.,
2007a










Xie etal.,
2007












Liu et al.,
2006b















C-3 5  DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Bladder and
liver/rat
(Fisher 344,
male)












Lung/mice
(C57BL/6J
Oggl+/+wtmice
and Oggl"'"
knockout mice,
both sexes, 14
weeks old at start
of treatment)
Liver cells/rat
(Sprague
Dawley)




Liver cells/rat
(Sprague
Dawley)









Skin/mouse
(homozygous,
strain Tg. AC,
female)


Arsenic
Species
MMAV

DMAV

TMAVO











DMAV







Asvas
Na2HAs
Cv
7H2O



Asvas
Na2HAs
Cv
7H2O








AsmSA




Dose in
Elemental Asa
(in Units Stated)
* 121ppm(DW)c

* 109 ppm (DW)C

* 110ppm(DW)c











* 115.3 ppm (DW)







* 0.24, 2.4, 24
ppm (DW)





* 0.24, 2.4, 24
ppm (DW)










200 ppm (DW)




Duration
of
Treatment
20 days
for all














4 weeks







1 month






4 months











4, 10 wk





LOELb


Results

Changes in gene expression observed in cDNA
microarray analysis: MMAV caused ft for 20 genes
and U for 1 gene in liver and ft for 5 genes and U for
5 genes in bladder. DMAV caused ft for 15 genes
and U for 2 genes in liver and ft for 13 genes and U
for 4 genes in bladder. TMAVO caused ft for 23
genes and U for 2 genes in liver and ft for 6 genes
and U for 7 genes in bladder. Groups of genes
affected by all arsenicals in both tissues included
genes related to xenobiotic metabolism, growth
factor receptors, and energy metabolism. In the
liver, phase I and II metabolizing enzymes were
induced to a lesser extent by MMAV and DMAV than
by TMAVO, and in the bladder they were induced
only by DMAV. CYP1A1 was only overexpressed
by TMAVO and in liver.
Results of an Affymetrix oligonucleotide microarray
analysis: a change in expression was found for 165
and 182 genes in male and female knockout Oggl"'"
mice, respectively. In DMAv-treated knockout
Oggl"7" mice, there was marked induction of Polal,
CYP7B1, NdfuaS, MMP-13 and other genes specific
to cell proliferation, cell signaling, and xenobiotic
metabolism.
Various






Various











200 ppm




Determination of mRNA levels of
cancer-related genes using real-time
quantitative RT-PCR:
ft cyclinDl at 2.4 only; ft p27Klpl at 2.4
only;
ft ILK at 0.24 only; U PTEN at 0.24
only; U 3-catenin at 24 only.
Determination of mRNA levels of
cancer-related genes using real-time
quantitative RT-PCR: ft cyclin Dl at 24
only; ft ILK at 0.24 and 2.4;
ft p27Klpl at 0.24 only; U PTEN at all
doses;
U 3-catenin at all doses. Results were
confirmed by protein levels.
(Histograms were assumed to be correct
for ILK and p27; the descriptions for
them appear to have become reversed in
the text.)
Kinetics of mRNA expression based on
RT-PCR:
EGFR and TNF-a: ft by week 10;
GM-CSF and TGF-a: ft by week 4; big
ft by week 10; c-myc: NSE.


Reference




Kinoshita et
al., 2007a












Kinoshita et
al., 2007b






Cuietal.,
2004b






Cuietal.,
2004b








Germolec et
al., 1998


C-36   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Heart/mouse
(C57BL/6NCr,
male)











Blood
plasma/mouse
(C57BL/6NCr,
male)
Tumors that
developed from
B16-F10(GFP)
melanoma tumor
cells/ mice
(NCr nu/nu,
male)




Arsenic
Species
AsmSA













AsmSA



AsmSA









Dose in
Elemental Asa
(in Units Stated)
0.05, 0.25, 0.5
ppm
(DW)











0.5 ppm (DW)



10, 50, 200 ppb
(DW)








Duration
of
Treatment
5, 10, 20
wk












20 wk



9wk










LOELb

Various













0.5 ppm



10 ppb










Results

mRNA levels determined by RT-PCR:
VEGF165: ft at 0.25 and 0.5 at wk 5; ft at
all doses at wk 10; NSE at wk 20;
VEGFR1 : NSE at wk 5 and 10; big ft at
0.25 and big U at 0.5 at wk 20;
VEGFR2: ft at 0.5 at wk 5; NSE at wk
10; ft at 0.05 and 0.25 and
U at 0.5 at wk 20;
PAI-1: NSE at wk 5; ft at 0.5 at wk 10;
ft at 0.25 and 0.5 at wk 20;
Endothelin-1: NSE at wk 5 and 10; ft at
0.05 and big ft at 0.25 at wk 20;
MMP-9: NSE at wk 5; ft at 0.5 at wk 10;
ft at all doses at wk 20.
PAI-1 protein levels determined by
ELISA assay:
ftto~1.33x.

Protein levels in primary melanoma
tumors determined by
immunohistochemical staining:
ft HIF-la at 10 and 50 only; ft VEGF at
10 and 200 only, ft for both proteins
was just locally around tumor blood
vessels. Western blot assay of whole
tumor lysates showed no more than
barely detectable ft of HIF-la at any
dose.


Reference




Soucy et al.,
2005








Soucy et al.,
2005





Kamat etal.,
2005





Apoptosis
Bladder and
liver/rat
(Fisher 344,
male)



Liver/rat
(Wistar, male)









MMAV

DMAV

TMAVO



AsmSA









* 121ppm(DW)c

* 109 ppm (DW)C

* 110ppm(DW)c


*0.03, 1.4,2.9
ppm (DW)









5, 10, 15,
and 20 days
for all




60 days










None

Various

Various


1.4










Apoptosis labeling index based on an
immunochemistry method of staining
single-stranded DNA:
Bladder: ft on day 20 to ~1.5x for
DMAV only;
Liver: ft on day 20 to ~3.3x for TMAVO
only.
Induced apoptosis (experimental -
control) based on TUNEL assay with PI
staining and analysis using fluorescence
microscopy:
0.03, 5.0; (NSE); 1.4, 14.9; 2.9, 22.3;
these results were consistent with DNA
ladder formation found by agarose gel
electrophoresis for which there was an ft
at 1.4; bigger ft at 2.9. There was also
microscopic evidence of cell death by
necrosis.


Kinoshita et
al., 2007a





Bashir et al.,
2006a







C-37   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Kidney,
leukocytes and
liver/rat
(albino Wistar,
male)


Kidney,
leukocytes and
liver/rat
(albino Wistar,
male)











Splenocytes and
thymocytes/mous
e
(C57BL/6,
female)







Brain and
liver/rat
(Wistar, male)










Arsenic
Species

AsmSA






AsmSA














Asvas
Na2HAs
Cv
7H2O









AsmSA










Dose in
Elemental Asa
(in Units Stated)
* 57.7 ppm (DW)






* 57.7 ppm (DW)















0.5, 5, 50 ppm
(DW)










* 3.6, 6.1,7.3
mg/kg
(gavage, with
animals being
killed 24 hr later
for sample
collection)





Duration
of
Treatment
30 days






30 days















8, 12 wk











One dose












LOELb

57.7 ppm






57.7 ppm















50 at 8 wk
for both
cell types









Various












Results

TNF-a levels: kidney, ft ~1.6x; leuko., ft
~2.2x; liver, ft ~1.9x;
caspase-3 levels: kidney, ft ~3.2x;
leuko., ft ~2.8x; liver, ft ~3.5x; effects
on both endpoints in all 3 tissues were
markedly reduced by co-treatment with
AA and/or a-Toc.
Induced percentage of DNA that was
fragmented
(experimental - control):
kidney, ft -17.6%; leuko., ft -17.4%;
liver, ft -2 1.8%.
Induced percentage of TUNEL positive
cells (experimental - control):
kidney, ft -6.7%; leuko., ft -5.1%; liver,
ft -8. 1%; effects on both endpoints in all
3 tissues were markedly reduced by co-
treatment with AA and/or a-Toc.
Confirmation of induced apoptosis in
leukocytes shown by finding typical
DNA ladders after agarose gel
electrophoresis; co-treatment with AA
and/or a-Toc abolished that effect.
Induced apoptosis (experimental -
control) determined by TUNEL method:
Splenocytes: 8 wk: 0.14% of cells at
dose of 50 (or 6.6x); 12 wk: 0.22% of
cells at dose of 50 (or 5.4x).
Thymocytes: 8 wk: 0.40% of cells at
dose of 50 (or 4.0x); 12 wk: 0.28%
(NSE) of cells at dose of 50 (or 2.5x).
For both cell types, the data suggested a
positive dose-response across all doses;
however, the other results showed much
variability.
Brain: caspase-3 activity: ft to ~1.4x
(NSE) at 3. 6,
to ~2.0x at 6.1, and to ~2.6x at 7.3;
Liver: caspase-3 activity: ft to ~1.8x at
3.6, to ~2.5x at 6.1, and to ~3.0x at 7.3.
Both brain and liver: agarose gel
electrophoresis showed DNA
"nucleosomal ladder," suggesting
induction of apoptosis; results were not
quantified. Histopathological
examination also showed evidence of
cellular necrosis.


Reference

Ramanathan
et al., 2005








Ramanathan
et al., 2005














Stepnik et al.,
2005













Bashir et al.,
2006b



C-3 8  DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species

Arsenic
Species
Dose in
Elemental Asa
(in Units Stated)
Duration
of
Treatment

LOELb


Results



Reference
Cancer Promotion
Skin/mouse
(homozygous,
strain Tg. AC,
female)










Skin/mouse
(hairless swiss-
bald strain, male)






















Lung/mouse
(ddY, male)












AsmSA













Asv
sodium
arsenate






















DMAV
assumed
to be
dimethy
lar-sinic
acid








200 ppm (DW)













* 11.4 ppm (DW)
























*217ppm(DW)













14 wk













25 wk
























25 wk













200 ppm













None, but
11.4 ppm if
also treated
with
DMBA




















217 ppm,
but only
following
4NQO
treatment









After low-dose application of TPA on 4
occasions over 2 weeks starting after 3 1
days of inorganic arsenic exposure, there
was a marked ft in the number of skin
papillomas compared to single
treatments, whereas no papillomas
developed in inorganic arsenic -treated
Tg.AC mice without TPA treatment or
in FVB/N mice with the combined
treatment. Injection of neutralizing
antibodies to GM-CSF after TPA
application reduced the number of
papillomas in Tg.AC mice. Inorganic
arsenic acted like a co-promoter.
PCNA protein levels determined by
Western blotting: no PCNA was present
following the inorganic arsenic
treatment alone, compared to the
baseline of 22 units of PCNA in the
control (set equal to x). When mice
were given 4 DMBA treatments (as an
initiating carcinogen) during the first 2
weeks of the inorganic arsenic treatment,
there was PCNA ft to ~5.3x. DMBA
treatment alone caused ft to only 2.9x.
Mice that were untreated or treated with
inorganic arsenic alone developed no
papillomas or skin tumors. DMBA
treatment alone induced development of
squamous cell papillomas. Combined
inorganic arsenic and DMBA treatment
caused development of well-
differentiated squamous cell carcinomas.
Inorganic arsenic acted as a skin tumor
promoter by promoting abnormal cell
proliferation. Findings suggest that
inorganic arsenic is toxic to normal skin
cells and that preneoplastic cells are
more resistant to inorganic arsenic.
Some of the mice were subcutaneously
injected with 10 mg/kg of 4NQO just
before the 25-wk DMA treatment began.
Some of the mice ate only feed
containing 0.05% of the antioxidant
EGCG. Number out of 10 mice in each
group bearing tumors: control, 0; DMA
alone, 0; 4NQO alone, 7; EGCG alone,
0; (4NQO + DMA), 10;
(4NQO + DMA + EGCG), 7. That last
group had only 0.89 tumor/mouse
compared to 3. 10 tumors/mouse in
4NQO group and 4.00 tumors/mouse in
the (4NQO + DMA) group.



Germolec et
al., 1998














Motiwale et
al., 2005





















Mizoi et al.,
2005









C-39   DRAFT—DO NOT CITE OR QUOTE

-------
Tissue or Cell
Type/Species
Arsenic
Species
Dose in
Elemental Asa
(in Units Stated)
Duration
of
Treatment
LOELb
Results
Reference
Cell Cycle Arrest or Reduced Proliferation
Heart/mouse
(C57BL/6NCr,
male)
AsmSA
0.5 ppm (DW)
5, 10, 20
wk
0.5 ppm at
20 wk
Density of microvessels of <12 um
diameter using histopathology and a
digital-imaging subroutine:
U to ~0.82x at 20 wk; hint of a U at 10
wk.
Soucy et al.,
2005
Cell Proliferation Stimulation
Bladder/rat
(F344, female)
Bladder/rat
(F344, female)
Liver/rat
(Fischer 344,
male)
(they used
normal-appearing
tissue)
Bladder and
liver/rat
(Fisher 344,
male)
DMAV
DMAV
TMAVO
MMAV
DMAV
TMAVO
* 54.3 ppm (food)
(assumes MW of
chemical used was
138.0)
* 54.3 ppm (food)
(assumes MW of
chemical used was
138.0)
*27.5, 110.2 ppm
(DW) Estimated
total intakes: 351
and 1363 mg
As/rat
* 121ppm(DW)c
* 109 ppm (DW)C
* 110ppm(DW)c
2wk
26 wk
104 wk
5, 10, 15,
and 20 days
for all
54.3 ppm
54.3 ppm
110.2 ppm
None
Various
Various
Stimulation of proliferation determined
by BrdU labeling assay:
ft to 3.9x; co-treatment with DMPS (a
chelator of trivalent arsenicals)
completely eliminated the effect.
Stimulation of proliferation determined
by BrdU labeling assay:
ft to 1.6x; co-treatment with DMPS (a
chelator of trivalent arsenicals)
completely eliminated the effect.
Histological examination showed simple
hyperplasia in 4 of 9 rats, compared to 0
of 10 rats in control and 1 of 10 rats with
co-treatment with DMPS.
Livers were stained for the analysis of
PCNA by an immunohistochemical
method, with the PCNA index being the
number of positive cells/100 cells: ft in
PCNA index to 2.0x, thereby suggesting
that cell proliferation in the normal-
appearing parenchyma was elevated.
The point estimate of the index was also
ft at lower dose, but the SE for it was
large.
PCNA labeling index based on an
immunochemistry method:
Bladder: ft on day 20 to ~1.8x for
DMAV only;
Liver: ft on day 20 to ~1.8x for TMAVO
only.
Cohen et al.,
2002
Cohen et al.,
2002
Shen et al.,
2003
Kinoshita et
al., 2007a
C-40   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Bladder/mouse
(only pregnant
CD1 females
drank the water,
male offspring
only)











Kidney/mouse
(only pregnant
CD1 females
drank the water,
male offspring
only)





Bladder/mouse
(only pregnant
CD1 females
drank the water,
female offspring
only)













Arsenic
Species
AsmSA
















AsmSA










AsmSA

















Dose in
Elemental Asa
(in Units Stated)
85 ppm (DW)
















85 ppm (DW)










85 ppm (DW)

















Duration
of
Treatment
10 days,
gestation
days
8 to 18













10 days,
gestation
days
8 to 18







10 days,
gestation
days
8 to 18















LOELb

85 ppm if
also treated
with DBS
orTAM













85 ppm










85 ppm if
also treated
with DBS
orTAM















Results

Some male offspring were also exposed
by subcutaneous injection to DBS or
TAM on the first 5 days after birth; all
male offspring were held for 90 wk
before examination. Induced (i.e.,
experimental - control) % of mice with
bladder hyperplasia: inorganic arsenic
alone, 9% (NSE); DBS alone, 12%
(NSE); TAM alone, 10% (NSE);
(inorganic arsenic + DBS), 45%;
(inorganic arsenic + TAM), 30%. All
induced percentages were the same for
total proliferative lesions, except for
(inorganic arsenic + TAM), which was
40%. The lesions induced by inorganic
arsenic with either DES or TAM
overexpressed ER-a.
Some male offspring were also exposed
by subcutaneous injection to DES or
TAM on the first 5 days after birth; all
male offspring were held for 90 weeks
before examination. Induced (i.e.,
experimental - control) % of mice with
cystic tubular hyperplasia: inorganic
arsenic alone, 23%; DES alone, 0%;
TAM alone, 0%; (inorganic arsenic +
DES), 24%; (inorganic arsenic + TAM),
7%.
Some female offspring were also
exposed by subcutaneous injection to
DES or TAM on the first 5 days after
birth; all female offspring were held for
90 wk before examination. Induced
(i.e., experimental - control) % of mice
with bladder hyperplasia: inorganic
arsenic alone, 12% (NSE); DES alone,
0% (NSE); TAM alone, -3% (NSE);
(inorganic arsenic + DES), 26%;
(inorganic arsenic + TAM), 23%. All
induced percentages were the same for
total proliferative lesions, except for
(inorganic arsenic + DES), which was
35%, and (inorganic arsenic + TAM),
which was 26%. Unlike in the male
offspring, inorganic arsenic did not
induce hyperplasia in kidneys.


Reference



Waalkes et
al., 2006b













Waalkes et
al., 2006b












Waalkes et
al., 2006a












C-41  DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Lung/mice
(C57BL/6J
Oggl+/+wtmice,
both sexes, 14
weeks old at start
of treatment)

Lung/mice
(C57BL/6J Oggl
'" knockout mice,
both sexes, 14
weeks old at start
of treatment)
Bladder/mouse
(C57BL/6,
female)


Bladder/mouse
(C57BL/6,
female)


Bladder/mouse
(C57BL/6,
female)






Blood
vessels/chicken
(Leghorn,
chorioallantoic
membranes of
10-day-old
chicken embryos)
Matrigel
implants/mouse
(C57BL/6NCr,
male)









Arsenic
Species
DMAV












AsmSA




AsmSA




AsmSA








AsmSA






AsmSA











Dose in
Elemental Asa
(in Units Stated)
* 115.3ppm(DW)












* 57.7 ppm (DW)




* 57.7 ppm (DW)




* 11.5, 57.7 ppm
(DW)







0.00033, 0.001,
0.0033,0.01,
0.033,0.1,0.33,
1.0,3.3, 10 uM



0.001,0.005,0.01,
0.05 ppm (DW)










Duration
of
Treatment
72 weeks












4wk




16 wk




16 wk








24 hr






5wk












LOELb

None




115.3







57.7 ppm




57.7 ppm




11.5 ppm








0.033 uM






0.001 ppm












Results

PCNA labeling index based on an
immunochemistry method, x = wt
control level:
wt with inorganic arsenic treatment: ft to
~3x (NSE).
Knockout Oggl"'" without inorganic
arsenic: ft to ~6x.
Knockout Oggl"7" with inorganic arsenic
treatment: ft to ~17x.
Results were confirmed in a study with
only a 4 week exposure.


All experimental mice developed mild
hyperplasia of the urinary bladder
epithelium, that being a 3- to 4-fold ft in
the thickness of the transitional cell
layer.
ft in PCNA-stained nuclei in the bladder
epithelium from 2% in control to 3 1% in
experimental group, an indication of big
ft in cell proliferation. Similar ft also
seen at 4 weeks.
Also consistent with ft in proliferation: ft
in DNA binding of the AP- 1
transcription factor to ~1.9x and ~4.7x at
the 2 doses, respectively. At one or both
doses (not specified): 38% and 76% of
the bladder cells stained positive for the
c-jun and c-fos immunoreactive proteins,
respectively, compared to only 2% in
control mice.
CAM assay to determine vascularity
(i.e., bloodvessel density): ft to ~2.2x
and remained at about that level to dose
of 1; U to ~0.28x at dose of 3.3 and
remained at about that level to dose of
10.

Blood vessel no. determined in Matrigel
implants surgically inserted during last 2
wk of inorganic arsenic treatment:
probable ft to ~1.8x at dose of 0.001;
statistically significant ft to ~2.4x at the
higher 3 doses. Implants were
supplemented with recombinant FGF-2;
inorganic arsenic-enhanced neovascular-
ization did not occur without FGF-2.
Data suggest that inorganic arsenic
potentiates, but does not directly cause,
neovascularization in Matrigel implants.


Reference




Kinoshita et
al., 2007b








Simeonova et
al., 2000



Simeonova et
al., 2000




Simeonova et
al., 2000






Soucy et al.,
2003







Soucy et al.,
2005







C-42   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Matrigel
implants/mouse
(C57BL/6NCr,
male)





Tumors that
developed from
B16-F10(GFP)
melanoma tumor
cells/mice
(NCr nu/nu,
male)


Tumors that
developed from
B16-F10(GFP)
melanoma tumor
cells/mice
(NCr nu/nu,
male)

Blood
vessels/chicken
(Leghorn,
chorioallantoic
membranes of
10-day-old
chicken embryos)



Skin/mouse
(homozygous,
strain Tg. AC,
female)





Arsenic
Species
AsmSA








AsmSA








AsmSA







AsmSA









AsmSA







Dose in
Elemental Asa
(in Units Stated)
0.05, 0.25, 0.5
ppm (DW)







10, 50, 200 ppb
(DW)
Note in ppb!






10, 50, 200 ppb
(DW)
Note in ppb!





0.33, 10 uM









200 ppm (DW)







Duration
of
Treatment
5, 10, 20
wk







8wk








8wk







48 hr









10 wk








LOELb

0.05 ppm
for each
duration






10 ppb








10 ppb







0.33 uM









200 ppm








Results

Blood vessel number determined in
Matrigel implants surgically inserted
during last 2 wk of inorganic arsenic
treatment: at 5 wk: ft to ~2.6x, ~4.4x,
and ~5.5x at the 3 doses in ascending
order. For each longer duration
treatment, there was still a strong ft at
dose of 0.05 but a somewhat diminished
ft at 2 higher doses.
Tumor volume and tumor growth rate,
after implantation of tumor cells (into
external surface at the base of right ear)
5 wk after inorganic arsenic treatment
began:
Volume: 10, ~1.4x (NSE); 50, ~2.2x;
200, ~3.0x.
Rate: 10, ~1.9x (NSE); 50, ~2.2x; 200,
~3.2x.
Mean no. of lung metastases/lobe, after
implantation of tumor cells (into external
surface at the base of right ear) 5 wk
after inorganic arsenic treatment began:
10, ~1.6x; 50, ~2.0x; 200, ~2.0x
(statistically significant at 10 and 200);
the metastases were significantly larger
at the 2 lower doses.
CAM assay to determine vascularity
(i.e., blood vessel density): ft to ~1.8x at
0.33 but big U at dose of 10. At dose of
0.33, co-treatment with YC-1 or SU5416
(inhibitors of HIF and VEGF receptor-2
kinase) eliminated inorganic arsenic
effect. 10 uM inorganic arsenic + YC-1
caused no change from control, but
inorganic arsenic alone, or in addition to
SU5416, resulted in U to ~0.28x.
By 10 weeks the skin showed
hyperkeratosis as well as ft in numbers
of proliferating cells. A kinetic study
with samples at weekly intervals
demonstrated ft in number of BrdU-
positive nuclei in skin after 4 weeks and
number remained elevated through 10
weeks.


Reference


Soucy et al.,
2005







Kamatetal.,
2005







Kamatetal.,
2005






Kamatetal.,
2005







Germolec et
al., 1998





Chromosomal Aberrations and/or Genetic Instability
Bone marrow/rat
(Rattus
norvegicus,
Charles foster
strain)

Asvas
disodiu
m
hydroge
n
arsenate
* 4.0 mg As/kg bw
(unspecified route
of administration)



15, 21 days





4.0 mg/kg





Chromosomal analysis of Giemsa-
stained cells, with few details provided:
induction of gross CAs for both periods
of treatment; induction of hyperploidy
detected as aneuploids for longer
treatment.

Dattaetal.,
1986



C-43   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Bone
marrow/mouse
(albino Swiss,
male)









Bone
marrow/mouse
(C57BL/6J/Han,
female)



Arsenic
Species
AsmSA












Asvas
Na2HAs
Cv
7H2O


Dose in
Elemental Asa
(in Units Stated)
* 1.44 mg/kg x 4,
5, and 6 times at
weekly intervals,
(gavage)









50, 200, 500
ppb (DW)
Note in ppb!



Duration
of
Treatment
Single dose
each week











3,6, 12
months





LOELb

1.44 x 4












None






Results

Significant ft in CA and probably also in
polyploidy after 4, 5, and 6 gavage
treatments. CA frequencies were
significantly higher than control in all 3
comparisons at 2.5x, 2.7x, and 4.4x,
respectively. Similar experiments with 7
and 8 exposures killed the mice. Daily
treatments by gavage with a black tea
infusion for one week before every
inorganic arsenic treatment caused a
significant reduction in the frequency of
CAs after 4 and 6 inorganic arsenic
treatments.
Half of the mice were maintained on a
low-Se diet. Mouse erythrocyte MN
test: inorganic arsenic caused no
induction of MN in PCEs and no change
in the PCE:NCE ratio at any dose at any
interval, with or without the low-Se diet.


Reference



Patra et al.,
2005









Palus etal.,
2006



Co-carcinogenesis
Skin/mouse
(Hairless mice,
strain Skhl)















Skin/mouse
(Hairless
CrL:SKl-hrBD,
female,
weanling)
Starting 3 wk
after inorganic
arsenic treatment
began; mice were
irradiated thrice
weekly with UV
at a dose of 1.0
kJ/m2
(i.e., -30% of
MED)
AsmSA

















AsmSA














*0.7, 1.4,2.9,5.8
ppm (DW)
















* 2.9 ppm (DW)














161 days
beginning
at 21 days
of age














29 wk














0.7 ppm

















2.9 ppm














Starting 21 days after the As111 treatments
began, mice had their dorsal skin
exposed to 1.0 kJ/m2 of solar spectrum
UV (a low nonerythemic dose) 3 times
weekly. Untreated control mice and
inorganic arsenic-treated mice
unexposed to UV developed no skin
tumors. Of mice exposed to UV, skin
tumor yields per mouse at the different
doses of inorganic arsenic were as
follows: 0, 2.40; 0.7, 5.40; 1.4, 7.21; 2.9,
11. 10; 5.8, 6.80. More than 95% of
tumors were squamous cell carcinomas.
Mice in all dose groups exposed to UV
and inorganic arsenic showed a 2.5-3x ft
in epidermal hyperplasia above that
caused by UV alone, with the highest
point estimate at 0.7.
Immunohistological determination of
oxidative DNA damage shown by
staining of 8-oxo-dG:
Control: no effect.
UV alone: very slight ft.
inorganic arsenic alone at 5.8 ppm
(earlier experiment): ft.
inorganic arsenic + UV (this
experiment): huge ft.
Co-treatment with vitamin E or p-XSC:
ft (i.e., a significant reduction in
inorganic arsenic + UV effect).
Above effects roughly paralleled those
for SCC induction, except that no tumors
were caused by arsenic alone.




Burns et al.,
2004












Uddin et al.,
2005













C-44   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species

Arsenic
Species
Dose in
Elemental Asa
(in Units Stated)
Duration
of
Treatment

LOELb


Results



Reference
Co-mutagenesis
Skin/mouse
(Fi offspring
from cross of
FVB/N carrying
Gil FLAP
transgene x
C57BL/6J, both
sexes)











AsmSA


















*5.8ppm(DW)


















10 wk


















None,
but 5.8 ppm
if
co-
treatment
withB[a]P













Frequencies of induction of PLAP+ cells
(result from frameshift mutations) in (A)
untreated control, (B) group with
inorganic arsenic treatment alone, (C)
group with skin painting with B [a]P 5
days/week during weeks 3-10 after start
of experiment, and (D) group with both
BandC:
A = x; B, ~1.9x, was a NSE; C, ~3.2x,
was a NSE; D, ~10.7x. Also,
significantly more of the individual
mutations arose as clusters in group D,
which suggests that more mutations
arose in stem cells. This assay in
bladder, spleen, lung, kidney, and liver
yielded no obvious effect. Oxidation of
guanosines in poly G tracts of G:C base
pairs is thought to be one cause of these
frameshift mutations.




Fischer etal.,
2005













Cytotoxicity
Bladder/rat
(F344, female)









Urothelium/rat
(F344, female)







DMAV










DMAV
as
sodium
cacodyl
ate-
trihydrat
e


* 54.3 ppm (food)
(assumes MW of
chemical used was
138.0)







*0.35, 1.4, 14,35
ppm (DW)







2wk










28 days








54.3 ppm










14 ppm








Evidence of cytotoxicity by SEM as
frequency of class-5 bladders, which
showed necrosis and piling up of
rounded urothelial cells: 6 of 10 rats,
compared to 0 of 10 in control. In group
with co-treatment with DMPS (a
chelator of trivalent arsenicals), only 1 in
10 rats had a class-1 bladder. In another
experiment with the same dose for 26
weeks, none of the rats had class-5
bladders.
By light and transmission electron
microscopy, no alterations were detected
at lower 2 doses. At higher 2 doses,
urothelial cells showed signs of
swelling, appearance of cytoplasmic
vacuoles and a decreased number of
mitochondria (all being signs of
cytotoxicity), with a positive dose-
response.



Cohen et al.,
2002








Sen etal.,
2005





DNA Damage
Liver/rat
(Fischer 344,
male)
(they used
normal-appearing
tissue)
TMAVO





*27.5, 110.2 ppm
(DW) Estimated
total intakes: 351
and 1363 mg
As/rat


104 wk




110.2 ppm





8-OHdG formation assessed by HPLC:
ft to ~1.22x; point estimate was also ft at
lower dose, but the SE for it was large.




Shen et al.,
2003



C-45   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Lung/mice
(C57BL/6J
Oggl+/+wtmice,
both sexes, 14
weeks old at start
of treatment)

Lung/mice
(C57BL/6J Oggl
'" knockout mice,
both sexes, 14
weeks old at start
of treatment)
Peripheral blood
leukocytes/mous
e
(C57BL/6J/Han,
female)














Lung/mouse
(ddY, male)






Liver/rat
(Fisher 344,
male)


Bladder/rat
(Fisher 344,
male)



Arsenic
Species
DMAV












Asvas
Na2HAs
04-
7H2O















DMAV
assumed
to be
dimethy
lar-sinic
acid


MMAV

DMAV

TMAVO
MMAV

DMAV

TMAVO
Dose in
Elemental Asa
(in Units Stated)
* 115.3ppm(DW)












50, 200, 500 ppb
(DW)
Note in ppb!
















*217ppm(DW)







* 121ppm(DW)c

* 109 ppm (DW)C

* 110ppm(DW)c
* 121ppm(DW)c

* 109 ppm (DW)C

* 110ppm(DW)c
Duration
of
Treatment
72 weeks












3,6, 12
months

















4wk







5, 10, 15,
and 20 days
for all


20 days
for all




LOELb

None



115.3








50 ppb


















2 17 ppm







None

None

110 ppm
None

109 ppm

None

Results

8-OHdG formation assessed by HPLC, x
= level of wt control:
wt with inorganic arsenic treatment: ft to
~1.6x(NSE);
knockout Oggl"'" without inorganic
arsenic: ft to ~7.8x;
knockout Oggl"7" with inorganic arsenic
treatment: ft to -13. Ix.





Half of the mice were maintained on a
low-Se diet.
Alkaline SCGE (comet assay) was used
to detect DNA fragmentation (SSBs) and
alkaline labile sites as well as oxidative
DNA base damage identified by using
FPG and Enm enzymes. The only
significant inorganic arsenic effects were
seen at 3 months, perhaps because water
consumption (and thus inorganic arsenic
consumption) was lower at the last 2
times sampled. An ft in DNA
fragmentation was observed only in the
mice with the low-Se diet, but there was
no positive dose-response. An ft in
oxidative DNA damage was observed
only in the mice with the normal-Se diet,
and again there was no positive dose-
response.
8-oxo-dG levels: ft to 1.42x;
subcutaneous injection of 10 mg/kg
4NQO just before 4-wk DMA treatment
had no significant effect on this level; it
was 1.3 8x. Use of feed containing
0.05% of the antioxidant EGCG was
tested. 8-oxo-dG level in the (4NQO +
DMA + EGCG) group was only 1.09x.
8-OHdG formation assessed by HPLC:
TMAVO: ft on day 15 to ~1.5x and on
day20to~1.82x.


8-OHdG formation assessed by HPLC:
DMAv:ftto~1.62x.





Reference




Kinoshita et
al., 2007b












Palus etal.,
2006














Mizoi et al.,
2005






Kinoshita et
al., 2007a



Kinoshita et
al., 2007a

C-46   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species

Arsenic
Species
Dose in
Elemental Asa
(in Units Stated)
Duration
of
Treatment

LOELb


Results



Reference
Effects Related to Oxidative Stress (ROS)
Brain, liver,
RBCs/rat
(Wistar, male)




Liver/rat
(Fisher 344,
male)









Kidney and
liver/rat
(Wistar, female)








Kidney and
liver/rat
(Wistar, male)









As111 as
SA





MMAV

DMAV

TMAVO








Asm
ATO









Asm
ATO









* 57.7 ppm (DW)






* 121ppm(DW)c

* 109 ppm (DW)C

* 110ppm(DW)c







* 30.3 mg/kg,
15 times (gavage)









* 30.3 mg/kg,
15 times (gavage)










12 wk






5, 10, 15,
and 20 days
for all









Every other
day for 30
days








Every other
day for 30
days









57.7 ppm






None

109 ppm

110 ppm







30.3 mg/kg
x!5









30.3 mg/kg
x 15










In liver and brain: U GSH levels; ft
GSSG levels; ft MDA levels.
InRBCs: U GSH levels; U ALAD levels;
ft MDA levels.
Some, but not all, of these effects were
mitigated by oral post-treatment with
NAC and/or DMSA.
Oxidative stress in microsomes shown
by elevation of total cytochrome P450
content and/or by ft in hydroxyl radical
levels:
DMAV for P450: ft on day 10 only to
DMAV for OH radicals: ft on day 15
only to ~1.18x.
TMAVO for P450: ft on days 10-20,
maximum ft on day 15 to ~1.25x.
TMAVO for OH radicals: ft on days 15
and 20, maximum ft on day 20 to
~1.33x.
Kidney: MDA level ft to 3.8x; GSH
level U to 0.78x;
GSSG level ft to 7.5x; GST activity U to
0.44x.
Liver: MDA level ft to 2.0x;
GSSG level ft to 5.3x; GST activity U to
0.52x.
Co-treatment with L-ascorbate reduced
the size of the inorganic arsenic-induced
effects (either ft or U) on all 4 endpoints
in kidneys and on all but GSH in livers.
Kidney: MDA level ft to 3.4x; GSH
level U to 0.62x;
GSSG level ft to 8.5x; GST activity U to
0.49x.
Liver: MDA level ft to 2.7x; GSH level
U to 0.82x;
GSSG level ft to 5.9x; GST activity U to
0.49x.
Co-treatment with L-ascorbate reduced
the size of the inorganic arsenic -induced
effects (either ft or U) on all 4 endpoints
in kidneys and on all but GSH in livers.


Flora, 1999







Kinoshita et
al., 2007a










Sohini and
Rana, 2007









Sohini and
Rana, 2007







C-47   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Blood, kidney,
liver/mouse
(albino Swiss,
male)








Liver/rat
(Wistar, male)











Liver/mouse
(BALB/c, male)









Liver/mouse
(BALB/c, male)












Arsenic
Species
AsmSA












AsmSA











AsmSA










AsmSA












Dose in
Elemental Asa
(in Units Stated)
* 57.7 ppm (DW)











*0.03, 1.4,2.9
ppm (DW)











50, 100, 150
ug/mouse/day for
6 days/week
(gavage)







50, 100, 150
ug/mouse/day for
6 days/week
(gavage)









Duration
of
Treatment
8wk











60 days












3 months










6 months













LOELb

57.7











Various












50 for ft
None
None
None
50 for ft
50 for ft





None
100 for ft
100 for U
100 for ft
100 for U
100 for U








Results

Blood: ALAD activity U to 0.32x; GSH
level U to 0.78x; ROS level ft to 2.82x.
Kidney: SOD activity U to 0.38x; CAT
activity U to 0.34x; TEARS level ft to
1.17x; GSH level U to ~0.39x; GSSG
level ft to ~2.5x;
GPx activity U 0.94x (NSE).
Liver: SOD activity U to 0.33 x; CAT
activity U to 0.54x; TEARS level ft to
1.25x; GSH level U to ~0.44x; GSSG
level ft to -3.1 x; GPx activity U 0.76x
(NSE); G-6-P activity U to ~0.73x.
Cytochrome P450 activity: ft to 1.41x
and l.Slx at 1.4 and 2.9, respectively.
MDA level: ft to 1.39x and 1.55x at 1.4
and 2.9, respectively.
GSH level: U to 0.59x, 0.47x, and 0.42x
at 3 doses in ascending order.
SOD activity: U to 0.76x, 0.60x, and
0.55x at 3 doses in ascending order.
U in activities of CAT, GPx, GR, G-6-P,
and GST, respectively, to 0.90x, 0.75x,
0.50x, 0.76x, and 0.6 Ix at 1.4 ppm and
to 0.54x, 0.66x, 0.42x, 0.64x, and 0.45x
at 2. 9 ppm.
Changes in various components of
antioxidant defense system:
GSH level: 50, 1.14x; 100, 1.17x; 150,
1.25x.
MDA level: NSE at any dose.
PSH level: NSE at any dose.
PC level: NSE at any dose.
GPx activity: 50, 1.12x; 100, 1.15x; 150,
1.24x.
CAT activity: 50, 1.06x; 100, l.OSx;
150, l.lOx.
Changes in various components of
antioxidant defense system:
GSH level: NSE at any dose.
MDA level: 50, NSE; 100, 1.39x; 150,
1.44x.
PSH level: 50, NSE; 100, O.Slx; 150,
0.75x.
PC level: 50, NSE; 100, 1.16x; 150,
1.30x.
GPx activity: 50, NSE; 100, 0.91x; 150,
0.90x.
CAT activity: 50, NSE; 100, 0.94x; 150,
0.92x.


Reference



Mittal and
Flora, 2006











Bashiretal.,
2006a











Das etal.,
2005









Das etal.,
2005







C-48   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Liver/mouse
(BALB/c, male)












Liver/mouse
(BALB/c, male)












Blood/rat
(Wistar, male)




Liver/rat
(Wistar, male)




Blood, kidney,
liver/rat
(Wistar, male)








Arsenic
Species
AsmSA













As111 SA














AsmSA





AsmSA




AsmSA









Dose in
Elemental Asa
(in Units Stated)
50, 100, 150
ug/mouse/day for
6 days/week
(gavage)










50, 100, 150
ug/mouse/day for
6 days/week
(gavage)










* 57.7 ppm (DW)





* 57.7 ppm (DW)






*1.15 mg/kg/day
(gavage)







Duration
of
Treatment
9 months













12 months













6 weeks





6 weeks





3 weeks










LOELb

50 for U
50 for ft
50 for U
50 for ft
50 for U
50 for U








50 for U
50 for ft
50 for U
50 for ft
50 for U
50 for U








57.7 ppm





57.7 ppm





Various










Results

Changes in various components of
antioxidant defense system:
GSH level: 50, O.SOx; 100, 0.77x; 150,
0.66x.
MDA level: 50, 1.97x; 100, 2.06x; 150,
2.16x.
PSH level' 50 0 80x' 100 0 75x' 150
0.71x.
PC level: 50, 1.64x; 100, 1.78x; 150,
1.94x.
GPx activity: 50, 0.95x; 100, 0.91x; 150,
0.87x.
CAT activity: 50, 0.95x; 100, 0.93x;
150, 0.92x.
Changes in various components of
antioxidant defense system:
GSH level: 50, 0.76x; 100, 0.72x; 150,
0.63x.
MDA level: 50, 2.20x; 100, 3.03x; 150,
3.97x.
PSH level: 50, 0.73x; 100, 0.63x; 150,
0.56x.
PC level: 50, 2.09x; 100, 2.91x; 150,
3.46x.
GPx activity: 50, 0.87x; 100, 0.84x; 150,
0.75x.
CAT activity: 50, 0.93x; 100, 0.92x;
150, 0.88x.
Effects on levels of biochemical
variables indicative of disturbances in
the heme synthesis pathway and
oxidative stress: ALAD U to 0.12x; GSH
U to 0.73x; RBC ROS ft to 1.35x; GPx
showed NSE.
Effects on levels of biochemical
variables indicative of oxidative stress:
GSH U to 0.69x; GSSG ft to 1.41x;
TEARS ft to 1.16x; catalase showed
NSE. There was NSE for any of these
parameters in the kidney.
ALAD activity: blood, 0.45x.
CAT activity: kidney, 1.12x (NSE);
liver, 1.1 6x.
GSH level: blood and kidney, NSE;
liver, 0.79x.
TEARS level: kidney, NSE; liver,
1.28x;.
Co-treatment with NAC (i.p. injection)
and/or zinc sulfate (oral) reduced some
effects, especially when used together.


Reference



Das etal.,
2005












Das etal.,
2005










Kaliaetal.,
2007




Kaliaetal.,
2007





Modi etal.,
2006






C-49   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Brain/rat
(albino Wistar,
male)








Brain/rat
(albino Wistar,
male)













Brain/rat
(albino Wistar,
male)












Kidney, liver,
RBCs/rat
(albino Wistar,
male)









Arsenic
Species

AsmSA










AsmSA















AsmSA














AsmSA










Dose in
Elemental Asa
(in Units Stated)
* 57.7 ppm (DW)










* 57.7 ppm (DW)















* 57.7 ppm (DW)














*5.8ppm(DW)











Duration
of
Treatment
60 days










60 days















60 days














12 weeks












LOELb

57.7 ppm










57.7 ppm















57.7 ppm














5.8 ppm,
but for only
some
effects









Results

Effects on levels of chemicals indicative
of oxidative stress in 5 regions of the
brain (hippocampus, cortex, striatum,
hypothalamus, and cerebellum): MDA ft
to from 1.64x to 2.2 Ix; GSH U to from
0.43x to 0.58x; GPx U to from 0.77x to
O.Slx; GR U to from 0.73x to 0.78x;
G6PDH U to from 0.70x to 0.84x.
Simultaneous treatment with DL-a-
lipoic acid markedly reduced all of these
effects.
Effects on levels of chemicals indicative
of oxidative stress in 5 regions of the
brain (hippocampus, cortex, striatum,
hypothalamus, and cerebellum):
ROS based on DCF assay ft to from
1.62x to 2.18x; total SOD U to from
0.56x to 0.77x; Mn SOD U to from
0.36x to 0.55x; Cu/Zn SOD U to from
0.53x to 0.62x; CAT U to from 0.67x to
O..80x.
Simultaneous treatment with DL-a-
lipoic acid markedly reduced all of these
effects.
(This is the same experiment as in the
previous row; findings not already listed
in that row are listed here.)
Measures of protein oxidation:
ft in protein carbonyl level: cerebellum,
1.23x; cortex, 1.32x; hippocampus,
1.48x; hypothalamus, 1.25x; striatum,
1.49x;
U in membrane protein sulfhydryl
content: cerebellum, 0.71x; cortex,
0.55x; hippocampus, 0.50x;
hypothalamus, 0.79x; striatum, 0.6 Ix;
essentially the same regional pattern of
inorganic arsenic -induced loss occurred
with total protein-bound sulfhydryls.
Co-treatment with DL-a-lipoic acid
mostly or completely abolished all of the
above effects.
MDA level: ft in kidney to -2. Ix, in
liver to ~1.7x, and in RBCs to ~1.4x.
CAT activity: U in kidney to -0.73 x, in
liver to -0.9 Ix (NSE), and in RBCs to
-0.78.
SOD activities were measured but with
NSE.
Co-treatment with cysteine, methionine,
AA, or thiamine usually decreased tissue
arsenic concentrations (especially in
kidney and liver) and blocked oxidative
damage to variable degrees.


Reference


Shila et al.,
2005a











Shila et al.,
2005b













Samuel etal.,
2005













Nandi et al.,
2005







C-50   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Kidney/rat
(albino Wistar,
male)






Liver/rat
(albino Wistar,
male)





RBCs/rat
(albino Wistar,
male)




Liver and
kidney/rat
(albino Wistar,
male)





Liver and
kidney/rat
(albino Wistar,
male)
















Arsenic
Species

AsmSA








AsmSA







AsmSA






AsmSA








AsmSA

















Dose in
Elemental Asa
(in Units Stated)
*5.8ppm(DW)








*5.8ppm(DW)







*5.8ppm(DW)






* 57.7 ppm (DW)








* 57.7 ppm (DW)


















Duration
of
Treatment
4, 8, 12
weeks







4, 8, 12
weeks






4, 8, 12
weeks





30 days








30 days



















LOELb

Various for
ft and U







Various for
ft and U






Various for
ft and U





57.7 ppm








57.7 ppm



















Results

MDA level: ft at 4 wk to ~1.27x (NSE),
at 8 wk to ~1.54x, and at 12 wk to
-2.1 Ix.
CAT activity: ft at 4 wk to ~1.72x, at 8
wk to ~1.18x (NSE) but U at 12 wk to
~0.75x.
SOD activity: ft at 4 wk to ~1.84x, at 8
wk to ~1.23x, but U at 12 wk to 0.91x
(NSE).
MDA level: ft at 4 wk to ~1.07x (NSE),
at 8 wk to ~1.46x, and at 12 wk to
~1.49x.
CAT activity: ft at 4 wk to ~1. 19x
(NSE), at 8 wk to ~1.52x but U at 12 wk
to -0.9 Ix (NSE).
SOD activity: ft at 4 wk to ~1.52x, at 8
wk to ~1.16x, but NSE at 12 wk.
MDA level: ft at 4 wk to ~1. 13x (NSE),
at 8 wk to ~1.28x, and at 12 wk to
~1.41x.
CAT activity: ft at 4 wk to ~1.36x, NSE
at 8 wk, and U at 12 wk to -0.7 Ix.
SOD activity: ft at 4 wk to ~1.81x, at 8
wk to ~1.59x, but NSE at 12 wk.
Level of ROS determined by DCFH
assay:
ft in liver to ~3.6x and in kidney to
~3.5x.
Level of MDA released per mg protein:
ft in liver to ~1.5x and in kidney to
Co-treatment with both DL-a-lipoic acid
and DMSA markedly reduced all of
these effects.
Activities of antioxidant enzymes:
U of SOD in liver to ~0.51x and in
kidney to ~0.55x.
U of CAT in liver to ~0.59x and in
kidney to ~0.58x.
U of GPx in liver to ~0.53x and in
kidney to ~0.56x.
Levels of non-enzymatic antioxidants:
U of GSH in liver to ~0.56x and in
kidney to ~0.67x.
U of AA in liver to ~0.48x and in kidney
to ~0.50x.
U of a-Toc in liver to ~0.49x and in
kidney to ~0.58x.
U of total sulfhydryls in liver to ~0.53x
and in kidney to ~0.59x.
Co-treatment with both DL-a-lipoic acid
and DMSA markedly reduced all of
these effects.


Reference



Nandi et al.,
2006






Nandi et al.,
2006






Nandi et al.,
2006





Kokilavani et
al., 2005










Kokilavani et
al., 2005












C-51  DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Blood (whole),
brain, kidney,
liver/mice
(Swiss albino,
male)









Blood (whole),
brain, kidney,
liver/rat
(Wistar, male)













Blood (whole),
brain/rat
(Wistar, male)

















Arsenic
Species

AsmSA













AsmSA
















AsmSA

















Dose in
Elemental Asa
(in Units Stated)
* 14.4 ppm (DW)













* 11.5ppm(DW)
















* 57.7 ppm (DW)


















Duration
of
Treatment
3 months













4wk
















10 wk



















LOELb

14.4 ppm













11.5 ppm
















57.7 ppm



















Results

Whole blood: U of ALAD activity to
0.37x; U of GSH level to 0.93x.
Brain: ft in TEARS level to ~2.2x; U in
GSH/GSSG ratio to ~0.96x.
Kidney: ft in TEARS level to 1.65x.
Liver: ft in TEARS level to 1.21x; U in
SOD activity to 0.76x; U in CAT activity
to 0.89x; U in GSH/GSSG ratio to 0.89x.
Post-treatments with 3 different extracts
ofHippophae rhamnoides L. (thought to
have antioxidant properties) showed
various levels of effectiveness in
reducing some of the above effects in all
but the kidney.
Whole blood: U of ALAD activity to
0.24x; U of GSH level to 0.86x; ft of
ZPP level to 1.30x.
Brain: ft in TEARS level to 1.89x; U in
GSH level to 0.85x; NSE on GSSG
level; U in SOD activity to 0.75x; U in
CAT activity to 0.75x.
Kidney: ft in TEARS level to 1.39x; U in
GSH level to 0.55x; ft in GSSG level to
1.59x.
Liver: ft in TEARS level to 1.96x; U in
GSH level to 0.6 Ix; ft in GSSG level to
2.00x; oral co-treatment with Centella
asiatica (thought to have antioxidant
properties) showed various levels of
effectiveness in reducing some of the
above effects.
Whole blood: ft of ROS level to 2.63x; U
of ALAD activity to 0.46x; U of GSH
level to 0.85x; U of Hb as grams/dL to
0.79x.
Brain: ft of ROS level to 4.03x; ft in
TEARS level to 1.50x; U in GSH level
to 0.82x; U in SOD activity to 0.92x
(NSE); U of ALAD activity to 0.58x; ft
of ALAS activity to 1.21 x; U of GPx
activity to 0.84x (NSE); ft of GST
activity to l.OSx (NSE); "considerable"
but unqualified ft in DNA
fragmentation (single-strand breaks) was
detected by polyacrylamide gel
electrophoresis.
Postreatment with the thiol chelating
agents DMSA, DMPS, and MiADMSA
showed various levels of effectiveness in
reducing some of the above effects.


Reference



Gupta and
Flora, 2005













Gupta and
Flora, 2006















Flora et al.,
2005













C-52   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Liver/mouse
(BALB/c, male)












Lung/mouse
(ddY, male)













Liver/rat
(Wistar, male)


















Arsenic
Species
Unspeci
fied
arsenica
l,but
from
discuss!
on
assumed
to be
AsmSA



DMAV
















AsmSA

















Dose in
Elemental Asa
(in Units Stated)
* 1.8ppm(DW)












*217.2ppm(DW)















* 3.6, 6.1,7.3
mg/kg
(gavage, with
animals being
killed 24 hr later
for sample
collection)












Duration
of
Treatment
3,6,9,12,
15 months











2, 4, 8, 15,
25 wk














One dose



















LOELb

1.8at>9
months for
MDA

1.8at>6
months for
GSH





217.2ppm
at 8 wk or
longer














Various



















Results

MDA cone: ft to ~1.7x at 9, ~1.9x at 12,
and~2.2xat 15.
GSH content: U to ~0.84x at 6, ~0.78x at
9, ~0.67x at 12, and ~0.58x at 15. U in
activities were also noted for G6PDH,
GPx, and
plasma membrane Na+/K+ ATPase at 6
months, for CAT at 9 months, and for
GST and GR at 12 and 15 months. It
seems likely that the activities remained
lower at later times than when each U
was noted, but that was not stated.
Immunohistochemical analysis of 4HNE
adducts showed that lipid peroxidation
occurred in 48.8%, 72.9%, and 77.6% of
terminal bronchiolar Clara cells by 8, 15,
and 25 weeks, respectively. (None
before that.) The modified proteins were
specifically in the secretory granules of
those cells. 8-OHdG adducts (showing
oxidative DNA) damage were also
demonstrated in the same cells. Clara
cells are the major target cell for DMA-
induced oxidative stress, and the authors
suggested that lipid peroxidation via the
formation of ROS is involved in
promotion of lung tumor (malignant
adenocarcinoma) formation following
initiation by 4NQO.
Significant dose-related ft in total
arsenic cone at all doses; cone in liver at
highest dose was ~22 times that in brain.

MDA cone: ft to 1.43x at 6.1 and 1.52x
at 7.3.
GSH level: U to 0.57x at 3.6, to 0.41x at
6.1,andto0.39xat7.3.
Total cytochrome P450 activity: ft to
1.46x at 6.1 and 1.54x at 7.3.
SOD level: U to 0.67x at both 6. 1 and
7.3.
CAT activity: U to 0.54x at 6. 1 and
0.49xat7.3.
GPx activity ft to 1. 15x at 3.6, 1.21x at
6.1,andl.27xat7.3.
GST activity: U to 0.72x at 6. 1 and 0.62x
at 7.3.
NSE on either GR or G6PD activity.


Reference
Mazumder,
2005














Anetal.,
2005















Bashir et al.,
2006b













C-53   DRAFT—DO NOT CITE OR QUOTE

-------
 Tissue or Cell
  Type/Species
Arsenic
Species
    Dose in
 Elemental Asa
(in Units Stated)
 Duration
    of
Treatment
LOELb
Results
                                               Reference
    Brain/rat
 (Wistar, male)
AsmSA
  * 3.6, 6.1,7.3
     mg/kg
  (gavage, with
  animals being
killed 24 hr later
   for sample
   collection)
                          One dose
            Various
           Significant ft in total arsenic cone at
                   both higher doses.
          MDA cone: ft to 1.48x at 6.1 and 1.56x
                        at 7.3.
          GSH level: U to 0.79x at 3.6, to 0.60x at
                 6.1,andto0.51xat7.3.
          SOD level: U to 0.73x at 6.1 and 0.70x
                        at 7.3.
            CAT activity: U to 0.58x at 6.1 and
                     0.51xat7.3.
         GPx activity ft to 1.17x at 6.1, and 1.26x
                        at 7.3.
         GST activity: U to 0.7Ix at 6.1 and 0.69x
                        at 7.3.
           NSE on either GR or G6PD activity.
                                                                                                   Bashir et al..
                                                                                                      2006b
   Kidney, rat
 (Wistar, male)
  Asm
  ATO
  * 30.3 mg/kg,
15 times (gavage)
                Every other
                 day for 30
                    days
            30.3 x 15
                GSH content U to ~0.59x.
                   GST activity: NSE.
                       Ranaand
                      Allen, 2006
Gene Mutations
   Skin/mouse
  (Aprt+/~ hybrid
mice of complex
genotype needed
  for assay: see
     paper)
AsmSA
         *5.7ppm(DW)
                   10 wk
             None
            Starting 2 wk after consumption of
           inorganic arsenic-contaminated water
            began, half of the mice were also
            exposed to B[a]P for 8 wk by skin
           painting. Skin was assayed for DAP-
          resistant (DAP1) colonies indicative of
          cells lacking Aprt activity as the result
          of loss of heterozygosity (LOH) at Aprt
           because of malsegregation or mitotic
           recombination in vivo. No  significant
            differences were found because of
         inorganic arsenic and/or B[a]P exposure.
            and thus there was no evidence that
              inorganic arsenic alone, or by
          enhancement of a known mutagen (but
           not one + in this assay), caused such
           genetic changes. Curiously, the point
            estimate for most LOH was in the
           control (45%); it was 38% for B[a]P
           alone, 8% for inorganic arsenic alone,
           and 30% for them together. Because
            there was much variability, these
           seemingly large differences were not
          	statistically  significant.	
                                                                                                   Fischer etal.
                                                                                                      2006
                                                     C-54   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Wing/Drosophila
melanogaster











Arsenic
Species
DMAV











Dose in
Elemental Asa
(in Units Stated)
0.05,0.1,0.25,0.5
mM
(in medium)









Duration
of
Treatment
72 hr












LOELb

0.25 mM,
regarding
total spots










Results

SMART (somatic mutation and
recombination test) wing spot assay:
positive dose-response was found, but
nature of induced mutations was
uncertain. Was earlier shown that
inorganic arsenic is inactive in this
assay. They showed no biomethylation
occurs in larvae or in growth medium.
Results suggest importance of
biomethylation as a determinant of
genotoxicity of arsenic compounds, at
least in Drosophila.


Reference


Rizkietal.,
2006








Hypermethylation of DNA
Lung/mice
(A/J, male)
















Asvas
Na2HAs
Cv
7H2O















* 0.24, 2.4, 24
ppm (DW)
















18 months

















The LOEL was 0.24 ppm. Extent of
hypermethylation of promoter regions of tumor
suppressor genes p\6WK4a and RASSF1A in lung
adenocarcinomas from inorganic arsenic exposed
mice compared to the control, based on methylation-
specific PCR: percentages of methylated promoters
of pl6INK43 in jung tumors of Q Q 24^ 2 4 and 24
ppm dose groups were 11%, 30%, 36%, and 42%,
respectively. Percentages of methylated promoters
of RASSF1A in lung tumors of the same dose groups
were 33%, 70%, 82%, and 89%, respectively.
Reduced expression, or lack of expression, of these 2
genes was correlated with the extent of
hypermethylation. There was constant expression of
these genes in lungs without tumors in both control
and inorganic arsenic-treated mice. They concluded
that epigenetic changes of tumor suppressor genes
are involved in inorganic arsenic-induced lung
carcinogenesis.




Cuietal.,
2006












Hypomethylation of DNA
Liver cells/mouse
(129/SvJ)









AsmSA








45 ppm (DW)









48 wk









45 ppm









There was global DNA
hypomethylation, as shown by 5-
methylcytosine content of DNA and by
using the methyl acceptance assay. In
particular, there was a marked U in
methylation within the ER-a gene
promoter region, which was statistically
significant in 8 of 13 CpG sites. Control
had 28.3% of ER-a sites methylated, but
experimental group had 2.9%.



Chen et al.,
2004b





C-55   DRAFT—DO NOT CITE OR QUOTE

-------

Tissue or Cell
Type/Species
Livers of
newborn
males/mouse
(only pregnant
C3H females
drank the water)









Arsenic
Species
AsmSA













Dose in
Elemental Asa
(in Units Stated)
85 ppm (DW)













Duration
of
Treatment
10 days,
gestation
days
8 to 18











LOELb

85 ppm














Results

Global DNA methylation status was not
significantly altered based on methyl
acceptance assay, which measures
methylation in both quiescent and active
areas of DNA. However, another assay
showed that GC-rich regions globally
were less methylated if they were from
livers of newborn males exposed in
utero to inorganic arsenic. Band
intensity showing the extent of
methylation was 0.20x after Rsal + Mspl
digestion and 0.40x after Rsal + Hpall
digestion. Mspl and Hpall are
methylation sensitive enzymes.


Reference
Xieetal.,
2007












Interference With Hormone Function
Kidney, rat
(Wistar, male)


Asm
ATO
* 30.3 mg/kg,
15 times (gavage)

Every other
day for 30
days
30.3 x 15


T3 and T4 levels in serum:
triodothyronine (T3) ft to ~4.8x;
thyroxine (T4) ft to ~1.7x.
Ranaand
Allen, 2006

Signal Transduction
Fetal
lungs/mouse
(only pregnant
C3H females
drank the water,
female offspring
only)













Adenomas and
adeno-
carcinomas from
lungs of adults
exposed in
wtero/mouse
(only pregnant
C3H females
drank the water,
female offspring
only)
AsmSA



















AsmSA










85 ppm (DW)



















85 ppm (DW)










10 days,
gestation
days
8 to 18
















10 days,
gestation
days
8 to 18







85 ppm



















85 ppm










ft in ER-a transcript (5.3x) and protein
levels; ft in expression of the following
estrogen-related genes: trefoil factor-3
(9.66x), anterior gradient-2 (3.21x); ft in
expression of the following steroid
metabolism genes: 17-p-hydroxysteroid
dehydrogenase type 5 (3.55x) and
aromatase (2.53x). (Expression of ER-a
and the ER-linked genes was unchanged
in male fetal lung as compared to
control.) The insulin growth factor
system was also activated, with
transcripts for
IGF-1, IGF-2, IGF-R1, IGF-R2, IGF-
BP1, and IGF-BP5 all being increased to
1.6-2.5x. Also, there was
overexpression of the following genes
that have been associated with lung
cancer: AFP (6.9x), EGFR (3.2x), L-
myc (1.9x), and metallothionein-1
Based on immunohistochemical
analysis:
intense and widespread ft in nuclear ER-
a expression; in contrast, normal adult
lung and DENA-induced lung
adenocarcinoma showed little evidence
of ER-a expression.









Shen et al.,
2007
















Shen et al.,
2007






C-56   DRAFT—DO NOT CITE OR QUOTE

-------
Tissue or Cell
Type/Species
Arsenic
Species
Dose in
Elemental Asa
(in Units Stated)
Duration
of
Treatment
LOELb
Results
Reference
a When doses were reported in mg arsenic/L or in ppm As, it was assumed that the doses included adjustment for the
amount of arsenic in solution. Because it was sometimes unclear from the papers whether a correction was needed, a "*"
was put front of the doses listed in the table if those doses were corrected to the amount of arsenic in the dose.
b Lowest observed effect level.
0 Estimates were based on the reported concentrations of MMAV, DMAV, and TMAVO in DW of 1.62, 1.45, and 1.47 mM,
respectively, and on their molecular weights (MWs) of 139.969, 137.997, and 136.025 g and on the atomic weight of
arsenic of 74.926 g. The paper stated that the concentrations of all arsenicals were 0.02% (or 200 ppm). For the
arsenicals themselves, the concentrations were actually 226, 200, and 200 ppm, respectively, if based on the MWs just
listed.
  Table C-3.  In vitro studies related to possible MOA of arsenic in the development of cancer



Type of
Cell/Tissue



Arsenic
Species



Concentration(s)
Tested (\iM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in jiM Unless Noted)




Reference
Aberrant Gene or Protein Expression
HaCaT cells




TRL 1215
cells (normal
rat liver)











Hepa-1 cells
(mouse
hepatoma)






As111 SA





As111 SA













As111 SA







0.5, 1.0





0.125,0.250,
0.500












1,3,10,30







20 passages





24 wk











30 min before
4hr
co-treatment
with 1 nM
TCDD





0.5



0.500 for
effects
noted here













1







ft intracellular GSH
quantities.
U keratins 5, 6, 7, 8, 10,
and 17.
Using Atlas Rat cDNA
expression microarrays,
-80 of the 588 genes
assayed were aberrantly
expressed — including
genes related to stress
and DNA damage, signal
transduction modulators
and effectors, apoptosis-
related proteins,
cytokines and cytokine-
related components, and
growth factors and
hormone receptors.

Results of Northern blot
analysis of mRNA: ft
TCDD-inducible levels
ofNqolmRNA;
response was much
higher at 3 and 10, but
decreased markedly at 30
to slightly more than was
present at 1.

Chien et
al., 2004



Chenetal.,
2001












Maier et
al., 2000






                                                  C-57   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Huh? cells










Huh? cells,
transfected for
use in the
DRE-CALUX
bioassay







PARP-1+/+
MEF cells

PARP-I-'-
MEF cells















Arsenic
Species

As111 SA











As111 SA










As111 SA
for both


















Concentration(s)
Tested (nM)

0.5,1,3,5, 10,20











0.5,1,3,5, 10,20










11.5
for both


















Duration of
Treatment

24 hr











24 hr










24 hr
for both


















LOECa
(HM)

3











3








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Following co-treatment
withlOnMTCDD:li
TCDD-inducible level of
CYP1A1 activation to
-45% of level without
inorganic arsenic, then
reached plateau of -18%
at doses of 5-15 (based
onEROD assay);
inorganic arsenic did not
affect CYP1A1
activation by itself .
Following co-treatment
withlOnMTCDD:
U TCDD-inducible
luciferase activity in the
DRE-CALUX bioassay
to -80% of level without
inorganic arsenic,
followed by a dose-
related U to 42% at dose
of 20.
In a microarray gene chip analysis that
analyzed the expression pattern of
more than 34,000 genes, —311 genes
were found to be differentially
expressed among the different groups
(i.e., control versus inorganic arsenic
treatment or in comparisons between
the 2 genotypes). Many of those
genes belonged to the following
groups: responders to stress and
external stimuli, genes related to cell
growth and maintenance, cell death, or
DNA metabolism. While some genes
were markedly up-regulated in both
genotypes (sometimes to widely
different amounts), other genes were
up-regulated for one genotype and
down-regulated for the other, and vice
versa.




Reference

Chao etal.,
2006b










Chao etal.,
2006b










Poonepalli
etal., 2005














C-58   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue







NB4 cells

































PRCCs


HEK293
cells








Arsenic
Species







As111 ATO

































As111 ATO
for both











Concentration(s)
Tested (nM)







0.5

































0.1


1









Duration of
Treatment





6, 12, 24, 48,
and 72 hr for
transcriptome
analysis;
12 and 48 hr
for proteomic
analysis





























10 min,
1, 6, 24 hr
for both










LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
In a microarray and 2-dimensional gel
electrophoresis (with mass
spectrometry) study aimed at
understanding effects of therapies with
ATO alone, retinoic acid alone, and
their combined therapy, the main
findings for ATO were as follows. At
the transcriptome level, ATO affected
regulation of 487 genes, many of
which were probably related to
essential
aspects of cell-activity
control such as induction of
differentiation antigens, modulation of
apoptosis regulators, and regulation of
genes involved in cell-cycle and
growth control. Other groups of
affected genes included those involved
with protein degradation, cell defense,
stress response, protein modification
and synthesis, and a group of 5 down-
regulated HLA-class I genes. At the
proteome
level, ATO affected 982
protein spots, and there was often a
time-dependent pattern of regulation,
with much
hr than at
lower protein levels at 48
12 hr after treatment. A
group of enzymes involved in
biochemical metabolism was found to
be significantly down-regulated, and
there was a strong reduction of
cytoskeleton proteins, implying a
considerable reorganization of the cell
nucleus and cytoplasmic structures.
By comparison with relatively minor
changes at many of the corresponding
genes at the transcriptome level, the
significant changes found at the
proteomic level suggest that ATO
particularly enhances mechanisms of
post-transcriptional/translational

0. 1 at 6 hr

Iat6hr







modification.
HMOX1 gene
expression (mRNA
levels measured by
quantitative PCR):
In PRCCs: NSE at 10
min or 1 hr; ~2.3x at 6
hr, ~2.8xat24hr.
HEK293:NSEatlOmin
or 1 hr; ~40x at 6 hr,
~54x at 24 hr.




Reference






Zheng et
al., 2005


































Sasaki et
al., 2007







C-59   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
PRCCs


HEK293
cells








PRCCs

HEK293
cells












HCT15 cells

HeLa cells

PLC/PR/5
cells

Chang cells








Arsenic
Species
As111 ATO
for both











As111 ATO
for both
















As111 SA
for all












Concentration(s)
Tested (nM)
0.1,0.5,2
for both











0.1
for both














278.33, the LC50

200.33, the LC50

376.66, the LC50

328.33, the LC50









Duration of
Treatment
24 hr
for both











10 min,
1, 6, 24 hr
for both
















24 hr
for all











LOECa
(HM)
0.1


0.5









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
HMOX1 gene
expression (mRNA
levels measured by
quantitative PCR):
In PRCCs: 0.1, 2.2x, 0.5,
11.7x;2, 33 5X
InHEK293:0.1, 1.2x,
0.5, 8.3x; 2, 224.9x.
Western blot analysis for
heme oxygenase 1
protein for dose of 1 for
24 hr: Huge ft in PRCCs
andbigflinHEK293.
Microarray analysis identified 73
genes whose expression changed in
both types of cells, and for many
expression increased in a time-
dependent manner. These included
HMOX1, Bax (involved in induction
of apoptosis), and genes involved in
many other biological processes
including intracellular protein
transport, signal transduction,
differentiation, GSH metabolism, and
protein complex assembly among
others. Data were presented that
suggest that heme oxygenase 1 protein
confers a cytoprotective effect against
inorganic arsenic treatment.
278.33

200.33

376.66

328.33







Western blot assay to
determine eIF4E protein
levels:
for all cell lines, there
was a reduction in the
protein level to roughly
50%-60%ofthe
corresponding control
level. There was also a
statistically significant,
but smaller, U after 16 hr
for all lines.




Reference
Sasaki et
al., 2007











Sasaki et
al., 2007
















Othumpan
gat et al.,
2005








C-60   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
HCT15 cells

HeLa cells

PLC/PR/5
cells

Chang cells

















HeLa cells





HeLa cells,
HCT15 cells,
CHO-K1 cells





TR9-7 cells
that were
released from
being mostly
synchronized
in G2 (using
Hoechst
33342)
shortly before
inorganic
arsenic
treatment
began




Arsenic
Species
As111 SA
for all























As111 SA





As111 SA
for all









As111 SA













Concentration(s)
Tested (nM)
278.33, the LC50

200.33, the LC50

376.66, the LC50

328.33, the LC50


















200





Various










5













Duration of
Treatment
24 hr
for all























24 hr







LOECa
(HM)
278.33

200.33

376.66

None


















200




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Quantitative real-time
PCR to determine eIF4E
mRNA levels: there was
a statistically significant
U only in lines HCT 15
and HeLa. Actual data
on gene expression, in
arbitrary units:
HCT15: no inorganic
arsenic, 0.099, with
inorganic arsenic, 0.049.
HeLa: no inorganic
arsenic, 0.041, with
inorganic arsenic, 0.029.
PLC/PR/5: no inorganic
arsenic, 0.051, with
inorganic arsenic, 0.028.
Chang: no inorganic
arsenic, 0.018, with
inorganic arsenic, 0.019.
(Judging from their SEs,
the result for PLC/PR/5
must have been of
borderline significance.)
Western blot assay to
determine protein levels:
Big U in cyclin Dl.
ft in cellular levels of
ubiquitin and in the
process of ubiquitination.
Additional experiments involving a genetically modified
cell line, an siRNA that targeted expression of eIF4E,
and proteasome inhibitors suggested (1) that the changes
seen in eIF4E protein levels played a role in the
observed cytotoxicity, (2) and that the inhibition of
cyclin D 1 is mediated through the inhibition of eIF4E,
and (3) that the inorganic arsenic stimulated
ubiquitination and the resulting proteolysis play an
important role in reducing eIF4E protein levels.



3-24 hr










Conclusions based on determining
protein levels using Western blot
assays until 24 hr of inorganic arsenic
treatment in cells made p53(+) or p53(-)
by controlling tetracycline levels: big
ft in ID 1, but it occurred only when
there was p53 protein present, arsenic
p53 protein level decreased, ID1
protein level decreased. The general
finding was confirmed by microarray
analysis. Work by others showed that
ID1 protects against apoptosis through
activation of the
NF-KB signaling pathway.




Reference

Othumpan
gat et al.,
2005





















Othumpan
gat et al.,
2005



Othumpan
gat et al.,
2005







McNeely
et al., 2006










C-61  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
TR9-7 cells
that were
released from
being mostly
synchronized
in G2 (using
Hoechst
33342)
shortly before
inorganic
arsenic
treatment
began





HeLa cells

















TRL1215
cells














Arsenic
Species



As111 SA













As111
ATO

















As111 SA















Concentration(s)
Tested (nM)



5














2

















0.125,0.250,
0.500














Duration of
Treatment



3hr














6 and 24 hr

















24 weeks















LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Conclusions based on microarray
analysis (done by hybridizing
fragmented cRNAs to U95Av2
GeneChips) in cells made p53(+) or
p53(-) by controlling tetracycline
levels: several genes were induced by
inorganic arsenic independently of p53
status, of which some of the biggest
effects were as follows (at both p53
conditions): HMOX1: huge ft by
>25x; MT2A: ft by >3x; SLC30A1: ft
by >3x. MKP-1 was induced only in
p53(+) cells, and
ubiquitin-conjugating enzyme E2N
was induced only in p53(-) cells.
In a cDNA microarray -based global
transcription profiling experiment that
compared the inorganic arsenic
treatment with a co-treatment of the
same inorganic arsenic dose with 30
uM emodin, the numbers of genes
with an expression level that differed
between the two treatments by more
than a factor of 2 at the 2 time points
were 793 and 480, respectively. The
affected genes included genes
involved in such things as cell
signaling, organelle functions, cell-
cycle control, redox regulation, and
apoptosis. The manner of data
presentation did not permit
identification of genes affected
exclusively by inorganic arsenic.



Various












mRNA levels determined
by real time RT-PCR:
effects on oncogenes
AFP: ft at 0.250, big ft at
0.500; WT-1: ft at 0.125,
big ft at 0.250 and 0.500.
c-jun: ft at 0.250, big ft
at0.500;H-ras:ftat
0.125, big ft at 0.250 and
0.500.
(By 24 weeks of
exposure at the higher 2
doses, these cells had
undergone malignant
transformation and were
called CAsE cells.)




Reference



McNeely
etal.,2006













Wanget
al., 2005















Liu et al.,
2006d












C-62   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



TRL 1215
cells
















TRL 1215
cells

















Arsenic
Species



As111 SA

















As111 SA


















Concentration(s)
Tested (nM)



0.125,0.250,
0.500
















0.125,0.250,
0.500

















Duration of
Treatment



24 weeks

















24 weeks


















LOECa
(HM)



Various

















Various















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
mRNA levels determined
by real time RT-PCR:
effects on stress-related
genes
HMOX-1: flat 0.125 and
0.250, big ft at 0.500;
SOD: flat 0.250, big flat
0.500.
MT-1: big ft at 0.250, ft
at 0.500; GSTjc: ft at
0.125, big ft at 0.250 and
0.500.
(By 24 weeks of
exposure at the higher 2
doses, these cells had
undergone malignant
transformation and were
called CAsE cells.)
mRNA levels determined
by real time RT-PCR:
effects on cell cycle
regulators
CyclinDl:ftat0.125,
then ft with dose to
0.500.
PCNA: ft at 0.250, big ft
at 0.500.
p21 : big U at 0.125, then
U with dose to 0.500.
p!6: Hat 0.125, big U to
-0% at 0.250 and 0.500.
(By 24 weeks of
exposure at the higher 2
doses, these cells had
undergone malignant
transformation and were
called CAsE cells.)




Reference



Liu et al.,
2006d
















Liu et al.,
2006d














C-63   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



TRL 1215
cells

















TRL 1215
cells















Arsenic
Species



As111 SA


















As111 SA
















Concentration(s)
Tested (nM)



0.125,0.250,
0.500

















0.125,0.250,
0.500















Duration of
Treatment



24 weeks


















24 weeks
















LOECa
(HM)



Various


















Various













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
mRNA levels determined
by real time RT-PCR:
effects on growth factor
genes
c-met: big ft at 0.125,
then ft with dose to
0.500.
HGF: ft at 0.125, big ft at
0.250 and 0.500.
FGFR1: huge U at 0.250,
then U to -0% at 0.500.
IGF-II: huge U to -0% at
all doses.
(By 24 weeks of
exposure at the higher 2
doses, these cells had
undergone malignant
transformation and were
called CAsE cells.)
Protein levels determined
using Western blots:
AFP: slight ft at 0.125
through 0.500; WT-1:
huge ft at 0.125 through
0.500.
CyclinDl:ftat0.125
through 0.500; p!6: huge
U at all doses.
p21 :li at 0.125, then U
with dose to 0.500.
(By 24 weeks of
exposure at the higher 2
doses, these cells had
undergone malignant
transformation and were
called CAsE cells.)




Reference



Liu et al.,
2006d
















Liu et al.,
2006d













C-64   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



TRL 1215
cells





















CL3 cells

























Arsenic
Species



As111 SA






















As111 SA

























Concentration(s)
Tested (nM)



0.500






















2

























Duration of
Treatment



24 weeks






















24 hr

























LOECa
(HM)



Various






















2






















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
mRNA levels determined
by real time RT-PCR:
effects of 72 -hr post-
treatment with 5 |iM 5-
aza-dC
(results were compared
to cells with inorganic
arsenic treatment alone)
Mr-lift 19xover
already elevated level.
p21: ft 15x over what
was a greatly reduced
level, and level then far
above that with no
inorganic arsenic
exposure
pl6andIGF-II:NSE.
(By 24 weeks of
exposure at the higher 2
doses, these cells had
undergone malignant
transformation and were
called CAsE cells.)
ft Nqol mRNA level to
1.7x control; ft Nqol
protein level to 6.4x
control.
Cells given this
inorganic arsenic
pretreatment became
more sensitive to MMC-
induced cytotoxicity and
less sensitive to ADM-
induced cytotoxicity.
Co-treatment with MMC
and the Nqol inhibitor
DIG resulted in big ft in
cell survival (even higher
than after MMC
treatment without an
inorganic arsenic
pretreatment). CL3R15
cells, which have much
higher levels of Nqo 1
activity than CL3 cells,
are also much more
sensitive to MMC-
induced cytotoxicity than
CL3 cells.




Reference



Liu et al.,
2006d





















Lin et al.,
2006





















C-65   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

H460 cells


CL3 cells


















CL3R15 cells

CL3R15 cells
co-treated
with 200 uM
DIC for 6 hr
to inhibit
>95%ofthe
high
endogenous
level of Nqo 1
activity



Arsenic
Species


As111 SA
for both



















As111 SA
for both













Concentration(s)
Tested (nM)

2.5,5, 10,20


1,2.5,5, 10


















50, 100, 200
for both













Duration of
Treatment


72 hr
for both



















6hr
for both













LOECa
(HM)

2.5


1



















100

50








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival determined
by SRB assay: LC50s:
H460, 9.0; CL3, 3.7;
H460 cell have ~30x
higher endogenous Nqol
activity than CL3 cells,
and unlike CL3 cells
they showed no
statistically significant
induction of Nqol after
24-hr treatments with
inorganic arsenic at
doses of 2, 5, or 10.
(Even at the highest level
of induction in CL3
cells, the endogenous
level of Nqol activity in
H460 cells was still ~15x
higher.) These findings
raised question whether
Nqol plays a role in
inorganic arsenic
resistance.
Cell survival determined
by colony-forming assay:
LC50s: with DIC, -35;
without DIC, 120.












Reference


Lin et al.,
2006



















Linetal.,
2006










C-66   DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue



SIK cells

















Arsenic
Species



As111 SA

















Concentration(s)
Tested (nM)



2

















Duration of
Treatment



1,3,5,7,9
days
















LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Changes in protein levels detected at
each of the 5 times using 2-
dimensional gel electrophoresis of
soluble
proteins, with proteins
identified by peptide mass mapping
and other methods: -300 distinct
protein spots were monitored with
-40% showing >2-fold ft or U in silver
staining intensity at every time point,
about as many ft as U, with at least as
many changes on day 1 as on other
days. There were some changes as to
the proteins
affected over time. Of 10
proteins identified as showing
prominent changes within first few
days of inorganic arsenic treatment,
enzymes of the glycolytic pathway
were seen to be substantially elevated.
This dose of inorganic arsenic
suppressed
differentiation but did not
cause cell loss.
Reference



Lee et al.,
2005
















C-67   DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue
TRL 1215
cells















Arsenic
Species
MMAV
DMAV
TMAVO















Concentration(s)
Tested (nM)
1300
700
10000















Duration of
Treatment
20 weeks
for all















LOECa
(HM)
1300
700
10000














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft GST activity to 2.6x
control, ft cellular GSH
protein level to 2.2x
control.
ft GST activity to 1.7x
control, U cellular GSH
protein level to 43% of
control.
ft GST activity to 1.8x
control, ft cellular GSH
protein level to 2.4x
control.
All 3 treatments
increased GST, MRP
and MDR at the mRNA
level, and all 3
treatments increased
GST, Mrps, and P-gp at
the protein level. GST
and MRP have several
forms. While not all
forms responded in the
same way, the overall
responses were as noted.
Experiments with
inhibitors of GSH, Mrps,
and P-gp led to the
conclusion that increased
arsenic excretion caused
the resistance to arsenic-
induced cytotoxicity that
resulted from these
treatments.
Reference
Kojima et
al., 2006















C-68   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




Ahr+/+MEFs
































Ahr+/+MEFs









AG06 cells

AG06 cells





Arsenic
Species




As111 SA
































As111 SA









As111 SA

As111 SA





Concentration(s)
Tested (nM)




1,2,5
































Duration of
Treatment




6hr
































LOECa
(UM)
2 for
Nqol only

None for
CYP1B1





























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
mRNA levels measured
by real-time RT-PCR: ft
NqolmRNAto4x
control; 5 uMB[a]P
increased Nqo 1 mRNA
to 8x control; there was a
synergistic interaction
between them such that
the dose of 2 of
inorganic arsenic plus
thedoseof5ofB[a]P
increased Nqo 1 mRNA
to 27x control. A
synergistic interaction to
20x control also occurred
with a dose of 1 of
inorganic arsenic. At a
dose of 5 of inorganic
arsenic, the interaction
became only additive.
The interaction between
inorganic arsenic and
B[a]P regarding
CYPlBlmRNAwas
never more than additive.
In Ahr A MEFs, there
was no interaction of
inorganic arsenic and
B(a)P regarding Nqol
mRNA; the combined
treatment did not ft Nqol
mRNA levels. Thus the
synergistic interaction
requires the wt Ahr gene.
Following treatment with 2 uM inorganic arsenic, 5 uM B[a]P, or both, for an
unspecified time, oligonucleotide microarray analysis of 13,332 sequences
from annotated mouse genes: they identified 64 genes that were up-regulated
or down-regulated by inorganic arsenic, B[a]P, or both; of these, 13 showed at
least a 2x up-regulation and 12 caused at least a 2-fold down-regulation in
gene expression because of the inorganic arsenic treatment alone. Many
different types of genes were affected. One of the major consequences of
exposure to these mixtures was the up-regulation of oxidative stress and
protein chaperone responses and the down-regulation of the TGF-(3 pathway.
Exposure to inorganic arsenic/B[a]P mixtures caused regulatory changes in
the expression of detoxification genes that ultimately affect the metabolic
activation and disposition of toxicants.

0.2, 1, 3, 10

3


24 hr

48 hr
48 hr


1

0.2
o
J



ft GSH concentration.

Specific activities: GSTrc
ftto~1.6xandyGCSto
~2.2xatdoseof3.




Reference



Kannet
al., 2005a
































Kannet
al., 2005a








Snow et
al., 1999
Snow et
al., 1999

C-69   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
AG06 cells

GM847 cells



GM847 cells







AG06 cells







WI38 cells





HaCaT cells



HaCaT cells





JB6 C141
cells






K562 cells







Arsenic
Species
As111 SA





As111 SA







As111 SA







As111 SA





As111 SA



As111 SA,
Asv

MMAV,
DMAV

As111 SA







As111 ATO







Concentration(s)
Tested (nM)
0.1,0.25,0.5, 1.0,
5, 10, 25




0.5, 1.0, 10, 25







0.2, 4, 20







0.3, 1.4,5.7,29




0.001,0.01,0.05,
0.1,0.5, 1.0



1.0


1.0


0.05, 0.2, 0.8,
3.125, 12.5,50,
200





2.5







Duration of
Treatment
24hr





24hr






3hr

24 hr






Not reported




2 days

14 days



2 days
for all



15 min







6hr







LOECa
(HM)
1.0

0.25



0.5 for ft

0.5 for U




0.2

0.2






0.3




0.1

0.01


1.0


None


0.8







2.5




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft GR protein level to
2.9xatl.

ft GR mRNA level to
1.3x and enzyme activity
to 2.0x at 0.25.
ft Trx, TrxR, GR mRNA
levels; for TrxR and GR:
ftto~2.7xby 10 and
thenlito~1.5xby25).
U GPx mRNA level to
~0.5xby land~0.2xby
25.
APE/Ref-1 mRNA
levels: at 3 hr: ft to ~2.7x
at 0.2 and then only
slight ft to ~3. Ox at 20.
At 24 hr: ft to ~3. Ox at
0.2butUto~0.9xat20.
(APE/Ref-1 is required
forBER.)
ft DNA Poly p level
(both cytoplasmic and
nuclear) to ~2xby 1.4
butUto~0.8xby29.
(DNA Poly p is required
forBER.)
U p53 protein; ft mdm2
protein.

U p53 protein; ft mdm2
protein.
U p53 protein; ft mdm2
protein; (much bigger
effect for As111).


No significant change.
ft Erk activation
resulting from Erk
phosphorylation; another
experiment showed that
overexpression of
dominant negative Erk2
blocks arsenite-induced
activation of Erk.
ft GlycoA, HLA-DR,
CD33, andCD34onthe
cell surface, indicating
maturation of myeloid
cells.




Reference

Snow et
al., 2001




Snow et
al., 2001





Snow et
al., 2001






Snow et
al., 2001




Hamadeh
etal., 1999



Hamadeh
etal., 1999



Huang et
al., 1999a






Li and
Broome,
1999


C-70   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
MCF-7 cells














H460 cells










Primary
cultures of rat
cerebellar
neurons



MC/CAR
(human
multiple
myeloma cell
line)














Arsenic
Species
As111 ATO













As111 ATO











As111 SA






As111 ATO


















Concentration(s)
Tested (nM)
3














10










10






2


















Duration of
Treatment
12hr














24hr










24 hr






72 hr


















LOECa
(HM)
3














10










10






2















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Microtubule
polymerization, with a
major effect on the
organization of the
cellular microtubule
network, resulting in the
formation of long
polymerized microtubule
bundles; ft p34cdc2/cyclin
B complex (both
activation and
accumulation); ft Bcl-2
phosphorylation.
The following changes
occurred only in mitotic
cells (definitely not in
interphase cells): ft
caspase-3 activation, ft
caspase-7 activation,
cleavage of PARP and (3-
catenin. These findings
suggest that arsenic-
induced mitotic arrest
may be a requirement for
the activation of
apoptotic pathways.
ft caspase activity
(apoptosis is blocked in
these cells if caspase is
inhibited; there was a
much bigger effect with
a 48-hr treatment).
ft caspase-3 activity,
p21,andCDKl;up-
regulation of cdc2
phosphorylation; U in
CDK6, cdc2, cyclin A,
and Bcl-2 levels; ft
binding of p21 with
CDK6, cdc2, and cyclins
A and E; U activity of
CDK6-associated kinase
and cdc2 -associated
kinase; loss of
mitochondria!
transmembrane potential
(Av|/m); no change in p27,
CDK2, CDK4, or cyclins
B1,D1, or E levels.




Reference

Ling et al.,
2002











Ling et al.,
2002










Namgung
and Xia,
2001





Parketal.,
2000













C-71  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
PCI-1 (human
head and neck
squamous cell
carcinoma
cell line)



(Human
myeloma-like
cell lines)
RPMI 8226
Karpas 707
U266
LAK effector
cells




WRL-68
(human
hepatic cell
line)
Human aorta
VSMCs
(vascular
smooth
muscle cells)
Human aorta
VSMCs
(vascular
smooth
muscle cells)
WI38 cells




WI38 cells










Arsenic
Species

As111 ATO







As111 ATO





As111 ATO





As111 SA



As111 SA



As111 SA




As111 SA




As111 SA










Concentration(s)
Tested (nM)

2






0.5






0.5




0.001,0.01,0.1,
10


2.5, 5, 10




2.5, 5, 10, 20




0.1

10, 20, 50
0.1


50









Duration of
Treatment

3 days






72 hr





72 hr





16hr



4hr




4hr




14 days

18 hr
14 days


18 hr









LOECa
(HM)

2






0.5





0.5





0.1

0.001

~5




~5




0.1

50
0.1


50






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft p21 and its binding
with cdc2; U protein
levels of cdc2 and cyclin
Bl;
U activity of cdc2 kinase;
no change in CDK2,
CDK4, CDK6 and
cyclins A, Dl, E.
ft CD38 and CD54
(molecules involved in
cell-cell interactions).



ftCDllaandCD31
(molecules involved in
cell-cell interactions, and
the ligands [i.e., counter-
receptors] of CD54 and
CD38, respectively).
ftGSH.

ft CK18.

ft p22phox mRNA
expression (p22phox is 1
of at least 7 subunits of
NADH oxidase.) U oc-
actin mRNA expression.
ft NADH oxidase
activity. The effect was
even stronger, with a
LOEC of 1, in
nonproliferating VSMCs.
ft p53 (3 -fold increase).

ft p53 (large increase).
ft cyclin Dl; also
treatment blocks ft in
p21 that occurs follow
exposure to 6 Gy of
ionizing radiation.
U cyclin D 1 ; also
treatment mostly blocks
ft inp21 that occurs
follow exposure to 6 Gy
of ionizing radiation.




Reference

Seoletal.,
1999





Deaglio et
al., 2001




Deaglio et
al., 2001




Ramirez et
al., 2000


Lynn et al.,
2000



Lynn et al.,
2000



Vogtand
Rossman,
2001

Vogtand
Rossman,
2001






C-72   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


Untransforme
dand
immortalized
RWPE-1 cells
(human
prostate
epithelial cell
line)
















PAEC from
freshly
harvested
vessels











Arsenic
Species



As111 SA






















As111
probably
ATO, but
called
arsenite










Concentration(s)
Tested (nM)



5






















5














Duration of
Treatment



Up to 30 wk






















15 minto 3 hr
depending on
endpoint












LOECa
(HM)



5






















5











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft MMP-9 activity (likely
biomarker of when
malignant transformation
occurred); U in DNA
methyltransferase
activity but no change in
DNA methyltransferase
mRNA levels; ft K-ras
mRNA and protein
levels. Time course
study suggested over-
expression of K-ras
preceded malignant
transformation. There
was no indication of
mutations being induced
in K-ras gene and no
indication that
hypomethylation of K-
ras promoter region
caused K-ras changes.
The cells became
tumorigenic after 29
weeks of treatment and
were then called the
CAsE-PE cell line.
ft NF-KB dependent
transcription, ft H2O2-
dependent tyrosine
phosphorylation (which
was blocked by CAT),
ftcSrc activation. MAP
kinases, extracellular
signal-regulated kinase,
and p3 8 were only
activated at a dose of
100, which causes cell
death.




Reference


Benbrahim
-Tallaa et
al., 2005






















Barchowsk
y et al.,
1999a








C-73   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


HeLa S3 cells
















TM3 cells













E7 cells









Arsenic
Species


As111 SA
















As111 SA













As111 ATO









Concentration(s)
Tested (nM)


5
















0.008, 0.77, 7.7













0.025,0.05,0.1,
0.25,0.51








Duration of
Treatment


24 hr
















70 days













4 weeks









LOECa
(HM)


5
















Various













0.005






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Changes in cells that
were arrested in mitosis
by As111: c-Mos was
hyperphosphorylated,
cyclin A was degraded,
cyclin B accumulated;
ftftp34cdc2/cyclinB
kinase activity. These
and numerous other
changes in mitotic
proteins were similar to
changes seen in cells
arrested in mitosis by
nocodazole, which is a
known microtubule
disassembly agent.
Changes in expression of
cell-cycle related genes:
U at 7.7 for Cyclin Dl;
for PCNA: flat 0.008, U
at 0.77 and 7.7.
Changes in expression of
DNA repair genes:
U at 0.77 and higher for
ERCC6andOGGl;
Uat7.7forXPC,MYH,
and DNA polymerase-p.
Changes in expression of
other genes:
U at 7.7 for, MnSoD, and
Bax;
for DNMT1: flat 0.008,
NSE at 0.77, U at 7.7.
ft Aurora-A protein
expression level, with a
positive dose-response,
reaching 4.2x control at
dose of 0.1; unreported
data showed ft Aurora-A
mRNA.




Reference


Huang and
Lee, 1998















DuMond
and Singh,
2007











Tseng et
al., 2006





C-74   DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue




BEAS-2B
cells




































Arsenic
Species




As111 AC





































Concentration(s)
Tested (nM)




1.25, 2.5, 5, 10, 20





































Duration of
Treatment




12 hr





































LOECa
(HM)




1.25





































Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft GADD45a protein
expression level, with a
positive dose-response;
however, only a marginal
ft in GADD45a
transcription;
pretreatment with NAC
completely blocked the ft
ofGADD45a. After
inorganic arsenic dose of
20 for 4-20 hr: transitory
activation of Akt and
transitory ft
phosphorylation of
FoxOSa. Inorganic
arsenic induced
accumulation of
GADD45a mRNA and
did not affect the
degradation of
GADD45a protein.
Inorganic arsenic
stabilized GADD45a
mRNA through
nucleolin; it induced the
binding of mRNA
stabilizing proteins,
nucleolin and less
potently, HuR, to
GADD45a mRNA.
Inorganic arsenic did not
affect the expression of
nucleolin; inorganic
arsenic treatment
resulted in redistribution
of nucleolin from
nucleoli to nucleoplasm.
Silencing of nucleolin
reversed inorganic
arsenic-induced
stabilization of the
GADD45a mRNA.
Reference




Zhang et
al., 2006




































C-75   DRAFT—DO NOT CITE OR QUOTE

-------
   Type of
 Cell/Tissue
 Arsenic
 Species
 Concentration(s)
   Tested QaM)
Duration of
 Treatment
LOECa
 (HM)
 Results (Compared
With Controls, With
 All Concentrations
       Being
in nM Unless Noted)
Reference
Gclm+/+ MEF
  cells and
 Gclrn7- MEF
  cells, from
   GCLM
  knockout
    mice
 As111 SA
  for all
  See rows under Apoptosis and Cytotoxicity for this citation for experimental
conditions.  Analysis of global gene expression profiles revealed up-regulation
or down-regulation of vast numbers of genes by inorganic arsenic. Significant
    changes were largely consistent with changes in the expression of DNA
   damage and repair genes, the suppression of TGF-(3 signals, inhibition of
  integrin-mediated cell adhesion, induction of multiple transcription factors,
    repression of co-repressors, and the  derailment of cell cycle regulatory
    functions.  Inorganic arsenic exposure also caused profound changes in
   protein levels in what appear to be conflicting regulatory changes. These
        changes go hand in hand with massive up-regulation of HSPs,
  metalloproteinases, and proteasome  components, and the authors suggested
 that inorganic arsenic induces critical changes in protein folding and structure
and that the cells mount a major effort to properly refold misfolded proteins or
  to eliminate them altogether.  Global gene expression profiles also indicated
  that tBHQ is significantly effective  in reversing inorganic arsenic-induced
     gene deregulation in Gclm+/+ but not in Gclm"7" MEFs.  These results
  suggested that regulation of GSH levels by GCLM determines the sensitivity
  to inorganic arsenic-induced apoptosis  and cytotoxicity by setting the overall
	ability of the cells to mount an effective antioxidant response.	
                                                                                      Kannet
                                                                                     al., 2005b
  NB4 cells
NB4-M-AsR2
    cells
As111 ATO
 for both
      0.5, 1
                  2,4
                                                   16hrs
                                                  for both
                  0.5
            JNK activation leading
           to phosphorylation of c-
           jun, after treatment with
              ATO alone and co-
            treatment with 100 uM
                   Trolox:
           At 0.5: slight ft alone, ft
                 with Trolox.
           At 1: big ft alone, huge ft
                 with Trolox.

            At 2: slight ft alone, ft
                 with Trolox.
           At 4: big ft alone, huge ft
                 with Trolox.
                                                                          Diazetal.,
                                                                            2005
  JB6C141
 PG13 cells

  JB6C141
 PG13 cells
 exposed to 4
kJ/m2ofUVB
   at end of
  inorganic
   arsenic
  treatment
 As111 SA
 for both
    1, 5, 10, 20
     for both
   24hrs
  for both
              in p53 activity with
            dose, reaching -30% of
             control at dose of 20.

             U in p53 activity with
            dose, reaching -5% of
              that with the UVB
            treatment alone at dose
              of 20. The UVB
              exposure strongly
           stimulated p53 activation
              (to ~9x the control
            level), and the inorganic
              arsenic treatment
            inhibited that increase,
             reducing it to  a point
           estimate less than that of
            the untreated control at
               the dose of 20.
                        Tang et al..
                           2006
                                                    C-76   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
JB6C141
P+l-l cells

JB6C141
P+l-l cells
exposed to 4
kJ/m2ofUVB
at end of
inorganic
arsenic
treatment




JB6C141
cells exposed
to 4 kJ/m2 of
UVB at end
of inorganic
arsenic
treatment











Arsenic
Species

As111 SA
for both













As111 SA
for both















Concentration(s)
Tested (nM)
1, 5, 10, 20


0.1, 1,5, 10











5, 10

1,5,10















Duration of
Treatment

24hrs
for both













24hrs
for both















LOECa
(HM)
5


5











5

5












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft in AP-1 activity to 2x
control at 5 and to 5x
control at 10, back to
control level at 20.

ft in AP-1 activity to
1.5x and 1.7x that with
the UVB treatment alone
at doses of 5 and 10,
respectively. It should
be noted that the UVB
exposure strongly
stimulated AP-1
activation (to ~6x the
control level).
llUVB-inducedp53
phosphorylation (at
serines 15 and 392);
bigger U at 10.

llUVB-inducedp53
DNA binding activity;
bigger U at 10.
Other experiments not
involving UVB showed
that inorganic arsenic
inhibited casein kinase
2a activity and decreased
p5 3 -regulated p21
protein expression.




Reference


Tang et al.,
2006














Tang et al.,
2006










C-77   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


SV-HUC-1
cells

























SVEC4-10
cells









Arsenic
Species
As111 SA

MMAm

DMA111
























As111 SA










Concentration(s)
Tested (nM)
0.5

0.05,0.1,0.2

0.2, 0.5
























5, 10, 20










Duration of
Treatment

Subcultured
twice weekly
for 25
passages
























24 hr










LOECa
(HM)
0.5

0.05

0.2
























5







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Results of cDNA
microarray analysis of
-2000 genes: 114 genes
were differentially
expressed among the 6
groups; DMA111 had a
substantially different
gene profile from other
2. Gene coding for IL-1
receptor, type II, was the
only gene with ft
expression by all
arsenicals. 11 genes had
U expression by all
arsenicals. For 2 of
those 11, transcription
was partially restored by
treatment with 5-aza-dC,
which suggests that the
suppression resulted
from epigenetic DNA
hypermethylation. The
treatments also caused
differential
morphological changes
affecting cell size, extent
of aggregation, and
adhesion ability.
Protein levels:
a7-nAChR: slight U at 5,
huge U at 10 and 20,
with only a trace present
at 20.
eNOS: slight U at 5,
huge U at 10 and 20,
with none present at 20.
ChAT: NSE.




Reference



Suetal.,
2006
























Hsu et al.,
2005






C-78   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


BEAS-2B
cells
















BEAS-2B
cells













BEAS-2B
cells







HT1 197 cells










Arsenic
Species


As111 ATO

















As111 ATO














As111 ATO








As111 SA










Concentration(s)
Tested (nM)


10, 20, 50

















10, 20, 50














10, 20, 50








10










Duration of
Treatment


12hr

















6hr














6hr








8hr










LOECa
(HM)


10

















10 for all














10








10







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
HSP70 protein ft: fold
increases over control by
Western blotting after
12-hr recovery period:
2.6x, 2.5x, and 7.9x at
doses of 10, 20, and 50,
respectively; alternative
ELISA analysis gave
similar response but with
much higher-fold
increases over the
control. Co-treatments
with large doses of
antioxidants CAT, SOD,
NAC, or SF considerably
reduced the arsenic
effect, with the NAC
treatment completely
eliminating it.
mRNA levels determined
by RT-PCR, with no
recovery time after
exposure,
fold ft over control:
At 10: HSP70A, 4.4x;
HSP70B, 4.3x; HSP70C,
3.6x.
After 4-, 8-, and 12-hr
recovery periods, mRNA
levels usually U to levels
closer to control and
often NSE; however, all
increases remained
significantly higher than
control at dose of 50.
Intracellular GSH levels:
U to 80% of control at
10, followed by dose-
related decrease to 70%
of control at dose of 50;
co-treatment with NAC
blocked this effect of
inorganic arsenic.
p53 protein levels: slight
ft; at 24 hr at this dose:
big ft to 4x control.
p21 protein levels: ft to
7.5x control; also at this
dose: at 12-20 hr, much
smaller increases; at 24
hr, big U; at 4 hr, 2.4x
control.




Reference


Han et al.,
2005

















Han et al.,
2005












Hanetal.,
2005






Hernandez
-Zavala
etal.,
2005





C-79   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


SVEC4-10
cells














RAW264.7
cells




















Arsenic
Species


As111 SA















As111 SA





















Concentration(s)
Tested (nM)

4, 8, 12, 16;
separase was
tested only at the
highest dose













2.5,5





















Duration of
Treatment


24 hr















24 hr





















LOECa
(HM)


Various















2.5


















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Effects on protein levels:
Securin: U at 12 to 23%,
U at 16 to 5%.
Separase: ft to 1.2x
control (of ?-able
significance).
Phospho-CDC2
(threonine-161): U at 16
to 34%.
CDC2: U at 12 to 73%, U
at!6to38%;cyclinBl:
U at 16 to 11%.
p53(DO-l):ftat4to2x
control with positive
dose-response reaching
8x control at dose of 16.
TRAP histochemistry
was done 3 days after the
end of the inorganic
arsenic treatment:
huge ft in TRAP activity
at both doses; this
increased activity
accompanied
multinucleated cell
formation and the
beginning of osteoclast
differentiation; the level
of effect at both doses
was comparable to (and,
at the dose of 2.5,
probably higher than)
that caused by a RANKL
treatment; co-treatment
with CAT blocked most
of the inorganic arsenic-
induced effect.




Reference


Chao etal.,
2006a














Szymczyk
etal., 2006

















C-80   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HCT1 16 cells
(securin +/+)



HCT1 16 cells
(securin -/-)












RKO cells
(p53 wt)

SW480 cells
(p53 mutant)



FGC4 cells









HepG2 cells












Arsenic
Species


As111 SA
for both
















As111 SA
for both






As111 SA









As111 SA












Concentration(s)
Tested (nM)


4, 8, 12, 16
for both

















8, 16
for both



50,65
Equivalent to <5%
and 20-25%
cytotoxicity






15,55
Equivalent to <5%
and 20-25%
cytotoxicity











Duration of
Treatment


24 hr
for both

















24 hr
for both




24 hr










24 hr












LOECa
(HM)

Various




Various













16

16




Various










Various









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Effects on protein levels:
Securin: U at 4, then U
with dose to -30% at 16.
Phospho-p53 (serine 15):
ft to 2x control at 4 and
then ft with dose to 6x
control at 16.
p53 (DO-1): ft to 2x
control at 12 and ft to
3. 4x control at 16.
No securin present at any
dose in -/- mutant.
Phospho-p53 (serine 15):
ft to 3.5x control at 4 and
then ft with dose to 7x
control at 16.
p53 (DO-1): ft to 1.8x
control at 4 and then ft
with dose to 3.2x control
at 16.
Effects on protein levels
of securin: rather similar
U in both, reaching 27%
and 13% of control in
RKO and SW480,
respectively.
Effects on protein levels
of SPs:
MT, HSP60andHSP90:
NSE at either dose.
HSP25: big ft at 50, big
ft at 65.
HSP40: big ft at 50, big
ft at 65.
HSP70: big ft at 50, huge
ft at 65.
Effects on protein levels
of SPs:
MT: NSE at 15, very
slight ft at 55.
HSP60andHSP90:NSE
at either dose.
HSP27: slight ft at 15, ft
at 55.
HSP40: slight ft at 15,
big ft at 55.
HSP70: ft at 15, big ft at
55.




Reference




Chao etal.,
2006a















Chao etal.,
2006a





Gottschalg
etal., 2006







Gottschalg
et al., 2006









C-81  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Rat
hepatocytes








HELP cells

















MDAH 2774
cells










UROtsa cells





UROtsa cells







Arsenic
Species

As111 SA









As111 SA

















As111 ATO











As111 SA





MMAm







Concentration(s)
Tested (nM)
10,20
Equivalent to <5%
and 20-25%
cytotoxicity







0.1,0.5, 1,5, 10

















1, 2, 5, 8











0.5, 5, 10, 25





0.05







Duration of
Treatment

24 hr








3, 6, 12, 24,
or48hr
















Probably
72hror
96 hr










24 hr





12 weeks







LOECa
(HM)

Various









Various

















lor 2











5





0.05




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Effects on protein levels
of SPs:
MT, HSP60andHSP90:
NSE at either dose.
HSP25: ft at 10, ft at 20.
HSP40: NSE at 10, ft at
20.
HSP70: NSE at 10, big ft
at 20.
HSP27 protein: ft at 0.5
and 1 after 12-hr
treatment, but U at 5 and
10 after
48-hr treatment; HSP27
was said to be a
chaperone whose
expression protects
against oxidative stress
and is anti-apoptotic.
HSP70 protein: U at 1
and 5 after 12-hr
treatment, but ft at 5 and
10 after 24-hr treatment;
an inducible form of
HSP70 was said to be
expressed at a high level
in various malignant
human tumors.
U topoisomerase Ila to
about half of control
value at dose of 5
(paralleling degree of
cytotoxicity) — there is
some question about this
result because band
densities were not
normalized to another
protein; decrease
possibly resulted from U
in cell number.
ft accumulation of high-
molecular-weight Ub-
conjugated proteins.
Co-treatment with BSO:
ftft in the same effect,
which was then seen
even at dose of 0.5.
Huge ft COX-2 protein,
with an even higher level
after 24 weeks and still
high level after 52
weeks.




Reference
Gottschalg
et al., 2006









Yanget
al., 2007
















Askar et
al., 2006










Bredfeldt
et al., 2004




Eblin et
al., 2007



C-82   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


UROtsa cells























UROtsa cells












Arsenic
Species
As111 SA


MMA111




















As111 SA



MMA111










Concentration(s)
Tested (nM)
1, 10


0.01,0.05,0.1




















1, 10



0.05,0.5,5










Duration of
Treatment

4hr
for both






















30 min
for both












LOECa
(HM)
1


0.01




















1



0.05







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Big ft COX-2 protein
level at both doses.

Regarding COX-2
protein level: huge ft at
0.01, big ft over control
at 0.05, ft over control at
0.1. Various
experiments, including
some with
pharmacological
inhibitors of various
signal transduction
pathways, led to the
conclusion that MMA111
appears to stimulate
ligand-independent
activation of EGFR,
subsequent ERK-1 and -
2 phosphorylation via
MEK-1 and -2, as well
as activation of PIS K,
which leads to elevated
levels of COX-2 protein.
ft HSP70 protein (similar
response at both doses;
with lower dose, the
level decreases from 60
to 240 min); ft MT
protein (much bigger ft
at higher dose).
ft HSP70 protein (strong
response at all doses).
ft MT protein (much
bigger ft at higher
doses).




Reference
Eblin et
al., 2007
























Eblin et
al., 2006








C-83   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




HaCaT cells















MEF cells













MEF cells













Arsenic
Species


As111 SA


Asv

MMAm

DMA111










As111 SA













As111 SA













Concentration(s)
Tested (nM)


2, 6, 10


1,5, 10

1,2,3

1,4,7










0.01,0.1,5, 10,
20,40












20 in most assays













Duration of
Treatment




24 hr
for all














5hr














LOECa
(HM)


6,2


None

20
,3

4










5











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Extent of selenium
incorporation into
selenoproteins
determined using
75Se-selenite:
LOECsof6and2forli
TrxRl and U cGpx,
respectively; big U at
higher dose(s).

NSE.
LOECsof2and3forft
TrxRl and U cGpx,
respectively.
ft of TrxRl and cGpx at
dose of 4 and decrease
for both proteins to near
control levels at higher
dose.
ft eIF2a
phosphorylation; ft
ATF4 protein; ft ATF3
protein. At doses >10: ft
GADD45a protein and ft
CHOP protein. All
effects showed
substantial
dose-related increases.
Effects were mostly
blocked by NAC
pretreatment. (ATF3
was not tested.)
GADD45a is a small protein implicated in the regulation
of the cell cycle, DNA repair, genome stability, innate
immunity, and apoptosis. Additional tests with
modulators and genetic variants of MEF cells showed
the following: ATF4 is required for an increase in
GADD45a mRNA following inorganic arsenic exposure,
and its induction is independent of p53. ATF4 binds to
a GADD45a promoter element in response to inorganic
arsenic stress. Exposure to inorganic arsenic reduces
proteasome activity, which permits the increase in
transcription of GADD45a to actually result in an
increase in the protein level of GADD45a, which is
labile.




Reference





Ganyc et
al., 2007













Jiang et al.,
2007












Jiang etal.,
2007









C-84   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Protein
extracts
(membrane
fraction)
derived from
BAEC cells



N-18 cells
















N-18 cells









Arsenic
Species
MMAm


As111 SA,
Asv,
MMAvor
DMAV



As111 SA















Asv
potas-
sium
arsenate







Concentration(s)
Tested (nM)
1,2.5,5,7.5, 10,
15


10





5, 10, 20, 50
















20









Duration of
Treatment

5 min
for all







6hr
















6hr









LOECa
(HM)
1


None





5 for first
effect
noted
















20






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
For MMAm only: U
eNOS activity, IC50 = 2.1
and a 5-min treatment at
dose of 10 caused -90%
U; co-treatment with
DTT substantially
blocked the MMAm
effect, resulting in only
-50% U.
ft synthesis of HSP
proteins of 50, 73, 78,
89, 98, and 104 kDa.
Other experiments
demonstrated: ft
activation of HSF1
DNA-binding (detected
by EMSA) by dose of 20
(lowest dose tested) in 2
hr; ft induction of
HSP70-luciferase
reporter gene expression
by dose of 20 (lowest
dose tested) in 6 hr; an ft
induction of HSP70
rnRNA by dose of 50
(lowest dose tested) in 1
hr.

ft induction of HSP70-
luciferase reporter gene
expression (point
estimates suggests
weaker response from
Asv than from same dose
ofAsmSA).




Reference



Sumi et al.,
2005





Khalil et
al., 2006















Khalil et
al., 2006





C-85   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



N-18 cells




















hsf^
immortalized
MEF cells

hsf7'
immortalized
MEF cells

hsf7'
immortalized
MEF cells
transfected
withHSFl
expression
vector







Arsenic
Species



As111
























As111 SA
for all
















Concentration(s)
Tested (nM)


2, 5, 10, 20, 50,
100, 200, 500
























5, 10, 20, 50, 100,
200, 500 for all
















Duration of
Treatment


0.5, 1,2,3,6,
or!2hr
























Ihr
for all
















LOECa
(HM)


Various






















50


None


50











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft induction of HSP70-
luciferase reporter gene
expression, with bell-
shaped dose-response
curves for each duration
of treatment; e.g. for 1-hr
treatment, the peak
occurred at dose of 200
(highest peak seen); for
6-hr treatment, the peak
occurred at dose of 20;
the bell-shaped curves
shifted to the left as the
duration increased.
Results on HSP70-firefly
luciferase activity were
normalized against that
of Renilla luciferase to
correct for differences in
transfection efficiency
and/or toxic and non-
specific effects of the
experimental treatment
conditions.
ft induction of HSP70-
luciferase reporter gene
expression:
ft with dose up to peak at
200; still big ft at 500.

No effect; clearly
inorganic arsenic
requires a functional
HSF1 gene to induce
HSP70-luciferase
reporter gene expression.

ft with dose up to peak at
200; still big ft at 500.
Generally similar results
were also found with
treatment durations of
0.5 and 2 hr.




Reference


Khalil et
al., 2006

























Khalil et
al., 2006












C-86   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


H1355 cells

















H1355 cells




























Arsenic
Species

As111 ATO

















As111 ATO





























Concentration(s)
Tested (nM)


5, 25, 50, 100, 200

















100




























Duration of
Treatment


24 hr

















24 hr




























LOECa
(HM)


Various

















Various

























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Phosphorylation of ERK
1/2: ft at 50, huge ft at
100 and 200.
Phosphorylation of INK:
slight ft at 50, huge ft at
100 and 200.
Phosphorylation of p38:
slight ft at 100, big ft at
200.
PARP cleavage: ft at 100
and 200.
Survivin protein level: U
at 100 and 200.
Ubiquitination in total
cell lysate: big ft at 100
(the only dose tested for
it).
Effects of pretreatments
with specific inhibitors
ofp38, JNK, MEK1/2
(upstream of ERK 1/2)
or ubiquitin-proteasome
showed that blockage of
either p3 8 or JNK
phosphorylation
attenuated the ATO-
induced down-regulation
of survivin and increase
of PARP cleavage;
however, blockage of
ERK 1/2 or ubiquitin-
proteasome did not
attenuate those same
effects. Also, only
inhibitors of p38 and
JNK affected ATO-
induced cytotoxicity,
which was just slightly
reduced (i.e., there was
~5%-8% more cell
survival). The specific
inhibitors of p38, JNK,
and MEK 1/2 did block
the phosphorylations of
p3 8, JNK, and ERK 1/2,
respectively.




Reference



Cheng et
al., 2006















Cheng et
al., 2006
























C-87   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue





A549 cells























A549 cells












Arsenic
Species




As111 ATO
























As111 ATO












Concentration(s)
Tested (nM)





2























2












Duration of
Treatment





48 hr























48 hr









Results (Compared
With Controls, With
All Concentrations
LOECa Being
(|oM) in |iM Unless Noted)
Protein levels and mRNA levels:
2 uM inorganic arsenic: NSE on
survivin. 200 uM sulindac: NSE on
survivin.
(Sulindac is a NSAID that inhibits
COX-2.)
(2 uM inorganic arsenic + 200 uM
sulindac): big U in survivin (by 72 hr
almost no survivin was protein
present). Protein levels only for
combined treatment:
big ft for p53 but NSE for XIAP,
cIAP-1, cIAP-2, andBcl-2. Inhibition
of p53 ft by siRNA blocked the down-
regulation of survivin by the (2 uM
inorganic arsenic + 200 uM sulindac)
treatment. (It is known that p53 binds
to the survivin promoter and
suppresses its transcription.)
Transfected cells with a survivin-
luciferase reporter also showed the big
U in survivin for the combined
treatment and NSE for single
treatments. Pretreatment with NAC
mostly (or entirely) blocked the
synergistic effect of a U of survivin
protein (was shown both by Western
blot and luciferase reporter assays).
More about the synergistic effect
between 2 uM inorganic arsenic and
200 uM sulindac: evidence that
changes in survivin levels are related
to synergistic big ft in cytotoxicity: (1)
if marked overexpression of survivin
by transfection, then U in cytotoxicity
by 1/3, (2) if inhibition of survivin
level by siRNA, then ft in cytotoxicity.
(Sulindac is a NSAID that inhibits
COX-2.)




Reference





Jinetal.,
2006b






















Jinetal.,
2006b








C-88   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


N-18 cells
















NHEK cells





NHEK cells





NHEK cells








Arsenic
Species


As111 SA
















As111 SA





As111 SA





As111 SA








Concentration(s)
Tested (nM)


2, 5, 10, 20, 50,
100, 200, 500















0.1, 1,5, 10





0.1, 1,5, 10





1








Duration of
Treatment


6hr
















72 hr





7 days





24, 48, 72 hr








LOECa
(HM)


10















0.1 for ft
Iforli




0.1 for ft
Iforli





1





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of HSP70-
luciferase reporter gene
expression: big ft at 10,
huge ft (peak) at 20, big
ft at 50, then NSE.
Effects of pretreatment +
co-treatment with
modulators:
DTT: almost entirely
blocked inorganic
arsenic effect; slight ft at
20 and 50, questionable
ft at 10 and 100.
NAC and GSH
(individually): ft at 10,
big ft at 20, huge ft
(peak) at 50, ft at 100,
then NSE.
Level of Pi-integrin
protein: after a possible
slight ft at 0.1, there was
all to
61-63% of control level
at other 3 doses.
Level of Pi-integrin
mRNA: after a possible
slight ft at 0. 1, a dose-
related 11 at other 3 doses
reaching 47% of control
at dose of 10.
Level of FAK protein
based on
immunofluorescence: ft
at 24 hr followed by U
below control level at
later times, with almost
none present at 72 hr.




Reference


Khalil et
al., 2006















Lee et al.,
2006b




Lee et al.,
2006b




Lee et al.,
2006b




C-89   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Normal
human
mammary
epidermal
keratinocytes












Swiss 3T3
mouse cells















UROtsa cells











Arsenic
Species

As111 SA
for all















As111 SA














As111 SA

MMAV

DMAV









Concentration(s)
Tested (nM)
0.005,0.5, 1,2.5


1,2.5,5














1, 2.5, 5, 10, 20,
40















5, 50 for all











Duration of
Treatment
4hr


8hr














16 hr
















2 hr for all











LOECa
(HM)
0.005


2.5














1
















5 for all








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft COX-2 mRNA (also at
8 and 24 hr).

ft COX- protein (also at
12 hr), also under the
same or similar
conditions: ft PGE2
secretion,
phosphorylation of
p42/44 MAPK, and
DNA synthesis. Tests
with various modulators
showed that inorganic
arsenic111 elevates COX-2
at the transcriptional and
translational levels.
ft GSH synthesis;
starting at 2. 5: cell
retraction and loss of
thick cables of actin
filaments, U cytoskeletal
protein synthesis;
starting at 20: ft in
protein sulfhydryl
content of both
cytoskeletal and
cytosolic protein
fractions, with the time
course showing a slight
decrease before the
increase. There was also
severe loss of
microtubules.
Increased DNA binding
of the AP-1 transcription
factor, which is often
associated with the
regulation of genes
involved in cell
proliferation. For all 3
chemicals the response
was higher at dose of 50;
the highest amount of
binding was with SA.




Reference

Trouba
and
Germolec,
2004













Li and
Chou,
1992














Simeonova
etal.,2000







C-90   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


UROtsa cells

















C-33A cells

HeLa cells

Jurkat cells

LCL-EBV
cells







HeLa cells















Arsenic
Species


As111 SA



















As111 SA
for all











As111 SA















Concentration(s)
Tested (nM)


10,50



















1, 10, 25, 50 for
all











100, 200, 400















Duration of
Treatment


2hr



















24 hr
for all











30 min















LOECa
(HM)


10

















None

10

1

1








100












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Use of a cDNA array
consisting of 588 human
genes, and other
methods:
At 10: ft activity of 7
genes; U activity in 6
genes.
At 50: ft activity of 15
genes; U activity in 6
genes.
Specifics:
Genes affecting cell
growth: ft for c-fos, c-
jun, Pig 7, EGR-1, and
Rho8.
Genes affecting cell
growth arrest: ft for
GADD45 and
GADD153.
p53 protein expression:
No ft, slight U at high
doses, very high basal
level.

ft, peak at 25, low basal
level.

ft, peak at 10, moderate
basal level.
ft, peak at 10, very low
basal level.
Decreases above peak
may result from cell
death.
ft GADD153 mRNA
expression (harvested for
RNA isolation after 4
hours of incubation
following the arsenite
treatment). This effect
was increased by
pretreatment with BSO,
PHEN (slight increase),
BCS, or mannitol (an
HO" scavenger). Effect
was completely blocked
by pretreatment with
NAC.




Reference


Simeonova
et al., 2000

















Salazar et
al., 1997












Guyton et
al., 1996











C-91  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
WI38 cells

Simian vims
40 (SV40)-
transformed
subline of the
above
parental
W138 line
with twice the
GPx specific
activity of
parental cells


JB6 C141
cells






JB6 C141
cells


HFW cells
(diploid
human
fibroblasts)




HFW cells
(diploid
human
fibroblasts)
HFW cells
(diploid
human
fibroblasts)

Both
HL-60 cells
and
HaCaT cells






Arsenic
Species



As111 SA
for both









As111 SA

Asv





As111 SA

Asv



As111 SA






As111 SA



As111 SA





As111 SA








Concentration(s)
Tested (jiM)



100, 200, 400
for both









3.125, 12.5,50,
200

3.125, 12.5,50,
200



200

200



5, 10, 20






1,2.5,5, 10,20



0.5,2, 10





0.5,20








Duration of
Treatment



30 min











3hr






0 min

60 min



24hr






24 hr



24 hr





3 days








LOECa
(HM)



100 for
both









50

50





200

200



5






1



See next
column




0.5





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft GADD153 mRNA
expression (harvested for
RNA isolation after 4
hours of incubation
following the arsenite
treatment). The increase
was cut approximately in
half (i.e., half the slope)
in the transformed cell
line. Other parts of this
study showed that AP-1
is critical to oxidative
regulation of GADD153.
ft activity of JNKs:
stronger response at 50
for Asv (sodium
arsenate); both forms
shown some response by
Ihr at dose of 200;
arsenic did not induce
p5 3 -dependent
transactivation.

ft phosphorylation of
JNKs: stronger response
for Asv (sodium
arsenate).
ft heme oxygenase
activity (arsenic -induced
synthesis of this enzyme
was blocked by co-
treatment with
antioxidants sodium
azideorDMSO);ft
ferritin.
ft GSH (by 20 level
drops to control level).


ft SOD activities, U
catalase and GPx
activities, with LOECs
being 0.5, 2, and 10,
respectively.
ft hTERT protein
expression;
however U hTERT
protein expression at 20
(i.e., significantly
inhibited at higher
concentration).




Reference



Guyton et
al., 1996









Huang et
al., 1999b







Huang et
al., 1999b



Lee and
Ho, 1995





Lee and
Ho, 1995


Lee and
Ho, 1995




Zhang et
al., 2003




C-92   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
HaCaT cells




HL-60 cells






NB4 cells












NB4 cells

NB4 cells

HeLa cells

LoVo cells
MCF7 cells



Arsenic
Species
As111 SA




As111 SA





As111 ATO













As111 ATO

As111 ATO


As111 ATO
for all




Concentration(s)
Tested (nM)
0.5, 10, 20




0.1,0.5, 1, 10,20






0.75












0.75

0.1,0.25


2
for all




Duration of
Treatment
3 days




3 days






8 days












2 days

12 days


14 days
for all




LOECa
(HM)
0.5




0.1






0.75












0.75

0.1


2
for all

Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft telomerase activity;
however, telomerase
activity was below
control level at 10 and
even lower at 20.
ft telomerase activity;
however, telomerase
activity was below
control level at 10 and
even lower at 20.
U telomerase activity; U
hTERT mRNA and
protein levels; U c-myc
mRNA and protein
levels; ft hTER mRNA
level; no change in p53
mRNA or protein level;
no change in Spl mRNA
or protein levels. Further
experiments showed that
arsenic inhibits
transcription of hTERT
and inhibits the function
of Spl in hTERT
transcription.
U hTERT mRNA.

U hTERT mRNA.

U hTERT mRNA and
U c-myc mRNA
for all.





Reference
Zhang et
al., 2003



Zhang et
al., 2003





Chou et
al., 2001











Chou et
al., 2001
Chou et
al., 2001
Chou et
al., 2001


C-93   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Normal
human
keratinocytes
treated with
50 mJ/cm2
UVB before
or after
inorganic
arsenic
treatment













NB4 cells









SHE cells

















Arsenic
Species


As111 SA
for both


















As111 ATO
for both









As111 SA

Asv















Concentration(s)
Tested (nM)

1, as pretreatment


1, as post-
treatment
begun 24 hr after
irradiation














1


0.5, 1.0, 1.5,2.0







6,8

50, 100, 150















Duration of
Treatment


24 hr for both



















2 days


3 days







48 hr
for both
















LOECa
(HM)

1



None
















1


1.0







—














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
No change from control
in procaspase-8 and
procaspase-9 protein
levels or in caspase-3,
caspase-8, and caspase-9
enzyme activities; this is
considered an LOEC
because the inorganic
arsenic-pretreatment
blocked the effects of
UVB described below.
U procaspase-8 protein
level, slight U
procaspase-9 protein
level; ft caspase-8
enzyme activity; ft
caspase-9 enzyme
activity; ft caspase-9
enzyme activity; effects
similar to with UVB
alone.
As a result of
permeability changes in
the outer mitochondria!
membrane:
slight release of
cytochrome c into
cytoplasm; complete
release by 3 days of
treatment.
ft Cpp32 (was activated)
as shown by U of its
precursor.
From among these
treatment groups, 5
neoplastic transformed
cell lines were produced
that were shown to be
tumorigenic. Of these:
all had ft c-Ha-ras
(oncogene) mRNA
expression;
4 had ft c-myc
(oncogene) mRNA
expression;
a few other arsenic-
treated cell lines also
showed the same effects.




Reference



Chenetal.,
2005b



















Jing et al.,
1999







Takahashi
etal.,2002













C-94   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Peritoneal
macrophages
(PMs) from
CDFi mice












U118MG
cells








HaCaT cells
(immortalized
non-
tumorigenic
human
keratinocyte
cell line)

arsenic-TL
cells (arsenic-
tolerant cells,
which are
HaCaT cells
that were
cultured for
28 weeks in
100 nM As111
SA)



Arsenic
Species



As111 SA

Asv

MMAV

DMAV

TMAV







As111 ATO













As111 SA
for both















Concentration(s)
Tested (nM)



1.25,2.5,5, 10

125, 250, 500,
1000

1.25,2.5,5, 10
mM

1.25,2.5,5, 10
mM
1.25,2.5,5, 10
mM




1,5,10,25













20 for both
















Duration of
Treatment






48 hr for all












24 hr













6 hr for both
















LOECa
(HM)



1.25

500

None

2.5 mM

5 mM







1 or 5













20 for
both












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Changes in release of
TNF-a from
macrophages in the
presence of both
lipopolysaccharide and
recombinant murine
interferon y, which are
two compounds known
to increase secretory
functions of PMs:
U at 1.25, no change
from control at 5; big ft
at 10.
big ft at 500 and much
bigger ft at 1000.
no effect.
U at 2.5, 5 and 10 mM.
U at 5 and 10 mM.
Changes in protein
expression:
p53:ftatl, U at 5 or
higher; Bcl-2: ft at 1 or
higher.
Bax: U at 1 or higher;
HSP70: ft at 5 or higher.
Co-treatment with lipoic
acid blocked all of these
effects at an inorganic
arsenic111 dose of 5.

ft caspase-3 activation.




Much smaller ft in
caspase-3 activation than
in HaCaT cells.













Reference





Sakurai et
al., 1998













Cheng et
al., 2007











Pi et al.,
2005












C-95   DRAFT—DO NOT CITE OR QUOTE

-------
   Type of
 Cell/Tissue
 Arsenic
 Species
Concentration(s)
  Tested QaM)
Duration of
 Treatment
LOECa
  Results (Compared
 With Controls, With
  All Concentrations
        Being
 in nM Unless Noted)
                                   Reference
HUVEC cells
As111 ATO
       20
    2hr
  20
ft expression of 1C AM-1;
  effect was similarly
    strong after 24-hr
  treatment but weaker
after 4- or 8-hr treatment
(yet still ft above control
   level). Effect was
completely blocked by a
 1-hr pretreatment with
 15 mMNAC followed
         by a
  co-treatment of NAC
with the Asin-treatment.
Griffin et
al., 2003
Apoptosis
  K562 cells
Asm ATO
      2.5
   12 hr
  2.5
    ft annexin V, an
   apoptotic marker.
 Li and
Broome,
  1999
 NCI (human
 myeloma cell
     line)
As111 ATO
                       24 hr
                               Apoptosis was
                            demonstrated by 4,6-
                                diamidino-2-
                          phenylindole staining, by
                            the demonstration of
                            typical DNA ladders
                              corresponding to
                              internucleosomal
                              cleavage, and by
                              annexin-V and PI
                              staining. Various
                           indications of induction
                            of apoptosis were also
                            presented (with less
                          detail) for at least 1 other
                            myeloma cell line and
                           for fresh myeloma cells.
                              In the NCI cells,
                               [3H]thymidine
                            incorporation was also
                               used to assess
                            proliferation: the 50%
                              growth-inhibitory
                            concentration (IC50) in
                           NCI cells was found to
                            be 0.3 uM, based on
                           concentrations tested of
                            0.05,0.1,0.5, 1,5, 10
                            over 72 hr.  Similar
                           testing of 3 other human
                             myeloma cell lines
                          yielded IC50s of 0.1 for 1
                            line and ~1 for 2 other
                            lines, with much less
                              detail presented.
                                    Rousselot
                                   etal., 1999
                                                   C-96   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

MGC-803
cells







Primary
cultures of rat
cerebellar
neurons


MC/CAR
(human
multiple
myeloma cell
line)



V79-C13
Chinese
hamster cell
line




HL-60 cells



HaCaT cells







Arsenic
Species
As111 ATO









As111 SA

DMAV



As111 ATO







As111 SA









As111 SA
for both








Concentration(s)
Tested (nM)

0.01-1








5, 10

5 mM



1,2,5,10







10








0.1,0.5, 1, 10,20,
40
for both








Duration of
Treatment
24 hr









12 hr

48 hr



72 hr







24 hr








5 days
for both









LOECa
(HM)
0.01









5

5 mM



2







10







lor
possibly
0.5

10




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis detected by
flow cytometry and by
agarose gel
electrophoresis of
genomic DNA showing
typical DNA ladder; at
various doses apoptosis
was also induced in 5
other human malignant
cell lines.
Demonstrated by "DNA
ladders" with agarose gel
electrophoresis and
microscopic examination
(nuclear fragmentation
and/or condensation).
Apoptosis was
demonstrated by an
analysis using a FACStar
flow cytometer and by
detection of cell
membrane changes by
labeling with annexin V-
FITC and annexin PI.
Apoptotic cells appeared
by 6 hr after treatment
began and included 40%
of cells by 24 hr;
frequency gradually
decreased during 48 hr
of observation after
treatment ended.
ByuseofHoechst/PI
staining assay:
ft in apoptosis for both;
for both cell lines, there
was the same general
response, but to a lesser
extent, when same
treatments were given
over 1 or 3 days.




Reference
Zhang et
al., 1999








Namgung
and Xia,
2001



Parketal.,
2000






Sciandrello
etal.,2002








Zhang et
al., 2003





C-97   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
HL-60 cells

HaCaT cells
SW13 cells

SW480 cells

HT1080 cells













TRL 1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
SOuMBSO

TRL 1215
cells







Arsenic
Species



As111 SA
for all


















MMAV
for both










DMAV








Concentration(s)
Tested (nM)



1, 10, 20, 40
for all


















5 mM
for both










5mM








Duration of
Treatment



5 days
for all


















24 hr
for both










24 hr








LOECa
(HM)
1

10
-20

-20

1













None



5 mM









5mM





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ByuseofHoechst/PI
staining assay:
ft in apoptosis in all.
SW13 and SW480 are
telomerase negative cell
lines, and they showed
much less apoptosis at
all concentrations than
the other 3 cell lines.
HT1080 is a telomerase
positive cell line, and it
was intermediate in the
amount of apoptosis at
all concentrations to HL-
60 (which was higher)
andHaCaT. Thus there
is a strong positive
correlation between
telomerase activity and
susceptibility to arsenic-
induced apoptosis.
Apoptosis demonstrated
by TUNEL staining:
there was little evidence
of induction of apoptosis
by MMAV alone;
however, the cells also
treated with BSO
showed considerable
apoptosis.




Apoptosis demonstrated
by TUNEL staining:
huge ft, much more
extensive that that of the
considerable level of
apoptosis reported in row
above for MMAV + BSO.




Reference



Zhang et
al., 2003



















Sakurai et
al., 2005a








Sakurai et
al., 2005a





C-98   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
TRL1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
50uMBSO













TRL 1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
SOuMBSO



Arsenic
Species

MMAV
for both
























MMAV
for both













Concentration(s)
Tested (nM)

5mM
for both
























5mM
for both













Duration of
Treatment

12, 24, 36, or
48 hr for both
























24 hr
for both













LOECa
(HM)
5mM



5 mM





















None



5 mM








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis demonstrated
by FACS analysis after
annexin-V and PI
staining:
5 mM MMAV alone
caused some apoptosis
after 48 hr; however, that
response was slight
compared to the response
oftheMMAv + BSO
group after only 24 hr,
andtheMMAv + BSO
group showed huge ft at
36 hr and even bigger ft
at48hr. After 48 hr, the
percentages of annexin-
positive cells were as
follows: control, 1.9%,
BSO alone, 6.7%; MMAV
alone, 10.6%; MMAV +
BSO, 64%. The PI
staining showed that by
48 hr there were also
numerous induced
necrotic cells in the
MMAV + BSO group.

Apoptosis demonstrated
by agarose gel
electrophoresis showing
induced
internucleosomal DNA
fragmentation:
substantial DNA
fragmentation in
MMAV + BSO group; no
effect with MMAV alone.






Reference



Sakurai et
al., 2005a























Sakurai et
al., 2005a









C-99   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
TRL1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
50uMBSO


TRL 1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
50uMBSO













Primary
keratinocytes
(in third
passage)
obtained from
foreskins of
adults



Arsenic
Species

DMAV
for both













MMAV
for both
























As111 SA








Concentration(s)
Tested (nM)

5mM
for both













5mM
for both
























1,5, 10








Duration of
Treatment

24 hr
for both













12 hr
for both
























48 hr








LOECa
(HM)
5mM


5mM











None


5mM























1





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

Apoptosis demonstrated
by agarose gel
electrophoresis showing
induced
internucleosomal DNA
fragmentation: massive ft
with DMAV alone (many
times more than with
MMAv + BSOin
previous row); slight ft in
DMAV + BSD group
(about the same as with
MMAv + BSOin
previous row).
Cellular caspase-3
activation: ft to ~1.6x in
MMAV + BSO group; no
effect
without BSO; other
experiments showed that
co-treatment with 150
uM
Z-DEVD-FMK (a
caspase 3 inhibitor)
during preincubation
period and during a 24-hr
MMAV treatment blocked
almost all or all of the
cytotoxicity detected by
AB assay (i.e., -35%
survival without
inhibitor, -92% survival
with inhibitor); with a
48-hr MMAV + BSO
treatment, Z-DEVD-
FMK caused cytotoxicity
to be markedly reduced
(i.e., -7% survival
without inhibitor, -42%
survival with inhibitor).
Apoptosis detected by
the presence of DNA
ladders after agarose gel
electrophoresis: much
bigger ft at two higher
doses, which showed a
similar effect.




Reference



Sakurai et
al., 2005a













Sakurai et
al., 2005a






















Liao etal.,
2004




C-100  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Primary
keratinocytes
(in third
passage)
obtained from
foreskins of
adults












HeLa cells


























Arsenic
Species


As111 SA
















As111 ATO



























Concentration(s)
Tested (nM)


1,5,10

















2


























Duration of
Treatment


48 hr

















3 days


























LOECa
(HM)


Various

















2























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Protein levels detected
by Western blotting:
F ADD : ft at 1, bigger ft
at 5 and 10.
Caspase-8 (p!8, active):
ft at 1, huge ft at 5 and
10.
Caspase-3 (p20, active):
huge ft at 5 and 10.
Cleaved PARP (85 kD):
ft at 5 and 10; additional
experiments with and
without modulators
confirmed the
involvement of the Fas-
associated pathway in
inorganic arsenic-
induced apoptosis.
Induced apoptosis
(experimental - control)
detected by Annexin
V/PI flow cytometry:
-13% for inorganic
arsenic alone; -3% for
10 uM emodin alone;
-41% for inorganic
arsenic plus 10 uM
emodin; -14% for
inorganic arsenic with
both 10 uM emodin and
l.SmMNAC. Other
experiments showed that
the effect of emodin in
enhancing inorganic
arsenic-induced
apoptosis involved a
decrease of
mitochondria! membrane
potential. Emodin was
used because it has a
semiquinone structure
that is likely to increase
the generation of
intracellular ROS.




Reference


Liao etal.,
2004
















Yi et al.,
2004






















C-101  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



HeLa cells

























AR230-S
cells,
AR230-r
cells,
KCL22-S
cells, KCL22-
r cells










Arsenic
Species



As111 ATO


























As111 ATO















Concentration(s)
Tested (nM)



2



























1














Duration of
Treatment



3 days



























24 hr














LOECa
(HM)



2



























None











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis
(experimental - control)
detected by Annexin V-
FITC/PI flow cytometry:
27.0% for inorganic
arsenic alone; 6.9% for
30 uM emodin alone;
44. 1% for inorganic
arsenic plus 30 uM
emodin; 20.4% for
inorganic arsenic with
both 30 uM emodin and
l.SmMNAC. Emodin
and inorganic arsenic
synergistically interacted
to greatly ft the ROS
level and to cause
cytotoxicity.
Pretreatment or co-
treatment with NAC
blocked the synergism
for both effects. A 2uM
inorganic arsenic
treatment of 90 min
caused an ft in ROS to
~2.0x (with wide
confidence limits) and, in
a treatment lasting 48 hr,
about 20% cytotoxicity.
Apoptosis detected by
Annexin V-FLUOS
staining kit and flow
cytometry: NSE in any
of the 4 cell lines with
ATOorlOOuMBSO
treatments alone. For the
combined treatment,
induced rates
(experimental - control)
were: AR230-S, -35%;
AR230-r, -35%;
KCL22-S, -10%;
KCL22-r, -13%.




Reference



Wang et
al., 2005


























Konig et
al., 2007










C-102  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


AR230-r
cells,
KCL22-r cells















U-937 cells

NB4 cells

HL-60 cells













Arsenic
Species


As111 ATO


















As111 ATO
for all















Concentration(s)
Tested (nM)


1

















1, 2, 4, 8

0.5, 1,2,4

1,2,4













Duration of
Treatment


24 hr


















24 hr
for all















LOECa
(HM)


None

















4

1

2










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Western blot analyses:
inorganic arsenic alone
caused NSE on protein
levels of tyrosine
phosphorylated Bcr-Abl
or total cellular Bcr-Abl
in either cell line. In
both cell lines, combined
treatment of inorganic
arsenic with 100 uM
B SO yielded huge U in
both proteins. In non-
imatinib resistant CML
cells, unlike in these 2
imatinib-resistant cell
lines, inorganic arsenic
alone had been shown to
suppress Bcr-Abl
activity.
Induced apoptosis
(experimental - control)
based on chromatin
fragmentation:
U-937 cells: 1, NSE; 2,
-2%; 4, -14%; 8, -85%.
NB4 cells: 0.5, NSE; 1,
-5%; 2, -33%; 4, -63%.
HL-60 cells: 1, NSE; 2,
-5%; 4, -22%.
Induction of apoptosis
was potentiated by co-
treatment with PI3K
inhibitors LY294002 and
wortmannin, and by the
Akt inhibitor Akt,5.




Reference


Konig et
al., 2007



















Ramos et
al., 2005










C-103  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




U-937 cells
























HK-2 cells













Arsenic
Species



As111 ATO
























As111 SA


Asv











Concentration(s)
Tested (nM)




4
























0.1, 1, 10
for both












Duration of
Treatment




Various
























6, 24 hr













LOECa
(HM)




4






















0.1 at 24
hr

Prob-
ably 1 at
24 hr







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
U in Akt phosphorylation
after 24 hr (not by 14 hr);
ft in caspase 3 activity to
~3x after 24 hr; ft in
cytochrome c protein
(released from
mitochondria) after 14
hr; big ft in activated
Bax after 14 hr; big ft in
HSP 27 after 14 and 24
hr;
big ft in HSP 70 after 14
and24hr. The
potentiation of apoptosis
by inhibitors mentioned
in prior row involved
more extreme changes in
the same direction for p-
Akt, caspase 3,
cytochrome c, and Bax
activation as well as
attenuation of HSP27
expression. It also
involved increased
disruption of the
mitochondria!
transmembrane potential.

To assess mitochondria!
function, depolarization
of mitochondria!
membrane was detected
using MitoTracker Red,
a mitochondrion
selective dye. Effect of
dose of lof As111
appeared equivalent to
that of dose of 10 of Asv
Effect increased with
dose and time.




Reference




Ramos et
a!., 2005























Peraza et
a!., 2006









C-104  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




HK-2 cells
























APL primary
cells

K562 cells
NB4 cells



Thymocytes
from adult
male
BALB/cByJ
mice












Arsenic
Species




As111 SA

























As111 ATO





As111 ATO

Asv















Concentration(s)
Tested (nM)




0.1, 1, 10,25

























3






5
for both















Duration of
Treatment




24 hr

























24 hr






3, 10, 22 hr
for both















LOECa
(HM)




0.1

























3 for all





None


None











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis
(experimental - control)
detected by Annexin V-
FITC/PI flow cytometry:
0.1, -36%; 1, -23%; 10,
-15%; 25, -15%.
Induced necrotic cells
(experimental - control)
detected by same
method: 0.1, -2.5%; 1,
-3%; 10, -6%; 25,
-24%. Apoptotic cells
detected in this way were
said to be in early
apoptosis. Examination
by transmission electron
microscopy showed that
most such cells failed to
complete apoptosis and
ultimately underwent
necrosis instead. They
suggested that inorganic
arsenic was so toxic to
mitochondria that they
lost "their ability to keep
the cell on course for
apoptotic cell death."
Apoptosis rates (control
rates were not provided),
detected by FITC-
annexin V and PI
double-staining:
52.2%
27.6%
56.6%
NSE at any time point
for induction of
apoptosis by any of the
following types of
analysis: (1) "Annexin
V-FITC positive"
without loss of
membrane impermeance
(i.e., "7-AAD negative")
to identify early
apoptotic cells, (2) DNA
loss, and (3) both
"Annexin V-FITC
positive" and "7-AAD
positive" for cells in the
final stages of cell death.




Reference




Peraza et
al., 2006
























Sahu and
Jena, 2005





Mondal et
al., 2005












C-105  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Jurkat cells

Namalwa
cells

NB4 cells

U937 cells








Jurkat cells

Namalwa
cells

NB4 cells

U937 cells


NB4 cells


U937 cells























Arsenic
Species


As111 ATO
for all














As111 ATO
for all






As111 ATO
for both

























Concentration(s)
Tested (nM)


1,2
for all














2
for all







1, 2, 4, 6
for both
























Duration of
Treatment


24 hr
for all














24 hr
for all







24 hr
for both
























LOECa
(HM)
None

2

1

None









None

2

2

None



1


4




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis
(experimental - control)
detected by fluorescence
microscope analysis after
staining with AO and
EB:
Namalwa cells: 1, -1%;
2, -16%.
NB4 cells: 1, -12%; 2,
-26%.
NSE at dose of 2 in
Jurkat and U937 cells.
Pretreatment with NAC
or Z-VAD-FMK blocked
induction of apoptosis in
Namalwa and NB4 cells.
Western blot analysis:
ft in PARP-cleavage and
U in procaspase-3 level
in both Namalwa and
NB4 cells but not in the
other two cell lines;
inorganic arsenic did not
induce JNK
phosphorylation.
Induced apoptosis
(experimental - control)
detected by fluorescence
microscope analysis after
staining with AO and
EB:
NB4 cells: 1, -6%; 2,
-30%; 4, -70%; 6, 85%.
U937 cells: 1, -0%; 2,
-4%; 4, -15%; 6, 12%.
NB4 cells showed more
severe cell growth
inhibition at doses of >2.
Also, Western blot
analysis showed that
inorganic arsenic
induced PARP cleavage
in a dose-dependent
pattern in NB4 cells. In
U937 cells there was
only very slight PARP
cleavage at the highest
dose. JNK
phosphorylation did not
occur in either cell line.




Reference



Chenetal.,
2006














Chenetal.,
2006







Chenetal.,
2006




















C-106  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

MEFs that are
wt


MEFs that are
DKOs for
Bax and Bak



















MEFs that are
wt or DKOs
for Bax and
Bak






Namalwa
cells

NB4 cells















Arsenic
Species


As111 ATO
both























As111 ATO









As111 ATO
for both

















Concentration(s)
Tested (nM)


10
for both























10, 125, 500, 1000









1
for both

















Duration of
Treatment


8hr
for both

























LOECa
(HM)

10


None





















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Various indicators of
apoptosis:
Induced (experimental -
control) DNA
fragmentation: wt, -7%;
DKO, NSE.
Cytochrome c release: ft
in wt, NSE in DKO.
Induced caspase-3
activity: wt, -140 units;
DKO, none. Caspase-3
activity was only
detected in DKO cells
when they were
permeabilized and
incubated for 1 hr in the
presence of 4 uM
exogenous cytochrome c.
These and other
experiments showed that
mitochondria! events
associated with apoptotic
cell death induced at
concentrations such as
10 or less required Bax
and/or Bak.
Results from several experiments suggested that
extramitochondrial thiol oxidation leading to changes in
intracellular Ca2+ compartmentalization plays a critical
role in inorganic arsenic-induced cytochrome c release.
At concentrations of 125 and higher, Bax and Bak
became irrelevant to the mechanism of cytotoxicity and
cell death resulted from oxidative stress that led to
necrosis. ROS seem to be implicated in a
concentration-dependent mechanistic switch between
apoptosis and necrosis.

24 hr
for both














1
for both
without
BSO













Induced apoptosis
(experimental - control)
detected by fluorescence
microscope analysis after
staining with AO and
EB:
Namalwa cells: inorganic
arsenic, -6%; inorganic
arsenic + 10 uM BSO,
-29%.
NB4 cells: inorganic
arsenic, -8%; inorganic
arsenic + 10 uM BSO,
-47%.
BSO treatments
markedly reduced GSH
levels.




Reference



Nuttetal.,
2005























Nuttetal.,
2005








Chenetal.,
2006













C-107  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Jurkat cells

U937 cells













Jurkat cells

Namalwa
cells

NB4 cells

U937 cells





Jurkat cells

U937 cells








Arsenic
Species

As111 ATO
for both















As111 ATO
for all










As111 ATO
for both









Concentration(s)
Tested (nM)

1
for both















1
for all










1
for both









Duration of
Treatment

48 hr
for both














24 hr
for Namalwa
and NB4 cells,
48 hr for other
2 lines








Various, for
6-72 hr









LOECa
(HM)
None
for both
without
BSO













1 for all
with BSO











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis
(experimental - control)
detected by fluorescence
microscope analysis after
staining with AO and
EB:
Jurkat cells: inorganic
arsenic, NSE; inorganic
arsenic + 10 uM BSO,
-25%.
U937 cells: inorganic
arsenic, NSE; inorganic
arsenic + 10 uM BSO,
-67%.
BSO treatments
markedly reduced GSH
levels.
Results of Western blot
analysis in all 4 cell lines
following co-treatment
of inorganic arsenic with
lOuMBSO:
Big ft in PARP-cleavage;
big U in procaspase-3
level.
Big ft in JNK
phosphorylation (the
latter effect was not seen
in absence of BSO co-
treatment).
Time course experiments for co-
treatment with 10 uM BSO showed ft
in PARP-cleavage;
U in procaspase-3 level; strong ft in
JNK phosphorylation. Induced
apoptosis increased to -85% and
-50% by 72 hr in U937 and Jurkat
cells, respectively.




Reference


Chenetal.,
2006
















Chenetal.,
2006








Chenetal.,
2006






C-108  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


U937 cells

















U937 cells























Arsenic
Species

As111 ATO


















As111 ATO























Concentration(s)
Tested (nM)

1


















1























Duration of
Treatment

48 hr


















48 hr























LOECa
(HM)
1, but only
with B SO
co-treat-
ment
















1, but only
with B SO
co-treat-
ment

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis
(experimental - control)
detected by fluorescence
microscope analysis after
staining with AO and
EB: -55% following the
co-treatment with B SO;
this ft was not
significantly decreased
by 4-hr treatments with
either 10 mM NAC or
200 units of catalase
even though those
treatments substantially
decreased H2O2 levels.
Moreover, NAC and
catalase did not block the
JNK activation caused
by the inorganic arsenic
+ B SO treatment.
Results of Western blot
analyses: huge ft inDR5,
huge U in Bid, and U in
IicBa following co-
treatment with 10 uM
BSO;NSEonthese3
proteins after inorganic
arsenic or BSO alone.
Experiments with
inhibitors suggested that
(1) both caspase- and
JNK-mediated pathways
(due to activation of NF-
KB) participate in the
induction of apoptosis
that occurs following co-
treatment with inorganic
arsenic and BSO and (2)
that JNK increases DR5
protein levels that in turn
mediate that apoptosis.




Reference


Chenetal.,
2006
















Chenetal.,
2006



















C-109  DRAFT—DO NOT CITE OR QUOTE

-------
   Type of
 Cell/Tissue
 Arsenic
 Species
 Concentration(s)
   Tested QaM)
Duration of
 Treatment
LOECa
 Results (Compared
With Controls, With
 All Concentrations
       Being
in nM Unless Noted)
                                   Reference
  NB4 cells,
NB4-AsR, and
 APL primary
     cells
As111 ATO
  A series of experiments was conducted involving 24-72 hr treatments with
 concentrations of inorganic arsenic of 0.125-10 uM. Tests of MEK1 mRNA
 knockdown using inorganic arsenic treatments and MEK1 inhibitors (namely,
 PD98059 at 40 uM and PD184352 at 1 uM) showed that MEK1 inhibitors and
 inorganic arsenic synergize to induce apoptosis.  Although inorganic arsenic
 induces apoptosis, it also causes ERK1/2 activation, which tends to decrease
     the extent of apoptosis by causing phosphorylation at Serll2 of the
       proapoptotic Bad protein.  Phosphorylated Bad protein does not
  heterodimerize with the Bel proteins. The only known function of the Bad
 protein is to bind (i.e., heterodimerize) with the death antagonist Bcl-2 family
  proteins, Bcl-2 and Bcl-xL, thereby blocking their antiapoptotic action by
 preventing them from binding to Bax/Bak. Because MEK1 inhibitors block
    this ERK1/2 activation and the phosphorylation of BAD, there is more
 nonphosphorylated Bad protein to heterodimerize with the Bcl-2 proteins and
   keep them from functioning to  block apoptosis. In this way, exposure to
  inorganic arsenic in the presence of MEK1 inhibitors greatly increases the
	extent of apoptosis.	
                                                    Lunghi et
                                                    al., 2005
Primary AML
blasts from 25
 patients with
  non-APL
    AML
As111 ATO
    In experiments involving 48-hr treatments that used concentrations of
       inorganic arsenic of 0.125-10 uM in the presence or absence of
  concentrations of the MEK1 inhibitor PD184352 of 0.1-10 uM, synergistic,
 additive, or antagonistic interactions in the induction of apoptosis were found
 in primary cells from 13, 8, and 4 patients, respectively. The p53-related gene
  p73 was shown to be the molecular target of importance in this interaction,
 and the synergism had the following basis.  Inorganic arsenic induced both the
     proapoptotic and antiproliferative TAp73 and the antiapoptotic and
 proproliferative ANp73 isoforms, with no net effect on apoptosis because the
 TAp73/ANp73 ratio did not change. The MEK1 inhibitor reduced the level of
  ANp73 and blunted the inorganic arsenic-induced up-regulation of ANp73,
    with the result that the TAp73/ANp73 ratio increased, leading to more
 apoptosis. At 1  uM, inorganic arsenic induced only p73, but at doses >2 uM)
    it also promoted accumulation of p53 protein levels, which also caused
	apoptosis.	
                                                                                   Lunghi et
                                                                                    al., 2006
                                                   C-110   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



CHO Kl cells
























Normal
human
keratinocytes
treated with
50 mJ/cm2
UVB before
or after
inorganic
arsenic
treatment














Arsenic
Species



As111 SA

























As111 SA
for both





















Concentration(s)
Tested (nM)



20, 40, 80
























1, as pretreatment


1, as post-
treatment
begun 24 hr after
irradiation

















Duration of
Treatment



4hr


























24 hr for both





















LOECa
(HM)



20
























1


None

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis detected by
flow cytometry and by
the presence of DNA
ladders from
internucleosomal DNA
degradation — ladder
effect did not appear
until 24 hr after
treatment. At dose of 40,
it took 8 hours after
treatment before
apoptosis could be
detected by flow
cytometry. Reduced
levels of apoptosis
resulted from treatment
with various modulators
(antioxidants, a copper
ion chelator, a protein
kinase inhibitor, and a
protein synthesis
inhibitor) either
simultaneously or, in
some instances,
immediately following
the arsenic treatment.
Apoptosis as detected by
PI staining and TUNEL
assay: the inorganic
arsenic treatment alone
did not induce a
significant increase in
apoptosis or cytotoxicity;
U in the level of UV-
induced apoptosis to
control levels, with a
corresponding U in
cytotoxicity to control
levels.
A similar amount of
apoptosis was seen as
with UVB alone, or
possibly apoptosis
increased slightly;
cytotoxicity was similar
to that with UVB
treatment alone or
possibly slightly more
extreme.




Reference



Wanget
al., 1996

























Chenetal.,
2005b

















C-111  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
A mouse
fibroblast cell
line as well as
various stable
transfectants
ofJB6C141
cells







NB4 cells

U937 cells

HL-60 cells

















Mouse
291.03C
keratinocytes




Arsenic
Species
As111 SA

Asv












As111 ATO
for all



















As111 SA
for both





Concentration(s)
Tested (nM)
Various















2 for all



















5

5




Duration of
Treatment
—














2 days
for all



















48 hr

60 hr




LOECa
(HM)
—













2

2

2

















5

5

Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Various tests indicated
that p53 is not involved
in arsenic -induced
apoptosis. The pathway
of JNKs was shown to
play an essential role in
arsenic-induced
apoptosis. For example,
such apoptosis was
blocked by expression of
the dominant-negative
mutant of JNK1.
Percentages of apoptosis
determined by
fluorescent microscopy,
and units of basal
activity of GSTjc, GPx,
and CAT, respectively:
67.5%, 94.0, 28.3, 25.8.
5.6%, 212.1,67.6, 170.5.
5.8%, 138.6, 55.5, 198.3.
These data and others
showed that the higher
the basal levels of these
3 enzymes, the less the
inorganic arsenic-
induced apoptosis.
Higher activities of these
enzymes decrease the
amount of H2O2 in cells.
Modulators that increase
activities of these
enzymes were shown to
decrease apoptosis and
vice versa.
Apoptosis measured by
flow cytometry:
ft by 4.20% over control,
which was 0.74%.
ft by 7.3 1% over control.




Reference
Huang et
al., 1999b















Jing et al.,
1999


















Wuetal.,
2005

C-l 12  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


Mouse
291.03C
keratinocytes
irradiated
immediately
after the
arsenic
treatment with
a single dose
of 0.30k J/m2
UV









HaCaT cells
(immortalized
,
non-
tumorigenic
human
keratinocyte
cell line)

arsenic-TL
cells (arsenic-
tolerant cells,
which are
HaCaT cells
that were
cultured for
28 weeks in
100 nM As111
SA)














Arsenic
Species




As111 SA
for all




















As111 SA
for both



























Concentration(s)
Tested (nM)


None (i.e., UV
only)

2.5

5.0


















20, 40, 60, 80 for
both



























Duration of
Treatment


—

24 hr

24 hr



















24 hr
for both



























LOECa
(HM)


—

2.5

5.0

















20



40























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis measured by
flow cytometry 24 hr
after the dose of UV:

ft by 26.87% over
control, which was
0.74%.

ft by 20.62% over
control.

ft by 9.78% over control.
Thus, both pretreatments
with As111 SA markedly
reduced the amount of
UV-induced apoptosis.
In parallel with the
above, UV-induced
caspase 3/7 activity was
also decreased by both
treatments.
Apoptosis detected using
flow cytometry
following staining with
Annexin V and PI:

ft in apoptosis.



Much smaller ft in
apoptosis. There was a
significant decrease in
apoptosis compared to
HaCaT cells at all 4 dose
levels. A similar
resistance by arsenic-TL
cells was seen to
apoptosis induction by
25J/cm2ofUVA, as
well as by cisplatin,
etoposide, and
doxorubicin. Arsenic-
TL cells showed greatly
increased stability of
nuclear P-PKB, and
pretreatment with
chemicals that inhibit
PKB phosphorylation
blocked inorganic
arsenic-induced acquired
apoptotic resistance.




Reference




Wuetal.,
2005




















Pi et al.,
2005
























C-113  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
MCF-7 cells


U-2OS cells











Arsenic
Species
As111 ATO


As111 SA











Concentration(s)
Tested (nM)
3


0.1,1, 10











Duration of
Treatment
36 hr


24 hr











LOECa
(HM)
3


0.1










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis detected based
on electrophoretic
analysis of DNA
fragmentation:
-18% of the cells were
apoptotic.
TUNEL staining assay
was used to detect
apoptotic cells after 0,
24, or 48 hr of post-
treatment culturing in
arsenic -free medium.
At dose of 0.1, apoptotic
cells were -0%, -0.3%,
and -3.6%, respectively.
At dose of 1, apoptotic
cells were -0%, -0.2%,
and -3.4%, respectively.
At dose of 10, apoptotic
cells were -0%, -0%,
and -0%, respectively.
Note that a 24-hr
treatment with SA
affected apoptosis only if
there was an additional
24-hr or longer period of
culturing in SA-free
medium between the end
of the SA treatment and
when the assay was
done.

Reference
Ling et al.,
2002


Komissaro
va et al.,
2005










C-l 14  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



U-2OS cells






















Undifferentiat
edPC12 cells










PARP-1+/+
MEF cells

PARP-I-'-
MEF cells






Arsenic
Species



As111 SA






















As111 ATO











As111 SA
for both









Concentration(s)
Tested (nM)



0.1, 1, 10






















8











11.5,23
for both









Duration of
Treatment



24 hr






















24 hr











24 hr
for both









LOECa
(HM)



0.1






















8











11.5

11.5





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Assay utilizing
activation of cellular
caspase-3 was used to
detect apoptotic cells
after 0, 24, or 48 hr of
post-treatment culturing
in arsenic-free medium:
At dose of 0.1, apoptotic
cells were -0%, -1.3%,
and -6.2%, respectively.
At dose of 1, apoptotic
cells were -0%, -0.3%,
and -5.4%, respectively.
At dose of 10, apoptotic
cells were -0%, -0%,
and -0%, respectively.
Note that a 24-hr
treatment with SA
affected apoptosis only if
there was an additional
24-hr or longer period of
culturing in SA-free
medium between the end
of the SA treatment and
when the assay was
done.
Induction of apoptosis
detected by annexin V
binding and caspase
activity:
-55% of cells with
apoptotic death, rest with
necrotic death; at 6 hrs,
-60% of dead cells were
apoptotic.
Induction of apoptosis
detected by PI and
RNase staining and flow
cytometry, visualized as
sub-Gl population and
reported as % of
apoptosis
(controls were always
-6% at 11. 5, -9% at 23
-11% at 11.5, -21% at
23.




Reference



Komissaro
va et al.,
2005




















Pigaetal.,
2007









Poonepalli
etal.,2005







C-115  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



PARP-1+/+
MEF cells

PARP-r7'
MEF cells





JB6C141
cells,
transfected
with IKKP-
KM to greatly
reduce COX-
2 induction
JB6C141
cells
transfected
with vector
only



JB6C141
cells, after
knockdown of
endogenous
COX-2
expression to
low levels by
its specific
siRNA
JB6C141
cells
transfected
with mock
vector for the
siRNA, with
normal COX-
2 expression







Arsenic
Species



As111 SA
for both










As111 SA
for both















As111 SA
for both


















Concentration(s)
Tested (nM)



11.5,23
for both










20,40
for both















10,20
for both


















Duration of
Treatment



48 hr
for both










24 hr
for both















36 hr
for both


















LOECa
(HM)



11.5

11.5








20



20










10





10













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of apoptosis
detected by PI and
RNase staining and flow
cytometry, visualized as
sub-Gl population and
reported as % of
apoptosis
(controls were always
0 /o):
-23% at 11. 5, -32% at
23.
-40% at 11. 5, -62% at
23.
Induction of apoptosis
detected by PI staining
and flow cytometry:
ftft in apoptosis: medium
alone, 0.83%; 20,
12.60%, 40, 41.33%;.


Slight ft in apoptosis:
medium alone, 1.03%;
20, 4.58%, 40, 7.23%.
Similar conclusion was
reached using TUNEL
assay and flow
cytometry.
Induction of apoptosis
detected by PI staining
and flow cytometry:
ftft in apopthosis:
medium alone, 4.14%;
10, 28.45%, 20, 49.22%.




Much smaller ft in
apoptosis: medium
alone, 1.86%; 10,
10.52%, 20, 26.60%.

Another experiment
showed that pretreatment
of normal JB6 C141 cells
with NS398, an inhibitor
of COX-2, markedly ft
amount of apoptosis.




Reference


Poonepalli
etal.,2005











Ouyang et
al., 2007















Ouyang et
al., 2007















C-116  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

MEF cells
that were
made IKKp-A
so that they
markedly
overexpressed
COX-2
MEF cells
that had the
vector only,
with normal
(low) level of
COX-2


SY-5Y cells

HEK 293
cells











Arsenic
Species




As111 SA
for both










As111 ATO
for both













Concentration(s)
Tested (nM)




20
for both











1
for both












Duration of
Treatment




36 hr
for both











24 hr
48 hr
72 hr











LOECa
(HM)


20



20










1
for all









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of apoptosis
detected by PI staining
and flow cytometry:

Slight ft in apoptosis:
medium alone, 0.68%;
20, 6.35%.

Big ft in apoptosis:
medium alone, 0.87%;
20, 49.62%.
Thus, COX-2 protects
cells from apoptosis.

Induction of apoptosis
detected by Hoechst
staining:
Response as % of control
in SY-5Y and HEK 293
cells, respectively, for
each duration of
treatment:
266%, 156%.
152%, 192%.
214%, 200%.
There was NSE on the
mitotic index at any
time.




Reference




Ouyang et
al., 2007











Florea et
al., 2007









C-117  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
PMs from
CDFj mice










TK6 cells



Arsenic
Species
As111 SA
Asv
MMAV
DMAV
TMAV








As111 SA
As111 ATO



Concentration(s)
Tested (nM)
10
ImM
10 mM
10 mM
10 mM








0.1, 1
for both



Duration of
Treatment
48 hr
for all










24 hr
for both



LOECa
(HM)
10
ImM
10 mM
10 mM
None








1
1


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis detected based
on electrophoretic
analysis of DNA
fragmentation and by
TUNEL staining. The
particular assay shown in
this row used cellular
morphological changes
to assess apoptosis and
the AlamarBlue assay to
measure cell death.
Approximate resulting
percentages of cell death
(listed first) and
apoptotic cells (listed
second) for the 5
compounds follow:
For As111 SA: 82% and
23%.
For Asv: 65% and 17%.
For MMAV: 10% and 7%.
For DMAV: 100% and
100%.
ForTMAv: 12% and
none.
Thus DMAV was unusual
in causing almost
entirely apoptotic cell
death, while the
inorganic arsenicals
caused mainly necrotic
cell death.
Apoptosis identified
using APO2.7 antibody:
ft to 5.0% from 3.6% in
control.
ft to 5.5% from 3.6% in
control.

Reference
Sakurai et
al., 1998










Hornhardt
etal.,2006


C-118  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

TK6 cells
irradiated
with 1,2, or 4
Gyof69
cGy/min
gamma
radiation at
beginning of
inorganic
arsenic
treatment






HCT1 16 cells
(securin +/+)

HCT1 16 cells
(securin -/-)









Arsenic
Species

As111 SA


As111 ATO














As111 SA
for both











Concentration(s)
Tested (nM)


0.1, 1
for both















16
for both











Duration of
Treatment


24 hr
for both















24 hr
for both











LOECa
(HM)

None


1













16


16







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Apoptosis identified
using APO2.7 antibody:
At dose of 1: 1 Gy,
9.1%;2Gy, 10.4%, 4
Gy, 22.6%; SA had no
significant effect on any
of them.

At dose of 1: 1 Gy,
12.5%; 2 Gy, 21.75%, 4
Gy, 38.6%; ATO caused
a significant increase
over the control (no
inorganic arsenic +
radiation) at all 3
radiation doses. This
was a synergistic
interaction.
Induced apoptosis (i.e.,
experimental - control)
detected using
fluorescent microscopy
after Hoechst staining:
securin +/+: -6%;
securin -/-: -10%;
with the amount of
apoptosis in the null
mutant being
significantly higher.




Reference



Hornhardt
etal.,2006















Chao etal.,
2006a







C-119  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
NB4 cells
NB4-M-AsR2
cells
DVI9 cells












Gclm"7" MEF
cells, from
GCLM
knockout
mice




Arsenic
Species
As111 ATO
for all













As111 SA
for both
durations





Concentration(s)
Tested (nM)
0.5, 1
for all













25
for both
durations





Duration of
Treatment
48 hr
for all













8hr
24 hr





LOECa
(HM)
0.5
1
0.5













25
for both
dura-tions





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis (i.e.,
experimental - control)
for ATO alone and for
ATO with 100 uM
Trolox, detected using PI
staining in binding
buffer:
At 0.5: -6% alone, -20%
with Trolox; at 1: -16%
alone, -55% with
Trolox.
At 0.5:0% alone, -11%
with Trolox; at 1: -14%
alone, -45% with
Trolox.
At 0.5: -1.5% alone,
-4% with Trolox; at 1:
-6% alone, -20% with
Trolox.
Additional support for
the conclusion that
Trolox enhanced ATO-
mediated apoptosis was
provided by an annexin
V-FITC staining assay
and by the observation
that Trolox significantly
enhanced the percentage
of cells with activated
caspase-3 and cleaved
PARP.
Induced apoptosis (i.e.,
experimental - control)
detected by staining with
FITC-labeled annexin-V
and PI:
At 8 hours: -5% early
apoptotic, -38% late
apoptotic, -8% necrotic.
At 24 hours: -3% early
apoptotic, -79% late
apoptotic, -5% necrotic.
Experiments in Gclm+/+
cells showed that co-
treatment or pretreatment
with tBHQ partially or
completely blocked
inorganic arsenic-
induced apoptosis.

Reference
Diaz et al.,
2005













Kannet
al., 2005b




C-120  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


MEFs













MEFs that
were wt


MEFs that
were Bax" "
and Bak"7"
double
knockout
(DKO) cells












Arsenic
Species

As111 ATO
















As111 ATO
for both


















Concentration(s)
Tested (nM)


2,3,5















10
for both


















Duration of
Treatment


3 days















12 hr
for both


















LOECa
(HM)


2














10


10














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis (i.e.,
experimental - control)
for ATO alone and for
ATO
co-treatment with
Trolox, detected by PI
staining using flow
cytometry:
ATO alone: 2, -9%; 3,
-22%; 5, -62%.
ATO and Trolox: 2,
-3%; 3, -3%; 5, -20%.
Thus, in contrast to what
happened in malignant
cells, Trolox blocked the
effects of ATO.
Induced apoptosis (i.e.,
experimental - control)
detected by PI staining
and F ACS:
-23% in wt and -7% in
DKO; the results at dose
of 500 are ignored here.
wt: large ft in release of
cytochrome c, which was
mostly blocked by
pretreatment with
BAPTA-AM; DKO:
trace ft in release of
cytochrome c.
Results showed that
cytochrome c release and
apoptosis occurred
largely via a Bax/Bak-
dependent mechanism.




Reference


Diaz et al.,
2005














Bustamant
e et al.,
2005














C-121  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Isolated rat
liver
mitochondria
loaded with
Ca2+














SVEC4-10
cells





RAW264.7
cells




RAW264.7
cells















Arsenic
Species

As111 ATO


















As111 SA






As111 SA





As111 SA
















Concentration(s)
Tested (nM)

10, 50, 100


















20






5,25





5,25
















Duration of
Treatment

2 min


















24 hr






24 hr





24 hr
















LOECa
(HM)

10


















20






5





5













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
There was a dose-
dependent, cyclosporin
A-sensitive release of
cytochrome c via
induction of
mitochondrial
permeability transition
and subsequent swelling
of mitochondria.
Mitochondrial GSH did
not seem to be a target
for inorganic arsenic
which, however,
appeared to cause
oxidative modification of
thiol groups of pore-
forming proteins, notably
adenine nucleotide
translocase.
Induced apoptosis (i.e.,
experimental - control),
apoptotic cells were
counted by
hemocytometer in a
fluorescence microscope:
-68%.
Apoptosis detected by
TUNEL assay; results
were presented as mean
densities of TUNEL
staining: there was a
positive dose-response.
Apoptosis detected by
fluorescence staining of
caspase-3 activation:
there was a positive
dose-response. A 30-
min pretreatment with
DPIC (which inhibits
H2O2 production)
completely blocked
caspase-3 activation at
both inorganic arsenic
doses, thus showing that
it prevented induction of
apoptosis by inorganic
arsenic.




Reference

Bustamant
e et al.,
2005
















Hsu et al.,
2005




Szymczyk
etal.,2006





Szymczyk
etal.,2006












C-122  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


NIH 3T3 cells





















HL-60 cells





















Arsenic
Species


As111 SA




















As111 ATO






















Concentration(s)
Tested (nM)


5, 10, 20, 50, 100,
200




















3





















Duration of
Treatment


6hr





















48 hr





















LOECa
(HM)


10





















3


















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of caspase 3/7
activity assayed using
Caspase-Glo™ assay (an
indicator of apoptosis):
units of activity at 0, 10,
50, 100, and 200 were
about 2.5, 4, 12, 17, and
36, respectively. Pre-
induction of HSP by
conditioning heat shock
(2 hr at 42°C on prior
day) or by constitutive
expression of HSP70
markedly reduced the
induction, as follows:
With heat: NSE at any
dose.
With constitutive
expression: at most a hint
of induction at highest 3
doses.
Induced apoptosis (i.e.,
experimental - control),
based on TUNEL assay:
15%. Effect of
intracellular AA (icAA):
(cells were loaded with 4
mM icAA by incubating
them with DHA prior to
inorganic arsenic
treatments, thus avoiding
generation of
extracellular ROS in
tissue culture media
caused by direct addition
to it of AA)
Induced apoptosis for
inorganic arsenic + icAA
= 1% (NSE).
Results using annexin
V/FITC assay gave a
consistent but milder
effect.




Reference


Khalil et
al., 2006




















Karasawa
s et al.,
2005
















C-123  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

H22 cells

BAEC cells















NB4 cells
















NB4 cells

















Arsenic
Species

As111 ATO
for both















As111 ATO
















As111 ATO


















Concentration(s)
Tested (nM)

0.5, 1,2,4
for both















3

















1

















Duration of
Treatment

24 hr, 48 hr

24 hr, 48 hr














48 hr

















48 hr

















LOECa
(HM)

1,0.5

1,1














3

















1














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis index
(i.e., experimental -
control), based on
TUNEL assay:
H22, 24 hr: 0.5, NSE; 1,
-8%; 2, -22%; 4, -35%.
H22, 48 hr: 0.5, -8%; 1,
-20%; 2, -36%; 4,
-45%.
BAEC, 24 hr: 0.5, NSE;
1, -6%; 2, -22%; 4,
-26%.
BAEC, 48 hr: 0.5, NSE;
1, -8%; 2, -28%; 4,
-40%.
% of cells with nuclear
fragmentation (NuFr):
-80%.
Effects of modulators at
high doses:
Co-treatments with
1000^000 uM DTT:
dose-related marked U in
NuFr reaching -20%.
Co-treatments with 100-
400 uM DMSA: dose-
related marked U in NuFr
reaching -20%.
Co-treatments with 50-
200 uMDMPS: dose-
related marked U in NuFr
reaching -27%.
% of cells with NuFr:
-20% for experiments
with DTT and DMSA;
about 12% in experiment
withDMPS.
Effects of modulators at
low doses:
Co-treatments with 12.5-
50 uM DTT: dose-
related marked ft in NuFr
reaching -90%.
Co-treatments with 10-
40 uM DMSA: dose-
related marked ft in NuFr
reaching -75%.
Co-treatments with 5-20
uM DMPS: dose-related
marked ft in NuFr
reaching -80%.




Reference


Liuetal.,
2006e















Jan et al.,
2006















Jan et al.,
2006













C-124  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



293 cells





















SV-HUC-1
cells


















Arsenic
Species



As111 ATO




















As111 ATO




















Concentration(s)
Tested (nM)



2




















2




















Duration of
Treatment



48 hr




















48 hr




















LOECa
(HM)



2




















2

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
% of cells with sub-Gl
DNA content: untreated
= -5%; dose of 2: big ft
to -53%.
Effects of co-treatment
(CoTr) with modulators
at high doses:
CoTr 200 uM DMSA: U
from inorganic arsenic
alone to -26%.
CoTr 100 uMDMPS:ll
from inorganic arsenic
alone to -37%.
Effects of CoTr with
modulators at low doses:
CoTr 20 uM DMSA: ft
from inorganic arsenic
alone to -83%.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -88%.
% of cells with sub-Gl
DNA content: untreated
= -6%; dose of 2: big ft
to -46%.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uM DMSA: U
from inorganic arsenic
alone to -22%.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -28%.
Effects of CoTr with
modulators at low doses:
CoTr 20 uM DMSA: ft
from inorganic arsenic
alone to -70%.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -72%.




Reference



Jan et al.,
2006




















Jan et al.,
2006















C-125  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


A549 cells

















A549 cells














WM9 cells
OM431 cells
LU1205 cells














Arsenic
Species

As111 ATO

















As111 ATO















As111 SA
for all















Concentration(s)
Tested (nM)


1, 2, 5, 10, 20, 50

















2















4















Duration of
Treatment


48hr

















48 hr















48 hr















LOECa
(HM)


5














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival determined
by MTT assay: LC50 =
-27.
Cell survival determined
by flow cytometry after
annexin V and PI
staining:
inorganic arsenic at dose
of 2: NSE.
200 uM sulindac: NSE.
(2 uM inorganic arsenic
+ 200 uM sulindac):
-40% cytotoxicity;
pretreatment with NAC
almost completely
blocked this synergistic
interaction.
Regarding caspase 3/7 protein levels:
2 uM inorganic arsenic: NSE.
200 uM sulindac: NSE.
(2 uM inorganic arsenic + 200 uM
sulindac): ft to ~1.4x.
Regarding caspase 9 protein levels:
2 uM inorganic arsenic: ft to 1.05x.
200 uM sulindac: NSE.
(2 uM inorganic arsenic + 200 uM
sulindac): ft to ~1.5x.
There was also a clear synergistic
interaction between these treatments in
causing big U of both procaspase-3
and procaspase-9 protein levels.
Pretreatment with NAC almost
entirely blocked the caspase 3/7 and
caspase 9 effects.


4












Induced apoptosis (i.e.,
experimental - control),
based on PI staining and
FACS analysis of hypo-
diploid content of DNA
in the pre-GO/Gl region:
WM9, ~32%;OM431,
-17%; LU1205, -18%.
Treatment with soluble
recombinant TRAIL was
effective in inducing
apoptosis; combined
treatment with inorganic
arsenic yielded no more
than an additive effect.




Reference


Jin et al.,
2006b
















Jinetal.,
2006b














Ivanov and
Hei, 2006











C-126  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Arsenic
Species



Concentration(s)
Tested (nM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)




Reference
Cancer Promotion


BALB/c 3T3
A3 1-1-1 cells
(derived from
mice)















BALB/c 3T3
A3 1-1-1 cells
(derived from
mice)










V79 cells












As111 SA

AsvDA

MMAV


DMAV











As111 SA

AsvDA

MMAV

DMAV




As111 SA


AsvDA


MMAV


DMAV








0.2, 0.5, 1, 2, 5

0.5, 1,2,5, 10

50, 100, 200, 500,
1000

10, 20, 50, 100,
200










1

5

500

50




0.15,0.3,0.7, 1.5,
2.5

0.5, 1.5,2.5,5, 10,
20

0.5, 1.5,2.5,5, 10,
20 mM

0.15,0.3,0.6, 1.3,
2.7, 5 mM










18 days for all


















18 days for all












72 hrs for all












0.5

1

200


None











1

5

500

None




0.7


5


5 mM


None








Caused promotion in a
two-stage transformation
assay; based on a
significant increase in
the number of
transformed cells after an
initiating treatment of
0.2 ug/mL MCA for 3
days followed by post-
treatment with an arsenic
compound for 18 days.
At doses above the
LOEC, the responses
increased no more than
slightly with dose. For
As111 SA there was a
humped dose-response
with a peak at the dose
ofl.
Caused promotion in a
two-stage transformation
assay; based on a
significant increase in
the number of
transformed cells after an
initiating treatment of
10 uM As111 SA for 3
days followed by post-
treatment with an arsenic
compound for 18 days.


Inhibited gap-junctional
intercellular
communication, which is
a mechanism linked to
many tumor promoters; it
is based on the metabolic
cooperation assay, which
detects chemicals that
inhibit the transfer of the
lethal metabolite of 6-
thioguanine from HPRT-
proficient to HPRT-
deficient cells, thereby
allowing recovery of the
6-thioguanine-resistant
(HPRT-deficient) cells.



Tsuchiya
etal.,2005

















Tsuchiya
etal.,2005










Tsuchiya
etal.,2005












C-127  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Arsenic
Species



Concentration(s)
Tested (nM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)




Reference
Cell Cycle Arrest or Reduced Proliferation
MGC-803
(human
gastric
cancer)



MC/CAR
(human
multiple
myeloma cell
line)










UROtsa cells









V79 cells














As111 ATO







As111 ATO









As111 SA

Asv

MMAmO

MMAV

DMAmI

DMAV



DMAV














0.01-1







1,2,3,4,5









0.1,0.5, 1,5

1,200

0.1,0.5, 1,5


1,200

0.1,0.5, 1,5

1,200


1,2, 5mM














24 hr







72 hr













24 hr
for all








12 hr














0.01







1









1

None

1


None

5

None


ImM














Growth inhibition
(growth measured by
MTT assay): at various
doses, growth inhibition
was also induced in 5
other human malignant
cell lines.
Growth inhibition
(growth measured by
MTT assay):
About 60% inhibition at
2; cells were arrested in
both Gl and G2-M
phases. Growth
inhibition was also
induced in 7 other human
multiple myeloma cell
lines to various degrees.
Extent of reduction of
cell proliferation based
on [3H]thymidine
incorporation:

Cell proliferation
reductions at dose of 5
were approximately as
follows: DMAmI, 15%;
As111, 30%; MMAmO,
85%.

Induction of mitotic
delay and formation of
aberrant mitotic spindles,
including tripolar and
quadripolar spindles:
-18% aberrant spindles
at 1 mM. y-tubulin was
co-localized with the
aberrant spindles. The
following things were
noted to occur after
exposure of V79 cells to
DMAV: multiple MTOCs,
multipolar spindles,
amoeboid cells,
multinucleated cells, and
cell death.
Zhang et
al., 1999






Parketal.,
2000












Drobna et
al., 2002








Ochi et al.,
1999a













C-128  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HeLa S3 cells















U937 cells


















Arsenic
Species

As111 SA














As111 SA
for both


















Concentration(s)
Tested (nM)

1, 3, 5, 10, 20














2.5,5, 10
for both
durations

















Duration of
Treatment

24hr













24 hr


48 hr

















LOECa
(HM)

o
5













2.5
for both
dura-tions















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cells arrested at mitotic
stage: At dose of 5, 35%
of cells were arrested in
that stage. Of 7 cell lines
tested in this way, two
others were almost as
sensitive to this effect.
Examination of cells
arrested in mitosis
showed abnormal mitotic
figures and spindles, as
well as deranged
chromosomal
congression.
Cell numbers counted
with a Coulter counter:
After 24 hr at the doses
of 2.5, 5, and 10, there
were approximately
71%, 56%, and 43% as
many cells as in the
control group,
respectively.
After 48 hr at the doses
of 2.5, 5, and 10, there
were approximately
54%, 38%, and 23% as
many cells as in the
control group,
respectively. There was
little if any cytotoxicity
even at 48 hr at the dose
of 5.




Reference

Huang and
Lee, 1998














McCollum
et al., 2005














C-129  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue




U937 cells

























Arsenic
Species




As111 SA

























Concentration(s)
Tested (nM)




5

























Duration of
Treatment




8hr

























Results (Compared
With Controls, With
All Concentrations
LOECa Being
(|oM) in |iM Unless Noted)
The LOEC was 5. Centrifugal
elutriation was used to enrich cells in
different phases of the cell cycle so
that the effect of inorganic arsenic
could be tested on them. Progression
of inorganic arsenic-treated cells from
each cell cycle stage to the next was
studied, and it was found that
inorganic arsenic slowed cell growth
in every phase of the cycle. For
example, in asynchronous populations
of untreated cells, DNA synthesis
lasted 10 to 12 hr. However, in cells
treated with 5 uM inorganic arsenic, it
lasted 16 hr. In the presence of
inorganic arsenic, cells in Gl entered
the S phase more slowly, etc. Cell
passage from any cell cycle phase to
the next was inhibited by 5 uM
inorganic arsenic arsenite. Clearly
there was not induction of cell-cycle
arrest at one specific checkpoint. The
biggest inorganic arsenic -induced
slowdown occurred between M and
Gl, and the next biggest was between
G2 and M. By looking at caspase
activity, they showed that inorganic
arsenic induced apoptosis specifically
in cell populations delayed in the
G2/M phase.
Reference




McCollum
et al., 2005
























C-130  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

PARP-1+/+
MEF cells


PARP-r7'
MEF cells




















CGL-2 cells




















Arsenic
Species


As111 SA
for both























As111 SA




















Concentration(s)
Tested (nM)


11.5,23
for both























1, 2, 3, 4, 5, 7, 10




















Duration of
Treatment


24 and 48 hr
for both























24 hr




















LOECa
(HM)

11.5
for both
at both
times






















1

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
inorganic arsenic caused
much disruption of cell
cycle as shown by PI and
RNase staining and flow
cytometry when
visualized as proportions
of cells that were in
G2/M, S, Gl,orsub-Gl
(i.e., apoptotic) under the
different conditions.
Disruption was more
extreme in PARP-l"'"
MEF cells. Results for
apoptosis, which are
easier to quantify, are
detailed in separate rows.
Especially at the highest
inorganic arsenic dose in
PARP-1"7" cells, the
proportion of G2/M cells
became especially small,
at least when the
comparison was made to
all cells and not just to
non-apoptotic ones.
Cell survival was
determined using a
colony-forming assay:
LC50= 1.7. arsenic
mitotic cells round-up,
they can be separated
from the attached
interface cells by using
the shake-off technique.
When that technique was
applied to a sample at the
dose of 2, 96% of the
attached cells were found
to be alive, and 96% of
the floating (i.e., mitotic)
cells were found to be
dead, thus indicating that
inorganic arsenic
induced mitosis-
mediated cell death.




Reference


Poonepalli
etal.,2005























Yihetal.,
2005
















C-131  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




CGL-2 cells




























CGL-2 cells




HeLa S3 cells










Arsenic
Species




As111 SA




























As111 SA




As111 SA










Concentration(s)
Tested (nM)




1, 2, 3, 4, 5, 10






























Duration of
Treatment




24 hr






























LOECa
(HM)




1



























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Treatments caused a shift
in percentages of cells in
Gl, S, andG2/M, with a
dose-dependent ft in
G2/M cells over the
range of doses of 0
(-25%) to 4 (-85%),
followed by a U above a
dose of 5 that reached
-50% at dose of 10.
G2/M cells were
predominantly mitotic
cells. Mitotic arrest was
associated with inorganic
arsenic -induced cell
death (see row
immediately above).
When synchronized cells
were treated with dose of
2, all cells, whether
treated in the Gl, S, or
G2 stage, progressed into
and arrested at mitosis,
where they were
demonstrated to contain
damaged DNA, as
demonstrated by the
appearance of the DNA
double-strand-break
marker phosphorylated
histone H2A.X (y-
H2AX).
Following on from row above, other experiments showed that inorganic
arsenic appears to inhibit activation of the G2 DNA damage checkpoint and
thereby allows cells with damaged DNA to proceed from G2 into mitosis.
The subsequent arresting of cells with damaged DNA in mitosis is thought to
enhance the induction of apoptosis.

5, 10, 20, 50








Ihr








10







Inhibition of mitotic exit
after cells were arrested
in mitosis by treatment
with nocodazole and the
nocodazole was removed
before arsenic treatment.
This shows that such a
dose interferes with
mitosis.




Reference




Yihetal.,
2005



























Yih et al.,
2005



Huang and
Lee, 1998






C-132  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HLFC cells

HLFK cells
(Ku70
deficient)











Human
primary
peripheral
blood
lymphocytes











Human
primary
peripheral
blood
lymphocytes










Arsenic
Species


As111 SA
for both












As111 SA
Asv
MMAm
MMAV
DMA111
DMAV









As111 SA
Asv
MMAm
MMAV
DMA111
DMAV











Concentration(s)
Tested (nM)


1,2.5,5, 10
for both












1.25-160
1.25-500
0.1-2.7
10-10000
0.11-12.26
10-10000









1.25-160
1.25-500
0.1-2.7
10-10000
0.11-12.26
10-10000











Duration of
Treatment


4hr
for both













24 hr
for all














24 hr
for all













LOECa
(HM)

2.5

2.5












2.5
50
1.5
10000
1.02
3000









20
150
1.8
None
1.02
300








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Following the 4-hr As111
treatment, cells were
incubated in arsenic-free
medium for 24 hr before
determining the
proliferation index and
the proportions of cells
in different parts of cell
cycle. Both cell types
had U proliferation index
and an ft in G0/Gi cells at
dose of 2.5. Both effects
were more extreme in
HLFK than in HLFC
cells at the 3 highest
doses.
Replicative index (RI):
All 6 compounds
induced significant
slowing of the cell cycle.
Methylated trivalent
arsenicals were 3 orders
of magnitude more
potent than the
methylated pentavalent
arsenicals. Inorganic
arsenic compounds were
substantially more toxic
than methylated
pentavalent arsenicals.
Mitotic index (MI):
u.
u.
u.
NSE.
ft to peak of 3x at 3. 07.
ft to peak of 6x at 1000.
Both decreased abruptly
near concentration at
which RI showed ft
proportion of first
division metaphases.
This is consistent with
disruption of spindle
integrity.




Reference


Liuetal.,
2007b














Kligerman
etal.,2003














Kligerman
etal.,2003









C-133  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Human
peripheral
lymphocytes







TR9-7 cells
that were
released from
being mostly
synchronized
in G2 (using
Hoechst
33342)
shortly before
inorganic
arsenic
treatment
began





PCI-1 cells










CHO cells
treated with
MMS before
or after
inorganic
arsenic
treatment





Arsenic
Species

As111 SA











As111 SA















As111 ATO











As111 SA










Concentration(s)
Tested (nM)

5











5















1,2,3,4










10, as
pretreatment

10, as
post-treatment







Duration of
Treatment

24 hr











1-24 hr















3 days










24 hr

24 hr









LOECa
(HM)

5









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
There was delayed cell
cycle progression. In
treated cells, 73% and
32% were still in the first
mitotic division at
fixation times of 72 and
96 hr, respectively,
whereas in untreated
cells up to 90% were in
second or subsequent
divisions at these times.
Conclusions based on mitotic indices
determined over the 24-hr period in
cells made p53(+) or p53(") by
controlling tetracycline levels:
inorganic arsenic delayed entry into
mitosis in both p53(+) and
p53(-) cells, with peak being delayed
by ~3 hr from that of cells unexposed
to inorganic arsenic. Mitotic exit
occurred at a normal rate in inorganic
arsenic-treated p53(+) cells but was
markedly delayed in p53(-) cells and
only reached the baseline level after
24 hr, by which time the inorganic
arsenic-treated p53(+) cells had already
reached that level and had begun to
cycle again.

2










10

10






Growth inhibition
(growth measured by
MTT assay):
About 50% inhibition at
2; cells were arrested in
the G2-M phases.
Growth inhibition was
also induced in 3 other
human head and neck
squamous cell carcinoma
cell lines.
Inhibition of mitosis and
cell proliferation:
U in inhibition of both
endpoints compared to
MMS alone.

ft in inhibition of both
endpoints compared to
MMS alone, synergistic
interaction.




Reference

Jha et al.,
1992










McNeely
etal.,2006














Seoletal.,
1999










Lee etal.,
1986






C-134  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
MCF-7 cells








Human
lymphocytes
















Chinese
hamster V79
cells


























Arsenic
Species
As111 ATO








As111 SA

















As111 SA

DMAV


























Concentration(s)
Tested (nM)
3







10'10, ID'8, ID'6,
10"4, 0.01, 1

















5

2mM


























Duration of
Treatment
24 hr








2hr

















24 hr
for both



























LOECa
(HM)
3








10-io

















5

2mM























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Treatment blocked the
cell cycle in mitosis,
resulting in a time-
dependent accumulation
of cells in G2/M, with
about 50% in G2/M at
this time.
Induction of mitotic
arrest:
4 of 5 donors showed
statistically significant
increase at lowest dose.
All showed significant
increase from dose of 10"
8 through 0.01. There
was much inter-
individual variation, but
there was a positive
dose-response within
data for each donor.
There was a almost no
response at dose of 1
because of cytotoxicity.
Accumulation of mitotic
cells and other abnormal
cells as follows
(approximate
percentages of cells of
each type present after
24-hr treatment): Control
(assumed same as
distribution at starting
time): 97%
mononucleated, 3%
metaphase.
As111: 75%
mononucleated, 11%
metaphase, 10%
binucleated, 4%
multinucleated.
DMAV: 24%
mononucleated, 52%
metaphase, 1%
binucleated, 23%
multinucleated. DMAV
caused disappearance of
microtubule network and
abnormalities of mitotic
microtubules (i.e.,
spindles) — there was a
big ft increase in
frequency of multipolar
and aster-like spindles.




Reference
Ling et al.,
2002







Vegaetal.,
1995
















Ochi et al.,
1999b
























C-13 5  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
SVEC4-10
cells

HCT1 16 cells
(securin +/+)
HCT1 16 cells
(securin -/-)







SVEC4-10
cells




HT1 197 cells



Arsenic
Species
As111 SA

As111 SA
for both








As111 SA





As111 SA



Concentration(s)
Tested (nM)
2, 4, 8, 16

4, 8, 12, 16
for both








20





1,5, 10



Duration of
Treatment
24 hr

24 hr
for both








24 hr





24 hr



LOECa
(HM)
4

12
4








20





5


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Fraction of cells in G2/M
phases of cell cycle:
slight ft at 4, big ft at 8
and 16. Also an effect
on rate of cell growth
(tested at 4, 8, 12, 16): U
at all doses, with a strong
dose-response.
Fraction of cells in G2/M
phases of cell cycle:
Similar ft at 12 and 16 to
-39%.
ft at 4 to -3 8% with a
positive dose-response,
reaching -49% at highest
dose. Consistent with
the conclusion, based on
the above data, that
securin protects against
arsenic-induced cell
cycle arrest, the -/- cells
also showed a much
bigger ft in the mitotic
index and in the fraction
of cells in
"anaphase/mitosis."
They also showed sister-
chromatid separation.
Cell numbers were
counted using a
hemocytometer: after 6
days of culturing after
the inorganic arsenic
treatment, there were
-25% as many cells as in
the control.
Complete inhibition of
cell proliferation
occurred eventually at
the dose of 10, with an
accumulation of cells in
S-phase. At the dose of
10, after 12 and 24 hr,
1.5x and 2. Ix more cells
were in S-phase than in
control, respectively,
with a large deficit of
cells inGl.

Reference
Chao etal.,
2006a

Chao etal.,
2006a








Hsu et al.,
2005




Hernandez
-Zavala
etal.,
2005


C-136  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
CL3 cells,
synchronous
atGl









Human
lymphoblastoi
d cells

















Lyophilized
bovine tubulin














Arsenic
Species
As111 SA







As111 SA


Asv


MMAm

MMAV


DMA111


DMAV








As111 SA
Asv
MMAm
MMAV
DMA111
DMAV











Concentration(s)
Tested (uM)
50







0.2, 0.4, 0.6, 1,
2.5, 5, 10 uM

0.5, 1,2.5,5,7.5,
10 mM

0.05,0.1,0.2,0.3,
0.4,0.5,1 uM
0.5, 1,2.5,5,7.5,
10 mM

0.05,0.1,0.2,0.3,
0.4, 0.5 uM

0.5, 1,2.5,5,7.5,
10 mM







0.1, 1, 10 mM
0.1, 1, 10 mM
1, 10, 100 uM
0.1, 1, 10 mM
1, 10, 100 uM
0.1, 1, 10 mM











Duration of
Treatment
3hr











6hr
for all


















Time course
over 1 hr














LOECa
(UM)
50







None


7.5 mM


0.4 uM

None


None


5mM








ImM
None
1 uM
0.1 mM
10 uM
0.1 mM








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell proliferation, based
on cell number: U to
-35% of control.
Survival was cut to
20%-25% by co-
treatment with PD98059
orU0126.
Effect on the mitotic
index:
NSE, but results were
confounded by high
toxicity.


Slight statistically
significant ft in slope.

Statistically significant ft
in slope.


Slight statistically
significant ft in slope.

Equivocal, highly
variable, effects probably
because of toxicity.
Statistically significant ft
in slope.
Effect on GTP-induced
polymerization of
lyophilized bovine
tubulin:

U at 1 mM, M at 10
mM.
NSE.
Slight ft at 1 uM, U at 10
uM, M at 100 uM.
Slight ft at 0.1 and 1
mM, ft at 10 mM.
U at 10 uM, M at 100
uM.
U at 0. ImM, NSE at 1
mM, ft at 10 mM.




Reference
Li et al.,
2006a










Kligerman
et al., 2005


















Kligerman
et al., 2005











Cell Proliferation Stimulation
C-137  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
K562 cells
(human
erythroleukem
ia cells)






NHEK cells











NHEK cells
















Arsenic
Species
As111 ATO






As111 SA


Asv,
MMAV,
DMAV









As111 SA
















Concentration(s)
Tested (nM)
2.5








0.001,0.005,0.01,
0.05,0.1,0.5, 1,5,
10 for all










0.2, 0.4, 0.8
















Duration of
Treatment
12hr

48 hr






24 hr
for all









Iday

2 days

3 days














LOECa
(HM)
2.5

2.5




0.005



None








0.2

0.4

0.4











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
-27% of cells are
mitotic.
(In control, only 4% of
cells are mitotic.)
-55% of cells are
mitotic.
Stimulation of cell
proliferation, but with
inhibition of cell
proliferation at > 0.05.
Stimulation was
measured as
incorporation of
3[H]thymidine into
cellular DNA.
No stimulation of cell
proliferation; inhibition
of cell proliferation at
0.05 or higher.
Increase in proliferation
based on cell counts:
ft of 32%, 58%, and
50%, respectively.

ft of 20% and 21% at
doses of 0.4 and 0.8,
respectively.
ft of 27%, only at dose of
0.4.
PI staining and FACS
analysis after 2 days
showed a significant
shift from cells in Gl to
cells in G2/S at both
doses that showed an ft
in proliferation.




Reference
Li and
Broome,
1999







Vegaetal.,
2001










Hwang et
al., 2006












C-13 8  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


HaCaT cells













JB6C141
cells
transfected as
described for
this assay



JB6C141
cyclinDl-Luc
massl cells








HaCaT cells











Arsenic
Species


As111 SA














As111 SA






As111 SA










As111 SA











Concentration(s)
Tested (nM)


1.25,2.5,5














1.25






5










0.5, 1.0











Duration of
Treatment


48hr














72 hr






24 hr










20 passages











LOECa
(HM)


1.25














1.25






5










—








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft in fraction of cells in S
phase: at doses of 0,
1.25, 2.5, and 5, the
percentages of cells in S
were 24.9%, 29.8%,
33.8%, and 38.7%,
respectively. Since there
was a corresponding ft in
fraction of cells in G2/M
phase, it was concluded
that inorganic arsenic
promoted the transition
fromGltoS. The 24-hr
treatment caused a
similar effect at the 2
higher doses.
Proliferation was
measured by using the
CellTiter-Glo®
Luminescent Cell
Viability Assay: ft in
proliferation index to
~1.62x.
Fraction of cells in S
phase and cell apoptosis
(i.e., cell sub-Gl phase)
were measured using PI
staining with flow
cytometry: ft in fraction
of cells in S from
-11.8% to -14.5%; there
was no induction of
apoptosis and no
evidence of cytotoxicity.
Not a significantly
increased growth rate,
but the trend was in that
direction with
accumulated population
doublings of 58 to 67 in
the control and 1.0
groups, respectively,
with the value being -6 1
in the 0.5 dose group.




Reference


Ouyang et
al., 2005












Ouyang et
al., 2006






Ouyang et
al., 2006









Chien et
al., 2004







C-139  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
C3H 10T1/2
cell line
(mouse cells
with
fibroblast
morphology
during routine
culture but
capable of
differentiation
into
adipocytes)




C3H 10T1/2
cell line
(mouse cells
with
fibroblast
morphology
during routine
culture but
capable of
differentiation
into
adipocytes)

BothHL-60
cells
and HaCaT
cells







UROtsa cells










Arsenic
Species



As111 SA















As111 SA









As111 SA










As111 SA










Concentration(s)
Tested (nM)



6















6









0.1,0.5, 1, 10,20,
40









2,4










Duration of
Treatment



8wk















8wk









5 days










72 hr










LOECa
(HM)



6















6








0.5 but
possibly
0.1









2







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

Marked increase in FBS-
stimulated DNA
synthesis (detected using
[3H]thymidine
incorporation) following
dexamethasone/insulin
treatment (to induce
differentiation), but only
after the arsenite
exposure has been
stopped — the increased
mitogenic response is
masked while the
arsenite treatment
continues.


Marked increase in cell
number compared to
control cells following
dexamethasone/insulin
treatment (to induce
differentiation), but
increase only occurs
after the arsenite
exposure has been
stopped.
By use of MTT assay: ft
in cell number, with peak
at 0.5; U in cell number
to below control level at
1, with a continuing
decrease at higher
concentrations.
(Same general response,
but to a lesser extent,
with same treatments
over 1 day or 3 days.)
Increase in cell
proliferation based on
statistically significant
increase in
[3H]thymidine
incorporation; also there
was a significantly
higher fraction of cells in
S-phase of cell cycle.




Reference


Trouba et
al., 2000














Trouba et
al., 2000









Zhang et
al., 2003








Simeonova
etal.,2000







C-140  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue






NHEK cells
















Arsenic
Species



As111 SA


MMAm


DMA111













Concentration(s)
Tested (nM)



0.2, 0.4, 0.6, 0.8,
1, 2, 4, 6, 12

0.1,0.2,0.4,0.5,
0.8, 1,2

0.1,0.2,0.4,0.5,
0.6, 0.7, 0.8, 1, 2,
3











Duration of
Treatment




24 hr
for all; index
was then
determined
immediately














LOECa
(HM)



2


0.5


0.6













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Proliferation index based
on MTT assay; the
statistical comparison
was with the untreated
control:

ft at 3 doses from LOEC
through 6.


ft at 2 doses from LOEC
through 0.8.
ft at 2 doses from LOEC
through 0.7.
Significant cytotoxicity
occurred at 12 uM and
higher for inorganic
arsenic and at 1 uM and
higher for the other
arsenicals. Cell cycle
distributions were
changed in many
different ways.
Reference





Mudipalli
etal.,2005
















C-141  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



NHEK cells
irradiated
with 100
mJ/cm2 of
UVB to arrest
94.5% of cells
in GO/G!
stages of cell
cycle while
only killing
2-3% of the
cells.

















Postconfluent
PAEC cells in
a
monolayer

PAEC cells in
mid-
exponential
growth in a
monolayer











Arsenic
Species


As111 SA


MMAm


DMA111

























As111 SA
for both

















Concentration(s)
Tested (nM)


0.2, 0.4, 0.6, 0.8,
1, 2, 4, 6, 12

0.1,0.2,0.4,0.5,
0.8, 1,2

0.1,0.2,0.4,0.5,
0.6, 0.7, 0.8, 1, 2,
3





















1, 2.5, 5, 10, 20




1, 2.5, 5, 10, 20















Duration of
Treatment



24 hr
for all; index
was then
determined
Immediately


























4 hr for both


















LOECa
(UM)


0.6


0.4


0.4























1




10












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Proliferation index based
on MTT assay; the
statistical comparison
was with the untreated
control:
ft at 6 doses from LOEC
through 6.


ft at 4 doses from LOEC
through 1.0.


ft at 5 doses from LOEC
through 0.8.
Significant cytotoxicity
occurred at 12 uM for
inorganic arsenic and at
1 uM and higher for the
other arsenicals. At all
doses showing a
significant effect on the
proliferation index after
arsenical exposure, the
point estimate was
always higher in the cells
with prior UVB
exposure. Cell cycle
distributions were
changed in many
different ways.
Incorporation of
[3H]thymidine into
genomic DNA:
ft at 1, 2.5, and 5,
indicating a mitogenic
response. Only the
response at 5 is
significantly higher, but
the 2 lower doses are
probably also higher;
there was no effect at
higher doses.
U in rate of DNA
synthesis. (In the
absence of any treatment,
such cells have a higher
rate of DNA synthesis
than the postconfluent
cells in a monolayer.)




Reference





Mudipalli
etal.,2005



























Barchowsk
y et al.,
1996













C-142  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


PAEC from
freshly
harvested
vessels


U-2OS cells



















SHE cells











TM3 cells








Arsenic
Species
As111
probably
ATO, but
called
arsenite



As111 SA



















DMAmI











As111 SA








Concentration(s)
Tested (nM)


1,5,10





0.01,0.05,0.1,
0.25,0.5, 1,2.5


















0.1,0.25,0.5, 1.0










0.000008,
0.00008, 0.0008,
0.008, 0.08, 0.77







Duration of
Treatment


24 hr





24 hr



















Iday











72 hr








LOECa
(HM)


1





0.01



















0.1











0.000008





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Extent of cell
proliferation was
estimated using
fluorescent Cyquant
assay:
ft at 1 and 5, but U at 10.
Cell survival was
determined using the
clonal survival treat-and-
plate method:
At doses of 0.01 and
0.05, clonal-forming
ability was stimulated to
120%-124%ofthe
control, p < 0.006.
There was no increase at
a 72-hr exposure or at
higher doses with a 24-hr
exposure. Similar results
were found with the
neutral red and MTT
assays, and sometimes
with those assays the
point estimates still
showed an increase at
the dose of 0.01 after the
72-hr exposure.
Cell growth (no. of
viable cells): ft at both
0.1 and 0.25, and also
big ft for them after 2
and 3 days. Increase by
1 day at dose of 0. 1 was
~8-fold.
At dose of 1.0, -40%
cytotoxicity. No clear
effect at 0.5 until after 3
days, then -40%
cytotoxicity.
Increase in cell
proliferation: a
statistically significant
increase at all doses
except 0.77; the peak of
-152% of control was at
0.00008.




Reference


Barchowsk
y et al.,
1999a



Komissaro
va et al.,
2005

















Ochi et al.,
2004









DuMond
and Singh,
2007




C-143  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


UROtsa cells







NHEK cells,
both with and
without
irradiation
with 100
mJ/cm2 of
UVB to arrest
94.5% of cells
in GO/G!
stages of cell
cycle while
only killing
2%-3%ofthe
cells














HELP cells












Arsenic
Species


MMAm
for all






As111 SA

MMAm

DMA111























As111 SA












Concentration(s)
Tested (nM)


0.05
for all






6

0.8

0.8























0.1,0.5, 1,5, 10












Duration of
Treatment


12 weeks
24 weeks
52 weeks







24 hr
for all
























24 hr












LOECa
(HM)


0.05
for all






6

0.8

0.8





















0.1 for ft
5forli










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Shortened cell
population doubling
times (hr) based on
counting cells in trypan
blue exclusion assay:
(control doubling time =
42 hr)
27 hr.
25 hr.
21 hr.
Examination of
expression profiles of
more than 10 cell cycle
and cell signaling
proteins that seem likely
to influence cell
proliferation showed that
many large changes
occurred following the
UVB and arsenic
treatments, arsenic
examples, all 3
arsenicals caused a big ft
in nuclear cyclin D 1 in
UVB irradiated cells,
and, for nuclear PCNA
in UVB-irradiated cells,
MMA and DMA caused
a big ft while inorganic
arsenic had no effect.
Activation of JNK
phosphorylation and
increased EOF
expression and
phosphorylation of the
EOF receptor occurred.
Cell proliferation
efficiency based on MTT
assay:
ft to 150% and 175% of
control at 0.1 and 0.5,
respectively; U to 60% of
control at 5; significant
stimulation of
proliferation was also
seen at dose of 0.5 after
treatments of 12 and 48
hr.




Reference


Bredfeldt
etal.,2006









Mudipalli
etal.,2005























Yanget
al., 2007








Chromosomal Aberrations and/or Genetic Instability
C-144  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HaCaT cells








HaCaT cells






Primary
Syrian
hamster
embryo cells
(HEC)
Primary
Syrian
hamster
embryo cells
(HEC)
Human
peripheral
lymphocytes
Human
peripheral
lymphocytes



Human
primary
peripheral
blood
lymphocytes







Arsenic
Species

As111 SA








As111 SA






As111 SA

Asv


As111 SA

Asv


As111 SA

Asv
As111 SA

Asv


As111 SA
Asv
MMAm
MMAV
DMA111
DMAV







Concentration(s)
Tested (nM)

0.5, 1.0








0.5, 1.0






0.38,3.8,7.7

3.2, 8, 16, 32


7.7

32


0.77, 1.9

16,32
7.7

32


1.25-160
1.25-500
0.1-2.7
10-10000
0.11-12.26
10-10000







Duration of
Treatment

20 passages








20 passages






24 hr for both




24 hr for both




48 hr for both


48 hr for both






24 hr
for all









LOECa
(HM)

0.5








0.5






0.38

16


7.7

32


0.77

16
7.7

32


None
None
None
1000
0.34
1000




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Comparative genomic
hybridization showed
that all 1 1 cell lines
derived from tumors (see
malignant
transformation) showed
significant loss of
chromosome 9q, and 7
lines showed significant
gain of chromosome 4q.
ft MN; detected using
the cytokinesis-block
micronucleus assay, and
scored only in
binucleated cells. There
was a positive dose-
response.
SCEs were induced;
slight upward trend with
dose.


CAs were induced:
mostly chromatid gaps
and breaks, but some
chromatid and
chromosome exchanges.
SCEs were induced;
dose-independent
response.
CAs were induced:
mostly chromatid and
chromosome gaps and
breaks, very few
exchanges.
SCE/metaphase
Top 3 in list were
negative. Potency of
others:
DMA111 > DMAV >
MMAV.
All were weak inducers
of SCE, with the most
potent inducing ~1
SCE/metaphase/|aM.




Reference

Chien et
al., 2004







Chien et
al., 2004





Larramend
y et al.,
1981


Larramend
y et al.,
1981


Larramend
y et al.,
1981
Larramend
y et al.,
1981




Kligerman
etal.,2003






C-145  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Human
primary
peripheral
blood
lymphocytes














Syrian
hamster
embryo cells






CHO Kl cells
in late Gl of
mitotic cycle




Human
peripheral
lymphocytes
Human
peripheral
lymphocytes



Arsenic
Species

As111 SA
Asv
MMAm
MMAV
DMA111
DMAV














As111 SA


Asv







As111 SA





As111 SA


As111 SA





Concentration(s)
Tested (nM)

1.25-160
1.25-500
0.1-2.7
10-10000
0.11-12.26
10-10000














0.8,3.0,6.2,10


10, 20, 64, 96






40






1, 5, 10


0.5, 1.0, 1.5,2.0





Duration of
Treatment


24 hr
for all


















24 hr for both








4hr






48 hr


48 hr





LOECa
(HM)

2.5
50
0.6
3000
1.35
3000














6.2


64






40






1


2.0


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Chromosomal
aberrations:
ft to 42.5% aberrant cells
at 10.0.
ft to 11.0% aberrant cells
at 80.0.
ft to 11.0% aberrant cells
at 1.2.
ft to 6.5% aberrant cells
at 3000.
ft to 22.0% aberrant cells
at 2.70.
ft to 57.0% aberrant cells
at 10000.
All 6 showed a positive
dose-response.
Chromatid and
isochromatid deletions
were most prevalent;
exchanges were
infrequent.
CAs and
endoreduplication
(also, with 48 hr
treatment, polyploidy).
Mainly chromatid gaps,
breaks, and exchanges,
but a few chromosome-
type aberrations
(fragments and
dicentrics).
High frequency of CAs
was induced; effect was
markedly reduced by
prior or simultaneous
(but not by subsequent)
treatment with 5 mM
GSH.
Induction of chromatid
aberrations; there was a
positive dose-response.
Induction of
chromosomal
aberrations.




Reference



Kligerman
et al., 2003


















Barrett et
al., 1989






Huang et
al., 1993





Jha et al.,
1992

Wiencke
and Yager,
1992
C-146  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


AS52 cells



















G12 cells





















CHO cells










Arsenic
Species


As111 SA

















MMAmO

DMAmI




















As111 SA

Asv









Concentration(s)
Tested (nM)


50, 100

















0.2, 0.4, 0.6, 0.8,
1.0

0.1,0.2,0.3,0.4



















0.01,0.1, 1, 10

0.01,0.1, 1, 10,
100








Duration of
Treatment


4hr


















3 days
for both




















12 hr
for both










LOECa
(HM)


100

















0.6


0.3



















1

100






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of gpt
mutations at the ypt
locus:
The mutation frequency
was twice that of the
spontaneous mutation
frequency at a high level
of cytotoxicity (15% of
the relative survival of
the control). Taken as
very weak evidence that
As111 is a gene mutagen;
results are grouped here
with CAs because most
or all of the induced
mutations were total
deletions of the gene,
perhaps caused by the
cytotoxicity.
Induction of mutations at
the gpt locus:
DMAmI: reached 5x
control mutant frequency
at 7% cell survival;
MMAmO: reached 5x
control mutant frequency
at 11% cell survival.
Taken as weak evidence
that the arsenicals are
gene mutagens with sub-
linear dose-responses;
results are grouped here
with chromosomal
aberrations because
-80% of the induced
mutations were deletions
of the gene, perhaps
caused by the
cytotoxicity. -11% of
non-deletion mutants
exhibited altered DNA
methylation.
Induction of
chromosomal
aberrations:
A positive dose-
response; 36.7% of cells
with aberrations at dose
of 10.
8.0% of cells with
aberrations at dose of
100.




Reference


Meng and
Hsie, 1996


















Klein et
al., 2007



















Kochhar et
al., 1996







C-147  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


CHO cells




CHO cells







MRC-5 cells










MRC-5 cells









Arsenic
Species

As111 SA

Asv


As111 SA

Asv






As111 SA










DMAV









Concentration(s)
Tested (nM)

0.01,0.1, 1, 10

0.01,0.1, 1, 10,
100

0.01,0.1, 1, 10

0.01,0.1, 1, 10






2.5, 5, 10










125, 250, 500









Duration of
Treatment

12 hr
for both



12 hr
for both







26 hr










26 hr









LOECa
(HM)

1

None


0.01

0.01






2.5










125






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of
endoreduplication:
A positive dose-
response; 22.0% of cells
with endoreduplication at
dose of 10.
Induction of SCEs:
10.94%/cell at lowest
dose; 14.08%/cell at
highest dose; slight
upward trend with dose.
11.38%/cell at lowest
dose; 12.84%/cell at
highest dose; no dose-
response.
Induction of SCEs
(frequencies):
0,3.24; 2.5, 5.23; 5, 6.2;
10, no surviving cells
could be found to
evaluate. There was also
much cytotoxicity at
dose of 5. High level of
cytotoxicity was also
reflected in the
proliferation index.
Induction of SCEs
(frequencies):
0, 4.25; 125, 5.89; 250,
5.95; 500, 5.91; thus no
dose-response for SCEs.
There was a significant
U in the proliferation
index at the highest dose.




Reference

Kochhar et
al., 1996




Kochhar et
al., 1996






Mouron et
al., 2006








Mouron et
al., 2005






C-148  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue




Human
lymphocytes



























Arsenic
Species




As111 SA




























Concentration(s)
Tested (nM)




10"10, 10'8, 10'6,
10'4, 0.01



























Duration of
Treatment




24 hr




























LOECa
(HM)




10-io




























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of hypoploid
and hyperploid cells:
There was a statistically
significant increase in
hyperploidy at all dose
levels in both 1st and
2nd division cells. There
was a positive (but
shallow) dose-response.
For example, in 2nd
division cells, the
frequency went from
2.3% at dose of lO'10 to
11.7% at dose of 0.01.
The 4 donors showed
variation, with 2
showing no effect at
lowest dose. It is unclear
at what dose level
induction of hypoploidy
became significant, but
there was a slight
positive dose-response
for it also. DataonCAs,
which were reported
only briefly, showed that
roughly 40% of cells had
CAs at the dose of 0.01.
A concentration of 1
only uM was highly
cytotoxic in these cells
with an exposure lasting
only 2 hr.
Reference




Vegaetal.,
1995



























C-149  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


Human
lymphocytes



















Primary
cultured
human
umbilical cord
fibroblasts























Arsenic
Species


As111 SA



















As111 SA
Asv
MMAV
DMAV
TMAV
























Concentration(s)
Tested (nM)


0.001,0.01,0.1



















0.8,2.3,3.8,7.7
16, 32, 64, 160,
321
1.4, 3.6, 7.1 mM
0.7, 1.4, 3.6 mM
3.7,7.6, 14.7 mM























Duration of
Treatment


24 hr





















24 hr for all


























LOECa
(HM)


0.001



















0.8 uM
16 uM
1.4 mM
0.7 mM
3.7 mM





















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Increase in hyperdiploid
frequency (based on
FISH analysis, there was
a statistically significant
dose-related increase for
each of the 2
chromosomes tested
from both donors).
There was also an
increase in hypodiploid
frequency, but it was
only seen (again at all
doses) in 1 of the 2
chromosomes tested and
in only 1 donor. A
related experiment
showed that As111 can
disrupt the microtubule
organization of
lymphocytes at a dose as
low as 0.001.
Induction of CAs:
The percentages of
abnormal cells at the
LOECs for the 5
chemicals in descending
order, as listed to the left,
were: 10%, 16%, 19%,
28%, and 26%.
Depletion of GSH by
pretreatment of cells
with B SO increased
induction of CAs by As111
SA, Asv, andMMAvbut
decreased it for DMAV.
In cells pretreated with
BSO before treatment
with DMAV, the presence
of 5 mM or higher GSH
in the medium markedly
increased induction of
CAs. Since GSH does
not enter the cells itself,
this suggests that some
clastogenic chemical is
generated in the medium
by interaction of DMAV
with GSH.




Reference


Ramirez et
al., 1997




















Oya-Ohta
etal., 1996






















C-150  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



CHO-9 cells















CHO-9 cells





Human
primary
peripheral
blood
lymphocytes

CHO Kl cells






Human
peripheral
lymphocytes



Arsenic
Species
As111 SA
Asv
MMAm
MMAV
DMA111
DMAV
TMAV









As111 SA
Asv
MMA111
MMAV
DMA111
DMAV
TMAV


As111 SA




As111 SA






As111 SA





Concentration(s)
Tested (nM)


10 to 10000 for all
















10 to 10000 for all





0.8




20






1, 5, 10





Duration of
Treatment


30 minfor all
















30 min





48 hr




6hr






48 hr





LOECa
(HM)
1000
1000
10
None
50
None
None









1000
1000
10
None
50
None
None


0.8




20






1


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of
chromosomal
aberrations:

Aberrations consisted
mainly of chromatid
exchanges and breaks;
dicentrics and rings
occurred rarely.
Frequencies of
aberrations per 100 cells
at the most effective
concentration for the 4
positive chemicals
ranged from 44 to 74x
that of the control.
Induction of SCEs:

For even the most potent
inducers of SCE, the
number of SCEs/cell was
less than double that of
the untreated control;
thus they were weak
inducers.
SCEs were induced;
simultaneous treatment
with SOD (an oxygen
radical scavenger)
blocked induction of
SCEs.
SCEs were induced;
simultaneous treatment
with squalene at from 40
to 160 uM significantly
and dose-dependently
inhibited induction of
SCEs.
SCEs were induced;
there was a positive
dose-response.




Reference



Dopp et
al., 2004














Dopp et
al., 2004




Nordenson
and
Beckman,
1991


Fanetal.,
1996





Jha et al.,
1992

C-151  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Human
peripheral
lymphocytes













CHO Kl cells









Mouse
lymphoma
cells
(L5178Y/Tk+A
-3.7.2Ccells)


















Arsenic
Species

As111 ATO















As111 SA









MMAm


DMA111



















Concentration(s)
Tested (nM)

0.00036, 0.00072,
0.0014














5, 10, 20, 40









0.19,0.28,0.38,
0.47, 0.52, 0.57

0.65,0.83, 1.29,
1.51


















Duration of
Treatment

24 hr















6hr









4 hr for both






















LOECa
(HM)

0.00036















5









0.28


1.51
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
SCEs were induced;
there was a positive
dose-response; co-
treatment with retinyl
palmitate at the highest
dose of As111 caused a
significant U to a SCE
frequency like that seen
at the middle dose; the
same thing also occurred
for PDT and ACT,
showing that retinyl
palmitate also reversed
some of the arsenic-
induced decrease in the
rate of cell proliferation.
Induction of MN in
binucleated cells, using
cytochalasin B after
arsenic treatment to
block cytokinesis:
simultaneous treatment
with 80 uM squalene
significantly reduced the
effect.
Mutations at Tk+/" locus
in mouse lymphoma agar
assay without exogenous
metabolic activation:
ft to 2.0x at 0.28, with a
positive dose-response,
reaching 7.2x at 0.57.
ft 2.4x control at
maximum concentration
tested.
Both compounds showed
large excess of small
colonies, which is
indicative of
chromosomal
aberrations; generally
similar results were
found in a mouse
lymphoma microwell
assay, which was
complicated by higher
toxicity.




Reference


Avaniand
Rao, 2007













Fan et al.,
1996








Kligerman
et al., 2003


















C-152  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




Human
peripheral
lymphocytes



















SY-5Y cells
HEK 293
cells









Mouse
lymphoma
cells
(L5178Y/Tk+/-
-3.7.2Ccells)








Arsenic
Species


As111 SA
Asv
MMAm

MMAV
DMAV
TMAVO


















As111 ATO
for all






As111 SA

Asv



MMAV


DMAV






Concentration(s)
Tested (nM)


0.5, 1,2,4
4, 8, 16, 32
0.01,0.05,0.1,
0.5,
1,2
50, 100, 250, 500
50, 100, 250
400, 800, 1000

















1
for all






2.3,5.4,7.7,8.5,
10.8, 14.6, 16.2
3.0, 15.2, 30.3,
45.5, 60.6, 75.8,
84.9

6.2, 12.3, 15.4,
18.5,24.7,30.9
mM

12.5,25.0,37.5,
50.0, 56.3, 62.5
mM



Duration of
Treatment





72 hr for all





















24hr
48 hr
72 hr








4hr
for all











LOECa
(HM)


2
8
1

100
250
None


















1
for all






8.5

45.5



18.5 mM

56.3 mM




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of MN in
binucleated lymphocytes
detected by the
cytokinesis-block assay
(using cytochalasin B):
ft in 2 donors at 2 and in
all 3 donors at 4.
ft in 1 donor at 8 and in
all 3 donors at 2 higher
doses.
ft in 1 donor at 1 and in
all 3 donors at 2.
ft in 2 donors at 100 and
250 and in all 3 donors at
500.
ft in 1 donor at 250.
NSE.
Further analysis of
MMAm showed ftft in
centromere-positive
micronuclei (-80% of
total), which is an
indicator of induced
aneuploidy.
Induction of MN
detected by Hoechst
staining; response in
comparison to control in
SY-5Y and HEK 293
cells, respectively, for
each duration of
treatment:
At24hr:3.70x, 3.35x.
At48hr:5.14x, 4.81x.
At 72 hr: 4.00x, 3.16x.


Mutations at Tk+/" locus
in mouse lymphoma agar
assay without exogenous
metabolic activation.
Very few, if any, large
colony mutants were
induced by all
compounds. Induction
of small colony mutants
is indicative of induction
ofCAs.




Reference




Colognato
et al., 2007




















Florea et
al., 2007










Moore et
al., 1997a








C-153  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Mouse
lymphoma
cells
(L5178Y/Tk+/-
-3.7.2Ccells)













SHE cells














V79 cells


















Arsenic
Species
As111 SA
Asv

MMAV


DMAV












As111 SA












As111 SA

Asv


















Concentration(s)
Tested (nM)
11.5, 13.1, 15.4
60.6, 69.7, 84.9

21.6, 24.7, 27.8
mM


50.0, 56.3, 62.5
mM










4,6,8














50, 100, 250, 500
for both

















Duration of
Treatment

4hr
for all
















24 hr














Ihr
for both

















LOECa
(HM)
11.5
60.6

24.7 mM

None












4 for first
two
effects











50

50















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of CAs:
Aberrations consisted
mainly of chromatid
exchanges and breaks;
all concentrations
reported showed
induction of CAs except
for DMAV, which gave
results of borderline
significance that were
considered negative by
the authors. Lower
frequencies of induction
were seen for MMAV
than for the inorganic
arsenics in spite of the
much higher doses.
Induction of CAs: 0, 1%;
4, 9%; 6, 15%; 8, 32%.
Induction of polyploidy
and endoreduplication: 0,
0%; 4, 6%; 6, 19%; 8,
27%.
Colony -forming
efficiency relative to
control after 7 days of
culturing post- As
treatment: 6, 77%; 8,
49%.
MI: 0, 9.2; 4, 10.9; 6,
8.7; 8, 1.3.
Induction of CAs (no. of
aberrations in 100
metaphase cells):
0, ~7; 50, -49; 100, -99;
250, -120; 500, -160.
0, -6; 50, -32; 100, -44;
250, -62; 500, -73.
Aberrations were mainly
chromatid breaks. Co-
treatment or pretreatment
with tea extracts reduced
aberration frequencies by
half or more, while post-
treatments also reduced
the level of effects,
which was suggestive of
enhanced repair. Tea
extracts induced CAT
and SOD activity.




Reference


Moore et
al., 1997a















Hagiwara
etal.,2006













Sinha et
al., 2005a














C-154  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




V79 cells





























Mouse
lymphoma
cells
(L5178Y/Tk+A
-3.7.2Ccells)







Arsenic
Species


As111 SA

Asv

DMAV


























As111 SA
Asv

MMAV


DMAV






Concentration(s)
Tested (nM)




50, 100, 250, 500
for all



























11.5, 13.1, 15.4
60.6, 69.7, 84.9

21.6, 24.7, 27.8
mM


50.0, 56.3, 62.5
mM




Duration of
Treatment




Ihr
for all




























4hr
for all










LOECa
(HM)



100 or
possibly
50 for all



























None
60.6

24.7 mM

None




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of micronuclei
(MN) in cytochalasin B
assay: (No. of MN per
1000 binucleated cells):
50, -105; 100, -110;
250, -170; 500, -300.

50, -80; 100, -105; 250,
-125; 500, -150.
50, 52; 100, 70; 250, 99;
500, 111.
Co-treatments with tea
extracts reduced MN
frequencies by two-thirds
or more for As111 and by
half or more for Asv and
DMAV. Pretreatments
with tea extracts also
caused a large U in MN
frequencies for all 3
arsenicals. Post-
treatments also reduced
MN frequencies, which
was suggestive of
enhanced repair. The
polyphenols EGCG and
TF extracted from tea
had similar effects in
reducing MN
frequencies. The LOECs
are uncertain because no
data were reported for
the untreated controls.

Induction of MN in
binucleated cells, using
cytochalasin B after
arsenic treatment to
block cytokinesis: As111
SA gave results of
borderline significance
that were considered
negative by the authors.




Reference




Sinha et
al., 2005b




























Moore et
al., 1997a







C-155  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Don Chinese
hamster cells









CHO cells









Human
lymphocytes

CHO cells



P388DJ
macrophage
cell line

Human
peripheral
lymphocytes
Human
peripheral
lymphocytes





Arsenic
Species
As111 SA

Asv
arsenic
pent-
oxide
Asv
disodium
arsenate




As111 SA

Asv








As111 SA


As111 SA


As111 SA

Asv

DMAV
As111 SA

DMAV
As111 SA







Concentration(s)
Tested (nM)
7.7


13.9


32.1






1,5, 10

50, 80, 100








0.5, 1.0, 5.0


1, 10


0.01,0.1, 1

0.1, 1, 10

1,10
1

1
0.5, 1.0, 1.5,2.0







Duration of
Treatment



28 hr
for all









24 hr
for all








48 hr


24 hr



48 hr
for all


48 hr

48 hr
48 hr







LOECa
(HM)
7.7


13.9


32.1






1

50








0.5


1


None

None

None
1

1
1.0




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

SCEs were induced for
all 3 chemicals at 1.56,
1.61, and 1.46 times the
control level,
respectively.
The concentrations
tested for SCEs for all 3
chemicals were the "50%
inhibition doses"
following culturing for
72 hours and using a
Giemsa test for viability.
Chromosome aberrations
(breaks and exchanges)
were induced by both
compounds with a dose-
response relationship;
As111 was 5-10 times
more effective than Asv
per unit dose; 80 (oM
was -50% growth
inhibition dose over 4
days for As111.
Chromosome aberrations
(breaks and exchanges)
were induced.
SCEs were induced with
a dose-response
relationship.

No more than slight hints
of induction of SCEs
under any of these
experimental conditions.
Induction of SCEs.

Induction of SCEs.
Induction of SCEs: in 2
of the 3 donors, the
LOECwasl.5. Cells
from one donor were
more sensitive.




Reference



Ohno et
al., 1982









Wan et al.,
1982








Wan et al.,
1982

Wan et al.,
1982



Andersen,
1983


Andersen,
1983
Wiencke
and Yager,
1992


C-156  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
BrdU-
substituted
replicating
human
lymphocytes






GO human
lymphocytes


2BS cells








V79-C13
Chinese
hamster cell
line














NB4 cells













Arsenic
Species
As111 SA

Asv







As111 SA

Asv


As111 SA








As111 SA

















As111 ATO













Concentration(s)
Tested (nM)
0.77, 1.54

13.5,26.9







1.54

26.9


1.0, 3.0, 5.0, 10








10

















0.75













Duration of
Treatment
24 hr

24hr







24hr

24hr


5hr








24 hr

















3wk













LOECa
(HM)
0.77

13.5







None

None


1.0








10

















0.75










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of SCEs, but
only in 2 of 4 subjects.

Induction of SCEs, but
only in 1 of 4 subjects; in
2 subjects (1 at lower
dose and 1 at higher
dose) there was a slight
but significant decrease
in SCEs.
No induction of SCEs
with either treatment.
(4 subjects in each
group.)
DNA-protein crosslinks
detected by alkaline
elution; peak effect at
3.0; no effect at 10.0;
further testing of DNA
showed the crosslinks to
be protein-associated
DNA-strand breaks.
Cells examined after 6,
12, 18, and 24 hr and
after 6, 24, and 48 hr of
recovery: by 6 hrof
treatment there were
multinucleated cells and
round cells, by 12 hr
there were giant cells.
Multinucleated cells
persisted at high levels to
48 hr after treatment.
Also saw abnormal
spindles and persistent
(i.e., up to 5 days
observed) aneuploidy
and hyperdiploidy, but
no statistically
significant changes in
CAs or MI.
Enlarged cells were
found that contained
chromosomal end-to-end
fusions. In 80
karyotypes, there were
an average of 2.4 fusion
events per cell, and 32
cells had polyploidy.
FISH analysis showed
that fusions are
associated with attrition
of telomeres.




Reference

Crossen,
1983








Crossen,
1983

Dong and
Luo, 1993








Sciandrello
et al., 2002
















Chou et
al., 2001









C-157  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

NB4 cells

HeLa cells




PARP-1+/+
MEF cells

PARP-I-'-
MEF cells






PARP-1+/+
MEF cells

PARP-I-'-
MEF cells











PARP-1+/+
MEF cells

PARP-1"'"
MEF cells





Arsenic
Species

As111 ATO
for both





As111 SA
for both









As111 SA
for both














As111 SA
for both








Concentration(s)
Tested (nM)

0.25

1




11.5,23
for both









11.5,23
for both














11.5,23
for both








Duration of
Treatment

4, 5, 6 wk

3,4wk




24 hr
for both









48 hr
for both














24 hr
for both








LOECa
(HM)

0.25

1




None

11.5








23

11.5













11.5

11.5




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Southern blot of digested
genomic DNA:
U telomere length at all 3
time points.
U telomere length at both
time points.
Telomere length
measured by flow FISH
assay (point estimate
comparisons were made
to unexposed cells of the
same genotype):
-98% of control at 11. 5,
-91% of control at 23;
both are NSEs.
-76% of control at 11. 5,
-71% of control at 23.
Telomere length
measured by flow FISH
assay (point estimate
comparisons were made
to unexposed cells of the
same genotype):
-99% of control at 11. 5,
-79% of control at 23;
the one at 11.5 was NSE.
-79% of control at 11. 5,
-41% of control at 23.
inorganic arsenic-
induced telomere
attrition was thus much
greater in PARP-l"7"
MEFs.
Induced (experimental -
control) % of MN in
binucleated cells (with
cytochalasin B post-
treatment to block
cytokinesis):
-4% at 11. 5, -5% at 23.
-18% at 11.5, -13% at
23.




Reference

Chou et
al., 2001




Poonepalli
et al., 2005










Poonepalli
etal.,2005













Poonepalli
et al., 2005






C-158  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


PARP-1+/+
MEF cells

PARP-r7'
MEF cells




PARP-1+/+
MEF cells

PARP-r7'
MEF cells









PARP-1+/+
MEF cells

PARP-I-'-
MEF cells










Arsenic
Species


As111 SA
for both







As111 SA
for both












As111 SA
for both













Concentration(s)
Tested (nM)


11.5,23
for both







11.5,23
for both












11.5,23
for both













Duration of
Treatment


48 hr
for both







24 hr
for both












48 hr
for both













LOECa
(HM)


11.5

11.5






None

11.5











None

11.5









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced (experimental -
control) % of MN in
binucleated cells (with
cytochalasin B post-
treatment to block
cytokinesis):
-6% at 11. 5, -6% at 23.
-27% at 11. 5, -15% at
23.
Induced (experimental -
control) frequency of
CAs per cell, using FISH
with a telomeric PNA
probe:
-0.04 at 11.5, -0.04 at
23;bothareNSEs.
-0.09 at 11. 5, -0.05 at
23; only the one at 11.5
was statistically
significant.
CAs included end-to-end
fusions, chromosome
breaks, and fragments.
Induced (experimental -
control) frequency of
CAs per cell, using FISH
with a telomeric PNA
probe:
-0.04 at 11. 5, -0.04 at
23;bothareNSEs.
-0.11 at 11. 5, -0.03 at
23; only the one at 11.5
was statistically
significant.
CAs included end-to-end
fusions, chromosome
breaks, and fragments.




Reference

Poonepalli
etal.,2005








Poonepalli
et al., 2005












Poonepalli
etal.,2005










C-159  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



293 cells























293 cells























Arsenic
Species


As111 ATO
























MMAm























Concentration(s)
Tested (nM)



2























2























Duration of
Treatment



24 hr























24 hr























LOECa
(HM)



2























2




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
No. MN/1000
binucleated cells (with
cytochalasin B during
treatments to block
cytokinesis): untreated =
-35; dose of 2: big ft to
-260.
Effects of co-treatment
(CoTr) with modulators
at high doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -155.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -170.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -605.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -670.
No. MN/1000
binucleated cells (with
cytochalasin B during
treatments to block
cytokinesis): untreated =
-35; dose of 2: big ft to
-230.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -130.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -155.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -465.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -470.




Reference



Jan et al.,
2006






















Jan et al.,
2006



















C-160  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



293 cells























SV-HUC-1
cells






















Arsenic
Species



DMA111























As111 ATO























Concentration(s)
Tested (nM)



2























2























Duration of
Treatment



24 hr























24 hr























LOECa
(HM)



2























2




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
No. MN/1000
binucleated cells (with
cytochalasin B during
treatments to block
cytokinesis): untreated =
-35; dose of 2: big ft to
-315.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -170.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -175.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -630.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -635.
No. MN/1000
binucleated cells (with
cytochalasin B during
treatments to block
cytokinesis): untreated =
-35; dose of 2: big ft to
-330.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -150.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -150.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -680.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -645.




Reference



Jan et al.,
2006






















Jan et al.,
2006



















C-161  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



SV-HUC-1
cells






















SV-HUC-1
cells






















Arsenic
Species



MMAm























DMA111























Concentration(s)
Tested (nM)



2























2























Duration of
Treatment



24 hr























24 hr























LOECa
(HM)



2























2




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
No. MN/1000
binucleated cells (with
cytochalasin B during
treatments to block
cytokinesis): untreated =
-35; dose of 2: big ft to
-270.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -145.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -150.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -570.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -470.
No. MN/1000
binucleated cells (with
cytochalasin B during
treatments to block
cytokinesis): untreated =
-35; dose of 2: big ft to
-400.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -160.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -145.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -620.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -650.




Reference



Jan et al.,
2006






















Jan et al.,
2006



















C-162  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


SHE cells









Primary rat
hepatocytes













CL3 cells,
synchronous
atGl

CL3 cells,
asynchronous
(asyn)

CL3 cells,
synchronous
atG2/M













Arsenic
Species

As111 SA

DMAmI








As111 SA

















As111 SA
for all



















Concentration(s)
Tested (nM)

3, 10

0.5, 1.0








0.25,0.5, 1,2.5,5,
7.5, 10
















50
for all



















Duration of
Treatment


48 hr for both









27 hr

















3hr
for all



















LOECa
(HM)

10

0.5








1

















50
for all
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Aneuploidy detected by
flow cytometry:
Slight It

Slight ft; iWl at 1.0.
Other experiments
showed that DMAmI
caused abnormalities of
mitotic spindles,
centrosomes, and
microtubule elongation.
Induction of MN (mean
no./lOOO cells):
17.4 at dose of 1,
increasing with dose to
24. 4 at dose of 7.5;
control = 13.7; too many
cells were dead at dose
of 10 to evaluate this
endpoint. Co-treatment
withlOor25uMSbmCl:
0 in micronucleus
frequency below
expectation of an
additive interaction; that
chemical also induced
MN.
Induction of MN;
inorganic arsenic
treatment was followed
by culturing with
cytochalasin B for 24 hr
to block cytokinesis):
induced no. of MN
(experimental -
controiyiOOO
binucleated cells: Gl,
-181; asyn, -141;
G2/M, -125; when Gl
cells were co-treated
with inorganic arsenic
and either PD98059 or
UO 126, this number U
from -181 to -75-80.
Percentages of
binucleated cells: Gl,
14%; asyn, 47%; G2/M,
39%.




Reference


Ochi et al.,
2004








Hasgekar
etal.,2006
















Li et al.,
2006a
















C-163  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




CL3 cells,
synchronous
atGl




























V79-C13
Chinese
hamster cells

















Arsenic
Species




As111 SA






























As111 SA



















Concentration(s)
Tested (nM)




50






























10



















Duration of
Treatment




3hr






























24 hr



















LOECa
(HM)




50






























10
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of MN;
inorganic arsenic
treatment was followed
by culturing with
cytochalasin B for 24 hr
to block cytokinesis):
induced frequency =
-18 1/1000 binucleated
cells (as in row above);
percentage of
binucleated cells: 14%
(as in row above).
Culturing of Gl cells
with cytochalasin B for
36-48 hr (instead of 24)
caused marked ft in
percentages of
binucleated cells and
marked U in induced
numbers of MN (1000
binucleated cells) from
181 to -40-70. Also,
when cultured with
cytochalasin B for 40 hr
(instead of 24 hr) after
the co-treatment of
inorganic arsenic with
PD98059orU0126,
these 2 structurally
dissimilar inhibitors of
MEK1/2 caused no
further U from inorganic
arsenic alone.
After being expanded
through 120 generations
in the absence of arsenic
and then being cloned,
acquired genetic
instability persisted and
often came to include
dicentric chromosomes
and telomeric
associations. These
same clones, which were
often aneuploid,
micronucleated and/or
multinucleated, were
affected by the DNA
hypomethylation that
was seen globally in the
cells immediately after
the 24-hr treatment.




Reference




Li et al.,
2006a





























Sciandrello
etal.,2004















C-164  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
CHO cell
lines:

Kl (parental
to the
following
line)

XRS-5 (X-ray
andH2O2
sensitive)



CHO cell
lines:

Kl (parental
to the
following
lines)

XRS-6 (X-ray
sensitive)

XRS-5 (X-ray
andH2O2
sensitive)
























Arsenic
Species


As111 SA
















As111 SA
































Concentration(s)
Tested (nM)



10, 20, 40, 80 for
both











20, 40, 80


20, 40, 60


10, 20, 30, 40, 60





























Duration of
Treatment



4 hr for both















4 hr for all
































LOECa
(HM)


80


10










40


20


10


























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

Induction of MN in
binucleated cells, using
cytochalasin B after
arsenic treatment to
block cytokinesis: the
much less responsive Kl
cells have 6 times as
much catalase activity as
XRS-5 cells; both lines
are similar in arsenic
uptake and release, in
GSH levels, and in GSH
S-transferase activity.
Frequencies of MN per
thousand binucleated
cells per uM of arsenic
for Kl, XRS-6, and
XRS-5 cells were 2.1,
4.5, and 10.8,
respectively.
(Cytochalasin B was
used after arsenic
treatment to block
cytokinesis.) Kl cells
have 5.8 times as much
catalase activity and 5.4
times as much GPx
activity as XRS-5 cells.
Kl cells have 3.7 times
as much catalase activity
and 2. 1 times as much
GPx activity as XRS-5
cells. The cells with
intermediate amounts
have an intermediate
response. Co-treatment
of XRS-5 cells with
catalase or GPx
eliminates induction of
MNbyAsmSA.
Treatment of Kl cells
with inhibitors of
catalase and GPx makes
them much more
sensitive to induction of
MNbyAsmSA;when
co-treated together, there
is a synergistic effect.




Reference


Wang and
Huang,
1994














Wanget
al., 1997




























C-165  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue

CHO-9 cells


HFW cells




HLFC cells
HLFK cells
(Ku70
deficient)





Arsenic
Species
As111 SA
Asv
MMAm
MMAV
DMAV
DMA111
TMAV
As111 SA
for both
durations




As111 SA
for both





Concentration(s)
Tested (nM)
1, 5, 10, 50, 100,
500 for both
1, 5, 10, 30
1, 5, 10, 30, 100,
500, 5000 for both
1,5,10
1, 5, 10, 5000
1.25,2.5,5, 10
5, 10, 20, 40, 80




1,2.5,5, 10
for both





Duration of
Treatment

Ihr
for all


24hr
4hr




24 hr
for both





LOECa
(HM)
None
None
10
5000
5000
1
5000
1.25
10




2.5
2.5




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of MN in
binucleated cells:
DMA111 was by far the
most potent.


Induction of MN, with
about 70% being
kinetochore-positive at
maximum induction
found at dose of 5.
Induction of MN, with
about 70% being
kinetochore-negative at
maximum induction
found at dose of 40.
Induction of micronuclei
(% of cells with MN):
Control, 5%; 1, 4%; 2.5,
8%; 5, 10%, 10, 15%.
Control, 4%; 1, 6%; 2.5,
10%; 5, 21%, 10,27%.
At the 2 higher doses the
% is significantly higher
in the HLFK cells. Ku70
is 1 of 3 subunits of
DNA-dependent protein
kinase, and the Ku70
protein plays an
important role in repair
of DNA double-strand
breaks.

Reference

Dopp et
al., 2004


Yihand
Lee, 1999




Liuetal.,
2007b




C-166  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HLFC cells

HLFK cells
(Ku70
deficient)















HFF cells












HL-60 cells

HaCaT cells












Arsenic
Species


As111 SA
for both

















As111 SA













As111 SA













Concentration(s)
Tested (nM)


1,2.5,5, 10
for both

















5













0.5, 10, 20













Duration of
Treatment


24 hr
for both

















24 hr













3 days













LOECa
(HM)

5

2.5

















5












10 for U

0.5 for ft,
10 U








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Formation of abnormal
nuclei (% of cells with
abnormal nuclei):
Control, 7%; 1, 9%; 2.5,
10%; 5, 19%, 10, 23%.

Control, 10%; 1, 12%;
2.5, 21%; 5, 37%, 10,
42%. At the 3 higher
doses the % is
significantly higher in
the HLFK cells. Ku70 is
1 of 3 subunits of DNA-
dependent protein
kinase, and the Ku70
protein plays an
important role in repair
of DNA double-strand
breaks.
cen+ and cen- MN
induced per 1000 cells:
cen- MN: control,
-10/1000; inorganic
arsenic, -17/1000.
cen+ MN: control,
-2/1000; inorganic
arsenic, -18/1000.
Co-treatment with 170
nM SAM essentially
eliminated induction of
cen+ MN without having
any effect on induction
of cen- MN.
Analysis of telomere
length by TRF analysis
using Southern blot
assay:
Telomeres were
shortened compared to
controls at 10 and 20.
Telomeres were
shortened compared to
controls at 10 and 20, but
in these cells only, the
telomeres were slightly
elongated at dose of 0.5.




Reference



Liuetal.,
2007b
















Ramirez et
al., 2007












Zhang et
al., 2003









C-167  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


Human-
hamster
hybrid AL
cells
















Human-
hamster
hybrid AL
cells






Human-
hamster
hybrid AL
cells





Arsenic
Species


As111 SA



















As111 SA









As111 SA








Concentration(s)
Tested (nM)


3.8,7.7, 15.4



















11.5, 15.4









3.8








Duration of
Treatment


1 day or 5 days



















24 hr









24 hr








LOECa
(HM)


Depends
on locus


















11.5









3.8





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of mutations at
both loci, with both
showing higher response
after
5 -day treatment than
after 1-day treatment.
After only 1 day of
treatment, the LOECs
were 3.8 at SI locus and
15.4 at the HPRT locus.
This effect is not
grouped with gene
mutations because most
mutations were large
deletions; about 28 times
as many mutations
occurred at the SI locus,
and co-treatment with
DMSO eliminated most
of the mutation
induction.
Induction of mutations at
CD59 locus (formerly
known as SI locus); this
effect is not grouped
with gene mutations
because most mutations
were large multilocus
deletions; co-treatment
with SOD or catalase
considerably reduced
mutation induction.
Induction of mutations at
CD59 locus;
pretreatment with BSO
(to reduce GSH levels)
increased mutation rate
about 3 -fold.




Reference


Heietal.,
1998


















Kessel et
al., 2002








Kessel et
al., 2002




C-168  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Enucleated AL
hybrid cells
treated with
As111 were
fused with
untreated
nuclei to form
reconstituted
AL hybrid
cells








AL hybrid
cells made
highly
deficient in
mitochondria!
DNA by long-
term
ditercalinium
treatment;
then called p°
cells



Human-
hamster
hybrid AL
cells



















Arsenic
Species


As111 SA

















As111 SA











As111 SA
for both





















Concentration(s)
Tested (nM)


15.4

















7.7, 11.5, 13.5,
15.4










1.9,3.8,7.7
for both





















Duration of
Treatment


3hr

















18 hr











16 days

30 days




















LOECa
(HM)


15.4

















None











1.9

1.9

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of mutations at
CD59 locus:
Mutant frequency >2x
the frequency in control
cells reconstituted from
untreated enucleated
cells and untreated
nuclei. Induction of
ROS was demonstrated
in inorganic arsenic-
treated enucleated cells
by using a fluorescent
probe. These results
suggest that
mitochondria may be
essential for induction of
CD59" mutations (in
nuclear DNA).

No increase in CD59"
mutations; there was a
dose-related increase in
cytotoxicity. Analysis of
DNA showed that
mtDNA was >95%
depleted in the p° cells.
Suggests that
mitochondria! function
may be necessary for
induction of CD59"
mutations by inorganic
arsenic.
Induction of mutations at
the CD59" locus:
increase in mutation
frequency at all doses,
with a positive dose-
response and at least a
doubling of the control
frequency at the higher
dose. These cells
showed a dose-related
increase in cytotoxicity,
with never less than a
60% surviving fraction.
After a 60-day exposure,
there was an almost 3-
fold increase in the
number of MN observed
over the untreated
control, but details were
not provided.




Reference


Liu et al.,
2005
















Liu et al.,
2005











Partridge
etal.,2007

















C-169  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human-
hamster
hybrid AL
cells






Arsenic
Species
As111 SA









Concentration(s)
Tested (nM)
0.8,3.8,7.7,15.4









Duration of
Treatment
24 hr









LOECa
(HM)
3.8






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of CD59
mutants:
(Addition of BSD, which
suppresses GSH,
increased mutant
frequencies more than 5-
fold.)




Reference

Liuetal.,
2001




Co-carcinogenesis
Rat lung
epithelial cell
line

Rat lung
epithelial cell
line exposed
to 100 nM
B[a]Pfor24
hr











As111 SA
for both

















1.5
for both















12 wk without
theB[a]P
treatment or
immediately
following that
treatment















1.5
for both















Transformation (i.e.,
anchorage-independent
growth in soft agar)
occurred with 12-wk
inorganic arsenic
treatment alone or with
B[a]P treatment alone.
There was a synergistic
interaction when the
B[a]P treatment was
followed by the 12-wk
inorganic arsenic
treatment, with the
transformation rate then
exceeding 500 and 200
times that of the
inorganic arsenic or
B[a]P treatments alone,
respectively.


Lauand
Chiu, 2006















C-170  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue
Rat lung
epithelial cell
line

Rat lung
epithelial cell
line exposed
to 100 nM
B[a]Pfor24
hr






















Arsenic
Species
As111 SA
for both






























Concentration(s)
Tested (nM)
1.5
for both






























Duration of
Treatment
12 wk without
theB[a]P
treatment or
immediately
following that
treatment


























LOECa
(HM)
1.5
for both






























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Changes in the proteome
of the transformed cells
detected by MALDI-
TOF-MS analysis and
other methods: inorganic
arsenic andB[a]P
treatments alone caused
changes in most of the
following proteins alone.
The combined treatment
often caused a
synergistic interaction on
the protein levels in the
same direction as one or
both treatments changed
them alone. Affected
proteins were as follows:
3 proteins belonging to
intermediate filaments
were down-regulated; 6
proteins belonging to
antioxidative stress-,
chaperone-, and
glycolytic proteins were
up-regulated. Also
phosph-ERKl/2 and a-
actinin, which are
associated with
promotion of cell
proliferation and de-
differentiation, were up-
regulated.
Reference

Lauand
Chiu, 2006





























C-171  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




GM04312C
cells


























Arsenic
Species




As111 SA



























Concentration(s)
Tested (nM)




10,50



























Duration of
Treatment




24 hr



























LOECa
(HM)




10
























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
BPDE-DNA adducts
were measured after a
30-min treatment with
O.SuMBPDEthat
followed the inorganic
arsenic pretreatment.
Compared to no
pretreatment, increases
in these adducts at the
doses of 10 and 50 were
1.4x and 1.6x,
respectively. In these
NER-deficient cells,
which could be used to
dissect induction of
DNA damage from DNA
repair, it was shown that
inorganic arsenic
markedly increased the
cellular uptake of BPDE
in a dose-dependent
manner. It was
concluded that this effect
contributes to the co-
carcinogenesis in
addition to arsenic's
"well demonstrated
inhibitory effect on DNA
repair."




Reference




Shen et al.,
2006























Co-mutagenesis
E. coli WP2
irradiated
with 5. 6 J/m2
of U Von
plates that
contained:










As111 SA

Asv













100, 250, 500, 750

100, 300, 500













—















100

None













Plating protocol for Trp+
revertants: synergistic
interaction in inducing
Trp+ revertants at lower
3 dose levels for SA
only, with peak effect at
250; synergistic
interaction was seen only
in a strain of E. coli that
can carry out excision
repair of pyrimidine
dimers. FourE1. coli
strains that did not meet
that criterion were tested,
with no synergism being
seen.
Rossman,
1981














C-172  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
CHO Kl cells
in late Gl of
mitotic cycle
exposed to 7
J/m2ofUV




CHO cells
exposed to 1,
2, 4, or 8 J/m2
ofUV



















CHO cells
exposed to 1,
2, 4, or 8 J/m2
ofUV

Human
peripheral
lymphocytes
simultaneousl
y treated with
6uMDEB




Arsenic
Species

As111 SA







As111 SA






















As111 SA





As111 SA








Concentration(s)
Tested (nM)

40







5, 10






















5, 10





0.5, 1.0, 1.5,2.0








Duration of
Treatment

2hr







24 hr






















24 hr





48 hr








LOECa
(HM)

40







5






















None





1.0





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

High frequency of
chromosome aberrations
was induced; effect was
markedly reduced by
prior or simultaneous
(but not by subsequent)
treatment with GSH.
Induction of
chromosomal
aberrations: synergistic
interaction was
demonstrated at all dose
levels of UVand
inorganic arsenic except
for 1 J/m2 with the 10
uM inorganic arsenic
treatment. At other UV
dose levels, the
responses at 10 uM
arsenic only slightly
exceeded those at 5 uM.
UV or inorganic arsenic
alone induced mainly
chromatid-type
aberrations, but in cells
treated with both agents
there was an apparent
increase of chromatid
breaks, chromatid
exchanges, chromatid
gaps, and chromosome
breaks.
Induction of SCEs: no
statistically significant
effect of the inorganic
arsenic treatment was
observed.

Induction of
chromosomal
aberrations: there was
synergistic interaction
between DEB and
inorganic arsenic.




Reference

Huang et
al., 1993






Lee et al.,
1985





















Lee et al.,
1985




Wiencke
and Yager,
1992



C-173  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human
peripheral
lymphocytes
simultaneousl
y treated with
6uMDEB



CHO cells
exposed to
2 or 4 J/m2 of
UV



CHO cells
exposed to
2 or 4 J/m2 of
UV
CHO Kl cells
exposed to
1.5 or 2. 5
J/m2ofUV






CHO cells
treated with
MMS before
or after
inorganic
arsenic
treatment




CHO cells
treated with
MMS before
or after
inorganic
arsenic
treatment





Arsenic
Species

As111 SA







As111 SA






As111 SA




As111 SA










As111 SA









As111 SA










Concentration(s)
Tested (nM)

0.5, 1.0, 1.5,2.0







5, 10






5, 10




10









10, as
pretreatment

10, as
posttreatment




5, 10, as
pretreatments

5, 10, as
posttreatments








Duration of
Treatment

48 hr







24 hr






24 hr




24 hr









24hr

24hr






24 hr


24 hr









LOECa
(HM)

-1.0







5






5




10









10

10






None


5






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of SCEs:
Unlike with CAs, there
was not a synergistic
interaction. Although no
statistical comparisons
were presented, the
trends suggested
additivity between the
two mutagens.
Induction of gene
mutations to 6-
thioguanine resistance:
synergistic interaction
was demonstrated at both
dose levels of UV and
inorganic arsenic.
Induction of gene
mutations to ouabain
resistance: inorganic
arsenic had no effect.
Induction of 6-TGr gene
mutations at the HPRT
locus: synergistic
interaction was
demonstrated at both
dose levels of UV;
inorganic arsenic at
doses of 10 to 40 had no
effect on the mutation
frequency by itself.
Induction of gene
mutations at the HGPRT
locus:
U compared to MMS
alone.

ft compared to MMS
alone, synergistic
interaction.
Induction of
chromosomal
aberrations:
No change from MMS
alone.


ft frequency compared to
MMS alone, synergistic
interaction with even
bigger effect at 10.




Reference

Wiencke
and Yager,
1992





Lee et al.,
1985





Lee et al.,
1985



Yanget
al., 1992








Lee et al.,
1986








Lee etal.,
1986







C-174  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

CHO cells
treated with
MMS before
or after
inorganic
arsenic
treatment

Human
peripheral
lymphocytes







VH16 cell
line (human
primary
fibroblasts)
exposed to 7.5
J/m2ofUV



V79 cells
treated with
MNU













Arsenic
Species


As111 SA






As111 SA










As111 SA







As111 SA















Concentration(s)
Tested (nM)

5, 10, as
pretreatments

5, 10, as
posttreatments



5










5







10

5













Duration of
Treatment

24 hr


24 hr



2 hr before
X-rays, 30 min
after
X-rays








24 hr







3hr

24 hr













LOECa
(HM)

None


None




5










5







10

5










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of SCEs:
No change from MMS
alone.


No change from MMS
alone.

Synergistic interaction in
causing dicentrics and
rings in both donors;
synergistic interaction in
causing deletions in one
of the donors and
approximately an
additive response in the
other; doses of X-rays
were 1 Gy or 2 Gy with
the dose rate unspecified.
inorganic arsenic
exposure increased the
frequencies of MN in
binucleated cells and of
SCEs over what they
would have been with
UV alone, but there was
not a synergistic effect
forMN.
Induction of gene
mutations at the HPRT
locus:
While neither inorganic
arsenic treatment
induced mutations by
itself, as a post-treatment
these inorganic arsenic
treatments both caused
an ft in the mutation
frequency compared to
MNU alone; there was a
synergistic interaction.




Reference



Lee et al.,
1986




Jha et al.,
1992









Jha et al.,
1992






Li and
Rossman,
1989a










C-175  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

V79 cells
exposed to 5-
15 J/m2 of
UVC









V79 cells
exposed to
55-220 KJ/m2
ofUVA








V79 cells
exposed to:

400-800 J/m2
ofUVB

200 J/m2 of
UVB





Mouse
291.03C
keratinocytes
irradiated
immediately
after the
arsenic
treatment with
a single dose
of0.30kJ/m2
UV



Arsenic
Species

As111 SA












As111 SA













As111 SA
for both











As111 SA











Concentration(s)
Tested (nM)

10












10













10

5, 10, 15










2.5, 5.0











Duration of
Treatment

3hr












3hr













3hr

24 hr










24 hr











LOECa
(HM)

10












10













None

10










5.0








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of gene
mutations at the HPRT
locus:
While the inorganic
arsenic treatment
induced no mutations by
itself, as a post-treatment
it caused an ft in the
mutation frequency
compared to UVC
irradiation alone; there
was a synergistic
interaction.
Induction of gene
mutations at the HPRT
locus:
While the inorganic
arsenic treatment
induced no mutations by
itself, as a post-treatment
it caused an ft in the
mutation frequency
compared to UVA
irradiation alone; there
was a synergistic
interaction.
Induction of gene
mutations at the HPRT
locus:
While the inorganic
arsenic treatments
induced no mutations by
themselves, the 24-hr
post-treatment caused an
ft in the mutation
frequency compared to
UVB irradiation alone;
there was a synergistic
interaction.
Effect on repair rate of
UV-induced
photodamage to genomic
DNA measured at 2 and
6 hr after the UV
exposure ended:
U in repair rate of 6-4PPs
by 48%, but no effect on
the repair of CPDs.






Reference

Li and
Rossman,
1991










Li and
Rossman,
1991










Li and
Rossman,
1991











Wuetal.,
2005







C-176  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


TK6 cells










TK6 cells
irradiated
with 1 or c3
Gyof69
cGy/min
gamma
radiation at
beginning of
inorganic
arsenic
treatment























Arsenic
Species

As111 SA

As111 ATO











As111 SA


As111 ATO




























Concentration(s)
Tested (nM)


0.1, 1, 10
for both









0.1,1, 10
for both
































Duration of
Treatment


24 hr
for both









24 hr
for both
































LOECa
(HM)

10

10











1


1

























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of MN using
flow cytometry assay:
ft to 24.7% from 3. 4% in
control.
ft to 17.4% from 3. 4% in
control; the text noted
that it was sometimes
difficult to distinguish
between the MN and
necrotic cell fragments
due to toxicity at the
dose of 10 for SA and
ATO.
Induction of MN using
flow cytometry assay:

At dose of 1: 1 Gy,
10.2%; 3 Gy, 12.2%;
12.2% was significantly
higher than 9. 8% in
control. There was a
statistically significant
(additive) effect.

At dose of 1: 1 Gy,
10.0%; 3 Gy, 16.3%;
16.3% was significantly
higher than 9. 8% in
control. There was a
statistically significant
(possibly slightly
synergistic) effect.
Interpretation of results
at dose of 10 was
complicated by difficulty
of distinguishing
micronuclei and necrotic
cell fragments.
Responses were
extremely different for
the 2 arsenicals at dose
of3Gy:30.2%forSA
and only 15. 9% for
ATO.




Reference

Hornhardt
etal.,2006










Hornhardt
etal.,2006





























Cytotoxicity


NHEK cells


As111 SA

Asv,
MMAV,
DMAV

0.001,0.005,0.01,
0.05,0.1,0.5, 1,5,
lOforall

24 hr

24hr


0.005

0.5


Extent of viability
determined by neutral
red assay; viability was
significantly reduced.



Vegaetal.,
2001

C-177  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HOS cells

AG06 cells

W138 cells




WI38 cells



HaCaT cells







HepG2 cells










17 human
cancer cell
lines:
4 bladder cell
lines,
2 lung cell
lines,
2 liver cell
lines,
1 leukemia
cell line,
and various
others




Arsenic
Species


As111 SA,
Asv






As111 SA



As111 SA




Dimethyl-
arsinate,
the usual
form of
DMAvin
this table

Thio-
DMAV
(i.e.,
Thio-
dimethyl-
arsinate)

As111
ATO












Concentration(s)
Tested (nM)


IC50
determinations






0.25,0.5, 1,2



0.5, 1.0








0.01,0.1,0.5, 1,5,
10, 50 mM
for both







IC50
determinations












Duration of
Treatment


100 hr for all







7 days



20 passages








48 hr
for both








96 hr













LOECa
(HM)


—







0.25



0.5







0.5 mM





0.1 mM















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Extent of viability
determined by neutral
red assay:
IC5os:3.5forAsin, 11 for
Asv

IC50s: 1.1 for As111, 16 for
Asv
IC50s: 8.8 for As111, 30 for
Asv
Clonal survival
determined by crystal
violet assay:
LD50: -1.85.
ft resistance to
cytotoxicity caused by
exposure to
concentrations of As111 of
!-16|aMfor72hr.



Cell survival was
determined by WST-8
assay:
LC50s: regular DMA,
-0.2 mM;
Thio-DMA, -0.02 mM.
At 0.1 mM, regular
DMA showed no
cytotoxicity, but thio-
DMA resulted in only
22% cell survival.
Viability determined by
sulphorhodamine B
method:
Bladder: IC50s: 0.34,
0.47,0.93, 1.38.
Lung: :IC50s: 3.27, 4.17.
Liver:IC50s:5.17, 7.17.
Leukemia: IC50s: 0.64.
All 17 lines: LC50 range
was 0.34-7. 17. There
was a strong positive
correlation between GSH
content of cells and
magnitude of IC50:




Reference


Hu et al.,
1998






Vogtand
Rossman,
2001

Chien et
al., 2004









Ramlet
al., 2007






Yanget
al., 1999









C-178  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

4 of above 17
human cancer
cell lines with
high levels of
GSH









Hepa-1 cells
(mouse
hepatoma)

NHEK cells




AG06 cells










Human-
hamster
hybrid AL
cells




Primary
cultures of rat
cerebellar
neurons



Arsenic
Species

As111 ATO














As111 SA


As111 SA




As111 SA










As111 SA







As111 SA

DMAV




Concentration(s)
Tested (nM)

IC50
determinations













2, 5, 10, 25, 50


IC50
determinations



0.1,0.3, 1,3










0.8,3.8,7.7, 15.4







5, 10, 15

1, 5, 30 mM




Duration of
Treatment

96 hr













12 hr

24 hr

72hr




48hr










24hr







12 hr

48 hr




LOECa
(HM)

—













None

10

—




3
3
0.3
0.1







3.8







5

5 mM

Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Viability determined by
sulphorhodamine B
method:
10 |oM BSD, which
depletes cellular GSH,
was incubated with cells
for 4 days, causing them
all to become very
sensitive to arsenic, as
follows:
IC50s without BSD: 0.47,
2.59, 2.08, 9.89.
IC50s with BSD: 0.19,
0.14,0.40,0.20,
respectively.

Viability determined by
LDH release method
Extent of viability
determined by neutral
red assay:
IC50: 10.8
Extent of viability
determined by neutral
red assay:
Values below at 3 :
-90% of cells viable if
no pretreatment (pt) to
change GSH level.
-85% of cells viable if
NACpt to ft GSH level.
-20% of cells viable if
BSD pt to U GSH level.
-20% of cells viable if
CHE pt to U GSH level.
No. of colonies counted
to determine surviving
fraction: LC50 = about
7.7.
(Addition of BSD, which
suppresses GSH
markedly, increased
cytotoxicity.)
Viability determined
using MTT metabolism
assay.





Reference


Yanget
al., 1999












Maier et
al., 2000

Snow et
al., 1999



Snow et
al., 1999









Liuetal.,
2001






Namgung
and Xia,
2001

C-179  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


Chang human
hepatocytes






Chang human
hepatocytes






Chang human
hepatocytes





Raji cells
(human B-
lymphocytes)






Jurkat cells








A549 cells






Arsenic
Species

As111 SA,
Asv,
MMAm,
MMAV,
DMAV



As111 SA,
Asv,
MMAm,
MMAV,
DMAV



As111 SA,
Asv,
MMA111,
MMAV,
DMAV




As111 SA
MMA111
DMA111






As111 SA
MMA111
DMA111



As111 SA


Asv

DMAV



Concentration(s)
Tested (nM)


LCso
determinations






LCso
determinations






LCso
determinations




0.2, 1, 10, 20, 40,
100 for all








0.2, 1, 10, 20, 40,
100 for all





0.016, 0.08, 0.4,
2.0, 10

30, 100, 300

2, 20, 200, 2000



Duration of
Treatment


24hr







24 hr







24 hr







4hr
4hr
2hr






4hr
4hr
2hr





7 days for all






LOECa
(HM)


—







—







—







10
40
10






40
0.2
10



0.016


30

None
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
LC50s using LDH
leakage assay in
phosphate media:
As111: 68.0.
Asv: 1,628.
MMA111: 6.0.
MMAV: 8,235.
DMAV: 9,125.
LC50s using K+ leakage
assay in phosphate
media:
As111: 19.8.
Asv: 1,006.
MMA111: 6.3.
MMAV: 9,283.
DMAV:4,109.
LC50s using the XTT
assay in phosphate
media:
As111: 164.
Asv: 3,050.
MMA111: 13.6.
MMAV: 42,000.
DMAV: 91,440.
Extent of viability
determined by trypan
blue assay:
Viabilities at maximum
dose for each:
As111: -85%.
MMA111: -85%.
DMA111: 60%.
Extent of viability
determined by trypan
blue assay:
Viabilities at maximum
dose for each:

As111: -95%.
MMA111: -52%.
DMA111: -58%.
Colony-forming
efficiency assay with
Giemsa staining:

LC50s: As111, -0.08; Asv,
-100.




Reference


Petrick et
al., 2000






Petrick et
al., 2000






Petrick et
al., 2000





Gomez et
al.,
2005






Gomez et
al., 2005







Mass and
Wang,
1997

C-180  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

CHO Kl cells











CHO cells:

Wild-type

V850

R120
















CHO Kl cells

CHO cell
lines:

Kl (parental
to the
following
line)

XRS-5 (X-ray
and H2O2
sensitive)





Arsenic
Species

As111 SA













As111 SA
for all



















As111 SA




As111 SA












Concentration(s)
Tested (nM)

10













5, 10, 15, 20, 30,
50, 75, 100 for
most


















20, 40, 80




10, 20, 40 for both












Duration of
Treatment

4hr













48 hr for all




















4hr




4hr












LOECa
(HM)

None













5

20

10 (lowest
for it)















20



40


20







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Clonogenic survival
assay for cytotoxicity:
12-hr pretreatment with
BSO depletes GSH; with
BSO at 50 and 400 uM,
survival was 9% and 1%,
respectively; other
experiments showed that
an increase in GSH
markedly reduced the
cytotoxicity of an As111
treatment following UV
irradiation.
Comparative inhibition
of cell growth was based
on numbers of cells
present compared to
control:
V 850 cells were adapted
to 850 uM H2O2 over
about 4 months of
exposures to increasing
concentrations; R 120
cells had then been
cultured 4 months
without exposure to
H2O2. IC50 values: Wild-
type, 17.2; V 850, 62.45;
R 120, 26.6. Results
after pretreatment with
BSO suggest that
intracellular thiol levels
(GSH mainly) may
account for the arsenic
resistance seen in V 850
cells.
Colony formation assay


Clonogenic survival with
crystal violet staining:
ID50s: line Kl, 37.8;
line XRS-5, 17.0; the
much less responsive Kl
cells have 6 times as
much catalase activity as
XRS-5 cells; both lines
are similar in arsenic
uptake and release, in
GSH levels, and in GST
activity.




Reference

Huang et
al., 1993













Cantoni et
al., 1994


















Wanget
al., 1996


Wang and
Huang,
1994








C-181  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
CHO cell
lines:

Kl (parental
to the
following
lines)

XRS-6 (X-ray
sensitive)

XRS-5 (X-ray
andH2O2
sensitive)







CHO-9 cells










BFTC905
cells and
NTUB1 cells






CHO Kl cells








Arsenic
Species




As111 SA













As111 SA
Asv
MMAm
MMAV
DMA111
DMAV
TMAVO







As111 SA
Asv
MMA111
MMAV
DMA111
DMAV



As111 SA








Concentration(s)
Tested (nM)


20, 40, 80, 160


20, 40, 80, 160

20, 40, 80, 160












0.1, 1, 10, 100,
500
for all










IC50
determinations






20








Duration of
Treatment




4hr















Ihr
for all











7 days







6hr








LOECa
(HM)


160


80

20










1
1
500
100
0.1
500
None








—







20





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)


Inhibition of cell growth:
ID50 values were: Kl,
235; XRS-6, 108; XRS-
5,33.
Kl cells have 5.8 times
as much catalase activity
and 5.4 times as much
GPx activity as XRS-5
cells. Kl cells have 3.7
times as much catalase
activity and 2.1 times as
much GPx activity as
XRS-5 cells. The cells
with intermediate
amounts have an
intermediate response.
Extent of viability
determined by trypan
blue assay:

DMA111 was by far the
most cytotoxic at all
concentrations tested,
with the percentages of
living cells at 1, 10, and
100 being approximately
45, 41, and 0%,
respectively.
Clonogenic survival in a
colony -forming assay,
IC50 values in BFTC905
and NTUB1 cells,
respectively:
0.13,0.16.
9.25, 9.00.
0.13,0.15.
3.04, 2.64.
0.52,0.58.
0.38,0.63.
Colony -forming assay:
this concentration caused
~32 % survival;
squalene at up to 160 uM
had no effect on
cytotoxicity.




Reference




Wanget
al., 1997















Dopp et
al., 2004










Wang et
al., 2007






Fan et al.,
1996




C-182  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

CHO cells
treated with
MMS before
or after
inorganic
arsenic
treatment


CHO Kl cells
exposed to
1.5 or 2.5
J/m2ofUV



C-33A cells
HeLa cells

Jurkat cells
LCL-EBV
cells


Jurkat cells
and human
lymphocytes













Mouse
291.03C
keratinocytes





Arsenic
Species


As111 SA








As111 SA






As111 SA
for all





As111 SA
















As111 SA






Concentration(s)
Tested (nM)

5, 10, as
pretreatments

5, 10, as post-
treatments





10






0.1, 1, 10,25,50
for all





01 1 10 25 50
for both















0.05,0.1,0.5, 1,5






Duration of
Treatment

24 hr


24 hr






24 hr






24 hr






24 hr
for both















7 days






LOECa
(HM)

None


5






10





10
50

0.1
10



0.1
for both















0.5



Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Colony -forming assay:
No change from MMS
alone.


ft in cytotoxicity
compared to MMS alone,
synergistic interaction
with even less survival at
10.
Colony-forming assay:
Synergistic ft in
cytotoxicity because of
the inorganic arsenic
post-treatment.
Cell viability determined
by Trypan blue
exclusion:
-35% viability at 50.

-75% viability at 50.
-55% viability at 50.

-60% viability at 50.
Cell viability determined
by Trypan blue
exclusion:
When both of these cell
types were transfected
with mutant p53 genes
(by electroporation) there
was substantially
increased cytotoxicity.
This ft was already
apparent at a dose of 0. 1
(i.e., theLOEQinthel
p53 mutation tested in
Jurkat cells and in 1 of 2
p53 mutations tested in
PHA-stimulated
lymphocytes.
Cytotoxicity based on
colony survival, using
crystal violet staining:
LC50 = 0.9; almost all
dead at dose of 5.




Reference



Lee et al.,
1986






Yanget
al., 1992





Salazar et
al., 1997





Salazar et
al., 1997















Wuetal.,
2005


C-183  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Chinese
hamster V79
cells


A2780 cells

H460 cells

MCF-7 cells



BALB/c 3T3
cells
(derived from
mice)













G12 cells










U-2OS cells







Arsenic
Species
As111 SA

DMAV



As111 ATO
for all



As111 SA
Asv
MMAV
DMAV
TMAV












As111 SA

MMAmO

DMAmI








As111 SA







Concentration(s)
Tested (nM)
1,2,5, 10

-0.8, 1, 2, 5, 10
mM


IC50
determinations




IC50
determinations














0.05,0.1,0.5, 1,
2.5,5, 10
0.2, 0.4, 0.6, 0.8, 1

0.1,0.2,0.3,0.4








0.01,0.05,0.1,
0.25,0.5, 1,2.5






Duration of
Treatment
24hrs
for both





72hr





18 hr
















72 hr










10 days







LOECa
(HM)
1

2mM




—





—














1

0.2

0.1








0.05




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cytotoxicity based on
number of viable cells
compared to control:
LC50s: As111, -5.5; DMAV:
-3.5 mM.
Cell survival was
determined using the
MTT assay:
IC50 values: A2780,
2.80; H460, 14.60;
MCF-7, 3.00.
Cell survival was
determined using the
MTT assay:
IC50 values: As111 SA,
16.9; Asv, 64; MMAV,
14.7 mM; DMAV, 4.35
mM;
TMAV, >74 mM.
Depletion of GSH in
cells by co-treatment
with 0.2 mM BSD
markedly reduced the
cytotoxicity of DMAV
even though it markedly
increased the
cytotoxicity of the other
4 compounds.
Cell survival was
determined using the
clonal survival assay:
LC50 values: As111 SA,
~8;MMAmO, 0.51;
DMAmI, 0.15.
The 2 methylated forms
were also tested at 4 and
24 hr and showed
cytotoxicity at both; for
MMAmO, cytotoxicity
was >50% at both times
at highest dose.
Cell survival was
determined using the
clonal survival assay:
LC50 = 0.68; 100% cell
killing at 2.5.




Reference

Ochi et al.,
1999b




Ling et al.,
2002




Ochi et al.,
1994















Klein et
al., 2007









Komissaro
va et al.,
2005


C-184  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



U-2OS cells










U118MG
cells










Undifferentiat
edPC12 cells









FGC4 cells

HepG2 cells

Rat
hepatocytes

SVEC4-10
cells




Arsenic
Species



As111 SA
for all








As111 ATO












As111 ATO











As111 SA
for all




As111 SA





Concentration(s)
Tested (nM)



LC50
determinations









1, 5, 10, 25, 50











1, 10, 100, 1000










50, 75, 100, 125

25, 50, 75, 100,
125

2, 10,25,35,45,
55
2, 4, 8, 12, 16





Duration of
Treatment


24hr

48hr

72hr







24hr











24 hr











24 hr
for all




24 hr





LOECa
(HM)


—

—

—







5











1










75

50

25


4


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using the
clonal survival treat-and-
plate (CSTP), neutral red
(NR), and MTT assays,
for the different
durations:
LC50:CSTP, 1.1; MTT,
3.8;NR, 4.8.
LC50: CSTP, 0.9; MTT,
0.99; NR, 1.05.
LC50: CSTP, 0.8; MTT,
0.8; NR, 0.84.
Cell survival was
determined using the
MTT assay: slightly >
50% survival at dose of
5; co-treatment with
lipoic acid blocked
cytotoxicity. Other tests
showed no ft in either
apoptotic cell death or
intracellular peroxide
levels; cell death was
shown to be autophagic.
Cell survival was
determined using the
MTT assay: LC50 = 8.
(At dose of 8, about 75%
cell survival at 12 hr.)
Effects of pretreatment
or co-treatment with
antioxidants on
cytotoxicity: NAC: big
but a-Toc, GSH, 17(3-
estradiol, orBO653:
NSE.
Cell survival was
determined by the NR
uptake assay:
LC50s: FGC4, -90;
HepG2, -70;
hepatocytes, -30.

Cytotoxicity determined
by the MTT assay: LC50
= -13.




Reference



Komissaro
va et al.,
2005








Cheng et
al., 2007










Pigaetal.,
2007










Gottschalg
etal.,2006




Chao etal.,
2006a

C-185  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
HCT1 16 cells
(securin +/+)

HCT1 16 cells
(securin -/-)



RKO cells
(p53 wt)

SW480 cells
(p53 mutant)







U-2OS cells




























Arsenic
Species

As111 SA
for both






As111 SA
for both








As111 SA
for all




























Concentration(s)
Tested (nM)

4, 8, 12, 16
for both






8, 16, 24, 32
for both








0.1, 1, 10





























Duration of
Treatment

24 hr
for both






24 hr
for both








24 hr





























LOECa
(HM)
4


4




8

8








1 or 10;
see
explana-
tion























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cytotoxicity determined
by the MTT assay:
LC50s = securin +/+,
-17; securin-/-, -11.
There was significantly
more cytotoxicity in null
mutant at doses of 8, 12
and 16.
Cytotoxicity determined
by the MTT assay:
LC50s = RKO, -20;
SW480, -27. There was
significantly more
cytotoxicity in
wt p53 cell line at doses
of 16, 24 and 32.
Trypan blue exclusion
assay to identify necrotic
cells (which take up
stain) after additional
periods of post-treatment
culturing of 0, 24, or 48
hr in arsenic -free
medium:
At dose of 0.1, no
increase in necrotic cells
at any time. At dose of
1, necrotic cells were
-0%, -20%, and -40%
of total cells,
respectively. At dose of
10, necrotic cells were
-70%, -95%, and -95%
of total cells,
respectively. Note that a
24-hr treatment with S A
affected the amount of
necrosis at a dose of 1
only if there was an
additional 24-hr or
longer period of
culturing in SA-free
medium between the end
of the SA treatment and
when the assay was
done.




Reference

Chao etal.,
2006a






Chao etal.,
2006a








Komissaro
vaetal.,
2005
























C-186  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HaCaT cells

CRL1675
cells

THP-1 +
A23 187 cells




HaCaT cells

CRL1675
cells

THP-1 +
A23 187 cells





Alveolar
macrophages
(AMs) from
CDFi mice

Peritoneal
macrophages
(PMs) from
CDFj mice



















Arsenic
Species

As111 ATO
for all










As111 ATO
for all









As111 SA
Asv
MMAV
DMAV
TMAV























Concentration(s)
Tested (nM)

LD10andLD25
determinations
for each cell line









LD10andLD25
determinations for
each cell line








IC50
determinations


























Duration of
Treatment


72 hr









72 hr under
chronic
exposure
conditions








48 hr



























LOECa
(HM)


—











—









—
























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cytotoxicity assessed
using fluorescein
diacetate assay:
LD10= 1.9;LD25= 15.2.

LD10=1.0;LD25=1.9.

LD10=1.9;LD25 = 3.8.
Testing for cytotoxicity
was preceded by
exposure to 1.0 uM As111
ATO for at least 8
passages to establish
chronic-exposure
conditions. Then,
following exposures to
various doses for 72 hr,
cytotoxicity was
assessed using
fluorescein diacetate
assay:
LD10 = 2.0 ; LD25 = 4.0.
LD10 = 0.5;LD25=1.3.
LD10 = 0.5;LD25 = 5.1.
Cell survival was
determined using the
AlamarBlue assay (said
to be similar to the MTT
assay):
IC50 values of AM cells:
As111 SA, 4; Asv, 400;
MMAV, >10 mM; DMAV,
5 mM; TMAV, »10
mM.
IC50 values of PM cells:
As111 SA, 5; Asv, 650;
MMAV, >10 mM; DMAV,
5 mM; TMAV, »10
mM.
DMAV caused almost
entirely apoptotic cell
death, while the
inorganic arsenicals
caused mainly necrotic
cell death. SOD, CAT.
and a peptide inhibitor
ICE inhibited the
cytotoxicity of As111 but
notofDMAv




Reference


Graham-
Evans et
al., 2004







Graham-
Evans et
al., 2004









Sakurai et
al., 1998























C-187  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


RHMVE cells



















HLFC cells

HLFK cells
(Ku70
deficient)







Arsenic
Species

MMAV

DMAV

TMAVO

















As111 SA
for both









Concentration(s)
Tested (nM)


0.25,0.5, 1,2.5,5,
10, 25, 50, 100
mM
for all

















5, 10, 20, 40, 80
for both









Duration of
Treatment


24 hr




















24 hr
for both









LOECa
(HM)

25 mM

ImM

None

















10
for both






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using a
modified MTT assay:
LC50s: MMAV, 33.6 mM;
DMAV, 2.54 mM;
TMAVO, cell number
increased by dose of 1
mM, reaching 135% of
control at dose of 25
mM. Another study
showed LC50s: As111, 36;
Asv, 220 (both uM).
Co-treatment with NAC
caused U in cellular
arsenic content and
cytotoxicity by DMAV
but not by MMAV Co-
treatment with B SO
caused big ft in
cytotoxicity of MMAV
but slight U in
cytotoxicity of DMAV.
Viability was determined
by trypsin blue exclusion
assay:
LC5os: HLFC, 27.38;
HLFK, 21.80;
cytotoxicity was
significantly greater for
HLFK than HLFC at
doses of 20, 40 and 80.




Reference



Hirano et
al., 2004


















Liu et al.,
2007b






C-188  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


NB4 cells

NB4-M-AsR2
cells

IM9 cells















MCF-7 cells,
T47D cells,
and
MDA-MB-
231 cells












Arsenic
Species



As111 ATO
for all


















As111 ATO
for all















Concentration(s)
Tested (nM)


0.5, 1

2,4

0.5, 1

















IC50
determinations














Duration of
Treatment


6 days
for all




















3 days















LOECa
(HM)


0.5

4

0.5

















—












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell viability (% of
control) for ATO alone
and for ATO with 100
uM Trolox, determined
using trypan blue
exclusion:
At 0.5: 75% alone, 43%
with Trolox; at 1: 30%
alone, 3% with Trolox.
At 2: 100% alone, -80%
with Trolox; at 4: -63%
alone, -30% with
Trolox.
At 0.5: -80% alone,
-70% with Trolox; at 1:
-50% alone, -25% with
Trolox.
Thus, Trolox enhanced
ATO-induced
cytotoxicity (or growth
inhibition) in all 3 cell
lines.
Cell viability for ATO
without and with 100 uM
Trolox co-treatment,
respectively, determined
using trypan blue
exclusion assay:
MCF-7: 2.07 and 1.02;
T47D: 3.22 and 1.56;
MDA-MB-23 1:2.27 and
0.98.
Thus, co-treatment with
Trolox enhanced ATO
growth inhibition
similarly to what was
seen in the row above.




Reference




Diaz et al.,
2005


















Diaz et al.,
2005











C-189  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human
PBMCs
cultured in
various ways











MDAH 2774
cells





SVEC4-10
cells

1RB3AN27
cells












BEAS-2B
cells

HT1 197 cells







Arsenic
Species
As111 ATO














As111 ATO






As111 SA


As111 SA













As111 ATO


As111 SA







Concentration(s)
Tested (nM)
1














1, 2, 5, 8






5, 10, 20, 40


0.1,0.5, 1,5, 10













10, 20, 50


1, 5, 10, 25, 50







Duration of
Treatment
15 days














72hr






24hr


72hr













24 hr


24 hr







LOECa
(HM)
1














lor 2






10


1













10


10




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Colony-forming ability
was assessed for ATO
alone and for co-
treatment with Trolox by
counting CFU-
erythrocytes, CFU-
granulocytes-
macrophages, and BFU-
erythrocytes. Biggest
effect of ATO alone:
62% U for CFU-
erythrocytes. In all 3
cases, co-treatment with
Trolox had little or no
effect.
Cytotoxicity assessed
using trypan blue
exclusion assay:
uncertainty about LOEC
exists because control
value was not reported:
LC50 = 5.
Cell survival was
determined using the
MTT assay: LC50 = -13.
Cell survival was
determined using the
MTT assay: there
probably was
cytotoxicity at dose of 1;
statistically significant
cytotoxicity at dose of 5;
LC50 = ~8; all
experiments on ROS or
induction of transcription
factors were at doses of
<10 for <4 hr, and under
those conditions, there
was no cytotoxicity.
Cell survival was
determined using the
MTT assay: LC50 = -15.
Cell survival was
determined using the
trypan blue exclusion
assay:
LC50 = -35.




Reference

Diaz et al.,
2005












Terek et
al., 2006





Hsu et al.,
2005

Felix et al.,
2005












Hanetal.,
2005

Hernandez
-Zavala
etal.,
2005

C-190  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue



HL-60 cells

U266 cells













































Arsenic
Species



As111 ATO
for both














































Concentration(s)
Tested (nM)



1,2,3,5,10
for both














































Duration of
Treatment



24 hr
for both














































LOECa
(HM)



2

1













































Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using the
trypan blue exclusion
assay:
LC50s = HL-60, ~7;
U266, ~2.
Effects of modulators in
both cell lines:
(Cells were loaded with
high concentrations of
intracellular AA [icAA]
by incubating them with
DHA prior to inorganic
arsenic treatments, thus
avoiding generation of
extracellular ROS in
tissue culture media
caused by direct addition
to it of AA.) icAA
caused big U in
cytotoxicity of inorganic
arsenic. GSH depletion
by BSO treatment caused
big ft in inorganic
arsenic-induced
cytotoxicity. icAA
caused big U in
cytotoxicity caused by
inorganic arsenic in
GSH-depleted cells.
Extracellular AA caused
big ft in inorganic
arsenic-induced
cytotoxicity, including
after GSH depletion.
Relatively limited data
from a methylcellulose
colony-forming assay in
both cell lines (with 48-
hr inorganic arsenic
treatment and 10-14 days
to form colonies) and
from cytotoxicity testing
of RPMI-8226 cells
supported some of the
above conclusions.
Effect of NAC was
tested in HL60 cells; it
caused big U in inorganic
arsenic-induced
cytotoxicity.
Reference



Karasawa
s etal.,
2005













































C-191  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Embryonic
mesenchymal
cells prepared
fromCF-1
mouse
conceptuses at
gestation day
11




HCT15 cells
HeLa cells

PLC/PR/5
cells

Chang cells

K562 cells,
AR230-S
cells, AR230-
r cells,
KCL22-S
cells, KCL22-
r cells, NB4
cells














Arsenic
Species


As111 SA










As111 SA
for all







As111 ATO



















Concentration(s)
Tested (nM)


5.8, 11.6, 15.4,
30.8









LCso
determinations







IC50
determinations
for all

















Duration of
Treatment


2hr










24 hr








3 days



















LOECa
(HM)


5.8










—








—
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using the
MTT assay: LC50 = -27;
15-min pretreatment with
0.5% (v/v) DMSO
completely blocked the
inorganic arsenic effect
at dose of 15.4, whereas
15-min pretreatment with
0.1% or 0.2% (v/v)
DMSO partially blocked
it.

Cell survival determined
by MTT cell
proliferation assay:
LC50s: HCT15, 278.33;
HeLa, 200.33;
PLC/PR/5, 376.66;
Chang, 328.33.
Antiproliferative activity
as determined by MTS
assay — some would
interpret such results as
cytotoxicity and present
results as LC50s:
IC50s: K562, 0.9;
AR230-S, 2.6; AR230-r,
6.9; KCL22-S, 2.6;
KCL22-r, 2.8; NB4, 0.4.
A dose of 2 represents
the upper margin of the
clinically useful range
for ATO. There was a
positive correlation
between GSH content of
cells and resistance to
the antiproliferative (i.e.,
cytotoxic) effect.




Reference

Perez-
Pasten et
al., 2006









Othumpan
gat et al.,
2005






Konig et
al., 2007















C-192  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

AR230-S
cells,
AR230-r
cells,
KCL22-S
cells, KCL22-
r cells


















H1355 cells









Arsenic
Species


As111 ATO























As111 ATO









Concentration(s)
Tested (nM)


1























3.125,6.25,12.5,
25, 50, 100, 200








Duration of
Treatment


2 days























24 hr









LOECa
(HM)


None























6.25






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined by trypan
blue assay:
100 uM BSO treatment
was shown to greatly U
GSH levels in all 4 cells
types both with and
without inorganic arsenic
exposure. In all 4 cell
types, the inorganic
arsenic + BSO treatment
caused big to huge U in
number of viable cells,
whereas untreated cells
or cells treated with
inorganic arsenic or BSO
showed ~2-fold H A
similar assay in primary
cultures of mononuclear
cells from 4 patients in
blast crisis with imatinib-
resistant CML also
showed maximum
cytotoxicity for the
combined inorganic
arsenic + BSO treatment.
Cell survival was
determined using the
MTT assay:
Cytotoxicity increased
with dose, with -57%
cell survival at dose of
200.




Reference


Konig et
al., 2007






















Cheng et
al., 2006





C-193  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue
TRL1215
cells = X in
this row

TRL1215
cells that had
been treated
with 1.3 mM
MMAV for 20
weeks prior to
acute arsenic
treatments =
Y in this row















Arsenic
Species


As111 SA,
Asv,
DMAV
for both






















Concentration(s)
Tested (nM)



LCso
determinations
for both






















Duration of
Treatment



48 hr
for both























LOECa
(HM)



—
























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival based on
AB assay:
LC50sforAsni:X, 16.3;
Y, 74.1.
LC50s for Asv:X, 157.1;
Y, 2743.8.
LC50s for DMAV: X,
2090; Y, 6950. Thus the
MMAV pretreatment
caused marked resistance
to cytotoxicity for all 3
arsenicals. Much of this
resistance was lost if Y
cells were cultured for 8
more weeks with no
arsenic in media. The
20-week pretreatment
caused no cytotoxicity,
gave the Y cells an
arsenic content of 135.4
+ 12.0 ng/mg cellular
protein, and did not
induce malignant
transformation.
Arsenicals were not
methylated or
demethylated in these
cells.
Reference



Kojima et
al., 2006























C-194  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue
TRL1215
cells = X in
this row

TRL1215
cells that had
been treated
with 0.7 mM
DMAV for 20
weeks prior to
acute arsenic
treatments =
Y in this row
























Arsenic
Species



As111 SA,
Asv,
DMAV
for both






























Concentration(s)
Tested (nM)




LC50
determinations
for both






























Duration of
Treatment




48 hr
for both































LOECa
(HM)




—
































Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival based on
AB assay:
LC50sforAsni:X, 16.3;
Y, 19.2.
LC50s for Asv:X, 157.1;
Y, 182.2.
LC50s for DMAV: X,
2090; Y, 4730.
Thus the DMAV
pretreatment caused
marked resistance to
cytotoxicity for only the
DMAV treatment, and the
slight differences for the
other 2 arsenicals were
not statistically
significant. WhenY
cells were cultured for 8
more weeks with no
arsenic in media, there
was no change regarding
the lack of resistance to
As111, but the resistance to
the other 2 arsenicals
increased substantially.
The 20-week
pretreatment caused no
cytotoxicity, gave the Y
cells an arsenic content
of41.8 + 2.5ng/mg
cellular protein, and did
not induce malignant
transformation.
Arsenicals were not
methylated or
demethylated in these
cells.
Reference




Kojima et
al., 2006































C-195  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue
TRL1215
cells = X in
this row

TRL1215
cells that had
been treated
with 10.0 mM
TMAVO for
20 weeks
prior to acute
arsenic
treatments =
Y in this row

















Arsenic
Species


As111 SA,
Asv,
DMAV
for both

























Concentration(s)
Tested (nM)



LCso
determinations
for both

























Duration of
Treatment



48 hr
for both


























LOECa
(HM)



—



























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival based on
AB assay:
LC50sforAsni:X, 16.3;
Y, 54.8.
LC50s for Asv:X, 157.1;
Y, 684.1.
LC50s for DMAV: X,
2090; Y, 4500. Thus the
TMAVO pretreatment
caused marked resistance
to cytotoxicity for all 3
arsenicals. Much of this
resistance was lost
regarding DMAV, and all
of it was lost regarding
the other 2 arsenicals, if
Y cells were cultured for
8 more weeks with no
arsenic in media. The
20-week pretreatment
caused no cytotoxicity,
gave the Y cells an
arsenic content of 543.8
+ 12.0 ng/mg cellular
protein, and did not
induce malignant
transformation.
Arsenicals were not
methylated or
demethylated in these
cells.
Reference




Kojima et
al., 2006

























C-196  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



NB4 cells
























293 cells





















Arsenic
Species
As111 SA

As111 ATO

MMAm

DMA111




















As111 ATO






















Concentration(s)
Tested (nM)
1,2,3,4

1,2,3,4


0.25,0.5, 1,2

2, 4, 6, 8



















0.5,1,2,3,4






















Duration of
Treatment



72 hr























12 days






















LOECa
(HM)
1

1


0.25

4



















1



















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using the
MTT assay:
LC50s: As111 SA, -3.4;
As111 ATO, -2.2; MMA111,
-1.2; DMA111, -5.8.
Co-treatment (CoTr)
with 3000 uM DTT
markedly decreased
cytotoxicity of all
arsenicals: Maximum
cytotoxicities with 3000
uM DTT CoTr:
As111 SA, -17%; As111
ATO, -12%; MMA111,
-25%; DMA111, -12%.
CoTr with 100 uM DTT
markedly increased
cytotoxicity of all
arsenicals:
LC50s with 100 uM DTT
CoTr: As111 SA, -2.2;
As111 ATO, -1.0;
MMA111, -0.28; DMA111,
-4.0.
Cell survival was
determined by colony-
forming assay (% of cells
forming colonies): -73%
at dose of 4; LC2s =
-3.6.
Co-treatment with 200
uM DMSA increased
survival: -87% at dose
of 4.
Co-treatment with 20 uM
DMSA decreased
survival: -61% at dose
of 4;
LC25 = ~1.6. Co-
treatment with 100 uM
DMPS increased
survival: -86% at dose
of 4. Co-treatment with
10 uM DMPS decreased
survival: -50% at dose
of4;LC25 = ~1.2.




Reference



Jan et al.,
2006























Jan et al.,
2006

















C-197  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


SV-HUC-1
cells
















HeLa cells



Primary rat
hepatocytes

A431 cells




RAW264.7
cells




Arsenic
Species


As111 ATO

















As111 SA



As111 SA


As111 ATO




As111 SA





Concentration(s)
Tested (nM)


0.5,1,2,3,4

















10, 100



2.5,5,7.5, 10, 15,
20, 25, 30, 40, 50

1.25,2.5,5, 10,20
for both



2.5, 5, 10, 25





Duration of
Treatment


12 days

















24hr



24hr


24hr

48 hr


24 hr





LOECa
(HM)


0.5

















10



7.5


2.5

1.25


2.5


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined by colony-
forming assay (% of cells
forming colonies): -62%
at dose of 4; LC25 =
-2.2. Co-treatment with
200uMDMSA
increased survival: -73%
at dose of 4; LC25 =
-3.5. Co-treatment with
20 uM DMSA decreased
survival: -43% at dose
of4;LC25 = ~1.4. Co-
treatment with 100 uM
DMPS increased
survival: -79% at dose
of 4. Co-treatment with
10 uM DMPS decreased
survival: -47% at dose
of4;LC25 = ~1.2
Cell survival determined
using a LIVE/DEAD
viability/cytotoxicity kit:
LC50: ~95.
Cell survival was
determined using the
MTT assay: LC50 = -18.
Cell survival was
determined using the
MTT assay:
At 24 hr: LC50 = -20.
At48hr:LC50 = ~3.
Cell survival based on
neutral red uptake assay:
LC50 = ~13.




Reference


Jan et al.,
2006
















Hansen et
al., 2006


Hasgekar
etal.,2006

Huang et
al., 2006



Szymczyk
et al., 2006

C-198  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
NIH 3T3 cells


















NHEK cells







BAEC cells














H22 cells

BAEC cells






Arsenic
Species
As111 SA


















As111 SA








As111 SA












As111 ATO
for both








Concentration(s)
Tested (nM)
5, 10, 20, 50, 100,
200

















0.2, 0.4, 0.8







1, 5, 10














0.5, 1,2,4
for both







Duration of
Treatment
6hr


















1, 2, 3, 4 days







24 hr

48 hr












24 hr, 48 hr

24 hr, 48 hr






LOECa
(HM)
20 for U

















0.2 for ft
on all
days






5

1












1,0.5

2,1



Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell viability assayed
using CellTiter-Glo
assay: possibly slight ft
at 5 and 10;
U at 20, LC50 = ~90.
Pre-induction of HSP by
conditioning heat shock
(2 hr at 42°C on prior
day) or by constitutive
expression of HSP70
markedly reduced the
cytotoxicity, as follows:
with heat: LOEC = 100
and -80% viability at
dose of 200, with
constitutive expression:
LOEC = 50 and -70%
viability at dose of 200.
Cell survival was
determined using the NR
uptake assay:
1ho~l.l-1.4x at doses
of 0.2 and 0.4 on all
days; point estimates at
dose of 0.8 were always
higher than control, but
the ft was always a NSE.
Cell survival was
determined using a
variation of the MTT
assay:LC50s:~7.5at24
hr, ~5.0at48hr. Unlike
co-treatment with Zn11,
Fen,orCun,
only co-treatment with
Mn11 increased inorganic
arsenic toxicity at
concentrations at which
it (the metal) did not
cause cytotoxicity alone.
Cell survival (also called
the proliferation index)
was determined using the
MTT assay:
LC50sforH22:~2.0at
24 hr, -1.2 at 48 hr.
LC50sforBAEC:~4.5at
24 hr, ~2 at 48 hr




Reference
Khalil et
al., 2006

















Hwang et
al., 2006







Bunderson
etal.,2006












Liu et al.,
2006e




C-199  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HEK 293
cells

HEK 293
cells
transfected
withOATP-C
















U937 cells








TRL 1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
50uMBSO



Arsenic
Species

As111 SA
Asv
MMAV
DMAV



















As111 SA
for all









MMAV
for both












Concentration(s)
Tested (nM)


LCso
determinations




















0.5, 1,2.5,5, 10,
20
for all








1.25,2.5,5, 10
mM
for both











Duration of
Treatment


72 hr




















24 hr

48 hr

72 hr







48 hr
for both












LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was determined using the
MTT assay:
LC50s without and with OATP-C,
respectively: As111: 10.9, 5.6; Asv:
98.1,53;
MMAV: 43 19.3, 421 1.6 (this
comparison: NSE);
DMAV: 994.1, 899.3 (this comparison:
NSE).
The OATP-C transfected cells
accumulated 43% more As111 and 34%
more Asv than the non-transfected
cells while they did not accumulate
more of the methylated arsenicals.
Co-treatment of the As111- or Asv-
treated cells with rifampin or
taurocholic acid eliminated the
difference between the two cell types.
OATP-C can transport inorganic
arsenic in a (GSH)-dependent manner
but this may not be the major pathway
for arsenic transport.

20

10

10





10 mM



2.5 mM








Cell survival was
determined using the Pi-
exclusion assay:
At 24 hr, -74% survival
at dose of 20.

At 48 hr, -62% survival
at dose of 20.
At 72 hr, -40% survival
at dose of 20, LC50:
-17.5.
Cell survival was
determined using the AB
assay:
Without BSO: -80% cell
survival at dose of 10
mM; at 5 mM, survival
may
have been higher than
that of control
LC50 with BSO: 3.2 mM.
Similar results were
obtained using CV assay.





Reference




Luetal.,
2006


















McCollum
etal.,2005










Sakurai et
al., 2005a








C-200  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Gclm+/+ MEF
cells

Gclm+/" MEF
cells

Gclm'7- MEF
cells, from
GCLM
knockout
mice







HeLa cells

U937 cells

Primary
human skin
fibroblasts
























Arsenic
Species


As111 SA
for all















As111 ATO
for all




























Concentration(s)
Tested (nM)


4, 8, 16, 32, 64
for all
















2 for all




























Duration of
Treatment


8hr
for all
















3 days




























LOECa
(HM)
16 or 64

16 or 64

4













2

2

None























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using the
MTT assay:
It was unclear which of
the first two genotypes
had the LOEC of 16; one
had an LOEC of 64, and
the LOEC for the other
one was 16;
LC50s: +/+, 86; +/-, 86; -
/-, ii;
pretreatment with tBHQ
protected Gclrn 7~ and
Gclm+/" MEF cells from
inorganic arsenic-
induced cytotoxicity in a
dose- and time-
dependent manner.
Cell survival was
determined using the
MTS assay:
-77% survival in HeLa
and -85% survival in
U937; no hint of
cytotoxicity in
fibroblasts. Co-
treatment with 10 uM
emodin apparently
sensitized HeLa and
U937 cells (but not
fibroblasts) to
cytotoxicity. The
addition of 1.5 mM NAC
to the co-treatment of
HeLa cells with 10 uM
emodin and 2 uM
inorganic arsenic
eliminated all
cytotoxicity; effect of
NAC was not tested in
U937 cells. Emodin was
used because it has a
semiquinone structure
that is likely to increase
the generation of
intracellularROS.




Reference



Kannet
al., 2005b















Yi et al.,
2004
























C-201  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


MCF-7 cells












MYP3 cells



























Arsenic
Species

As111 ATO










As111 SA

Asv
MMAm

DMA111
DMAV
TMAVO























Concentration(s)
Tested (nM)


0.5, 1, 2, 4, 8, 16









2,3

35,40
1,1.5

0.6, 1
0.6 mM, 1 mM
15 mM, 20 mM























Duration of
Treatment

24 hr, 48 hr, or
96 hr












7 days
for all


























LOECa
(HM)
2 at 24 hr;
Iat48
and
96 hr








2

35
1

0.6
0.6 mM
15 mM




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

Cell survival was
determined using the
MTTassay:LC50sat24,
48, and 96 hr were 8.6,
3. 3, and 1.86,
respectively. Apoptosis
was shown to be the
mechanism of cell death
after treatment with a
dose of 5 for 3 days.
Cell survival was
determined using the
MTT assay:
-33% at 2, -9% at 3.
-37% at 35, -28% at 40.

-60% at 1, -7% at 1.5.
-28% at 0.6, -10% at 1.
-45% at 0.6 mM, -28%
at 1 mM.
~28%atl5mM,~18%
at 20 mM.
Co-treatments with
antioxidants that work by
different mechanisms
yielded the following
results: melatonin
slightly inhibited
cytotoxicity of As111.
NAC inhibited
cytotoxicity of MMAm,
DMA111, DMAV and
TMAVO. Vitamin C
inhibited cytotoxicity of
As111, Asv, MMAm and
DMA111. Tironand
Trolox did not affect
cytotoxicity of any
arsenical.




Reference


Ye et al.,
2005












Wei et al.,
2005






















C-202  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Primary
keratinocytes
(in third
passage)
obtained from
foreskins of
adults





Huh? cells









CL3 cells,
synchronous
atGl

CL3 cells,
asynchronous
(asyn)

CL3 cells,
synchronous
atS

CL3 cells,
synchronous
atG2/M




Arsenic
Species


As111 SA









As111 SA













As111 SA
for all













Concentration(s)
Tested (nM)


0.001,0.01,0.1, 1,
5, 10, 100, 1000








0.5,1,3,5, 10,20













50, 100
for all













Duration of
Treatment
24hr

48hr

72hr







24 hr













3hr
for all













LOECa
(HM)
SforU

lft;5U

0.1 ft;
5U





1 for ft
20 for U













—











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined using the
XTT assay:
ft in viability
(proliferation) to 1.1 8x
and 1.32x at dose of 1 at
48 and 72 hr,
respectively.
LC50s at 24, 48, and 72
hr were -160, -10, and
-4.2, respectively.
Cell survival was
determined using the
MTT assay: ftto-l.lxat
doses of 1 and 3; U to
58% at dose of 20. In
co-treatments with lOnM
TCDD, inorganic arsenic
doses of 5 and 10 caused
0% and 10%
cytotoxicity,
respectively.


Cell survival was
determined using a
colony-forming assay:
% survival at dose of 50:
Gl, 45%; asyn, 35%;
S, 29%; G2/M, 17%.
Survival at dose of 50 in
Gl cells was cut from
45% to 25% to 30% by
co-treatment with
PD98059orU0126,
which are these 2
structurally dissimilar
inhibitors of MEK1/2.




Reference


Liao etal.,
2004








Chao etal.,
2006b












Li et al.,
2006a










C-203  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human
neuroblastom
a cell lines:
IMR-32
SK-N-DZ
SK- N-BE(2)
SK-N-AS
SH-SY5Y
All 4 lines ±
co-treatment
with 25 uM
DCHA












HaCaT cells



















Arsenic
Species



As111 ATO















As111 SA



Asv




MMAm



DMA111











Concentration(s)
Tested (uM)



1














0.5,1,1.5,2.5,4,
6, 7, 8, 10, 12, 13,
14, 16, 18, 20, 22

10, 20, 30, 40, 50,
60, 80, 100, 120,
160, 200, 240,
280, 320, 360

0.1,0.5, 1,2,2.5,
3,3.5,4,4.5,5,
5.5,6,6.5,8, 10


0.1,0.5, 1,2,3,4,
5,6,7,8,9, 10, 11










Duration of
Treatment



48 hr




















24 hr
for all


















LOECa
(UM)
-or +
DCHA
None, 1
None, 1
1,1
1,1
1,1











0.5 for ft
12 for U

10 for ft
100 for U


0.1 for ft
2. 5 for U

None for
ft
3 for U










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival (% of
control) was determined
using the MTT assay:
inorganic arsenic alone.
DCHA alone, (inorganic
arsenic + DCHA)
NSE, NSE, 35%.
NSE, NSE, 45%.
73%, NSE, 41%.
56%, NSE, 39%.
61%, NSE, 40%.
co-treatment of
(inorganic arsenic
+DCHA) with vitamin E
blocked much of the
cytotoxicity in line IMR-
32.
Cell survival was
determined using the
MTT assay, with a
proliferative effect being
seen at lower doses:
Peak of 141% at dose of
1; first point estimate
below 100% at dose of 8;
about 50% cytotoxicity
at 22.


Peak of 145% at dose of
10; first point estimate
below 100% at dose of
80; about 50%
cytotoxicity at 320.
Peak of 160% at dose of
1; first point estimate
below 100% at dose of
2.5; about 50%
cytotoxicity at 4. 5.
About 60% cytotoxicity
at 11.




Reference




Lindskog
etal.,2006


















Ganyc et
al., 2007















C-204  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




Mouse
lymphoma
cells
(L5178Y/Tk+/-
-3.7.2Ccells)









V79 cells
treated with
MNU






V79 cells
exposed to
UVA, UVB,
orUVCover
a wide range
of doses



Human-
hamster
hybrid AL
cells



Arsenic
Species
As111 SA




Asv




MMAV


DMAV




As111 SA








As111 SA







As111 SA






Concentration(s)
Tested (nM)
2.3,3.1,4.6,6.2,
7.7, 9.2, 10.8,
11.5, 13.1, 13.9,
14.6, 15.4

6.1, 15.2,30.3,
45.5, 48.5, 54.6,
60.6, 66.7, 69.7,
72.8, 78.8, 84.9

12.3, 15.4, 18.5,
21.6, 24.7, 27.8
mM

12.5, 18.8,25.0,
31.3,37.5,43.8,
50.0, 56.3, 62.5
mM
10
5







10







3.8,7.7, 15.4






Duration of
Treatment





4hr
for all











3hr
24 hr







3hr







1 day or 5 days






LOECa
(HM)
4.6




15.2




12.3 mM


18.8 mM




10
5







10







3.8



Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)


Cell survival, as percent
of relative total growth
compared to the vehicle
control:
Estimates of LC5os:
As111 SA, -7.3 uM; Asv,
-50.3 uM;
MMAV, ~16.1mM;
DMAV, -38.8 mM.







Cell survival, percent of
control: both inorganic
arsenic treatments caused
4% or less cytotoxicity;
however, as post-
treatments they both
considerably increased
the cytotoxicity caused
by the MNU treatments.
Cell survival, percent of
control:
The inorganic arsenic
treatments caused 8% or
less cytotoxicity;
however, as post-
treatments they increased
the cytotoxicity caused
by the UV treatments.
Colony-forming assay;
-55% survival with 1-
day treatment at 7.7.





Reference





Moore et
al., 1997a











Li and
Rossman,
1989a







Li and
Rossman,
1991





Heietal.,
1998


C-205  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


1T1 cells

MYP3 cells













AG06 cells






Human cells:
AG06
(keratinocytes
)
AG06
(keratinocytes
)
HaCaT
(keratinocytes
NHEK
(keratinocytes
GM847
(fibroblasts)
WI38
(fibroblasts)

AG06 cells



K562 cells





Arsenic
Species

As111 SA
Asv
MMAmI2
MMAV
DMAmI
DMAV
TMAVO










As111 SA







As111 SA
MMAm
As111 SA
As111 SA
As111 SA
As111 SA









As111 SA

MMAm

As111 ATO





Concentration(s)
Tested (nM)



LC50
determinations for
all












0.2, 4, 20









IC50
determinations












1, 5, 10, 20, 30


2.5





Duration of
Treatment



7 days














24 hr
pretreatment








48 hr













5hr


12 hr





LOECa
(HM)



—













0.2 for ft


4forU







—













—


2.5


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
LC50s based on trypan
blue assay for viability:
AsmSA:4.8inlTl;0.4
inMYP3.
Asv:31.3inlTl;5.3in
MYP3.
MMAmI2: 1.0 in IT 1; 0.8
inMYP3.
MMAV: 1.7mMinboth
lines.
DMAmI: 0.8 in 1T1; 0.5
inMYP3.
DMAv:0.50mMinlTl;
LlmMinMYPS.
TMAVO: l.VmMin
lTl;4.5mMinMYP3.
Extent of viability
determined by NR assay:
ft in viability over that
seen for MNNG alone at
-1-15 (oM MNNG.
U in viability below that
seen for MNNG alone at
-1 5-40 (oM MNNG
(synergistic interaction).
Extent of viability
determined by NR assay:
IC50: 7.2.
IC50: -7.5.
IC50: 11.6.
IC50: 12.3.
IC50: 10.7.
IC50: 11.2.







Extent of viability
determined by NR assay:
-20 kills 20% of cells.

-20 kills 50% of cells.
-50% of cells die.






Reference



Cohen et
al., 2002













Snow et
al., 1999








Snow et
al., 2001











Snow et
al., 2001


Li and
Broome,
1999
C-206  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human
kidney
carcinoma
cell lines:
UOK123
UOK109
UOK121

Human lung
carcinoma
cell line:
A549

HFW cells
(diploid
human
fibroblasts)








HFW cells
(diploid
human
fibroblasts)
V79-C13
Chinese
hamster cell
line



Syrian
hamster
embryo cells











Arsenic
Species




As111 SA








As111 SA











As111 SA



As111 SA





As111 SA


Asv











Concentration(s)
Tested (nM)




IC50
determinations







2.5, 5, 10, 20











1.25,2.5,5, 10

5, 10, 20, 40, 80

5, 10, 20, 30, 40,
50,60




-0.7, 1.4, 2, 3, 4,
5,6

~5, 10, 20, 50, 75,
100, 130, 160, 200










Duration of
Treatment




7 days








6hr











24 hr

4hr

24 hr






7 days
for all












LOECa
(HM)




—








2.5











1.25

-10

10





0.7ft, 5U


10ft, lOOli








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Extent of viability
determined by colony-
formation efficiency
assay:

0.020.
0.021.
0.020.



0.4.
Cytotoxicity determined
by a colony-forming
assay; co-treatment with
catalase (but not heat-
inactivated catalase) at
100 ug/mL markedly
reduced cytotoxicity;
increasing GSH levels
with B-mercaptoethanol
reduced cytotoxicity;
decreasing GSH levels
with B SO increased
cytotoxicity.
Cytotoxicity determined
by a colony-forming
assay.

Cytotoxicity determined
by a colony-forming
assay: survival at 10 was
76.3 ±2.61% of control;
IC50: ~20.
Cytotoxicity determined
by measuring CFE:
Small but reproducible ft
from 0.7 to about 1.5
followed by a
logarithmic decrease in
CFE with a linear
increase in dose.
Small but reproducible ft
from 10 to 50 followed
by a logarithmic
decrease in CFE with a
linear increase in dose.




Reference



Zhong and
Mass,
2001







Lee and
Ho, 1995










Yihand
Lee, 1999


Sciandrello
etal.,2002





Barrett et
al., 1989









C-207  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

UROtsa cells













UROtsa cells









UROtsa cells









BALB/c 3T3
A3 1-1-1 cells
(derived from
mice)




TK6 cells




HL-60 cells






Arsenic
Species

As111 SA













MMAm
for all




As111 SA

Asv

MMAmO

MMAV

DMAmI

DMAV

As111 SA

AsvDA

MMAV

DMAV

As111 SA


As111 ATO



As111 ATO






Concentration(s)
Tested (nM)

0.1, 10,25,50,
100, 200












0.5, 1,2,5, 10
for all




0.1,0.5, 1,5

1,200

0.1,0.5, 1,5


1,200

0.1,0.5, 1,5

1,200
2, 5, 10, 15, 20

10, 15, 20, 25, 30

1,2,5, 10 mM

0.5, 1, 2, 5mM

0.1,0.5, 1, 10,
100, 1000

0.1, 1, 10, 100


0.2, 0.4, 0.8, 1.6,
3.1,6.3, 13,25,
50, 100





Duration of
Treatment

24 hr













24 hr

48 or 72 hr







24 hr
for all









72 hr for all





24 hr
for both




24 hr






LOECa
(HM)

50













5

5



None

None

5


None

None

None
5

10

5 mM

ImM

10


10



0.8



Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Viability determined
using MTT assay: IC50 =
-100; doses <10 were
said to stimulate
mitochondria! activity
(i.e., the curve went up;
the assay tests
mitochondria! function),
but the stimulation was
not statistically
significant.
Co-treatment with BSO:
big ft in cytotoxicity,
withIC50 = ~15.
Viability determined
using MTT assay:
IC50 = ~5.

All cells (or almost all
cells) were dead at
LOEC.
Viability determined
using MTT assay:

WithMMAmO: 50%
cytotoxicity was
estimated to result from
dose of about 2.5, with
about 90% cytotoxicity
at dose of 5.



Cytotoxicity based on
percent cell growth
compared to treatment
with distilled water:
IC50 values: As111 SA, 4.8;
AsvDA, 17;
MMAV, 9.8 mM; DMAV,
3.2 mM.
Cytotoxicity based on
trypan blue exclusion
assay:

For both: LC50 between 3
and 4.

Viability determined
using MTT assay:
LC50 = 32.





Reference

Bredfeldt
etal.,2004












Bredfeldt
etal.,2006








Drobna et
al., 2002








Tsuchiya
etal.,2005





Hornhardt
etal.,2006




Yedjou
and
Tchounwo
u, 2007
C-208  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

IEC cells
(primary
culture)

IEC-6 cells

















MDAH 2774
cells








HPBMs
exposed to
M-CSF for 7
days and
considered
M-
macrophages
HPBMs
exposed to
GM-CSF for
7 days and
considered
GM-
macrophages



Arsenic
Species


As111 SA
for both



















As111 ATO









As111 SA
Asv
MMAV
DMAV



As111 SA
Asv
MMAV
DMAV






Concentration(s)
Tested (nM)


7.7, 15, 38, 77,
116, 154 for both



















1, 2, 5, 8










LCso
determinations





LCso
determinations







Duration of
Treatment


24 hr
for both



















72 hr










48 hr






48 hr








LOECa
(HM)

15


15
















Ior2
(uncer-
tain since
control
not
shown)







—






—





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Viability determined
using MTT assay:
At dose of 77: IEC,
-45% dead; IEC-6,
-55% dead; the
cytotoxicity of the 2 cells
types was almost
identical at most doses;
based on this and their
rather similar
concentration-dependent
declines in membrane
enzymes and constituents
(e.g., alkaline
phosphatase, hexose,
sialic acid, cholesterol,
and phospholipid), the
primary and established
cultures gave
approximately similar
toxic responses.

Cytotoxicity estimated
by XXT proliferation
assay
and alternatively by
trypan blue dye-
exclusion assay (for
which treatment time
was either 72 or 96 hr —
it was unclear from
methods): IC5o by both
methods: 5.
Viability based on AB
assay:
LC50 values: As111, 7.0;
Asv, 1900;
MMAV, 2500; DMAV,
800.

Viability based on AB
assay:
LC50 values: As111, 5.8;
Asv, 2800;
MMAV, 2000; DMAV,
2000.





Reference



Upreti et
al., 2007


















Askar et
al., 2006









Sakurai et
al., 2006





Sakurai et
al., 2006




C-209  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
The following
cell lines: HL-
60, U-937,
TIG-112,
CRL-1609,
RAW264.7,
mouse normal
embryo cells,
mouse
embryo cells
that were MT
+/+and
MT -/-, and
the following
3 types of
human
immune cells:
peripheral
T-
lymphocytes,
immature
dendritic cells
and multi-
nucleated
giant cells
GM04312C
cells








Arsenic
Species






As111 SA


















As111 SA









Concentration(s)
Tested (nM)






LC50
determinations

















2.5, 10, 50









Duration of
Treatment






48 hr


















24 hr









LOECa
(HM)






—


















2.5






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)



Viability based on AB
assay:
LC50 values: HL-60, 13;
U-937, 12;
TIG-112, 25; CRL-1609,
17; RAW264.7, 25;
MT +/+ cells, 4.8; MT -
/- cells , 5.8;
T-lymphocytes, 3.3;
dendritic, 8.2; giant, 2.3.












Viability based on
neutral red assay: LC50 =
—20. However, when
viability was based on
colony-forming assay:
LC5o = ~6withLOECof
2.5




Reference





Sakurai et
al., 2006


















Shen et al.,
2006





C-210  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


Primary
mouse
hepatocytes



















SV-HUC-1
cells



JB6C141
cells

JB6C141
cells exposed
to 0.1, 0.2,
0.5, 1,2,3,4,
5, 6,7, or 8
kJ/m2ofUVB
at end of
pretreatment
with inorganic
arsenic






Arsenic
Species



As111 SA


















As111 SA

MMAm

DMA111




As111 SA
for both















Concentration(s)
Tested (nM)



60, 100, 200


















0.5, 1,2,5, 10

0.1,0.25,0.5, 1

0.25,0.5, 1,2,5


0.1, 1,5, 10,20,
50, 100, 500, 1000


10














Duration of
Treatment



24 hr




















3 days for all






24 hr for both
















LOECa
(HM)



60


















1

0.25

0.5


5



10











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Viability determined
using MTT assay: LC50
= -200 (LC50 = 30 for
48-hr treatment).
Pretreatment with SFN
caused big U in
cytotoxicity. SFN
activates transcription
factor Nrf2 and causes
significant ft of protein
expressions responsible
for excretion of arsenic
into extracellular space.
SFN caused big ft in
intracellular GSH levels
and big U in intracellular
arsenic levels. Also,
pretreatments with BSO,
EA, orMK-571,whichft
arsenic accumulation in
hepatocytes, caused big
ft in cytotoxicity.

Viability determined by
SRB assay:
LC50 values: As111, 2.91;
MMAm, 0.46;
DMA111, 1.59.
Viability determined by
MTS assay:
LC50 = -15, decreased
with dose until reached
-12% of control at top 3
doses.



Probably some
cytotoxicity at UVB dose
of 5, and there was
significant cytotoxicity at
UVB dose of 6.
Viability was -70% of
control at highest UVB
dose.




Reference



Shinkai et
al., 2006



















Suetal.,
2006






Tang et al.,
2006











DNA Damage
WRL-68
(human
hepatic cell
line)

As111 SA




0.001,0.01,0.1,
10



16 hr




0.001




Induction of DNA-
protein crosslinks
(methylated forms of
arsenic could not be
detected in the cells).
Ramirez et
al., 2000



C-211  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human aorta
VSMCs
(vascular
smooth
muscle cells)



PHA-
stimulated
and
unstimulated
human
lymphocytes















L-132 cells
(human
diploid
alveolar
epithelial type
II cells)





L-132 cells
(human
diploid
alveolar
epithelial type
II cells)









Arsenic
Species

As111 SA






AsmATO



















As111 SA

MMAV

DMAV






As111 SA

MMAV

DMAV











Concentration(s)
Tested (nM)

-1.2,2.5,5, 10







10


















100

100

5, 10, 100







100
for all













Duration of
Treatment

4hr







2hr



















6hr
for all









3hr
for all













LOECa
(HM)

-1.2







10


















None

None

5






None

None

100








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
DNA strand breaks
(double and single strand
breaks and alkali-labile
sites) detected by comet
assay; the effect was
similar in
nonproliferating VSMCs.
Oxidative damage to
DNA measured by the
comet assay, including
SSBs — after digestion
with FPG, arsenic-
induced base damage
was converted to a large
increase in SSBs and
some FPG-created
DSBs. (FPG cleaves
purines including 7,8-
dihydro-8-oxoguanine
(8-oxoG),
formamidopyrimidines,
and AP sites.) Like the
damage induced by
H2O2, arsenic-induced
DNA damage was
repaired more slowly in
unstimulated
lymphocytes.

Induction of DNA S SB
resulting from inhibition
of repair polymerization
by polymerization
inhibitors aphidicolin
and hydroxyurea. DMAV
induced them in a dose-
dependent manner
(measured by alkaline
elution).
Induction of DNA repair
synthesis using the BrdU
photolysis assay (single-
strand DNA breaks
induced by UV-
irradiation were
measured by alkaline
elution). Follow-up
experiment with same
DMAV treatment for 1, 3,
or 6 hr showed increases
with longer durations of
treatment.




Reference

Lynn et al.,
2000






Li et al.,
2001


















Yamanaka
etal., 1997









Yamanaka
etal., 1997










C-212  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

L-132 cells
(human
diploid
alveolar
epithelial type
II cells)











4>X174 RF I
DNA
Naked
double-
stranded
circular DNA






Human
primary
peripheral
blood
lymphocytes


















Arsenic
Species


MMAV













As111 SA

MMAm


DMA111


DMAV




As111 SA
Asv
MMAm
MMAV
DMA111
DMAV


















Concentration(s)
Tested (nM)

100 with 10 mM
SAM present













0.1, 1, 10, 100,
300 mM
10, 15, 20, 25, 30,
60 mM

40, 80, 150, 250
(oM

0.1, 1, 10, 100,
300 mM



1-1000
1-1000
1.25-80
Not reported-875
1.4-91
Not reported- 1000


















Duration of
Treatment


6hr















2hr
for all










2hr
for all





















LOECa
(HM)


100













None

30 mM


150 (oM


None




Not
reported
for any of
them

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of DNA repair
synthesis using the BrdU
photolysis assay (single-
strand DNA breaks
induced by UV-
irradiation are measured
by alkaline elution).
This and other evidence
strongly suggests that the
DNA damage was not
directly induced by
MMAV but by
dimethylated arsenic that
was produced
metabolically by reaction
ofMMAvwithSAM.



Nicked DNA in DNA
nicking assay.





Breaks and/or alkali-
labile lesions in DNA
detected in the single-
cell gel comet assay —
the relative potencies
based on slopes are
shown below (the larger
the number, the bigger
the effect):
As111 1
Asv 1.
MMAm 77
MMAV <1
DMA111 386
DMAV <1
As111 and Asv caused a
significant effect, and
they were not
significantly different
from each other. MMAm
and DMA111 were thus 77
and 386 times more
potent in causing DNA
damage than SA.




Reference

Yamanaka
etal., 1997
















Mass et
al., 2001










Mass et
al., 2001

















C-213  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
E. coli
WP2s(X)
(lonn, sulAi,
trpE65,
uvrA155,
lamB+)





Raji cells
(human
B-
lymphocytes)










Jurkat cells








Arsenic
Species
MMAm

DMA111






As111 SA

MMA111

DMA111









As111 SA

MMA111

DMA111






Concentration(s)
Tested (nM)
0.01,0.10, 1.0, 10
for all









0.2, 1, 10, 20, 40,
100 for all










0.2, 1, 10, 20, 40,
100 for all









Duration of
Treatment
Overnight
for all







4hr

4hr

2hr









4hr

4hr

2hr






LOECa
(HM)
None

None






10

0.2

10









10

0.2

10



Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Assay to test for
induction of prophage
with and without
exogenous metabolic
activation:
No statistically
significant induction of
prophage by either
compound.
Extent of DNA damage
detected by single-cell
gel electrophoresis
(comet) assay:
At 0.2 and 1: MMA111 »
DMA111 = As111.
At 100: all 3 chemicals
had roughly the same
level of DNA damage as
MMA111 had at 0.2, but
MMA111 still has
significantly more DNA
damage than the other
two chemicals.
Extent of DNA damage
detected by single-cell
gel electrophoresis
(comet) assay:
At 0.2 and 1: MMA111 »
DMA111 = As111.
At 40 and 100: DMA111 >
MMA111 > As111.




Reference
Kligerman
et al., 2003









Gomez et
al., 2005












Gomez et
al., 2005




C-214  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
NB4 cells

HL-60 cells

CHO-K1 cells



















Human
peripheral
blood
lymphocytes
from 2
donors, with
results
reported
separately















Arsenic
Species
As111 SA
for all





















As111 SA

MMAm


DMA111



















Concentration(s)
Tested (nM)
0.25,0.5, 1

0.25,0.5, 1

0.25,0.5, 1,2


















5, 10

2.5, 5, 10, 20, 40,
80, 100

2.5, 5, 10, 20, 40,
80


















Duration of
Treatment
4 hr for all
























4 hr for all






















LOECa
(HM)
0.25

0.25

None


















None

2.5


5
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
The LOECS shown are
for DNA strand breaks
(termed ADSB by the
authors) detected by the
comet assay without any
additional treatments of
DNA to digest and
reveal ODA or DPC.
They also treated the
damaged DNA with FPG
or PK to yield estimates
of ODA or DPC,
respectively. TheLOEC
was 0.25 for all 3 cell
types for ODA, DPC, or
ODA+DPC. Clearly
much more DNA
damage is revealed by
treatments with FPG,
PK, or both. DNA
damage was induced at
levels causing no
cytotoxicity.
The LOECs apply to the
extent of DNA damage
detected by SCGE
(comet) assay at pH >
13. There was no
cytotoxicity at doses up
to 20. Much lower
responses for all
arsenicals were seen in
comet assay at pH of
12.1, with the difference
between this and pH 13
being defined as alkaline
labile sites. DNA
damage by both
methylated arsenicals
was markedly reduced
by co-exposures to the
antioxidants Se-Met or
vitamin C. DNA-double
strand breaks were not
induced.




Reference



Wanget
al., 2001




















Soto-
Reyes et
al., 2005

















C-215  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
MRC-5 cells







MRC-5 cells














Arsenic
Species
As111 SA







As111 SA














Concentration(s)
Tested (nM)
2.5, 5, 10







2.5, 5, 10














Duration of
Treatment
2hr







2hr














LOECa
(HM)
2.5





2.5 for
SSBs
10 for
protein-
DNA
adducts













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
DNA SSBs detected by
the standard alkaline (pH
> 13) comet assay: ft
with dose of both tail
length and tail moment at
doses of 2.5 and 5, but a
U for both effects at dose
of 10 to less than effect
seen at dose of 2.5. NSE
on cytotoxicity at any of
the tested doses.
Protein-DNA adducts
and DNA SSBs detected
by alkaline (pH > 13)
comet assay done with
and without
posttreatment with
proteinase K,
respectively:
Experiment without
proteinase K: ft of both
tail length and tail
moment at doses of 2.5
and 5, but a U of both
effects at dose of 10 to
less than effect seen at
other doses.
Experiment with
proteinase K: ft of both
tail length and tail
moment at doses of 2.5
and 5, and a further large
ft in both parameters at
dose of 10. NSE on
cytotoxicity at any of the
tested doses in either
experiment. Evidence
for protein-DNA adducts
(or crosslinks) came
from ft observed at dose
of 10, which is thus the
LOEC for that effect.
Proteinase K breaks the
crosslinks that hinder the
DNA fragmentation
caused by the DNA
SSBs.

Reference
Mouron et
al., 2006







Mouron et
al., 2006













C-216  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


MRC-5 cells















MRC-5 cells



























Arsenic
Species


DMAV















DMAV



























Concentration(s)
Tested (nM)


125, 250, 500















125, 250, 500



























Duration of
Treatment


2hr















2hr



























LOECa
(HM)
500 for U
in SSBs
(see row
below)











125 for
both
protein-
DNA
adducts
and SSBs






















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
DNA SSBs detected by
the standard alkaline (pH
> 13) comet assay: slight
ft in tail moment (TM) at
dose of 125 (a NSE);
point estimates of TM
were below control at 2
higher doses, with that at
500 being significantly
below it; actual data:
TMs: 0, 13.4; 125, 14.6;
250, 13.1; 500, 9.7. NSE
on cytotoxicity at any of
the tested doses.
Protein-DNA adducts
and DNA SSBs detected
by alkaline (pH > 13)
comet assay done with
and without
posttreatment with
proteinase K,
respectively:
Experiment without
proteinase K (buffer
only): progressive U in
tail moment (TM) with
increasing dose; actual
data: TMs: 0, 7.7; 125,
6.7; 250, 5.3; 500, 4.9.
Experiment with
proteinase K: ft in TM,
with a positive dose-
response; actual data:
TMs: 0,8.3; 125, 11.9;
250, 22.2; 500, 23.3.
NSE on cytotoxicity at
any of the tested doses in
either experiment.
Proteinase K breaks the
crosslinks that hinder the
DNA fragmentation
caused by the DNA
SSBs.




Reference


Mouron et
al., 2005














Mouron et
al., 2005























C-217  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

L-132 cells
(human
alveolar type
II cells)











HepG2 cells








NB4 cells














HL-60 cells










Arsenic
Species

DMAV














As111 SA






As111 SA
As111 ATO
MMAm
DMA111












As111
ATOc










Concentration(s)
Tested (nM)

5,7.5, 10 mM














7.5








0.5














12.5, 25, 50










Duration of
Treatment

12 hr














24 hr








30 min














24 hr










LOECa
(HM)

5mM














7.5








0.5














12.5







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
DNASSB detected by
alkaline elusion: there
was a dose-response.
Early in the exposure
period, there was marked
suppression of
replicative DNA
synthesis, and the chain
length of the nascent
DNA was shorter than
that of the control, which
suggests that the
template DNA was
modified by more than
just strand breaks.
Induction of DNA DSBs
by immunodetection of
yH2A.Xfoci:
ft to ~6x control level;
co-treatment with 170
nM SAM did not change
the induced DSB
frequency.
Experiments with EN111,
FPG and NE (from NB4
cells) as well as
experiments using
immunodepletion of NE
with antibodies directed
against proteins known
to be involved in
excision repair suggest
that these trivalent
arsenicals induce only
oxidative DNA adducts
and that OGG1,MYH
and APE are involved in
the excision of the
oxidative DNA adducts.
DNA damage detected
by alkaline SCGE
(comet) assay: while the
response was barely
statistically significant at
the lowest dose, it was
strong at the other 2
doses, with a positive
dose-response.




Reference

Tezuka et
al., 1993













Ramirez et
al., 2007







Puetal.,
2007












Yedjou
and
Tchounwo
u, 2007





C-218  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


PARP-1+/+
MEF cells

PARP-r7'
MEF cells







PARP-1+/+
MEF cells

PARP-1"'"
MEF cells







HaCaT cells

CRL1675
cells

THP-1 +
A23 187 cells






Arsenic
Species


As111 SA
for both










As111 SA
for both











As111 ATO
for all










Concentration(s)
Tested (nM)


11.5,23
for both










11.5,23
for both







72-hr LD10 and
LD25 for each cell
line:
1.9, 15.2

1.0, 1.9

1.9,3.8








Duration of
Treatment


30 min
for both










24 hr
for both












72 hr for all










LOECa
(HM)


11.5

11.5









11.5

11.5









15.2

None

1.9





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Extent of DNA damage
detected by SCGE
(comet) assay at pH >13,
reported as induced
damage (experimental -
control) in units of TM
length:
-0.4 at 11. 5, -0.7 at 23.
-2.9 at 11. 5, -3. 4 at 23.
All 4 estimates were
statistically significant.
Extent of DNA damage
detected by single-cell
gel electrophoresis
(comet) assay at pH >13,
reported as induced
damage (experimental -
control) in units of tail
moment length:
-2.0 at 11. 5, -3. 6 at 23.
-4.8 at 11.5, -5. 5 at 23.
All 4 estimates were
statistically significant.
DNA single-strand
breaks detected by
SCGE (comet assay)
following alkaline
treatment:

NSE at LD10; U at LD25
(perhaps stimulates
repair).

NSEatLD10;NSEat
LD25.
ft at LD10; ft at LD25.




Reference



Poonepalli
et al., 2005









Poonepalli
et al., 2005











Graham-
Evans et
al., 2004






C-219  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue






HaCaT cells

CRL1675
cells

THP-1 +
A23 187 cells









293 cells






















Arsenic
Species






As111 ATO
for all













As111 ATO























Concentration(s)
Tested (nM)
72-hr LD10 and
LD25 for each cell
line under
chronic-exposure
conditions, as
follows:
2.0, 4.0

0.5, 1.3

0.5,5.1











1






















Duration of
Treatment



Under chronic
exposure
conditions:

72 hr for all














6hr






















LOECa
(HM)






2.0

1.3

0.5











1



















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Testing for DNA single-
strand breaks was
preceded by exposure to
1.0uMAsmATOforat
least 8 passages to
establish chronic-
exposure conditions.
Then, following
exposures to various
doses for 72 hr, DNA
single-strand breaks
were detected by single-
cell gel electrophoresis
(comet assay) following
alkaline treatment:
ftatLD10;ftftatLD25.
NSEatLD10;ftatLD25.
ftft at LD10; ftft at LD25.
DNA damage reported in
units of tail moment in a
comet assay that used
nuclear extraction
incubation: untreated =
-11 units; dose of 1 : big
ft to -58 units.
Effects of co-treatment
(CoTr) with modulators
at high doses:
CoTr 200 uMDMSA:li
from inorganic arsenic
alone to -38 units.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -39 units.
Effects of CoTr with
modulators at low doses:
CoTr 20 uM DMSA: ft
from inorganic arsenic
alone to -104 units.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -84 units.




Reference




Graham-
Evans et
al., 2004















Jan et al.,
2006


















C-220  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




SV-HUC-1
cells




















UROtsa cells







UROtsa cells








E. coli strain
WP2S(X)







Arsenic
Species



As111 ATO





















As111 SA

MMAm





As111 SA

MMAm







As111 SA








Concentration(s)
Tested (nM)



1





















1, 10

0.05,0.5,5





1, 10

0.05,0.5,5







Up to 3.2mM








Duration of
Treatment



6hr





















30 min
for both






60 min
for both








20 hr








LOECa
(HM)



1





















1

0.05





10

0.05







None





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
DNA damage reported in
units of tail moment in a
comet assay that used
nuclear extraction
incubation: untreated =
-10 units; dose of 1: big
ft to -49 units.
Effects of CoTr with
modulators at high
doses:
CoTr 200 uMDMSA:ll
from inorganic arsenic
alone to -34 units.
CoTr 100 uMDMPS:U
from inorganic arsenic
alone to -35 units.
Effects of CoTr with
modulators at low doses:
CoTr20uMDMSA:ft
from inorganic arsenic
alone to -99 units.
CoTrlOuMDMPS:ft
from inorganic arsenic
alone to -89 units.
Detection of 8-oxo-dG
(measure of oxidative
DNA damage):
ft to 3 x control at 1, ft to
2x control at 10.
ft to 5x control at 0.05, ft
to 4x control at 0.5, NSE
at 5.
Detection of 8-OHdG
formation (measure of
oxidative DNA damage):
NSE at 1, big U from
control at 10.
ft to 3x control at 0.05, ft
to 3. 3x control at 0.5, ft
to 4.3 x control at 5.
Thus MMAm showed a
time delay just as it did
for ROS production.
No induction of I phage
(part of "SOS" system)
using 8 serial 2-fold
dilutions from a
concentration that
inhibits growth.




Reference




Jan et al.,
2006




















Eblin et
al., 2006






Eblin et
al., 2006







Rossman
etal., 1984




C-221  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human-
hamster
hybrid AL
cells


HaCaT cells











TK6 cells














4>X174 RF I
DNA
Naked
double-
stranded
circular DNA



Arsenic
Species
As111 SA




As111 SA

Asv









As111 SA

As111 ATO











As111 un-
specified

MMAm

DMA111





Concentration(s)
Tested (nM)
30.8




5, 10, 20, 30

10, 20, 30, 50, 100









0.1, 1, 10
for both












10|aM-30 mM in
log increments

10, 20, 30, 40, 50

37.5, 75, 150, 300,
1000




Duration of
Treatment
24 hr




24 hr











22 hr
for both















24 hr
for all






LOECa
(HM)
30.8




10

20









1


10











None

10

37.5


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induction of 8-OHdG;
co-treatment with SOD
or catalase considerably
reduced induction of this
oxidative DNA damage.
Induction of 8-OHdG;
pre-incubation with
SOD, CATorDMSO
almost completely
blocked this.
Oxidative DNA damage
by 20 uM As111 SA: pre-
incubation with
MnTMPyP, Z-NAME or
FeTMPyP substantially
blocked such damage.
Induction of DPCs
detected by a decrease in
DNA damage detected in
the comet assay when an
arsenic treatment was
followed by exposure to
lor2Gyof69cGy/min
gamma radiation. The
DPCs kept the damaged
DNA from moving
during electrophoresis.
While both SA and ATO
caused a significant
effect, the effect was
more pronounced for SA.



Nicked DNA in DNA
nicking assay.







Reference
Kessel et
al., 2002




Ding et al.,
2005









Hornhardt
etal.,2006














Nesnow et
al., 2002




C-222  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




Supercoiled
DNA
(plasmid pBR
322);
similar results
were found
for plasmid
cpX174, but
details were
not reported














Arsenic
Species
As111 SA
Asv
MMAm
MMAV
Mono-
methyl-
arsine

DMA111
DMAV

Dimethyl-
arsine
Tri-
methyl-
arsine












Concentration(s)
Tested (nM)
>5mM
>5mM
>5mM
>5mM

> 5 mM


<5 mM
>5mM


<0.5mM

<0.5mM













Duration of
Treatment






2hr
for all




















LOECa
(HM)
None
None
>5mM
None

> 5 mM


<5 mM
None

<0.5mM

<0.5mM











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)



Damage to DNA
detected by agarose gel
electrophoresis:
The arsines were
produced in aqueous
reaction mixtures of
sodium borohydride and
the appropriate arsenical.
Trimethylarsine and
dimethylarsine were
about 100 times more
potent than DMA111.
WhenNADHor
NADPH, which are
biological hydride
donors, were incubated
with DMA111 for 2 hr,
DNA damage was
increased by at least 10-
fold, possibly because of
the generation of
dimethylarsine.




Reference






Andrewes
et al., 2003

















DNA Repair Inhibition or Stimulation


CHO Kl cells

















As111 SA

















5, 10, 20, 40, 80

















6 hrs

















5















DNA single-strand
breaks detected by
alkaline elution: those
induced by MMS were
repaired after incubation
in a drug-free medium
for 6 hr; however,
posttreatment with
sodium arsenite
accumulated MMS-
induced breaks with a
dose-response for the
arsenite exposure. Both
alkali-labile sites and
frank breaks were
enhanced, with the latter
occurring at higher doses
of MMS and arsenite.


Lee-Chen
etal., 1993














C-223  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

V79 cells,
strain 743 -3 -6














HeLa S3 cells


























Arsenic
Species

As111 SA
for both










MMAm


MMAV


DMA111


DMAV





















Concentration(s)
Tested (nM)

10 uM











0.0001,0.001,
0.01,0.1, 1

0.01,0.1, 1, 10,
100, 500

0.0001,0.001,
0.01,0.1

0.01,0.1, 1, 10,
100, 250




















Duration of
Treatment

Stu-












Por all:
18hr + 5
min more
while also
being treated
with
100 uM
H202






















LOECa
(UM)

10 uM











0.001


None


0.001


None


















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Similar decreases in
inducible total nuclear
DNA ligase activity and
in inducible nuclear
DNA ligase II activity
were demonstrated after
arsenic treatments given
before or after MNU
treatments, thereby
demonstrating that most
of the inhibited activity
was DNA ligase II.
Effect on H2O2-induced
poly(ADP-ribosyl)ation:
U with dose, 59% of
control at dose of 1.


NSE.


U with dose, 49% of
control at dose of 0. 1 .

NSE.
Other experiments
showed that the above
effects were real
decreases (not merely
delayed responses). All
above measurements
were at dose levels with
little to no cytotoxicity.
After 18 hr incubation,
these arsenicals had NSE
on the extent of gene
expression of PARP-1 at
doses up to 0.1 and 100
for methylated and
pentavalent arsenicals,
respectively.




Reference

Li and
Rossman,
1989b














Walter et
al., 2007





















C-224  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Isolated
recombinant
PARP-1











Jurkat cells








Jurkat cells













HLFC cells

HLFK cells
(Ku70
deficient)





Arsenic
Species

As111 SA

MMAm

DMA111











As111 SA








As111 SA














As111 SA
for both







Concentration(s)
Tested (nM)


10, 50, 100, 200,
500
for all












0.01,0.1, 1,5, 10








1














1,2.5,5, 10
for both







Duration of
Treatment
For all:
10 min
preincubation
before
PARP-1
reaction with a
nicked plasmid
as substrate









24hr








24 hr














2hr
for both







LOECa
(HM)

10

10

10











0.01








1













2.5

1




Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Effect on activity of
PARP-1:
U with dose, 58% of
control at dose of 500.

U with dose, 24% of
control at dose of 500.

U with dose, 15% of
control at dose of 500.
These data suggest that
trivalent arsenicals
inhibit cellular
poly(ADP-ribosyl)ation
by reducing PARP-1
activity.
U ERCC1 mRNA level;
not said to be statistically
significant until dose of
1, but means + SDs
suggest 45% U at 0.01
and 60% U at 0.1.
Decreases of 60%, 95%,
and 85% at doses of 1, 5,
and 10, respectively
U in repair following a 2-
hr in vitro treatment with
4 uM 2-AAAF
immediately after the
inorganic arsenic
treatment. DNA damage
measured by SCGE
(comet) assay:
inorganic arsenic group
had ft DNA damage after
2-hr 2-AAAF treatment
and following a
4-hr repair period. No
difference in DNA
damage before 2-AAAF.
DNA DSB damage as
measured with neutral
SCGE assay:
This type of damage was
significantly greater for
HLFK than HLFC at all
4 doses.




Reference



Walter et
al., 2007











Andrew et
al., 2006








Andrew et
al., 2006













Liu et al.,
2007b




C-225  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
HLFC cells


HLFK cells
(Ku70
deficient)











CHO-K1 cells





GM847 cells

HaCaT cells






W138 cells
for both












Arsenic
Species

As111 SA
for both














As111 SA





As111 SA
for both






As111 SA
for both













Concentration(s)
Tested (nM)

5 for both















0.1,0.5, 1,5, 10





0.1,0.5, 1,5, 10
for both






0.1,0.5, 1,5, 10
for both













Duration of
Treatment

2hr
for both














24 hr





24 hr
for both






24 hr

48 hr












LOECa
(HM)
5


5













0.1





0.1

0.1





0.1

0.1









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
The LOECs are for
induction of DNA DSBs.
After the 2-hr As111
treatment, cells were
incubated in arsenic-free
medium to measure
repair of DNA DSBs
using the neutral SCGE
assay at 0.5, 1, 1.5, and 2
hr. At all time points
there was significantly
and substantially less
repair in HLFK, showing
that the Ku70 deficiency
decreases the efficacy of
repair of arsenic-induced
DSBs.
DNA polymerase (3
promoter activity: big ft
at 0.1; slight ft at 0.5; no
effect at 1; big U at 5 and
10.
DNA polymerase (3
protein levels:
Big ft at 0.1 and 0.5,
slight ft at 1, no effect at
SandbigUatlO.
Big ft at 0.1 and 0.5, no
effect at 1, big U at 5 and
10.
DNA ligase activity:
ft at 0.1, big ft at 0.5,
huge ft at 1, U at 5, big U
at 10.
No effect at 0.1, big ft at
0.5 and 1, no effect at 5,
big U at 10.
Two other experiments
of 72 and 96 hr duration
showed generally even
more subdued increases
and decreases than the
48-hr experiment.




Reference


Liu et al.,
2007b













Snow et
al., 2005




Snow et
al., 2005







Snow et
al., 2005









C-226  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Arsenic
Species



Concentration(s)
Tested (nM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)




Reference
Effects Related to Oxidative Stress (ROS)
Hepa-1 cells
(mouse
hepatoma)
stably
transformed
with
pEpREpgeo






WI38 (human
fibroblasts)






Purified
thioredoxin
enzyme from
mouse liver;
to test the
NADPH-
dependent
reduction of
DTNB


Primary
culture of rat
hepatocytes


Human-
hamster
hybrid AL
cells





As111 SA











As111 SA







As111 SA
MMAm
DMA111
Asv
MMAV
DMAV






AsmSA

MMAm

As111 SA








0.1,1,5,25,50








0.05,0.5,5
(24 hr
pretreatment)
followed by 60
min
exposure to H2O2
at
1, 10 or 50 mM
for 1 hr and then
24 hr to recover

-0.2-800
-0.2-800
-0.2-800
-10-6000
-10-6000
-10-6000






1-50

0.1-10

30.8








6hr










24 hr
pretreatment






—












30 min
for both


Within 5 min








5










0.05








-100
-0.2
~3
-300
	
—






—



30.8







Activated a |3-
galactosidase gene
reporter system: suggests
there was induced
oxidative stress — 5.6-
fold response;
progressively and
markedly decreasing
responses at 2 higher
doses.
Extent of viability
determined by NR assay:
Compared to control
cells exposed to H2O2,
with no pretreatment:
ft viability at 1 mM H2O2
only. At dose of 5, there
was an ft in viability at
10 mM H2O2 but a U in
viability at 50 mm H2O2.
Approximate IC50s
(inhibition of enzyme
activity):
-200.
-0.4.
-30.
-3000.
Never more than —80%
inactivation.
Never more than —80%
inactivation.
Decreased thioredoxin
enzyme activity (the
NADPH-dependent
reduction of DTNB)
IC50: » 100.
IC50: ~3.
Production of ROS,
measured by ESR and
with about a 3 -fold
increase in amplitude of
signals; concurrent
treatment with the
radical scavenger DMSO
eliminates the effect.

Maier et
al., 2000










Snow et
al., 2001








Linetal.,
1999








Lin et al.,
2001



Liu et al.,
2001





C-227  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human aorta
VSMCs
(vascular
smooth
muscle cells)



HFW cells
(diploid
human
fibroblasts)










Jurkat cells






















Arsenic
Species

As111 SA






As111 SA











As111 SA

MMAm

DMA111




















Concentration(s)
Tested (nM)

-1.2,2.5,5, 10





1.25,2.5,5, 10

5, 10, 20, 40, 80












0.2, 10, 20, 100
for all





















Duration of
Treatment

4hr





24 hr

4hr










2hr

2hr

2hr




















LOECa
(HM)

-1.2





1.25

20










None

10

10

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Numerous experiments
in this study led to the
conclusion that arsenite
activates NADH oxidase
to produce superoxide,
which then causes
oxidative DNA damage.
Micronuclei were
induced in both
protocols; the yield of
micronuclei was greatly
reduced by the presence
of the antioxidants
catalase or NAC (the
precursor of GSH),
which suggests that
oxidative stress was
involved in the induction
of micronuclei.
Level of intracellular
peroxides determined by
flow cytometry using
cell permeable
fluorogenic marker
DHR123:
At 10 and 20: DMA111 »
MMAin»Asm.
AtlOO:MMAin>DMAm
about equal to As111.
(Cell lysis may explain
the reduction of DMA111
at dose of 100 to 1/3
level seen at 20.)
Control value was not
reported. If control
value was actually 0 (and
thus the baseline in the
figure), then the LOEC
for all 3 arsenicals would
have been 0.2, with a
rather similar slight
response for all of them.




Reference

Lynn et al.,
2000





Yihand
Lee, 1999












Gomez et
al., 2005


















C-228  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Whole blood
lymphocytes
from 2 human
donors, with
results
reported
separately










HaCaT cells












L-132 cells










Arsenic
Species
MMAm


DMA111












As111 SA

Asv











DMAV










Concentration(s)
Tested (nM)

2.5, 5, 10, 20 for
both














5, 10, 15, 20
for both










10 mM alone

10 mM + 0.5 mM
PQ








Duration of
Treatment

4 hr for all















24hr











2hr


Ihr








LOECa
(HM)
2.5


10












5

10










None


10 mM





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Levels of MDA as lipid
peroxidation marker in
human plasma:
For MMA111 both donors
showed significant
increase over control at
all doses except 10, for
which only 1 was
significant. For DMA111
both donors showed
significant increase over
control at 20, but only 1
did at 10. There was no
cytotoxicity at the dose
levels tested.
Induction of 3 -NT,
which is a diagnostic
marker for RNS in vivo;
pre-incubation with
SOD, MnTMPyP, L-
NAMEorFeTMPyP
almost completely
blocked this protein
damage by 20 uM As111
SA; pre-incubation with
CAT or DMSO had no
effect, in sharp contrast
to what happened for
ROS-damage to DNA.

DNA single-strand
breaks detected by
alkaline elusion: co-
exposure with PQ or
sequential exposures of 1
hr (with either one first)
yielded a strong
response.




Reference

Soto-
Reyes et
al., 2005













Ding et al.,
2005











Kawaguchi
etal., 1996






C-229  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

PAEC cells
harvested
from freshly
isolated
vessels













NB4 cells









PAEC from
freshly
harvested
vessels
HFW cells







Cell free
buffer











Arsenic
Species


As111 SA















As111 ATO








As111
probably
ATO, but
called
arsenite

As111 SA







DMAmI












Concentration(s)
Tested (nM)


5, 10
















1









5



5, 10, 20







—












Duration of
Treatment


Ihr
















4hr









5-15 min



24 hr







—












LOECa
(HM)


5
















1









5



5







—









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Various experiments
showed that inorganic
arsenic activates a
NADPH-dependent
oxidase located in the
plasma membrane that
results in superoxide
accumulation. Both the
subunits of the oxidase
were shown to be
essential for the
response, and the
oxidase is dependent on
exogenous NAD(P)H for
activity. The peak effect
occurred within 1 hr and
was higher at a dose of 5
than 10.
Generation of ROS led
to decrease (and eventual
loss, with continued
treatment) of
mitochondria! membrane
potential, with
subsequent outer
mitochondria! membrane
permeability changes.


ff in superoxide and
H2O2 accumulation.


DCF fluorescence to
indicate formation of
cellular oxidants; co-
treatment with BHT (a
radical scavenger)
completely blocked this
effect.
Oxidative damage was
induced in thymine to
form cis-thymine glycol.
SOD and CAT did not
alter this reaction. Other
tests suggest that the
reaction requires the
formation of a reactive
arsenic peroxide,
probably dimethylated
arsenic peroxide.




Reference


Smith et
al., 2001















Jing et al.,
1999








Barchowsk
y et al.,
1999b

Lee and
Ho, 1995






Yamanaka
etal.,2003








C-230  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Postconfluent
PAEC cells in
a monolayer









Human-
hamster
hybrid AL
cells























Arsenic
Species

As111 SA











As111 SA


























Concentration(s)
Tested (nM)

1, 2.5, 5, 10, 20











11.5, 15.4


























Duration of
Treatment

30 min











24 hr


























LOECa
(HM)

1











11.5























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
DCF fluorescence as a
direct measure of
intracellular oxidant
concentrations (i.e.,
accumulation of ROS):
likely ft at all doses, with
a peak at 5 that is -45%
higher than control, a
difference that is
statistically significant.
Induction of CD59"
mutations: dose-related
increase in mutation
frequency; pretreatment
+ co-treatment with L-
NMMA (a nitric oxide
synthase inhibitor)
substantially reduced the
mutation frequencies at
both doses. Similar
treatment with D-
NMMA (the inactive
enantiomer) had no
effect. These findings
were taken as evidence
that peroxynitrites have
an important role in
inorganic arsenic-
induced genotoxicity.
That conclusion was
supported by a Western
blot analysis of
nitrotyrosine-modified
proteins induced by
inorganic arsenic
treatments and mostly
blocked by L-NMMA.




Reference

Barchowsk
y et al.,
1996









Liu et al.,
2005






















C-231  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



HepG2 cells



























NB4 cells

NB4-M-AsR2
cells

IM9 cells













Arsenic
Species


As111 ATO




























As111 ATO
for all

















Concentration(s)
Tested (nM)


20




























0.5, 1

2,4

0.5, 1














Duration of
Treatment


6hr




























24hrs
for all

















LOECa
(HM)


20




























0.5

2

0.5











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Analysis of 481 selected
genes in a DNA
microarray experiment:
hierarchical clustering
analysis showed that
inorganic arsenic
exposure closely
resembled DMNQ
exposure (and was
extremely different from
DMN or phenol
exposure) regarding
patterns of genes that
were up-regulated and
down-regulated. In
phase 1 of this
experiment, DMNQ was
selected as a model
chemical that generates
ROS and is known to
induce genes associated
with cell proliferative
responses. Exposure to
inorganic arsenic caused
significant up-regulation
of 38 genes and down-
regulation of 20 genes;
dose used had >80% cell
viability.
HMOX-1 protein (a
stress-responsive
protein) levels after
treatment with ATO
alone and co-treatment
with 100 uM Trolox:
At 0.5: slight ft alone,
big ft with Trolox; at 1:
ft alone, huge ft with
Trolox.
At 2: slight ft alone, big
ft with Trolox; at 4: ft
alone, huge ft with
Trolox.
At 0.5: slight ft alone,
big ft with Trolox; at 1:
ft alone, huge ft with
Trolox.




Reference


Kawata et
al., 2007




























Diazetal.,
2005













C-232  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


NB4 cells







BFTC905
cells and
NTUB1 cells












BFTC905
cells and
NTUB1 cells






















Arsenic
Species

As111 ATO
for all





As111 SA
Asv
MMAm
MMAV
DMA111
DMAV









As111 SA
Asv
MMAm
MMAV
DMA111
DMAV





















Concentration(s)
Tested (nM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Regarding row above, other indications that Trolox potentiates ATO-mediated
oxidative stress: bigger ft in protein carbonyls (indicator of oxidative damage
to proteins) and 8-iso-PGF2a (indicator of lipid peroxidation) by combined
ATO and Trolox treatment(s) than by ATO treatment(s) alone. Other
experiments showed that the synergistic effect of Trolox on ATO-mediated
apoptosis was not related to extracellular H2O2 production. ATO was shown
to induce the formation of Trolox phenoxyl radicals by electronic spin
resonance spectroscopy.


0.2
for all













0.2
for all






















24 hr
for all













24 hr
for all




















0.2
0.2
0.2
None
0.2
None









0.2
0.2
0.2
0.2
0.2
None


















Relative extent of
oxidative damage
(peroxidation) in lipids,
measured as
malonaldehyde
formation; ranking of
those with statistically
significant ft over control
(i.e., unranked arsenicals
hadNSE):
In BFTC905 cells:
Asin>DMAin>MMAm»
Asv
in NTUB1 cells:
DMAIII»MMAIII>Asm.
Relative extent of
oxidative damage
(carbonylation) in
proteins; ranking of
those with statistically
significant ft over control
(i.e., unranked arsenicals
hadNSE):
In BFTC905 cells:
MMAni>Asin>DMAm»
Asv
In NTUB1 cells:
Asin>MMAin>DMAm»
Asv>MMAv.
Consistent with these
effects, increased levels
of nitric oxide,
superoxide ions,
hydrogen peroxide, and
the cellular free iron pool
were consistently
detected in both cell
lines after treatments by
the 3 trivalent arsenicals.




Reference


Diaz et al.,
2005






Wanget
al., 2007














Wanget
al., 2007



















C-233  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


BFTC905
cells and
NTUB1 cells


















Gclm"'" MEF
cells, from
GCLM
knockout
mice



NB4 cells












1RB3AN27
cells








Arsenic
Species
As111 SA
Asv
MMAm
MMAV
DMA111
DMAV

















As111 SA
for all






AsmATO












As111 SA









Concentration(s)
Tested (nM)


0.2
for all






















Duration of
Treatment


24 hr
for all






















LOECa
(HM)
0.2
0.2
0.2
0.2
0.2
0.2

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Relative extent of
oxidative damage (comet
assay) in DNA; ranking
of those with statistically
significant ft over control
(i.e., unranked arsenicals
hadNSE):
Without enzyme
digestion:
In BFTC905 cells: Asm =
MMAm>MMAv> DMAV.
In NTUB1 cells: Asm =
MMAln>DMAm =
MMAm = DMAV.
WithEnm + FPG
digestion:
In BFTC905 cells:
Asni>MMAin>DMAm>
MMAV.
In NTUB1 cells:
Asln>MMAln>DMAm>M
MAV>
DMAV = Asv.
See rows under Apoptosis and Cytotoxicity for this citation for experimental
conditions. The high level of arsenic-induced oxidative stress from some
treatments was not significantly decreased by tBHQ. Yet, tBHQ pretreatment
or co-treatment greatly decreased inorganic arsenic induced apoptosis and
cytotoxicity.



0.75












0.1,0.5, 1,5, 10






Results were obtained from various experiments,
including Affymetrix oligonucleotide microarray
analysis using a chip that contained 22,000 open reading
frames from the human genome. Treatment for 10 days
increased the expression of a set of genes responsible for
ROS production. Genes were identified that responded
to inorganic arsenic and H2O2 but whose response to
inorganic arsenic was reversed by NAC. It was found
that 26% of the genes significantly responsive to
inorganic arsenic might have been directly altered by
ROS. Inorganic arsenic treatment induced ROS, which
in turn oxidized the Spl transcription factor, with a
corresponding decrease in its in situ binding to the
promoters of the 3 genes hTERT, C17, and c-Myc, with
the result that their expressions were significantly
suppressed (e.g., hTERT: U expression to < 1% normal).
2hr






0.5






ROS production using
DCFH-DA assay: ft with
a positive dose-response;
the increase at dose of 1
was blocked by co-
treatment with either
NAC or a-Toc.




Reference




Wanget
al., 2007


















Kannet
al., 2005b





Chou et
al., 2005











Felix etal.,
2005





C-234  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


BEAS-2B
cells















Embryonic
mesenchymal
cells prepared
fromCF-1
mouse
conceptuses at
gestation day
11










RAW264.7
cells





HELP cells









Arsenic
Species

As111 ATO



















As111 SA















As111 SA






As111 SA









Concentration(s)
Tested (nM)


5, 10, 20


















5.8, 11.6, 15.4















2.5,5, 10,25






0.1,0.5, 1,5, 10









Duration of
Treatment


24hr


















2hr















3hr






4hr









LOECa
(HM)


5


















5.8















5






0.5






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Production of 8-
isoprostane, a by-product
of lipid peroxidation: ft
with a positive dose-
response; 2x control at 5,
6x control at 20. In
addition, electron spin
resonance studies
(involving co-treatments
with CAT, SF, NAC, or
NADPH) and confocal
microscope studies
showed that inorganic
arsenic can produce
ROS, such as H2O2 and
'OH, as a result of
reduction reactions
within cells.
Production of ROS
detected by a variant of
the DCF assay using
CM-H2DCFDA:
Induced RFUs (i.e.,
experimental - control):
5.8, -950; 11.6, -2050;
15.4, -2700.
15-min pretreatment with
0.2or0.5%(v/v)DMSO
blocked all or almost all
inorganic arsenic-
induced production of
ROS at dose of 15.4,
whereas 15-min
pretreatment with 0.1%
(v/v) DMSO partially
blocked it.
Extracellular H2O2
production quantified
using the Amplex Red
Hydrogen Peroxide
Assay: there was a
positive dose-response,
reaching
~1.4x control.
Production of ROS
detected by the DCFH-
DA assay in the 15 min
after inorganic arsenic
treatment: ft with dose to
>2x control at dose of
10.




Reference


Hanetal.,
2005
















Perez-
Pasten et
al., 2006













Szymczyk
etal.,2006






Yanget
al., 2007





C-23 5  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


HELP cells


















NB4 cells


























Arsenic
Species


As111 SA


















As111 ATO


























Concentration(s)
Tested (nM)


0.1,0.5, 1,5, 10


















1,3


























Duration of
Treatment


3, 6, 12, 24,
or48hr

















16 hr


























LOECa
(UM)


Various


















1























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
SOD activity after 24 hr:
ft at 0.5, U at 5 and 10;
hint of similar change of
direction in response also
in treatments of some
other durations.
GPx activity after 24 or
48 hr: U at 5 and 10; but
hints of ft at lower doses
and U at higher doses in
treatments of some
durations.
MDA content (measure
of LPO) after 24 or 48
hr: ft at 5 and 10; tended
to increase with time and
dose in treatments of all
durations.
Effect on cellular total
antioxidant capacity
determined using the
ABTS assay
(Troiloc -equivalent
antioxidant capacity in
units of nmol/mg
protein):
Control = -420;
inorganic arsenic at dose
of 3: -150; inorganic
arsenic at dose of 1 :
-240.
Effects of co-treatment
(CoTr):
inorganic arsenic at 3 +
CoTr with 1000 uM
DTT: -275.
inorganic arsenic at 3 +
CoTr with 2000 uM
DTT: -340.
inorganic arsenic at 1 +
CoTr with 25 uMDTT:
-150.
inorganic arsenic at 1 +
CoTr with 50 uM DTT:
-125.




Reference


Yanget
al., 2007

















Jan et al.,
2006






















C-236  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

NB4 cells












BFTC905
cells and
NTUB1 cells
















A549 cells

















Arsenic
Species
As111 ATO













DMAV


















As111 ATO

















Concentration(s)
Tested (nM)

0.5












1,2


















2

















Duration of
Treatment

2hr












24 hr
for all

















48 hr

















LOECa
(HM)

0.5










1 in at
least one
cell line
for all 3
effects














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Intracellular H2O2 level
(units of Amplex red
assay): control = -20;
inorganic arsenic -45.
Effects of co-treatment
(CoTr):
CoTr with 80 uMDTT:
-72.
CoTr with 100 uM
DMSA: -67.
CoTr with 20 uM
DMPS: -72.
ft oxidative damage
(peroxidation) in lipids,
measured as
malonaldehyde
formation: at both doses
in BFTC905 cells, at
doseof2inNTUBl
cells.
ft oxidative damage
(carbonylation) in
proteins: at higher dose
inBFTC905cells,at
lower dose in NTUB1
cells.
ft oxidative damage
(comet assay) in DNA,
without enzyme
digestion: at both doses
in both cell lines.
Loss of MMP determined by flow
cytometry using JC-1: 2 uM inorganic
arsenic: ft to ~1.25x;
200 uM sulindac: ft to ~1.15x; (2 uM
inorganic arsenic + 200 uM sulindac):
ftto~1.9x.
There was also a synergistic
interaction between these treatments in
causing big ft in cytochrome C protein
level in the cytosol, which is thought
to result from damage to
mitochondria! membranes that permits
cytochrome C release to the cytosol.
Pretreatment with NAC almost
entirely blocked the MMP and
cytochrome C effects. (Sulindac is a
NSAID that inhibits COX-2.)




Reference

Jan et al.,
2006











Wang et
al., 2007

















Jinetal.,
2006b













C-237  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

A549 cells













HeLa cells












BAEC cells


BAEC cells


HEK 293
cells and
SV-HUC-1
cells











Arsenic
Species

As111 ATO













As111 SA












As111 SA


As111 SA


As111 SA
MMAm
DMA111












Concentration(s)
Tested (nM)

2













10, 100












5, 10


10


0.2
for all













Duration of
Treatment

48hr













4hr












Ihr


24 hr


24 hr
for all













LOECa
(HM)

2











10 for
Trxl and
Trx2;
none for
GSH/
GSSG









5


10


0.2
0.2
0.2









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Production of ROS using
carboxy-H2DCFDA
assay: control, -0.8 unit;
2 uM inorganic arsenic,
-4.2 units; 200 uM
sulindac: ~4.5x; (2 uM
inorganic arsenic + 200
uM sulindac): ~7.5x.
Thus there was only
additivity. Pretreatment
with NAC before
combined treatment: U to
-3.9 units.
Effects on Trxl and Trx2
redox states determined
using Redox Western
blot methods:
Trxl : ft in oxidation at
10, slightly bigger ft at
100.
Trx2: huge ft in
oxidation at 10, slightly
bigger ft at 100.
In contrast, inorganic
arsenic had little effect
on the GSH/GSSG redox
state, as determined by
HPLC.
ft in peroxynitrite to
~1.4xand~1.6xat 5 and
10, respectively.
ft in nitrotyrosine
formation to ~1 . 15x.
Relative extent of
oxidative damage
(peroxidation) in lipids,
measured as
malonaldehyde
formation; ranking of
those with statistically
significant ft over control
(all were significant):
In HEK 293 cells:
Asin»MMAin>DMAm.
In SV-HUC-1 cells:
Asin>DMAm>MMAni.




Reference

Jinetal.,
2006b












Hansen et
al., 2006











Bunderson
etal.,2006

Bunderson
etal.,2006

Wanget
al., 2007










C-23 8  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



HEK 293
cells and
SV-HUC-1
cells











TRL1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSOfor24hr
to deplete
GSH levels
and then co-
treated with
50uMBSO
TRL 1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSOfor24hr
to deplete
GSH levels
and then co-
treated with
SOuMBSO




Arsenic
Species


As111 SA
MMAm
DMA111















MMAV
for both











MMAV
for both













Concentration(s)
Tested (nM)


0.2
for all
















5mM












5 mM














Duration of
Treatment


24 hr
for all
















24 hr












48 hr














LOECa
(HM)


0.2
0.2
0.2













None


5mM









None


5 mM










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Relative extent of
oxidative damage (comet
assay) in DNA; ranking
of those with statistically
significant ft over control
(all were significant):
Without enzyme
digestion:
In HEK 293 cells:
Asni>MMAm = DMA111.
In SV-HUC-1 cells:
MMAm = As111 = DMA111.
WithEnm + FPG
digestion:
In HEK 293 cells: As111 =
MMAm = DMA111.
In SV-HUC-1:
Asni>DMAm = MMAm.
Cellular ROS levels
based on DCFH-DA
assay:
MMAV: NSE.
MMAV + BSD: ft to
~2.22x .







Cell survival determined
by AB assay:
MMAV: 100% survival.
MMAV + BSO:~3%
survival.
Co-treatment with 10
mM DMPO during the
MMAV + BSO treatment
blocked most of the
cytotoxicity, resulting in
-72% survival. DMPO
effectively scavenged
cellular radical
molecules.




Reference



Wanget
al., 2007















Sakurai et
al., 2005a












Sakurai et
al., 2005a









C-239  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
TRL1215
cells

TRL 1215
cells
pretreated
with 50 uM
BSO for 24 hr
to deplete
GSH levels
and then co-
treated with
50uMBSO

HeLa cells















HeLa cells













Jurkat cells
Namalwa
cells

NB4 cells

U937 cells




Arsenic
Species


MMAV
for both










As111 ATO















As111 ATO














As111 ATO
for all








Concentration(s)
Tested (nM)


5mM











2















2















2








Duration of
Treatment


48 hr











Various up
to24hr














Ihr














24 hr
for all








LOECa
(HM)
None


None
with
DMPO








2















2













None
2

2

None


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Caspase 3 activity
(related to apoptosis):
MMAV: NSE.
MMAV + BSO: ft to
~1.66x.
Co-treatment with 10
mM DMPO during the
MMAV + BSO treatment
completely blocked the
ft of caspase 3 activity.
DMPO effectively
scavenged cellular
radical molecules.
ROS levels were shown
by DCFH-DA assay to
be significantly elevated
by inorganic arsenic and
to ft roughly 3x higher
than for inorganic
arsenic alone following a
combined inorganic
arsenic plus 10 uM
emodin treatment; the
addition of 1.5 mMNAC
as a co-treatment
attenuated (but did not
completely block) that ft
in ROS levels.
Analysis of GSH/GSSG
ratios showed that co-
treatment of inorganic
arsenic with emodin had
a major oxidative impact
on the cellular redox
state, as shown by
following ratios: control,
-62; inorganic arsenic,
-52; 10 uM emodin,
-34; inorganic arsenic
plus 10 uM emodin, -13;
pretreatment with 1.5
mM NAC attenuated
(but did not completely
block) this effect.

ft in H2O2 levels as
detected by FACS after
staining with DCFH-DA:
large effect seen in
Namalwa and NB4 cells
only; NSE in other cell
lines.




Reference



Sakurai et
al., 2005a









Yi et al.,
2004














Yietal.,
2004













Chenetal.,
2006





C-240  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
U937 cells




HEK293 cells











Arsenic
Species
As111 ATO




As111 ATO











Concentration(s)
Tested (nM)
1




2











Duration of
Treatment
24 hr




48 hr











LOECa
(HM)
None




2










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft in H2O2 levels as
detected by FACS after
staining with DCFH-DA:
large effect was seen
only following a co-
treatment with B SO for
24 hr; the ft was
substantially decreased
by a 4-hr treatment with
either 10 mM NAC or
200 units of catalase.
Cell survival was
determined by the WST-
1 cell proliferation assay:
inorganic arsenic
treatment resulted in
-22% cell survival; co-
treatment with 1 mM
Tironor400U/mLCAT
significantly ft cell
survival although more
than 60% of the cells still
died; co-treatment with
200 U/mL SOD
markedly U cell survival.
These and other data
suggested that inorganic
arsenic induced both
superoxide anion and
H2O2 through the
activation of NAD(P)H
oxidase. Presence of
superoxide anion in cells
that resulted from
inorganic arsenic
treatment was confirmed.

Reference
Chenetal.,
2006




Sasaki et
al., 2007









C-241  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

PRCCs

HEK293
cells
















NB4 cells

HL60 cells

KMS12BM
cells

U266 cells























Arsenic
Species

As111 ATO
for both




















As111 ATO
for all



























Concentration(s)
Tested (nM)

0.01,0.05,0.1,
0.5,1,5, 10,20
for both


















-0.01,0.05,0.1,
0.5, 1, 2, 5, 10, 50
for first three


-0.05,0.1,0.6,
1.2,6























Duration of
Treatment


48 hr
for both



















48 hr
for all



























LOECa
(HM)

0.5

0.1

















-0.1

~5

-0.2

-0.6





















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival was
determined by the WST-
1 cell proliferation assay:
LC50sinPRCC:
inorganic arsenic, -10;
co-treatment of inorganic
arsenic with 10 uM a-
lipoic acid, -25.
LC50sinHEK293:
inorganic arsenic, -1; co-
treatment of inorganic
arsenic with 10 uM a-
lipoic acid, -7. In both
cell types, this
antioxidant markedly
attenuated inorganic
arsenic's cytotoxicity,
andinHEK293cellsit
was shown to suppress
superoxide anion
generation.
Cell survival was
determined by the WST-
1 cell proliferation assay:
LC50s: NB4, -0.2; HL60,
~8;KMS12BM, -0.3;
U266, -0.3. In all 4 cell
lines, co-treatment of
inorganic arsenic with 10
uM a-lipoic acid resulted
in a remarkably similar
dose-related pattern of
cell survival to that seen
with inorganic arsenic
alone, this being in sharp
contrast to the
attenuation of
cytotoxicity caused by it
that was seen in PRCCs
and HEK293 cells. Note
that the LOEC is higher
than the estimated LC50
of 0.3 for U266 cells
because the next lower
dose of 0. 1 had no effect,
and the LC50 was
estimated from the dose-
response curve that was
presented.




Reference


Sasaki et
al., 2007




















Sasaki et
al., 2007























C-242  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
JAR cells




JAR cells






BEAS-2B
cells




Undifferentiat
edPC12 cells












Arsenic
Species
As111 ATO




As111 ATO






As111 SA





As111 ATO













Concentration(s)
Tested (nM)
5




5




1, 2.5, 5 for
mRNA


2.5, 5 for protein



8













Duration of
Treatment
2, 4, 6, 16,
24hr



6hr




8hr
for both






Various up to
6hr












LOECa
(HM)
5




5




Ifor
mRNA

2.5 for
protein




8









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft in HMOX-1 protein
level in cytoplasm by 2
hr, with time-related
response becoming huge
by 16 hr.
Intracellular H2O2 level
detected by DCFH-DA
and flow cytometry
assay:
ft to 2x.
BigftinHMOX-1
mRNA level at 1, bigger
ft of the same at 2. 5,
huge ft of the same at 5.

BigftinHMOX-1
protein level at 2.5, huge
ft of the same at 5.
Detection of ROS shown
by increase of DCF-
fluorescence in DCFH-
DA assay:
ft to ~2x control for
several time points
during first hr; no hint of
effect at 3-6 hr;
fluorescence was
observed before the
onset of cell death.




Reference
Massrieh
etal.,2006



Massrieh
et al., 2006




O'Hara et
al., 2006






Pigaetal.,
2007








C-243  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue





UROtsa cells


























Human-
hamster
hybrid AL
cells














Arsenic
Species

As111 SA



MMAm


























As111 SA
For both
















Concentration(s)
Tested (nM)

1, 10, 100



0.05,0.5,5


























7.7

















Duration of
Treatment

10 min



50 min


























60 days

















LOECa
(HM)

1



0.05


























7.7














Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Detection of ROS using
CM-H2DCFDA assay:
Slight ft at 1, big ft at 10,
huge ft at 100. When
quantified at dose of 10
over 10 min: 20 RFU by
4.5 min, 110 RFU by 10
min. Pretreatment with
PEG-SOD or PEG-CAT
blocked most ROS
production.
ft at 0.05, huge ft at 0.5,
slightly weaker response
at dose of 5 than at dose
of 0.05. When
quantified at dose of 0.5
over 10 min: 0 RFU.
When quantified at dose
of 0.5 over 50 min: 10
RFU by 42 min, 65 RFU
by 50 min. Pretreatment
with PEG-CAT blocked
most ROS production,
and co-treatment with
PEG-SOD blocked some
ROS production; less
effect for both than for
inorganic arsenic111,
suggesting a difference
in the ROS they produce.
Effects related to
mitochondria:
fluorescence microscopy
showed that arsenic
treatment led to
considerable variation in
the distribution of
mitochondria between
cells and caused the
fraction of them with
elongated morphology to
increase from 6% to
66%; -50% U in COX
activity; -40% U in
oxygen consumption;
-40% ft in citrate
synthase activity.




Reference





Eblin et
al., 2006

























Partridge
et al., 2007













C-244  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Human-
hamster
hybrid AL
cells








Splenic
lymphocytes
from
SodltmlLeb
knockout
mice











Lyophilized
bovine tubulin





















Arsenic
Species

As111 SA
for both











As111 SA















DMA111






















Concentration(s)
Tested (nM)
1.9,3.8,7.7


1.9,3.8,7.7










50, 100, 200















50






















Duration of
Treatment
60 days


Iday










2hr














Time course
over 1 hr






















LOECa
(HM)
3.8 for
copy #

1.9 for
deletions









50















50



















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
mtDNA copy number: U
to~0.84xat3.8;Uto
~0.65x at 7.7.


SA induced large
heteroplasmic deletions
in mitochondrial DNA,
and the frequencies of
induction increased with
dose and time of
exposure.
Breaks and/or alkali-
labile lesions in DNA
detected in the single-
cell gel (comet) assay:
big ft in effect in the
SOD -/- mice, which
were also shown to have
big U in levels of SOD in
spleens (and also in
livers and kidneys).
SOD +/- mice were
intermediate in SOD
levels and DNA damage.
Results suggest ROS
may be involved in Asm-
induced DNA damage.
Big U in GTP-induced
polymerization of
lyophilized bovine
tubulin.
Effects of modulators:
NAC blocked the
inhibition by DMA111,
while AA, CAT, DMSO,
Tiron, or Trolox® had
NSE on it, which
suggests that ROS is not
involved in the
inhibition. Premixingof
. V
inorganic arsenic ,
MMAV, or DMAV for 2 hr
with a 5 -fold molar
excess of GSH greatly
decreased the
polymerization of tubulin
(i.e., increased the
inhibition).




Reference

Partridge
etal.,
2007









Kligerman
and
Tennant,
2007













Kligerman
and
Tennant,
2007
















C-245  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue
W138 cells
and HaCaT
cells










Arsenic
Species
As111 SA











Concentration(s)
Tested (|oM)
0.5











Duration of
Treatment
24 hr











LOECa
(HM)
0.5











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ROS (peroxide) levels
based on DCF assay :U in
both cell lines compared
to control, and less in
W138 than in HaCaT.
The average activities of
3 important intracellular
redox agents, GSH, GR,
and GST are ~3X higher
in WI38 cells than in
HaCaT cells. After the
inorganic arsenic
treatment, there was a
60-min menadione
treatment followed by a
60-min recovery period.
During this 120 min,
ROS levels in W138
cells never reached
control levels, while the
control level was
substantially exceeded in
HaCaT cells after 60 min
of the menadione
treatment and later.
Reference
Snow et
al., 2005










C-246  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

NB4 cells



HL-60 cells





























PAEC cells
isolated from
freshly
harvested
vessels









CHO Kl cells





Arsenic
Species



As111 SA
for all































As111 SA












As111 SA





Concentration(s)
Tested (|oM)

2 for all assays,
which tested
effects of various
co-treatments
described in
Results column





























5












20, 40, 80, 160





Duration of
Treatment



4hr
for all































Up to 20 min












4hr





LOECa
(HM)

—



—






























5












40


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
This row relates only to
the effects seen after co-
treatments in an attempt
to learn how SA causes
DNA damage. They
assayed DNA strand
breaks (AD SB) detected
using the comet assay.
In the absence of a co-
treatment, a significant
increase would be
expected with a dose of
only 0.25. Conclusions
always were supported
by data on ODA and
DPC individually.
Chemicals used
individually in co-
treatments were:
catalase, calcium
chelators, and inhibitors
of nitric oxide synthase,
SOD, and
myeloperoxidase. On
the basis of the large
reduction in DNA strand
breaks seen following
the co-treatments, they
concluded that arsenite
induces DNA adducts
through calcium-
mediated production of
peroxy nitrite,
hypochlorous acid, and
hydroxyl radicals.
ft oxygen consumption
associated with ft
superoxide (O2~)
formation;
ft extracellular
accumulation of H2O2,
with same time and dose
dependence as
superoxide formation.
Pretreatment of the cells
with DPI, apocynin, or
SOD abolished arsenite-
stimulated superoxide
(O2~) formation.
ft intracellular peroxide
level (strong hint of same
effect at dose of 20)




Reference



Wang et
al., 2001































Barchowsk
y et al.,
1999b










Wang et
al., 1996

C-247  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
4>X174 RF I
DNA
Naked
double-
stranded
circular DNA
in presence of
ROS
inhibitors
















Both
HL-60 cells
and
HaCaT cells










HL-60 cells













Arsenic
Species

MMAm


DMA111





















As111 SA












As111 SA













Concentration(s)
Tested (nM)

10, 20, 30, 40, 50


37.5, 75, 150, 300,
1000




















0.1,0.5, 1, 10,20,
40











10













Duration of
Treatment


24 hr
for all






















5 days












3 days













LOECa
(HM)

—


—





















0.5 but
possibly
0.1










—










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
This row relates only to
the effects seen after co-
treatments in an attempt
to learn how SA causes
DNA damage.
Significant (and usually
complete) reduction in
nicked DNA (in DNA
nicking assay) was found
when ROS inhibitors
Trolox, melatonin, or
Tiron were present
individually during the
arsenic treatment. Spin
trap agent DMPO was
also effective in
preventing DNA nicking
by these compounds.
Thus, production of ROS
by these chemicals is
associated with their
DNA-cutting activity.
Genotoxicity is an
indirect effect via the
generation of ROS.
By use of MTT assay, in
presence of 2.5 mM
DMPO: ft in cell
number, with peak at 0.5
(DMPO has no effect); U
in cell number to below
control level at 1 for HL-
60 and at 10 for HaCaT,
and DMPO significantly
lessens reduction in cell
number at >10 (possibly
1) for HL-60 and at >20
(possibly 10) for HaCaT.
Analysis of TRF using
Southern blot assay in
presence of 2.5 mM
DMPO:
With DMPO present,
telomere length was
longer than it was with
arsenic alone; interpreted
to mean that DMPO
provided some protection
against arsenic-induced
telomere shortening.




Reference


Nesnow et
al., 2002






















Zhang et
al., 2003











Zhang et
al., 2003









C-248  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
HL-60 cells



HaCaT cells








Arsenic
Species

As111 SA
for both










Concentration(s)
Tested (|oM)

0.1,0.5, 1, 10,20,
40 for both










Duration of
Treatment

5 days
for both










LOECa
(HM)
Ibut
possibly
0.5

10





Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ByuseofHoechst/PI
staining assay, in
presence of 2.5 mM
DMPO:
ft in apoptosis for both;
however, DMPO
significantly reduced the
amount of apoptosis at
>1 for HL-60 and at >10
forHaCaT.




Reference


Zhang et
al., 2003






Enzyme Activity Inhibition
AG06 cells
were
pretreated for
24 hr with
unspecified
low dose of
As, and then
extracts of the
cells were
tested for
activity of:
GSH
peroxidase
and
ligase










Cell-free
system using
purified
human
enzymes













As111 SA




















As111 SA

Asv















IC50
determinations




















IC50
determinations














Rate over 6
min




















Rate of
reaction over 6
min















—





















—













IC50s:
2.0 (was 0. 13 mM for
purified enzyme with no
arsenic pretreatment)
14.5 (was 6.5 mM with
no arsenic pretreatment).
The same paper
presented the IC50s for a
similar treatment with
Asv for GSH peroxidase,
and it was 173 uM. The
paper also presented
IC50s for numerous
purified enzymes with
both SA and Asv, but
they were almost all far
above a physiologically
interesting range and are
thus not presented here.
Most were in the range
of 6.3 to 381 mM for SA
and usually even higher
for Asv.
Inhibition of PDH: IC50s:
5.6 (oM for inorganic
arsenic111, 206 mM for
Asv;
7 other enzymes
involved in aspects of
DNA repair and/or
cellular stress response
hadIC50sforAsniof6.3-
381mM. Only PDH,
with its lipoic acid
cofactor, was inhibited
by physiologically
relevant, micromolar
concentrations of As111.





Snow et
al., 1999



















Hu et al.,
1998











C-249  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Cell-free
system using
purified
human
enzymes

Cell-free
system using
purified
porcine heart
PDH

Cell-free
system using
hamster
kidney PDH




Arsenic
Species
As111 SA



Asv



As111 SA


MMA111


As111 SA

MMA111




Concentration(s)
Tested (nM)
-0.0007, 0.001,
0.007, 0.01, 0.07,
0.1

-0.01,0.07,0.1, 1,
10, 25, 75, 100,
125

25, 75, 100, 200
(all approximate)

8, 16, 30, 50, 100


-20 to -400

-20 to -400




Duration of
Treatment

Rate of
reaction over 1
min




30 min
for both



30 min
for both






LOECa
(HM)
-0.001



-25



-25


o
— o



—


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)


Inhibition of PDH.




Inhibition of PDH
(IC50s):
106.1.


17.6.
Inhibition of PDH
(IC50s):
115.7.

61.0.




Reference


Hu et al.,
1998




Petrick et
al., 2001



Petrick et
al., 2001



Gene Amplification


Mouse 3T6
cells



AG06 cells





AG06 cells











Human
osteosarcoma
TE85 (HOS)
cells





As111 SA


Asv



As111 SA





As111 SA











As111 SA








0.2, 0.4, 0.8, 1.6,
3.2,6.4

1, 2, 4, 8, 16



7, 10, 17, 20





6











0.0125, 0.025,
0.05, 0.1 for both
durations








Not reported




3.5 hr




Assay's
maximal
response time










6 wk

8wk






0.4


2



None





6











0.025

0.0125






Gene amplification of
dhfr gene detected by
MTX-selection assay:
Both compounds showed
positive dose-response
extending to highest
concentrations tested.
Amplification of SV40:
none observed at
concentrations causing
from 40% to 98%
cytotoxicity.
Amplification of
endogenous dhfr genes
(determined by MTX-
selection assay): highly
effective at this
concentration, which
caused 50% survival.
"Amplification factor"
was -3 even though it
was 1 (i.e., no induction)
for same concentration
for amplification of
SV40.
Amplification of
endogenous dhfr genes
(determined by MTX-
selection assay): dose-
response was the same
for both durations
beginning with 0.025; it
increased with dose to
0.05 and then plateaued.


Barrett et
al., 1989



Rossman
and
Wolosin,
1992


Rossman
and
Wolosin,
1992









Mure et
al., 2003






C-250  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

SHE cells















Arsenic
Species
As111 SA

Asv














Concentration(s)
Tested (nM)
6,8

50, 100, 150














Duration of
Treatment

48 hr
for both














LOECa
(HM)
6

50











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
From among these
treatment groups, 5
neoplastic transformed
cell lines were produced
that were shown to be
tumorigenic. Of these: 3
had c-Ha-ras (oncogene)
gene amplification; 2 had
c-myc (oncogene) gene
amplification;
a few other arsenic-
treated cell lines also
showed this same gene
amplification.




Reference

Takahashi
etal.,2002











Gene Mutations
E. coli
(several
strains)







V79 cells









G12 cells








As111 SA









As111 SA









As111 SA








Up to 25 mM








0.5

5, 20, 100







5, 10, 15

10, 25, 50







Various








2 days

Uptol.Shr







24 hr

3hr







None








None

None







None

None







Several assays (spot
tests, treat and plate
protocols, and
fluctuation tests) for Trp+
revertants yielded no
evidence of induction of
gene mutations. Also,
there was no induction of
I prophage.
In several assays,
ouabain resistance and
thioguanine resistance
were used as genetic
markers. No evidence
was found of induction
of gene mutations. Only
the dose of 100 caused
cy totoxicity (3 3 . 1 % the
survival of the control).
No statistically
significant induction of
mutations at the gpt
locus in an assay that can
detect multilocus
deletions, point
mutations, and small
deletions (tested up to
cy totoxicity of 61.9% of
cells killed).
Rossman
etal., 1980







Rossman
etal., 1980








Li and
Rossman,
1989a







C-251  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Salmonella
typhimurium
strains TA98,
TA100,
TA104




Syrian
hamster
embryo cells

Human
osteosarcoma
TE85 (HOS)
cells




TM3 cells












Arsenic
Species
As111 SA
Asv
MMAm
MMAV
DMA111
DMAV




As111 SA

Asv

As111 SA

MMA111





As111 SA
for both











Concentration(s)
Tested (|oM)
Tested up to
concentrations
limited by
cytotoxicity or to
the limit
concentration for
the assay



-0.8, 1.6, 3, 3.5, 5

~8, 16, 32, 64, 128

0.0125, 0.025,
0.05,0.1

0.00625, 0.0125,
0.025, 0.05




0.008, 0.77, 7.7
for both











Duration of
Treatment

3 days
for all








Not reported



8 wk for all





-25 days

-75 days











LOECa
(HM)

None








None

None

0.0125

None




0.008

0.008








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Salmonella mutagenicity
plate incorporation assay
with and without
exogenous metabolic
activation:
There was no indication
of any induction of gene
mutations over
background levels by
any of the compounds.

Gene mutation assays for
the Na+/K+ ATPase and
HPRT loci.
Mutations in the HPRT
gene: positive dose-
response to highest
concentration for As111;
no increase until almost
15 generations of
continuous exposure.
Detection of DNA
changes by RAPD-PCR:
gain or loss of loci and
changes in the intensity
of loci were detected at
the DNA sequence level;
although the nature of
the "mutations" and
whether they were actual
gene mutations is
unknown.




Reference


Kligerman
et al., 2003







Barrett et
al., 1989

Mure et
al., 2003





Singh and
DuMond,
2007








C-252  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



Arsenic
Species



Concentration(s)
Tested (|oM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)




Reference
Hypermethylation of DNA
Human
kidney
carcinoma
cell lines:
UOK123
UOK109

Human lung
carcinoma
cell line:
A549


















A549 cells
(human
adenocarcino
ma)






As111 SA
for all






















As111 SA

Asv

DMAV






0.010, 0.020,
0.050
0.007, 0.021,
0.093



0.08, 0.4, 2.0

















0.08, 0.4, 2.0

3, 10, 30, 100, 300

2, 20, 200, 2000






4wk
4wk



2wk





















7 days for all








<0.050
<0.093



<2.0



















0.08

30

None




The number of specific
DNA sequences shown
to undergo
hypermethylation
changes by methylation
sensitive AP-PCR
following exposure to
SA:
lfromlineUOK123,4
from line UOK 109, and
1 from line A549.
The concentrations used
to treat these lines were
known to be the IC30,
IC50, and IC80
concentrations for UOK
cells and the IC2o, IC50,
and IC8o concentrations
for A549 cells. It was
not reported which
concentrations yielded
the hypermethylation
changes, but the LOECs
could not be higher than
the maximum
concentration used for
each cell line.
Hypermethylation within
a 341 -base-pair fragment
of the promoter of p53 .

For the two inorganic
forms, there was a
positive dose-response
throughout the range of
concentrations tested.



Zhong and
Mass,
2001






















Mass and
Wang,
1997





Hypomethylation of DNA
TRL 1215
cells (normal
rat liver)



As111 SA




0.125,0.250,
0.500



19 wk




0.125



Global DNA
hypomethylation,
thought to be caused by
continuous methyl
depletion.

Zhao etal.,
1997


C-253  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
Human
kidney
carcinoma
cell line:
UOK121
Human lung
carcinoma
cell line:
A549










Untransforme
dand
immortalized
RWPE-1 cells
(human
prostate
epithelial cell
line)



Arsenic
Species
As111 SA
for all













As111 SA




Concentration(s)
Tested (|oM)
0.009, 0.020,
0.074
0.08, 0.4, 2.0












5




Duration of
Treatment
4wk
2wk












30 wk




LOECa
(HM)
<0.074
<2.0












5



Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
The number of specific
DNA sequences shown
to undergo
hypomethylation
changes by methylation
sensitive AP-PCR
following exposure to
SA:
1 from line UOK121 and
1 from line A549.
The concentrations used
to treat these lines were
known to be the IC30,
IC50, and IC80
concentrations for
UOK121 cells and the
IC20, ICso, and IC80
concentrations for A549
cells. It was not reported
which concentrations
yielded the
hypermethylation
changes, but the LOECs
could not be higher than
the maximum
concentration used for
each cell line.
Global hypomethylation
of DNA (up to 131%
increase in unmethylated
DNA compared to the
control);
hypomethylation still
present 6 weeks after end
of exposure. The cells
became tumorigenic after
29 weeks of treatment
and were then called the
CAsE-PE cell line.

Reference
Zhong and
Mass,
2001













Benbrahim
-Tallaa et
al., 2005


C-254  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


SHE cells


















TM3 cells















HaCaT cells












Arsenic
Species
As111 SA


Asv
















As111 SA
for both















As111 SA












Concentration(s)
Tested (|oM)
6,8


50, 100, 150
















0.008, 0.77, 7.7
for both















0.2












Duration of
Treatment


48 hr
for both















-25 days

-75 days













For 10 serial
passages in
folic-acid-
depleted media











LOECa
(HM)


—
















0.008

0.008















0.2









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
From among these
treatment groups, 5
neoplastic transformed
cell lines were produced
that were shown to be
tumorigenic. Testing of
them using the
methylation-sensitive
restriction endonuclease
isoschizomers Hpall and
Mspl revealed
hypomethylation of c-
myc and c-Ha-ras in the
5'-CCGG sequence.
Both of these oncogenes
were often shown to
exhibit gene
amplification and
ft mRNA expression.
Detection of methylation
changes in DNA by
RAPD-PCR using
methylation-sensitive
restriction endonuclease
isoschizomers Hpall and
Mspl: methylation
changes were detected at
18 loci, with some
showing
hypomethylation and
others hypermethylation.
Some loci were only
affected by the shorter-
term exposure, and vice-
versa.
Genomic
hypomethylation as
demonstrated by a 27%
U in the level of
5-methyl-dCMP
compared with cells
cultured for the same
number of passages in
medium without As111.
This dose was too low to
have much, if any, effect
on the proliferation rate.




Reference


Takahashi
etal.,2002
















Singh and
DuMond,
2007














Reichard
etal.,2007








C-255  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

HaCaT cells














Arsenic
Species

As111 SA














Concentration(s)
Tested (|oM)

0.5, 1.5,5














Duration of
Treatment

72 hr














LOECa
(HM)

Various











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
J]DNMTlmRNAat0.5,
and progressively larger
decreases at 2 higher
doses;
U DNMT3A mRNA at
1.5, and larger U at dose
of 5. These cells did not
show any detectable
quantities of the other 2
mammalian DNA
methyltransferases.
Big ft HMOX-1 RNA at
1.5 with very big ft at 5.




Reference

Reichard
et al., 2007










Immune System Response
(Human
myeloma-like
cell lines)
RPMI 8226
Karpas 707
U266






HPBMs co-
exposed to
M-CSF








HPBMs co-
exposed to
GM-CSF








As111 ATO









As111 SA
Asv
MMAV
DMAV







As111 SA
Asv
MMAV
DMAV









0.5, 1,2









IC50
determinations









IC50
determinations










72 hr









7 days










7 days











0.5









—










—









Induction of cell lysis by
LAK effector cells was
apparent by 36 hours and
maximal at 72 hours.
The extent of lysis was
determined by the 51Cr
release assay. At these
concentrations, arsenic
trioxide had no effect on
viability (using trypan-
blue assay) or apoptosis.
Viability of M-type
macrophages based on
AB assay was used to
estimate the arsenic
concentration at which
maturation into M-type
macrophages was
inhibited by 50%: IC50
values: As111, 0.06; Asv,
200;
MMAV, 750; DMAV, 300.
Viability of GM-type
macrophages based on
AB assay was used to
estimate the arsenic
concentration at which
maturation into GM-type
macrophages was
inhibited by 50%: IC50
values: As111, 0.38; Asv,
300;
MMAV, 700; DMAV, 550.

Deaglio et
al., 2001









Sakurai et
al., 2006









Sakurai et
al., 2006








C-256  DRAFT—DO NOT CITE OR QUOTE

-------
  Type of
 Cell/Tissue
Arsenic
Species
 Concentration(s)
   Tested QaM)
Duration of
 Treatment
LOECa
 (HM)
  Results (Compared
 With Controls, With
  All Concentrations
        Being
 in nM Unless Noted)
Reference
 HPBMs co-
 exposed to
GM-CSF and
    IL-4
Asm SA
    determination
   7 days
            Viability of immature
           dendritic cells based on
            AB assay was used to
             estimate the arsenic
           concentration at which
          maturation into immature
             dendritic cells was
           inhibited by 50%: IC50
                value: 0.70.
                         Sakurai et
                         al., 2006
 HPBMs co-
 exposed to
GM-CSF and
    IL-4
Asm SA
 IC50 determination
  14 days
                Viability of
             multinucleated giant
           cells based on AB assay
           was used to estimate the
           arsenic concentration at
            which maturation into
             multinucleated giant
            cells was inhibited by
            50%: IC50 value: 0.33.
                         Sakurai et
                         al., 2006
 HPBMs co-
 exposed to
  GM-CSF
As111 SA
 With regard to 4 rows immediately above this one, SA at doses of 0.05 to 0.5
    induced abnormal morphological changes in the HPBMs to form small
    nonadhesive and CD14-positive cells called arsenite-induced cells that
 displayed a dendritic morphology with delicate membrane projections. This
     response was not produced by treatments with many other metallic
 compounds (e.g., chromium, mercury, and zinc) including inorganic arsenicv,
	MMAV and DMAV. This effect was not seen at doses exceeding  1.	
                                                   Sakurai et
                                                    al., 2006
 HPBMs co-
 exposed to
  GM-CSF
As111 SA
       0.5
   7 days
  0.5
  In comparison to the
  cells not treated with
 inorganic arsenic, there
    was 43.3% less
metabolic activity, 0.6%
as much adherent ability,
  a 76% higher cellular
  GSH concentration,
  256% as much NO2
  production, 185% as
 much IL-la production
in the supernatant, 412%
    as much IL-la
production in the lysate,
  and 576 ng/g cellular
 protein of IL-12 in the
lysate even though none
was detected in arsenic-
    untreated cells.
 Sakurai et
 al., 2006
  HUVECs
As111 SA
       0.5
    3hr
  0.5
  Both HUVECs and
 PMNs were pretreated
  for24hrwithGLN
  (glutamine) at 0, 300,
600, or 1000 uM. Those
  HUVECs were then
  exposed to the same
 concentration of GLN
  with or without the
Hou et al.,
   2005
                                                  C-257  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue





















































Arsenic
Species





















































Concentration(s)
Tested (|oM)





















































Duration of
Treatment





















































LOECa
(HM)





















































Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
inorganic arsenic
treatment for 3 hr. The
pretreated PMNs were
added to wells and
allowed to migrate
across the pretreated
HUVECs for 2 hr, after
which surface
expressions on HUVECs
of 1C AM- 1 and VCAM-
1 were measured, with
the following results:
ICAM-1: ft in inorganic
arsenic only group and
huge ft at all 3 dose
levels of GLN; VCAM-
1 : NSE in inorganic
arsenic only group and ft
at all 3 dose levels of
GLN, with largest ft at
300 uM. Clearly
HUVECs were activated
by inorganic arsenic.
Also at this time, PMN
expressions of CD 1 Ib
and IL-8 receptor were
measured, with the
folio wing results: CD 11-
b: ft in inorganic arsenic
only group and bigger ft
at all 3 dose levels of
GLN; IL-8 receptor: ft in
inorganic arsenic only
group and at all 3 dose
levels of GLN. Clearly
PMNs were activated by
the inorganic arsenic
treatment of the
HUVECs.
Effects on PMN
migration:
In absence of GLN
pretreatment, inorganic
arsenic caused slight U
from 36% to 30%
migrated. In the
inorganic arsenic + 300
uM GLN group: ft from
-40% (for GLN alone) to
-50% migrated (for
inorganic arsenic +
GLN), which was the
most migration observed.
Reference





















































C-258  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




PBMCs co-
treated
with GM-CSF































PBMCs co-
treated
withM-CSF







Arsenic
Species




As111 ATO

































As111 ATO









Concentration(s)
Tested (|oM)




0.125,0.25,0.5, 1,
2
































1









Duration of
Treatment




6 days

































6 days









LOECa
(HM)




0.125

































1






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis (i.e.,
experimental - control)
detected using A5/SG
assays: approximate
induced frequencies at
the 5 doses: 6%, 20%,
29%, 48%, and 62%,
respectively, with all
being statistically
significant except first
one. Induced frequency
of necrotic cells was
-20% at the highest
dose, and there were
smaller numbers of
necrotic cells induced at
the lower doses.
After dose of 1 for 3
days: ft caspase-3
activity, ft caspase-8
activity, big ft in active
caspase-3 subunitpl?.
ATO was shown to
reduce DNA binding of
the transcriptionally
active p65 NF-KB
subunit to the KB
consensus sites in GM-
CSF treated PBMCs,
which was thought to be
important in
development of
apoptosis. Other
experiments showed that
ATO inhibited
macrophagic
differentiation of
PBMCs.
Induced apoptosis (i.e.,
experimental - control)
detected using A5/SG
assays:
-44%. The induced
frequency of necrotic
cells was -23%.




Reference




Lemarie et
al., 2006a
































Lemarie et
al., 2006a





C-259  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


U937 cells
co-treated
with PMA















U937 cells
co-treated
with PMA
PBMCs co-
treated
with GM-CSF


Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment


Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF for
6 days before
inorganic
arsenic
treatment



Arsenic
Species


As111 ATO

















As111 ATO


As111 ATO






As111 ATO












As111 ATO










Concentration(s)
Tested (|oM)


1,4

















4


1





1,4

1,4











1










Duration of
Treatment


4 days

















4 days


3 days





3 days

6 days











6 days










LOECa
(HM)


4

















4


1





4

4











None for
u






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Induced apoptosis (i.e.,
experimental - control)
detected using A5/SG
assays: approximate
induced frequencies at
the 2 doses: 3% and
35%, respectively, with
the higher one being
statistically significant.
Induced frequency of
necrotic cells was ~9% at
the highest dose. Other
experiments showed (1)
that ATO induced
apoptosis through
inhibition of NF-KB
signals and (2) that ATO
inhibited macrophagic
differentiation of U937
cells.
U FLIPL protein level, U
XI AP protein level.

U FLIPL protein level
and U FLIPL mRNA
level;
U XIAP protein level and
U XIAP mRNA level.

Induced apoptosis (i.e.,
experimental - control)
detected using A5/SG
assays:
No induced apoptosis at
dose of 1 at either time.
At dose of 4: -22% and
-50% after 3 and 6 days,
respectively; thus these
cells are resistant to
induction of apoptosis by
ATO at low doses.

NSE regarding FLIPL
protein level;
big ft XIAP protein
level.









Reference


Lemarie et
al., 2006a
















Lemarie et
al., 2006a

Lemarie et
al., 2006a





Lemarie et
al., 2006a











Lemarie et
al., 2006a






C-260  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF for
6 days before
inorganic
arsenic
treatment
Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment















Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment







Arsenic
Species


As111 ATO









As111 ATO

























As111 ATO















Concentration(s)
Tested (|oM)


4









0.25,0.5, 1

























0.5, 1,2,4















Duration of
Treatment


3









6 days

























6 days















LOECa
(HM)


4









0.25

























2












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)


Big U FLIPL protein
level; big U XIAP
protein level.





Major alterations in the
morphology, adhesion,
and actin organization
with the impression that
inorganic arsenic "de-
differentiated"
macrophages back into
monocytic cells. The
effect was time-
dependent with rounded
and contracted
morphology first
observed at dose of 1
after only 8 hr. By 6
days at dose of 1 only
3 1% as many cells were
adherent as in control.
Inorganic arsenic
induced a reorganization
of theF-actin
cytoskeleton. The series
of experiments suggested
that the effects occurred
because of the activation
ofaRhoA/ROCK
pathway.

Induced apoptosis (i.e.,
experimental - control)
detected using A5/SG
assays: approximate
induced frequencies at
the 4 doses: 0%, 0%,
20%, and 50%,
respectively. Induced
frequency of necrotic
cells was ~4% at the
highest dose.
18 days of treatment at
dose of 1 caused no
cytotoxicity.




Reference

Lemarie et
al., 2006a









Lemarie et
al., 2006b
























Lemarie et
al., 2006b











C-261  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment








Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment



















Arsenic
Species


As111 ATO



















As111 ATO


























Concentration(s)
Tested (|oM)


1



















1


























Duration of
Treatment


6 days



















6 days


























LOECa
(HM)


1



















1























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

Changes in surface
markers: CD 14: ft 5. Ix;
CD71: U to 45% of
control;
CD29: U to 49% of
control; CD lib: Uto
42% of control.
Changes in major
functions: marked U in
endocytosis and
phagocytosis. Changes
in surface markers and
morphology were shown
to be reversible when
inorganic arsenic was
removed and cells were
cultured with GM-CSF
for 6 days.
Ability to secrete
inflammatory cytokines
in response to co-
treatment of inorganic
arsenic (dose of 1) and
200 ng/mL LPS for 8 or
24 hr
(control = macrophages
treated with LPS only):
TNF-a secretion: ft
~3. Ox and ~3. Ox at 8 and
24 hr, respectively.
IL-8 secretion: ft ~3x
and ~4.5x at 8 and 24 hr,
respectively.
Much more extreme
potentiation was
demonstrated for both
cytokines at the mRNA
level at 8 hr. The text
implies that the
potentiation of both
secretion and mRNA
production does not
occur without the
6-day inorganic arsenic
treatment.




Reference


Lemarie et
al., 2006b


















Lemarie et
al., 2006b






















C-262  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment















Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days
before
inorganic
arsenic
treatment







Arsenic
Species


As111 ATO

























As111 ATO















Concentration(s)
Tested (|oM)


1

























1















Duration of
Treatment


6 days

























8hr















LOECa
(HM)


1

























1












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
The inorganic arsenic-
treated macrophages
differentiated into
dendritic-like cells when
exposed to GM-CSF and
IL-4 in the absence of
inorganic arsenic for 6
days. This conclusion
was based on the ~9x
increase in the
expression of the typical
dendritic marker CD la.
The increase was similar
to that seen in PBMCs
treated with GM-CSF
and IL-4 for 6 days, and
in both cases the
dendritic-like cells were
nonadherent. In contrast,
fully differentiated
macrophages (i.e.,
PBMCs treated with
GM-CSF for 6 days
without inorganic
arsenic) did not show
this response.

ft GTP-binding fraction
ofRhoA;
ft phospho-Moesin
protein level.
(Phosphorylated-Moesin
is a major cytoskeleton
adaptor protein involved
in RhoA regulation.
RhoA is a small GTPase
protein known to
regulate the actin
cytoskeleton in the
formation of stress
fibers.)




Reference


Lemarie et
al., 2006b
























Lemarie et
al., 2006b











C-263  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Differentiated
macrophages
developed
from PBMCs
treated with
GM-CSF
for 6 days and
then
pretreated
with ROCK
inhibitor Y-
27632 for 2 hr
before
inorganic
arsenic
treatment

HepG2 cells











Arsenic
Species



As111 ATO













As111 SA











Concentration(s)
Tested (|oM)



1













0.04, 0.4, 4, 40











Duration of
Treatment



72 hr












48-hr
pretreatment











LOECa
(HM)



1













4








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)


Pretreatment with the
ROCK inhibitor
prevented both the F-
actin reorganization and
cellular rounding of
macrophages treated
with inorganic arsenic.
It also considerably
blunted damage to the
phagocytosis function
caused by the inorganic
arsenic treatment.


After the inorganic
arsenic pretreatment,
there was a 30-min
treatment with IL-6,
which induced STATS
activity unless inhibited
by the pretreatment.
Level of STATS activity:
huge U at 4; no activity
at 40.




Reference


Lemarie et
al., 2006b













Cheng et
al., 2004







C-264  DRAFT—DO NOT CITE OR QUOTE

-------
Type of
Cell/Tissue



HepG2 cells
























Arsenic
Species



As111 SA
























Concentration(s)
Tested (|oM)



0.04, 0.4, 4, 40,
400























Duration of
Treatment


48-hr
pretreatment
























LOECa
(HM)



40
























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
After the inorganic
arsenic pretreatment,
there was a 30-min
treatment with IL-6,
which induced both
STATS tyrosine
phosphorylation and
STATS serine
phosphorylation. Only
the tyrosine
phosphorylation was
inhibited by the
inorganic arsenic
pretreatment, with slight
U at 40 and huge U at
400. Inorganic arsenic is
thought to inactivate the
JAK-STAT signaling
pathway by means of
inhibition of STATS
tyrosine phosphorylation.
Other inflammatory
stimulants, stress agents,
and cadmium failed to
induce similar effects on
the tyrosine
phosphorylation of
STATS.
Reference



Cheng et
al., 2004























C-265  DRAFT—DO NOT CITE OR QUOTE

-------
   Type of
 Cell/Tissue
Arsenic
Species
Concentration(s)
  Tested QaM)
        Duration of
        Treatment
                LOECa
   Results (Compared
  With Controls, With
   All Concentrations
         Being
  in nM Unless Noted)
                                                   Reference
 HepG2 cells
As111 SA
   4, 40, 400
          30-min
        pretreatment
         and 1-hr
        co-treatment
         with IL-6
  Huge U in Cis mRNA
 and in SOCS mRNAs for
 5 of 6 forms tested (U for
 the other form); the U at
 higher doses was usually
 the same or more severe;
 UinSTATmRNAsfor4
 of 6  forms tested, the U
   at higher doses was
 usually the same or more
  severe.  The decreases
 for STAT mRNAs were
 very slight compared to
  those for SOCS. The
 inhibition of induction of
 SOCS mRNA confirmed
    that JAK-STAT
   signaling had been
    turned off. Other
 experiments showed that
  the effect of inorganic
  arsenic on JAK-STAT
      inactivation is
  independent of ligand-
 receptor action and is a
 result of the direct action
 of arsenic on the JAK1
	protein.	
                                                    Cheng et
                                                    al., 2004
 HepG2 cells
As111 SA
 0.04, 0.4, 4,
      400
40,
    inorganic arsenic (in co-treatment with IL-6 for
 unknown duration) activated all 3 subfamilies of MAP
  kinases (i.e., there was phosphorylation of ERK 1/2,
   p38, and JNKs) with LOECs of 40, 0.04, and 0.04
 respectively. Such activation was independent of IL-6
 stimulation at least at higher doses. Experiments with
  specific inhibitors of the 3 MAP kinases showed that
  inorganic arsenic selectively targeted JAK tyrosine
 kinase and that the inhibition of JAK-STAT activity by
inorganic arsenic did not require the participation of any
                  MAP kinases.
                          Cheng et
                          al., 2004
   PBMCs
 treated with
1000 U/mL of
M-CSFatthe
 same time as
with inorganic
   arsenic
As111 SA
  0.005, 0.010,
0.050,0.10,0.50
          7 days
                 0.050
      Cell survival
 demonstrated by trypan
  blue exclusion assay:
  LC50: 0.22; about 25%
  survival at dose of 0.5.
 The cells differentiated
  into adhesive M-type
 macrophages that were
   elongated and had a
 spindle-like morphology.
Sakurai et
al., 2005b
                                                  C-266   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
PBMCs
treated with
5000 U/mL of
GM-CSF at
the same time
as with
inorganic
arsenic











PBMCs
treated with
5000 U/mL of
GM-CSF at
the same time
as with
inorganic
arsenic
















Arsenic
Species


As111 SA


















As111 SA





















Concentration(s)
Tested (|oM)


0.005, 0.010,
0.050,0.10,0.50

















0.50





















Duration of
Treatment


7 days


















7 days





















LOECa
(HM)


0.10


















0.50


















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cell survival
demonstrated by trypan
blue exclusion assay:
-85% survival at 2
highest doses; up to dose
of 0.050, all cells
differentiated into GM-
Mp, which had a round
saucer-like appearance;
at dose of 0.10, -80% of
living cells were GM-Mp
and the rest were
abnormal "arsenite-
induced cells"; at dose of
0.50, -10% of living
cells were GM-Mp and
the rest were "arsenite-
induced cells."
In comparison to
controls (i.e., PBMCs
treated with 5000 U/mL
of GM-CSF and no
inorganic arsenic), the
resulting
morphologically,
phenotypically, and
functionally altered
"arsenite-induced cells"
had:ftHLA-DRto5.0x;
U CD lib to 0.7 Ix;
ftCD14tol.4x;UCD54
to 42% of control; big U
in phagocytic ability; ft
in effectiveness in
inducing allogeneic or
autologous T-cell
responses; and huge ft in
response to bacterial LPS
by inflammatory
cytokine release.




Reference


Sakurai et
al., 2005b
















Sakurai et
al., 2005b


















C-267  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

PBMCs
treated with
5000 U/mL of
GM-CSF at
the same time
as with
inorganic
arsenic











PBMCs
treated
with 1000
U/mL of
M-CSF or
5000 U/mL of
GM-CSF at
the same time
as with
inorganic
arsenic



Arsenic
Species


As111 SA



















Asv











Concentration(s)
Tested (|oM)


0.50



















LCso
determinations










Duration of
Treatment


7 days



















7 days











LOECa
(HM)


0.50



















>1








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
The resulting high
numbers of "arsenite-
induced cells" were
markedly reduced by co-
treatment with DMPO,
DMSO, orBHT, all of
which are membrane-
permeable radical
trapping reagents.
Further indication that
ROS might impact
development of the
"arsenite-induced cells"
was that by using
DCFH-DA it was shown
that ROS levels were
much higher throughout
the 7 days of culturing
and >2x higher on days
1-4 of that period.

Cell survival
demonstrated by trypan
blue exclusion assay:
LC50: 300 for simple
cytotoxicity for both
treatments and with no
toxic effect on
differentiation into
macrophages up to dose
of 1.




Reference


Sakurai et
al., 2005b


















Sakurai et
al., 2005b







C-268  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


PBMCs
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment




















PBMCs
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment
















Arsenic
Species



As111 SA





























As111 SA






















Concentration(s)
Tested (|oM)



1,2,3,4,5





























1,2,3,4,5






















Duration of
Treatment



120 hr





























120 hr






















LOECa
(HM)



1





























1



















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Number of rounds of cell
division estimated using
CFSE dilution assay with
FACS (control had 6
rounds):
5, 4, 3, 2, and 1 rounds
of cell division were
observed after doses of
1, 2, 3, 4, and 5,
respectively; there was a
marked dose-related U in
both proliferation and the
percentage of divided
cells. Additional
staining with 7-AAD
revealed that, at even the
higher doses, most cells
were viable but unable to
divide. The reduced
proliferation resulted
from an ft in the fraction
of non-dividing cells and
a delay in the cell cycle,
with only a comparative
small ft in the number of
dead cells. At the
highest dose, 63% of the
cells were non-dividing,
and 2/3 of them were
alive.
Expression of CD4 and
CDS molecules was
determined using CFSE
staining during the
inorganic arsenic
treatment and then, after
the 96 hr incubation, by
indirect
immunofluorescence
using OKT4 or OKT8
hybridoma supernatants
and goat anti-mouse
IgG-PE, followed by 7-
AAD staining and FACS
analysis. The dose of 1
slightly modified the
expression of both CD4
and CDS. At doses >3: a
marked U in number of
cells expressing CD4; at
doses >4: a marked U in
number of cells
expressing CDS.




Reference




Tenorio
and
Saavedra,
2005
























Tenorio
and
Saavedra,
2005

















C-269  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
PBMCs
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment


Human CD4+
cells
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment

Human CD8+
cells
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment



Arsenic
Species


As111 SA













As111 SA
for both

















Concentration(s)
Tested (|oM)


1,2,3,4,5













1,2,3,5
for both

















Duration of
Treatment


120 hr













120 hr
for both

















LOECa
(HM)


1











1
for both:
a slight
but
signifi-
cant effect












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Evaluation of blast
transformation of both
CD4+ and CD8+ T cells
suggested that they have
different sensitivities to
inorganic arsenic. There
was an accumulation of
resting CD8+ cells with a
positive dose-response;
that accumulation was
not seen for CD4+ cells.


Number of rounds of cell
division estimated using
CFSE dilution assay with
FACS (control CD4+and
CD8+ cells had 6 and 5
rounds, respectively):
At a dose of 1: only 5
rounds in CD4+ but no U
in rounds in CD8+;
however, CD8+ cells had
U in cell number in the
last 3 rounds, arsenic
doses increased in both
cell types: decreasing
numbers of cell divisions
and of numbers of cells
in each round. Effects
were generally more
extreme in CD8+ cells.




Reference

Tenorio
and
Saavedra,
2005










Tenorio
and
Saavedra,
2005













C-270  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Human CD4+
cells
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment

Human CD8+
cells
stimulated
with PHA for
96 hr starting
24 hr after the
beginning of
the inorganic
arsenic
treatment



PBMCs
stimulated
with PHA
during the
inorganic
arsenic
treatment,
CD4+ and
CD8+T cells
were analyzed
separately



















Arsenic
Species





As111 SA
for both



















As111 SA



























Concentration(s)
Tested (|oM)





1,2,3,4,5
for both



















1,2,3,4,5



























Duration of
Treatment





120 hr
for both

















24 hr

48 hr

72 hr

























LOECa
(HM)


1
for both:
a slight
but
signifi-
cant effect
















2

1

1






















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)




CFSE dilution assay with
7-AAD staining and
FACS:
In both cell types there
were apparent
differences from the
control at the dose of 1,
and there was a
progressive U in viable
proliferating cells with
increasing dose, arsenic
doses increased from 0 to
3, there was a much
faster ft in the fraction of
resting cells that was
alive among CD8+ cells
than among CD4+ cells,
and that fraction
remained higher.
LOECs were based on
FACS patterns that
seemed substantially
different as to kinetics of
expression of CD25 and
CD69 in CD4+ T cells.
Inorganic arsenic
delayed both the
expression of CD25 and
the down-regulation of
CD69, suggesting that
inorganic arsenic delays
the activation kinetics of
CD4+T cells. CD4+T
cells exposed to the
highest dose for 72 hr
showed a very similar
pattern to that seen in
non-inorganic arsenic-
exposed cells stimulated
for only 24 hr. A similar
analysis of CD8+T cells
showed similar results;
however, with them there
were somewhat more
CD25"CD69~ cells (i.e.,
cells unable to activate)
as dose increased.




Reference





Tenorio
and
Saavedra,
2005
















Tenorio
and
Saavedra,
2005






















C-271  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



SV-HUC-1
cells























SV-HUC-1
cells














Arsenic
Species



As111 SA
























As111 SA















Concentration(s)
Tested (|oM)



2, 4, 8, 10, 40
























2, 4, 8, 10, 40















Duration of
Treatment



48 hr
























48 hr















LOECa
(HM)


2 for
all effects























2 for
all effects












Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Labeling indices (LIs)
for immunochemistry of
cells:
Bcl-6: ft at 2, increases
with dose as follows: LIs
ofO, 1.04,3.05,6.01,
8.24, and 23. 94 for
control and doses listed,
respectively.
JAK2: U at 2, decreases
with dose as follows: LIs
of 100, 58.1,48.9, 13.0,
5.1, and 0.8 for control
and doses listed,
respectively.
p-STAT3 (Tyrosine
705): ft at 2 with peak at
dose of 4 before
decreasing, as follows:
LIs of 100, 111.7, 151.0,
125.2, 119.0, and 50.8
for control and doses
listed, respectively. All
experimental LIs above
differed from control,
p<0.05.
Effects on protein levels
determined by Western
blotting:
Bcl-6: ft at 2 and
increases with dose.
JAK2:Uat2and
decreases with dose.
P-STAT3 (Tyrosine
705): ft at 2, peak at 4,
less than control at 40.
Results at different doses
were highly consistent
with results obtained
using immunochemistry,
as shown in row above.




Reference



Huang et
al., 2007b























Huang et
al., 2007b











C-272  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


SV-HUC-1
cells
















BAEC cells
















Thymocytes
(freshly
isolated)
Splenocytes
(freshly
isolated)



Arsenic
Species


As111 SA

















As111 SA
















Asv


Asv





Concentration(s)
Tested (|oM)


2, 4, 8, 10, 40

















10
















67, 150,315,680,
1000, 2000

67, 150,315,680,
1000, 2000




Duration of
Treatment


48hr

















48hr
















24 hr


24 hr





LOECa
(HM)

2 for
all effects

















10
















315


150


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Microscopy and
immunochemistry
showed Bcl-6 and p-
STAT3 (Tyrosine 705)
to be localized in the
nucleus and JAK-2 to be
localized in the
cytoplasm.
Morphological changes
began to appear at dose
of 2. At dose of 4, cells
became round and
exhibited nuclear
condensation. At highest
two doses, there was
cellular shrinkage and
cytoplasmic
vacuolization.
ftinLTE4to~5x. Co-
treatment with 50 uM
Mn11, which caused ~9x
ft by itself, caused an
approximately additive ft
to~12x. Addition of L-
NAME to the inorganic
arsenic/Mn co-treatment
boosted LTE4 level to
~24x. Addition of ETU
to inorganic arsenic/Mn
co-treatment boosted
LTE4 level to slightly
above that of inorganic
arsenic/Mn combination.
Addition of AA-861 to
inorganic arsenic/Mn co-
treatment reduced LTE4
level by -80%.
Cell survival determined
using XTT assay:
LC50: 442.
Cell survival determined
using XTT assay: LC50:
427.




Reference


Huang et
al., 2007b
















Bunderson
etal.,2006















Stepnik et
al., 2005

Stepnik et
al., 2005

C-273  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Thymocytes
(freshly
isolated)












Splenocytes
(freshly
isolated)














Arsenic
Species

Asv














Asv
















Concentration(s)
Tested (jiM)

67,315,680














67,315,680
















Duration of
Treatment

24 hr














24 hr
















LOECa
(HM)

315














315













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Point estimates of
induced apoptosis
(experimental minus
control) determined by
TUNEL staining: 5% at
67(NSE); 16% at 3 15
(NSE); and 24% at 680.
27% of control cells
were apoptotic. Agarose
gel electrophoresis of
DNA showed high (and
indistinguishable) levels
of apoptosis in control
group and at the 3
experimental dose levels.
Point estimates of
induced apoptosis
(experimental minus
control) determined by
TUNEL staining: 1% at
67 (NSE); 16% at 3 15;
and 33% at 680. 29% of
control cells were
apoptotic. Agarose gel
electrophoresis of DNA
showed high (and
indistinguishable) levels
of apoptosis in control
group and at the 3
experimental dose levels.




Reference

Stepnik et
al., 2005













Stepnik et
al., 2005












Inhibition of Differentiation
C3H 10T1/2
cell line
(mouse cells
with
fibroblast
morphology
during routine
culture but
capable of
differentiation
into
adipocytes)



As111 SA











6











8wk











6










Complete inhibition of
differentiation into
adipocytes induced by
dexamethasone/insulin
(dexl) treatment. The
effect is the same if
arsenic is removed just
before the dexl
treatment.



Trouba et
al., 2000








C-274  DRAFT—DO NOT CITE OR QUOTE

-------

Type of
Cell/Tissue
C3H 10T1/2
cell line
(mouse cells
with
fibroblast
morphology
during routine
culture but
capable of
differentiation
into
adipocytes)
SIK cells
treated in
surface
cultures
beginning
when they
reached
confluence,
which is when
their rate of
division
decreases as
differentiation
increases


hEp cells
treated in
surface
cultures
beginning
when they
reached
confluence,
which is when
their rate of
division
decreases as
differentiation
increases



Arsenic
Species


As111 SA






As111 SA




As111 SA







Concentration(s)
Tested (|oM)


0.1, 1,3,6, 10






2




2







Duration of
Treatment


48 hr






Various




Various







LOECa
(HM)


3






2




2






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)


Dose-related inhibition
of differentiation into
adiopocytes induced by
dexamethasone/insulin
(dexl) treatment. These
concentrations do not
cause cytotoxicity.




CFE based on assay
using Rhodanile blue
staining: on 1 day post-
confluence both
experimental and control
groups had CFEs of
-11%, by 4 days their
CFEs were -9.2% and
-5.2%, and by 14 days
they were -4.7% and
-0.6%, respectively.
Thus, inorganic arsenic
decreased the exit of
cells from the
germinative
compartment under
conditions that promote
differentiation.

CFE based on assay
using Rhodanile blue
staining: at 4 days post-
confluence experimental
and control groups had
CFEs of -1.1% and
0.25%, by 11 days their
CFEs were -1.0% and
-0.05%, and by 14 days
they were -1.0% and
-0%, respectively. Thus,
inorganic arsenic
decreased the exit of
cells from the
germinative
compartment under
conditions that promote
differentiation.

Reference


Trouba et
al., 2000






Patterson
etal.,2005




Patterson
etal.,2005






C-275  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
SIK cells,
with inorganic
arsenic
treatment
beginning 1
day before
suspension
and
continuing
while cells
were in
suspension,
which drives
such cells
prematurely
into the
differentiation
pathway

hEp cells,
with inorganic
arsenic
treatment
beginning 1
day before
suspension
and
continuing
while cells
were in
suspension,
which drives
such cells
prematurely
into the
differentiation
pathway
SIK cells,
with inorganic
arsenic
treatment
beginning
when they
were put into
suspension,
which drives
such cells
prematurely
into the
differentiation
pathway



Arsenic
Species




As111 SA


















As111 SA
















As111 SA













Concentration(s)
Tested (|oM)




2


















2
















0.1,0.2,0.5,2













Duration of
Treatment
1,2, 3, 4 or 5
days, when
including the 1
day of
treatment
before being
put into
suspension











1 or 2 days,
when
including the 1
day of
treatment
before being
put into
suspension













4 days













LOECa
(HM)




2


















2
















0.1










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
CFE based on assay
using Rhodanile blue
staining comparison with
control (C):
At 1 day : C, -11.0%,
inorganic arsenic,
-10.8%.
At 2 days: C, -0.5%;
inorganic arsenic,
-2.3%.
At 3 days: C, -0.1%;
inorganic arsenic,
-2.0%.
At 4 days: C, -0%;
inorganic arsenic,
-1.3%.
At 5 days: C, -0%;
inorganic arsenic,
-0.8%.


CFE based on assay
using Rhodanile blue
staining comparison with
control (C):
At 1 day: C, -1.15%,
inorganic arsenic,
-1.37%.
At 2 days: C, -0.08%;
inorganic arsenic,
-0.68%.








CFE based on assay
using Rhodanile blue
staining (control CFE =
-0.03%):
Experimental CFEs: 0.1,
-0.10%; 0.2, -0.23%;
0.5, - 0.40%; 2, -
0.80%.








Reference



Patterson
etal.,2005

















Patterson
etal.,2005















Patterson
etal.,2005










C-276  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
RACs from
either the SIK
or hEp cell
line
(did not
specify
which)
treated in
surface
culture

SACs from
either the SIK
or hEp cell
line
(did not
specify
which)
treated in
surface
culture

SIK cells,
with inorganic
arsenic
treatment
beginning
when cultures
reached 90%
confluence













Arsenic
Species

As111 SA










As111 SA










As111 SA



















Concentration(s)
Tested (nM)

2










2










2



















Duration of
Treatment

4,7, 11, or 14
days









4,7, 11, or 14
days









3 days



















LOECa
(HM)

2










2










2
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Mean no. of colonies
present, comparison with
control (C):
At 4 days: C, -3.0;
inorganic arsenic, -18.2.
At 7 days: C, -0.8;
inorganic arsenic, -10.5.
At 11 days: C, -0.7;
inorganic arsenic, -7.8.
At 14 days: C, -0.4;
inorganic arsenic, -5.0.
Mean no. of colonies
present, comparison with
control (C):
At 4 days: C, -1.3;
inorganic arsenic, -5.5.
At 7 days: C, -0.5;
inorganic arsenic, -1.8.
At 11 days: C, -0.6;
inorganic arsenic, -1.5.
At 14 days: C, -0.3;
inorganic arsenic, -1.3.
Relative CFEs based on
Rhodanile blue assay,
with values relative to
the CFE of untreated
cells in medium normally
contained insulin (was
set at 1):
ft to -2.6; if inorganic
arsenic + EGF in
medium: ft to -4.1.
If EGF alone: -1.9; if no
insulin in medium (±
EGF): -3.5. Thus,
inorganic arsenic delays
differentiation and
preserves the
proliferative potential of
keratinocytes.




Reference

Patterson
etal.,2005









Patterson
etal.,2005










Patterson
and Rice,
2007













C-277  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
hEp cells,
with inorganic
arsenic
treatment
beginning
when cultures
reached 90%
confluence











SIK cells,
with inorganic
arsenic
treatment
beginning
when cultures
reached 90%
confluence

















Arsenic
Species

As111 SA


















As111 SA























Concentration(s)
Tested (|oM)

2


















2























Duration of
Treatment

3 days


















9 days























LOECa
(HM)

2


















2




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Relative CFEs based on
Rhodanile blue assay,
with values relative to
the CFE of untreated
cells in medium normally
contained insulin (was
set at 1):
ft to -2.6; if inorganic
arsenic + EOF in
medium: ft to -4.1; if
EOF alone: -2.1; if
neither EOF nor insulin:
-2.1; if EOF but no
insulin: -5.3. Thus,
inorganic arsenic delays
differentiation and
preserves the
proliferative potential of
keratinocytes.
Relative CFEs based on
Rhodanile blue assay,
with values relative to
the CFE of untreated
cells in medium normally
contained insulin (was
set at 1):
ft to -3.8; if inorganic
arsenic + EGF in
medium: ft to -5.1; if
EGF alone: -1.3; if no
insulin in medium: -5.5.
In the absence of
insulin, EGF
substantially augmented
CFE while inorganic
arsenic had no effect.
Thus, inorganic arsenic
delays differentiation and
preserves the
proliferative potential of
keratinocytes.




Reference


Patterson
and Rice,
2007
















Patterson
and Rice,
2007

















C-278  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

SIK and
hEPcells, with
inorganic
arsenic
treatment
beginning
when cultures
reached 90%
confluence









Arsenic
Species



As111 SA















Concentration(s)
Tested (|oM)



2















Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
With regard to effects in the previous 3 rows, EGFR
inhibitors AG1478 and PD 158780 kept inorganic
arsenic from increasing the CFE regardless of the
addition of EOF. They did not affect CFE in the
presence of insulin but largely prevented the ft in CFE
resulting from insulin removal. Inorganic arsenic
treatment caused big ft in active Ras protein, a
downstream effector of EGFR; co-treatment with
AG1478 blocked that effect. Other experiments showed
that the inorganic arsenic treatment blocked the U in
active EGFR protein and the U in active (3-catenin that
normally occur after confluence as cells exit the
proliferative pool and differentiate. Also, expression of
a dominant negative (3-catenin suppressed the ft in
colony -forming ability and yield of putative stem cells
induced by inorganic arsenic and EOF.




Reference


Patterson
and Rice,
2007











Interference With Hormone Function


EDR3 cells
transfected as
described in
paper














COS-7 cells
transfected as
described in
paper








As111 SA

















As111 SA










0.045,0.09,0.18,
0.27, 0.36, 0.45,
0.54, 0.675, 0.9,
1.8,2.7















0.1,0.5, 1.0,2.0,
3.0










-18 hr

















-18 hr











-0.09

















None









ft of hormone-activated
OCR-mediated gene
transcription of reporter
genes containing TAT
glucocorticoid response
elements in the presence
of activated OCR; peak
response was at -0.5;
however, inorganic
arsenic was inhibitory at
doses of 1.8 and 2.7.
Other experiments
showed a similar effect
on the endogenous TAT
gene and also that the
central DNA binding
domain of the OCR is
the minimal region
required for the arsenic
effect.
inorganic arsenic had no
effect on transcriptional
repression by OCR.
That is, arsenic had no
effect on the ability of
hormone-activated OCR
to inhibit API expression
or
NF-KB-mediated gene
expression.


Bodwell et
al., 2004
















Bodwell et
al., 2004








C-279  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

EDR3 cells
transfected as
described in
paper






EDR3 cells
transfected as
described in
paper









EDR3 cells
transfected as
described in
paper











NHEK cells








Arsenic
Species

As111 SA









As111 SA












As111 SA













As111 SA

Asv,
MMAV,
DMAV





Concentration(s)
Tested (|oM)
0.045,0.09,0.18,
0.27, 0.36, 0.45,
0.54, 0.675, 0.9,
1.8,2.7






0.045,0.09,0.18,
0.27, 0.36, 0.45,
0.54, 0.675, 0.9,
1.8,2.7










Duration of
Treatment

~18hr









~18hr












LOECa
(HM)

-0.09









-0.09









Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft of hormone-activated
OCR-mediated gene
transcription of reporter
genes containing TAT
response elements in the
presence of activated PR;
peak response was at
-0.5; however, inorganic
arsenic was inhibitory at
doses of 0.9, 1.8 and 2.7.
ft of hormone-activated
OCR-mediated gene
transcription of reporter
genes containing TAT
response elements in the
presence of activated
MCR; peak response was
at -0.5; however,
inorganic arsenic was
inhibitory at doses of 1.8
and 2.7.
For all 3 steroid receptors tested (OCR, PR and MCR — see 3 rows
immediately above this one), the degree of stimulation at lower inorganic
arsenic concentrations or repression at higher inorganic arsenic concentrations
was highly dependent on, and inversely related to, the amount of activated
steroid receptor within cells. The relative increases in transcription noted
above, which were up to ~2x or more above control levels, were at the lowest
levels of activated steroid receptor within cells that were tested. Other studies
showed that iA (1) had no significant effect on cellular steroid levels or on
binding of steroid to the receptor, (2) did not activate or act as an agonist for
OCR, (3) did not act as a competitive antagonist, (4) did not appear to affect
the ability of the hormone to bind to or activate OCR, (5) did not appear to
affect hormone-stimulated nuclear translocation of OCR, and (6) did not
significantly alter the level of OCR for either cells expressing endogenous
OCR or those expressing stably integrated OCR. Dimerization is not critical
for the response to inorganic arsenic. In summary, it is clear that inorganic
arsenic can simultaneously disrupt multiple hormone receptor systems.


0.001,0.005,0.01,
0.05,0.1,0.5, 1,5
for all




24 hr


24 hr




0.001


0.001-0.5



For cytokines GM-CSF,
TNF-a, andIL-6:
substantial ft at 0.001-
0.01, but no change or U
(sometimes markedly) at
0.05-5.
No change or U
(sometimes markedly).




Reference

Bodwell et
al., 2006








Bodwell et
al., 2006












Bodwell et
al., 2006












Vegaetal.,
2001




C-280  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Hormone-
responsive
H4IIE (rat
hepatoma cell
line)











Arsenic
Species

As111 SA














Concentration(s)
Tested (nM)

0.3, 1.0,2.0,3.3














Duration of
Treatment

2hr














LOECa
(HM)

0.3











Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
U in hormone-inducible
expression of GRE2-Luc
with a 2-hr As111
pretreatment before an
18-hr Dex treatment.
The pretreatment did not
block the normal Dex-
induced nuclear
translocation of
glucocorticoid receptor.
As111 selectively inhibited
glucocorticoid-receptor-
mediated transcription.




Reference

Kaltreider
etal.,2001










Malignant Transformation or Morphological Transformation
HaCaT cells







TRL 1215
cells (normal
rat liver)



JB6 C141
cells
simultaneousl
y treated with
10 ng/mL
EOF




JB6 C141
cells







Primary
Syrian
hamster
embryo cells
(HEC)
As111 SA








As111 SA




As111 SA

Asv







As111 SA








Asv




0.5, 1.0







0.125,0.250,
0.500




25, 50, 200

12.5, 50, 200







25, 50, 100








13,27




20 passages







18 wk





14 days for
both







4 wk followed
by 4 wk at
lower
concentration






7-8 days




0.5







0.250





50

12.5







25








13




Cells became
tumorigenic; tumors
were produced by 2
months after injection of
cells into Balb/c nude
mice; cells from tumors
were much more
malignant.
Transformed cells
produced aggressive
tumors capable of
metastasis after
inoculation into nude
mice.
Inhibition of EGF-
induced cell
transformation:
The effect was much
stronger for Asv (sodium
arsenate) with complete
blockage of
transformation at 50 and
200.
Transformed cells, as
shown by growth of
colonies in soft agar;
transformation did not
occur at the 2 higher
doses; SA-induced
transformation was
blocked by introduction
of dominant negative
Erk2.
Morphologically
transformed cells.



Chien et
al., 2004






Zhao etal.,
1997




Huang et
al., 1999b








Huang et
al., 1999a







DiPaolo
and Casto,
1979


C-281  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Syrian
hamster
embryo cells





Human
osteosarcoma
TE85 (HOS)
cells




Untransforme
dand
immortalized
RWPE-1 cells
(human
prostate
epithelial cell
line)



SHE cells






SHE cells










Arsenic
Species
As111 SA

Asv




As111 SA



MMAm





As111 SA








As111 SA

DMAmI




As111 SA

Asv









Concentration(s)
Tested (|oM)
-0.8, 1.6, 3, 3.5, 5

~8, 16, 32, 64, 128




0.0125, 0.025,
0.05,0.1


0.00625, 0.0125,
0.025, 0.05




5








1,3, 10

0.1,0.2,0.4, 1.0




3, 6, 8, 10

50, 100, 150









Duration of
Treatment
7 days
for all






6 and 8 wk
for both







29 wk








48 hr
for both





48 hr
for both










LOECa
(HM)
0.8

8



0.025 at 8
wk; 0.05
at
6 wk

None





5








1

0.1




6

50






Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Morphological
transformation:
For both chemicals: a
positive dose-response
throughout the dose
range tested.


Transformation to
anchorage-independence
in soft agar As111: positive
dose-response to highest
concentration; 8 weeks
was -40 generations;
MMAm was more toxic
than inorganic arsenic111.

Aggressive tumors were
produced after cells
showing ft secretion of
MMP-9 were inoculated
into nude mice.


Morphological
transformation (% of
surviving colonies
transformed at each
concentration):
1,0. 11%; 3, 0.23%; 10,
0.48%.
0.1, 0.28%; 0.2, 0.51%;
0.4, 3.41%, 1.0, 3.35%.
Neoplastic
transformation based on
anchorage-independent
growth and/or
tumorigenicity in
newborn hamsters. All 5
anchorage-independent
cultures tested were
tumorigenic.




Reference
Barrett et
al., 1989







Mure et
al., 2003






Achanzar
etal.,2002








Ochi et al.,
2004




Takahashi
etal.,2002







C-282  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

UROtsa cells
















NIH 3T3 cells































Arsenic
Species

MMAm
















As111 SA































Concentration(s)
Tested (|oM)

0.05
















2, 5, 10, 20, 50,
100, 200






























Duration of
Treatment

52 weeks
















7 days































LOECa
(HM)

0.05
















2




























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Anchorage-independent
growth as detected by
colony formation in soft
agar; cells from those
colonies showed
enhanced tumorigenicity
in SCID mouse
xenografts. After only
24 weeks there was also
much anchorage-
independent growth, but
those cells did not show
the enhanced
tumorigenicity.
Anchorage-independent
growth in soft agar
assayed using
AlamarBlue dye assay
and microscopic
examination: ft to ~1.4x
control at 2 and 5 ; NSE
at 10, marked dose-
related U at higher doses.
A daily 2-hr 42°C heat
shock (which would
induce HSPs) boosted
induction of anchorage-
independent growth for
up to 3 repetitions, but 5
heat-shock repetitions
markedly reduced such
growth. When the same
experiment was repeated
in R-3T3 (transformed)
cells, there was NSE by
inorganic arsenic or heat
shock on the already
high level of anchorage-
independent growth;
inorganic arsenic caused
U at dose of 20, and at
higher doses the U
became marked, as it did
also at all doses
following 5 daily
repetitions of the heat-
shock treatment.




Reference

Bredfeldt
etal.,2006















Khalil et
al., 2006



























C-283  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue


BALB/c 3T3
A3 1-1-1 cells
(derived from
mice)































Arsenic
Species
As111 SA

AsvDA

MMAV

DMAV






























Concentration(s)
Tested (nM)
2, 5, 10, 15, 20

10, 15, 20, 25, 30

1,2,5, 10 mM

0.5, 1,2, 5 mM






























Duration of
Treatment



72 hr for all

































LOECa
(HM)
5

15

10 mM

ImM



























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Caused initiation in a
two-stage transformation
assay; based on a
significant increase in
the number of
transformed cells after an
initiating treatment with
an arsenic compound for
72 hr followed by post-
treatment with 0.1
ug/mL TPA for 18 days.
Except for As111 SA,
responses were stronger
at higher doses; with it,
the peak response was at
10, with a steep decline
by 20. Slight but
significant
transformation occurred
even without TPA at the
2 highest doses of As111
SA and for 2 mM DMAV.
The ranges of positive
effects in foci/dish in the
two-stage transformation
assay (from the LOEC to
the peak) for each
arsenical were as
follows: As111 SA, 1.80-
3.90; AsvDA, 1.20-
2.90; MMAV, 1.10 (only
1 positive value); DMAV,
1.0-3.10. The control
value was 0.30.




Reference


Tsuchiya
et al., 2005






























Signal Transduction
MGC-803
(human
gastric
cancer)



Primary
cultures of rat
cerebellar
neurons
As111 ATO






As111 SA

DMAV

0.01-1






10

5mM

48hr






4hr

8hr

0.01






10

5mM

Increase in intracellular
Ca2+ as measured by a
Ca2+ sensitive
fluorescent probe Indo-
1/AM in flow cytometric
assays, which parallels
the effect on apoptosis.

For both: ft in activated
p38MAPkinase.

Zhang et
al., 1999





Namgung
and Xia,
2001

C-284  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
Primary
cultures of rat
cerebellar
neurons




SY-5Y cells

HEK 293
cells



















Arsenic
Species

As111 SA






As111 ATO
for both





















Concentration(s)
Tested GoM)

10






0.1, 1
for both





















Duration of
Treatment

Ihr






~lhr
for both





















LOECa
(HM)

10






0.1

0.1

















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft JNK3 MAP kinase.
No change in JNK1 and
JNK2 MAP kinases.
(Blocking the p38 and
JNK signaling pathways
inhibited arsenite-
induced apoptosis.)
The Ca2+ concentration
in cells was substantially
increased (and by rather
similar amounts) by both
doses; inorganic arsenic
triggered 3 different
kinds of Ca2+ signals:
slow (sustained),
transient elevations, and
calcium spikes. The
irreversible increases
were independent of
extracellular Ca2+ and
dependent on internal Ca
stores, which could
become depleted. Little
or no cytotoxicity
resulted from these doses
during the time of
measuring Ca2+
concentrations.




Reference
Namgung
and Xia,
2001






Florea et
al., 2007

















C-285  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue




UROtsa cells























Postconfluent
PAEC cells in
a monolayer





















Arsenic
Species
As111 SA

Asv

MMAmO

MMAV

DMAmI

DMAV


















As111 SA






















Concentration(s)
Tested (|oM)
0.1,0.5, 1,5

1, 10, 100

0.1,0.5, 1,5

1, 10, 100

0.1,0.5, 1,5

1, 10, 100


















0.5,2,5






















Duration of
Treatment



Up to
2hr
for all























Ihr






















LOECa
(HM)
Various

None

Various

None

Various

None


















2



















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)

Phosphorylation of
ERK2: ft with potencies:
MMAmO » DMAmI »
As111
AP-1 binding activity:
for As111: U at 0.1 and 0.5;
forMMAmO,bigftat
0.1,0.5, 1.0 but no
increase at 5; for
DMAmI: UatO.l, 0.5,
and 1, and big ft at 5.
Phosphorylation of c-
Jun: for As111: Hat 0.1
and 0.5 and ft at 1 and 5;
forMMAmO,bigftatl
and5;forDMAmI:ftat
0.1, big ft at 5.
Also trivalent arsenicals
caused changes in Fra-1
and induced AP-1
dependent gene
transcription. There was
no effect on c-Jun N-
terminal kinases or p38
kinases.
EMS A analysis: ft
nuclear retention of NF-
KB binding proteins; ft
nuclear translocation of
NF-KB binding proteins.
Supershift analysis
showed that p65/p50
heterodimers accounted
for the majority of
proteins binding
consensus KB sequences
in cells treated with As111
or oxidants. These and
other experiments
suggest that As111 initiates
vascular dysfunction by
activating oxidant-
sensitive endothelial cell
signaling. Increased
binding of proteins to
genomic KB sites could
induce a mitogenic or
inflammatory response.




Reference



Drobna et
al., 2002























Barchowsk
y et al.,
1996


















C-286  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

Gclm+/+ MEF
cells





Hepa-lclc7
cells












1RB3AN27
cells














1RB3AN27
cells





Arsenic
Species
As111 SA
for all






As111 SA













As111 SA















As111 SA






Concentration(s)
Tested (|oM)



Duration of
Treatment



LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
See rows under Apoptosis and Cytotoxicity for this citation for experimental
conditions. Inorganic arsenic inhibits NFicB activation and nuclear
translocation. Co-treatment or pretreatment with tBHQ appears to reverse the
inorganic arsenic -mediated inhibition of NF-KB translocation, and it triggers
the nuclear accumulation of the transcription factor Nrf2. tBHQ may cause its
cytoprotective effects by inducing gene expression changes though activation
of at least NF-KB and Nrf2.

6, 12, 25, 50













0.1,0.5, 1,5, 10















0.1,0.5, 1




Ihr













240 min















120 min




6













Various















0.1



ft AhR nuclear
translocation, with a
positive dose-response;
other experiments
showed that the
translocation occurs by
different mechanisms
from those followed by
ligands and that AhR-
dependent gene
expression is only
weakly up-regulated by
inorganic arsenic.
Activation of nuclear
transcription factors
detected by EMSA: NF-
KB: slight ft at 0.5, flat
1, huge ft at 5 and 10;
effect at dose of 1 was
considerably suppressed
by co-treatment with
NAC. Inorganic arsenic-
induced degradation of
IicBawas demonstrated
in the cytosolic fraction.
AP-1: ft at 0.1, slight ft
at 0.5, huge ft at 1, 5, and
10; effect at dose of 1
was completely blocked
by co-treatment with
NAC.
Phosphorylation
(activation) of ERK
detected by EMSA: huge
ft at all 3 doses.




Reference

Kannet
al., 2005b





Kannet
al., 2005a












Felix etal.,
2005














Felix et al.,
2005


C-287  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

JB6C141
cells

















Hepa-lclc7
cells































Arsenic
Species

As111 SA
for both

















As111 SA
































Concentration(s)
Tested (|oM)
20,40

40














2, 5, 10

0.1, 1,2,5, 10


2, 5, 10

2, 5, 10




























Duration of
Treatment
12 hr
for both















5hr

48hr


5hr

5hr




























LOECa
(HM)
20

40














2

0.1


None

2

























Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
ft COX-2 protein level.

ft COX-2 transcription.
However, deletion of
NF-KB binding sites
from the
COX-2 promoter
blocked this effect.
Other experiments,
including some in MEF
cells, confirmed the
requirement of the
IKKJ3/NF-KB pathway
for the induction of
COX-2 by As111 (shown
at protein and
transcription levels).
ftNqolmRNA
expression, with a
positive dose-response.

ft Nqol enzyme activity,
with a slightly higher
and rather similar
response at doses 1-10.
NSEonNrf2mRNA
levels.
ft Nrf2 protein level,
with a positive dose-
response.
These and other
experiments showed that
Nqol induction occurred
through the Nrf2/ARE
pathway with the
following important
steps: (1) inorganic
arsenic markedly
stabilizes Nrf2; (2)
inorganic arsenic
disrupts the Nrf2-Keapl-
Cul3 complex in the
nucleus, and (3)
inorganic arsenic recruits
NrfZandMaftothe
ARE of Nqol. Inorganic
arsenic does not recruit
Keapl, Cul3, ubiquitin,
c-Jun, or c-Fos to the
ARE of Nqol.




Reference

Ouyang et
al., 2007

















He et al.,
2006




























C-288  DRAFT—DO NOT CITE OR QUOTE

-------
  Type of
 Cell/Tissue
 Arsenic
 Species
Concentration(s)
  Tested QaM)
  Duration of
  Treatment
LOECa
  Results (Compared
 With Controls, With
  All Concentrations
        Being
 in nM Unless Noted)
                                   Reference
 WM9 cells
OM431 cells
K1735-SW1
    cells
  and other
 melanoma
  cell lines
 As111 SA
     2,4,6
    A series of experiments (usually with durations of
 treatment between 30 min and 16 hr) demonstrated that
    inorganic arsenic up-regulated TRAIL-mediated
 apoptosis. Inorganic arsenic up-regulated surface levels
 of death receptors, TRAIL-R1 and TRAIL-R2, through
 increased translocation of these proteins from cytoplasm
   to cell surface.  Furthermore, activation of cJun and
  suppression of NF-KB by inorganic arsenic caused up-
    regulation of the endogenous TRAIL and down-
    regulation of cFLIP gene expression, which was
 followed by cFLIP protein degradation and, finally, by
 acceleration of TRAIL-induced apoptosis.  cFLIP is one
                    of the main
	anti-apoptotic proteins in melanomas.	
                                   Ivanov and
                                    Hei, 2006
 HeLa cells
 As111 SA
      100
     4hr
  100
       Big ft in
  autophosphorylation
  (activation) of ASK 1
     determined by
    autoradiography.
Hansen et
 al., 2006
 A431 cells
As111 ATO
       20
   Many experiments, usually at dose of 20, and over
 various durations, and often involving inhibitors or other
   modulators, yielded the following information and
 conclusions: inorganic arsenic had following effects: ft
 p21 promoter activity, ft p21 mRNA level, ft p21 protein
 level.  Transfection with a p21 siRNA reduced inorganic
    arsenic-induced p21 expression and reduced the
  inorganic arsenic-induced cytotoxicity after 24 hr by
   half.  Conclusions: inorganic arsenic induced p21
  activation via the EGFR-Ras-Raf-ERKl/2 pathway.
   ERK1/2 and JNK may differentially contribute to
 inorganic arsenic-induced p21 expression via the EGFR-
   Ras-Raf-ERKl/2 pathway. The ERK 1/2 and JNK
   pathways play opposing roles in inorganic arsenic-
	induced cytotoxicity.	
                                    Huang et
                                    al., 2006
NHEK cells
               As111 SA
                  0.4
                     1,3,5,7
                   0.4
                 on days 3
                and 5 only
            CyclinDmRNA level:
           ftto~3.2xonday3;ftto
           ~1.5x on day 5; NSE on
                 other days.
                         Hwang et
                          al., 2006
NHEK cells
 As111 SA
      0.4
    1,3,5,7
  0.4
on day 3
  only
 Binding of transcription
factors to their respective
binding motifs within the
 cyclin D1 promoter by
 demonstrated by EMSA:
 ft for API to 1.9x;NSE
   on other days; ft for
 CREEP to 1.6x; NSE on
  other days.  Note the
correspondence with ft in
   mRNA level in row
 above; there was a hint
 of an ft for both on day
          7.
Hwang et
 al., 2006
                                                 C-289  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

NHEK cells




DU145 cells






















HaCaT cells,
trans-fected
for use in a
luciferase
reporter assay

HaCaT cells



















Arsenic
Species
As111 SA





As111 SA























As111 SA




As111 SA



















Concentration(s)
Tested (|oM)

0.1,0.2,0.4




50, 100, 200























1.25,2.5,5




0.31, 1.25,5



















Duration of
Treatment

1,2,3,4,7




Ihr























12 hr




12 hr



















LOECa
(HM)
0.4
on days
2-7



50























1.25




0.31
















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Cyclin D protein level: ft
to~1.35xondays2-4;ft
to 2x on day 7.
AMPK activation: ft at
50, big ft at 100, ft at
200;
activation was blocked
by preincubation with
CAT, GSH, orNAC;
tests with a dominant
negative form of AMPK
showed that AMPK
activity is necessary for
inorganic arsenic-
induced VEGF
expression. Other
experiments showed that
the arsenic -induced
AMPK signaling
pathway is independent
ofthep38MAPkinase
and
PI-3 kinase pathways
and that the blocking of
AMPK activation
markedly increased
cytotoxicity from
inorganic arsenic
exposures of 50 or 100.

ft cyclinDl transcription
to 1.9x and then ft with
dose to 2.4x at highest
dose.
Protein levels determined
by Western blot assay: ft
cyclin D 1 and then ft
with dose to highest
dose; other experiments
showed that induction of
cyclinDl required
activation of NF-KB and
also required IKK|3. It
was suggested that the
inorganic arsenic-
induced stimulation of
the transition from Gl to
S phase that was
reported in this paper
occurred through a
IKK|3/NF-KB/cyclin D 1 -
dependent pathway.




Reference
Hwang et
al., 2006




Lee et al.,
2006c






















Ouyang et
al., 2005



Ouyang et
al., 2005















C-290  DRAFT—DO NOT CITE OR QUOTE

-------
   Type of
 Cell/Tissue
 Arsenic
 Species
Concentration(s)
  Tested QaM)
  Duration of
  Treatment
LOECa
   Results (Compared
  With Controls, With
   All Concentrations
         Being
  in nM Unless Noted)
                                    Reference
   Primary
keratinocytes
   (in third
   passage)
obtained from
 foreskins of
    adults
 As111 SA
   0.1, 1,5, 10
     48 hr
Various
    Enzyme activities
  detected by luciferase
   assays: NF-KB: ft to
      -l.SxatO.l,
 ft to ~5x at 1, ft to ~3.8x
 at5,NSEatlO.  AP-1:
 NSEatO.l, ftto~1.7xat
           1,
   ft to ~2.7x at 5, ft to
 ~4.5x at 10. All results
  were confirmed at the
	protein level.	
Liao etal.,
   2004
 H9c2 cells
 As111 SA
    1,2.5,5
 for 1 or 2 days
 There was a large dose-related U in cell migration rates
 at the 2 higher doses at both durations.  NSE on viability
 (based on MTT assay) at these 3 doses, but at the dose
 of 10, which was not tested for other effects, there was a
   U in viability.  There was a dose-related U in focal
        adhesions per cell at all doses and a U in
 F-actin content of cells at the dose of 5. At doses of 2.5
   and 5, there was a U in tyrosine phosphorylation of
 FAK, a U in phosphorylation of FAK at phosphotyrosine
   397, and a U in tyrosine phosphorylation of FAK's
   substrate paxillin. Inorganic arsenic affected focal
 adhesion structure or formation and not the turnover or
 amounts of focal adhesion proteins. Focal adhesions are
 involved in integrin signaling, and the inorganic arsenic-
    induced changes may disrupt cell contraction and
   signaling. It was concluded that inorganic arsenic
   decreases cell migration through an effect on focal
 adhesions and by disruption of cell interactions with the
                 extracellular matrix.
                                    Yancy et
                                    al., 2005
 MEFs from
 wild type or
  Ikk(3 gene
  knockout
   As111
 (AsCl3)
Various between
   1.25 and 50
   mouse
  embryos
  In a series of experiments lasting for 2-32 hr, the main
  findings were as follows.  In knockout MEFs, which
  exhibit NF-KB inhibition due to IKKp deficiency, (1)
    there was a big ft in basal levels of mRNAs of the
    following genes: gadd45a, gadd45p, gadd45y and
   gadd!53; (2) there was a big ft in inorganic arsenic-
 induced (usually at 20 uM for 4 hr) levels of mRNAs for
 gadd45a and gadd!53; and (3) there was no induction by
  arsenic (same conditions) of mRNAs for gadd45p and
     gadd45y.  It appears that NF-KB activation is an
   inhibitory signal for the expression of gadd45a and
  gadd!53. C-myc expression was reduced in knockout
   cells, and IKKp deficiency did not affect p53 or Akt
	signaling and the expression of FoxO3a.	
                                     Zhang et
                                     al., 2005
  JAR cells
As111 ATO
 0.5, 1,2.5,5, 10
      6hr
  0.5
  Big ft in nuclear Nrf2
 protein level, with dose-
 related ft becoming huge
   by dose of 10; also
 similar ft in cytoplasmic
   Nrf2 protein level.
 Massrieh
et al., 2006
                                                   C-291   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
JAR cells
















BEAS-2B
cells






SVEC4-10
cells














SVEC4-10
cells










Arsenic
Species
As111 ATO
















As111 SA







As111 SA















As111 SA











Concentration(s)
Tested (|oM)
5
















5







10















10











Duration of
Treatment
2, 4, 6, 16,
24 hr















4hr







3 min to 4 hr















3 min to 4 hr











LOECa
(HM)
Various
















5







10















10








Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Big ft in nuclear Nrf2
protein level at first 4
time points, but small ft
at24hr. Slight ft in
MafF protein level at
many time points, but
NSE on 2 other
dimerization partners of
Nrf2, namely MafG and
MafK. Experiments
done in part in HEK293T
cells suggested that in
JAR cells there is an ft in
binding of endogenous
Nrf2/smallMafDNA-
binding complexes to a
StRE site.
Huge ft in nuclear Nrf2
protein level. Other
experiments showed
inorganic arsenic caused
ft in Nrf2 transcriptional
complex binding to the
HMOX-1 ARE cis
element.
inorganic arsenic
induced actin filament
reorganization to form
lamellipodia and
filopodia structures at the
leading edge of the cells
and rosette-like
structures in the cell
bodies. Effects were
noted after only 3 min;
longer treatments did
more damage.
Reorganization of actin
filament occurred
through the activation of
Cdc42.
Huge ft in activation of
Cdc42 already after 3
min and level of
activation stayed almost
as high for at least 1 hr;
by 4 hr the level of
activation was similar to
that of control. See
comment in row above.




Reference
Massrieh
et al., 2006















O'Hara et
al., 2006






Qian et al.,
2005














Qian et al.,
2005







C-292  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
SVEC4-10
cells
















JB6C141
cyclinDl-Luc
massl cells



JB6C141
cyclin
Dl-Luc
massl cells
JB6C141
cells
















Arsenic
Species
As111 SA

















As111 SA





As111 SA



As111 SA

















Concentration(s)
Tested (|oM)
10

















5





2



0.1,0.5, 1,5, 10

















Duration of
Treatment
Various

















24hr





12 hr



Ihr

















LOECa
(HM)
Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
As more information about series of
experiments described in 2 rows
immediately above, it was shown that
inorganic arsenic-stimulated Cdc42-
induced actin filament reorganization
regulated the activation of NADPH
oxidase. Authors suggested that the
formation of superoxide anion radicals
observed after inorganic arsenic
treatment occurred through the
activation of NADPH oxidase. Rac
activities were required for Cdc42-
mediated superoxide anion radical
production, and NADPH oxidase
activity was involved in inorganic
arsenic-stimulated cell migration via
Cdc42-mediated actin filament
reorganization.
5





2



Various














Protein level determined
by Western blot assay:
huge ft in cyclin Dl;
separate treatments with
vanadate, cadmium, or
NiCl2: NSE.
mRNA level determined
by luciferase reporter
assay:
ft in cyclin Dl to~3.2x.
Protein levels determined
by Western blot assay:
Phosphorylation of Akt
Ser473:ftat0.1-5,bigft
at 10.
Phosphorylation of Akt
Thr308: slight U at 0.1, U
at 0.5, ft at 1 and 5, big ft
at 10.
Phosphorylation of
p70S6KThr389:bigftat
0.1-10.
Phosphorylation of
p70S6KThr421/Ser424:ft
at 0.1-10.




Reference
Qian et al.,
2005
















Ouyang et
al., 2006




Ouyang et
al., 2006


Ouyang et
al., 2006













C-293  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue
JB6C141
cells














HEL cells




















A549 cells













Arsenic
Species
As111 SA














As111 ATO





















As111 SA













Concentration(s)
Tested (|oM)
5













0.25,0.5, 1,2,5,
8, 10forP-STAT3

0.5, 1, 2, 4, 6, 8,
10
forHSPVO

















1, 5, 10, 20













Duration of
Treatment
20min














6hr
for both




















24 hr













LOECa
(HM)
5













0.5


1



















10










Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Protein level determined
by Western blot assay:
big ft in PI-3K activation
as shown by big ft in
PIPS; inhibition
experiments showed that
inorganic arsenic
triggered a
PI-3K/Akt/IKKp/NF-KB
signal cascade that
played essential roles in
inducing cyclin D 1
expression.
Western blot analysis:
U P-STAT3 protein level
(IC50s=1.334);3HSP90
inhibitors all markedly
potentiated the effect
with iC50s of 0.0468,
0.395, and 0.745.
ft HSP70 protein level.
Dose of -2.9 doubled the
control level. 3 HSP90
inhibitors all markedly
potentiated the effect.
HSP70 inhibits
apoptosis. Much more
inorganic arsenic was
needed to kill half the
cells in 6 hr (LC50, e t =
80) than to
down-regulate P-STAT3
by 50% in 6 hr (1.334, as
above). The trypan blue
assay was used to
determine cell survival.
inorganic arsenic
activated the binding of
IRP-1 to IRE: ft to 1.35x,
with smaller ft to 1.25x
at dose of 20; 10 and 20
uM inorganic arsenic
caused a slight ft in HIF-
1 a protein level (only
-2% above control).
Inorganic arsenic at dose
of 20 had NSE on ferritin
protein level.




Reference
Ouyang et
al., 2006














Wetzler et
al., 2007



















Li et al.,
2006b









C-294  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

CL3 cells,
synchronous
atGl












CL3 cells,
synchronous
atGl


































Arsenic
Species

As111 SA














As111 SA




































Concentration(s)
Tested (|oM)

50, 100














5, 10, 25, 50




































Duration of
Treatment

3hr














3hr




































LOECa
(HM)

50














Varied

































Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
At 50, ft in
phosphorylation
(activation) of ERK1/2
to 1.7x right after
treatment and to 2.0x
after a 24-hr recovery
period. At 100: ft to
2.3x immediately; no
activation of ERK1/2
occurred following co-
treatment with PD98059
orU0126.
Dose-related ft in
phosphorylation
(activation) of ERK1/2
to~1.45xat25and
~1.8xat50,LOEC=10.
Dose-related U in
efficiency of synthesis of
NER to -50% and -41%
of control at doses of 25
and 50, respectively.
LOEC = 25;forboth
ERKl/2andNER,the
changes at 5 and 10 were
NSE. NER synthesis
efficiency was
determined based on
whole cell extracts of
treated cells in an assay
with UV-irradiated
pUC19 plasmids. Co-
treatment of inorganic
arsenic with either
PD98059orU0126
blocked much of the
phosphorylation of
ERK1/2 and restored
50%-80%oftheNER
synthesis efficiency. In
summary, co-treatments
of inorganic arsenic with
inhibitors that blocked
activation of ERK1/2 did
the following: (1) ft NER
synthesis efficiency, (2)
U induction of
micro nuclei, (3) U
survival, and (4) U rate
of proliferation.




Reference

Li et al.,
2006a














Lietal.,
2006a































C-295  DRAFT—DO NOT CITE OR QUOTE

-------
  Type of
Cell/Tissue
 Arsenic
 Species
Concentration(s)
  Tested QaM)
 Duration of
  Treatment
LOECa
 (HM)
 Results (Compared
With Controls, With
 All Concentrations
       Being
in nM Unless Noted)
Reference
                                            Effects detected by Western blot assay following 30-min
                                                treatment: increases in p-EGFR, p-ERK, p-JNK,
A431 cells
As111 ATO
                                             pp-38, andp21
                                                          WAFl/CIPI
                                                   (an immunoblot assay was used
       20
    for p21). NSE for INK.  A series of experiments
  involving modulators and reporter genes showed that:
   (1) EGFR activation can mediate inorganic arsenic-
   induced p21 expression, (2) activation of EGFR by
 inorganic arsenic occurred later than by EOF, (3) c-Src
    was involved in inorganic arsenic-induced ERK
         activation and p21 expression, (4) the
  EGFR/Ras/Raf/ERK pathway is involved in inorganic
  arsenic-induced p21 gene expression, (5) Spl binding
 sites in the promoter are essential for inorganic arsenic-
 induced p21 activation, and (6) a post-transcriptional or
  post-translational stabilization mechanism is essential
                        for
	inorganic arsenic-induced p21 expression.	
                                     Liu and
                                     Huang,
                                      2006
MDA-MB-
 435 cells
 As111 SA
1, a non-cytotoxic
      dose
    0.5 hr,
     Ihr
     2hr
     4hr
     6hr
            Effects on the nuclear
           binding of the following
            4 transcription factors,
           relative to control and in
           sequential order of the 5
                time periods:
             AP-LNSE, NSE, ft
             2.5x, ft 2.5x, NSE.
           NF-KB: NSE, U 0.5x, ft
             3.5x, ft 3.5x, ft 1.5x.
            Spl:li0.5x, Uo.5x, ft
               3x, NSE, NSE.
           YB-1:NSE,NSE, ft 9x,
                 ft 3x, ft 3x.
             Another experiment
           using a highly cytotoxic
            dose of 100 resulted in
             markedly different
           patterns over time within
           approximately the same
               ranges of effect.
                        Kaltreider
                       etal., 1999
                                                 C-296   DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



H41 IE cells

















SIK cells,
with inorganic
arsenic
treatment
beginning 1
day before
suspension
and
continuing
while cells
were in
suspension














Arsenic
Species



As111 SA



















As111 SA























Concentration(s)
Tested (|oM)



0.33, a
non-cytotoxic
dose

















2























Duration of
Treatment

0.5 hr,
Ihr
2hr
4hr
6hr














2 or 5 days,
when
including the 1
day of
treatment
before being
put into
suspension



















LOECa
(HM)



—



















2




















Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Effects on the nuclear
binding of the following
4 transcription factors,
relative to control and in
sequential order of the 5
time periods:
AP-LNSE, ft5.5x, ft
8.5x, NSE, U 0.5x.
NF-KB:ft 1.3x,NSE, ft
1.5x,NSE,NSE.
Spl: NSE at any time.
YB-1: ft 3x, ft 3x, ft
3.2x,NSE, Uo.5x.
Another experiment
using a highly cytotoxic
dose of 333 resulted in
markedly different
patterns over time within
approximately the same
ranges of effect.
Protein levels determined
by immunoblotting
assay: (3-catenin: ft to
3.2x on day 2 and to 3.6x
on day 5; (31-integrin: ft
to 2.7x on day 2 and to
4.0xonday5;p-GSK3p
(the inactive form): ft to
2.5x on day 2 and to 2.2x
on day 5.
On day 1, in cells
harvested before
suspension, there was ft
inp-GSK3(3to l.Sxand
NSE for other two
proteins. Consistent
with inorganic arsenic
maintaining the cell's
proliferative potential,
levels of these 3 proteins
decreased much less
rapidly during the 4 days
in suspension if treated
with inorganic arsenic.




Reference



Kaltreider
etal., 1999


















Patterson
et al., 2005



















C-297  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

SIK cells
treated while
being
maintained in
surface
cultures


















B16-F10 cells













Arsenic
Species


As111 SA





















As111 SA
for all













Concentration(s)
Tested (|oM)


2




















0.01,0.1, 1, 10

0.01,0.1, 1, 10


0.01,0.033,0.1










Duration of
Treatment


Various




















4hr

72 hr


7 days










LOECa
(HM)


2




















None

0.01


0.01







Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Protein levels determined
by immunoblotting
assay: nuclear (3-catenin:
ft to 3. 4x on day 9;
cytoplasmic (3-catenin: ft
to 2.5x on day 11; p-(3-
catenin: U to 0.45x on
day 1; pl-integrin: ft to
1.6x on day 7 and to 4.5x
on day 11;
P-GSK3P (the inactive
form): ft to 1.8x on day 7
and to 3. Ix on day 11.
Consistent with
inorganic arsenic
maintaining the cell's
proliferative potential,
inorganic arsenic
decreased the rate of
post-confluent loss of all
of these proteins except
P-p-catenin.
HIF-la protein levels:
No ft seen; no mention if
there were decreases.

Small ft, but decreased to
no change from control
at 0.1, and at higher
doses a
U from control.
Big ft, but decreased to
almost 2 times control at
0.033, and there was no
change from control at
0.1.




Reference


Patterson
et al., 2005





















Kamat et
al., 2005









C-298  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue



B16-F10 cells













J82 cells

HMEC-1 cells

SMC cells



J82 cells

HMEC-1 cells

SMC cells
H22 cells





Arsenic
Species



As111 SA
for all













As111 SA
for all






As111 SA
for all


As111 ATO





Concentration(s)
Tested (|oM)

0.01,0.1, 1, 10

0.01,0.1, 1, 10



0.01,0.1









0.01,0.1

0.01,0.1

0.01,0.1



0.01,0.1

0.01,0.1

0.01,0.1
0.5, 1,2,4





Duration of
Treatment

4hr

72 hr



7 days










7 days
for all






7 days
for all


36 hr





LOECa
(HM)

0.1

0.01



0.01









0.01

0.01

0.01



0.01

0.01

None
0.5


Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
VEGF secretion:
Small ft seen at two
middle doses.

Big ft; much smaller
increase over control at
0.1; no change from
control at 1.0, and a U
from control at 10.
Big ft, but U from control
at 0.1. Also, both 7-day
treatments showed ft
HRE transactivation that
was mostly or
completely blocked by
co-treatment with YC-1 .
HIF- la protein levels:
ft, but U from control at
0.1.

ft, but U from control at
0.1.
Big ft at both dose
levels.
VEGF secretion:
ft, but no change from
control at 0.1.

ft, but no change from
control at 0.1.
Huge ft in VEGF protein
level in cell lysates, with
similar response at all
doses (Westen blot
assay).




Reference



Kamat et
al., 2005














Kamat et
al., 2005






Kamat et
al., 2005

Liu et al.,
2006e


C-299  DRAFT—DO NOT CITE OR QUOTE

-------



Type of
Cell/Tissue

MCF-7 cells















MCF-7 cells










HeLa cells




















Arsenic
Species

As111 ATO














As111 ATO










As111 ATO





















Concentration(s)
Tested (nM)

0.5, 2, 5, 10















2,5










2




















Duration of
Treatment

60 min
















LOECa
(UM)

Various













Results (Compared
With Controls, With
All Concentrations
Being
in |iM Unless Noted)
Results of Western blot
analysis: ft
phosphorylation (i.e.,
activation) of ERK1/2 at
2 with a dose-related
increase to 10; ft
phosphorylation (i.e.,
activation) of p38 at 2
with a dose-related
increase to 10; NSE on
JNK1/2. Thus, inorganic
arsenic activates the
prosurvival mitogen-
activated MEK/ERK
pathway.
The findings in the row above prompted investigation
whether a combination treatment with either of the MEK
inhibitors U0126 (at 10 uM) and PD98059 (at 20 uM)
could lead to enhanced growth inhibition and apoptosis.
They did, augmenting apoptosis approximately 2x
compared to the effects of inorganic arsenic and either
inhibitor alone, based by Hoechst 33258 or annexin V/PI
staining and flow cytometry. Treatment with a p38
inhibitor did not prevent inorganic arsenic-induced
apoptosis; there was a slight but nonsignificant ft in
apoptosis.


Various



















None

















Experiments showed that
NF-KB and AP-1
activation served as
prosurvival or
antiapoptotic forces and
that their activation by
PMA was suppressed by
co-treatment with 10 uM
emodin plus inorganic
arsenic, whereas emodin
or inorganic arsenic
alone had rather little or
no suppressive effect.
The synergistic
suppression was thought
to be caused, at least in
part, by cellular ROS
because pretreatment or
co-treatment with NAC
reduced the effect.




Reference

Yeetal.,
2005














Yeetal.,
2005









Yietal.,
2004
















a Lowest observed effect concentration.
                                          C-300  DRAFT—DO NOT CITE OR QUOTE

-------
                               APPENDIX D. IMMUNOTOXICITY
 1          Arsenic has been observed to affect the immune system.  While changes to the immune
 2    system are not directly related to any specific disease or cancer endpoint, disruptions to the
 3    immune function can impact the individual and likely increase their risk for developing a disease
 4    or cancer outcome. This may, in part, be why there are so many cancers and diseases associated
 5    with arsenic exposure.  In addition, arsenic's effects on the immune response may play some role
 6    in acting as a co-carcinogen with other compounds. Therefore, immunotoxicity is listed as a key
 7    event in this review, and many studies are detailed in the MOA section (4.4.1). The effects,
 8    however, on the immune system are important to note in and of themselves, and a few are
 9    detailed here.
10           Gonsebatt et al. (1994) selected two populations from Comarca Lagunera (Mexico) to
11    study the labeling, mitotic, and replication indexes (LI, MI, and RI, respectively)  of peripheral
12    blood lymphocytes. The exposed population  consisted of 33 individuals from Santa Ana,
13    Coahuila, who had arsenic levels in the drinking water ranging from 375 to 392 ppb (92% in the
14    form of Asv and 8% in the form of As111) for several years.  Approximately 50% of the exposed
15    individuals had cutaneous signs of arsenic poisoning. The control population consisted of 30
16    individuals (selected based on similar proportions of age and sex as the exposed population)
17    from Nazareno, Durango, who had arsenic levels from the preceding 2 years ranging from 19 to
18    26 ppb (>95% as Asv). Urine and blood samples were obtained  from  all subjects. Average total
19    arsenic in the urine and blood of the control group was 36.7 and  37.2 ug/L, respectively, and
20    758.4 and 412.0 ug/L, respectively, in exposed subjects. Peripheral blood lymphocyte  counts
21    were significantly  greater in the exposed population (3.1 x  106 cells) compared to the control
22    population (2.6 x 106  cells), with a greater increase noted in females.  In females, the average
23    36-hour LI was lower in the exposed population than the control population, regardless of the
24    presence or absence of skin lesions (Table D-l below).  Only exposed males with skin lesions,
25    however, exhibited a lower average 36-hour LI; males without skin lesions had an increase in LI.
26    MI were significantly increased in both sexes after a 72-hour incubation period (Table D-l), but
27    were not after 48 or 60 hours of incubation. Exposed females had a significantly  lower RI after
28    48, 60, and 72 hours of incubation, while males were unaffected.
                                               D-1   DRAFT—DO NOT CITE OR QUOTE

-------
     Table D-l. Lymphocyte counts and labeling, mitotic, and replication indexes (mean ± se) in
                the peripheral blood lymphocytes in populations exposed to low (control) and
                high (exposed) levels of arsenic (Gonsebatt et al., 1994)

Lymphocyte
Count (xlO6
cells/ml)
Labeling Index
(36 hours)
with skin lesions
without skin
lesions
Mitotic Index
48 hours
60 hours
72 hours
Replication Index
48 hours
60 hours
72 hours
Control
Males
2.7±1.2
3.32±1.06
1.15±0.23
2.53±0.30
3.52±0.37
1.07±0.01
1.43±0.03
1.93±0.09
Females
2.4±1.1
4.77±1.06
2.52±0.48
4.90±0.79
3.96±0.52
1.16±0.02
1.54±0.04
2.08±0.04
Total
2.6±1.1
3.37±0.61
1.89±0.30
3.85±0.50
3.78±0.34
1.12±0.01
1.49±0.03
2.01±0.04
Exposed
Males
3.0±1.2
2.14±0.86
4.05±1.31
1.59±0.29
3.35±0.39
6.00±0.55b
1.05±0.01
1.39±0.04
1.89±0.09
Females
3.1±0.8a
2.74±0.32
3.90±0.49
1.59±0.26
3.65±0.48
6.60±0.69b
1.10±0.02a
1.37±0.04a
1.84±0.05a
Total
3.1±1.0a
2.42±0.49a
3.95±0.56
1.59±0.20
3.50±0.32
6.34±0.45b
1.08±0.01a
1.38±0.02a
1.86±0.05a
     a Statistically different (p < 0.05) from the control (two-tailed Mann-Whitney U-test).
     b Statistically different (p < 0.001) from the control (two-tailed Mann-Whitney U-test).

 1          A previous study by Ostrosky-Wegman et al. (1991), in which cell-cycle kinetics of
 2   peripheral blood lymphocytes showed a significantly longer average generation time (AGT) for
 3   the highly exposed group as compared to the control group.  The AGT was 19.90 hours in the
 4   low-exposure group compared to 28.70 hours in the high-exposure group. The AGT in the
 5   control group was 19.02 hours. The exposed group consisted of 11 individuals (9 females and 2
 6   males) from Santa Ana, Coahuila, with drinking water containing an average of 390 ppb (98% as
 7   AsV). The control group consisted of 13 individuals (11 females and 2 males) from Nuevo
 8   Leon, Coahuila (drinking water concentrations not  reported). Average urine arsenic/creatinine
 9   levels were 0.121 ug/mL in the control group and 1.565 ug/ml in the exposed group. There were
10   no incidence of skin lesions in the control subjects, but 4 of the 11 exposed subjects had skin
11   lesions (i.e., hyperkeratosis, hypo- and hyperpigmentation, and skin horns).
12          IgG and IgE levels were significantly elevated in arsenic-exposed individuals who
13   presented clinical symptoms of arsenic exposure (i.e., skin lesions) (Islam et al.,  2007). As
14   exposure duration increased, so did the severity of the  skin lesions.  The level of IgE also was
                                               D-2    DRAFT—DO NOT CITE OR QUOTE

-------
 1    greater with longer durations of arsenic exposure. IgG levels were highest during the initial
 2    stages of skin lesions. There was a smaller, but significant, increase in IgA in individuals with
 3    arsenicosis compared to the control group, but no change was observed in IgM levels.  Arsenic-
 4    exposed individuals also had a greater incidence (63% of subjects) of respiratory complications,
 5    such as chest sounds, shortness of breath and breathing complications, irritation of the upper and
 6    lower respiratory tract, cough,  bronchitis, and asthma than the control group (7%). The IgE level
 7    in individuals with respiratory  complications was greater than in arsenic-exposed individuals
 8    without complications. Because the difference in IgE levels could not be explained by
 9    differences in eosinophil levels, it was suggested that the reason may be inflammatory reactions
10    due to arsenic exposure.
11          Yu et al. (1998) found that patients with Bowen's disease (skin carcinoma in situ) from
12    an arsenic-endemic area in the southwest coast of Taiwan had significantly decreased T-cells,
13    increased B-cells, decreased T-helper cells, decreased IFN-y release, decreased TNF-a release,
14    increased IL-2 release, decreased soluble IL-2 receptor release, and changes in soluble CD4 and
15    soluble CDS release (increases in spontaneous release, but decreases in phytohaemagglutinin-
16    induced release) in comparison to normal controls, as well as non-Bowen's patients in the
17    endemic region. Results indicate a depressed cell-mediated immunity in patients with Bowen's
18    disease.  The deficient immune response appears to be related to an impairment of the  membrane
19    IL-2R expression in lymphocytes after stimulation.  This study, however, could not associate
20    arsenic with these changes because individuals without Bowen's disease who lived in the
21    endemic region did not demonstrate the same effects. In addition, a cause and effect relationship
22    could not be determined.  Since arsenic has been demonstrated to affect the immune response in
23    other studies, it is possible that individuals developing Bowen's disease were more susceptible to
24    the effects of arsenic on the immune system.
25          The development of skin lesions is a typical symptom of arsenic-exposed individuals.
26    However, not all individuals exposed,  even those within the same family, develop skin lesions.
27    Mahata et al. (2004)  examined some effects on peripheral lymphocytes in arsenic-exposed
28    individuals with or without skin lesions and compared those results to a group of subjects that
29    were unexposed. Six individuals (3 males and 3 females) were selected from each group:
30    symptomatic (with lesions and exposure), asymptomatic (without lesions and exposure), and
31    unexposed.  Where possible, symptomatic and asymptomatic were selected from the same
32    family. The 6 controls (3  males and 3  females) were selected for similar socio-economic status,
33    age, and gender.  Levels of arsenic in urine, nail, and hair demonstrated that the control group
34    had little exposure to arsenic. Individuals with skin lesions were noted to have less arsenic in
35    their urine and more  in their hair and nails.  This indicated individual differences in distribution
36    and excretion (for more information on this  see Section 4.7.3.1 on genetic polymorphism) that
37    may be related to the individual's susceptibility to developing skin lesions. When the blood of

                                                D-3    DRAFT—DO NOT CITE OR QUOTE

-------
 1    the individuals from all three groups was exposed to further arsenite in vitro, all groups
 2    demonstrated a dose-dependent increase in chromosomal aberrations in the lymphocytes, with a
 3    significant increase observed even at the lowest concentration (1 uM). Untreated lymphocytes,
 4    however, had a greater level of chromosome aberrations in arsenic-exposed individuals. In
 5    addition, individuals with skin lesions had a 1.7-fold increase in "background" chromosomal
 6    aberrations compared to asymptomatic individuals. Although the arsenic-exposed individuals
 7    had more chromosomal aberrations in the absence of additional arsenic exposure in vitro, the in
 8    vitro exposure to arsenite caused a smaller increase in chromosome aberrations in lymphocytes
 9    of exposed individuals compared to unexposed individuals, indicating a greater sensitivity in the
10    control lymphocytes to the in vitro effects of As111.
11          The JAK-STAT pathway is essential in mediating the normal functions of different
12    cytokines in the hematopoietic and immune systems.  In vitro studies by Cheng et al. (2004)
13    suggest that arsenic exposure in the range of 0.4 to 400 uM caused inactivation of the JAK-
14    STAT signaling pathway in HepG2 cells (a human hepatocarcinoma cell line). This inactivation
15    was caused by the direct interaction of arsenic with JAK tyrosine kinase and was independent of
16    arsenic activation of mitogen-activated protein (MAP) kinase. Exposure to sodium arsenite
17    abolished the STAT activity-dependent expression of cytokine signaling suppressors by
18    inhibiting IL-6-inducible STAT3 tyrosine phosphorylation. This effect on the STAT3 tyrosine
19    phosphorylation induced by arsenic was not observed with other inflammatory stimulants, stress
20    agents, or cadmium (metal).
21          Harrison and McCoy (2001) performed an in vitro study to examine the role of apoptosis
22    and enzyme inhibition in arsenic's suppression of the immune response. Cysteine cathepsins are
23    lysosomal  enzymes that are critical in antigen processing. Because of As111 interactions with
24    sulfhydryl  groups, cathepsin L (a member of the papain superfamily of cysteine proteases and a
25    major lysosomal protease involved in cleaving exogenous protein antigens into peptide
26    fragments) was examined as a potential target  for arsenic inhibition. As111 caused a dose-
27    dependent  inhibition of cathepsin L, both as a purified enzyme and in active murine B cells.
28    Inhibition occurred  in TA3 cells even at concentrations that did not affect cell viability; greater
29    inhibition was obtained with the purified enzyme. Addition of DTT caused a complete reversal
30    of the inhibition.  AsV was not able to inhibit cathepsin L, therefore, indicating the involvement
31    of the sulfhydryl groups.  Although cathepsin L was inhibited by 4 hours, exposure for 18 hours
32    led to increases in apoptosis even at the lowest concentration (0.8 uM). Apoptosis was observed
33    at concentrations about 100-fold lower than those inhibiting cathepsin L, suggesting that
34    apoptosis likely plays a more important role in immunosuppression than inhibition of lysosomal
35    cathepsins.
36          De  La Fuente et al. (2002) found a significant increase in apoptosis in PMBCs from
37    healthy donors at concentrations of 15  uM As111 after 48 hours of exposure; an increase also was

                                                D-4   DRAFT—DO NOT CITE OR QUOTE

-------
 1    noted at 5 uM, but did not reach statistical significance. Results did not show a dose-response;
 2    instead apoptosis levels were similar between 15 and 75 uM of arsenite. Lower concentrations
 3    of As111 (i.e., 1 uM and 2.5 uM) also were able to increase apoptosis levels, but required at least
 4    96 hours of exposure compared to only 16 hours of exposure needed with 15 uM of As111.
 5    Measuring the levels of apoptosis in the PMBCs of children chronically exposed to arsenic
 6    (urinary levels of arsenic between 94 and 240 ug/g of creatine) also demonstrated an increase in
 7    apoptosis when compared to the control group. However, exposing the cells of chronically
 8    exposed children to arsenic in vitro demonstrated a decrease in apoptosis when compared to
 9    controls.  Therefore, these data support the data of Mahata et al. (2004), which suggested that
10    control PMBCs are more sensitive to in vitro arsenic exposure.
11           In contrast, Gonzalez-Rangel et al. (2005) found the opposite response. Although their
12    data also show an increase in basal apoptosis in PMBCs lymphocytes and monocytes (but  not
13    natural killer [NK] cells) in an exposed individual compared to six non-exposed individuals, the
14    data also show an increased sensitivity to in vitro arsenite-mediated apoptosis in lymphocytes
15    and NK cells in the exposed individual. This study, however, used a higher concentration  of
16    arsenite (30 uM) compared to the other studies (which used at most 15 uM) and only used one
17    exposed individual compared to 6 unexposed individuals. Therefore, results could be different
18    due to dose or because of inter-individual variation.
19           Although Harrison and McCoy (2001) and De La Fuente et al. (2002) observed an
20    increase in the apoptosis in PMBCs, Chen et al. (2005b) did not observe any effect on the
21    apoptosis of human keratinocytes (obtained from the adult foreskin through routine
22    circumcision) with 1 uM of arsenite.  When cells were exposed to As111  for 24 hours prior to
23    exposure to UVB, however, the As111 protected against UVB-induced apoptosis, indicating a
24    possible mechanism for arsenic's observed co-carcinogenic nature.  Exposing the cells to As111
25    after UVB exposure did not cause a reduction in apoptosis and possibly increased the level of
26    apoptosis.
27           Galicia et al. (2003) isolated PBMC from  healthy, non-smoking, males (22-40 years old)
28    who were not exposed to arsenic to examine the effects of As111 on cell proliferation.  Although
29    there was individual variability, a dose-dependent decreased in cell proliferation in PHA-induced
30    cells was observed.  In all cases, the highest concentration used (1 uM) decreased the percent of
31    dividing cells, with a reduction of 12% to 54%. Although cell proliferation was affected, there
32    was relatively little affect on cell viability. After further examination, it was suggested that
33    proliferation of T lymphocytes was affected and there was a reduction in CD3+ cells producing
34    IL-2. Although arsenic  prevents cells from entering the cell cycle and slows down the
35    progression through the cell cycle, no alteration in the expression of CD69 or CD25 activation
36    molecules was observed. Thus, it was concluded that the reduction in T cell proliferation was
37    caused by a decrease in  the production and secretion of IL-2, thereby blocking the T cells in Gl.

                                                D-5    DRAFT—DO NOT CITE OR QUOTE

-------
 1          Di Gioacchino et al. (2007) examined the effects of arsenic compounds (i.e., As111, AsV,
 2    MMAV, and DMAV) on PBMC proliferation and cytokine release. As111 had the greatest effect
 3    on the cells. At 10-4 M, As111 caused the greatest decrease in PHA-induced cell proliferation and
 4    the greatest reduction in IFN-y and TNF-a release. At 10-4 M, the effect on cell proliferation by
 5    compound was Asin>AsV>DMAv>MMAv. At 10-7 M, however, As111 caused a significant
 6    increase in cell proliferation.  DMAV also caused a significant increase in cell proliferation at 10-
 7    7 M, but had no effect on cell proliferation at 10-4 M.  DMAV and AsV caused a significant
 8    decrease in IFN-y release at 10-4 M, but did not affect TNF-a release. Although the text
 9    indicates that dose-response analyses were performed, the article provides no data. It was
10    concluded that the immunotoxicity of arsenic was dependent on the chemical speciation of
11    arsenic.
12          AsV (0.5, 5, or 50 mg As/L)  administered for 12 weeks via drinking water to female
13    C57BL/6J/Han mice (8-12 weeks old) was determined to decrease NO and superoxide
14    production (Arkusz et al., 2005).  While there was a concentration-dependent decrease in ROS
15    production, the decrease observed in NO was similar across the three doses. The AsV did not
16    appear to affect TNF-a production. It should be noted, that in testing for the NO and superoxide
17    production, 2 x 105 cells/well were plated.  Therefore, a cell to cell comparison was made
18    between the isolated macrophages from the control group and arsenate-treated mice.  The AsV
19    treatment, however, was noted to cause a concentration-dependent increase in the number of
20    peritoneal macrophages isolated.  The percent increase compared to control (55%, 77%, and
21    101%, respectively) was such that it may have compensated for the changes noted in NO and
22    superoxide production.  This, however,  was not tested.
23          Nayak et al. (2007), however, did test the immune response in zebra fish to virus or
24    bacterial infection. Zebra fish embryos (one-cell stage) were exposed to 2 or 10 ppb As111 in
25    water until 4 days post-fertilization.  Seven days later fish were infected or left uninfected.  Viral
26    and bacterial loads were then examined. Viral load was significantly increased in both As111
27    treatment groups compared to the infected control group, with a concentration-dependent
28    increase in the viral load. There also was a significant increase in the bacterial load in treated
29    fish; however, the increase was similar across both treatments.  As111 was also found to decrease
30    ROS burst, interferon, MX mRNA expression, IL-lp, and TNF-a mRNA levels. The maximum
31    response for these cytokines was  also found to be delayed compared to the controls.
                                               D-6   DRAFT—DO NOT CITE OR QUOTE

-------
      APPENDIX E. QUANTITATIVE ISSUES IN THE CANCER RISK ASSESSMENT FOR
                                     INORGANIC ARSENIC
 1          As discussed in Section 5.3, the arsenic cancer risk assessment involved two distinct
 2    phases. The first phase involved the derivation and fitting of dose-response models using the
 3    Taiwanese epidemiological data of Chen et al. (1988a, 1992) and Wu et al. (1989). The outputs
 4    of this phase of the analysis were arsenic dose-response coefficients that described the
 5    relationship between estimated arsenic intake in the Taiwanese population and proportional
 6    increases in age-specific lung and bladder cancer mortality risk.  A key assumption underlying
 7    this relative risk model is that the risk of arsenic-related cancer is a constant multiplicative
 8    function of arsenic dose and the "background" age profile of risks.
 9          The second phase of the risk assessment involved the estimation of arsenic-related cancer
10    risks in a (hypothetical) U.S. population exposed to arsenic at varying levels in drinking water.
11    This phase of the analysis involved the application of the dose-response coefficients for arsenic
12    derived from the Taiwanese data to the age-specific background population risks for the U.S.
13    population.  In addition, the risk estimates were converted from mortality-based values to
14    incidence-based estimates.  The following sections describe each of these phases.
      E.I. CANCER RISK ASSESSMENT FOR THE TAIWANESE POPULATION
15          The calculation of cancer risks from the Taiwanese epidemiological  data was performed
16    using Excel workbook files.  The files contained the input data for the dose-response models and
17    spreadsheets to accept user-specified inputs, perform calculations, and summarize outputs from
18    the assessment. Input data included male and female lung and bladder cancer mortality and
19    person-years at risk (PYR) data for arsenic-exposed populations from 42 villages obtained from
20    Morales et al. (2000), village water arsenic concentrations (minimum, median, and maximum
21    data sets), and southwest  Taiwan and all Taiwan reference population mortality and PYR data.
22          The user first specifies drinking water consumption and body weights for the Taiwanese
23    population in the "Poisson Model" page of the risk calculation files. Solver® is then invoked to
24    estimate the age coefficients (al, a2, and a3) and the arsenic dose-response coefficient (b) in
25    equation E-l by maximizing the likelihood function that is coded into the spreadsheets.  Solver is
26    then reconfigured to calculate the upper confidence limit (UCL) on "b" using the profile
27    likelihood method (see below).  The resulting UCL value is then input to the "BEIR Model"
28    sheet and the LEDM for cancer incidence is calculated, again using Solver®.  The LED0i value is
29    transferred to the "Summary" sheet, where other risk metrics (unit risk, cancer risks at different
30    drinking water concentrations, and the drinking water concentration corresponding to 10-4
31    lifetime risk) are calculated.  Risk metrics are calculated based on user-specified drinking water
32    intake and body weight for the U.S. population. Likelihood calculations for most of the


                                               E-1    DRAFT—DO NOT CITE OR QUOTE

-------
 1    endpoints were replicated using a different optimizer in the Non-Linear Estimation module of the
 2    Statistica® software package.
      E.2. MLE ESTIMATION OF DOSE-RESPONSE PARAMETERS
 3          The Taiwan risk model spreadsheets calculate the dose-response parameters for the
 4    Poisson model, fitting separate models for each endpoint:
 5
 6                       h(x,t) = exp(ai + a2 x age + a3 x age2) x (l + b x dose)      (Equation E-l)
 7
 8          In this model, the midpoints of the age group strata are normalized (placed on a "z-
 9    scale") before risk is  estimated; age is thus treated as a "nuisance parameter" in the model.
10    Dose, as noted above, is calculated from dietary arsenic and village water concentrations and is
11    expressed in terms of mg/kg-day. Each age-dose group is represented by a row on the
12    spreadsheet.  There are 42 villages with arsenic well water data and the reference population,
13    each divided  into 13 age strata, for a total of 559 population groups.  The model begins with
14    randomly selected values for the four parameters and then calculates the Poisson log likelihood
15    values for each group:
16
17                     log likelihood = observed x  ln[hcuRRENi(x,t)] - predicted     (Equation E-2)
18
19    where:
20          observed =          the number of cancer deaths in groups age t, exposed at dose x
21          hcuRRENx(x,t)  =      the estimated total cancer risk in age group t at dose x, based on
22                              the current parameter estimates
23          predicted =          the predicted number of cancer deaths in age group t at dose x, =
24                              hCuRRENx(x,t) x PYR, where PYR = person-years at risk
25
26                 The sum of the log likelihood across all the age groups is then maximized using
27    standard optimization methods (Excel Solver®)  to provide the MLE estimates of the age and
28    dose parameters.
29    E.2.1.  Estimation of Upper Confidence Limits on the Arsenic Dose-Response Parameters
30          ED01 values are derived based on the MLE dose-response parameter estimates. The
31    LEDoi estimates are derived from the 95% upper  confidence limits (UCLs) on the dose-response
32    parameters.  The UCLs on the dose-response "b" parameters were estimated using the "profile
33    likelihood" method (Venson and Moolgavkar, 1988).  In this approach, the value of the dose
34    parameter, b, was varied from its estimated mean value, and the changes in log-likelihood were
35    calculated. The ratio of the log likelihood for the best-fit model to the log likelihood for other
36    values of "b" is known to follow an approximate chi-squared distribution with one degree of
37    freedom.  Thus, the 5th and 95th confidence limits on the dose coefficient "b" correspond to the
38    values where the likelihood ratio is equal to 1.92. Upper and lower confidence limits were

                                               E-2   DRAFT—DO NOT CITE OR QUOTE

-------
 1    calculated using Solver®.  The fact that the profile likelihood method ignores the likelihood
 2    impact of the age "nuisance parameters" implies that the calculated confidence limits are only
 3    approximate. Confidence limit calculations using other methods (empirical Bayesian simulation
 4    and "bootstrap-t") gave similar values (within a few percent).

      E.3. ESTIMATION OF RISK FOR U.S. POPULATIONS EXPOSED TO ARSENIC IN
         DRINKING WATER
 5          LEDoi values were calculated using a life-table method that is a variation on the "BEIR
 6    IV" model recommended by NRC (2001). Specifically, the approach includes a modification
 7    suggested by Gail et al. (1999) for obtaining more accurate estimates of incidence within multi-
 8    year age strata.  The BEIR IV relative risk model takes as its inputs the arsenic dose-response
 9    "b" coefficient from the Poisson model, background cancer incidence data, along with age-
10    specific mortality data to directly estimate lifetime bladder and lung cancer incidence for the
1 1    target (U. S.  adult) population. Lung and bladder cancer incidence reference data for the years
12    2000-2003 were obtained from the National Cancer Institute's SEER program (NCI, 2006).
13    U.S. gender and age-specific population data and all-causes mortality data came from the
14    National Center for Health Statistics (NCHS, 2000).
15          Formulas for calculating LEDOT values were implemented on separate Excel spreadsheets
16    for each endpoint.  The following calculations were implemented in separate lines on each
17    spreadsheet. In all of the equations, the subscript "i" refers to age group:
18
                     L(x) = lifetime  risk of cancer  incidence at dose x
19                      i                                                      (Equation E-3)
20
21    Numerator Terms:
22
23    q(x) = cancer incidence hazard at dose (x), age interval (i)
24
25    c;(x) = Ci(0) x (1 + beta x dose) (beta comes from the linear Poisson model)
26
27    c;(0) = background cancer incidence; / cancer free population;
28
29    Background cancer incidence c ; comes from SEER, cancer-free population;, see (7)
30
31    b; = background cancer incidence; / alive population; (SEER data)
32
33    Y; = exp (- 5b;)
34
35    F; = initial estimated background probability of survival without cancer to the
36    end of interval (i - 1)
                                               E-3   DRAFT—DO NOT CITE OR QUOTE

-------
                                                                                 (Equation E-4)
 2
 3    G; = initial estimated background probability of survival without cancer to
 4    the middle of interval (i)
 5
                                          f^-    rr-(-2.5*bi)
 6                                        CzZ = PI                              (Equation E-5)
 1
 8    Cancer-free population; = alive population; x  G;
 9
10    Denominator Terms:
11
12    s;(x) = total noncancer mortality and cancer incidence hazard, at dose (x)
13    in age interval (i)
14
15    s;(x) = background noncancer mortality (x, i) + cancer incidence hazard (x, i)
16
17    s;(x) = (d; - h;) + Ci(0) x (1 + beta x dose)
18
19    d; = total mortality (background) in age interval (i)
20
21    d; = total deaths; / population;  (Census, Vital Stat. U.S.)
22
23    h; = cancer deaths; / population; (Census, Vital Stat. U.S.)
24
25    Survival (Ti and n) Estimation:
26
27    T; = probability of survival without cancer to end of interval (i - 1)
28
                                      ;'=!-!    ;=2-i     ;=2-i
                                  Ti =  Y\ri= [] Wr; * [] Wib
29                                    >'=1     >=1     >=1                          (Equation E-6)
30
31    r; = probability of survival cancer free through interval (i), given survival to beginning of
32    interval  (i)
33
34    n = Wi x Wib
35
36    W; = exp (-5d; + 5h; - 5c;)
37
38    Wib = exp (-5c; x Beta x x)
39
40          To calculate ED0i values, the value of the daily arsenic dose used to calculate h;(x), and
41    hence L(x), was varied until L(x) = 0.01 (1%). For the MLE estimation, Solver was used to
42    estimate LED0i values in the model spreadsheets.
                                                 E-4   DRAFT—DO NOT CITE OR QUOTE

-------
          APPENDIX F. RISK ASSESSMENT FOR TOWNSHIPS AND LOW-EXPOSURE
                                  TAIWANESE POPULATIONS

      F.I. RECENT STUDIES OF THE TAIWANESE POPULATIONS THAT DO NOT FIND
          CONSISTENT EXPOSURE-RESPONSE RELATIONSHIPS
 1          As discussed in Section 5.3.8.5, several recently published studies have called into
 2    question the strength and significance of the exposure-response relationship for arsenic in the
 3    Taiwanese population studies (Chen et al., 1988a, 1992; Wu et al., 1989). This appendix
 4    provides a brief analysis of some of these concerns.
 5          Based on "graphical and regression analysis," Lamm et al. (2003) found no significant
 6    dose-response relationship for arsenic-related bladder cancer in the subset of the Taiwanese
 7    population with median drinking water well concentrations less than 400 ppb (ug/L).
 8    Significant, positive dose-response slopes were found for villages with median well
 9    concentrations above 400 ppb. They also observed that all of the villages "solely dependent" on
10    artesian wells had median arsenic concentrations above 350 ppb, and that the median well
11    concentrations in villages not solely dependent on artesian wells were generally below this
12    value. Based on these observations, Lamm et al. (2003) suggested that the nature of the villages'
13    water sources (artesian vs. non-artesian), rather than arsenic concentration, explained the
14    observed variations in bladder cancer risk in the Taiwanese population.
15          Kayajanian (2003) also argued that EPA is misinterpreting the data from the Taiwanese
16    population.  Kayajanian stratified median well arsenic concentration into 10 ranges from 10 to
17    934 ppb.  The author then calculated combined mortality rates for lung,  bladder,  and liver cancer
18    for each stratum of the population.  They calculated that crude (age-unadjusted) cancer mortality
19    for both males and females was significantly elevated in the lowest exposure groups, decreased
20    to minimums for villages with water arsenic concentrations between 42  and 60 ppb, and then
21    again increased with increasing arsenic exposure. They argued on this basis (and based on the
22    analysis of cancer mortality data from another epidemiological study) that health standards for
23    arsenic should be set in the vicinity of 50 ug/L (ppb) in order to minimize the risk of arsenic-
24    associated cancer, and that lower exposures would actually result in increased risk in the U.S.
25    population.
26          In a more recent study, Lamm et al. (2006) reported additional analyses of the
27    relationship between cancer risks and drinking water arsenic in the same Taiwanese population.
28    In this analysis, they divided the epidemiological  data according to six "township" designations
29    provided by the original Chinese investigators (townships 0, 2, 3, 4, 5, and 6).1  They stratified
30    the data into two groups: townships that (by their characterization) "showed a significant dose-
31    response relationship" with arsenic (2, 4, 6) and townships "that did not" (0, 3, and 5).  They
      1 Each township included subsets of the 42 "villages" used as the basic units of analysis in the current assessment.
                                                F-1   DRAFT—DO NOT CITE OR QUOTE

-------
 1    then applied linear regression to characterize the relationship between combined bladder and
 2    lung cancer standardized mortality ratios (SMRs) and arsenic exposures in the Taiwanese
 3    villages.  They found that (1) dummy variables related to township were significant (along with
 4    arsenic well  concentration) when all the townships were included in the analysis, and (2) the
 5    dose-response parameter for arsenic exposure became insignificant for arsenic well
 6    concentrations less than 151 ppb when only townships 2, 4, and 6 were included in the
 7    regression.  Based on these results, they concluded that location (township) was an important
 8    explanatory variable for cancer risks and that 151 ppb represented a "threshold" well arsenic
 9    concentration below which no exposure-response relationship for arsenic could be detected.
      F.2. LIMITATIONS OF THE RECENT STUDIES
10           The studies discussed above all have significant limitations, relating both to the methods
11    used to select or stratify data for the risk assessment and to the methods used in analyzing
12    exposure-response data.  In the first place, it is important to recognize the complexity and
13    limitations of the data. Cancer mortality and person-years at risk observations are provided for a
14    large number (559) of relatively small age- and village-stratified populations (median person-
15    years at risk ~ 340 for both males and females). Most population groups have zero cancer
16    deaths, and the data are very "noisy." Cancer mortality is strongly age-dependent, and
17    simultaneously evaluating the age-and dose-dependence of cancer mortality based on a data set
18    in which cancer deaths are "rare events" requires appropriately structured models. All of these
19    features of the data drove the selection of the Poisson regression methods described in Section 5,
20    and the use of simpler models (linear regression, for example) can (and did) lead to misleading
21    results.
22           With regard to the Lamm et al. (2003) paper,  it is likely that the use of linear regression
23    and the failure to account for the age-dependency of bladder cancer risks combined to make it
24    impossible to detect a significant exposure-response relationship in villages with water arsenic
25    levels less than 400 ppb. In addition, it should be noted that Lamm et al. (2003) did not have
26    data regarding the actual sources of drinking water in the various villages;  instead they relied on
27    the arsenic concentration to assess the degree of dependency of specific villages on artesian
28    (generally high-arsenic) versus shallow (low-arsenic) wells.  When defined in this circular
29    fashion, it is inevitable that including the degree of "artesian well dependence" in a multiple
30    regression along with arsenic concentration would deprive the latter variable of much of its
31    explanatory power and statistical significance. Finally, the rationale for excluding valid data on
32    southwestern or all-Taiwan reference populations from the analysis is highly questionable, and
33    again lowers the likelihood of detecting significant exposure-response relationships.
34           The major limitation of Kayajanian's (2003) analysis of the Taiwanese data is that it
35    breaks the data into strata that are too small to be used to calculate reliable mortality risks, and
36    that it is very sensitive to the specific way that the data are stratified. The relatively high cancer
                                                 F-2   DRAFT—DO NOT CITE OR QUOTE

-------
 1    mortality risks seen in the low-dose strata are associated with a small number of villages that
 2    happen to have a (relatively) large number of deaths.  The observed trend in cancer mortality
 3    versus arsenic dose would be very different if only few cancer deaths were misclassified, or if
 4    the pattern of cancer deaths had been slightly different by chance.  Again, failure to use a model
 5    that adequately addresses the distribution of cancer deaths as rare events (or that incorporates
 6    age dependence) resulted in results that are misleading.
 7          Lamm et al.'s (2006) failure to find a significant exposure-response relationship in
 8    villages with arsenic water concentrations below 151  ppb can also be explained by (1) the use of
 9    linear regression without age-adjustment and (2) the omission of data from three of the six
10    townships from  some of the regressions. Lamm et al. (2006) did not explain the specific criteria
11    for determining  if a township "showed a dose-response relationship," but based on the
12    description of their methods provided in the article, it may be assumed that they used linear
13    regression to characterize the relationship between SMRs and arsenic exposure in each village in
14    the various townships. Given the small number of villages in each township, this approach and
15    the rationale for leaving townships 0, 3, and 5 out of the analysis appear arbitrary and
16    unjustified.    In the following sections, we present alternative analyses that further investigate
17    the nature of arsenic exposure-response relationships  in the various townships and in villages
18    with low arsenic drinking water concentrations.
      F.3. CALCULATIONS OF RISKS FOR TOWNSHIP GROUPS
19          To address the issues raised by Lamm et al. (2003, 2006), EPA compared the patterns of
20    cancer risks for  subjects in the two groups of townships (0, 3, and 5 vs. 2, 4, and 6) to see
21    whether there were any differences. As noted above,  it is not believed that Lamm et al.'s
22    approach to omitting townships because they lack an internal dose-response relationship is valid,
23    so EPA did not do so.
24          First, to get a rough idea of the patterns in  cancer incidence versus exposure, the crude
25    cancer risks (population-weighted deaths per person-year for all age groups) and population-
26    weighted average arsenic exposure concentrations were calculated for each of the six villages.
27    The results are shown in Figure F-l.  This figure simply illustrates that, even without age-
28    adjustments, arsenic dose-response relationships across the villages are quite robust.
                                                F-3    DRAFT—DO NOT CITE OR QUOTE

-------
      2.E-03
"re
£     2.E-03
"re  -o
r-2  1.E-03
            o> c
            o =  5.E-04
            re
           O
                            -*- Villages 0,3,5
                            -•-Villages 2,4,6
                            A Reference Population
                 O.E+00
             t
                         0       100      200     300      400     500     600
                      Population-Weighted Drinking Water As, micrograms/L
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
              Figure F-l. Lifetime crude total cancer risk (male + female) for the
                         low- and high-exposure villages
       For both sets of villages (low- and high-exposure), crude cancer risks increase with
average arsenic drinking water concentration. Age distributions were very similar in all cohorts,
so the lack of age-adjustment did not seriously bias the results. While total cancer risks are
dominated by male lung cancer, the other endpoints showed generally the same pattern. This
finding suggests that the positive exposure-response relationship for arsenic is not being
seriously confounded by a "village effect."  Given the small populations, populations at risk, and
numbers of cancer deaths in the individual villages, it is not clear that analyzing exposure-
response relationships within these villages (as defined by Lamm et al.) is justified.
       Exposure-response relationships in the various townships were also investigated using a
variant of the multiple regression method applied by Lamm et al. (2006). In this analysis,
however, the non-linear relationship between cancer risk and age was explicitly recognized, and
the analysis was conducted for township both "with" and "without" significant exposure-
response relationships by Lamm et al.'s definition.  First, male and female combined cancer
mortality risks (bladder + lung) were regressed against the same non-linear age dependency
incorporated into the Poisson model shown in Equation 5-2.  That is, the following equation was
fit to both the male and female cancer data from the various age groups in the low- and high-
exposure villages:
              risk (age) = exp(ai + a2 x age +
                                                          age2)
       Then, the residuals from these regressions (the cancer risks with the effect of age
removed) were regressed against estimated arsenic dose levels. The dose levels were calculated
                                          F-4   DRAFT—DO NOT CITE OR QUOTE

-------
 1    assuming a nonwater arsenic intake of 10 |J,g/day for exposed and reference populations, which
 2    is consistent with the assumptions outlined in Section 5.3.5.  The regressions were population-
 3    (person-year) weighted, in effect giving a linear regression of age-adjusted cancer risks versus
 4    arsenic dose. The results are shown in Table F-l.
      Table F-l. Coefficients from linear regressions of age-adjusted cancer risk versus arsenic
                 doses for townships identified by Lamm et al. (2006)
Township Numbers
Reference
Population"
Male arsenic dose
coefficient (p-value)
Female arsenic dose
(p-value)
All Townships
Included
0.035
(0.043)
0.12
(0.0002)
Excluded
0.032
(0.068)
0.12
(0.0004)
Townships 2, 4, and 6
Included
0.092
(0.0002)
0.11
(0.0001)
Excluded
0.091
(0.001)
0.12
(0.0001)
Townships 0, 3, and 5
Included
-0.0093
(0.787)
0.14(0.015)
Excluded
-0.002
(0.487)
0.13 (0.026)
       aSouthwest Taiwan.
 5           The estimated dose coefficients for age-adjusted women's cancer risk (the linear "slope"
 6    of the relationship between cancer mortality, with the effect of age removed, and arsenic dose2)
 7    are positive and statistically significant for all combinations of townships.  Coefficients for male
 8    age-adjusted  cancer risk are positive and significant when all townships are included (although
 9    marginally significant when the reference population is excluded). Similarly, age-adjusted male
10    cancer risk coefficients are positive and highly significant for townships 2, 4, and 6, with or
11    without the reference population. In contrast, the arsenic dose-response coefficients for male
12    age-adjusted  cancer risks are negative, but very small and not significant, for townships 0, 3, and
13    5.
14           This analysis illustrates that, even using the less-desirable linear regression approach,
15    when the cancer risk for the genders separated, and with proper age adjustment, female arsenic
16    dose-response relationships are robust and significant for both village groups. For males, the
17    arsenic dose-response relationships are significant when a reference population is included,
18    except for townships 0, 3, and 5. As noted above, the rationale for analyzing groups of
19    townships separately is questionable, as is the omission of a reference population.  The results
20    showing apparently insignificant associations between male cancer risks and arsenic exposure
       This approach is not particularly desirable from the standpoint of finding the best fit to the data because it restricts
      the effect of arsenic on cancer risk to being linear, and assumes that regression residuals are normally distributed,
      which is unlikely to be true. This approach has been used to illustrate that even using simple models, positive dose-
      response relationships can be detected in the data. Due to the different form of this model, the slope coefficients
      derived in this section are also not comparable to those shown in Tables 5-3 and 5-4.
                                                  F-5    DRAFT—DO NOT CITE OR QUOTE

-------
 1    more than anything reflect the limitations of this less-than-optimal approach to risk modeling for
 2    these data.
      F.4. CALCULATION OF ARSENIC-RELATED CANCER RISKS FOR LOW-
          EXPOSURE VILLAGES
 3          Rather than stratify the Taiwanese population by township, a better way to test the
 4    significance of exposure-response relationships at low doses is to simply restrict the analysis to
 5    the villages with low arsenic water concentrations, but use the appropriate Poisson regression
 6    methodology. In the analysis summarized in Table F-2, the Poisson model shown in Equation 5-
 7    2 was fit to data from the approximately one-half of subject groups with median arsenic drinking
 8    water concentrations less than 150 ppb.  Lamm et al. (2006) considered this concentration to be a
 9    natural breakpoint because the median arsenic concentrations in the Wu et al. (1989) and Chen et
10    al. (1992) population cluster into two groups,  one  group with 10-126 ppb and the other with
11    256-934 ppb. Arsenic "b" coefficients (the dose coefficients in the Poisson model) were
12    estimated separately for lung and bladder cancer and for both endpoints combined, for men and
13    women.
      Table F-2. Arsenic dose coefficients for study populations with median well water arsenic
                concentrations less than 127 ppb
Endpoint
Male lung
Male bladder
Male combined
Female lung
Female bladder
Female combined
Arsenic "b" Coefficient
(95% UCL, LCL)
85.7(13.1, 172.1)
586 (335, 877)
160 (83.4, 247)
615 (412, 836)
2639 (2021, 3307)
924(721, 1139)
14
15          For all of the endpoints, the arsenic dose coefficients are positive with lower confidence
16    limits that are also positive.3 This finding indicates that for population groups with water arsenic
17    concentrations less than or equal to 126 ppb, the dose-response relationships are positive and
18    statistically significant.
19          On the whole, the analyses presented in this section provide support for statistically
20    significant dose-response relationships for arsenic-related cancer, even in the population groups
21    with relatively low exposures. When the data are artificially stratified, when no reference
22    population is included, and when inappropriate statistical models are employed, it is possible to
       As in Section 5.3.8, the upper and lower confidence limits were calculated using profile likelihood; similar results are
      obtained using bootstrap methods.
                                                F-6   DRAFT—DO NOT CITE OR QUOTE

-------
1   find insignificant or negative dose-response relationships for arsenic for some portions of the
2   data. When appropriate models are used, however, the Taiwanese data show robust and
3   significant positive associations between arsenic exposures and cancer risks for all of the
4   endpoints analyzed, even in low-exposure groups. No evidence was found that either 400 ppb or
5   150 ppb represent "threshold" arsenic concentrations in drinking water below which cancer risks
6   are not increased. Likewise, the analyses do not support the existence of a "village effect"
7   related to the degree of dependence on artesian versus shallow wells.
                                               F-7   DRAFT—DO NOT CITE OR QUOTE

-------