&EPA
TOXICOLOGICAL REVIEW
OF
INGESTED INORGANIC ARSENIC
(CAS No. 7440-38-2)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
July 2005
NOTICE
This document is an IRIS review draft for the Science Advisory Board (SAB). It has not
been formally released by the U.S. Environmental Protection Agency and should not at this
stage be construed to represent Agency position on this chemical. It is being circulated for
review of its technical accuracy and science policy implications.
U.S. Environmental Protection Agency
Washington DC
DRAFT - DO NOT CITE OR QUOTE
-------
DISCLAIMER
This document is a preliminary draft for review purposes only and does not constitute U.S.
Environmental Protection Agency policy. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.
DRAFP - DO NOT CITE OR QUOTE
-------
TABLE OF CONTENTS - TOXICOLOGICAL REVIEW for INORGANIC ARSENIC
(CAS No. 7440-38-2)
FOREWORD viii
AUTHORS, CONTRIBUTORS, AND REVIEWERS viii
1. INTRODUCTION 1
2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO
ASSESSMENTS 3
3. TOXICOKINETICS :. .- 5
3.1. ABSORPTION 5
3.2. DISTRIBUTION 6
3.3. METABOLISM 10
3.4. .ELIMINATION 14
3.5. PHYSIOLOGICALLY BASED TOXICOKJNETIC MODELS 15
4. HAZARD IDENTIFICATION 17
4.1. STUDIES IN HUMANS 17
4.2. PRECHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS
IN ANIMALS - ORAL AND INHALATION 37
4.2.1. Prechronic and Chronic Studies 37
4.2.2. Cancer Bioassays 38
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES - ORAL AND
INHALATION 41
4.3.1. Oral , 41
4.4. OTHER STUDIES 42
4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS51
4.5.1. Oral 51
4.5.2. Inhalation 51
4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER
CHARACTERIZATION 51
4.6.1. Summary of Overall Weight-of-Evidence 51
4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence 51
4.6.3. Mode of Action Information 53
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 56
4.7.1. Possible Childhood Susceptibility 56
4.7.2. Possible Gender Differences 58
4.7.3. Other 59
5. DOSE-RESPONSE ASSESSMENTS 59
DRAFT - DO NOT CITE OR QUOTE iii
-------
5.1. ORAL REFERENCE DOSE (RfD) 59
5.1.1. Choice of Principal Study and Critical Effect 59
5.1.2. Methods of Analysis 59
5.1.3. RfD Derivation 59
5.1.4. Previous Oral Assessment 60
5^ INHALATION RLTCRENCE DOSE (RfC) 60
5.3. CANCER ASSESSMENT (Oral Exposure) 60
5.3.1. Choice of Study / Data - with Rationale and Justification .60
5.3.2. Dose-response Data 61
5.3.3. Dose Conversion 62
5.3.4. Extrapolation Method(s) ' 62
5.3.5. Oral Slope Factor 64
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND
DOSE RESPONSE 67
6.1. HUMAN HAZARD POTENTIAL 67
6.2. DOSE RESPONSE 68
7. REFERENCES 69
APPENDIX A. Summary of External Peer Review and Public Comments and Disposition
99
APPENDIX B 100
DRAFT - DO NOT CITE OR QUOTE iv
-------
List of Tables
Table 2-1. Chemical and Physical Properties of Arsenic and Selected Inorganic Arsenic
Compounds ' 3
Table 4-1. SMRs in Taiwan by Sex 19
Table 4-2. Age-Adjusted Mortality Rates (per 100,000) in Taiwan 20
Table 4-3. Model Parameters and Potency Index ; 21
Table 4-4. Estimated values for Chen et al. (1992) 22
Table 4-5. Statistically Significant (p<0.05) SMRs for Cancers in a Blackfoot Disease
Endemic Area of Taiwan Compared to Local and National Controls 24
Table 4-6. Risk Statistics of Best Fitting Models from Morales et al. 2000 26
Table 4-7. SMRs in Japanese Cohort : . . 29
Table 4-8. NRC 2001 Maximum Likelihood Estimates of Excess Lifetime Risk 37
Table 5-4. Calculation for Male Lung Cancer Risk from Arsenic in Drinking Water 66
Table 5-5. Calculation for Female Lung Cancer Risk from Arsenic in Drinking Water ... 66
Table 5-6. Calculation for Male Bladder Cancer Risk from Arsenic in Drinking Water ... 67
Table 5-7. Calculation for Female Bladder Cancer Risk from Arsenic in Drinking Water . 67
Table B-l. Poisson modeling options. (Modified from Morales et al., 2000) 100
DRAFT - DO NOT CITE OR QUOTE
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and
rationale for the hazard and dose-response assessment in IRIS pertaining to chronic exposure
to inorganic arsenic. It is not intended to be a comprehensive treatise on the chemical or
toxicological nature of inorganic arsenic.
In Section 6, EPA has characterized its overall confidence in the quantitative and
qualitative aspects of hazard and dose response. Matters considered in this characterization
include knowledge gaps, uncertainties, quality of data, and scientific controversies. This
characterization is presented in an effort to make apparent the limitations of the assessment
and to aid and guide the risk assessor in the ensuing steps of the risk assessment process.
For other general information about this assessment or other questions relating to
IRIS, the reader is referred to EPA's IRIS Hotline at 301-345-2870.
AUTHORS, CONTRIBUTORS, AND REVIEWERS
Chemical Manager/Author
Elizabeth A. Doyle, Ph.D.
Office of Water
U.S. Environmental Protection Agency
Washington, DC
Robyn B. Blain, Ph.D.
Gregory M. Blumcnthal, Ph.D.
Welford C. Roberts, Ph.D.
ICF Consulting
Fairfax, VA
Reviewers
This document and summary information on IRIS have received peer review both by
EPA scientists and by independent scientists external to EPA. Subsequent to external review
and incorporation of comments, this assessment has undergone an Agency-wide review
process whereby the IRIS Program Director has achieved a consensus approval among the
Office of Research and Development; Office of Air and Radiation; Office of Prevention,
Pesticides, and Toxic Substances; Office of Solid Waste and Emergency Response; Office of
Water; Office of Policy, Economics, and Innovation; Office of Children's Health Protection;
Office of Environmental Information; and the Regional Offices.
DRAFT - DO NOT CITE OR QUOTE vi
-------
List of Abbreviations
8-OHdG
AIC
ATSDR
As"1
Asv
BFD
BMD
BMD.o
BMDL
BMDS
BMI
BrdU
CA
CASRN
CCR
CHO
CI
COPD
DAAC
DHLP
DMA1"
DMAV
DMA
DMPS
DMSO
DTT
ED
EKG
eNOS
EPA
GAPDH
GI
GM-CSF
GSH
GSSG
HSDB
IGF-1
IGFBP-1
IRIS
ISHD
8-hydroxydeoxyguanosine
Akaike Information Criterion
Agency for Toxic Substances and Disease Registry
arsenite
arsenate
blackfoot disease
benchmark dose
benchmark dose at 10% effect
lower 95% confidence limit on the benchmark dose
benchmark dose software
body mass index
bromodeoxyuridine
chromosome aberrations
CAS registry number
chromale copper arsenate
Chinese hamster ovary
confidence interval
chronic obstructive pulmonary disease
diseases of the arteries, arterioles and capillaries
dihydrolipoic acid
dimethylarsenous acid
dimethyl arsinic acid
dimethyl arsenic-used when the oxidative slate is unknown or not specified
2,3-dimercaptopropane-l-sulfonicacid
dimethyl sulfoxide
dithiothreitol
effective dose
electrocardiogram
endothelial nitric oxide synthase
Environmental Protection Agency
glyceraldehyde-3-phosphate dehydrogenase
gastrointestinal
granulocyte macrophage-colony stimulating factor
glutathione
oxidized glutathione
Hazardous Substance Data Base
insulin-like growth factor 1
IGF-1 binding protein
Integrated Risk Information System
ischemic heart disease
DRAFT - DO NOT CITE OR QUOTE
vn
-------
JNCV Joint National Committee on Detection, Evaluation, and Treatment of High
Blood Pressure
LD50 lethal dose 50%
LOAEL lowest-observed-adverse-effect-level
MH-PR Mantel Haenzel-weighted prevalence ratios
MMA1" monomethylarsonous acid
MMA'"0 methylarsine oxide
MMAV monomethyl arsonic acid
MMA monomethyl arsenic-used when oxidative state is unknown or not specified
MN micronuclei
MNU jV-methyl-./V-nitrosourea
NCI National Cancer Institute
NO nitric oxide
NOAEL no-observed-adverse-effect-level
NRC National Research Council
ODC ornithine decarboxylase
PBPK model physiologically based pharmacokinetic model
PKG phosphoglycerate kinase
PNP purine nucleoside phosphorylase
RBCs red blood cells
RfD oral reference dose
RFC inhalation reference concentration
ROS reactive oxygen species
RR relative risk
SAM S-adenosylmethionine
SCE sister chromatid exchange
SMR standard mortality ratio
SOD superoxide radical dismutase
TGF-a transforming.growth factor-alpha
TMAO trimethylarsirie oxide
TNF-cc tumor necrosis factor-alpha
TPA 12-O-tetradecanoyl phorbol-13-acetate
USGS U.S. Geological Survey
UVR ultraviolet radiation
DRAFT - DO NOT CITE OR QUOTE
Vlll
-------
1. INTRODUCTION
This assessment is based on the EPA-sponsored reviews "Arsenic in Drinking
Water" and "Arsenic in Drinking Water, 2001 Update" published by the National Research
Council in 1999 and 2001, respectively. The NRC arsenic committee took into consideration
presentations at the committee's public meetings, submitted public comments, and the
comments made by technical experts on the draft NRC arsenic reports. The conclusions,
recommendations and final content of the NRC (1999, 2001) reports rest entirely with the
committee and the National Research Council. This IRIS document has undergone review by
EPA health scientists from several program offices and regional offices.
The goal of this IRIS assessment is to provide a summary, in the format needed for
the IRIS system, of the key science analysis and recommendations for risk assessment
methodology developed by the NRC. This assessment also incorporates EPA's review of
literature on arsenic's health effects published after the publication of the NRC 2001 update.
This new information is incorporated into the assessment text as appropriate, but did not
indicate a need for any fundamental changes vis-a-vis the recommendations developed by the
NRC. The quantitative cancer risk assessment in this assessment tracks NRC
recommendations on cancer dose response modeling and risk assessment. This assessment
does include additional sensitivity analysis on effects of modeling assumptions on estimated
cancer risk. The NRC text recommends such sensitivity analysis. The recommended cancer
unit risk estimate in this IRIS assessment is based on the NRC's preferred approach (as
presented in the Summary of the 2001 Update). The NRC did not develop quantitative risk
assessments for noncancer health effects of arsenic. By comparison this IRIS assessment
does develop a reference dose (RfD) to address noncancer health effects of arsenic.
This document presents background and justification for the hazard and dose-
response assessment summaries in EPA's Integrated Risk Information System (IRIS). IRIS
Summaries may include an oral reference dose (RfD), inhalation reference concentration
(RfC), and a carcinogenicity assessment. The majority of arsenic exposures discussed in this
Toxicology Review are from oral waterbome arsenic exposure. Although inhalation
exposure is already in IRIS, it will not be addressed or revised in this report.
The RfD provides quantitative information for noncancer dose-response assessments.
It is based on the assumption that thresholds exist for certain toxic effects such as cellular
necrosis but may not exist for other toxic effects such as some carcinogenic responses. It is
expressed in units of mg/kg-day. In general, the RfD is an estimate (with uncertainty
spanning perhaps an order of magnitude) of a daily exposure to the human population
(including sensitive subgroups) that is likely to be without an appreciable risk of deleterious
noncancer effects during a lifetime.
The carcinogenicity assessment in this report provides information on the
carcinogenic hazard potential of the substance in question and quantitative estimates of risk
DRAFT - DO NOT CITE OR QUOTE
-------
from oral exposure. The information includes a weight-of-evidence judgment of the
likelihood that the agent is a human carcinogen and the conditions under which the
carcinogenic effects may be expressed. Quantitative risk estimates are presented in three
ways. The slope factor is the result of application of a low-dose extrapolation procedure and
is presented as the risk per mg/kg/day. The unit risk is the quantitative estimate in terms of
either risk per ug/L drinking water or risk per ug/m3 air breathed. Another form in which
risk is presented is a drinking water or air concentration providing cancer risks of 1 in 10,000;
1 in 100,000; or 1 in 1,000,000.
Development of these hazard identification and dose-response assessments for
inorganic arsenic has followed the general guidelines for risk assessment as set forth by the
National Research Council (NRC, 1983). EPA guidelines that were used in the development
of this assessment may include the following: Guidelines for the Health Risk Assessment of
Chemical Mixtures (U.S. EPA, 1986a), Guidelines for Mutagenicity Risk Assessment (U.S.
EPA, 1986b), Guidelines for Developmental Toxicity Risk Assessment (U.S. EPA, 1991),
Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
Neurotoxicity Risk Assessment (U.S. EPA, 1998a), Guidelines for Carcinogen Risk
Assessment (U.S. EPA, 2005), Supplemental Guidance for Assessing Susceptibility from
Early-Life Exposure to Carcinogens (U.S. EPA, 2005), Recommendations for and
Documentation of Biological Values for Use in Risk Assessment (U.S. EPA, 1988),
(proposed) Interim Policy for Particle Size and Limit Concentration Issues in Inhalation
Toxicity (U.S. EPA, 1994a), Methods for Derivation of Inhalation Reference Concentrations
and Application of Inhalation Dosimetry (U.S. EPA, 1994b), Use of the Benchmark Dose
Approach in Health Risk Assessment (U.S. EPA, 1995), Science Policy Council Handbook:
Peer Review (U.S. EPA, 1998b, 2000a), Science Policy Council Handbook: Risk
Characterization (U.S. EPA, 2000b), Benchmark Dose Technical Guidance Document (U.S.
EPA, 2000c) and Supplementary Guidance for Conducting Health Risk Assessment of
Chemical Mixtures (U.S. EPA 2000d).
The literature search strategy employed for this compound was based on the CASRN
and at least one common name. At a minimum, the following databases were searched:
RTECS, HSDB, TSCATS, CCRIS, GENE-TOX, DART/ETIC, EMIC, TOXLINE,
CANCERLIT, and MEDLINE. Any pertinent scientific information submitted by the public
to the IRIS Submission Desk also was considered in the development of this document. The
relevant literature was reviewed through December, 2004.
DRAFT - DO NOT CITE OR QUOTE
-------
2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO
ASSESSMENTS
Properties
Arsenic is a metalloid that can exist in the -3, 0, +3, and +5 oxidation states. The arsenite
(+3) and arsenate (+5) forms are the primary forms found in drinking water. The chemical
and physical properties of arsenic are listed in Table 2-1 below.
Table 2-1. Chemical and Physical Properties of Arsenic and Selected Inorganic
Arsenic Compounds (ATSDR, 2000; Merck Index, 1989)
As
CAS No. . 7440-38-2
Valance 0
Molecular Weight 74.9
Synonyms Metallic
arsenic, gray
arsenic
Physical State Solid
(25-C).
Boiling Point ('C) 6 1 3 (sublimes)
Melting Point (*C) 817@28atm
Density 5.727
Vapor Pressure —
(20'C)
Water Solubility Insoluble
(g/lOOmL)
Log Octanol/Water —
Partition -
Coefficient (log
Taste Threshold
Odor Threshold
Conversion Factor —
ASA
1327-53-3
+3
197.8
Arsenic
trioxide,
arsenolite,
white arsenic
(+3)
Solid
465
312
3.738
—
3.7@20'C;
11.5@100'C
—
—
_.
As,O5 NaAsOj
1303-28-2 7784-46-5
+5 +3
229.8 129.9
Arsenic Sodium
pentoxide, arsenite (+3)
arsenic acid
anhydride (+5)
Solid Solid
...
315
(decompose)
4.32 1.87
---
150@16'C; Very soluble
76.7 @ 100'C
—
—
—
NajHAsO,
7778-43-0
+5
185.9
Disodium
arsenate (+5)
Solid
_.
86.3
1.87
...
Very soluble
...
—
—
— No data available.
DRAFT - DO NOT CITE OR QUOTE
-------
Uses
The metalloid, arsenic, is used as an alloying constituent in metallurgy for hardening
copper and lead alloys (HSDB, 2005). It also is used in glass manufacturing (decolorizing
and refining agent), as component of electrical devices, in the semiconductor industry, and as
a catalyst in the production of ethylene oxide. Other arsenic compounds are used in
preserving hides, as a mordant in the textile industry, medicinals, pesticides, pigments, and
wood preservatives. Approximately 90% of the domestic consumption of arsenic is in the
production of wood preservatives (e.g., chromate copper arsenate [CCR]; ATSDR, 2000),
which is currently being phased out.
Occurrence
Because arsenic naturally comprises 0.0005 % (2 ng/g) of the earth's crust, where it is
the twentieth most abundant element, background levels of arsenic occur in all environmental
media (ATSDR, 2000; Merck Index, 1989). Therefore, low concentrations of arsenic are a
natural component of water, food, soil, and air. However, industrial activities such as
refining and smelting operations have increased the concentration of arsenic in the
environment, often resulting in toxic concentrations of arsenic in soil, air, and water (Adams
et al., 1994). In addition, certain geographic areas have high levels of arsenic in their
underground rock formation, which can be leached and cause high arsenic concentrations in
drinking water (ATSDR, 2000).
The highest background arsenic levels found in the environment are in soils, at a
mean concentration of approximately 5,000 parts of arsenic per billion parts of soil (ppb)
(ATSDR, 2000). Food and drinking water also typically contain arsenic concentrations of 20
to 140 ppb (highest in shellfish and other marine foods) and <50 ppb, respectively (ATSDR,
2000). The majority of surface and ground waters contain less than 10 u.g arsenic/L
(although levels of 1,000-3,400 u.g/L have been reported, especially in areas of the western
U.S.), and several studies suggest that most (>90%) ground and surface drinking water
systems contain less than 5 fig/L of arsenic (ATSDR, 2000). Mean arsenic concentrations in
ambient air have generally been found to range from <0.001 to 0.003 micrograms per cubic
meter (u.g/m3) in remote areas and from 0.02 to 0.03 jig/m3 in urban areas, but can range
much higher, e.g., up to 2.5 u.g/m3 near nonferrous metal smelters (ATSDR, 2000).
Environmental Fate
Arsenic as a free element (0-oxidation state) is rarely encountered in the environment
(HSDB, 2005). Under normal conditions, arsenic is present as soluble inorganic arsenate
(+5-oxidation state) because it is more thermodynamically stable in water than arsenite (+3-
oxidation state).
DRAFT - DO NOT CITE OR QUOTE
-------
Arsenic is largely immobile in agricultural soils, and tends to remain and concentrate
on the soils surface indefinitely (ATSDR, 2000). However, the downward migration of
arsenic is much greater in sandy soil compared to clay loam (ATSDR, 2000). The most
influential parameter affecting arsenic mobility is the iron content of the soil. Leaching of
arsenic from polluted soil is generally low (ATSDR, 2000).
3 TOXICOKINETICS
3.1. ABSORPTION
The majority of arsenic exposures discussed in this Toxicology Review are from oral
waterborne arsenic exposure. Although inhalation exposure also is common, it will not be
addressed specifically in this report. Dermal exposure and exposure from food, however,
need to be addressed as they are possible confounding variables in epidemiology studies.
There is a lack of information on the bioavailability of inorganic arsenic in various types of
food (NRC, 1999; 2001). Although there have been no studies performed on the rate of
inorganic arsenic absorption through intact human skin, systemic toxicity in people dermally
exposed to inorganic arsenic solutions indicates the skin as an exposure route (Hostynek et
al., 1993). The systemic absorption via the skin appears to be low (NRC, 1999) as is
demonstrated by an in vitro study by Wester et al. (1993) where 2-6% of radiolabeled
arsenate (as a water solution) was absorbed by human and rhesus monkey skin over a 24-hour
period. Mouse dorsal skin was demonstrated to absorb 30% (Rahman et al., 1994) through
similar in vitro testing. This study also demonstrated that 60-90% of the absorbed arsenic
was retained in the skin. NRC (1999) suggests this indicates that inorganic arsenic binds
externally to skin and hair.
Water soluble forms of inorganic arsenic (both trivalent and pentavalent) are readily
absorbed (about 80-90%) in experimental animal models as well as humans (Pomroy et al.,
1980; Freeman et al., 1995). Monomethyl arsenic acid (MMAV) and dimethyl arsinic acid
(DMAV) also appear to be well absorbed in humans and experimental animals (75-85%;
Stevens et al., 1977; Buchet et al., 1981; Yamauchi and Yamamura, 1984; Marafante et al.,
1987; Yamauchi et al., 1988). Gastrointestinal absorption of low-solubility arsenic
compounds such as arsenic trisulfide, lead arsenate, arsenic selenide gallium arsenide
(Mappes, 1977; Marafante et al., 1987; Webb et al., 1984; Yamauchi et al., 1986), and of
arsenic-contaminated soil (Freeman et al., 1995), is much lower than soluble inorganic
arsenic. However, the degree of absorption for arsenic-contaminated soil is dependent on the
form of arsenic in the soil and on the type of soil.
Harrington et al. (1978) compared a group of people in Fairbanks, Alaska who had
arsenic contaminated water (345 ug/L) in their home, but drank only bottled water, to a group
of people who had less than 50 jig arsenic/L in their home water. The results demonstrated
that the group with high arsenic in their water had the same average concentration of total
arsenic metabolites in their urine (i.e., 43 |ig/L) as the group with less than 50 |j.g/L in their
DRAFT - DO NOT CITE OR QUOTE 5
-------
home water (i.e., 38 ng/L in urine), indicating external absorption via bathing or other
exposure sources. In addition, the group with high arsenic in their water had elevated hair
arsenic concentrations associated with bathing, suggesting that arsenic was bound externally
to hair and probably also to skin during washing with arsenic-rich water.
3.2. DISTRIBUTION
The retention and distribution of arsenic are determined mainly by its chemical
properties. While both arsenite (As111) and arsenate (Asv) bind to sulfhydryl groups, As1" has
approximately a 10-fold greater affinity for sulfhydryl groups than Asv (Jacobson-Kram and
Montalbano,1985), which has properties similar to those of phosphate. Cellular uptake rates
and resulting tissue concentrations are substantially lower for the pentavalent than for the
trivalent forms of arsenic, and in terms of cellular efflux of methylated species, DMA appears
more readily excreted than MMA (NRC, 2001). Liu et al. (2002) found arsenite to be
transported into cells by aquaglycoporins 7 and 9, which have been associated with the
transport of glycerol. Arsenate, however, has been suggested to be transported by the
phosphate transporter (Huang and Lee, 1996). Retention of arsenic can vary not only by its
form, but also by tissue (Thomas et al., 2001). Other factors that affect the retention and
distribution of arsenic include: the species, dose level, methylation capacity, valence form,
and route of administration.
Transportation in Blood
Once arsenic is absorbed it is transported in the blood to organs throughout the body.
It is generally bound to sulfhydryl groups of proteins and low-molecular-weight compounds
such as glutathione (GSH) and cysteine (NRC, 1999). Binding of As1" to GSH has been
demonstrated by several investigators (Anundi et al., 1982; Scott et al., 1993; Delnomdedieu
et al., 1994a; Delnomdedieu et al., 1994b). Because of species differences in binding
characteristics of arsenic, the retention in the blood varies among species. Arsenic
elimination in humans is triphasic with half-times of 1 hour, 30 hours, and 200 hours (Mealey
et al., 1959; Pomroy et al., 1980). Rats retain arsenic in the blood considerably longer than
other species because DMA"1 accumulates in red blood cells, apparently bound to
hemoglobin (Odanaka et al., 1980; Lerman and Clarkson, 1983; Vahter, 1983; Vahter, 1984).
Shiobara et al. (2001) demonstrated that the uptake of DMA in the blood was
dependent both on the chemical forms and animal species. DMA1" and DMAV were
incubated with rat, hamster, mouse, and human red blood cells (RBCs). DMAV was only
minimally taken up or the uptake was very slow in all animal species tested. DMA1", on the
other hand, was efficiently taken up by the RBCs in the following order: rats > hamsters >
humans. Mice RBCs were less efficient at the uptake of DMA1" and also followed a different
pattern from the other species. Rat RBCs retained the DMA1", but hamsters effluxed DMA1"
in the form of DMAV. Humans also effluxed DMA1" as DMAV, but the rate of uptake of
DMA"1 and efflux of DMAV was much slower than in hamster RBCs.
DRAFT - DO NOT CITE OR QUOTE
-------
In CHO cells, the rate of uptake was DMA1" > MMA1" > As"1 (Dopp et al., 2004); the
pentavalent forms are taken up at a much lower rate than the trivalent forms. Stevens et al.
(1977) calculated a half-time of 90 days in rat RBCs after a single oral dose of 200 mg/kg.
Lanz et al. (1950) also reported a high retention of arsenic in the blood of cats, although less
than in the rat; however, they did not determine if DMA was the cause. A single iv dose of
5.8 u.g As/kg (in the form of 73Asv) body weight administered to two male chimpanzees had a
half-life plasma elimination rate of 1 hour and a half-life elimination rate from red blood cells
of about 5 hours (Vahter et al., 1995b).
The concentration of arsenic in plasma and red blood cells will vary depending on
heath status and exposure level. Heydorn (1970) reported that heathy, presumably unexposed
people, had similar arsenic concentrations in their plasma and whole blood (2.5 ^g/L). This
indicates that there was very little accumulation in the red blood cells. However, Taiwanese
people exposed to arsenic-rich water had plasma levels of 15 ng/L and whole blood levels of
22 (ig/L. Blackfoot disease patients and their families had levels of 30 \igfL and 60 u.g/L in
the plasma and red blood cells, respectively, indicating an increasing accumulation of arsenic
in the red blood cells. De Kimpe et al. (1993) showed that chronic hemodialysis patients had
greater serum (12 ng/L) and erythrocyte (9.5 ng/L) arsenic levels than that of controls (0.38
\ig/L and 3.2 (ig/L, respectively). Similarly, Zhang et al. (1996) reported that mean total
arsenic levels were greater in hemodialysis patients (6.5 ng/L), compared to other types of
patients (5.1 ng/L), and controls (0.96 ug/L). The ratio between plasma and red blood cell
arsenic concentrations also may depend on the exposure form of arsenic (NRC, 1999).
Delnomdedieu et al. (1995) demonstrated that As1" is taken up more readily by red blood
cells in rabbit than Asv, MMA, or DMA.
Tissue Distribution
The distribution of inorganic arsenic (mainly As"1) in the skin, hair, oral mucosa, and
esophagus is most likely due to the binding of the inorganic form of arsenic with sulfhydryl
groups of keratin in these organs. In studies using rats and mice where the transfer of methyl
groups from S-adenosylmethionine (the proposed mechanism of arsenic metabolism; see
Metabolism section below) was chemically inhibited, the concentration of arsenic in most
tissues (especially the skin) increased (Marafante and Vahter, 1984). This also is evident in
the tissue distribution in the marmoset monkey, which does not methylate inorganic arsenic
(Vahter et al., 1982). In contrast to rodent models, however, marmoset monkeys also
accumulate arsenic in their testes, mainly in the spermatogenetic epithelium, and liver
(Vahter etal., 1982).
The longest retention of inorganic arsenic in mammalian tissues during experimental
studies has been observed in the skin (half-time more than a month; Marafante and Vahter,
1984), hair, squamous epithelium of the upper gastrointestinal tract (oral cavity, tongue,
esophagus, and stomach wall), epididymis, thyroid, skeleton, and the lens of the eye
(Lindgren et al., 1982, 1984; Vahter et al., 1982). All the aforementioned tissues, with the
exception of the skeleton, contained higher concentrations of As"1 than Asv within a short
DRAFT - DO NOT CITE OR QUOTE 7
-------
period after the individual addition of either. The calcified areas of the skeleton (in mice) had
an immediate accumulation and long-term retention of Asv, most likely a reflection of the
similarities between Asv and phosphate causing a substitution of phosphate by Asv in the
apatite crystals in bone.
Hughes et al. (2003) estimated that a steady-state of whole-body arsenic was
established after nine repeated daily doses of 0.5 mg As/kg as radioactive arsenate in adult
female B6C3F1 mice. Twenty-four hours dosing, the whole-body burden of arsenic
following 9 days of exposure was about twice that observed after a single dose; however the
rate of elimination was slower after the repeated dose. Accumulation of radioactivity was
highest in the bladder, kidney, and skin. The loss of radioactivity was greatest in the lungs,
but was slowest in the skin. Atomic absorption spectrometry revealed an organ-specific
distribution of arsenicals. MMA was detected in all tissues except the bladder. DMA was
found in the highest percentage in the bladder and lung after a single per os exposure, with
increases after repeated exposure. Inorganic arsenic was predominant in the kidney. After a
single per os exposure of arsenate (0.5 mg As/kg), DMA was the predominant form of
arsenic in the liver, but after nine repeat exposures the percentage of DMA decreased while
the percentage of inorganic arsenic increased. A trimethylated form of arsenic also was
detected in the liver.
Kenyon et al. (2005) examined the time course for tissue distribution of different
arsenic species after a single oral dose of 0, 10, or 100 micromole As/kg as sodium arsenate
to adult female B6C3F1 mice. The concentration of all forms of arsenic were lower in the
blood compared to the other organs across all doses and time points. The concentration of
inorganic arsenic in the liver and kidney were similar at both dose levels with peak
concentrations observed 1 hour after dosing. For the first 1 to 2 hours, inorganic arsenic was
the predominant form in the liver and kidney for both dose groups; however, at the later
times, DMA was the predominant form. Kidney measurements 1 hour after dosing
demonstrated that MMA was 3-4 times higher compared to the other tissues. DMA
concentrations in the kidney reached their peak 2 hours after dosing. DMA was the
predominant form measured in the lungs at all time points following exposure to 10
micromole As/kg as arsenate. DMA concentration in the lung were greater than or equal to
those of the other tissues beginning at 4 hours. The study could not distinguish the different
valence states of the MMA or DMA compounds.
Human subjects also have demonstrated a high concentration of arsenic in tissues
with a high content of cysteine-containing proteins, including the hair, nails, skin, and lungs.
Levels in these tissues of human subjects exposed to background levels of arsenic ranged
from 0.01 to 1.0 mg/kg of dry weight (Liebscher and Smith, 1968; Cross et al., 1979; Das et
al., 1995). Benign and malignant skin lesions from 14 patients with a minimum of 4 years
exposure to inorganic arsenical medication had a greater arsenic level (0.8 to 8.9 ppm) than
normal skin or malignant skin lesions from 6 subjects with no history of arsenic intake (0.4 to
1.0 ppm; Scott, 1958). In West Bengal, India, where the district average of arsenic in the
drinking water is 200-700 ng/L, arsenic concentrations in the skin scale, hair, and nails were
DRAFT - DO NOT CITE OR QUOTE 8
-------
1.9-5.5, 3.6-9.6, and 6.1-23 mg/kg dry weight, respectively (Das et al., 1996). However, the
amount of arsenic in the skin, hair, and nails resulting from an external exposure cannot be
determined.
Marmoset monkeys do not accumulate arsenic in the ocular lens or the thyroid
(Vahter et al., 1982). However, administration of MC labeled DMA to mice demonstrated
accumulations in these tissues. In fact, the tissues with the longest retention of DMA were
the lens of the eyes, thyroid, lungs, and intestinal mucosa (Vahter et al., 1984). Methylated
arsenic (both MMA and DMA), however, has a shorter tissue retention than inorganic arsenic
with 14C labeled-DMA concentration decreasing rapidly in most tissues in mice (Vahter et al.,
1984).
Trivalent and pentavalent inorganic arsenic, as well as methylated metabolites have
been demonstrated to cross the placenta at all stages of the gestational period (Hanlon and
Perm, 1977; Lindgren et al., 1984; Hood et al., 1987) with tissue distribution of arsenic
similar between the mother and the fetus. The marmoset monkey (known not to methylate
arsenic) had somewhat less placental transfer after administration of arsenite than mice
(Lindgren et al., 1984). The arsenic concentrations in the cord blood was similar to that
observed in maternal blood (an average of 9 p.g/L) in pregnant women living in a village in
northwestern Argentina, where the arsenic in the drinking water is approximately 200 ng/L
(Concha et al., 1998b). This also was noted in pregnant women with no known arsenic
exposure where an average of 2-3 u.g/L was measured in the cord blood and maternal blood
(Kagey et al., 1977). Women living near smelters also have been determined to have an
increased concentration of arsenic (Tabacova et al., 1994). Although the fetus is exposed to
arsenic, it may be more in the form of DMA (at least in late gestation), as 90% or more of the
arsenic in the urine and plasma of newboms and mothers (at time of delivery) was in the form
of DMA. In any case, the fetal toxicity of arsenic has yet to be clarified.
Intracellular Distribution
Rabbits and mice exposed to radiolabeled arsenic had the majority of the arsenic in
the nuclear and soluble fractions when measured in the liver, kidneys, and lungs (Marafante
et al., 1981; Marafante and Vahter, 1984). The marmoset monkey had a different
intracellular distribution with approximately 50% of the arsenic in the microsomal fraction in
the liver (Vahter et al., 1982; Vahter and Marafante, 1985). However, chemical inhibition of
arsenic methylation in rabbits did not alter the intracellular binding of arsenic (Marafante and
Vahter, 1984; Marafante et al., 1985). An increase in tissue arsenic concentration (especially
in the liver) was associated with an increase in the arsenic concentration in the microsomal
fraction in the liver in rabbits fed diets with low concentrations of methionine, choline, or
proteins, which leads to a decrease in arsenic methylation (Vahter and Marafante, 1987). The
levels of arsenic in the microsomal fraction of the liver in these rabbits was similar to those
observed in the marmoset monkey (Vahter et al., 1982), therefore, indicating nutritional
factors may play a role in the subceilular distribution of arsenic.
DRAFT - DO NOT CITE OR QUOTE
-------
3.3. METABOLISM
Upon entering the system arsenate can be reduced to arsenite. 'A substantial fraction
of absorbed Asv is rapidly reduced to As"1, probably mainly in the blood, in mice, rabbits, and
marmoset monkeys (Vahter and Envall, 1983; Vahter and Marafante, 1985; Marafante et al.,
1985). This reduction also might occur in the stomach or intestines, but quantitative
experimental data are not available. GSH may play a role in the reduction of Asv(NRC,
1999). Cysteine and dithiothreitol (DTT) have been shown to reduce pentavalent arsenic in
purified rabbit liver enzyme preparations (Zakharyan et al., 1995) and rat liver cytosols
(Buchel and Lauwerys, 1985; Styblo et al.,' 1996).
Although pentavalent arsenic can be directly reduced, arsenate reductase enzymes
have been detected in the human liver (Radabaugh and Aposhian, 2000), which has been
. subsequently characterized as a purine nucleoside phosphorylase (PNP) (Gregus and Nemeti,
2002; Radabaugh et al., 2002). This enzyme requires a thiol and a heat-stable cofactor for
activation. According to Radabaugh et al. (2002), dihydrolipoic acid (DHLP) is most active
naturally occurring thiol in mammalian systems required for the enzymatic reduction of
arsenate to arsenite.
Nemeti et al. (2003), however, observed the reduction of arsenate to arsenite by PNP
in vitro only. PNP did not appear to a major player in the reduction of arsenate to arsenite in
either human erythrocytes or in rats in vivo. Ne"meti and Gregus (2004; 2005) further
demonstrated that human erythrocytes contain a PNP-independent Asv-reducing mechanism
that requires a supply of GSH, NAD, and a substrate to either one or both of the following
enzymes: glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or phosphoglycerate kinase
(PGK). This also was demonstrated to be the mechanism of reduction in rat liver cytosol
(Ne"meti and Gregus, 2005), however, another unidentified enzyme in the liver cytosol also
had the capacity to reduce Asv. A further study (Gregus and N6meti, 2005) demonstrated
that GAPDH was capable of Asv reductase activity, but that PGK served as an auxiliary
enzyme when 3-phosphoglycerate is the glycolic substrate.
The oxidative methylation of trivalent arsenicals, as well as, the reduction of
pentavalent arsenicals is catalyzed by an arsenic (+3 oxidation state) methyltransferase,
which is encoded by the AS3MT gene (Waters et al., 2004a). In chimpanzees, which do not
methylate arsenic, Li et al. (2005) found a frameshift mutation in the gene code that caused a
deletion in the gene for arsenic (+3 oxidation state) methyltransferase leading to an inactive
truncated protein. The addition of GSH increased the yield on mono- and di-methylated
arsenicals, but suppressed the production of TMAO (Waters et al., 2004a). Thomas et al.
(2004) discovered a similar arsenic methyltransferase in the rat liver, which they designated
cyt!9 because an orthologous cyt!9 gene encodes an arsenic methyltransferase in the mouse
and human genome. Glutathione alone does not support recombinant rat cyt!9 catalytic
function, but when added to the reaction mixture containing other reductants, the rate of
arsenic methylation increases (Waters et al., 2004b). According to Waters et al. (2004b),
cyt!9 may posses both As1" methyltransferase and Asv reductase activities. In the presence of
DRAFT - DO NOT CITE OR QUOTE 10
-------
exogenous or physiological reductant, cyt!9 was found to catalyze the entire sequence to
convert arsenite to its methylated metabolites. Drobna et al. (2004) linked the genetic
polymorphism ofcyt]9, with other cellular factors and to the interindividual variability in the
capacity of primary human hepatocytes to retain and metabolize As'".
Dithiols (e.g., reduced lipoic acid) have been indicated in recent years to be more
active than GSH for providing a reducing environment for methylation by MMA1"
methyltransferase (Zakharyan et al., 1999). An MMAV reductase also has been detected in
rabbit liver (Zakharyan and Aposhian, 1999), hamster tissues (Sampayo-Reyes et al., 2000),
and human liver (Zakharyan et al., 2001). In any case, arsenate is rapidly reduced to arsenite;
therefore, chronic exposure to arsenite and arsenate should result in fairly similar metabolite
distribution in the body unless the reducing capacity is exceeded by acute high-dose
exposures (Vahter, 1981; Lindgren et al., 1982).
Humans and most experimental animal models methylate inorganic arsenic to MMA
and DMA, with the amounts differing by species as determined by urinary metabolites. The
methylated metabolites historically have been considered less acutely toxic, less reactive with
tissue constituents, less cytotoxic, and more readily excreted in the urine than inorganic
arsenic (Buchet et al., 1981; Vahter and Marafante, 1983; Vahter et al., 1984; Yamauchi and
Yamamura, 1984; Marafante et al., 1987; Moore et al., 1997a; Rasmussen and Menzel, 1997;
Concha et al., 1998a; Hughes and Kenyon, 1998; Sakurai et al., 1998). Monomethylarsonous
acid (MMA'") and dimethylarsenous acid (DMA111), however, have recently been
demonstrated to be more cytotoxic in Chang cells (a human liver cell line; Petrick et al.,
2000, 2001), CHO (Dopp et al., 2004), and cultured primary rat hepatocytes (Styblo et al.,
1999b, 2000) than As1", Asv, MMAV, or DMAV. '
Methylation is important in the distribution in tissues and in the excretion of arsenic.
It has been demonstrated that inhibition of arsenic methylation results in increased tissue
concentrations of arsenic (Marafante and Vahter, 1984; Marafante et al., 1985).
Arsenic Methylation
In vitro studies using rat liver preparations indicate that the methylating activity is
localized in the cytosol with S-adenosylmethionine (SAM) as the main methyl donor for As"1
(Marafante and Vahter, 1984; Buchet and Lauwerys, 1985; Marafante et al., 1985; Styblo et
al., 1995, 1996; Zakharyan et al., 1995). The proposed metabolic pathway for arsenic is
demonstrated in Figure 3.1.
DRAFT - DO NOT CITE OR QUOTE 11
-------
Figure 3-1. Proposed metabolic pathway for the methylation of inorganic arsenic in
humans. R = reductase; MT = methyltransferase; e" = electron
MT (MMAV) R (MMA111)
CH3+ 2 e-
(CH3)2AsO2- (CH3)2AsO-
MT (DMAV) R (DMA111)
The main products of methylation are MMAV and DMAV, which are readily excreted
in the urine (Marcus and Rispin, 1988). MMA1" and DMA1" have recently been detected in
human urine (NRC, 2001). However, most publications do not differentiate the valence state
of MMA or DMA. Le et al. (2000a,b) as well as Del Razo et al, (2001) noted that the
concentration of trivalent metabolites in the urine may be underestimated because they are
easily oxidized after collection. However, Le et al. (2000a,b) discovered 32-127 u^g/L of
MMA"1 and 21-38 ^ig/L of DMA1" in the urine of a population from Inner Mongolia, China,
who were exposed to 510-660 ngTL (0.46 uM) of arsenic via the drinking water. MMA1" and
DMA"1 have recently been demonstrated to be more cytotoxic in Chang cells (a human liver
cell line; Petrick et al., 2000) or cultured primary rat hepatocytes (Styblo et al., I999b, 2000)
than As1", Asv, MMAV, or DMAV. Petrick et al. (2001) estimated LD50s of 112 u.mol/kg and
29.3 umol/kg for As1" and MMA1", respectively, therefore, indicating the toxicity of
intermediates in the metabolism of arsenic.
A small percent of DMA1" may further be methylated to trimethylarsine oxide
(TMAO) in mice, hamsters, and humans (for review, see Kenyon and Hughes, 2001).
TMAO is detected in the urine following DMA exposure, but has not been detected in the
blood or tissues of mice exposed intravenously to DMA (Hughes et al., 2000) or in the urine
of mammals exposed to inorganic arsenic. This may be due to rapid clearance of DMA and
MMA from cells (Styblo et al., 1999a; Lin et al., 2001); however, most analytical methods
are not optimized for the detection of TMAO and thus it could be present and not detected.
Styblo et al. (1999b) reported that DMA was the only metabolite detected in rat or human
hepatocytes incubated with DMA or a glutathione complex of DMA.
Although the kinetics of arsenic methylation in vivo are not fully understood, it is
believed that the liver may be the primary source of arsenic methylation. Marafante et al.
(1985) discovered that DMA appeared in the liver prior to any other tissue in rabbits exposed
to inorganic arsenic. It also has been demonstrated that oral administration of inorganic
arsenic favors methylation more than either subcutaneous or intravenous administration
(Charbonneau et al., 1979; Vahter, 1981; Buchet et al., 1984) presumably because the arsenic
will pass through the liver first after oral administration. However, liver disease (i.e.,
alcoholic, postnecrotic, or biliary cirrhosis, chronic hepatitis, hemochromatosis, and steatosis)
DRAFT - DO NOT CITE OR QUOTE 12
-------
increased the ratio of DMA to MMA in the urine following a single injection of sodium
arsenite (Buchet et al., 1984; Geubel et al., 1988), indicating a more efficient methylation of
arsenic. In addition, the site of methylation may depend on the rate of reduction of Asv to
As111. Isolated rat hepatocytes readily absorbed and methylated As1", but not Asv (Lerman et
al., 1983). Kidney slices, on the other hand, produced five times more DMA produced from
Asv than As1" (Lerman and Clarkson, 1983). Therefore, it is likely that any arsenate not
initially reduced can be methylated in the kidney for urinary excretion.
There have been unequivocal results with in vitro studies on the methylation capacity
of different tissues. Buchet and Lauwerys (1985) reported the liver in rats as the main organ
for methylation with the methylating capacities in the red blood cells, brain, lung, intestine,
and kidneys as insignificant in comparison. Arsenite methyltransferases from mouse tissues
demonstrated that the testes had the highest amount of methylating activity, followed by the
kidney, liver, and lung (Healy et al., 1998). Aposhian (1997) determined that the amount of
methyltransferases vary in different tissues and animal species, methylating capacities in
vitro, however, do not necessarily reflect in vivo methylation (NRC, 1999). Arsenite bound
to tissue component can be methylated and released. This may explain the initial rapid phase
(immediate methylation and excretion) followed by a slow elimination phase (continuous
release of bound arsenite through methylation) (Marafante et al., 1981; Vahter and Marafante,
1983) as described below under Elimination.
Species Differences in the Methylation of Arsenic
There is considerable variation in inorganic arsenic methylation among mammalian
species (NRC, 1999). Humans, rats, mice, dogs, rabbits, and hamsters have been shown to
have efficient methylation of arsenic to MMA and/or DMA. Rats and hamsters also appear
to methylate administered DMA to TMAO more efficiently than other species (Kitchin et al.,
1999; NRC, 1999; Yamauchi and Yamamura, 1984) with about 10% urinary arsenic present
as TMAO after exposure to DMA. in the drinking water (100 mg/L) of male rats (Yoshida et
al., 1998). Humans (mainly exposed to background levels or exposed at work) have been
estimated through a number of studies to excrete 10-30% of the arsenic in its inorganic form,
10-20% as MMA, and 55-75% as DMA (see Hopenhayn-Rich et al., 1993 for review). A
study of urinary arsenic in a population in northern Argentina exposed to arsenic via drinking
water demonstrated an average of only 2% MMA in the urine (Concha et al., 1998a; Vahter
et al., 1995a). This may indicate variations in methylation depending on route of exposure,
level of exposure, and possible nutritional or genetic factors.
The rabbit (Marafante ct al., 1981; Vahter and Marafante, 1983; Maiorino and
Aposhian, 1985) and hamster (Charbonneau et al., 1980; Yamauchi and Yamamura, 1984,
1985; Marafante and Vahter, 1987) are more comparable to human with respect to arsenic
methylation than other experimental animals. However, in general rabbits and hamsters
excrete more DMA and less MMA than humans. However, Flemish giant rabbit (De Kimpe
et al., 1996) and New Zealand rabbits (Bogdan et al., 1994) excrete MMA in amounts similar
to humans. Mice and dogs, efficient methylators of arsenic, excrete more than 80% of the
DRAFT - DO NOT CITE OR QUOTE 13
-------
administered dose as DMA within a few days (Charbonneau et al., 1979; Vahter, 1981).
Guinea pigs (Healy et al., 1997), marmoset monkeys (Vahter et al., 1982; Vahter and
Marafante, 1985), and chimpanzees (Vahter et al., 1995b) do not appear to methylate
inorganic arsenic. In addition, no methyltransferase activity was detected in these species
(Zakharyan et al., 1995, 1996; Healy et al., 1997; Vahter, 1999b). Li et al. (2005) found a
frameshift mutation in the gene code of chimpanzees that caused a deletion in the gene for
arsenic (+3 oxidation state) methyltransferase leading to an inactive truncated protein. Healy
et al. (1999) determined marked variations in the activity of methyltransferases and Vahter
(1999a) determined a difference in methylation efficiency between different species. The
variations in methyltransferase activity and methylation efficiency are probably the
underlying reason for the cross-species variability in methylation ability as all the species had
ample arsenate reductase activity (Vahter, 1999b; NRC, 2001). However, methyltransferases
have only recently been detected in human hepatocytes (Zakharyan et al., 1999; Styblo et al.,
1999a).
3.4. ELIMINATION
The major route of excretion for most arsenic compounds by humans is via the
urine with a biological half-time of about 4 days, which is slightly shorter following
exposure to Asv than to As1" (Yamauchi and Yamamura 1979; Tam et al. 1979; Pomroy et
al. 1980; Buchet et al. 1981). Six human subjects, who ingested radiolabeled
74As-arsenate, excreted 38% of the dose in the urine within 48 hr and 58% within 5 days
(Tam et al., 1979). The data indicate a three compartment exponential function, with 66%
excreted with a half-time of 2.1 days, 30% with a half-time of 9.5 days, and 3.7% with a
half-time of 38 days (Pomroy et al., 1980). Buchet et al. (1981) supported this information
by exposing three subjects orally to 500 u.g of arsenic in the form of arsenite in water.
Buchet's results indicate about 33% of the dose was excreted in the urine within 48 hr, and
45% within 4 days. The methylated metabolites, MMA and DMA, are excreted in the
urine faster than the inorganic arsenic. In humans, about 78% of MMA and 75% of DMA
were excreted in the urine within 4 days of ingestion (Buchet et al., 1981). In mice, the
. half-time of MMA and DMA was found to be about 1 hr following iv administration
(Hughes and Kenyon, 1998).
Rats have a slow whole body clearance of DMA (Vahter et al., 1984), even though
they are efficient at methylating inorganic arsenic to DMA, because a significant portion of
the DMA produced is retained in the erythrocytes (Odanaka et al., 1980; Lerman and
Clarkson, 1983; Vahter, 1983; Vahter et al., 1984). The rat also has extensive biliary
excretion of inorganic arsenic which is about 800 times greater than observed in the dog
and 37 times that of the rabbit.
Although absorbed arsenic is removed from the body mainly via the urine, small
amounts of arsenic are excreted via other routes (e.g., skin, sweat, hair, and breast milk).
Although arsenic has been detected to a low degree in the breast milk of women in
northwestern Argentina (i.e., 2 fig/kg), breast-feeding caused a decrease in the
DRAFT - DO NOT CITE OR QUOTE 14
-------
concentration of arsenic in the urine of the newborn child (Concha et al., 1998c). Women
in the Philippines were determined to have about 19 ^g/kg arsenic in their breast milk. The
average concentration of arsenic in sweat induced in a hot and humid environment was 1.5
u.g/L, and the hourly loss was 2 u.g (Vellar, 1969). With an average arsenic concentration
in the skin of 0.18 mg/kg, Molin and Wester (1976) estimated that the daily loss of arsenic
through desquamation was 0.1-0.2 (ig in males with no known exposure to arsenic.
3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
A physiologically based pharmacokinetic (PBPK) model for exposure to inorganic
arsenic (orally, intravenously, and intratracheally) in hamsters and rabbits has been
developed by Mann et al. (1996a). It contains tissue compartments for lung (naso-pharynx,
tracheo-bronchial, pulmonary), plasma, RBCs, liver, gastrointestinal (GI) tract, skin, lungs,
kidney, keratin, and combined other tissues. Oral absorption of As111, Asv, and DMA
(pooled 111 and V oxidation states) is modeled as a first-order transport directly from the Gl
contents into the liver. Distribution to tissues is diffusion-limited, based upon literature
values for capillary thickness and pore sizes for each tissue! Reductive metabolism of Asv
to As1" is modeled as a first-order process in the plasma. Oxidative metabolism of As1" to
Asv is modeled as first-order processes in the plasma and kidneys. Methylation of
inorganic As species to MM A (pooled III and V oxidation states) and then to DMA is
modeled as saturable Michaelis-Menten processes in the liver. Urinary, biliary, and fecal
excretion of As1", Asv, MMA, and DMA are modeled as first-order processes. Parameters
for absorption, tissue partition, metabolism, and biliary excretion were estimated by fitting
the model to literature data on the urinary and fecal excretion of total arsenic from rabbits
and hamsters administered various arsenic compounds by iv, oral gavage, or intratracheal
instillation (Marafante et al, 1985; Marafante et al., 1987; Charbonneau et al., 1980;
Yamauchi and Yamamura, 1984). The model was found to simulate accurately the
excretion of arsenic metabolites in the urine of rabbits and hamsters, and was additionally
validated with reasonable fits to liver, kidney, and skin concentrations in rabbits and
hamsters (Marafante et al, 1985; Marafante and Vahter, 1987; Yamauchi and Yamamura,
1984).
Mann et al. (1996b) extended the PBPK model for humans by adjusting
physiological parameters (organ weights, blood flows) accordingly and re-estimating
absorption and metabolic rate constants by fitting the model to literature data on the urinary
excretion of total arsenic following a single oral dose of As1" or Asv in human volunteers
(Buchet et al., 1981; Tarn et al., 1979). The extended model was validated against
empirical data on the urinary excretion of the different metabolites of inorganic arsenic
following repeated oral intake of arsenite, intake of inorganic arsenic via drinking water,
and occupational exposure to arsenic trioxide (Buchet et al., 1981; Vahter et al., 1986;
Harrington et al., 1978; Valentine et al., 1979). The model predicted a slight decrease (i.e.,
about 10% with an increase in dose of about 1,000 u,g) in the percentage of DMA in urine
with increasing single-dose exposure (highest dose of arsenic at 15 |ig/kg of body weight),
DRAFT - DO NOT CITE OR QUOTE 15
-------
especially following exposure to As1", and an almost corresponding increase in the
percentage of MMA. The model demonstrated that adults drinking water containing 50
u,g/L had a higher urinary excretion of arsenic than an occupational exposure of 10 u.g/m3
(Mannetal., 1996b).
Yu (1999a,b) also developed a PBPK model for arsenic in humans. It contains
tissue compartments for lung, skin, fat, muscle, combined kidney and richly perfused
tissues, liver, intestine, GI contents, stomach contents, and bile. Oral absorption of As1".,
Asv, and DMA (pooled III and V oxidation states) is modeled as a first-order transport
from the GI contents into the intestinal.tissue. Distribution to tissues is flow-limited.
Reductive metabolism of Asv to As"1 is modeled as a first-order, glutathione-dependent
process in the intestinal tissue, skin, liver and kidney/rich tissues. Oxidative metabolism of
As1" to Asv is not modeled. Methylation of inorganic As species to MMA (pooled III and
V oxidation states) and then to DMA is modeled as saturable Michaelis-Menten processes
in the liver and kidney/rich tissues. Urinary, biliary, and fecal excretion of As1", Asv,
MMA, and DMA are modeled as first-order processes. Parameters for absorption, tissue
partition, metabolism, and biliary excretion were estimated by fitting the model to literature
data on tissue concentrations of total arsenic from a fatal human poisoning (Saady et al.,
1989) and blood, urine, and fecal elimination of total arsenic following oral administration
(Odanaka et al., 1980; Pomroy et al., 1980). The model was not validated against external
data and fits to the data sets used for parameter estimation were not provided.
Gentry et al. (2004) adapted the model proposed by Mann et al. (1996a) to different
mouse strains by adjusting physiological parameters (organ weights, blood flows)
accordingly and re-estimating absorption, partition, and metabolic rate constants by fitting
the model to literature data on urinary excretion of various arsenic species following iv
administration of MMA to B6C3F1 mice (Hughes and Kenyon, 1998) or single oral
administration of Asm or Asv to B6C3F1 mice (Kenyon et al., 1997; Hughes et al., 1999).
Additionally, the description of methylation in the model was refined to address the
uncompetitive inhibition of the conversion of MMA to DMA by As"1. The PBPK model
was then validated using data from a single oral administration of Asv (Hughes et al., 1999)
and 26-week drinking water exposure of As1" to C57Black mice (Moser et al., 2000). This
data was adequately fit by the model without further parameter adjustment. Ng et al.
(1999) had found arsenic-induced tumors in C57C1/6J mice, while numerous other mouse
strains (Swiss CR:NIH[S], C57Bl/6p53[+/-], and C57Bl/6p53[+/+], and Swiss CD-I) have
not resulted in a significant increase in arsenic-induced tumors. The Gentry et al. model
was unable to explain the different outcomes in the mouse bioassay.
The Mann et al. (1996a,b) and Gentry et al. (2004) models are well-documented,
validated against external data, and appear to capture the salient features of arsenic
toxicokinetics in rodents and humans. The information provided by these different models
may help understand the mode(s) of action involved in carcinogenesis along with possible
reasons that humans are apparently more susceptible to the carcinogenic effects of arsenic.
DRAFT - DO NOT CITE OR QUOTE 16
-------
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS
The toxicily of arsenic is dependent on its valence state (i.e., trivalent or
pentavalent) with trivalent arsenic appearing to be 4 times more acutely toxic than
pentavalent arsenic. However, symptoms are not dependent on valence slate. There have
been a number of epidemiology studies conducted on the health effects of arsenic. Some
of the limitations of the epidemiology studies include not specifying the form of arsenic,
difficulties with the exposure assessment, and/or not examining possible confounding
variables. However, the pentavalent form of inorganic arsenic predominates in water and
soil (Buchet and Lison, 2000). Although many of the epidemiology studies attempt to
examine major confounders such as age or smoking status, few if any address nutritional
effects such as selenium and/or zinc deficiencies, which may enhance arsenic toxicity
(Gebel, 2000). This section will focus mainly on the epidemiology studies conducted on
the effects of chronic ingestion of arsenic in drinking water. The studies address the
concentration of arsenic in the well water, but few if any address exposure through bathing
or possible exposure through food consumption.
Cancer Effects
There are many epidemiology reports that examined the association between
arsenic and cancers. Each study is conducted in a different manner (i.e., prevalence
studies, cross-sectional studies, case-controls, cohort, and ecological studies) with its own
limitations, but the majority found an association between arsenic and cancer (type varied
by study). Table 4-1 summarizes the epidemiology studies giving their type and cancer
results. Arsenic has been associated with cancer as far back as 1887 when Hutchinson
reported an unusual number of skin tumors in patients treated with arsenicals, The
association between skin cancer and arsenic has been reported in a number of studies since
then (Neubauer, 1947; Sommers and McManus, 1953; Roth, 1957; Robson and Jelliffe,
1963; Fiertz, 1965; Tseng et al., 1968; Tseng, 1977; Kjeldsberg and Ward, 1972; Jackson
and Grainge, 1975; Popper et al., 1978; Prystowsky et al., 1978; Reymann et al., 1978;
Nagy et al., 1980; Falk et al., 1981; Roat et al., 1982; Robertson and Low-Beer, 1983; Yu,
1984; Yuetal., 1984; Chen etal., 1985; Wuet al., 1989; Chen etal., 1988; Cuzicket al.,
1992; Hopemhayn-Rich et al., 1996a and 1998; Smith et al., 1998; Ahmad et al., 1999;
Hinwood et al., 1999; Ma etal., 1999; Tsai etal., 1999; Karagas etal., 2001; Kuorkawaet
al., 2001; and Tucker et al., 2001).
Several studies have demonstrated an association between arsenic and cancer
and/or disease in populations in Taiwan. This population has been extensively studied due
to the switch from surface water wells to artesian (ground water) wells for drinking water
in the 1920s to improve the microbiology and reduce salinity of the water. The artesian
wells have since been discovered to be contaminated with naturally occurring arsenic
resulting in widespread exposure.
DRAFT - DO NOT CITE OR QUOTE 17
-------
Tseng et al. (1968) and Tseng (1977) demonstrated an increase in the occurrence of
skin lesions with high level (^0.60 ppm) of exposure to arsenic from artesian wells in
villages of southwestern Taiwan. The increases were seen across both sexes and different
age groups. This was an ecological study examining the prevalence of skin cancer,
keratosis, and hyperpigmentation in 37 villages in Taiwan containing a total of 40,421
inhabitants who consumed drinking water from artesian wells, which unknowingly was
contaminated with high concentrations of naturally occurring arsenic. Three exposure
groups were investigated: 0.0-0.29, 0.30-0.59, and ^ 0.60 ppm. Hyperpigmentation
(183.5/1000) had the highest prevalence, followed by keratosis (71.0/1000), and skin
cancer (10.6/1000). There was a clear-cut increase in the prevalence of skin cancer from
low to high exposure groups in both sexes of three age groups. In 20-39 year old the
prevalence were 1.3, 2.2, and 11.5 from the low- to high-dose group. For the same dose
groups the prevalence was 4.9, 32.6, and 72.0 in the 40-59 year age group and 27.1, 106.2,
and 192.0 in the 60 years and over age group.
Chen et al. (1985) studied 84 villages in Southwestern Taiwan known to be
endemic for Blackfoot-disease and other types of skin lesions. Mortality rates of the 84
villages from 1968-1986 were compared to national and sex-specific rates. Of these
villages, 31 used only artesian wells (100 to 200 m deep), 27 used both artesian and
shallow wells, 24 used only shallow wells, and 2 used surface water. Although no
exposure measurements were taken, the authors cited an earlier study (Chen et al. 1962)
that reported the arsenic content of artesian well water measured in BFD-endemic areas in
1959 to 1962 to range from 0.35 ppm to 1.14 ppm with a median of 0.78 ppm and the
shallow well water to have arsenic content between 0.00 and 0.30 ppm, with a median of
0.04 ppm (Chen et al. (1962) actually reported the arsenic content of BFD-endemic shallow
wells to be between 0.00 and 0.25 ppm, with a median of 0.03ppm).
Standardized Mortality Ratios (SMRs) of the 84 villages for the period 1968-1982
were computed for various cancers by gender. National Taiwan rates were used to derive
the standard. The SMRs of the villages, regardless of exposure, for various cancers
separated by sex are listed in Table 4-1. Cancers of the small intestine, esophagus,
nasopharynx, rectum, stomach, and thyroid were not statistically significantly associated
with arsenic exposure.
DRAFT - DO NOT CITE OR QUOTE 18
-------
Table 4-1. SMRs in Taiwan by Sex. From Chen et al. (1985).
(95% confidence intervals in parentheses.)
Males Females
Bladder cancer 11.0(9.3-12.7) 20.1(17.0-23.2)
Kidneycancer 7.7 .(5.4-10.1) 11.2(8.4-14)
Skin cancer 5.3 (3.8-6.9) 6.5(4.7-8.4)
Lung cancer 3.2 (2.9-3.5) 4J (3.6-4.7)
Livercancer 1.7 (1.5-1.9) 2.3(1.9-2.7)
Colon cancer. 1.6 (1.2-2.0) 1.7(1.3-2.1)
Leukemia 1.4 (1.0-1.8) 0.9(0.5-1.3)
Chen et al. (1985) also found that the age-specific cumulative mortality ratios for
cancers of the bladder, kidney, skin, lung, liver, and colon in the BFD-endemic area were
higher than in Taiwan for all age groups and that the SMRs in the BFD-endemic areas
generally increased with increasing BFD relevance rates. Finally, the authors observed a
dose-response association between BFD endemicity (defined as follows: hyperendemic
when BFD rate was greater than 5.0/1000; endemic when BFD prevalence rate was
between 0.1 and 5.0/1000; and nonendemic where no BFD cases were reported) and the
SMRs for cancer of the bladder, kidney, skin, lung and liver, but not colon cancer.
Although cigarette smoking has been related to cancers of the kidney and bladder, the high
SMRs observed in the BFD-endemic regions are not readily explained by the relatively
higher smoking rate in the BFD-endemic area than in Taiwan (40% versus 32%).
Chen et al. (1988) analyzed the mortality of residents in the BFD-endemic area
based on 899,811 person-years observed from 1973 to 1986. Age-adjusted mortality rates
(per 100,000) for various cancers were calculated, with results displayed in Table 4-2.
DRAFT - DO NOT CITE OR QUOTE 19
-------
Table 4-2. Age-Adjusted Mortality Rates (per 100,000) in Taiwan
from Chen etal. (1988)
Cancer site
All Sites
Liver
Lung
Skin
Prostate
Bladder
Kidney
Sex
M
F
M
F
M
F
M
F
M
M
F
M
F
BFD-endemic area
>0.6*
434.7
369.4
68.8
31.8
87.9
83.8
28.0
15.1
8.4
89.1
91.5
21.6
33.3
0.30-
3.59*
258.9
182.6
42.7
18.8
64.7
40.9
10.7
10.0
5.8
37.8
35.1
13.1
12.5
O.30*
154.0
113.3
32.6
14.2
35.1
26.5
1.6
1.6
0.5
15.7
16.7
5.4
3.6
General
population
in Taiwan
128.1
85.5
28.0
8.9
19.4
9.5
0.8
0.8
1.5
3.1
1.4
1.1
0.9
*Well-water arsenic concentrations (ppm).
Significantly higher age-adjusted mortality from the various cancers occurred
among residents in the BFD-endemic region than in the general Taiwanese population. In
addition, according to Chen et al. (1988), there was a significant dose-response relation
between the arsenic level in drinking water and age-adjusted mortality for cancers of the
bladder, kidney, skin, prostrate, lung, and liver.
Wu et al. (1989) investigated 42 villages in southwestern Taiwan with
contaminated drinking water sources. These included 27 of the villages in the four
townships investigated by Chen et al. (1985) plus an additional 15 villages in two other
townships. Water samples were collected from 155 wells of the 42 villages where shallow
and/or deep well water was being used in 1964-1966. The arsenic content of well water
samples ranged from 0.010 ppm to 1.752 ppm with two clusters at levels of 0.05 ppm-0.25
ppm and 0.45-0.65 ppm.
Death certificates of people who died from cancer during 1973-1986 were coded
for underlying causes of death, and age-adjusted mortality rates were computed. The
resulting mortality rates were categorized into three groups according to their median
arsenic levels of well water (<0.30 ppm (20 villages); 0.30-0.59 ppm (15 villages); and
DRAFT - DO NOT CITE OR QUOTE
20
-------
^0.60 ppm (7 villages). Mortality from a number of different types of cancers was found
to increase with increasing arsenic concentrations for both sexes among residents aged 20
years or more. The authors performed a Mantel-Haenszel x2 test to test for significance of
the trends. Table 4-3 shows the age-adjusted mortality rates (per 100,000) from various
cancers and all sites combined that were significant. No significant trends were found for
mortality from leukemia, and cancers of the nasopharynx, esophagus, stomach, colon and
uterine cervix with increasing arsenic levels.
Table 4-3. Model Parameters and Potency Index (Wu et al. 1989)
Cause
of
death
All
sites
Bladder
Kidney
Skin
Lung
Liver
Prostat
e
Median arsenic levels in well water (ppm) and significance levels
Males (467,173 person-years)
<0.30
224.56
22.64
8.42
2.03
49.16
47.78
0.95
0.30-
0.59
405.12
61.02
18.90
14.01
100.67
67.62
9.00
;> 0.60
534.61
92.71
25.26
32.41
104.08
86.73
9.18
P
t
t
t
t
t
I
t
Females (431,633 person-
years)
0.30
162.22
25.60
3.42
1.73
36.71
21.40
-
0.30-
0.59
277.20
57.02
19.42
14.75
60.82
24.18
-
* 0.60
487.20
111.30
57.98
18.66
122.16
31.75
-
P
t
t
t
t
t
N
S
-
test.
t < 0.05; t < 0.001, based on the trend test of the extension of the Mantel-Haenszel
NS: not significant
Chen et al. (1992) used the Armitage-Doll model to develop cancer potency indices
for residents in an endemic area of chronic arsenism. The study area and population were the
same as in Wu et al. (1989). The age-specific mortality rate for a disease was calculated as:
I(t,d) = B(t) + H(t,d), where I(t,d) was the age and cause-specific mortality rate, B(t)
the background mortality rate, where the reference was the general Taiwan population, and
H(t,d) the mortality rate due to the arsenic exposure at age t and dose rate d, which was
assumed to have the form H(t,d) = a x d x tk.
The authors estimated the parameters a and k using maximum likelihood methods.
Table 4-4 shows the results. The potency index represents the excess lifetime risk from an
DRAFf - DO NOT CITE OR QUOTE
21
-------
intake of 10 ng/kg/day1" of arsenic. The authors point out that the risk for males and females
was within a factor of two indicating no gender difference and also that the potency of the
four internal cancers were comparable, differing only by a factor of four.
Table 4-4. Estimated values for Chen et al. (1992)
Disease
Liver cancer
Lung cancer
Bladder cancer
Kidney cancer
Sex
Male
Female
Male
Female
Male
Female
Male
Female
a
3.3 E-8
1.0 E-8
2.2E-10
2.2E-9
6.3E-11
1.4E-11
1.1E-12
2.4E-9
k
2.6
2.8
4.0
3.4
4.5
4.9
5.2
3.3
Potency
index
0.0043
0.0036
0.012
0.013
0.012
0.017
0.0042
0.0048
Quo et al. (1997) used tumor registry data and information provided by the
nationwide water-quality survey conducted during 1974-1976. Data on arsenic
concentrations were gathered from 243 township containing 11.4 million residents.
Regression models were created to compare the incidence of bladder, kidney, and
subcategories of those cancers diagnoses and six variables for the proportions of wells in
each of six categories of arsenic concentration (<0.05 ppm; 0.05-0.08 ppm; 0.09-0.16 ppm;
0.17-0.32 ppm; 0.33-0.64 ppm; and >0.64 ppm) in each township. Adjustments for age,
urbanization index, and annual number of cigarettes sold per capita were included in the
sex-specific models. For high arsenic concentration levels (>0.64 ppm), there was a strong
association with transitional-cell carcinomas of the bladder, kidney, ureter, and all urethral
cancers combined for both sexes. A significant association was not found for arsenic
concentrations below 0.64 ppm. Guo et al. (1997) did not present relative risk estimates so
the results cannot be directly compared with other studies. Additionally, no association
was found with cigarette sales. Yet, there was a positive association found with
urbanization. A crude incidence rate of 2.15 per 100,000 was determined for bladder
cancer, which is far below that of the comparable Asian population, suggesting under-
ascertainment of newly diagnosed bladder cancer in the voluntary national cancer registry.
In another study by Guo (2004), arsenic exposure was associated with an increase
in mortality from lung cancer. Arsenic was measured in 138 villages from a census survey
conducted by the government. Certificates of death that occurred in the village between
1 Chen et al. (1992) give the units as (Jg kg day; it is assumed that what is meant is ^g/kg/day, although
Hg/day is an alternate possibility.
DRAFT - DO NOT CITE OR QUOTE 22
-------
January 1, 1971 and December 31, 1990 were reviewed, and 673 male and 405 female
mortality cases of lung cancer were identified. Multi-variate regression models were
applied to assess association between arsenic levels in drinking water and mortality of lung
cancers using village as a unit. After adjusting for age, arsenic levels above 0.64 mg/L
were associated with a significant increase in the mortality of lung cancer in both genders,
but no significant effect was observed at lower levels. Post-hoc analyses confirmed such a
dose-response relationship.
Tsai et al. (1999) used SMRs for deaths occurring from 1971 to 1994 in the
blackfoot disease endemic area of southwest Taiwan. The median arsenic concentration in
the artesian well water was stated to be 0.78 ppm (range 0.25 to 1.14 ppm). Soil in this
area also contained high arsenic levels (median -7.2 mg/kg, range 5.3 to 11.2 mg/kg).
Comparison between both local (similar habits) and national (more stable age and sex
rates) reference groups demonstrated an association between arsenic and liver, respiratory,
bone, skin, bladder, and kidney cancer, as well as lymphoma in both males and females
(see Table 4-5 below). In addition, colon, leukemia and prostate cancers were increased in
males.
DRAFT - DO NOT CITE OR QUOTE 23
-------
Table 4-5. Statistically Significant (p<0.05) SMRs for Cancers in a Blackfoot
Disease Endemic Area of Taiwan Compared to Local and National Controls (Tsai et
al., 1999)
Cause
Liver
.Respiratory
Nasal
Laryngeal
Lung
Bone
Skin
Bladder
Kidney
Lymphoma .
Colon
Leukemia
Prostate
* 95% con
Local Reference SMR*
Males
1.83
(1.69-1.98)
3.00
(2.14-4.09)
1.78
(1.20-2.55)
3.10
(2.88-3.34)
2.46
(1.77-3.34)
4.83
(3.74-6.15)
8.92
(7.96-9.96)
6.76
(5.46-8.27)
1.63
(1.23-2.11)
1.49
(1.20-1.83)
1.34
(1.04-1.70)
2.52
(1.86-3.34)
fidence interval
Females
1.88
(1.64-2.14)
4.98
(3.33-7.15)
4.76
(2.53-8.15)
4.13
(3.77-4.52)
2.25
(1.56-3.14)
5.68
(4.41-7.21)
14.07
(12.51-15.78)
8.89
(7.42-10.57)
1.70
(1.18-2.37)
1.42
(1.13-1.76)
Not
significant
-
National Reference SMR*
Males
1.83
(1.69-1.98)
3.66
(2.61-4.98)
1.76
(1.19-2.52)
2.64
(2.45-2.84)
2.33
(1.67-3.16)
5.97
(4.62-7.60)
10.50
9.37-11.73)
6.80
(5.49-8.32)
1.42
(1.07-1.84)
1.35
(1.09-1.66)
1.34
(1.04-1.70)
1.96
(1.44-2.59)
Females
1.87
(1.64-2.14)
5.10
(3.41-7.32)
3.76
(2.00-6.43)
3.50
(3.19-3.84)
2.18
(1.51-3.05)
6.81
(5.29-8.63)
17.65
(5.70-19.79)
10.49
(8.75-12.47)
1.43
(1.00-1.99)
Not
significant
Not
significant
Morales et al. (2000) calculated excess lifetime risk estimates for the arsenic-
exposed populations in southwestern Taiwan using the same data set used by Wu et al.
(1989) and Chen et al. (1992) for residents from 42 villages. However, Morales et al.
limited their investigation to the study of bladder, liver, and lung cancer. It was assumed
that the number of deaths due to cancer follow a Poisson distribution. The models they
considered for the hazard function were of the form: H(x,t) = h0(t) x g(x), where h0(t)
represents the instantaneous hazard of dying of cancer at time t for the unexposed
population and g(x) is a modifying function to account for an exposure x. The authors
developed nine different generalized linear models (GLMs) by considering different forms
DRAFT - DO NOT CITE OR QUOTE
24
-------
for the age effect h0(t) (linear, regression, and spline), the dose effect g(x) (linear,
quadratic, exponential linear, and exponential quadratic), and dose transformation (linear,
logarithmic, and square root); a tenth model considered was the multistage-Weibull
(MSW). Each of the ten models was fit to the data to determine the constants implicit to
the form used for the age and dose effect functions; this process was completed three times,
once without consideration of a comparison population, once with considering all of
Taiwan as the comparison population, and once considering southwestern Taiwan as the
comparison population. The resultant models were compared and ordered using the
Akaike information criterion (AIC).
The authors completed their analysis by displaying, for each of the three cancer
types, the concentrations of arsenic that would result in a number of different risk
measures, including a 1% excess risk (ED01), a 5% excess risk (EDU5), and a margin of
exposure MOE0,(50) which represents the ratio of a point of departure, here taken to be
ED01, to an environmental exposure of interest, here taken to be 50 u.g/L. The authors state
that "the concentrations are reported in U.S. equivalent concentrations of arsenic in
drinking water, based on conversions that account for the average weight and average
water intake for a male living in the United States compared to a male living in Taiwan."
No further details are provided on how this conversion was carried out.
The risk statistics of the best-fitting models vary from model to model, as well as
according to whether a comparison population was used. Table 4-6 illustrates the range of
estimates for each of the three possible comparison populations.
DRAFT - DO NOT CITE OR QUOTE 25
-------
Table 4-6. Risk Statistics of Best Fitting Models from Morales et al. 2000
Concentrations (ug/L) without Comparison Population
ED01
MOE0,
(50)
ED05
Bladder
M
351-395
7.0-12.7
1181-
1439
F
244-365
4.9-7.3
796-828
Lung
M
227-364
4.5-7.3
1171-
1345
F
256-396
5.1-7.9
879-898
Liver
M
573-864
11.5-
17.3
-
F
657-824
13.1-
16.5
-
Combined
M
163-169
3.3-3.4 •
703-720
F
120-267
2.4-5.3
492-605
Concentrations (ng/L) with Taiwanese Comparison Population
ED01
MOBOI
(50)
ED05
Bladder
M
22-164
0.4-3.3
504-852
F
17-88
0.3-1.8
293-455
Lung
M
11-196
0.2-3.9
925-
1276
F
8-116
0.2-2.3
448-579
Liver
M
239-895
4.8-17.9
1089
F
331-551
6.6-11.0
-
Combined
M
3-106
0.1-2.1
111-544
F
2-53
0-1.1
54-273
Concentrations (ng/L) with Southwestern Taiwanese Comparison Population
ED0,
MOE0,
(50) '
ED05
Bladder
M
21-185
0.4-3.7
649-959
F
19-136
0.4-2.7
452-624
Lung
M
10-181
0.2-3.6
768-936
F
10-113
0.2-2.3
520-608
Liver
M
119-779
2.4-15.6
1608
F
455-597
0.2-2.3
-
Combined
M
3-98
0.1-2.0
93-506
F
2-55
0-1.1
63-284
Data from Tables 8, 9, and 10 of Morales et al. (2000); includes best-fitting of the GLM models plus
the MSW model. . ' '.
Kayajanian (2003) performed a re-analysis of the Taiwan data presented in NRC
(1999) (data stated to be from pgs 308-309, Table A10-1). In this analysis, villages were
grouped by fives: the lowest five, the next lowest five, etc. For each exposure group the
total number of lung, liver, and bladder cancers were put in the numerator, while the
denominator contained the totaled man-years associated with these cancers. The results
were expressed as fraction as cancers/1000 man-years. The lowest-dose group was found to
have a 3 times higher ratio than the next lowest group (1.65 compared to 0.53). The
difference was statistically significant with a p<0.001. The number of cancers/1000
woman-years was 3 times higher in the lowest dose group compared to the second and
third lowest exposure groups (1.62; 0.51, and 0.52, respectively; p<0.001). The arsenic
concentration in the lowest dose group was 10 to 32 ug/L, 42 to 60 u^g/L in the second
lowest dose group, and 65 to 110 ug/L in the third lowest dose group. In Morales et al.
(2000), the lowest dose group was considered 0 to 100 ug/L, which would not allow for the
same results to be noted.
DRAFT - DO NOT CITE OR QUOTE
26
-------
Kayajanian (2003) ftirther tested the observation using mortality from cancer (lung,
liver, bladder, kidney, colon, and melanoma) in Utah with four exposure groups (0 to <25;
25 to <75; 75 to <150; and 150tol75 |ag/L. Although there were no significant differences
in the males, the second and third lowest doses were lower than the lowest dose group
(2.68, 1.99, and 1.83 cancers per 100 people, respectively). In females, the 25 to <75 ug/L
group had results significantly lower than the 0 to <25 ng/L group (2.63 compared to 0.642
cancers per 100 people, respectively). Data by Cuzick et al. (1982) was also re-evaluated
and demonstrated that regardless of time from exposure (subjects were intentionally dosed
for medicinal purposes), the <500 mg dose group had fewer cancer mortalities than
expected. As the doses increased, however, the observed outweighed the expected.
Chiou et al. (2001) conducted a cohort study to evaluate the association between
transitional cell carcinoma (TCC) of the bladder and arsenic exposure from drinking water.
The study population consisted of 8,102 residents (4,056 men and 4,046 women) in
northeastern Taiwan, where each subject's individual arsenic exposure was estimated
based on the arsenic concentration in his or her own well water. A total of 3,901
households provided well water samples (85.1%). Arsenic concentrations ranged from
undetectable (<0.15 ng/L) to 3.59 mg/L, However, 1136 subjects in 685 households no
longer had wells and provided no data on arsenic exposure. Cancer occurrence was
ascertained from annual personal interviews, and through hospital records, national death
certification, and cancer registries. The study was initiated in October 1991 and follow-up
concluded on December 31, 1996. During the follow-up period, there were 18 new cases
of urinary tract cancers, including 11 TCCs. There were 15 cases with urinary cancer and
10 cases with TCC that had data on arsenic exposure. Standardized incidence ratios (SIRs)
were estimated from the general population of Taiwan. The exposure data were stratified
into 3 duration groups (<20.0, 20.1-29.9, and ^40.0 years) and 4 arsenic levels (si0.0,
10.1-50.0, 50.1-100.0, and > 100.0 (Ag/L), Cox's proportional hazards regression analysis
was conducted to estimate multivariate-adjusted relative risks and 95% confidence
intervals. Overall, there was a significantly increased incidence of urinary cancers of the
study cohort compared to the general population (SIR=2.05, 95% CI 1.22-3.24). A
significant dose-response relationship was also observed after adjustment for age, sex, and
cigarette smoking. The multivariate-adjusted relative risks for TCC were 1.9 (95% CI 0.1 -
32.5), 8.2 (95% CI 0.7-99.1), and 15.3 (95% CI 1.7-139.9), for arsenic concentrations of
10.1-50.0, 50.1-100, and > 100 ug/L, respectively, compared to the reference level of s 10.0
ug/L. The multivariate-adjusted relative risks for total urinary cancers were 1.5 (95% CI
0.3-8.0), 2,2 (95% CI 0.4-13.7), and 4.8 (95% CI 1.2-19.4), for arsenic concentrations of
10.1-50:0, 50.1-100, and > 100 ug/L, respectively, compared to the reference level of <;10.0
Hg/L. The duration of exposure was not associated with urinary tract cancers. Since each
household in this study had its own drinking water well, individual exposure estimates
were much more precise than previous studies in other endemic areas of Taiwan.
Although there were a number of studies performed on populations in Taiwan,
there also were a number of studies that examined other geographic regions and found
associations between arsenic and cancer. There was a significant (p<0.001), dose-related
DRAFT - DO NOT CITE OR QUOTE 27
-------
increase in mortality for both males and females in C6rdoba, Argentina for low-, medium-,
and high-exposure groups (Hopenhayn-Rich et al., 1996a). The exposure levels for the
groups were not specified; however, the average level for the high exposure was 178 ug/L.
Water usage data also was not available. In males, respective cancer mortalities (and 95%
CI) were: kidney, 0.87 (0.66-1.10), 1.33 (1.02-1.68), and 1.57 (1.17-2.05); lung, 0.92 (0.85-
0.98), 1.54 (1.44-1.64), and 1.77 (1.63-1.77); and bladder, 0.80 (0.66-0.96), 1.42 (1.14-
1.74), and 2.14 (1.78-2.53). In females the respective mortalities (and 95% CI) were:
kidney, 1.00(0.71-1.37), 1.36 (0.94-1.89), and 1.81 (1.19-2.64); lung, 1.24(1.06-1.42),
1.34 (1.12-1.58), and 2.16 (1.83-2.52); and bladder, 1.21 (0.85-1.64), 1.58 (1.01-2.35), and
1.82(1.19-2.64). SMRs were established for deaths occurring between 1986 and 1991 in
regions that were determined to have high, medium, or low arsenic in the drinking water up
through the 1970s (in the 1970s aqueducts from rivers low in arsenic were built for water
consumption) compared to all of Argentina.
The SMRs for liver cancer in all three exposure levels in males and females were
significantly elevated, but the dose-response trend was less apparent (males, 1.54, 1.80, and
1.84 for low, medium and high exposure, respectively, p=0.06; females, 1.69, 1.87, and
1.92, respectively, p=0.14). The SMR for skin cancer was elevated in high exposure
females (i.e., 2.78, 95% CI 1.61-4.44) and all male exposure groups (i.e., 2.04, 95% CI
1.38-2.89; 1.49, 95% CI 0.83-2.45; and 1.49, 95% CI 0.71-2.73 for low, medium, and high
exposures, respectively), but did not follow a dose-related trend. Since there was only a
small number of deaths from skin cancer and most skin cancer is nonfatal, this may
underestimate the relationship between arsenic and skin cancer in this study. There was no
relationship observed between arsenic exposure and stomach tumors. Although the study
could not account for any possible confounding factors due to lack of data on individuals,
the study attempted to account for smoking by examining chronic obstructive pulmonary
diseases (COPD) since 80% of COPD mortality is associated with smoking. Since there
was no association between COPD and arsenic, it was concluded that smoking was not a
significant confounder.
Cuzick et al. (1992) studied the excessive mortality of bladder cancer in 478
patients medically treated for skin complaints, malaria, anemia, epilepsy or anxiety with
potassium arsenite for various lengths of time (2 weeks to 12 years) during 1945-1969 in
England and Wales. There was an excessive occurrence of bladder cancer deaths than
expected (p <0.05; 5 observed/1.6 expected) adjusting for age-, sex-, and calendar-year.
No bladder cancer deaths were found in the first 5 years. The SMR for patients receiving
doses greater than 500 mg of arsenic was calculated to be 5.00 (95% CI 2,0-15). There
were no excessive lung cancer deaths observed. Smoking was ruled out as a possible
factor in the excess of bladder-cancer deaths as the SMR for bladder cancer in relation to
smoking was 0.91 (95% CT 0.74-1.1) using circulatory disease as indicators.
A cohort of 113 people exposed to arsenic concentrations greater than 1.0'mg/L by
industrially contaminated drinking water for 5 years in villages of Niigata Prefecture, Japan
was used by Tsuda et al. (1995) to determine the SMRs in Table 4-7. The expected
DRAFT - DO NOT CITE OR QUOTE 28
-------
number of deaths were based on sex-, age-, and cause-specific mortality from 1960 to
1989.
Table 4-7. SMRs in Japanese Cohort (Tsuda et al. 1995)
Cancer Observed deaths/Expected deaths SMR
Urinary-tract 3/0.10 31.18 (95% CI = 8.62-91.75)
Lung 8/0.51 15.69 (95% CI = 7.38-31.02)
Liver 2/0.28 7.17 (95% CI = 1.28-26.05)
Uterine 2/0.15 13.47 (95% CI = 2.37-48.63)
Bladder cancer risks were determined from 117 bladder cancer cases and 266
population-based controls in Utah after relatively low-level exposure to arsenic in drinking
water (Bates et al., 1995). Study participants were surveyed in 1978 for the National
Bladder Cancer Study sponsored by the National Cancer Institute. Based on residential-
history information and water-sampling information from public water sources, individual
exposures to arsenic in drinking water were determined, and it was reported that 92% (81
towns) of the towns in Utah had arsenic concentrations of less than 10 [ig/L; only one town
had concentrations greater than 50 ug/L. There were no associations between bladder
cancer and arsenic exposure levels. Odds ratios were calculated using cumulative dose
(defined as the product of micrograms per liter and years of exposure) and were 1.00; 1.56,
95% CI 0.8-3.2; 0.95, 95% CI 0.4-2.0; and 1.41, 95% CI 0.7-2.9 for a relative lifetime
exposure of less than 19 mg, 19 to 33 mg, 33 to 53 mg, and greater than 53 mg,
respectively. The study suggests that smokers had increased risk in time-window analyses
of exposures for years 20-29 (16 cases in the highest exposure group) and 30-39 years (9
cases in the highest exposure group) prior to the surveys.
Arsenic exposure was related to an increased risk of mortality in both males and
females 30 years of age or older from Northern Chile when compared to the National
mortality data (Smith et al., 1998). The SMRs (and 95% CI) for cancer mortalities in males
and females were: bladder, 6.0 (4.8-7.4) and 8.2 (6.3-10.5), respectively (p<0.001); lung,
3.8 (3-5-4.1) and 3.1 (2.7-3.7), respectively (p<0.001), kidney, 1.6 (1.1-2.1) and 2.7 (1.9-
3.8), respectively (p=0.012 and O.001, respectively); and skin 7.7 (4.7-11.9) and 3.2 (1.3-
6.6), respectively (p<0.001 and =0.016, respectively). There was no increased risk of
mortality from liver cancer in either males or females (SMR = 1.1 with a 95% CI of 0.8-1.5
for both; p>0.3) associated with arsenic exposure. Concentrations of arsenic in drinking
water were well-documented and had been high in all major population centers of Region
II, especially before 1975. The population-weighted average in the years 1950-1974 was
420 ug/L with a maximum of 870 ug/L measured in Antofagasta (the largest city) between
1955 and 1969. The arsenic exposure in all cities or towns (8 examined) exceeded the
standard of 50 ng/L from 1950 until 1990. The mortality rate was obtained during the
period of 1989 and 1993. The reported number of excessive lung cancer deaths among
men (400.8) was about 5 times that for bladder cancer (77.5). Among women, the number
of excessive lung cancer deaths (105) was about twice that for bladder cancer (56.2). The
study examined the SMRs by age and demonstrated that the greatest excessive risk for lung
DRAFT - DO NOT CITE OR QUOTE 29
-------
cancer and COPD occurred in the 30-39 age group for both males and females
(observed/expected = 5.0 and 60.0, respectively), which would most likely be associated
with exposure during childhood. In addition, females had the greatest excessive risk of
death from liver (observed/expected = 2.5) and kidney (observed/expected = 30) cancers in
this age group. The study had limited information on smoking and could not separate the
results by age or sex, but did not demonstrate any regional differences in smoking habits
that may explain the increased cancer risk. The study estimates that 9.3% of all male and
4.9% of all female deaths from bladder, kidney, lung, and skin cancer may be attributable
to arsenic exposure. Although this study has the limitations of an ecological study design,
several strengths were identified. There were well-defined population exposures,
compared with several other exposed populations, and mortality to COPD was low,
indicating low smoking rates in this population.
Kurttio et al. (1999) established an association between arsenic and bladder cancer
by evaluating the relative risks (RR) in 61 bladder cancer cases diagnosed between 1981
and 1995 in Finland who drank water from drilled wells between 1967 and 1980 compared
to an age and sex balanced reference cohort of 275 subjects. The dose-dependent
relationship between arsenic and bladder cancer was only noted when exposure was
calculated as beginning of the use of the well until 2 years before diagnosis (short latency;
RR ratios of 1, 1.53 (95% CI 0.75-3.09), and 2.44 (95% Cl 1.11-5.37) for concentrations in
the water of <0.1, 0.1-0.5, and *0.5 ng/L, respectively; RR ratios of 1,1.34 (95% CI 0.66-
2.69), and 1.84 (95% CI 0.84-4.03) for daily doses of arsenic of <0.2, 0.2-1.0 and * 1.0
(j,g/day). When exposure was considered from the beginning of the use of the well until 10
years before diagnosis (long latency), there was no association with an increase in risk for
bladder cancer. Smoking also was associated with bladder cancer with a possible
synergistic effect observed between arsenic and smoking on the risk of developing bladder
cancer. There was no relationship between kidney cancer and arsenic exposure.
Misclassification of exposure was possible during the study due to recall bias since the
study was interested in water consumption from the 1970s. In addition, the study did not
attempt to measure arsenic intake from dietary sources or nutritional factors which may
effect the metabolism and carcinogenicity of arsenic.
Chiou et al. (2001) performed a cohort study to examine the relative risk of
transitional cell carcinomas in relation to ingested arsenic of 8,102 residents in northeastern
Taiwan. Estimations of individual exposure was based on the arsenic concentration in
their well water and questionnaire information on duration of consumption. The
occurrence of urinary tract cancers were ascertained by follow-up interviews, community
hospital records, national death certification profile, and cancer registry profile.
Multivariate-adjusted relative risk and 95% confidence intervals (CI) were estimated using
Cox proportional hazards regression analysis. There was a dose-dependent trend for both
transitional cell carcinomas (p<0.05; 1, 1.9, 95% CI 0.1-32.2; 8.1, 95% CI 0.7-98.2; and
15.1, 95% CI 1.7-138.5 for arsenic concentrations in the well water of 0-10.0, 10.1-50.0,
50.1-100.0, and > 100.0, respectively) and all urinary cancers (p<0.01; 1, 1.6, 95% CI 0.3-
8.4; 2.3, 95% CI 0.4-14.1; and 4.9, 95% Cl 1.2-20.0, respectively).
DRAFT - DO NOT CITE OR QUOTE 30
-------
A cohort of residents from Millard County, Utah was assembled using historical
documents of the Church of Jesus Christ of Latter-Day Saints for establishing an
association between arsenic in the drinking water and mortality outcome (Lewis et al.,
1999). SMRs were calculated using residence history and median drinking water arsenic
concentrations. Lewis et al. (1999) evaluated the SMR by low (<1000 ppb-year), medium
(1000-4999 ppb-year), and high (sSOOO ppb-year) exposures per sex, as well as all
exposures combined by sex. Arsenic exposure took into consideration the number of years
living in a certain area and the concentration of arsenic for that specific area to obtain a
cumulative estimate of exposure. The level of arsenic in the drinking water for the
different towns ranged from a mean of 18.1 to 190.7 ppb (ng/L). The low dose was
selected based on 20 years of exposure for most cancers to manifest with exposure to 50
ppb for a cumulative arsenic exposure of 1000 ppb-years. Although all endpoints were
examined, only cancer endpoints are referred to at this point.
Lewis et al. (1999) conducted a retrospective cohort mortality'study of 4,058
residents of Millard County, Utah. Drinking water in this region is derived from wells
where arsenic concentrations range from undetectable up to a few hundred micrograms per
liter. Lewis et al. (1999) calculated SMRs using residence history and median drinking
water arsenic concentrations (i.e., low, <1000 ppb-year; medium, 1000-4999 ppb-year; and
high, 2:5000 ppb-year), exposures by sex as well as all exposures combined by sex.
Arsenic exposure took into consideration the number of years living in a certain area and
the concentration of arsenic for that specific area to yield a cumulative estimate of
exposure. For several causes of death, statistically significant increases in SMRs occurred
in a generally uneven pattern based on exposure level. Such as the significant (psO.05)
associations between arsenic and mortality overall in males from prostate cancer (i.e.,
SMR=1.07, 1.70, and 1.65 for low, medium, and high exposure). Females had a
significant association between melanomas and arsenic exposure. The SMR for all females
was 1.82 (95% CI, 0.50-4.66), and was 5.30 (no Cl provided) in the low exposure group.
Results were not provided for the medium and high exposure groups. While there were no
other significant associations between specific cancers and arsenic exposure in females,
there were increased SMR values for biliary passage and liver cancer (i.e., SMR=1.42 all
females; 95% Cl 0.57-2.93), kidney cancer (i.e., SMR=1.60 all females; 95% CI 0.44-
4.11), and all other malignant neoplasms (i.e., SMR=1.34; 95% CI 0.84-2.03). This study
did not look at any other possible confounding factors, such as smoking, on the mortality.
However, smoking was rare in this population due to the churches prohibition of the use of
tobacco, alcohol, or caffeine. NRC (2001) details several limitations to this study mainly
on the exposure estimates. In addition, comparison rates used in the analysis were for the
state of Utah with the study cohort composed of Mormons who have strict religious
prohibitions. Although smoking rates in the state of Utah are low (12-13%), this rate is
expected to increase the rates of several cancers (e.g., lung). That fact and the rural setting
of the cohort were possible contributors to the deficits observed in SMRs for urinary and
pulmonary cancers. .
DRAFT - DO NOT CITE OR QUOTE 31
-------
A case-control study using patients diagnosed with lung cancer between 1994 and
1996 in northern Chile and frequency-matched hospital controls provides evidence that
consumption of inorganic arsenic from drinking water is associated with human lung
cancer (Ferreccio et al., 2000). Control selection was considered a major weakness by the
study authors and included two types of control. First, patients were selected who had been
diagnosed with cancers that are generally not suspected to be related to arsenic and second
patients were selected with diagnosis other than cancer. This limitation of the controls
likely biased the estimates towards lower odds ratios. There was a clear dose-related
increase in the odds ratio when adjustments were just made for age and sex (1; 1.5, 95% CI
0.4-4.6; 4.0, 95% CI 1.4-12.1; 4.6, 95% CI 2.2-10.0; and 8.0, 95% CI 3.8-17.0 for arsenic
water concentrations of 0-10, 10-29, 30-49, 50-199, and 200-400, respectively) or when
adjustments also included smoking, socioeconomic status, and working in copper smelting
(1; 1.6, 95% CI 0.5-5.3; 3.9, 95% CI 1.2-12.3; 5.2, 95% CI 2.3-11.7; and 8.9, 95% CI 4.0-
19.6, respectively). The study also demonstrated a greater than additive effect between
arsenic exposure and smoking (odds ratios in never smoked for arsenic concentrations s49
[considered the referent category], 50-199, and ;>200 ug/L were 1; 5.9, 95% CI 1.2-40.2;
and 8.0, 95% CI 1.7-52.3, respectively; odds ratios in ever smoked group were 6.1, 95% CI
1.31-39.2; 18.6, 95% CI 4.13-116.4; and 32.0, 95% CI 7.22-198.0, respectively). In
addition, peak exposure in the cases mainly occurred 20-40 years prior to diagnosis. This
study has individual estimates of exposure on all subjects for more than 40 years. Because
of the time to onset of lung cancer, this is a study strength for causal association. Other
strengths include an acceptable response rate, unbiased ascertainment of exposure,
information of individual data on potential confounding factors for lung cancer, appropriate
analysis of study data, and adequate study size.
A number of studies on the association between arsenic exposure and cancer have
appeared in the peer-reviewed literature after the NRC (2001) report was published. These
are discussed below.
Tucker et al. (2001) analyzed data describing prevalent skin cancer from a cross-
sectional study in a region of Inner Mongolia with increased concentrations of arsenic in
drinking water. The 1992 study examined a total of 3,179 persons in three villages and the
well-water-use histories for these individuals. Water samples were collected from 184 of
the 187 local wells and analyzed for arsenic content. The median age of the participants
was 29 years with an average well-use history of 25 years. Arsenic in drinking water
ranged from below detection (10 ng/L) to 2,000 |ig/L. Skin cancer was observed in eight
subjects. Several statistical models (frequency weighted, simple linear regression, hockey
stick, and maximum likely estimate) were used to analyze data. Two measures of exposure
were used with each model. A dose-response relationship was found for skin cancer.
Karagas et al. (2001) measured arsenic levels in toenail clippings from subjects
with basal cell carcinoma (587), squamous cell carcinoma (284), or controls (524). The
study demonstrated that there was only an increase in squamous cell carcinomas (odds
ratio= 2.07 with a 95% CI of 0.92 to 4.66) and basal cell carcinomas (odds ratio=l .44 with
DRAFT - DO NOT CITE OR QUOTE 32
-------
a 95% CI of 0.74 to 2.81) in the subjects with the highest (i.e., 0.345-0.81 jig/g) levels of
toenail arsenic. Toenail arsenic levels from 0.009-0.344 ug/g did not have elevated odds
ratios. However, arsenic measured in toenail clippings may result from external exposure
as well as internal exposure and typically relates exposure from a two week period
occurring approximately a year prior to sampling. Therefore, the exposure measured in the
toenail clippings did not occur during the critical period for development of skin cancer
(NRC.2001).
A re-analysis of the Wu et al. (1989) and Chen et al. (1992) data sets by Lamm et
al. (2003) used exposure to arsenic as a continuous variable, while simultaneously
stratifying on the type of well used as a drinking water source (i.e., artesian, shallow, or
mixed). Artesian wells (14 villages) all had medians > 325 ng/L, while shallow wells (19
villages) had medians < 325 jig/L and mixed wells (9 villages) had medians both greater
than and less than 325 jig/L. This approach determined that there was no increase in
bladder cancer mortality rates below 400 ng/L in any of the well types, and that a
significant dose-response relationship between arsenic exposure and bladder cancer rates
was present in only the artesian wells (median levels ranging from 350-934 jig/L). The
median arsenic levels for the shallow and mixed wells ranged from 10-717 u.g/L. The
authors suggest two possible reasons for their findings: 1) arsenic is acting as a high-dose
carcinogen, or 2) a co-carcinogenic factor is present only in artesian wells and not the
shallow/mixed wells. However, this study has several limitations of its own. First the
classification into village well type was based on median arsenic concentrations such that
well type and arsenic concentration are not independent variables. In addition, the study
did not consider age or smoking status as factors since the data were not available. These
limitations were suggested by the U.S. EPA (2004a). In addition, the U.S. EPA (2004a)
suggests that the use of linear regression may not be appropriate because bladder cancer
rates are not normally distributed, and that a Poisson regression would have been more
appropriate.
Because Morales et al.'s (2000) analysis of the Taiwan data suggest that excessive
cancer mortality may occur in many populations where the drinking water standard for
arsenic is set at 50 u.g/L (the current drinking water standard for arsenic in the U.S.) and
Lamm et al. (2003) suggest that this was not the case, Lamm et al. (2004) performed an
ecological study examining the relationship between bladder cancer mortality in white
males from 133 counties in the United States and arsenic exposure through drinking water.
The study was designed to be analogous to the Wu et al. (1989) Southwest Taiwan Study.
Arsenic exposure from groundwater arsenic data collected by the U.S. Geological Survey
(USGS) for 133 US counties. These counties exclusively used groundwater as a source of
drinking water, had arsenic concentrations ^3 ug/L, and reported at least one white male
bladder cancer death. White male bladder cancer mortality data from 1950-1979 were
extracted from National Cancer Institute (NCI) and EPA (NCI/EPA 1983): The study
included over 4500 US white male bladder cancer deaths, and the data suggest that there
was no increase in bladder cancer mortality in the 3-60 u.g/L range of arsenic exposure.
Standardized mortality ratios (SMR) that were stratified by median arsenic groundwater
DRAFT - DO NOT CITE OR QUOTE 33
-------
concentrations, ranged from 0.95 (95% CI, 0.89-1.01) in the lowest exposure group (3.0-
3.9 ug/L) to 0.73 (95% CI, 0.41-1.27) in the highest exposure group (50.0-59.9 ug/L). The
composite totals found an SMR of 0.94 with 95% CI of 0.90-0.98.
A major limitation in the Lamm et al. (2004) study was the lack of information
about the counties or specifics indicating that the counties were similar in nature. If the
counties with higher concentration of arsenic in their groundwater had younger males or
better medical facilities, results would tend to show lower mortality in these areas. Other
potential confounding factors that were not addressed include: diet, use of alternative water
sources, variation of intake, age, duration of exposure, occupation, income, smoking status,
and migration. In addition, exclusion of counties without bladder cancer mortality may
also bias the results. U.S. EPA (2004a) discusses these limitations and also states that the
authors overanalyzed the data and that the method of calculating the expected number of
cases was unconventional.
Steinmaus et al. (2003) conducted a case-control study of bladder cancer in relation
to arsenic in the drinking water in six counties of western Nevada and Kings County in
California, selected because the counties contained the largest population historically
exposed to approximately 100 jig/L of arsenic in their drinking water. The 181 cases
presented in the article were 20-85 years old with primary bladder cancer diagnosed
between 1994 and 2000. Controls (328 identified, 248 of which were interviewed) were
matched to cases by 5-year age groups and gender and were selected by random digit
dialing (for controls under 65) or using the Health Care Financing Administration (for
controls over 65 years old). Arsenic exposure was assessed for each subject by linking the
residence within the study area to a water arsenic measurement for that residence. Fluid
intake (L/day), estimated by interviewing cases and controls or their relatives, was
multiplied to the arsenic concentration (ug/L) to determine average daily arsenic intake
(ug/day).
Overall, Steinmaus et al. (2003) did not find an association between bladder cancer
risk and arsenic exposure, with an adjusted odds ratio of 0.94 (95% CI, 0.56-1.57; p=0.48)
for intakes greater than 80 ug/day. The adjusted odds ratios were not significant when the
data were stratified by exposures using the highest 1-year, 5-year, 20-year averages or .
cumulative exposures, or by the length of exposures (5 years, 20 years or 40 years).
However, for smokers with arsenic exposures of 40 or more years to > 80 ug/day (median
intake 177 |ag/day), the odds ratio was 3.67 (95% CI, 1.43-9.42; p<0.01). These results
support the hypothesis that arsenic and cigarette smoke act synergistically in causing
bladder cancer, and suggest that the latency of arsenic-caused cancer may be greater than
40 years. This study had several limitations, particularly in non-matching characteristics of
the cases and controls involving income, education, and smoking. Other limitations
involved possible exposure misclassification (because more next-of-kin of cases were
interviewed than controls) and lack of exposure data for 31% of the private wells
(accounting for 11% of the person-years). Because of the long latency period for bladder
cancer and the small number of cases, it is unlikely that a statistically significant risk for
DRAFT - DO NOT CITE OR QUOTE 34
-------
bladder cancer would be detected in the low levels of arsenic exposure among never
smokers. U.S. EPA (2004) reported that for a moderate (2.0-3.0 relative risk) association,
a sample size of 178,000 to 535,000 would be needed. In addition, cases that chose not to
participate in the study were more likely to live in areas of known arsenic contamination
than nonparticipating controls. The authors did not have arsenic levels for 31% of the
private wells, which accounted for 11% of the person-years. The authors gave this group a
value of zero for the wells. Zero exposure also was applied to residents living outside the
study area and those using bottled water or filters to remove arsenic. In doing so they bias
the results towards the null effect. Because of the long latency period for bladder cancer
and the small number of cases, it is unlikely that a statistically significant risk for bladder
cancer would be detected in the low levels of arsenic exposure among never smokers. The
sample size needed to detect a very weak (1.15 relative risk) association would be
17,045,000. U.S. EPA (2004) reported that for a moderate (2.0-3.0 relative risk)
association, a sample size of 178,000 to 535,000 would be needed.
Tollestrup et al. (2003) conducted a retrospective cohort study to determine whether
childhood exposure to ambient arsenic was associated with increased mortality rates. The
cohort was comprised of children who had lived within 2 miles of a copper smelter and
arsenic refinery (American Smelting and Refining Company) in Ruston, Washington, for at
least 2 years from 1907-1932. The subjects were identified from school census records,
and included 1,827 boys and 1,305 girls with an age limit of 14 years. Exposure intensity
was calculated as the total number of days spent at a residence within 1 mile of the smelter
stack, and grouped by the number of years spent at the residence: 0 s 1.0 year, 1.0-3.9
years, 4.0-4.9 years, and * 10.0 years. A total of 3,336 potential subjects were identified,
and 196 were excluded because they had worked at the smelter. Crude mortality rates were
based on person-years of follow-up, and calculated for 10 general causes of death. The
highest crude mortality rate for boys was for ischemic heart disease in all exposure
intensity groups, but no evidence of a dose-response relationship was found. The 2nd
highest mortality rate for boys was for malignant neoplasms, with a range of 12.5/10,000
person-years to 21.9/10,000 person-years. A dose-response was observed only for the
mortality rate for "external causes," such as motor vehicle accidents. Cox proportional
hazard ratios adjusted for year of birth found only one exposure group (* 10.0 years) for
which the mortality rations were significantly higher than 1.00. These included all causes
of death (1.52, 95% CI 1.23-1.86), ischemic heart disease (1.77, 95% CI 1.21-2.58), and
external causes (1.93, 95% CI 1.03-3.62). Although girls also had the highest crude
mortality rates for malignant neoplasms and ischemic heart disease, no dose-response
relationships were observed. This study did not find consistent patterns of adverse health
effects from childhood exposure to ambient arsenic at levels much lower than occupational
settings.
A deficiency of the Tollestrup et al. (2003) study was the truncation of the study
period to 1932 which could result in exposure misclassification. Other limitations include
ambiguous exposure data (exposures to arsenic were not chronic and were unknown since
air and soil levels were not quantified), poor follow-up (34.7% of boys and 46.5% of girls
DRAFT - DO NOT CITE OR QUOTE 35
-------
were not found after their last date of exposure), the use of crude mortality rates, and lack
of information on smoking within the cohort and on family members who worked at the
smelter and could have brought arsenic into the household.
Chen et al. (2004) investigated the relationship between ingested arsenic and lung
cancer and the effect of smoking on the relationship. A total of 2,503 residents in
southwestern and 8,088 residents in northeastern Taiwan were followed for an average
period of 8 years. These were areas where residents had been drinking well water
contaminated with high concentrations of arsenic until the establishment of public water
systems. Questionnaires were administered to all participants in the study eliciting
information on residential and occupational history, history of drinking well water,
cigarette smoking and alcohol consumption. Water measurements taken in the 1960s of
shared artesian wells in the southwestern area were used in conjunction with information
derived from the questionnaire to derive an average arsenic concentration for each
participant, which was used as an exposure metric in subsequent analysis. Average arsenic
concentrations for participants in the northeastern region, who derived their drinking water
from shallow wells, were determined by direct measurement of individual wells. The
incidence of lung cancer was ascertained from national registry data for the period the
period January 1985-December 2000. During the follow up period of 83,783 person-years,
139 lung cancer cases were diagnosed.
After adjusting for cigarette smoking and other risk factors such as age, alcohol
consumption, and years of schooling, a significant (p O.001) increasing trend in lung
cancer was shown to result from increasing average levels of arsenic in well water. With
levels <10 ug/L as the referent, relative risks (with 95% confidence intervals) for those
consuming drinking water with arsenic concentrations of 10-99, 100-299, 300-699, and
*700 ug/L, were respectively, 1.09 (0.63-1.91), 2.28 (1.22-4.27), 3.03 (1.62-5.69), 3.29
(1.60-6.78). It was further shown that 32% to 55% of lung cancer cases were attributable
to both arsenic exposure and cigarette smoking. The synergism was shown to be additive;
multiplicative interaction was not statistically demonstrated.
NRC (2001) made a series of recommendations for conducting the dose-response
assessment:
• Since the formula used by Morales et al. (2000) to compute ED0| did not
easily accommodate the incorporation of the baseline cancer risk based on
the U. S. population, the ED calculation should be done according to a
formula presented in Appendix II of the BEIR IV analysis of lung cancer
associated with radon exposure (NRC 1988).
Morales et al. (2000) estimated risk using no comparison population, the
entire Taiwanese population as a comparison population, and the
southwestern Taiwanese population as a comparison population. The NRC
recommended using a comparison population in risk estimation.
DRAFT - DO NOT CITE OR QUOTE 36
-------
• Because of the biological plausibility and consistency with other approaches
to quantitative risk assessment, the additive Poisson model with a linear
dose should be used in the risk assessment.
• Since the dietary intake of arsenic in the Taiwanese population is higher
than in the United States population, a constant concentration of arsenic
should be added to the exposure rates for all individuals in the study
villages. The Committee recommended use of a background rate in food of
30 ng/day, assuming a 50 kg weight for a Taiwanese person.
With these assumptions, the Subcommittee used the southwestern Taiwanese data
set (Chen et al., 1985, 1982; Wu et al., 1989) to perform maximum likelihood estimates of
the excess lifetime risk (incidence per 10,000 people) of lung and bladder cancer for
populations exposed to various concentrations of arsenic in drinking water. National
Taiwan data was used as background incidence, The results are shown in Table 4-8, where
it was further assumed that the typical U.S. resident weighs 70 kg, compared with 50 kg for
the typical Taiwanese, and that the typical Taiwanese drinks just over 2 L of water per day,
compared with 1 L per day in the United States.
Table 4-8. NRC 2001 Maximum Likelihood Estimates of Excess Lifetime Risk
(per 10,000 people)
Arsenic
Concentration
(ngflO
3
5
10
20
Bladder Cancer
Females
2.3
3.8
7.5
15
Males
2.0
3.2
6.8
13
Lung Cancer
Females
1.8
3.0
6.2
12
Males
1.7
3.0
6.1
12
Confidence Intervals not provided.
4.2. PRECHRONIC AND CHRONIC STUDIES AND CANCER
BIOASSAYS IN ANIMALS - ORAL AND INHALATION
4.2.1. Prechronic and Chronic Studies
Wei et al. (1999 and 2002) demonstrated that 10-week-old male F344/DuCrj
rats (36/group) administered 12.5, 50, or 200 ppm dimethylarsenic acid (DMA; a major .
metabolite of inorganic arsenic) in their drinking water for 104 weeks had no affect on the
morbidity, mortality, body weights, hematology, or serum biochemistry. Reductions in
electrolyte concentrations in the urine were related to an increase in urinary volume
resulting from an increased water consumption in the 50- and 200-ppm groups. There was
no difference in the urinary pH between control and treated rats.
DRAFT - DO NOT CITE OR QUOTE
37
-------
4.2.2. Cancer Bioassays
Cancer bioassays with inorganic arsenic have obtained negative results with
mice, rats, hamsters, rabbits, beagles, and cynomologus monkeys (for review see Kitchin,
2001; NRC, 1999). However, the following studies have observed increases in tumors in
animals exposed to arsenic.
Transplacental-Mice
Timed pregnant female C3H/HeNCr (C3H) mice (10/group) were administered
0 (control), 42.5, or 85 ppm arsenite in their drinking water ad libitum from Day 8 to Day
18 of gestation (Waalkes et al, 2003). Strain and doses used in the experiment were
determined through preliminary short-term testing that determined C3H mice to be the
most sensitive to arsenic toxicity of the three strains tested (i.e., C3H, C57BL/6NCr, and
B6C3Fl/NCr), and the preliminary test indicated that a dose of 100 ppm was unpalatable
and resulted in approximately 10% reduced growth in the offspring. The doses used in this
study did not affect maternal water consumption or body weight in the dams. It was
estimated that the pregnant females consumed 9.55 to 19.13 mg arsenic/kg/day for a total
dose of 95.6 to 191.3 mg arsenic/kg.
Offspring were weaned at 4 weeks and received no additional exposure to
arsenic. Male and female offspring (25/sex/group) were observed for the next 74 or 90
weeks, respectively. Males were sacrificed at 74 weeks due to high mortality in the high-
dose group beginning at 52 weeks. Both the 42.5- and 85-ppm males had a significant
increase in the incidence of hepatocellular carcinomas (12.5% in controls vs. 38.1% in
42.5-ppm and 60.9% in 85-ppm groups) arid adrenal cortical tumors (37.5% in controls vs.
66.6% in 42.5-ppm and 91.3% in 85-ppm groups), which followed a significant (p^0.001),
dose-related trend. In addition, the 85-ppm group had a significant increase in the
multiplicity (tumor/mouse) for both hepatocellular carcinomas (0.13, 0.42, and 1.30,
respectively) and adrenal tumors (0.71, 1.10, and 1.57, respectively) which also had a
significant (psO.02), dose-related trend. Although there was no differences in the
incidence of hepatocellular adenomas in males, the multiplicity of hepatocellular adenomas
(0.71, 1.43, and 3.61, respectively) followed a significant (pO.OOOl), dose-related trend.
Males and females had an increase in lung tumors (8.0%, 13.0%, and 25.0%,
respectively, in females, 0%, 0%, and 13.0%, respectively, in males), which followed a
significant (p^O.03), dose-response trend. In addition, females had increases in the
incidence of benign ovarian tumors, which reached statistical significance in the 85-ppm
group. Although a significant increase was not observed in malignant ovarian tumors, the
total incidence (benign plus malignant) of ovarian tumors was significant in the 85-ppm
group and followed a significant (p=0.015), dose-related trend (8% in control vs. 26% in
42.5-ppm and 37.5% in 85-ppm groups). There was an increase in uterine tumors that was
not significant and did not follow a dose-response, but was accompanied by a significant
(p=0.0019), dose-related increase in hyperplasia with a significant increase occurring at
DRAFT - DO NOT CITE OR QUOTE 38
-------
both doses. Females also had a dose-related increase in hyperplasia of the oviduct. The
number of tumor-bearing males and the number of males bearing malignant tumors was
significantly increased in both dose groups and followed a significant (p=0.0006 and
0.0001, respectively), dose-related trend. Females had a slight increase in tumor-bearing
animals, which did not reach statistical significance and did not appear to be dose-related.
The number of females bearing malignant tumors was significantly increased for both dose
groups, but not in a dose-dependent manner.
Waalkes et al. (2004a) followed this same procedure (except offspring were
observed for 104 weeks), but 25 male and 25 female offspring from each exposure group
(0, 42.5, or 85 ppm in the drinking water from gestational days 8 to 18 with no additional
exposure after birth) were exposed to acetone or 12-0-tetradecanoyl phorbol-13-acetate
(TPA; 2 u.g/0.1 ml in acetone) twice a week to a shaved area of dorsal skin for 21 weeks
after weaning in an attempt to promote skin tumors. However, very few skin lesions
occurred and were not associated with arsenic exposure either in the absence or presence of
TPA. As was noted in Waalkes et al. (2003), there was a dose-dependent increase in the
incidence and multiplicity of hepatocellular adenomas and carcinomas in treated males,
both in the absence (adenomas: 41.7, 52.2, and 90.5% for the 0-, 42.5-, and 85-ppm
exposure groups, respectively; carcinomas: 12.5, 34.8, and 47.6%, respectively; total
incidence:50, 60.9, and 90.5%, respectively; multiplicity: 0.75, 1.87, and 2.14 respectively)
and presence (adenomas: 34.8, 52.2, and 76.2%, respectively; carcinomas: 8.7, 26.0, and
33.3%, respectively; total incidence:39.1, 65.2, and 85.7%, respectively; multiplicity: 0.61,
1.44, and 2.14 respectively) of TPA with a statistically significant increase noted with 85-
ppm arsenic. Arsenic only caused a dose-dependent increase in hepatocellular adenomas
and carcinomas in the presence of TPA in females (adenomas: 8.3, 18.2, and 28.6% for the
0, 42.5, and 85 ppm exposure groups with TPA exposure, respectively; carcinomas: 4.2,
9.1, and 19.0%, respectively; total incidence: 12.5, 27.3, and 38.1%, respectively;
multiplicity: 0.13, 0.32, and 0.71 respectively) with a statistically significant increase in
total incidence and multiplicity for the 85-ppm group.
There also was an increase in ovarian adenomas in treated female offspring
regardless of whether they were treated with TPA (0, 22.7, 19.0%, respectively) or acetone
(0, 17.4, and 19.0%, respectively). There was no effect on the incidence of ovarian
carcinomas. This was accompanied by increases in the incidence of uterine epithelial
hyperplasia (cystic) and total uterine proliferative lesions, which increased in severity with
dose. There also was a dose-dependent increase in oviduct hyperplasia.
Male offspring exposed to arsenic had an increase in the incidence and
multiplicity of cortical adenomas of the adrenal glands. The increases were statistically
significant for both arsenic exposure groups, but was only related to dose in the absence of
TPA (p=0.020). Incidences were as follows: 37.5, 65.2, and 71.4% for the 0-, 42.5-, and
85-ppm dose groups, respectively, in the absence of TPA and 30.4, 65.2, and 57.1%,
respectively, with TPA treatment. Multiplicities also were statistically significantly
increased in arsenic exposed male offspring with a significant dose-dependent trend both in
DRAFT - DO NOT CITE OR QUOTE 39
-------
the absence (0.58, 2.13, and 2.19, respectively; p=0.0014) or presence (0.54, 1.65, and
1.62, respectively; p=0.016) of TPA.
Lung adenomas were increased in a dose-dependent manner in females exposed
to TPA (4.2, 9.1, and 28.6%, respectively; p=0.018), but not in the absence of TPA (4.2,
8.7, and 9.5%, respectively; not significant). Males only had a statistically significant
increase (5-fold increase) in lung adenomas in the 42.5-ppm group exposed to TPA.
Arsenic caused a statistically significant increase in the tumor multiplicity of all
tumors in males (with or without TPA), which was not dependent on dose. Although
females also had an increase in the tumor multiplicity of all tumors, the only statistically
significant increase occurred in the 85-ppm group exposed to TPA. The increase in
females exposed to TPA also appeared to be dose-dependent The statistically significant
increase observed in the multiplicity of malignant tumors in males was greater in the
absence of TPA, but was dose-dependent in the presence of TPA. In females, there was
also an increase in the multiplicity of malignant tumors in arsenic treated mice (regardless
of TPA exposure), but the results did not reach statistical significance nor were they dose-
dependent.
Rat-Oral
Wei et al. (1999 and 2002) demonstrated that 10-week-old male F344/DuCrj
rats (36/group) administered 50, or 200 ppm dimethylarsenic acid (DMA; a major
metabolite of inorganic arsenic) in their drinking water for 104 weeks developed bladder
tumors (mainly carcinomas) and papillary or nodular hyperplasia in a dose-dependent
manner. Controls and rats administered 12.5 ppm did not develop any bladder tumors or
hyperplasia. There was a significant (p<0.05) increase in 5-bromo-2'-deoxyuridine (BrdU)
labeling of morphologically normal epithelium of the bladder in the 50- (p<0.05) and 200-
ppm (p<0.01) groups (Wei et al., 2002). There was no significant increase in any other
tumor type related to DMA treatment. There appeared to be a dpse-related increase in
subcutis fibromas (i.e., 4% in controls, 12% in 12.5-ppm group, and 16% in both the 50-
and 200-ppm groups). Data indicate that multiple genes are involved in the stages of
DMA-induced urinary bladder tumors. Wei et al. (2002) further indicate that reactive
oxygen species (ROS) may play an important role during the early stages of DMA
carcinogenesis.
Shen et al. (2003) administered TMAO, an organic metabolite of inorganic
arsenic, to male F344 rats for 2 years via their drinking water at concentrations of 0, 50, or
200 ppm. Daily intakes were estimated to be 0, 638, and 2475 mg/kg, respectively. From
87 weeks of treatment on, there was an increase in the incidence and multiplicity of
hepatocellular adenomas in rats sacrificed or dead. Incidences of 14.3, 23.8, and 35.6%,
respectively, were reported. The respective multiplicities were 0.21, 0.33, and 0.53. The
results were statistically significant in the 200-ppm dose group.
DRAFT - DO NOT CITE OR QUOTE 40
-------
Other
Transgenic models also have been developed to examine arsenic carcinogenesis.
Arsenic exposure (200 ppm sodium arsenite in drinking water for 4 weeks) in transgenjc
(Tg.AC) mice containing activated H-ras did not induce skin tumors alone, but in the group
also administered subsequent skin painting with TPA an increase in the number of
papillomas was noted in arsenite treated mice compared to TPA alone. Thus, it was
suggested that arsenite may be a "co-promoter" in skin carcinogenesis (Germolec et al.,
1997; Luster et al., 1995).
Ten ppm of either sodium arsenite or DMAV (cacodylic acid) administered for 5
months in the drinking water of K6/ODC transgenic mice induced a small number of skin
papillomas (Chen et al., 2000a). K6/ODC transgenic mice have hair follicle keratinocytes,
(likely targets for skin carcinogens), which over express ornithine decarboxylase (ODC).
ODC is involved in polyamine synthesis, which is needed in S phase. Over expression of
ODC is sufficient to promote papilloma formation without administration of TPA, which
has been demonstrated to induce ODC (O'Brien et al., 1997).
Rossman et al. (2001) administered sodium arsenite (10 ppm) in the drinking water
of hairless Skh 1 mice for 26 weeks in conjunction with 1.7 kJ/m2 solar UVR (ultraviolet
radiation; considered a low, nonerythemic dose) three times weekly (duration unspecified
in Rossman, 2003) demonstrated a 2.4-fold increase in the yield of skin tumors than in
mice administered UVR alone. A second experiment by the same group (Burns et al,
2004), demonstrated a 5-fold increase in skin tumors using 5 mg/L arsenite with 1 kJ/m2
solar UVR, but also observed a significant increase with 1.25 ppm arsenite. The skin
tumors (mainly squamous cell carcinomas) occurred earlier, were larger, and were more
invasive in mice administered arsenite. Arsenite alone did not induce skin tumors.
Rossman (2003) suggests that this demonstrates arsenite enhances the onset and growth of
malignant skin tumors induced by a genotoxic carcinogen in mice. Rossman (2003) also
suggests that the increased tumor incidence observed by Waalkes et al. (2003) may be due
to the same enhancement as C3H mice have a high background of spontaneous tumors and
suggests the need for examining the transgenic effects in another strain of mice with a
lower background tumdrgenicity.
A critical review of the inhalation data was not conducted as part of this evaluation
discussed in this report.
4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES - ORAL AND
INHALATION
Not addressed in this draft.
4.3.1. Oral
DRAFT - DO NOT CITE OR QUOTE 41
-------
Not addressed in this draft.
4.4. OTHER STUDIES
Possible Modes or Mechanisms of Action
As discussed above, the metabolism of inorganic Arsenic in humans occurs through
alternating steps of reduction and methylation including formation of DMAV, DMA1",
monomethylarsonic acid (MMAV and MMA1"), and trimethanearsinic oxide (TMAO) [19].
Many of the major metabolites MMA, DMA and TMAO have been subjected to a variety
of toxicological tests in vivo and in vitro. The various forms of arsenic have been shown
to differ in toxicity. The trivalent species MMA1" and DMA1" have recently been identified
as the most toxic and genotoxic forms in several assay systems (Thomas et al., 2001). The
relative contributions of the organic metabolites, together with inorganic arsenic to the
toxicity and carcinogenesis of inorganic arsenic are uncertain. Each of the arsenical
metabolites exhibits its own toxicity, possibly via similar and/or separate MOAs that are
responsible for inorganic arsenic toxicity and tumor formation (Kitchin, 2001).
Inorganic arsenic is unique as it is one of a small group of demonstrated
individual chemicals that is carcinogenic to humans and produces tumors at multiple sites
(bladder, lung, skin, liver, and possibly kidney). Rodents are generally nonresponsive to
the tumorigenic effects of iAs except for a recent transplacental mouse study where
arsenite gave liver, lung, ovarian, and adrenal cortical tumors. After decades of research on
arsenic, we have learned a great deal about how arsenic interacts with biological systems
and is affected by biological systems, but we still do not know how arsenic induces human
cancer. Humans are more responsive to arsenic in terms of breadth of effects than any
single rodent species. To date, we do not have enough information to explain these
differences.
The biotransformation and pharmacodynamics of iAs are complex in mammalian
systems with arsenite being biotransformed through a series of reduction and methylation
steps in a cascade to form the final urinary metabolite, trimethylarsine oxide, and possibly
its reduced form, trimethylarsine. Arsenical forms of greater instability are produced
within each step, and these forms have greater reactivity toward biological and biochemical
intermediates, and biological macromolecules. Each intermediate arsenical form has the
potential to induce cancer (genotoxicity) or to affect the promotion and progression of
cancer such as affecting signal transduction pathways and gene expression. Many of these
forms have been detected in the urine of humans exposed to iAs and in rodents exposed to
inorganic and organoarsencials. Moreover, the exposure of mammalian cells and organs to
mixtures of these intermediates brings to the forefront potential synergistic interactions
between these forms that could enhance the tumorigenesis process. To even further
complicate these processes, there is a growing body of evidence that implicates arsenic-
induced reactive oxygen species (ROS) and the downstream effects of arsenic-induced
oxidative damage and oxidative stress in the mechanisms of cellular injury, toxicity, and
DRAFT - DO NOT CITE OR QUOTE 42
-------
carcinogenic activity. This implies that some (if not many) of the toxicological effects of
arsenic are mediated indirectly through ROS. ROS are known to induce DNA damage,
lipid peroxidation, and protein oxidation. ROS themselves are not stable forms. ROS can
interconvert between themselves, can react with nitric oxide to become reactive nitrogen
species (RNS) which have their own spectra of biological activities, and high-energy ROS
can cascade down to lower-energy forms and in that process can radicalize other biological
molecules. Moreover, ROS and sequelae radicals are affected by cellular defenses that can
ameliorate their activities.
Therefore, the metabolic, pharmacokinetic, pharmacodynamic, and cellular
processes that are taking place within these cascades are seemingly complex and it is
believed to be difficult, if not impossible, to apportion risk at this time to any one of these
arsenical intermediates by any known scientifically defensible method. Inorganic arsenic
represents a mixtures issue, with numerous metabolites, each of which has its own
spectrum of toxicity via similar or different modes of action.
The mode of action of inorganic arsenic has been the topic of significant research
effort. The literature on this topic is summarized below.
Genotoxicity
Chromosomal abnormalities have been observed in humans as well as experimental
animals. Chromosomal aberrations and sister chromatid exchange have been observed in
patients treated with Fowler's solution (Burgdorf et al., 1977; Nordenson et al., 1979),
which contains arsenic, and in subjects occupationally exposed (Beckman et al.; 1977).
Although Vig et al. (1984) did not find an increase in either endpoint in the lymphocytes of
subjects exposed to arsenic (109 |J.g/L) via drinking water, other authors have demonstrated
that exposure to drinking water containing 400 ng/L caused an increase in chromosome
aberrations in peripheral lymphocytes (Beckman et al., 1977; Nordenson et al., 1978;
Petres et al., 1977). An increase in micronuclei in exfoliated cells also was noted in
humans exposed to arsenic via the drinking water (Warner et al., 1994; Gonsebatt et al.,
1997).
Liou et al. (1999) conducted a nested case-control study in the blackfoot endemic
area of Taiwan to explore whether incidences of chromosome aberrations (CA) or sister
chromatid exchange (SCE) could predict cancer development. Venous blood samples were
taken from a cohort of 686 residents at the beginning of the study, 22 of which developed
some form of cancer during the 4-year follow-up period. Cytogenetic analyses were
performed on lymphocytes taken from the stored blood samples of the 22 cancer cases and
22 matched controls. It was found that chromosome-type CAs, but not chromatid-type
CAs or SCEs, were significantly higher (p<0.05) in the cases than in the controls. The
cancer risk odds ratio (OR) for subjects with > 0 chromosome-type CAs was 5.0 (95%
confidence interval 1.09-22.82).
DRAFT - DO NOT CITE OR QUOTE 43
-------
Basu et al. (2002) examined the incidences of micronuclei (MN) in a group of 45
individuals from West Bengal, India, who had cutaneous signs of arsenicism and drank
water having a mean arsenic content of 368.11 \ig/L. The control group consisted of 21
asymptomatic individuals with a drinking water supply having a mean arsenic
concentration of only 5.49 ng/L. For the exposed and control groups, arsenic
concentrations in the urine, nails and hair were 24.45, 4.88 ug/L, and 12.58 and 0.51 ng/g,
6.97 and 0.34 ng/g, respectively. In the exposed group, the frequencies of MN per 1000
cells were highly elevated over those of the control group: 5.15 vs. 0.77 in the oral mucosa,
5.74 vs. 0.56 in urothelial cells, and 6,39 vs. 0.53 in lymphocytes, respectively. This was
supported by Moore et al. (1997b) who reported an increase in the prevalence of MN in
Chilean males chronically exposed to 600 [ig As/L. Thus, individuals exposed to high
levels of arsenic via drinking water appear to be sustaining significant cytogenetic damage.
Rats orally exposed to arsenate (4 mg As/kg/day) for 2-3 weeks developed major
chromosomal abnormalities in the bone marrow (Datta et al., 1986). Arsenite-induced
micronuclei in the bone marrow of mice in vivo (Tinwell et al., 1991).
Sodium arsenate has been demonstrated to transform Syrian hamster embryo cells
(Dipaolo and Casto, 1979) and to produce sister chromatid-exchange in DON cells, CHO
cells, and human peripheral lymphocytes exposed in vitro (Larramendy et al., 1981; Ohno
et al, 1982; Wan et al., 1982; Andersen, 1983; Crossen, 1983). Sodium arsenate and
sodium arsenite are mutagenic at concentrations of 10-14 (ig/mL and 1-2 ng/mL,
respectively (Harrington-Brock et al., 1993; Moore et al., 1995). MMA and DMA were
slightly less potent with 2.5-5 mg/mL and 10 mg/mL, respectively, to induce a genotoxic
response. The study authors judged the mutations to be chromosomal rather than point
mutations due to the small colony size.
Arsenic has generally failed as a mutagen in bacteria and has only been observed as
a weak mutagen at the hprt locus in Chinese hamster V79 cells at toxic concentrations (Li
and Rossman, 1989a). Arsenic does not appear to cause point mutations in standard
assays, but instead causes large deletion mutations (Rossman, 1998). These large deletions
can cause lethality when closely linked to essential genes. Therefore,-the mutations are not
easily observed in standard bacterial and mammalian cell mutation assays. However, even
in transgenic cell lines, which were tolerant of large deletions, arsenic was still only weakly
mutagenic at toxic doses (Rossman, 2003).
It has been suggested that arsenic acts as an aneugen at low doses, but as a
clastogen at high doses (Rossman, 2003). This was suggested because a low-dose protocol
(5 uM arsenite for 24 hours in normal human fibroblasts) results in mainly kinetochore
positive (1C) MN (usually derived from whole chromosomes), while a high-dose protocol
(20 u-M for 4 hours) resulted in mainly kinetochore negative (K") MN (derived from
chromosomal fragments). Both the protocols caused the same level of arsenic
accumulation and toxicity in normal human fibroblasts (Yih and Lee, 1999 as cited in
Rossman, 2003). Other clastogenic agents, such as X-rays, also induce high levels of K'
DRAFT - DO NOT CITE OR QUOTE 44
-------
MN (Fenech et al., 1999) and other agents causing aneuploidy by interfering with spindle
function, such as vinblastine, to induce K+ MN (Eastmond and Tucker, 1989). V79 cells
exposed to lOuM arsenite disrupted mitotic spindles and persistent aneuploidy (5 days after
removal of arsenic) without notable chromosome aberrations in surviving cells
(Sciandrello et al., 2002; Warner et al., 1994).
Zhang et al. (2003) determined that low concentrations of arsenic (<1 uM)
increased telomerase activity, maintained or elongated telomere length, and promoted cell
proliferation in cultured HL-60 and HaCaT cells, while high concentrations (l-40uM) of
arsenic decreased telomerase activity, reduced telomere length, and induced apoptosis.
Telomeres are located at the end of chromosomes and play a critical role in maintaining
chromosome and genomic stability. Results of the study indicate that telomerase was
involved in arsenic-induced apoptosis. Data, however, also suggest that reactive oxygen
species (ROS) may be involved in the shortening of telomeres and apoptosis induced
arsenic. Chou et al. (2001) report that arsenic trioxide (0.75 uM) inhibited telomerase
activity in NB4 cells after 8 days of exposure with significant numbers of fusion
chromosomes observed in 2-3 weeks. Chou et al. (2001) also demonstrated that the
suppressed telomerase activity in NB4 cells correlated with a decrease in ATERT mRNA
and protein. Zhang et al. (2003) propose that the increase in telomerase activity leading to
promotion of cell proliferation leads to its carcinogenic effects; however, its
anticarcinogenic effects are related to oxidative stress leading to telomeric DNA attrition
and apoptosis.
Sodium arsenite induces DNA-strand breaks associated with DNA-protein
crosslinks in cultured human fibroblasts at 1-5 mM (3 mM was the most effective), but not
at 10 mM (Dong and Luo, 1993).
Injection (intraperitoneal) of DMAV (10.6 mg/kg/day) for 5 consecutive days only
weakly enhanced the frequency of lacZ gene mutants (at most 1.3-fold greater than control)
in cells from the lung of Muta™ Mouse transgenic mouse (Noda et al., 2002), even though
such treatment was known to induce DNA damage in the lung via the formation of various
peroxyl radicals. Mutant frequencies were not significantly enhanced at the lacZ locus in
the bladder or bone marrow. A marginal increase in G:C to A:T transitions occurred at the
cJIgene locus in the lung. Arsenic trioxide (7.6 mg/kg/day) also failed to enhance lacZ
mutant frequencies in the lung, kidney, bone marrow or bladder. MN formation in
peripheral blood reticulocytes was enhanced by arsenite, but not by DMAV. The study
authors concluded that the assay system may not be sensitive enough to detect DMAv's
genotoxicity.
To study the effect of methylation state on arsenic-induced genotoxicity, Yamanaka
et al. (1997) exposed cultures of human alveolar epithelial type II (L-132) cells to arsenite,
MMA or DMA and assayed for DNA damage using a DNA repair synthesis inhibition
protocol. DNA single strand breaks, resulting from the inhibition of repair, were induced
by 5-100 uM DMA, but not by arsenite or MMA, even at 100 uM concentrations.
DRAFT - DO NOT CITE OR QUOTE 45
-------
However, DMA exposure also induced DNA repair using a bromodeoxyuridine (BrdU)-
photolysis assay. Again, these results were not observed with either arsenite of MM A.
However, when cells were co-exposed to MMA and the methyl-group donor SAM, cellular
levels of DMA increased and DNA repair was induced, emphasizing the importance of
methylation to arsenic's ability to damage DNA in these human epithelial cells.
Further exploring the impacts of methylation and valence state on the genotoxicity
of arsenical, Mass et al. (2001) assessed the DNA-damaging capacities of As1", Asv,
MMAV and DMAV in a supercoiled .XI74 RFI DNA nicking assay. The methylated
trivalent arsenicals were the only arsenic compounds observed to damage naked DNA.
Concentrations of up to 300 mM As"1, 1 M Asv, 3 M MMAV and 300 mM DMAV were
without nicking activity. DMA"1 had nicking activity at 150 uM. The compounds also
were assessed in the SCO comet assay using human lymphocytes. The activity of Asv was
only slightly greater than that of As"1, while MMAV and DMAV were without any
demonstrable activity. DMA1" was 386 times more potent than As1".
Nesnow et al. (2002) used the supercoiled 0X174 DNA nicking assay to study the
involvement of ROS on MMA1" and DMA'"-induced DNA damage. MMA1" completely
fragmented the 0X174 DNA at concentrations of 30-50 mM. Significant DNA-nicking
also was observed with 37.5 |o,M of DMA1". ROS inhibitors Tiron, melatonin, and the
vitamin E analogue Trolox all inhibited the DNA-nicking activity of both MMA"1 and
DMA"1 at low (0.01 mM) concentrations. The authors also suggest that the formation of
DMPO-hydroxyl free radical adducts was dependent on time and the presence of DMA1".
The formation of DMPO-hydroxyl free radical adduct was completely inhibited by Trion
and Trolox. Data indicate that the DNA-damaging action of DMA1" is an indirect
genotoxic effect mediated by ROS formed concomitantly with the oxidation of DMA1" to
DMAV.
Kashiwada et al. (1998) examined the cytogenetic effects in mouse bone marrow
cells after a single intraperitoneal injection of DMAV. Their data suggest that DMAV may
cause mitotic arrest both in vivo and in vitro. The data also suggest that DMAV is a
significant inducer of aneuploidy (44.9 % versus 6.0 % in controls) with over 80% of the
aneuploid cells hyperploids containing 1 or 2 extra chromosomes. The authors suggest that
DMAV aneuploid induction could be at least one of the mechanisms underlying arsenic's
carcinogenicity.
Kligerman et al. (2003) exposed human peripheral blood lymphocytes,
L5178Y/Tk(+/-) mouse lymphoma cells, Salmonella, and E. coli to Asv, As"1, MMAV,
MMA1", DMAV, or DDM"1. MMA1" and DMA1" were found to be the most potent
clastogens in human lymphocytes, as well as, the most potent mutagens at the Tk (+/-)
locus in mouse lymphoma cells; however, they did not induce prophage. None of the
arsenic compounds, however, were potent inducers of SCE nor did they cause gene
mutations in the three strains of Salmonella tested.
DRAFT - DO NOT CITE OR QUOTE 46
-------
Dopp et al. (2004) found a significant (p<0.05) increase in the number of
micronucleus (MM), chromosome aberrations (CA), and sister chromatid exchanges (SCE)
in CHO cells exposed to MMA1" and DMA1". Arsenate and arsenite induced CA and SCE,
but not MN. TMAOV, MMAV, and DMAV were not found to be genotoxic in the range of
concentrations tested (s5 mM).
Aberrant gene/protein expression
Arsenite has been demonstrated to induce gene amplification, which is a sign of
gene instability, at the dhfr locus in human and rodent cells; however, it was unable to
cause amplification of SV40 sequences in SV40-transformed human keratinocytes or
Chinese hamster cells (Mure et al., 2003; Barrett et al., 1989;Rossman and Wolosin, 1992).
These data suggest that arsenite feeds into checkpoint pathways common to those
involving p53, whose disruption leads to cellular gene amplification, instead of signaling ,
typical DNA-damaging agents, which tend to amplify SV40 (Livingstone et al., 1992).
DNA hypo- or hypermethylation also have been indicated as a mechanism for
carcinogenesis. Cytosine methylation in the p53 promoter in human adenocarcinoma A549
cells was one of the first indications of arsenite induced methylation changes (Mass and
Wang, 1997). However, both hypo- and hypermethylation of different genes after exposure
to arsenite have been noted in human kidney UOK. cells (Zhong and Mass, 2001).
Waalkes et al. (2004b) observed a 32% decrease in insulin-like growth factor 1
(IGF-1) expression and a 3 -fold increase in the expression of IGF- 1 binding protein
(IGFBP-l) in arsenic-induced hepatocellular carcinomas. Scharf et al. (2001) also
observed this in patients with hepatocellular carcinomas. However, Waalkes et al. (2004b)
also noted a 2-fold increase in hepatic IGFBP-l expression in nontumor tissues of arsenic-
exposed animals.
Cytokeratin-8 and cytokeratin-18 were both found increased in arsenic-induced
hepatocellular carcinomas (Waalkes et al., 2004b). In arsenic-exposed nontumorous tissue,
cytokeratin-8 was increased and cytokeratin-18 was decreased compared to normal
unexposed tissue. This is notable because cytokeratins have been suggested to play
essential "guardian" role in the liver with aberrant expression associated with liver disease
and hepatocellular carcinoma formation (Omary et al., 2002).
Betaine-homocystein methyltransfcrase was reduced by 68% in arsenic-induced
hepatocellular carcinomas compared to the nontumor tissue (Waalkes et al., 2004b).
However, the nontumor tissue in arsenic exposed animals had similar levels of betaine-
homocystein as in spontaneously developing hepatocellular carcinomas, which was lower
than normal unexposed tissue. Avila et al. (2000) also have observed a reduction in this
enzyme in human liver cirrhosis and hepatocellular carcinomas. Waalkes et al (2004b)
suggests that this may be associated with abnormal DNA methylation status and aberrant
DRAFT - DO NOT CITE OR QUOTE 47
-------
gene expression, both of which have been suggested mechanisms of arsenic carcinogenesis
(Huang et al., 2004; Rossman, 2003).
Possible involvement of reactive oxygen species (ROS1
Reactive oxygen species (ROS) have been postulated to be involved in both the
initiation and promotional stages of carcinogenesis (Shackelford et al., 2000; Zhong, et al.,
1997; Bolton, et al. 2000; Bolton et al., 1998; Chen et al., 2000b; Khan et al., 2000; and Xu
et al., 1999). Arsenite does not react with DNA itself, but oxidate damage to the DNA of
arsenite treated cells has been observed (Rossman, 2003). Low levels of ROS can
modulate gene expression by acting as a secondary messenger, while high doses of ROS
can cause oxidative injury leading to cell death (Perkins et al., 2000). ROS also has been
suggested to damage cells by the following mechanisms: lipid peroxidation; DNA and
protein modification, as well as causing structural alterations in DNA including base-pair
mutations, rearrangements, deletions, insertions, and sequence amplifications (but not point
mutations); involvement in the signaling of the cell transformation response; affecting
cytoplasmic and nuclear signal transduction pathways that regulate gene expression; and
increasing the expression of certain genes (e.g., MDM2 protein, a key regulator of the
tumor suppression gene p53) (Rossman et al., 1980; Rossman et al., 1975; Li et al., 1998;
Sen, et al., 1996; Lander, 1997; and Hamadeh et al., 1999).
ROS have been measured in cells during arsenic metabolism (Barchowsky et al.,
1999). Arsenic has been demonstrated to increase oxygen consumption, superoxide (O2'~)
formation, and extracellular accumulation of hydrogen peroxide (H202). Data by Wang et
al. (1996) suggest elevated intracellular peroxide levels as well. Using a confocal
scanning microscope with a fluorescent probe, Liu et al. (2001) reported a 3-fold increase
in intracellular oxyradical production. However, co-treatment of cultures with dimethyl
sulfoxide (DMSO), an oxygen radical scavenger, caused the fluorescent intensity to be
reduced to control levels indicating a reduction in oxyradical production. Increases in
serum lipid peroxides have been observed in a Chinese population chronically exposed to
arsenic from their drinking water (Pi et al., 2002).
Further support for the involvement of ROS in arsenic carcinogenesis, is that
situations that cause a reduction of ROS also cause reductions in the genotoxic effects of
arsenic, while situations that are associated with increases in ROS cause an increase in
arsenic induced genotoxic response. Examples of this include, inhibition of arsenic-
induced sister chromatid exchange in human cells when superoxide radical dismutase
(SOD) is added to the culture medium (Rossman, 1998); protection from arsenic
genotoxicity in the presence of SOD, elevated GSH, Vitamin E, catalase, and squalene (Hei
et al, 1998; Lee and Ho, 1995; Nordenson and Beckmam, 1991; Huang et al., 1993; Wang
and Huang, 1994; Fan et al., 1996; Kesscl et al., 2002); H2O2-resistant CHO cells
exhibiting cross-resistant to arsenite (Cantoni et al., 1994); and blockage of the
mutagenicity of arsenite in AL cells by dimethyl sulfoxide (DMSO), a free radical
scavenger,(Hei et al., 1998). Depletion of GSH in cell culture increases the toxic and
DRAFT - DO NOT CITE OR QUOTE 48
-------
clastogenic effect of arsenite (Oya-Ohta et al., 1996). GADD153, a DNA damage-
inducible protein, is induced by arsenite, but this is suppressed by the antioxidant N-acetyl
cysteine (Guyton et al., 1996). Rossman (1998, 2003) and Wang et al. (1997) also note
that XRS-5 cells, which are deficient in catalase, are hypersensitive to arsenic cytotoxicity
as well as arsenic-induced MN formation.
GSH depletion, increased oxidized glutathione (GSSG), and elevated
malondialdehyde in the liver and brain were observed in rats administered arsenite for 12
weeks (Flora etal., 1999).
Rin et al. (1995) postulates that DMA forms a stable adduct with DNA. These
adducts may be more susceptible to single strand DNA breaks by 02".
Shen et al. (2003) tested the contribution of ROS to TMAO induced hepatocellular
adenomas in male F344 rats administered 0, 50, or 200 ppm TMAO in their drinking water
for 2 years (estimated average daily intakes of 0, 638, and 2475 mg/kg/day, respectively).
Formation of 8-hydroxydcoxyguanosine (8-OHdG) was measured using high performance
liquid chromatography. The 8-OHdG values were significantly (p<0.05) higher in the 200-
ppm group compared to the controls. TMAO also was reported to increase cell
proliferation in normally appearing parenchyma as measured by proliferating cell nuclear
• antigen index. The results indicate that TMAO possibly causes liver tumorigenicity via
oxidativc DNA damage and enhanced cell proliferation.
DNA repair inhibition
Arsenite appears to inhibit the DNA repair process by inhibiting both excision and
ligation (Jha et al., 1992; Lee-Chen et al., 1993). DNA repair enzymes are inhibited in
arsenic exposed cells (Simeonova and Luster, 2000; Maier et al., 2002). Arsenite has been
demonstrated to decrease DNA ligase activity, but did not directly affect the enzyme level
(Li and Rossman, 1989b; Hu et al., 1998). Hu et al. (1998) suggest that arsenite indirectly.
inhibits DNA ligase activity either by altering cellular redox levels or by affecting signal
transduction pathways and phosphorylation of proteins related to DNA ligase activity.
LI 32 cells exposed to 10 mM DMAV increased the frequency of single strand
breaks in DNA, reduced DNA replication, and shortened the length of the nascent chain of
DNA (Tezuka et al., 1993), suggesting that both replication and repair of DNA are
affected. Kawaguchi et al. (1996) suggests an interaction between DMA and paraquat (a
generator of O2') in the production and persistence of DNA damage.
This inhibition of DNA repair could explain why arsenic compounds enhance the
carcinogenic effect of various compounds. Arsenic has been found to enhance the
mutagenicity of UV in E. coli (Rossman, 1981); enhance the mutagenicity and/or
clastogenicity of UV, Af-methyl-N-nitrosourea (MNU), diepoxybutane, X-rays, and
methylmethane sulfonate in mammalian cells (Li and Rossman, 1989a; Li and Rossman,
DRAFT - DO NOT CITE OR QUOTE 49
-------
1991; Wiencke and Yager, 1992; Lee et al, 1985; Lee et al., 1986; Yang et al., 1992; Jha et
al, 1992); and enhanced skin tumors(mainly squamous cell carcinomas in hairless Skh 1
mice after exposure to 1.7 kJ/m2 solar UVR [ultraviolet radiation; considered a low,
nonerythemic dose], three times weekly with the onset earlier and tumors larger in mice
exposed to arsenic, while arsenite alone did not induce tumors) (Rossman et al., 2003;
Burns et al., 2004).
Signal transduction
Arsenic also has been speculated to modulate gene expression by activating signal
transduction pathways (Snow, 1992; Huang et al., 1999). However, data are conflicting
and generally relate to the DNA repair inhibition, gene expression, or gene amplification
mentioned above. P53, a crucial tumor suppressor gene, involvement has been
controversial (Huang et al., 2004; Rossman, 2003). Exposure to human cells to arsenite
(0.1-lOOuM) for 24 hours caused a dose-dependent increase in the level ofp53 protein
expression, especially in the presence of cells carrying wild type p53 genes (Salazar et al.,
. 1997). However, when immortalized human keratinocyte HaCaT cells, which normally
over express p53 protein, was exposed to arsenite (0.01-1 U.M) a dose- and time-dependent
decrease in p53 protein levels was observed (Hamadeh et al., 1999). While yet another
study demonstrated no effect of arsenite (12.5-200 u.M) on p53-'dependent transcription
activity in p53 promoter-transfected JB6 C141 mouse cells (Huang et al., 1999).
Activation of signal transduction pathways which enhance cell proliferation, reduce
antiproliferative signaling, and override checkpoint controlling cell division after genotoxic
insult also have been considered as possible mechanisms of arsenic's co-carcinogenic
properties (see Rossman, 2003 for review).
Other
Bladder Carcinogenesis
Drobnd et al. (2002) used a UROtsa cell line, which is an SV40 immortalized cell
line derived from normal human urothelium. This cell line was determined not to
methylate arsenic. Therefore, it is useful in examining the effects of inorganic arsenic and
methylated arsenic. Cells were incubated with inorganic (As111 and Asv) or methylated
arsenicals (MMAV, MMA11^, DMAV, and iododimethylarsine [DMA111!]) for 2 or 24 hours.
Because trivalent arsenic has been determined to be more toxic than pentavalerit arsenic,
the doses were lower for trivalent arsenic compounds (0.1-5 u-M) than for pentavalent
compounds (1-200 u.M). Pentavalent arsenic was reported not to affect the UROtsa cells,
and the data was not presented in the study report. Cell viability was only decreased with a
dose of 5u.M methylarsine oxide (MMA'"O) after both 2 and 24 hours. Cell proliferation
also was affected to a lesser extent with 5 u.M DMAs'"I. All the trivalent arsenic
compounds examined increased AP-1 DNA binding activity in UROtsa cells, but the
methylated compounds were much more potent activators of AP-1. The AP-1 DNA
DRAFT - DO NOT CITE OR QUOTE 50
-------
binding activity was determined to act through the ERK-dcpendent induction of c-Jun
phosphorylation with the possible involvement of Fra-1 phosphorylation. The order of
potency of c-Jun/Fra-1 phosphorylation is as follows from most potent to least potent:
MAs"1 O > DMAs'"l > As"1. These results suggest that methylation of arsenic is involved
in the toxicity and carcinogenicity of arsenic to urinary bladder cells.
Cohen et al. (2002) demonstrated that dimethylarsenic acid (DMAV) is cytotoxic to
the urinary bladder via formation of DMA1". Female F344 rats (4 weeks old) were
administered 100 ppm DMAV via the diet with or without co-administration of 2,3-
dimercaptopropane-1-sulfonic acid (DMPS), a chelator of trivalent arsenicals for 2 or 26
weeks. Controls were fed basal diets only. Bladder weights were unchanged by week 2,
but were significantly greater by week 26. At 2 weeks, one of 10 rats administered DMAV
had simple hyperplasia, while 4 of 9 had developed simple hyperplasia by 26 weeks. None
of the controls developed hyperplasia at 2 or 26 weeks. Co-administration of DMPS
caused a reduction in the development of simple hyperplasia. DMA1" also was measured in
the urine on days 1, 71, and 175. In vitro experiments demonstrated that arsenite
compounds (i.e., arsenite, MMA1", and DMA1") were more cytotoxic to rat (i.e., MYP3)
and human (1T1) urinary bladder cell lines than arsenate compounds (i.e., arsenate,
MMAV, DMAV, and TMAO).
4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER
EFFECTS
4.5.1. Oral
Not addressed in this document.
4.5.2. Inhalation
A review and assessment of relevant inhalation studies is not addressed in this
report.
4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER
CHARACTERIZATION
4.6.1. Summary of Overall Wcight-of-Evidcncc
Based upon current EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA,
2005) inorganic arsenic is determined to be "carcinogenic to humans" due to convincing
epidemiological evidence of a causal relationship between oral exposure of humans to
inorganic arsenic and cancer.
4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence
DRAFT - DO NOT CITE OR QUOTE 51
-------
There have been numerous studies examining the carcinogenic potential of
inorganic arsenic via oral exposure (see NRC 2001 Table 2-1). The majority of these
studies are ecological in nature and are therefore subject to the biases of ecological studies
(e.g., lack of individual exposure). Ecological epidemiologic studies examine differences
in disease rates among populations in relation to age, gender, race, and differences in
temporal or environmental conditions. Although previous draft cancer guidelines have
considered these studies insufficient to ascertain the causal agent or degree of exposure, the
current guidelines (US EPA, 2005) do not provide guidance on the issue. Given the
detailed information available on the drinking water exposure concentrations in the studies
selected for arsenic, they are considered, for the purposes of this report, to be sufficient for
both ascertaining the causal agent and determining a dose-response relationship.
The strength of the associations observed in the studies mentioned above support a
relation between oral exposure to inorganic arsenic and cancer. Each study was conducted
differently and contain their own biases (e.g., lack of confounding variables, possible recall
bias), but the combination of all the study results provides support of an association
between oral exposure to inorganic arsenic and cancer including bladder, kidney, lung,
liver, and prostate. Different populations were assessed. Therefore, it is unlikely that any
single attribute (e.g., nutritional habits) associated with a single population is responsible
for the increased cancer rates. Although many studies did not account for confounding
variables (e.g., cigarette smoking in association with lung cancer), the positive associations
also were observed in-those studies that did account for certain confounding variables (e.g.,
habits, age, socioeconomic status).
Twelve studies examined the association between arsenic concentrations in
drinking water and skin cancer. All 12 had positive associations. Of the 12 studies, 9 were
ecological study designs with which there was one each of cross sectional, case-control,
and cohort. The major flaw in the association between oral exposure to inorganic arsenic
and skin cancer was the lack of data and/or analysis of dermal contact with the water
containing the inorganic arsenic through bathing.
Eighteen studies have examined arsenic in the drinking water and bladder cancer.
Fifteen of the studies (9 ecological, 2 case-control, and 4 cohort studies) found a positive
association. The studies that failed to observe an association between oral exposure to
inorganic arsenic and bladder cancer (Bates et al., 1995; Lamm et al., 2004; Steinmaus et
, al., 2003; 1 ecological and 2 case-control) were those which examined low exposures.
Bates et al. (1995) and Steinmaus et al. (2003), however, found an association between
arsenic exposure (exposures 20 years or more) and bladder cancer in people who had ever
smoked.
Nine studies (6 ecological and 3 cohort) found an association between liver cancer
and arsenic concentrations in the drinking water. Smith et al. (1998) was the only study
that examined liver cancer that did not find an association with arsenic. This study was
DRAFT - DO NOT CITE OR QUOTE 52
-------
ecological in nature using mortality as an endpoint. Its major limitations were that there
was limited information on smoking and the results could not be separated by age or sex.
Respiratory cancers have been associated with arsenic exposure via drinking water
in 11 of!2 studies. The majority of the studies were ecological (i.e., 8), one was a case-
control and 3 were cohort. One of the cohort studies demonstrated negative findings
(Cuzick et al, 1992). Mortality was the endpoint in the negative study. Eight studies (7
ecological and 1 cohort) examined and found an association between arsenic in the
drinking water and kidney tumors. Three studies (2 ecological and 1 cohort) found an
association between arsenic concentrations in the drinking water and prostate cancer.
There also have been an occasional study that has associated arsenic exposure with
leukemia and lymphomas, colon cancer, and bone cancer (Chen et al., 1985, 1988; Tsai,
1999; Wu, 1989).
Lung and liver tumors also were observed in mice administered inorganic arsenic
for a short duration transplacentally with possible additional exposure while nursing
(pregnant dams were exposed for 10 days during gestation only). Therefore, increasing the
evidence that lung and liver cancers are associated with oral exposure to inorganic arsenic.
4.6.3. Mode of Action Information
The carcinogenic mode of action for inorganic arsenic is not known and consensus
in the scientific community has yet to be reached. Kitchin (2001) provides a review of nine
different possible modes for arsenic carcinogenesis, including chromosome abnormalities;
oxidative stress; increased transcription of mRNA and secretion of transforming growth
factor-alpha (TGF-a); granulocyte macrophage-colony stimulating factor (GM-CSF) and
tumor necrosis factor-alpha (TNF-a); cell proliferation, promotion and/or progression in
carcinogenesis; altered DNA repair; p53 gene suppression; altered DNA methylation
patterns; and gene amplification. The article suggests that the three with the most positive
evidence in both animals and human cells are chromosomal abnormalities, oxidative stress,
and a continuum of altered growth factors leading to increased cell proliferation which
leads to the promotion of carcinogenesis. Rossman (2003), Huang et al. (2004), and
Simeonova and Luster (2000) also provide reviews examining the aforementioned modes
of arsenic carcinogenesis. In addition, it is noted that humans excrete more MMA than
other species, and they are more prone to arsenic-induced carcinogenesis; therefore, it is
likely that MMA (probably the trivalent form) is involved in the carcinogenesis of arsenic.
One possible carcinogenic process involves the metabolism of inorganic arsenic to
DMA or TMAO. Wei et al. (1999; 2002) demonstrated that DMA was carcinogenic in
male F344 rats. Shen et al. (2003) demonstrated the carcinogenicity of TMAO in male
F344 rats. Arsenite and arsenate, however, have been negative in standard carcinogenic
bioassays (NRC, 1999). Both Wei et al. (2002) and Shen ct al. (2003) further established
that reactive oxygen species (ROS) are likely involved in the carcinogenesis (further
DRAFT - DO NOT CITE OR QUOTE 53
-------
description below). Wei efal. (2002) also demonstrated that multiple genes are involved at
different stages of DMA-induced tumor development. The in vitro studies by Nesnow et
al. (2002) demonstrated that MMA1" and DMA"1 had DNA-damaging effects that were
inhibited by chemically diverse ROS inhibitors. DMA, but not MMA or arsenite, also was
demonstrated to cause DNA damage (Yamanaka et al., 1997). The DNA damage
accompanied by possible increased cell proliferation (Shen et al., 2003) provide additional
mechanistic insight. The association of ROS and oxidative stress with a wide variety of
cancers (e.g., bladder, lung, liver, prostate, etc.) indicates either a number of different
mechanisms or a mechanism generalizable over multiple organs (Klaunig and Kamendulis,
2004).
Zhang et al. (2003) determined that low concentrations of arsenic (<1 uM) increase
telomerase activity, maintained or elongated telomere length, and promoted cell
proliferation in cultured HL-60 and HaCaT cells, while high concentrations (l-40uM) of
arsenic decreased telomerase activity, reduced telomere length and induced apoptosis.
Telomeres are located at the end of chromosomes and play a critical role in maintaining
chromosome and genomic stability. Results of the study indicate that telomerase was
involved in arsenic-induced apoptosis. Data, however, also suggest that reactive oxygen
species (ROS) may be involved in the shortening of telomeres and apoptosis induced
arsenic. Chou et al. (2001) reported that arsenic trioxide (0.75 uM) inhibited telomerase
activity in NB4 cells after 8 days of exposure with significant numbers effusion
chromosomes observed in 2-3 weeks. Chou et al. (2001) also demonstrated that the
suppressed telomerase activity in NB4 cells correlated with a decrease in ftTERT mRNA
and protein. Zhang et al. (2003) propose that the increase in telomerase activity leading to
promotion of cell proliferation leads to its carcinogenic effects, but its anticarcinogenic
effects are related to oxidative stress leading to telomeric DNA attrition and apoptosis.
The following include alterations of cellular pathways that may be involved in the
process of arsenic carcinogenesis. Neither a step by step process nor details of how each
process is involved have been developed at this time. Data, however, implicate a role for
each mode of action.
Genotoxicity
There are large amounts of information available on the in vitro and in vivo
genotoxicity of arsenicals with chromosomal abnormalities observed in both humans and
animals (Basu et al., 2001; Kligerman et al. 2003). Briefly, arsenicals are stronger
clastogens than point mutations, and methylated trivalent arsenicals are usually much more
potent genotoxins than are inorganic arsenic or pentavalent methylated arsenicals
(Kligerman et al., 2003). However, there is no connection between the clastogenicity of
arsenic and the carcinogenicity in various organs in humans. Some data indicates that the
clastogenic effects of arsenic also are mediated via free radicals (Kitchin, 2001; Kligerman
et al., 2003).
DRAFT - DO NOT CITE OR QUOTE 54
-------
It also has been suggested that arsenic acts as an aneugen at low doses, but as a
clastogen at high doses (Rossman, 2003). This was suggested because a low-dose protocol
(5 uM arsenite for 24 hours in normal human fibroblasts) results in mainly kinetochore
positive (K+) MN (usually derived from whole chromosomes), while a high-dose protocol
(20 u.M for 4 hours) resulted in mainly kinetochore negative (K") MN (derived from
chromosomal fragments).
Aberrant gene/protein expression
Arsenite has been demonstrated to induce gene amplification, which is a sign of
gene instability at the dhfr locus in human and rodent cells; however, it was unable to
cause amplification of SV40 sequences in SV40-transformed human keratinocytes or
Chinese hamster cells (Mure et al., 2003; Barrett et al., 1989; Rossman and Wolosin,
1992). These data suggest that arsenite feeds into checkpoint pathways common to those
involving p53, whose disruption leads to cellular gene amplification, instead of signaling
typical DNA-damaging agents, which tend to amplify SV40 (Livingstone et al., 1992).
DNA hypo- or hypermethylation also have been indicated as a mechanism for
carcinogenesis. Cytosine methylation in the p53 promoter in human adenocarcinoma A549
cells was one of the first indications of arsenite induced methylation changes (Mass and
Wang, 1997). However, both hypo- and hypermethylation of different genes after exposure
to arsenite have also been noted in human kidney UOK cells (Zhong and Mass, 2001).
Possible involvement of reactive oxygen species (ROS)
Reactive oxygen species (ROS) have been postulated to be involved in both the
initiation and promotional stages of carcinogenesis (Shackelford et al., 2000; Zhong et al.,
1997; Bolton, et al. 2000; Bolton et al., 1998; Chen et al., 2000b; Khan et al., 2000; and Xu
et al., 1999). Arsenite does not react with DNA itself, but oxidate damage to the DNA of
arsenite-treated cells has been observed (Rossman, 2003). Low levels of ROS can
modulate gene expression by acting as a secondary messenger, while high doses of ROS
can cause oxidative injury leading to cell death (Perkins et al., 2000). ROS also has been
suggested to damage cells by the following mechanisms: lipid peroxidation; DNA and
protein modification and causing structural alterations in DNA including base-pair
mutations, rearrangements, deletions, insertions, and sequence amplifications (but not point
mutations); involvement in the signaling of the cell transformation response; affecting
cytoplasmic and nuclear signal transduction pathways that regulate gene expression; and
increasing the expression of certain genes (e.g., MDM2 protein, a key regulator of the
tumor suppression gene p53) (Rossman et al., 1980; Rossman et al., 1975; Li et al., 1998;
Sen, et al., 1996; Lander, 1997; and Hamadeh et al., 1999).
DNA repair inhibition
DRAFT - DO NOT CITE OR QUOTE 55
-------
Arsenite appears to inhibit the DNA repair process by inhibiting both excision and
ligation (Jha et al., 1992; Lee-Chen et al., 1993). DNA repair enzymes are inhibited in
arsenic exposed cells (Simeonova and Luster, 2000; Maier et al., 2002). Arsenite has been
demonstrated to decrease DNA ligase activity, but did not directly affect the enzyme level
(Li and Rossman, 1989b; Hu et al., 1998). Hu et al. (1998) suggest that arsenite indirectly
inhibits DNA ligase activity either by altering cellular redox levels or by affecting signal
transduction pathways and phosphorylation of proteins related to DNA ligase activity.
This inhibition of DNA repair could explain why arsenic compounds enhance the
carcinogenic effect of various compounds.
Signal transduction
Arsenic also has been speculated to modulate gene expression by activating signal
transduction pathways (Snow, 1992; Huang et al., 1999). However, the data is conflicting
and generally relates to the DNA repair inhibition, gene expression, or gene amplification
mentioned above. P53, a crucial tumor suppresser gene, involvement has been
controversial (Huang et al., 2004; Rossman, 2003). Activation of signal transduction
pathways which enhance cell proliferation, reduce antiproliferative signaling, and override
checkpoint controlling cell division after genotoxic insult also have been considered as
possible mechanisms of arsenic's co-carcinogenic properties (see Rossman, 2003 for
review).
Low-dose extrapolation
According to the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA,
2005), a linear extrapolation to low doses is to be used when there are mode of action data
to indicate that the dose-response curve is expected to have a linear component below the
point of departure. Examples of such data include DNA-reactivity or direct mutagenic
activity. Additionally, a linear extrapolation is used as a default approach when the
available data are insufficient to establish the mode of action for a tumor site. It has been
suggested, but not proven, that arsenic may be directly genotoxic. Therefore, a linear
extrapolation is recommended.
4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.7.1. Possible Childhood Susceptibility
Although children are exposed to arsenic through the same sources as adults (i.e.,
air, water, food, and soil), their habits may cause children to have a greater exposure than
adults. Children tend to eat less of a variety than adults; therefore, exposure to
contaminated food or juice or infant formula prepared with contaminated water may
accumulate to a significant exposure. In addition, children are more likely to ingest
arsenic-contaminated soil either intentionally or via putting dirty hands in their mouths.
DRAFT - DO NOT CITE OR QUOTE 56
-------
Accidental poisoning also is more likely to occur due to children's curiosity and aptness to
taste things they find.
Although there is no data on the absorption of arsenic from the gastrointestinal tract
of children to compare to adult absorption, some evidence suggests that children are less
efficient at methylating (ATSDR, 2000); therefore, children may have a different tissue
distribution and longer retention times which may increase their susceptibility. As noted
above adults have been demonstrated to excrete 40-60% of the arsenic as DMA, 20-25% as
inorganic arsenic, and 15-25% as MMA, but Concha et al. (1998b) determined that
children ingesting 200 ug/L arsenic in their drinking water excreted about 49% as
inorganic arsenic and 47% as DMA. Women in the same study were determined to excrete
66% of the arsenic as DMA and 32% as inorganic arsenic.
When considering infants and children, NRC (2001) addressed primarily the issues
of exposure differences (due to increased water consumption rates) and the intrinsic
susceptibility associated with differences in arsenic metabolism. Exposure differences are
not an issue addressed in this Toxicological Review; however, age-related differences in
arsenic metabolism are relevant. The NRC reported that previous studies indicate that
young individuals may have a lower rate of methylation of inorganic arsenic than adults,
but there is variation among the data and between the populations assessed. The NRC
concluded that it is not known what impact such differences in methylation with age would
have on arsenic toxicity. It is plausible that different or multiple modes of carcinogenic
action are operative at different life-stages due to intrinsic age-related differences.
The cancer data used in the current analysis is derived from various studies of the
arseniasis-endemic area of southwest Taiwan (Chen, et al., 1985. 1988, 1991; Wu, 1989).
Data from this area illustrate that childhood cancer is limited (Chen, et al., 1992); this
analysis used 271,530 person-years for individuals <30 years of age, and identified no
cases of bladder cancer and five cases of lung cancer in this population (as well as no cases
of kidney cancer and six cases of liver cancer). These numbers and the accompanying
mortality rates (per 100,000) are low when compared with higher rates in older individuals.
It has been illustrated that there is an increasing dose-response relationship
between increased cancer mortality and increased years of exposure to the high-arsenic
artesian well water of Southwest Taiwan (Chen, et al., 1986), so it is important to consider
the extent to which the exposures in assessed population included childhood exposures.
The analysis of cancer potential in the same population (Chen, et al.,-1992) included "only
residents who had lived in the study area after birth," and assumed that the arsenic intake of
each person continued from birth to the end of the follow-up period (1973-1986)'. No
information was provided on the exposure of pregnant women in this population to the
1 The artesian wells were introduced in 1910-1920, prior sources of fresh water included ponds, streams, or
rainwater (Tseng, 1968).
DRAFT - DO NOT CITE OR QUOTE 57
-------
artesian well water, and arsenic is believed to pass through the placenta (Hanlon and Perm,
1977; Lindgren et al., 1984; Hood et al., 1987; Concha et al., 1998b).
Chen, et al. (1992) stated that their results on cancer potential may somewhat
underestimate risk in this population because tap water (with lower arsenic concentrations)
was introduced into the.area in 1956, and was available for almost 75% of the residents in
the 1970s. As such, the actual arsenic ingestion may be lower than estimated as residents
switched away from the high-arsenic artesian wells to alternate water sources. Also, this
study is based on mortality records (1973-1986) from the study region, so cancer incidence
among individuals exposed during childhood and early adulthood, who then migrated from
the region would not be captured. Chen, et al. (1986) reported that the 1982 migration rate
for this area was 27%, with primarily the youths and young adults leaving the area to move
to cities, and those 45+ years emigrated at a rate lower than 6%. There is limited migration
to this region, as it has been reported than more than 90 percent of the local residents lived
in the study area all their lives (Wu, et al., 1989).
The recent EPA Supplemental Guidance for Assessing Susceptibility from Early-
Life Exposure to Carcinogens (EPA, 2005b) identifies age-dependent adjustment factors to
be applied to carcinogens with a mutagenic mode of action when chemical-specific data are
unavailable. As outlined above, chemical-specific early-life data are being used in the
current assessment of arsenic carcinogenicity because the epidemiological data include
childhood exposures of the southwest Taiwanese study population.
4.7.2. Possible Gender Differences
Some differences in methylation have been noted between men and women
(Hopenhayn-Rich et al., 1996b). Men had higher MMA:DMA ratios than women (0.23 vs
0.17, respectively). In addition, Concha et al. (1998a) demonstrated that pregnant women
in their third trimester excrete more than 90% DMA in plasma and urine. This indicates
possible hormonal effects of arsenic methylation.
4.7.3. Other
Genetic polymorphism
Although most humans excrete 10-30% of consumed arsenic as inorganic arsenic,
10-20% as MMA, and 60-80% as DMA, some populations seem to have a somewhat
different distribution. A study of urinary arsenic in a population in northern Argentina
exposed to arsenic via drinking water demonstrated an average of only 2% MMA in the
urine (Vahter et al, 1995; Concha et ah, 1998). Studies on populations in San Pedro and
Toconao in northern Chile demonstrated that differences in the ratio of MMA:DMA
excretion between the two populations (Hopenhayn-Rich et al., 1996b). However, Chiou
ct al. (1997) found that in a population in northeastern Taiwan, 27% of the arsenic
consumed was excreted as MMA. Although these variations have not been unequivocally
DRAFT - DO NOT CITE OR QUOTE 58
-------
linked with genetic factors as opposed to environmental factors, human polymorphism has
been reported for other methyltransferases (e.g., thiopurine S-methyltransferase; Yates et
al., 1997).
Nutritional Status
In many of the studies listed above concerning high levels of arsenic exposure in
relation to severe arsenic-related health effects (i.e., southwestern Taiwan and northern
Chile) the inhabitants were of a low socioeconomic level and had poor nutritional sjtatus.
Many of the studies listed above have related poor nutritional status with possible increases
in adverse health effects. Mazumder et al. (1998) demonstrated that people in and around
West Bengal who had body weights below 80% for their age and sex had an increase (2.1
for females and 1.5 for males) in the prevalence of keratosis. In addition, selenium has
been demonstrated to reduce the teratogenic, clastogenic, and cytogenic effects of arsenic
(ATSDR, 1993). Kreppel et al. (1994) demonstrated that zinc protects mice against acute
arsenic toxicity.
5. DOSE-RESPONSE ASSESSMENTS
5.1. ORAL REFERENCE DOSE (RfD)
5.1.1. Choice of Principal Study and Critical Effect
Not addressed in this document.
5.1.2. Methods of Analysis
Not addressed in this document.
5.1.3. RfD Derivation
Not addressed in this document.
5.1.4. Previous Oral Assessment
Not addressed in this document.
5^ INHALATION REFERENCE DOSE (RfC)
Not iudrcsscu in this document;
DRAFT - DO NOT CITE OR QUOTE 59
-------
5.3. CANCER ASSESSMENT (Oral Exposure)
5.3.1. Choice of Study / Data - with Rationale and Justification
As stated by NRC (2001), there are few animal carcinogenicity bioassays on arsenic
with no positive animal models for dose-modeling. There is, however, a large database on
human exposure to arsenic. It was concluded by NRC (2001) that the southwestern Taiwan
data was the critical data set for conducting a quantitative risk assessment for exposure to
arsenic in drinking water. Data provided in Chiou et al. (2001) and Ferreccio et al. (2000)
were also considered, but were not used due to the large confidence intervals, which are not
precise enough for quantified risk assessment.
Morales et al. (2000) established bladder, lung, and liver cancer risk for mortality in
the southwestern Taiwanese population based on the same dataset that was analyzed by
NRC (1999). For the purpose of this report, a re-evaluation of Morales et al. (2000) was
performed. Morales et al. (2000) used the generalized linear model and the multi-stage-
Weibull model on data derived from an arseniasis-endemic area of southwest Taiwan that
was obtained from Chen et al. (1988), Wu et al. (1989), and Chen et al. (1992) to establish
risk estimates for cancers of the bladder, liver, and lung from arsenic exposure; NRC
(2001) performed a separate analysis based on the data in Morales et al. (2000). Therefore,
the data presented in Morales et al. (2000), with an unexposed (assumed zero dose of
arsenic in drinking water) southwest Taiwanese population as a control group, was chosen
as the data for analysis.
Lung and bladder cancer have been chosen as the endpoint because they are the
internal cancers most consistently seen and best characterized in epidemiology studies
(NRC, 2001). The choice of this endpoint is consistent with the NRC reports (1999,
2001). It has been decided that the oral slope factor will combine the lung and bladder
cancer results. Because cause of death is listed as only from one cause (e.g., either lung or
bladder cancer), there is no double counting; therefore, combining the two cancers will
account for deaths from both types of cancer. There are several possible modes of action
that make it plausible that different types of cancers occur. Although liver cancers were
also examined by Morales et al. (2000), they were not examined in the current report
because liver cancers were considered to be affected by a lot of confounding factors. Skin
cancer has also been noted in the Taiwan population, but skin cancer is generally an
incidence instead of mortality. In addition, skin cancer may be influenced by external
exposure through bathing.
Because there is a high general background level of inorganic arsenic in food, it has
been suggested that effective exposures to arsenic in Taiwan are higher than represented
simply by the amount of well water drunk. For this paper, the issue of intake of arsenic
from food (e.g., dry rice, sweet potatoes) has been distinguished from the issue of intake of
arsenic from drinking water and intake of arsenic from water used in cooking, such as
water used to boil rice and potatoes. To account for background levels of arsenic in food,
DRAFT - DO NOT CITE OR QUOTE 60
-------
U.S. EPA (2001) assumed that the inorganic arsenic consumption due to food in Taiwan
was 50 (ig/day, compared to 10 jig/day in the United States. NRC (1999) cited results of
Schoof et al. (1998) as estimating Taiwanese daily intake from yams as 31 jig/day and rice
as 19 ug/day. NRC (1999) also noted that the Li et al. (1979) study found 95% of the rice
crop to contain arsenic primarily 100 to 700 ug/kg, with some up to 1.43 mg/kg. The soil
had probably been treated with arsenical pesticides. The NRC (2001) found little evidence
to support EPA's assumption that food contributed 50 mg/day of inorganic arsenic to
the Taiwanese diet. NRC addressed the issue by determining how sensitive the
calculation of ED0, was to the consumption rate. NRC (2001) found that changing
the consumption rate from 50 ug/day to 30 ng/day did not change the calculated
ED0: significantly (about 1% difference). This lack of sensitivity was not
unexpected, since the southwestern Taiwanese population, which was used as a
comparison group, had a similar dietary intake as the exposed population. It was
concluded that the model should be run over a range of arsenic consumption rates
to confirm that the calculated risk is insensitive to this value.
The NRC (1999, 2001) concluded that the selection of the drinking water
consumption rates should consider, as possible, the uncertainty associated with
trying to accurately determine the mean consumption rate of the populations, and of
the variability of individuals within the populations. This document agrees that the
model is highly sensitive to the term selected. Review of relevant literature
suggests that the mean adult drinking water consumption rate is between 1 to 4.6
L/day. Based on an informed judgment about the most plausible values for water
intake in the Taiwanese study population, a narrower range of 2 to 3.5 L/day may
be appropriate.
5.3.2. Dose-response Data
Male data from Morales et al. 2000 (Table 1)
Hg/L of
arsenic
<100
100-299
300-599
*600
person-years at
risk
95,455
47,268
72,068 •
42,179
Bladder
cancer
17
9
32
27
Liver
cancer
31
23
39
29
Lung cancer
28
30
53
33
Female data from Morales et al. 2000 (Table 1)
Hg/L of
arsenic
<100
100-299
300-599
*600
person-years at
risk
86,975
43,212
64,903
38,869
Bladder
cancer
21
11
30
28
Liver
cancer
12
14
13
12
Lung cancer"
29
19
36
38
DRAFT - DO NOT CITE OR QUOTE
61
-------
5.3.3. Dose Conversion
Conversion was done by assuming that equivalent doses of arsenic in the US
and Taiwan should correspond to the same amount in units of u.g/kg/day. Drinking
water concentrations of arsenic in units of ng/L (ppb) in Taiwan were adjusted for
dietary intake of 0, 10, 30, or 50 fig/day by dividing the dietary intake by
Taiwanese daily drinking water intake (2 or 3.5 L/day; stated in NRC, 2001 to be
the estimated consumption for women and men in Taiwan, respectively) and adding
it to the nominal drinking water concentration. All nonlinear optimization was
calculated based upon this adjusted Taiwanese concentration. Results were then
converted to fig/kg/day using Taiwanese body weight (50 kg) and drinking water (2
or 3.5 L/day) assumptions and converted back to US-relevant drinking water
concentrations in u.g/L using US body weight (70 kg) and drinking water (2 L/day)
assumptions.
Cus = CMdj + 50 kg x vdw x 70 kg * 2 L/day
where C,ndJ is the adjusted Taiwanese arsenic concentration, C, is the nominal
Taiwanese arsenic concentration, As,-ood is the assumed daily dietary arsenic intake
for Taiwanese, Vdwis the assumed daily drinking water consumption for Taiwanese,
and Cus is the U.S. -relevant arsenic concentration.
In the Morales et al. (2000) and NRC (2001) analyses, the control group was
assumed to have no dietary intake of arsenic. In contrast, this analysis assumes that
the control and study groups are exposed to the same level of dietary arsenic.
5.3.4. Extrapolation Method(s)
In 2001, an update of Arsenic in Drinking Water (NRG. 2001) was
published, which described the statistical methodology that they used to calculate
cancer risk estimates for arsenic from epidemiological data. This is a condensed
discussion of the adaptation, application, and refinement of the methodology.
Analysis of human epidemiological cancer data needs to 1) adjust for effects
of age on cancer incidence, and 2) adjust for differences in age-related cancer
incidence between the study population (in this case, southwest Taiwanese) and the
intended reference population (i.e., US population).
In order to facilitate these adjustments, a specific functional form is used to
fit the data and extrapolate for low-level cancer risk. This form is the Poisson
model:
h(x,t) = h0(t) x g(x)
where h(x,t) is the risk at dose x and age t, h0(t) is the baseline risk of the reference
population at age t, and g(x) is the additional risk attributable to exposure at dose
x. Neither h0(t) nor g(x) is a single value; both are mathematical functions. When
these h0(t) and/or g(x) are best described by nonlinear functions, as is the case for
arsenic, complex statistical programming is required to solve this equation and
calculate risk.
DRAFT - DO NOT CITE OR QUOTE 62
-------
In order to provide the best fit for the data, an additive linear model
represented by the following functions were used for h0(t) and g(x):
Age effect
ho(t)
Quadratic
exp(a0+a,t+a2t2)
Dose
transformatio
n x
Identity
x=ppb
Dose effect
g(x)
Linear
1+P, x
where cc0, a,, a2, and p are constants to be estimated, and x is the arsenic dose
adjusted for dietary arsenic exposure. Adjustments for dietary exposure were made
at 30 ug in food per day. The study reference population was assumed to be
southwest Taiwanese. Nonlinear optimization was done using the statistical
programming language R (version 1.8.0) and the optimization routine "optim."
This is in contrast to the use of S-plus and "nlminb" in Morales et al. (2000) and
NRC (2001). It is conceivable that differences between the two optimization
routines could lead to differences in both the mean values and lower confidence
limits obtained.
Once the fitted constants are calculated, the resulting benchmark doses at
1% and 5% incidence may be calculated for the southwest Taiwanese population.
However, the southwest Taiwanese population is not the population of interest. In
order to obtain benchmark doses relevant to an American population, the values
must be adjusted based upon both an unexposed (assumed zero dose of arsenic in
cooking and drinking water) southwest Taiwanese population as a control group
and 1996 U.S. lifetables for males and females (Vital Statistics of the U.S., 1996).
This process is described in detail in Arsenic in Drinking Water (2001). The BEIR-
IV forrhula, corrected for two errors in NRC (2001), is
R(x) =
lifetime risk of death at dose x
(1)
where:
s,
age group (in 5-year intervals: 20-25, ..., 85-90)
cancer mortality hazard, age /, dose x
total mortality hazard, age /, dose x
/•(survive ;-th time interval | survive (i-l)-th time
interval) ^
exp(-5 /J,'(x))
P(survive up to end of (M)-th time interval)
/-i
(2)
(3)
Given a linear relative hazard model:
h,(x) =
DRAFT - DO NOT CITE OR QUOTE
(4)
63
-------
h,(x) = h,(l+ftx) (4)
where ht =/i,(0) and p is the relative risk slope estimated by regression. The
additional dose-related cancer hazard is hjfix, so the total mortality hazard
increases at the same time to h'(x) = h'+h^x where again h' = h, (0). Therefore,
(1) becomes
The two typographical errors in NRC (2001) were:
h, instead of h' in equation (2);
• (1 -
-------
Oral Slope Factor1: 5.7E-6 (mg/kg-day)"'
Drinking Water Unit Risk: 1.6E-04 (ng/L) •'
Drinking Water Concentrations at Specified Risk Levels, using default assumptions
of 70 kg body weight and 2 L/day water ingestion:
E-4(l in 10,000) 6.3E-1 ug/L
E-5 (1 in 100,000) 6.3E-2 ug/L
E-6 (1 in 1,000,000) 6.3E-3 ug/L
During its deliberations on arsenic, there were significant discussions among
the various Offices about the water and food consumption parameters that should
be used to calculate the slope factor(s) for arsenic. The assumption of 2 L/day
water consumption was used. For inorganic arsenic exposure from the diet, three
values were considered; 1) 0 ug As/day, 2) 30 ug As/day (NRC selection), and 3)
50 ng As/day (average inorganic arsenic exposure from diet of Taiwanese; Schoof
et al., 1998). The National Research Council selected 2 L/day and 30 ng As/day,
although no reason was given for their selection of the food value. For clarity, the
slope factors for each of these permutations are given in Table 5-4 through 5-7.
The Excel spreadsheet for the calculations is available at
www. epa. go v/waterscience.sab. The various assumptions for dietary arsenic
intake have slight effects on the derived slope factor, while the two different
assumptions for drinking water consumption lead to two-fold differences in the
derived slope factor. The EPA has chosen 2 L drinking water/day and 30 ug dietary
As/day in derivation of the recommended oral slope factor for inorganic arsenic, as
is consistent with NRC (2001).
In the Morales et al. (2000) and NRC (2001) analyses, the control group was
assumed to have no dietary intake of arsenic. The effects of different assumptions
of dietary intake were limited to the dosed groups. Age-dependent and
nonspecified effects on cancer incidence were largely fixed by the large control
population. Therefore, NRC (2001, p. 196) found: "approximately a 1% increase in
ED (effective dose) estimates" when they added 30 ug/day of arsenic in diet. This
analysis follows the Morales et al. (2000) assumption, with no dietary intake of
arsenic for the control population.
The oral slope factor was calculated using an assumption of 30ug/day for dietary intake and 2L/day for drinking
water consumption
DRAFT - DO NOT CITE OR QUOTE 65
-------
Table 5-4. Calculation for Male Lung Cancer Risk from Arsenic in the
Drinking Water
dietary
adjust.
(Hg/day)
50
30
0
drinking
water
volume
(M
2
2
2
ED0,
(Hg/L)
96
90
90
LED01
(M/L)
87
82
83
slope
factor (x
10"')
(mg/kg-
day)'1
40
43
42
unit risk
(xIO-5)
(Hg/L)"'
11
12
12
ppb
at
1E-4
risk
0.87
0.82
0.83
Table 5-5. Calculation for Female Lung Cancer Risk from Arsenic in the
Drinking Water
dietary drinking
adjust. water
(u,g/day) volume
(L)
ED0
LED0
50
30
0
82
80
. 79
76
72
70
slope
factor (x
10'')
(mg/kg-
day)'1
46
49
50
unit risk
(xlO's)
13
14
ppb
at
1E-4
risk
0.76
0.72
0.70
DRAFT - DO NOT CITE OR QUOTE
66
-------
Table 5-6. Calculation for Male Bladder Cancer Risk from Arsenic in the
Drinking Water
dietary drinking ED0, LED,,
adjust.
slope
(Hg/day) volume
(L)
50
30
0
Id'7)
(mg/kg-
2
2
2
320
314
304
286
282
271
day)'1
12
12
13
unit risk ppb
water (ug/L) (ug/L) factor (x (xlO'5)
(Hg/L)'1
3.5
3.5
3.7
at
1E-4
risk
2.86
2.82
2.71
Table 5-7. Calculation for Female Bladder Cancer Risk from Arsenic in the
Drinking Water
dietary drinking
adjust. water
(jag/day) volume
(L)
50
30
0
>6
e
2
2
2
EDOI
(ug/L)
474
472
465
LEDOI
(Ug/L)
437
421
415
slope
factor (x
ID'7)
(mg/kg-
day)"'
8.0
8.3
8.4
unit risk
(xlO5)
((jg/L)'1
.23
2.4
2.4
ppb
at
1E-4
risk
4.37
4.21
4.15
Table 5-8. Combined Bladder and Lung Cancer Risk from Arsenic in Drinking Water
Sex
Site
Female Bladder
Lung.
Combined
Male Bladder
Lung
Combined
slope
factor
(x 10-')
(mg/k
g-day)-1
8.3
49
57
12
43
55.0
unit risk
(xlO'5)
(ug/L)-'
2.4
14
16
3.5
12
16
ppb at
1E-4
risk
4.21
0.72
0.63
2.82
0.82
0.63
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD
AND DOSE RESPONSE
6.1. HUMAN HAZARD POTENTIAL
There have been numerous studies examining the carcinogenic potential of
inorganic arsenic via oral exposure. The majority of these studies are ecological in
nature and are therefore subject to the biases of ecological studies (e.g., lack of
individual exposure). The strength of the associations observed in the studies
mentioned above determine an association between oral exposure to inorganic
arsenic and cancer. Each study was conducted differently and contained their own
biases (e.g., lack of confounding variables, possible recall bias), but combination of
DRAFT - DO NOT CITE OR QUOTE
67
-------
all the study results supports an association between oral exposure to inorganic
arsenic and cancer including bladder, kidney, lung, liver, and prostate. Leukemia
and lymphomas, colon cancer, and bone cancer have been associated with arsenic
exposure in a few studies (Chen et al., 1985, 1988; Tsai, 1999; Wu, 1989).
Different populations were assessed, therefore, it is unlikely that any single
attribute (i.e., nutritional habits) associated with a single population is responsible
for the increased cancer rates. Although many studies did not account for
confounding variables (e.g., cigarette smoking in association with lung cancer), the
positive associations also were observed in those studies that did account for
certain confounding variables (e.g., habits, age, socioeconomic status). Details on
the specific studies are provided above in Section 4.1.
The major flaw in the association between oral exposure to inorganic arsenic
and skin cancer was the lack of data and/or analysis of dermal contact with the
water containing the inorganic arsenic through bathing. The studies that failed to
observe an association between oral exposure to inorganic arsenic and bladder
cancer (Bates et al., 1995; Lamm et al., 2004; Steinmaus et al., 2003) examined low
exposures. However, Bates et al. (1995) and Steinmaus et al. (2003) did find an
association between arsenic exposure (exposures 20 years or more) and bladder
cancer in people who had ever smoked. Lung and liver tumors also were observed
in mice administered inorganic arsenic for a short duration transplacentally with
possible additional exposure while nursing (pregnant dams were exposed for 10
days during gestation only) (Waalkes et al., 2003, 2004a). Therefore, increasing
the evidence that lung and liver cancers are associated with oral exposure to
inorganic arsenic.
The confidence in the data is strong because the assessment is based on lung
cancer mortality in humans (females). The data demonstrates a definite dose-
related effect in humans. Uncertainty lies in exact exposure by subjects primarily
due to the fact that this is an ecological study. The calculations did integrate the
possibility of arsenic exposure from other sources (i.e., food; assumption of 30
Hg/day). There is also uncertainty in the amount of water consumed/day by
Taiwanese females (2.0 or 3.5 L).
6.2. DOSE RESPONSE
Cancer has been selected as the most sensitive endpoint The oral slope
factor varies depending on assumptions of the volume of water consumed over the
course of a day and the amount of arsenic consumed through the diet. Tables 5-4
through 5-7 provide estimates for assumptions of 2 L/day of water consumed with
0, 30 or 50 jig/day of arsenic consumed via the diet, with 2 L/day and 30 ug/day
being the assumptions recommended by the NAS (NRC, 2001). Changes in these
assumptions may change the oral slope factors (this is spelled out in more detail in
the cancer section), and would lead to changes in the concentration of arsenic in the
drinking water associated with an increased risk of cancer. Cancer assessment via
inhalation exposures was not assessed in this report.
DRAFT - DO NOT CITE OR QUOTE 68
-------
7. REFERENCES
Adams, MA, Bolger, PM, and Gunderson, EL (1994) Dietary intake and hazards of
arsenic. In: Chappell,.WR, Abernathy, CO, and Cothern, CR, eds. Arsenic:
Exposure and Health Northwood, U.K.: Science and Technology Letters. Pp. 41-49
in, W.R.(as cited in U.S. EPA, 2004b).
Ahmad, S, Sayed, MH, Barua, S, et al. (2001) Arsenic in drinking water and
pregnancy outcomes. Environm. Health. Perspect. 109(6):629-631.
Ahmad, S, Anderson, WL, and Kitchin, KT (1999) Dimethylarsenic acid effects on
DNA damage and oxidative stress related biochemical parameters in B6C3F1 mice.
Cancer Lett. 139(2):129-135 (as cited inNRC, 2001).
Ahsan, H, Perrin, M, Rahman, A, et al. (2000) Associations between drinking water
and urinary arsenic levels and skin lesions in Bangladesh. J. Occup. Environ. Med.
42(12):1195-1201.
Andersen, 0 (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 (as cited in U.S. EPA, 1998c).
Anundi, I, Hogberg, J, and Vahter, M (1982) GSH release in bile as influenced by
arsenite. FEBS Lett/ 145:285-288 (as cited in NRC, 1999).
Aposhian, HV (1997) Enzymatic methylation of arsenic species and other new
approaches to arsenic toxicity. Annu. Rev. Pharmacol. Toxicol. 37:397-419 (as
cited in NRC, 1999).
Aschengrau, A, Zierler, S, and Cohen, A (1989) Quality of community drinking
water and the occurrence of spontaneous abortion. Arch. Environ. Health
44(5):283-290.
ATSDR (2000) Arsenic. Toxicological Profile (Update). Agency for Toxic
Substances and Disease Registry, Public Health Service, US Department of Health
and Human Services, Atlanta, GA.
ATSDR (1993) Arsenic. Toxicological Profile (Update). Agency for Toxic
Substances and Disease Registry, Public Health Service, US Department of Health
and Human Services, Atlanta, GA.
Avila, MA, Berasain, C., Torres, L., et al. (2000) Reduced mRNA abundance of the
main enzymes involved in methionine metabolism in human liver cirrhosis and
hepatocellular carcinoma. J. Hepatol. 33:907-914 (as cited in Waalkes et al.,
2004b).
Barchowsky, A., Klei, LR, Dudek, EJ, et al. (1999) 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 (as cited in Huang et al.,
2004).
DRAFT - DO NOT CITE OR QUOTE 69
-------
Barrett, JC5 Lamb, PW, Wang, TC, and Lee, TC (1989) Mechanism of arsenic-
induced cell transformation. Biol. Trace Element Res 21:421-429 (as cited in
Rossman, 2003).
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.
Mutation Research 516:29-40.
Basu, A, Mahata, J, Gupta, S, and Giri, AK. (2001) Genetic toxicology of a
paradoxical human carcinogen, arsenic: a review. Mutat. Res. 488(2): 171-194.
Bates, MN, Smith, AH, and Cantor, KP (1995) Case-control study of bladder
cancer and arsenic in drinking water. Am. J. Epidemiol. 141:523-530. (California
EPA, 2003)
Beckman, G, Beckman, L, and Nordenson, 1(1977) Chromosome aberrations in
workers exposed to arsenic. Environ. Health Perspect. 19:145-146 (as cited in U.S.
EPA,1998c).
Bogdan, GN, Sampayo-Reyes, A, and Aposhian, HV (1994) Arsenic binding
proteins of mammalian systems: I. Isolation of three arsenite-binding proteins of
rabbit liver. Toxicology 93:175-193 (as cited in NRC, 1999).
Bolton, JL, Trush, MA, Penning, TM, et al. (2000) Role of quinones in toxicology.
Chcm. Res. Toxicol. 13:135-160 (as cited in Huang et al., 2004).
Bolton, JL, Pisha, E., Zhang, F, and Qiu, S (1998) Role of quinoids in estrogen
carcinogenesis. Chem. Res. Toxicol. 11:1113-1127 (as cited in Huang etal., 2004).
Borzsonyi, M, Berecsky, A, rudnai, P, et al. (1992) Epidemiological studies on
human subjects exposed to arsenic in drinking water in southeast Hungary. Arch.
Toxicol. 66(l):77-78.
Buchet, JP, Lauwerys, R, and Reels, 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 (as cited in
ATSDR, 2000; NRC, 1999).
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 (as cited in
NRC, 1999).
Buchet, JP and 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 (as cited in NRC, 1999).
Buchet, J-P and Lison, D (2000) Clues and uncertainties in the risk assessment of
arsenic in drinking water. Food and Chemical Toxicology 38:S81-S85.
Burgdorf, W, Kurvink, K, and Cervenka, J (1977) Elevated sister chromatid
exchange rate in lymphocytes of subjects treated with arsenic. Hum. Genet.
36(l):69-72 (as cited in U.S. EPA, 1998c).
DRAFT-DO NOT CITE OR QUOTE 70
-------
Burns, 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 (as cited in Rossman, 2003)
California EPA (2003) Public Health Goal for Arsenic in Drinking Water. Pesticide
and Environmental Toxicology Section Office of Environmental Health Hazard
Assessment California EPA. March 2003.
Cantoni, 0, Hussain, S, Guidarelli, A, et al. (1994) Cross-resistance to heavy
metals in hydrogen peroxide-resistant CHO cell variants. Mutat. Res. 324:1-6 (as
cited in Rossman, 2003).
Charbonneau, SM, Hollins, JG. Tam, GKH, et al. (1980) Whole-body retention,
excretion, and metabolism of [ As]arsenic acid in the hamster. Toxicol. Lett.
5:175-182 (as cited in NRC, 1999).
Charbonneau, SM, Tam, GKH, Bryce, F, et al. (1979) Metabolism of orally
administered inorganic arsenic in the dog. Toxicol. Lett. 3:107-114 (as cited in
NRC, 1999).
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 (as cited in
Rossman, 2003).
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.
Chem. Res. Toxicol. 13:342-350 (as cited in Huang et al., 2004).
Chen, C-J, Chiou, H-Y, Chiang, M-H, et al. (1996) Dose-response relationship
between ischemic heart disease mortality and long-term arsenic exposure.
Arterioscler. Thromb. Vase. Biol. 16:504-510.
Chen, C-J, Hsueh, Y-M, Lai, M-S, et al. (1995) Increased prevalence of
hypertension and long-term arsenic exposure. Hypertension 25(1):53-60.
Chen, C-J, Chen, C-W, Wu, M-M, and Kuo, T-L (1992) Cancer potential in liver,
Jung, bladder and kidney due to ingested inorganic arsenic in drinking water. Br. J.
Cancer 66(5):888-892 (as cited in U.S. EPA, 1998c).
Chen, CJ and Wang, CJ (1990) Ecological correlation between arsenic level in well
water and age-adjusted mortality from malignant neoplasms. Cancer Res. 50:5470-
5474.
Chen, C-J, Wu, M-M, Lee, S-S, et al. (1988) Atherogenicity and carcinogenicity of
high-arsenic artesian well water. Arteriosclerosis 8:452-460.
Chen, CJ, Chuang, YC, Lin, TM, and 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. (as cited in NRC, 1999; U.S. EPA,
1998c)
DRAFT - DO NOT CITE OR QUOTE 71
-------
Chen, KP, Wu, HY, and Wu, TC (1962) Epidemiologic studies on blackfoot disease
in Taiwan. 3. Physicochemical characteristics of drinking water in endemic
blackfoot disease areas. Mem. Coll. Med. Natl. Taiwan Univ. 8:115-129 (as cited
inNRC, 1999).
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.
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 (as cited in
NRC, 1999).
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. (As cited in NRC, 1999)
Chou, W-C, Hawkins, AL, Barrett, JF, et al. (2001) Arsenic inhibition of
telomerase transcription leads to genetic instabilty. J. Clin. Inves. 108:1541-1547.
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.
Concha, G,Nermell3 B, and Vahter, M (1998a) Metabolism of inorganic arsenic in
children with chronic high arsenic exposure in northern Argentina. Environ. Health
Perspect. 106:355-359 (as cited in NRC, 1999).
Concha, G, Vogler, G, Lezeano, D, et al. (1998b) Exposure to inorganic arsenic
metabolites during early human development. Toxicol. Sci. 44:185-190 (as cited in
NRC, 1999).
Concha, G, Vogler, G, Nermell, B, and 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, and Smith, H (1979) Industrial exposure to
arsenic. J. Radioananl. Chem. 48:197-208 (as cited in NRC, 1999).
Crbssen, PE(1983) Arsenic and SCE in human lymphocytes. Mut. Res. 119:415-
419 (as cited in U.S. EPA 1998c).
Cuzick, J, Evans, S, Oilman, M, and Prince Evans, D (1982) Medicinal arsenic and
internal malignancies.. Br. J. Cancer 45:904-911 (as cited in Kayajanian, 2003).
Cuzick, J, Sasieni, P, and Evans, S (1992) Ingested arsenic, keratoses, and bladder
cancer. Am. J. Epidemiol. 136:417-421 .
Das, D., G. Samantha, B.K. Mandal, T.R. Chowdhury, C.R. Chanda, P.P.
Chowdhury, G.K. Basu and D. Chakraborti. (1996) Arsenic in ground water in six
districts in West Bengal, India. Environ. Geochem. Health 18:5-15 (as cited in
NRC, 1999).
DRAFT - DO NOT CITE OR QUOTE 72
-------
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 (as cited in NRC,
1999). . .
Datta, S, Talukder, G, and Sharma, A (1986) Cytotoxic effects of arsenic in dietary
oil primed rats. Sci. Culture. 52:196-198 (as cited in U.S. EPA, 1998c;NRC, 1999;
ATSDR.2000).
De Kimpe, J, Cornells, R, Mees, L, and Vanholder, R (1996) Basal metabolism of
intraperitonelly injected carrier-free74As-labeled arsenate in rabbits. Fundam.
Appl. Toxicol. 34:240-248 (as cited in NRC, 1999).
De Kimpe, J. Cornells, R, Mees, L, et al. (1993) More than tenfold increase of
arsenic in serum and packed cells of chronic hemodialysis patients. Am. J. Nephrol.
13:429-434. (As cited in NRC, 1999)
Del Razo, LM, Styblo, M, Cullen, WR, and Thomas DJ (2001) Determination of
trivalent methylated arsenicals in biological matrices. Toxicol. Appl. Pharmacol.
174:283-293.
Delnomdedieu, M, Styblo, M, and Thomas, DJ (1995) Time dependence of
accumulation and binding of inorganic and organic arsenic species in rabbit
erythrocytes. Chem.-Biol. Interact. 98:69-83 (as cited in NRC, 1999).
Delnomdedieu, M, Basti, MM, Otvos, JD, and Thomas, DJ (1994a) Reduction and
binding of arsenate and dimethylarsinate by glutatione: a magnetic resonance study.
Chem.-Biol. Interact. 90:139-155 (as cited in NRC, 1999).
Delnomdedieu, M, Basti, MM, Styblo, M, et al. (1994b) Complexation of arsenic
species in rabbit erythrocytes. Chem. Res. Toxicol. 7:621-627 (as cited in NRC,
1999).
Ddsi, I (1992) Arsenic contamination of drinking water in south-east Hungary.
Georgr. Med. 22:45-53 (as cited in NRC, 1999).
DiPaolo, JA and Casto, BC (1979) Quantitative studies of in vitro morphological
transformation of Syrian hamster cells by inorganic metal salts. Cancer Res.
39:1008-1013 (as cited in U.S. EPA, 1998c).
Dong, J-T, and 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 (as cited in U.S. EPA, 1998c).
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.
DRAFT - DO NOT CITE OR QUOTE 73
-------
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, Jaspers, I, Thomas, DJ, and Styblo, M (2002) Differential activation of
AP-1 in human bladder epithelial cells by inorganic and methylated arsenicals.
FASEBJ17(l):67-69.
Eastmond, DA and Tucker, JD (1989) Kinetochore localization in micronucleated
cytokinesis-blocked Chinese hamster ovary cells a new and rapid assay for
identifying aneuploidy-inducing agents. Mutat Res 224:517-525 (as cited in
Rossman, 2003).
Engel, RR and Smith, AH (1994) Arsenic in drinking water and mortality from
vascular disease: an ecological analysis in 30 counties in the United States.
Archives of Environmental Health 49(5):418-427.
Falk, H, Herbert, J, Crowley, S, et al. (1981) Epidemiology of hepatic
angiosarcoma in the Unites States: 1964-1974. Environ. Health Perspect. 41:107-
113 (as cited in NRC, 1999).
Fan, S-R, Ho, I-C, Yeph, FL-F, et al. (1996) Squalene inhibits sodium arsenite-
induced sister chrbmatid exchanges and micronuclei in Chinese hamster ovary-K 1
cells. Mutat. Res. 368:165-169 (as cited in Rossman, 2003).
Fenech, M, Crott, J, Turner, J, and Brown, S (1999) Necrosis, apoptosis, cytostasis
and DNA damage in human lymphocytes measure simultaneously within the
cytokineses-block micronucleus assay: description of the method and results for
hydrogen peroxide. Mutagenesis 14:605-612 (as cited in Rossman, 2003).
Perm, VH and Hanlon, DP. (1985) Constant rate exposure of pregnant hamsters to
arsenate during early gestation. Environ. Res. 37:425-432 (as cited in NRC, 1999).
Perm, VH, Saxon, A, and Smith, BM (1971) The teratogenic profile of sodium
arsenate in the golden hamster. Arch. Environ. Health 22:557-560 (as cited in NRC,
1999).
Perm, VH, and Carpenter, SJ (1968) Malformations induced by sodium arsenate. J.
Repro. Pert. 17:199-201 (as cited in NRC, 1999).
Ferreccio, C, Gonzalez, C, Milosavjlevic, et al. (2000) Lung cancer and arsenic
concentrations in drinking water in Chile. Epidemiology 1:673-679.
Fiertz, U (1965) Katamnestische untersuch uber die nebenwirkungen der therapie
mit anorganiscnem arsen bei hautkrankheiten. Dermatologica 131:41-58 (as cited in
California EPA, 2003).
Flora, SJ (1999) Arsenic-induced oxidative stress and its reversibility following
combined administration of N-acetylcystein and meso 2,3-dimercaptosuccinic acid
in rats. Clin. Exp. Pharmacol. Physiol. 26:865-869 (as cited in Huang et al., 2004).
DRAFT - DO NOT CITE OR QUOTE 74
-------
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 (as cited in ATSDR,
2000 and California EPA, 2003).
Gebel, T (2000) Confounding variables in the environment toxicology of arsenic.
Toxicology 144:155-162.
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, Doorman, GA, et al. (1997) Arsenic can mediate skin
neoplasia by chronic stimulation of keratinocyte-derived growth factors. Mutat.
Res. 386:209-218 (as cited in Rossman, 2003).
Geubel, AP, Mairlot, MC, Buchet, JP, and Lauwerys, R (1988) Abnormal
methylation capacity in human liver cirrhosis. Int. J. Clin. Pharmacol. Res. 8:117-
122 (as cited in NRC, 1999).
Gonsebatt, ME, Vega, L, Salazar, AM (1997). Cytogenic effects in human exposure
to arsenic. Mutat. Res. 386:219-228 (as cited in Kitchin, 2001).
Gregus, Z and Nemeti, B. (2005) The glycolytic enzyme glyceraldehyde-3-
phpsphate dehydrogenase works as an arsenate reductase in human red blood cells
and rat liver cytosol. Toxicol. Sci. 85(2):859-869.
Gregus, Z and Nemeti, B. (2002) Purine nucleoside phosphorylase as a cytosolic
arsenate reductase. Toxicol. Sci. 70:13-19.
Guo, MR, Chiang, MS, Hu, H, et al. (1997) Arsenic in drinking water and incidence
of urinary cancers. Epidemiology 8:545-550 (as cited in NRC, 1999).
Guo, HR (2004) Arsenic level in drinking water and mortality of lung cancer
(Taiwan). Cancer Causes Control 15(2): 171-177.
Guyton, KZ, Xu, QB, and Holbrook, NJ (1996) Induction of the mammalian stress
response gene GADD153 by oxidative stress: role of AP-1 element. Biochem. J.
314:547-554 (as cited in Rossman, 2003).
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 (as cited in Huang et al., 2004; Rossman, 2003).
Hanlon, DP and Perm, VH (1977) Placental permeability of arsenate ion during
early embryogenesis in the hamster. Experientia 33(9):1221-1222 (as cited in
ATSDR, 2002).
Harrington-Brock, K, Smith, TW, Doerr, CL, et al. (1993) Mutagenicity of the
human carcinogen arsenic and its methylated metabolites monomethylarsonic and
dimethylarsenic acids in L5178Y TK+/- mouse lymphoma cells (abstract). Environ.
Mol. Mutagen. 21(Supplement 22):27 (as cited in U.S. EPA 1998c).
DRAFT - DO NOT CITE OR QUOTE 75
-------
Harrington, JM, Middaugh, JP, Morse, DL, and 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 (as cited in NRC, 1999).
Hcaly, SM, Wildfang, E, Zakharyan, RA, and Aposhian, HV (1999) Diversity of
inorganic arsenite biotransformation. Biol. Trace Elem. Res. 68(3):249-266 (as
cited in NRC, 2001).
Mealy SM, Casarez, EA, Ayala-Fierro, F, and Aposhian, HV (1998) Enzymatic
methylation of arsenic compounds: V. Arsenite methyltransferase activity in tissues
of mice. Toxicol. Appl. Pharmacol. 148:65-70 (as cited in NRC, 1999).
Healy, SM, Zakharyan, RA, and 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 (as cited
in NRC, 2001 and ATSDR, 2002).
Hei, TK, LIU, SX, and Waldren, C. (1998) Mutagenicity of arsenic in mammalian
cells: role of reactive oxygen species. ProcNatl Acad Sci USA 95:8103-8107 (as
cited in Rossman, 2003).
Heydorn, K (1970) Environmental variation of arsenic levels in human blood
determined by neutron activation analysis. Clin. Chem. Acta 28:349-357 (as cited
in NRC, 1999).
Hinwood, AL, Jolley, DJ, and 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 (as cited in NRC, 2001).
Holson, JF, Stump, DO, Clevidence, KJ, et al. (2000) Evaluation of the prenatal
developmental toxicity of orally administered arsenic trioxide in rats. Food Chem.
Toxicol. 38:459-466 (as cited in ATSDR, 2002).
Hood, RD (1998) Developmental effects of methylated arsenical metabolites in
mice. Bull. Environ. Contamin. Toxicol. 61:231-238 (as cited in Thomas et al.,
2001; ATSDR, 2002).
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 (as cited in NRC, 1999 and
ATSDR, 2002).
Hood, RD, Thacker, GT, and Patterson, BL (1977) Effects in the mouse and rat of
prenatal exposure to arsenic. Environ. Health Perspect. 19:219-222 (as cited in
ATSDR, 2002).
Hopenhayn, C, Ferreccio, C, Browning, SR, et al. (2003) Arsenic exposure from
drinking water and birth weight. Epidemiology 14(5):593-602.
DRAFT - DO NOT CITE OR QUOTE 76
-------
Hopenhayn-Rich, C., Browning, SR, Hertz-Picciotto, I, et al. (2000) Chronic
arsenic exposure and risk of infant mortality in two areas of Chile. Environmental
Health Perspectives 108(7):667-673.
Hopenhayn-Rich, C, Hertz-Picciotto, I, Browning, S, et al. (1999) Reproductive
and developmental effects associated with chronic arsenic exposure. In: Chappell,
WR, Abernathy, CO, and Calderon, RL, eds. Arsenic Exposure and Health Effects.
Oxford: Elsevier.
Hopenhayn-Rich, C, Biggs, ML, Smith, AH (1998) Lung and kidney cancer
mortality associated with arsenic in drinking water in C6rdoba, Argentina.
International Journal of Epidemiology. 27:561-569.
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 (as
cited in NRC, 1999).
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, Smith, AH, and Goeden, HM (1993) Human studies do not
support the methylation threshold hypothesis for the toxicity of inorganic arsenic.
Environ. Res. 60:161-177 (as cited in NRC, 1999).
Hostynek, JJ, Hinz, RS, Lorence, CR, etal. (1993) Metals and the skin. Crit. Rev.
Toxicol. 23:171-235 (as cited in NRC, 1999).
HSDB (2005) Hazardous Substance Data Bank. National Library of Medicine.
Records for element arsenic and arsenic compounds. Retrieved from the Internet on
July 18, 2005. Bethesda, MD:NLM, HSDB.
Hu, Y, Su, L, and 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 (as cited in Huang et al., 2004).
Huang, C, Ke, Q, Costa, M, and Shi, X (2004) Molecular mechanisms of arsenic
carcinogenesis. Molecular and Cellular Biochemistry 255:57-66.
Huang, Q, Ma, WY, Li, J, et al. (1999) Requirement of Erk, but not JNK, for
arsenite-induced cell transformation. J. Biol. Chem. 274:14595-14601 (as cited in
Huang et al., 2004 and Rossman, 2003).
Huang, R.N. and T.C. Lee. (1996) Cellular uptake of trivalent arsenite and
pentavalent arsenate in KB cells cultured in phosphate-free medium. Toxicol. Appl.
Pharmacol. 136(2):243-249 (as cited in NRC, 2001; Rossman, 2003).
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 (as cited in Rossman, 2003)
DRAFT - DO NOT CITE OR QUOTE 77
-------
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 (2002) Arsenic toxicity and potential mechanisms of action.
Toxicology Letters 133:1-16.
Hughes, MF, Del Razo, LM, and 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 (as cited inNRC, 2001).
Hughes, M, Kenyon, E, Edwards, B, et al. (1999) Strain-dependent disposition of
inorganic arsenic in the mouse. Toxicology 137:95-108 (as cited in Gentry et al.,
2004).
Hughes, MF and 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 53:95-112 (as cited inNRC, 1999).
Hutchinson, J (1887) Arsenic cancer. Br. Med. J. 2:1280-1281 (as cited in NRC,
1999).
Jackson, R and Grainge, JS (1975) Arsenic and cancer. Can. Med. Assoc. J.
113:396-401. (As cited in NRC, 1999).
Jacobson-Kram, D, and Montalbano, D (1985) The reproductive effects assessment
group's report on the mutagenicity of inorganic arsenic. Environ. Mutagen. 7:787-
804 (as cited in California EPA, 2003)
Jensen, GE and Hansen ML (1998) Occupational arsenic exposure and glycosylated
haemoglobin. Analyst 123(l):77.
Jha, An, Noditi, M, Nilsson, R, and Natarajan, AT (1992) Genotoxic effects of
sodium arsenite on human cells. Mutat. Res. 284(2):215-221 (as cited in ATSDR,
2002; Rossman, 2003).
Joint National Committee on Detection, Evaluation, and Treatment of High Blood
Pressure. (1993) The Fifth Report of the Joint National Committee on Detection,
Evaluation, and Treatment of High Blood Pressure (JNCV). Arch. Intern. Med.
153:154-183 (as cited in Rahman et al., 1999a).
Kagey, BY, Bumgarner, JE, and 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 1 llh Annual Conference on Trace
Substances in Environmental Health. Columbia, Mo: University of Missouri Press.
Pp. 252-256 , (as cited in NRC, 1999).
Karagas, MR, Stukel, TA, Moriss, JS, et al. (2001) Skin cancer risk in relation to
toenail arsenic concentrations in a US population-based case-control study. Am. J.
Epidemiol'. 153:559-565.
DRAFT - DO NOT CITE OR QUOTE 78
-------
Kashiwada, E, Kuroda, K, and Endo, G (1998) Aneuploidy induced by
dimethylarsinicacid in mouse bone marrow cells. Mutat. Res. 413:33-38.
Kawaguchi, K, OKU, N, Rin, K, et al. (1996) Dimethylarsenics reveal DNA
damage induced by superoxide anion radicals. Biol. Pharm. Bull. 19:551-553 (as
cited in Thomas, 2001).
Kayajanian, G (2003) Arsenic, cancer, and thoughtless policy. Ecotoxicology and
Environmental Safety 55:139-142.
Kenyon, EM, Del Razo, and Hughes, MF. (2005) Tissue distribution and urinary
excretion of inorganic arsenic and its methylated metabolites in mice following
acute oral administration of arsenate. Toxicol. Sci. 85(l):468-475.
Kenyon, EM and Hughes, MF (2001) A concise review of the toxicity and
carcinogenicity of dimethylarsinic acid. Toxicology 160(l-3):227-236 (as cited in
NRC,2001).
Kenyon, E, Hughes, M, and Levander, 0 (1997) Influence of dietary selenium on
the disposition of arsenate in female B6C3F1 mouse. J. Toxicol. Environ. Health
51:279-299 (as cited in Gentry et al., 2004).
Kessel, M, Liu, SX, Xu, A, et al. (2002) Arsenic induces oxidative DNA damage in
mammalian cells. Mol Cell Biochem 234-235:301-308 (as cited in Rossman, 2003).
Khan, AU, Kovacic, D, Kolbanovski, A (2000) The decomposition of peroxynitrite
to nitroxyl anion (NO-) and singlet oxygen in aqueous solution. Proc Natl Acad Sci
USA 97:2984-2989 (as cited in Huang et al., 2004).
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, Del Razo, LM, Brown, JL, et al. (1999) An integrated
pharmacokinetic and pharmacodynamic study of arsenate action. 1. Heme
oxygenase induction in rats. Teratog. Carcinog. Mutagen. 19(6):385-402 (as cited
inNRC,2001).
Kjeldsberg, CR and Ward, HP (1972) Leukemia in arsenic poisoning. Ann. Intern.
Med. 77:935-937 (as cited in NRC, 1999).
Klaunig, JE and Kamendulis, LM. (2004) The role of oxidative. stress in
carcinogenesis. Annu. Rev. Pharmacol. Toxicol. 44:239-267.
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. Environmental and Molecular
Mutagenesis 42:192-205
Kreppel, H, Liu, J, Liu, Y., et al. (1994) Zinc-induced arsenite tolerance in mice.
Fundam. Appl. Toxicol. 23:32-37 (as cited in ATSDR, 2000).
DRAFT - DO NOT CITE OR QUOTE 79
-------
Kuorkawa, M., Ogata, K., Idemori, M., et al. (2001) Investigation of skin
malformations of arsemicism due to intake of arsenic-contaminated groundwater in
residents of Samta, Jessore, Bangladesh. Arch. Dermatol. 137(1):102-103 (as cited
inNRC,2001).
Kurttio, P, Pukkala, E, Kahelin, II, et al. (1999). Arsenic concentration in well
water and risk of bladder and kidney cancer in Finland. Environ. Health Perspect.
107:705-710.
Lai, MS, Hsueh, YM, Chen, CJ, et al. (1994) Ingested inorganic arsenic and
prevalence of diabetes mellitus. Am. J. Epidemiol. 139:484-492 (as cited by
Tseungetal.,2000).
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. Journal of Occupational and Environmental Medicine
:1-16.
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. Biomedical and Environmental Sciences 16:355-368.
Lander, HM (1997) An essential role for free radicals and derived species in signal
transduction. Faseb J. 11:118-124 (as cited in Huang et al., 2004).
Lanz, H. Jr, Wallace, PC, and Hamilton, JG (1950) The metabolism of arsenic in
laboratoy
animals using As7" as a tracer. Univ. Calif. Publ. Pharmacol. 2:263-282 (as cited in
NRC, 1999).
Larramendy, ML, Popescu, NC, and 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 (as cited in U.S. EPA
,1998c).
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 (as cited in California. EPA, 2003).
Le, XC, Lu, X, Ma, M, et al. (2000b) Speciation of key arsenic metabolic
intermediates in human urine. Anal. Chem. 72:5172-5177 (as cited in California
EPA, 2003)
Lee, M-Y, Jung, B-I, Chung, S-M, et al. (2003) Arsenic-induced dysfunction in
relaxation of blood vessels. Environmental Health Perspectives 111 (4):513-517.
Lee, M-Y, Bae, 0-N, Chung, S-M, et al. (2002) Enhancement of platelet
aggregation and thromus formation by arsenic in drinking water: a contributing
factor to cardiovascular disease. Toxicology and Applied Pharmacology 179:83-88.
DRAFT - DO NOT CITE OR QUOTE 80
-------
Lee, T-C and Ho, I-C (1995) Modulation of cellular antioxidant defence activities
by sodium arsenite in human fibroblasts. Arch. Toxicol. 69:498-504 (as cited in
Rossman, 2003).
Lee, TC, 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 (as cited in Rossman,
2003).
Lee, T-C, Huang, RY, and Jan, KY (1985) Sodium arsenite enhances the
cytotoxicity clastogeriicity and 6-thioguanine-resistant mutagenicity of ultraviolet
light in Chinese hamster ovary cells. Mutat. Res. 148:83-89 (as cited in Rossman,
2003).
Lee-Chen, SF, Gurr, JR, Lin, IB, and 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.
Lerman, SA and Clarkson, TW (1983) The metabolism of arsenite and arsenate by
the rat. Chem.-Biol. Interact. 45:401-406 (as cited in NRC, 1999).
Lerman, SA, Clarkson, TW, and Gerson, RJ (1983) Arsenic uptake and metabolism
by liver cells is dependent on arsenic oxidation state. Chem.-Biol. Interact. 45:401-
406 (as cited in NRC, 1999).
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, 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-H, and Rossman, TG. (1991) Comutagenesis of sodium arsenite with
ultraviolet radiation in Chinese hamster V79 cells. Biol. Met. 4:197-200 (as cited
in Rossman etal. 2003).
Li, JJ, Oberley, LW, Fan, M, et al. (1998) Inhibition of AP-1 and NF-kappaB by
manganese-containing superoxide dismutase in human breast cancer cells. Faseb J.
12:1713-1723 (as cited in Huang et al., 2004).
Li, J-H and Rossman, TG (1989a) Mechanism of comutagenesis of sodium arsenite
with N-methyl-N-nitrosourea. Biol. Trace Element Res. 21:373-381 (as cited in
Rossman, 2003).
Li, JH and Rossman, TG (1989b) Inhibition of DNA ligase activity by arsenite: A
possible mechanism of its comutagenesis. Mol. Toxicol 2:1-9 (as cited in Huang et
al.,2004).
Liebscher, K. and H. Smith. (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. (As cited in NRC, 1999)
DRAFT - DO NOT CITE OR QUOTE 81
-------
Lin, S, Del Razo, LM, Styblo, M, et al. (2001) Arsenicals inhibit thioredoxin
reductase in cultured rat hepatocytes. Chem. Res. Toxicol. 14(3)305-311 (as cited
inNRC, 2001).
Lindgren, A, Danielsson, BRG, Dencker, L, and Vahter, M (1984) Embryotoxicity
of arsenite and arsenate: Distribution in pregnant mice and monkeys and effects on
embryonic cells in vitro. ActaPharmacol. Toxicol. 54:311-320 (as cited inNRC,
1999 and ATSDR, 2002).
Lindgren, A, M. Vahter, and L. Dencker. (1982) Autoradiographic studies on the
distribution of arsenic in mice and hamster administered 74As-arsenite or -arsenate.
Acta Pharmacol. Toxicol. 51:253-365 (as cited in NRC, 1999).
Liou, SH, Lung, JC, 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, Z, Shen, J, Carbrey, JM, et al. (2002) Arsenite transport by mammalian
aquaglyceroporins AQP7 and AQP9. Proc. Natl. Acad. Sci. 99:6053-6058 (as cited
inRossman, 2003).
Liu, SX, Athar, M, Lippai, I, et al. (2001) Induction of oxyradicals by arsenic:
Implication for mechanism of genotoxicity. Proc. Natl. Acad. Sci. USA 98:1643-
1648 (as cited in Huang et al., 2004).
Livingstone, LR, White, A, Spouse, J, et al. (1992) Altered cell cycle arrest and
gene amplification potential accompany loss of wild-type p53. Cell 70:923-935 (as
cited in Rossman, 2003).
Luster, MI, Wilmer, JL, Germolec, DR, et al. (1995) Role of keratinocyte-derived
cytokines in chemical toxicology. Toxicol. Lett. 82-83:471-476 (as cited in
Rossman, 2003).
Ma, HZ, Xia, YJ, Wu, KG, et al. (1999) Human exposure to arsenic and health
effects in Bayingormen, Inner Mongolia. In: Chappell, WR, Abernathy, CO, and
Calderon, RL, eds. Arsenic Exposure and Health Effects. Oxford: Elsevier. Pp.
127-131.,
Maier, A, Schumann, BL, Chang, X, et al. (2002) Arsenic co-exposure potentiates
benzo[a]pyrene genotoxicity. Mutat. Res. 517:101-111 (as cited in Huang et al.,
2004).
Maiorino, RM, and Aposhian, HV (1985) Dimercaptan metal-binding agents
influence the biotransformation of arsenite in the rabbit. Toxicol. Appl. Pharmacol.
77:240-250 (as cited in NRC, 1999).
Mann, S, Droz, PO, and Vahter, M (1996a) A physiologically-based
pharmacokinetic model for arsenic exposure: I. Development in hamster and
rabbits. Toxicol. Appl. Pharmacol. 137:8-22.
DRAFT - DO NOT CITE OR QUOTE 82
-------
Mann, S, Droz, PO, and Vahter, M (1996b) A physiologically based
pharmacokinetic model for arsenic exposure: II. Validation and application in
humans. Toxicol. Appl. Pharmacol. 140:471-486.
Mappes, R (1977) Experiments on the excretion of arsenic in urine. Int. Arch.
Occup. Environ. Health 40:267-272 (as cited in ATSDR, 2002).
Marafante, E, Vahter, M, Norin, H, et al. (1987). Biotransformation of
dimethylarsinic acid in mouse, hamster, and man. J. Appl. Toxicol. 7:111-117 (as
cited in ATSDR, 2000; NRC, 1999)
Marafante, E, Vahter, M, and Envall, J (1985) The role of methylation in the
detoxication of arsenate in the rabbit. Chem.-Biol. Interact. 56:225-238 (as cited i"
NRC, 1999).
Marafante, E, and Vahter M (1987) Solubility, retention and metabolism of
intratracheally and orally administered inorganic arsenic compounds in the
hamster. Environ Res 42:72-82 (as cited in ATSDR, 2000)
Marafante, E, and Vahter, M (1984) The effect of methyltransferase inhibition on
the metabolism of (74As)arsenite in mice and rabbits. Chem.-Biol. Interact. 50:49-
57 (as cited in NRC, 1999).
Marafante, E, Rade, J, Sabbioni, E, et al. (1981) Intracellular interaction and
metabolic fate of arsenite in the rabbit. Clin. Toxicol. 18:1335-1341 (as cited in
NRC, 1999).
Marcus, WL, and Rispin, AS (1988) Threshold carcinogenicity using arsenic as an
example. In: Cothern, CR, and Mehlman, MA, eds. Advances in Modern
Environmental Toxicology. Vol. XV. Risk Assessment and Risk Managment of
Industrial and Environmental Chemicals. Princeton, NJ: Princeton Scientific
Publishing Co., Inc. Pp. 133-158.
Mass, MJ, Tennant, A, Roop, BC, et al (2001) Methylated trivalent arsenic species
are genotoxic (abstract). Chem. Res. Toxicol. 14:355-361.
Mass, MJ and 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 (as cited in Rossman, 2003).
Mazumder, DNG, Haque, R, Ghosh, N, et al. (2000) Arsenic in drinking water and
the prevalence of respiratory effects in West Bengal, India. International Journal of
Epidemiology 29:1047-1052.
Mazumder, DNG, Haque, R, Ghosh, N, et al. (1998) Arsenic levels in drinking
water and the prevalence of skin lesions in West Bengal, India. International
Journal of Epidemiology. 27:871-877.
Mealey, J, Brownwll, GL, and Sweet, WH (1959) Radioarsenic in plasma, urine,
normal tissue, and intracranial neoplasms. Arch. Neurol. Psychiatr. 81:310-320 (as
cited in NRC, 1999).
DRAFT-DO NOT CITE OR QUOTE 83
-------
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.l llh Ed.
Rahway, NJ: Merck & Co., Inc. Pp.126-127, 8515.
Molin, L, and Wester, PO (1976) The estimated daily loss of trace elements from
normal skin by desquamation. Scand. J. Clin. Lab. Invest. 36:679-682 (as cited in
NRC, 1999).
Moore, MM, Harrington-Brock, K., and Doerr, CL (1997a) Relative genotoxic
potency of arsenic and its methylated metabolites. Mutat. Res. 386:279-290 (as
cited in NRC, 1999).
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 (as cited in
ATSDR, 2000).
Moore, MM, Harrington-Brock, K., and Doerr, CL (1995) Genotoxicity of arsenic
and its methylated metabolites. Mutagenesis and Cellular Toxicology Branch,
Genetic Toxicology Division, Health Effects Research Laboratory, U.S.
Environmental Protection Agency, Research Triangle Park, NC (as cited in U.S.
EPA, 1998c).
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 \np53 +/-
mice. Denver, CO; AWWA Research Foundation and the American Water Works
Association.
Mure, K, Uddin, A, Lopez, LC, et al. (2003) Arsenite induces delayed mutagenesis
and transformation in human osteosarcoma cells at extremely low concentrations.
Environ Mol Mutagen 41:322-331 (as cited in Rossman, 2003).
Nadeenko, VG, Lenchenko, V, Genkina, SB, et al. (1978) The influence of
tungsten, molybdenum, copper and arsenic on the intrauterine development of the
fetus. TR-79-0353. Farmakologiya i Toksikologiya. 41:620-623. (As cited in U.S.
EPA, 1998c).
Nagy, G, Nemeth, A, Bodor, F, and Fiscor, E (1980) Cases of bladder cancer
caused by chronic arsenic poisoning (in Hungarian). Orv. Hetil. 121:1009-1011 (as
cited in NRC, 1999).
NCI/EPA. (1983) U.S. Cancer Mortality Rates and Trends, 1950-1979; Volume II
(as cited in Lamm et al., 2004)
NRC (2001) Arsenic in drinking water. 2001 update. National Academy Press,
Washington, DC. http://www.nap.edu/openbook/0309076293/html/Rl.html.
DRAFT-DO NOT CITE OR QUOTE 84
-------
NRC( 1999) Arsenic in drinking water. National Academy Press, Washington, DC.
http://www.nap.edu/openbook/0309063337/htrnl/Rl.html.
Nemec, MD, Holson, JF, Farr, CH, and Hood, RD. (1998) Developmental toxicity
assessment of arsenic acid in mice and rabbits. Reprod Toxicol 12:647-658 (as
cited in Thomas etal, 2001).
Nemeti, B and Gregus, Z. (2005) Reduction of arsenate to arsenite by human
erythrocyte lysate and rat liver cytosol - Characterization of a Glutathionc- and
NAD-Depcndent Arsenate Reduction Linked to Glycolysis. Toxicol. Sci.
85(2):847-858.
Nemeti, B,Csanaky, I, and Gregus, Z. (2003) Arsenate reduction in human
erythrocytes and rats - Testing the role of purine nucleoside phosphorylase.
Toxicol. Sci. 74:22-31.
Nemeti, B and Gregus, Z. (2004) Glutathione-dependent reduction of arsenate in
human erythrocytes-a process independent of purine nucleoside phosphorylase.
Toxicol. Sci. 82(2):419-428.
Nesnow, S, Roop, BC, Lambert, G, et al. (2002) DNA damage induced by
methylated trivalent arsenicals is mediated by reactive oxygen species. Chem. Res.
Toxicol. 15:1627-1634.
Neubauer, 0 (1947) Arsenical cancer: a review. Br. J. Vancer 1:192-251 (as cited
inNRC, 2001).
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, and Gothern,
CR, eds. Arsenic: Exposure and Health Northwood, U.K.: Science and Technology
Letters. Pp. 217-223 (as cited in Gentry et al., 2004).
Noda, Y, Suzuki, T, Kohara, A, et al. (2002) In vivo genotoxicity evaluation of
dimethylarsinic acid in Muta™ Mouse. Mutat. Res. 513:205-212.
Nordenson, 1 and Beckman, L. (1991) Is the genotoxic effect of arsenic mediated by
oxygen free radicals? Hum. Hered. 41:71-73 (as cited in Rossman, 2003).
Nordenson, I, Salmonsson, S, Brun, E, et al. (1979) Chromosome aberrations in
psoriatic patients treated with arsenic. Hum. Genet. 48:1-6.-
Nordenson, 1, Beckman, G, Beckman, L, and Nordstrom, S (1978) Occupational
and environmental risk in and around a smelter in Northern Sweden. II.
Chromosomal abberations in workers exposed to arsenic. Hereditas. 88:263-267 (as
cited in Kitchin, 2001).
O'Brien, TG, Megosh, LC, Gillard, G, and Soler, AP (1997) Ornithine
decarboxylase over expression is a sufficient condition for tumor promotion in
mouse skin. Cancer Res. 57:2630-2637 (as cited in Rossman, 2003).
DRAFT - DO NOT CITE OR QUOTE 85
-------
Odanaka, Y, Matano, 0, and Goto, S (1980) Biomethylation of inorganic arsenic by
the rat and some laboratory animals. Bull. Environ. Contamin. Toxicol. 24:452-459
(as cited in NRC, 1999)
Ohno, H, Hanaoka, F, and Yamada, M (1982) Inducibility of sister chromatid
exchanges by heavy-metal ions. Mutat. Res. 104(1-3): 141-146 (as cited in U.S.
EPA1998c).
Omary, MB, Ku, NO, and Toivola, DM (2002) Keratins: guardians of the liver.
Hepatology 35:251-257 (as cited in Waalkes et al., 2004b).
Oya-Ohta, Y., T. Kaise, T. Ochi (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 (as cited
in Rossman, 2003).
Perkins, C, Kim, CN, Fang, G, and 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 (as cited in Huang et al.,
2004).
Petres, J, Baron, D, and Hagedorn, M (1977) Effects of arsenic cell metabolism.and
cell proliferation: Cytogenic and biochemical studies. Environ. Health Perspect.
19:223-227 (as cited in Kitchin, 2001).
Petrick, JS, Jagadish, B, Mash, EA, and Aposhian, HV (2001).
Monomethylarsonous.acid(MMAm) and arsenite: LD50 in hamsters and in vitro
inhibition of pyruvate dehydrogenase. Chem. Res. Toxicol. Web Release date May
31, 2001 (as cited in Thomas et al., 2001; NRC, 1999).
Petrick, JS, Ayala-Fierro, F, Cullen, WR, et al. (2000) Monomethylarsonous acid
(MMA(III)) is more toxic than arsenite in Chang human hepatocytes. Toxicol.
Appl. Pharmacol. 163:203-207 (as cited in Thomas et al., 2001; NRC, 1999).
Pi, J, Yamauchi, H, Kumagai, Y, et al. (2002) Evidence for induction of oxidative
stress caused by chronic exposure of Chinese residents to arsenic contained
drinking water. Environ Health Perspect 110:331-336 (as cited in Rossman, 2003).
Pomroy, C., Charbonneau, SM, McCullough, RS, and Tarn, GKH (1980) Human
retention studies with 74As. Toxicol. Appl. Pharmacol. 53:550-556 (as cited in
NRC, 1999).
Popper, H, Thomas, LB, Telles, NC, et al. (1978) Development of hepatic
angiosarcoma in man induced by vinyl chloride, thorotrast, and arsenic. Am. J.
Pathol. 92:349-369 (as cited in NRC, 1999).
Prukop, JA, and Savage, NL (1986) Some effects of multiple, sublethal doses of
monosodium methanearsonate (MSMA) herbicide on hematology, growth, and
reproduction of laboratory mice. Bull. Environ. Contamin. Toxicol. 36:337-341 (as
cited in ATSDR, 2000).
DRAFT - DO NOT CITE OR QUOTE 86
-------
Prystowsky, SD, Elferibein, GJ, and Lamberg, SI (1978) Nasopharyngeal carcinoma
associated with long-term arsenic ingestion. Arch. Dermatol. 114:602-603 (as cited
inNRC,1999).
Radabaugh, TR, Sampayo-Reyes, A, Zakharyan, RA, and 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. Re.s. Toxicol. 15(5):692-698.
Radabaugh, TR, and 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 (as cited in NRC, 2001).
Rahman, M, Tondel, M, Ahmad, SA, et al. (1999a) Hypertension and arsenic
exposure in Bangladesh. Hypertension 33:74-78.
Rahman, M, Tondel, M, Chowdhury, IA, and Axelson, 0. (1999b) Relations
between exposure to arsenic, skin lesions, and glucosuria. Occup. Environ. Med.
56:277-281.
Rahman, M, Tondel, M, Ahmad, SA, and Axelson, 0 (1998) Diabetes mellitus
associated with arsenic exposure in Bangladesh. American Journal of Epidemiology
148(2):198-203.
Rahman, M, Wingren, G., and Axelson, 0. (1996) Diabetes mellitus among
Swedish art glass workers - an effect of arsenic exposure? Scand J Work Environ
Health 22:146-149 (as cited in Rahman et al., 1999b).
Rahman, M, and Axelson, O. (1995) Diabetes mellitus and arsenic exposure: a
second look at a case-control data from Swedish copper smelter. Occup Environ
Med 52:773-774 (as cited in Rahman et al., 1999b).
Rahman, MS, Hall, LL, and Hughes, MF (1994) In vitro percutaneous absorption of
sodium arsenate in B6C3F1 mice. Toxicol. In Vitro 8:441-448 (as cited in NRC,
1999).
Rasmussen, R.E. and D.B. Menzel. (1997) Variation in arsenic-induced sister
chromatid exchange in human lymphocytes and lymphoblastoid cell lines. Mutat.
Res. 386:299-306 (as cited in NRC, 1999).
Reymann, F, Moller, R, and Nielsen, A (1978) Relationship between arsenic intake
and internal malignant neoplasms. Arch. Dermatol. 114:378-381 (as cited in NRC,
1999).
Rin, K, Kawaguchi, K., Yamanaka, K, et al. (1995) DNA-strand breaks induced by
dimethylarsinic acid, a metabolite of inorganic arsenic, are strongly enhanced by
superoxide anion radicals. Biol. Pham. Bull. 18:45-48 (as cited in Thomas et al.,
2001).
DRAFT - DO NOT CITE OR QUOTE 87
-------
Roat, JW, Wald, A, Mendelow, H., and Pataki, KT (1982) Hepatic angiosarcoma
associated with short-term arsenic ingestion. Am. J. Med. 73:933-936 (as cited in
NRC, 1999).
Robertson, DAF and Low-Beer, TS (1983) Long term consequences of arsenical
treatment for multiple sclerosis, Br. Med. J. 286:605-606 (as cited in NRC, 1999).
Robson, AO and Jelliffe, AM (1963) Medicinal arsenic poisoning and lung cancer.
Br. Med. J. 5351:207-209 (as cited in NRC, 1999).
Rogers, EH, Chernoff, N, and Kavlock, RJ. (1981) The teratogenic potential of
cacodylic acid in the rat and mouse. Drug Chem. Toxicol. 4:49-61 (as cited in
Thomas etal., 2001).
Rose, GA, Blackburn, H, Gillum, RF, and Prineaas, RJ. (1982) In: Cardiovascular
Survey Methods. 2nd ed. Geneva, Switzerland: World Health Organization. Pp. 82-
85 (as cited in Rahman et al., 1999a)
Rossman, TG. (2003) Review: Mechanism of arsenic carcinogenesis: an integrated
approach. Mutat Res 533:37-65.
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 AppI Pharmacol 176:64-71 (as cited in Rossman, 2003).
Rossman, TG. (1998) Molecular and genetic toxicology of arsenic. In: Rose, J, ed).
Environmental Toxicology: Current Developments. Amsterdam: Gorden and
Breach Science Publishers. Pp. 171-187 (as cited in Huang et al., 2004).
Rossman, TG and Wolosin, D. (1992) Differentialsusceptibility to carcinogen-
induced amplification of SV40 and dhfr sequences in SV40-transformed human
keratinocytes. Mol. Carcinog. 6:203-213 (as cited in Rossman, 2003).
Rossman, TG, Stone, D, Molina, M, and Troll, W (1981) Absence of arsenite
mutagenicity in E. coli and Chinese hamster cells. Environ. Mutagen. 2:371-379
(as cited in Huang et al., 2004; Rossman, 2003).
Rossman, T.G., D. Stone, M. Molina, W. Troll. (1980) Absence of arsenite
mutagenicity in E. coli and Chinese hamster cells. Environ. Mutagen. 2:371-379 (as
cited in Huang et al., 2004).
Rossman, T, Meyn, MS, and Troll, W. (1975) Effects of sodium arsenite on the
survival of UV-irradiated Escherichia coli: Inhibition of a recA-dependent
function. Mutat Res 30:157-162 (as cited in Huang et al., 2004).
Roth, F (1957) The sequelae of chronic arsenic poisoning in Moselle vinters. Germ.
Med. Monthly 2:172-175 (as cited in ATSDR).
Rudnai. P, Csanady, M, Saekany, E, et al. (1996) Association between arsenic
levels of drinking water in reproductive outcomes in Hungary (abstract).
DRAFT - DO NOT CITE OR QUOTE 88
-------
International seminar on arsenic exposure: health effects, remediation methods in
treatment costs. Uni versidad de Chile, Oct. 8-10 (as cited in NRC, 1999).
Saady, JJ, Blanke, RV, and Poklis, A (1989) Estimation of the body burden of
arsenic in a child fatally poisoned by arsenite weedkiller. J Anal Toxicol 13:310-
312 (as cited in ATSDR, 2000).
Sakurai, T, Kaise, T, and Matsubara, C (1998) Inorganic and methylated arsenic
compounds induce cell death in murine macrophages via different mechanisms.
Chem. Res. Toxicol. 11:273-283 (as cited in NRC, 1999).
Salazar, AM, Ostrosky, WP, Menendez, D, et al. (1997) Induction of p53 protein
expression by sodium arsenite. Mutat. Res. 381:259-265 (as cited in Rossman,
2003).
1 Sampayo-Reyes, A., R.A. Zakharyan, S.M. Healy and H.V. Aposhian. (2000)
Monomethylarsonic acid reductase and monomethylarsonic acid in hamster tissue.
Chem. Res. Toxicol. 13(11):1181-1186 (as cited in NRC, 2001).
Scharf, JG, Dombrowski, F, and Ramadori, G. (2001) The IGF axis and
hepatocarcinogenesis. Mol Pathol. 54:138-144 (as cited in Waalkes et al., 2004b).
Schroeder, HA, and Mitchner, M (1971) Toxic effects of trace elements on the
reproduction of mice and rats. Arch. Environ. Health. 23(2):102-106 (as cited in
ATSDR, 2000; NRC, 1999).
Sciandrello, G, Barbaro, R, Caradonna, F, et al. (2002) Early induction of genetic
instability and apoptosis by arsenic in cultured Chinese hamster cells. Mutagenesis
17:99-103 (as cited in Rossman, 2003).
Scott, N, Hatlelid, KM, MacKenzie, NE, and Carter, DE (1993) Reactions of
arsenic(III) and arsenic(V) species with glutathione. Chem. Res. Toxicol. 6:102-
106 (as cited in NRC, 1999).
Scott, A (1958) The retention of arsenic in the late cutaneous complications of its
administration. Br. J. Dermatol. 70:195-200.
Sen, CK and Parker, L. (1996) Antioxidant and redox regulation of gene
transcription. Faseb J. 10:709-720 (as cited in Huang et al., 2004).
Shackelford, RE, Kaufmann, WK, and Paules, RS. (2000) Oxidative stress and cell
cycle checkpoint function. Free Radic. Biol. Med. 28:1387-1404 (as cited in Huang
et al., 2004).
Schoof, RA, Yost, LJ, Crecelius, C, Irgolic, K, Goessler, W, Guo, H-R, and
Greene, H. (1998) Dietary arsenic intake in Taiwanese districts with elevated
arsenic in drinking water. Hum. Ecol. Risk Assess. 4:117-135.
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. Carcinogenesis 24(11): 1827-1 835.
DRAFT - DO NOT CITE OR QUOTE 89
-------
Shiobara, Y, Ogra, Y, and Suzuki, KT. (2001) Animal species difference in the
uptake of dimethylarsinous acid (DMA(III)) by red blood cells. Chem. Res.
Toxicol. 14(10):1446-1452.
Simeonova, PP and Luster MI. (2000) Mechanisms of arsenic carcinogenicity:
Genetic or epigenetic mechanisms? Environ Pathol Toxicol Oncol 19:281 -286.
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. American Journal of Epidemiology. 147(7):660-669.
Snow, ET. (1992) Metal•carcinogenesis: Mechanistic implications. Pharmacol.
Ther. 53:31-65 (as cited in Huang etal., 2004).
Sommers, SC and McManus, RG (1953) Multiple arsenical cancers of skin and
internal organs. Cancer 6:347-359 (as cited in NRC, 1999).
Steinmaus, C, Yuan, Y, Bates, MN, and Smith AH (2003) Case-control study of
bladder cancer and drinking water arsenic in the western United States. American
Journal of Epidemiology 158(12):! 193-1201.
Stevens, JT, Hall, LL, Farmer, JD, et al. (1977) Disposition of 14C and/or 7"As-
cacodylic acid in rats after intravenous, intratracheal or peroral administration.
Environ. Health Perspect. 19:151-157.
Styblo, M., L.M. Del Razo, L. Vega, D.R. Germolec, E.L. LeCluyse, G.A.
Hamilton, W. Reed, C. Wang, W.R. Cullen and D.J. Thomas. (2000) Comparative
toxicily of trivalent and pentavalent inorganic and methylated arsenicals in rat and
human cells. Arch. Toxicol. 74(6):289-299 (as cited in NRC, 1999; Thomas et al.,
2001; Huang etal.,2004).
Styblo, M, Del Razo, LM, LeCluyse, EL, et al. (1999a) Metabolism of arsenic in
primary cultures of human and rat hepatocytes. Chem. Res. Toxicol. 12(7):560-565
(as cited in NRC, 2001).
Styblo, M, Vega, L, Germolec, DR, et al. (1999b) Metabolism and toxicity of
arsenicals in cultured cells. In: Chappell, WR, Abernathy, CO, Calderon, RL, eds.
Arsenic Exposure and Health Effects,. Oxford: Elsevier Pp. 311-323. (As cited in
NRC, 1999; NRC, 2001; Thomas etal.,2001).
Styblo, M, Delnomdedieu, M, and Thomas, DJ (1996) Mono- and dimethylation of
arsenic in rat liver cytosol in vitro. Chem.-Biol. Interact. 99:147-164 (as cited in
NRC, 1999).
Styblo, M, Yamauchi, H, and Thomas, DJ (1995) Comparative in vitro methylation
of trivalent and pentavalent arsenicals. Toxicol. Appl. Pharmacol. 135:172-178 (as
cited in NRC, 1999)
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 (as cited in NRC, 1999).
DRAFT - DO NOT CITE OR QUOTE 90
-------
Tarn, GKH, Charbonneau, SM, Bryce, F, et al. (1979) Metabolism ofinorganic
arsenic (74As) in humans following oral ingestion. Toxicol. Appl. Pharmacol.
50:319-322 (as cited in NRC, 1999).
Tezuka, M, Hanioka, K, Yamanaka, K, and 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 (as cited in Thomas, 2001).
Thomas, DJ, Waters, SB, and Styblo, M. (2004) Elucidating the pathway for
arsenic methylation. Toxicol. Appl. Pharmacol. 198(3):319-326.
Thomas, D.J., M. Styblo and S. Lin. (2001) The cellular metabolism and systemic
toxicity of arsenic. Toxicol. Appl. Pharmacol. 176:127-144.
Tinwell, H, Stephens, SC, and Ashby, J. (1991) Arsenite as the probable active
species in the human carcinogenicity of arsenic: mouse micronucleus assays on Na
and K arsenite, orpiment, and Fowler's solution. Environ. Health Perspect. 95:205-
210 (as cited in U.S. EPA 1998c; Rossman, 2003).
Tondel, M, Rahmann, M, Magnuson, A, et al. (1999) The relationship of arsenic
levels in drinking water and the prevalence rate of skin lesions in Bangladesh.
Environ. Health Perspect. 107(9):727-729 (as cited in NRC, 2001).
Tsai, S-M, Wang, T-N, Ko, Y-C (1999) Mortality for certain diseases in areas with
high levels of arsenic. Archives of Environmental Health 54(3):186-193.
Tseng, C-H, Tai, T-Y, Chong, C-K, et al. (2000) Long-term arsenic exposure and
incidence of non-insulin-dependent diabetes mellitus: A cohort study in arseniasis-
hyperendemic villages in Taiwan. Environ. Health Perspect. 108(9):847-851.
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.
Tsuchiya, K (1977) Various effects of arsenic in Japan depending on type of
exposure. EnvironmentalHealth Perspectives 19:35-42.
Tsuda, T, Babazono, A, Yamamoto, E, ct al. (1995) Ingested arsenic and internal
cancer: a historical cohort study followed for 33 years. Am. J. Epidemiol. 141:198-
209 (as cited in California EPA, 2003).
Tucker, SB, Lamm, SH, Li, FX, et al. (2001) Relationship between consumption of
arsenic-contaminated well water and skin disorders. In: Huhhot, Inner Mongolia.
Inner Mongolia Cooperative Arsenic Project (IMCAP) Study. Final Report.
Department of Dermatology, University of Texas, Houston, and Department of
Environmental Epidemiology, Institute of Environmental Engineering, Chinese
Academy of Preventive Medicine, Beijing, China. July 5, 2001 (as cited in NRC,
2001).
DRAFT - DO NOT CITE OR QUOTE 91
-------
U.S. EPA (2005) Guidelines for Carcinogen Risk Assessment. Office of Research
and Development, Risk Assessment Forum, Washington, DC. EPA/630/P-03/001F.
U.S. EPA (2004a) ORD Response for Critical Reviews of Manuscripts Cited in
AX-030820. U.S. EPA Office of Research and Development, National Health and
Environmental Effects Research Laboratory, Human Studies Division, RTP, NC.
April, 2004.
U.S. EPA (2004b) Arsenic Drinking Water Health Advisory. Office of Water, U.S.
Environmental Protection Agency. Office of Water, Health and Ecological Criteria
Division, Washington, DC. July 2004.
U.S. EPA (2001) National Primary Drinking Water Regulations; Arsenic and
Clarifications to Compliance and New Source Contaminants Monitoring. 66 FR
14:6975-7022. January 2001.
U.S. EPA. (2000a) Science policy council handbook: peer review. Second edition.
Prepared by the Office of Science Policy, Office of Research and Development,
Washington, DC. EPA 1OO-B-00-001.
U.S. EPA. (2000b) Science policy council handbook: Risk characterization.
Prepared by the Office of Science Policy, Office of Research and Development,
Washington, DC. EPA 100-B-00-002.
U.S. EPA. (2000c) Benchmark Dose Technical Support Document. External
Review Draft, EPA/63 O/R-00/001, October. Office of Research and Development,
Risk Assessment Forum, Washington, DC.
U.S. EPA. (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. (1998a) Guidelines for neurotoxicity risk assessment. Federal Register
63(93):26926-26954.
U.S. EPA. (1998b) Science policy council handbook: Peer review. Prepared by the
Office of Science Policy, Office of Research and Development, Washington, DC.
EPA 100-B-98-001.
U.S. EPA (1998c) Integrated Risk Information System for Inorganic Arsenic. U.S.
Environmental Protection Agency. Carcinogenesis Assessment last revised 1998.
bttp:/Avww. epa.gov/iris/subst/0278. htm.
U.S. EPA. (1996) Guidelines for reproductive toxicity risk assessment. Federal
. Register 61(212):56274-56322.
U.S. EPA. (1995) Use of the benchmark dose approach in health risk assessment.
EPA/63 O/R-94/007.
U.S. EPA. (1994a) Interim policy for particle size and limit concentration issues in
inhalation toxicity: notice of availability. Federal Register 59(206):53799.
DRAFT - DO NOT CITE OR QUOTE 92
-------
U.S. EPA. (1994b) Methods for derivation of inhalation reference concentrations
and application of inhalation dosimetry. EPA/600/8-90/066F.
U.S. EPA. (1991) Guidelines for developmental toxicity risk assessment. Federal
Register 56(234):63798-63826.
U.S. EPA. (1988) Recommendations for and documentation of biological values for
use in risk assessment. EPA 600/6-87/008, NTIS PB88-179874/AS, February 1988.
U.S. EPA. (1986a) Guidelines for the health risk assessment of chemical mixtures.
Federal Register 51(185):34014-34025.
U.S. EPA. (1986b) Guidelines for mutagenicity risk assessment. Federal Register
51(185):34006-34012.
Vahter, M (1999a) Variation in human metabolism of arsenic. In: Chappell, WR,
Abernathy, CO, Calderon, RL, eds. Arsenic Exposure and Health Effects. Oxford:
Elsevier. Pp. 267-279 , (as cited in NRC, 2001).
Vahter, M (1999b) Methylation of inorganic arsenic in different mammalian
species and population groups. Sci. Prog. 82(Pt. l):69-88 (as cited in NRC, 2001)
Vahter, M, Concha, G, Nermell, B, et al. (1995a) A unique metabolism of inorganic
arsenic in native Andean women. Eur. J. Pharmacol. 293:455-462 (as cited in NRC,
1999)
Vahter, M, Couch, R, Nermell, B, and Nilsson, R (1995b) Lack of methylation of
inorganic arsenic in the chimpanzee. Toxicol. Appl. Pharmacol. 133:262-268 (as
cited in NRC, 1999).
Vahter, M (1994) Species differences in the metabolism of arsenic compounds.
Appl. Organomet. Chem. 8:175-182.
Vahter, M, and Marafante, E(1987) Effects of low dietary intake of methionine,
choline or proteins on the biotransformation of arsenite in the rabbit. Toxicol. Lett.
37:41-46 (as cited in NRC, 1999).
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 (as cited in Mann et al., 1996b)
Vahter, M, and Marafante, E (1985) Reduction and binding of arsenate in marmoset
monkeys. Arch. Toxicol. 57:119-124 (as cited in NRC, 1999).
Vahter, M, Marafante, E, and Dencker, L (1984) Tissue distribution and retention
of 74As-dimethylarsinic acid in mice and rats. Arch. Environ. Contam. Toxicol.
13:259-264 (as cited in NRC, 1999)
Vahter, M (1983) Metabolism of inorganic arsenic in relation to chemical form and
animal species. Doctoral Thesis. Karolinska Institute!, Stockholm (as cited in NRC,
1999)
DRAFT - DO NOT CITE OR QUOTE 93
-------
Vahter, M, and Envall, J (1983) In vivo reduction of arsenate in mice and rabbits.
Environ. Res. 32:14-24 (as cited in NRC, 1999).
Vahter, M and Marafante, E (1983) Intracellular interaction and metabolic fate of
arsenite and arsenate in mice and rabbits. Chem.-Biol. Interact. 47:29-44 (as cited
in NRC, 1999).
Vahter, M, Marafante, E, Lindgren, A, and Dencker, L (1982) Tissue distribution
and subcellular binding of arsenic n marmoset monkeys after injection of 74As-
arscnite. Arch. Toxicol. 51:65-77 (as cited in NRC, 1999).
Vahter, M (1981) Biotransformation of trivalent and pentayalent inorganic arsenic
in mice and rats. Environ. Res. 25:286-293 (as cited in NRC, 1999).
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 (as cited
in ATSDR, 2000).
Vellar, OD (1969) Nutrient losses through sweating. Thesis. Oslo University, Oslo,
Norway (as cited in NRC, 1999).
Vig, BK, Figueroa, ML, Cornforth, MN, and Jenkins, SH (1984) Chromosome
studies in human subjects chronically exposed to arsenic in drinking water. Am. J.
Ind. Med. 6(5):325-338 (as cited in ATSDR, 2000).
Waalkes, MP, Ward, JM, and 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, and Diwan, BA (2004b) Animal models for
carcinogenesis: inorganic arsenic is a transplacental carcinogen in mice.
Toxicology and Applied Pharmacology 198:377-384.
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. Toxicology and Applied Pharmacology
186:7-17.
Wan, B, Christian, RT, and Soukup, SW (1982) Studies of cytogenetic effects of
sodium arsenicals on mammalian cells in vitro. Environ. Mutag. 4:493-498 (as
cited in U.S. EPA 1998c)
Wang, S-L, Chiou, J-M, Chen, C-J, et al. (2003) Prevalence of non-insulin-
dependent diabetes mellitus and related vascular diseases in Southwestern
arseniasis-endemic and nonendemic areas in Taiwan. Environmental Health
Perspectives 111(2):155-159.
Wang, C-H, Jeng, J-S, Yip, P-K, et al. (2002) Biological gradient between long-
term arsenic exposure and carotid atherosclerosis. Circulation 105:1804-1809.
DRAFT - DO NOT CITE OR QUOTE 94
-------
Wang, TS, Shu, YF, Liu, YC, et al. (1997) Glutathione peroxidase and catalase
modulate the genotoxicity of arsenite. Toxicology 121 (3):229-237.
Wang, TS, Kuo, CF, Jan, KY, and Huang, H. (1996) Arsenite induces apoptosis in
Chinese hamster ovary cells by generation of reactive oxygen species. J. Cell
Physiol. 169:256-268 (as cited in Huang et al., 2004).
Wang, TS and Huang, H. (1994) Active oxygen species are involved in the
induction of micronuclei by arsenite in XRS-5 cells. Mutagencsis 9:253-257 (as
cited in Rossman, 2003).
Wanstall, JC, Jeffrey, TK, Gambino, A, ct al. (2001) Vascular smooth muscle
relaxation mediated by nitric oxide donors: a comparison with acetylcholine, nitric
oxide and nitroxy ion. Br J Pharmacol 134(3):463-472.
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 Epid. Biomarkers Prevent. 3:583-590 (as cited in Kitchin,
2001; Rossman, 2003).
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, and Carter, DE (1984) In vitro solubility and in vivo toxicity
of gallium arsenide. Toxicol. Appl. Pharmacol. 76:96-104 (as cited in ATSDR,
2000).
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, Wanibuchi, H, Morimura, K, et al. (1999) Urinary bladder carcinogenicity
of dimethylarsinic acid in male F344 rats. Carcinogenesis. 20(9):1873-1876.
Wester, R.C., Maibach, HI, Sedik, L, Melendres, J, and Wade M. (1993) In vitro
and in vivo percutaneous absorption and skin decontamination of arsenic from
water and soil. Fundam. Appl. Toxicol. 20:336-340 (as cited in NRC, 1999).
Wiencke, JK, and Yager, JW. (1992) Specificity of arsenite in potentiating
cytogenic damage induced by the DNA crbsslinking agent diepoxybutane. Environ
Mol Mutagen 19:195-200 (as cited in Rossman, 2003).
Wu, MM, Kuo, TL, Hwang, YH, and 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 .
DRAFT - DO NOT CITE OR QUOTE 95
-------
Xu, Y, Krishnan, A, Wan, XS, et al. (1999) Mutations in the promoter reveal a
cause for the reduced expression of the human manganese superoxide dismutase
gene in cancer cells. Oncogene 18:93-102 (as cited in Huang el al., 2004).
Yamanaka, K, Hayashi, H, Tachikawa, M, et al. (1997) Metabolic methylation is a
possible genotoxicity-enhancing process of inorganic arsenics. Mutat. Res. 394(1-
3):95-101.
Yamauchi, H, Yamato, N, and Yamamura, Y (1988) Metabolism and excretion of
orally and intraperitoneally administered methylarsonic acid in the hamster. Bull.
Environ. Contam. Toxicol. 40:280-286 (as cited in AT SDR, 2000).
Yamauchi, H, Takahashi, K, and Yamamura, Y (1986) Metabolism and excretion of
orally and intraperitoneally administered gallium arsenide in the hamster.
Toxicology 40:237-246 (as cited in ATSDR, 2000)
Yamauchi, H, and Yamamura, Y (1984) Metabolism and excretion of orally
administered dimethylarsinic acid in the hamster. Toxicol. Appl. Pharmacol.
74:134-140.
Yamauchi, H, and 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 (as cited in NRC, 1999).
Yang, C-Y, Chang, C-C, Tsai, S-S, et al. (2003) Arsenic in drinking water and
adverse pregnancy outcome in an arseniasis-endemic area in northeastern Taiwan.
Environmental Research 91:29-34.
Yang, L, Peterson, PJ, Williams, WP, et al. (2002a) The relationship between
exposure to arsenic concentrations in drinking water and the development of skin
lesions in farmers from inner Mongolia, China. Environmental Geochemistry and
Health 24:293-303.
Yang, L, Wang, W, Hou, S, et al. (2002b) Arsenism clinical stages and their
relation with hair arsenic concentration of residents of Bayinmaodao rural district,
inner Mongolia, China. Environmental Geochemistry and Health 24:337-348.
Yang, JL, Chen, MF, Wu, CW, and Lee, TC. (1992) Posttreatmerit with sodium
arsenite alters the mutation spectrum induced by ultraviolet light irradiation in
Chinese hamster ovary cells. Environ Mol Mutagen 20:156-164 (as cited in
Rossman, 2003).
Yates, CR, Krynetski, EY, Loennechen, T, et al. (1997) Molecular diagnosis of
thiopurinc S-methyltransferase deficiency: Genetic basis for azathioprinc and
mercaptopurine intolerance. Ann. Intern. Med. 126:608-614.
Yih, J-H, and 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 (as cited in Rossman, 2003).
DRAFT - DO NOT CITE OR QUOTE 96
-------
APPENDIX A. Summary of External Peer Review and Public Comments and
Disposition
To Be Added
DRAFT - DO NOT CITE OR QUOTE 99
-------
APPENDIX B:Quantitative Issues in the Cancer Risk Assessment for
Inorganic Arsenic
1. Description of the data set used for modeling
The data set chosen for analysis is the southwest Taiwanese study by Chen et
al. (1988), which had previously been analyzed by Morales et al. (2000) and by
NRC (2001). Briefly, mortality due to lung and bladder cancer was recorded in a set
of Taiwanese villages with measured arsenic concentrations in their well water.
This information was stratified by village, age, and sex. A set of southwest
Taiwanese villages without measurable well water arsenic was used as a
background population.
2. Choice of model
The model selected for this analysis is an additive (linear) Poisson model
with a linear term for dose and quadratic adjustment for age. The basic formula for
a Poisson model is:
h(x,t).= h(f>* * g(x)
where:
h(x,t} = Cancer hazard function
h(tY ~ a function describing the baseline risk in the absence of
exposure to the chemical in question. Although the baseline function can
incorporate a number of factors (e.g., age, gender, smoking/non-smoking, etc.), the
Taiwanese data used by Morales et al. only contained gender and age (t)
information.
g(x) = a function describing the effect of exposure to the chemical.
The term x is the chemical dose.
The use of a Poisson model provides great flexibility regarding the inclusion
of covariates (such as age) without the introduction of restrictive assumptions
about the data. Previous analyses had used either additive (linear) or multiplicative
(exponential) Poisson models; linear, logarithmic, or square-root terms for dose;
and linear, quadratic, or spline adjustments for age (Table B-l) (Morales et al.,
2000).
Table B-l. Poisson modeling options. (Modified from Morales et al., 2000)
Age effect h(f)* Dose Dose effect g(x)
transformation
Linear
exp(a0 + a,/)
Quadratic
exp(«0 + a,f +
Linear
x = ppb°
Logarithmic
x = log(l + ppb)
Linear
l+'P,*
Quadratic
te + tS
DRAFT - DO NOT CITE OR QUOTE 1 00
-------
Regression Square root Exponential
spline linear
exp[cc0 + a,«s(/)]A x = /ppb cxp(p,x)
Exponential
quadratic
"Represents exposure concentration in parts per billion, which is equivalent to micrograms per liter.
*flj(0 represents a natural spline applied to t.
' The previous EPA draft assessment for arsenic (USEPA, 2001) used a
multiplicative (exponential) model with a linear term for dose and a quadratic age
effect. The multiplicative model was chosen based upon lack of necessity for an
external control population, which EPA believed at that time would skew the
results in a non-linear fashion. The linear term for dose and quadratic age effect
were selected as representing the best fit to data by Akaike information criterion
(AIC). NRC (2001) recommended the use of an additive (linear) model with a
southwest Taiwanese background population. The rationale for this choice was that
no objective criteria could be found for excluding the use of the additive model
with background, and the additive model was a better fit to the data by AIC. The
current analysis accepts the NRC recommendation.
3. Low-dose extrapolation method
According to the Guidelines for Carcinogen Risk Assessment (USEPA,
2005):
Linear extrapolation should be used in two distinct
circumstances: (1) when there are data to indicate that the'dose-
response curve has a linear component below the POD, or (2) as a
default for a tumor site where the mode of action is not established
(see Section 3.3.1). For linear extrapolation, a line should be drawn
from the POD to the origin, corrected for background. This implies
a proportional (linear) relationship between risk and dose at low
doses. (Note that the dose-response curve generally is not linear at
higher doses.)
The slope of this line, known as the slope factor, is an
upper-bound estimate of risk per increment of dose that can be used
to estimate risk probabilities for different exposure levels. The
slope factor is equal to 0.01/LEDQ| if the LEDQ| is used as the POD.
As the mode of action for the carcinogenicity of inorganic arsenic is not
established, a linear low-dose extrapolation from the LED01 is used for this
analysis.
4. Selection of background population
For this analysis, a background population from southwest Taiwanese
villages without detected arsenic in their drinking water supply was used. The
DRAFT - DO NOT CITE OR QUOTE 101
-------
previous EPA draft assessment for arsenic (USEPA, 2001) did not use any
comparison population. This was due to concerns regarding the tendency of large
background populations to skew dose-response curves in a non-linear fashion. NRC
(2001) recommended the use of a southwest Taiwanese background population. The
rationale for this choice was that no objective criteria could be found for excluding
the use of this background population, that the presence of a background population
was not a unique determinant of the linearity or non-linearity of the dose-response,
and that the absence of a background population could easily lead to an
underestimate of the slope of the dose-response. The current analysis accepts the
NRC recommendation.
5. Adjustment for background cancer risks
The hazard (or relative risk) of cancer due to arsenic exposure has been
estimated based on Taiwanese data, including the Taiwanese background rate. The
results are then used to calculate lifetime risk for the U.S. population, using the
BEIRIV formula (NRC, 2001) and the U.S. background cancer rates. The BIER-IV
formula, corrected for two errors in NRC (2001), is
R(x) = lifetime risk of death at dose x
/2,(*)0/
7T7TS,l-<7,) CD
where:
/ = age group
h,{x) = cancer mortality hazard, age «', dose x
ht (x) = total mortality hazard, age i, dose x
qt = /'(survive /-th time interval | survive (M)-th time
interval)
exp(-5 V(*)) (2)
S, = /'(survive up to end of (f'-l)-th time interval)
/-I
<7y (3)
The <7,'s and S,'s also depend on x through equations (2) and (3), but the
dependence is not shown, and in fact is sometimes ignored when h, is small relative
to h', so that /»,*(*) = A/(0) for A: of interest.
NRC (2001, p. 186) printed two errors related to (1):
• they wrote h, in place of h* in equation (2);
• they wrote (l-<7,) in place of q^ in equation (3).
Now we use a linear relative hazard model
hfr) - )
where h, =A,(0) and ft is the relative risk slope that we estimate by regression. The
additional dose-related cancer hazard is h,px, so the total mortality hazard
DRAFT - DO NOT CITE OR QUOTE 102
-------
increases at the same time to h'(x) = h'+h,ftx where again h' = h'(Q). Therefore,
(1) becomes
and again the q,'s and S's also depend on x.
The previous EPA draft assessment for arsenic (USEPA, 2001) did not
adjust the Taiwanese results to the U.S. population. NRC (2001) recommended the
use of the BEIR IV formula for this adjustment. The current analysis accepts the
NRC recommendation, with corrections.
6. Adjustment for dietary arsenic intake
In addition to arsenic intake from drinking and cooking water, measurable
amounts of dietary arsenic intake have been suggested to occur in southwest
Taiwan, with estimates in the range of 50 ng/day (Schoof et al. 1998). For this
analysis, arsenic concentrations for the exposed (but not background) southwest
Taiwanese populations have been adjusted by adding either 0, 30, or 50 ug/day as
follows:
amtdi,t
whe/e:
concndj = the diet-adjusted concentration of arsenic
concdw = the actual well-water arsenic concentration
amtdiet = the daily dietary arsenic intake
voldw = the daily drinking water intake
volcw = the daily cooking water intake
As is evident in Table B-2, the different assumptions as to dietary arsenic
intake have no significant effect upon the calculation of cancer slope factors to 2
significant figures. The previous EPA draft assessment for arsenic (USEPA, 2001)
used a value of 50 ug/day for dietary arsenic intake. Although NRC (2001) found
little or no support for that assumption, they stated that the results would be
insensitive to the assumption. The current analysis supports the NRC assertion and
uses their recommended value of 30 [ig/day for dietary arsenic intake.
DRAFT - DO NOT CITE OR QUOTE 103
-------
Table B-2. Sensitivity of cancer slope factor to dietary arsenic assumptions
site
lung
lung
lung
lung
lung
lung.
bladder
bladder
bladder
bladder
bladder
bladder
sex
. female
female
female
male
male
male
female
female
female
male
male
male
dietary
arsenic
we/day
0
30
50
0
30
50
0
30
50
0
30
50
tu01
u-g/mL
79
77
72
92
91
100
364
369
375
260
255
260
LbL)0,
|ig/mL
67
69
65
82
84
82
332
346
340
233
240
237
slope
(ug/mL)-'
1.5E-04
1.4E-04
1.5E-04
1.2E-04
1.2E-04
1.2E-04
3.0E-05
2.9E-05
2.9E-05
4.3E-05
4.2E-05
4.2E-05
7. Adjustment for drinking water volumes
The calculated risk associated with arsenic exposure through drinking water
is, as expected, inversely linearly related to the assumed volumes of drinking water
for the study population. For this analysis, the total drinking water volume
(including cooking water) is assumed to be 2 L/day for Taiwan and 1 L/day for the
US. When combined with the differences in mean body weights between the
Taiwanese and US populations (50 kg vs.70 kg), the resulting dose adjustment
factor is 3. This approach is different from that of the previous EPA draft
assessment for arsenic (USEPA, 2001), which used a Monte Carlo analysis to
generate a distribution of exposures. NRC (2001) noted the broad interindividual
variability in water intake and the lack of data specific to the southwest Taiwanese
population. The current analysis does not directly address this issue.
8. Adjustment for cooking water
In addition to arsenic intake from drinking water and dietary sources,
measurable amounts of arsenic intake from cooking water have been suggested to
occur in southwest Taiwan. Estimates of the volume of water consumed by this
route are estimated at 1 L/day (USEPA, 1989). For this analysis, 1 L/day is added
to the Taiwanese drinking water volume to account for cooking water. This is in
accordance, with what was done in the previous EPA draft assessment for arsenic
(USEPA, 2001). NRC (2001) agreed with EPA's method for accounting for the
extra water consumption due to use of drinking water in cooking food, although
they noted that the rationale for using 1 L/day was not documented.
9. Adjustment for mortality
The hazard (or relative risk) of cancer due to arsenic exposure has been
estimated based on Taiwanese mortality data, as opposed to incidence data. The
results are then used to calculate lifetime risk for the U.S. population, using the
DRAFT - DO NOT CITE OR QUOTE
104
-------
BEIR IV formula (NRC, 2001) and the U.S. background cancer mortality rates. The
BEIR IV formula adjusts for US background mortality rates, resulting in a relative
risk. Therefore, no additional adjustment is needed between incidence and
mortality. The previous EPA draft assessment for arsenic (USEPA, 2001) did not
adjust the Taiwanese results to the U.S. population, and therefore required an
adjustment between mortality and incidence. NRC (2001) recommended the use of
the BEIR IV formula for thi's adjustment. The current analysis accepts the NRC
recommendation.
10. Model Implementation Details
The additive Poisson models for each of the endpoints (lung and bladder
cancer in males and females) and their related BEIR IV calculations are
implemented using two Microsoft Excel 2002 SP3 spreadsheets: MCcancerfit.xls
and BEIR.xls. A listing of the contents of these spreadsheets is in Table B-3. The
process for running the calculations is a multi-step algorithm:
1) In MCcancerfit.xls, for each of flung (female lung cancer), mlung (male lung
cancer), fblad (female bladder cancer), and mblad (male bladder cancer), run
the Solver add-in to maximize sum (cell N2) by adjusting al, a2, a3, and b
(cells Gl :G4). This provides starting values for the Bayesian regression.
2) In MCcancerfit.xls, for each of MC flung, MC mlung, MC fblad, and MC
mblad, paste the starting values into al .start, a2.start, a3.start, and b.start
(cells N1 :N4). Force a recalculation (key F9 on Windows PCs). Random
values for al, a2, a3, and b are calculated and placed into cells E:H.
4) In.MCcancerfit.xls, for each of MC flung, MC mlung, MC fblad, and MC
mblad, run the respective macro MCflung, MCmlung, MCfblad, or
MCmblad. These macros calculate the log-likelihood of the data given the
parameter values by placing the random parameter values (cells E:H) in the
model calculation sheets (flung, mlung, fblad, or mblad) in cells Gl :G4 and
paste the resulting log-likelihood (cell N2) back into the respective MC
sheet (column I). Likelihoods (scaled by the maximum likelihood to avoid
numeric instability) are then calculated. The mean and standard deviation
values for beta, weighted by the likelihood function, are calculated and
placed in cells N8 and N9, respectively. Upper and lower bounds (95%
confidence limits) on beta are calculated by adding or subtracting 2 x the
weighted standard deviation to the weighted mean (cells N10 and Nil).
5) In BEIR.xls, for each of flung, mlung, fblad, and mblad, the beta values
(mean, upper, or lower) calculated in MCcancerfit.xls are placed in b (cell
T10). Risks at given arsenic concentrations are calculated by adjusting x
(cell U10) and reading the total relative risk at sum(er) (cell Tl 5).
Concentrations corresponding to a given risk level can be calculated using
either Solver or Goal Seek to seek out the value of x (cell U10) that gives a
particular value to sum(er) (cell Tl 5).
DRAFT - DO NOT CITE OR QUOTE 105
-------
There are several user issues involving implementation of these
spreadsheets:
1) These spreadsheets use macros. Security levels on machines used to run
these spreadsheets must be adjusted to allow signed macros to run.
2) Macro run time is expected to be several minutes on a 3 gigahertz Pentium 4
PC with 512 KB RAM running Windows XP SP2 and Microsoft Excel 2002
SP3. Other computers may take more or less time. These spreadsheets have
not been tested for compatibility with other CPUs, operating systems, or
versions of Microsoft Excel.
3) The Solver add-in is required for setting initial parameter values in
MCcancerfit.xls. Solver is not used in the calculation of final parameter
values. Solver may be used in BEIR.xls calculations, but it is not required.
4) Automatic spreadsheet calculation is switched off in MCcancerfit.xls to
prevent unintended recalculation of random parameter values.
The intent of these spreadsheets is to make the calculations underlying the
arsenic cancer risk assessment as transparent as possible, thereby facilitating peer
review and subsequent modification. This is not intended to be the most rapid,
computationally efficient implementation possible. Nor is it intended to run on the
platforms of choice of most statisticians (SAS, S-PLUS, R, etc.). By programming
these calculations in Microsoft Excel, they are available, accessible,
understandable, and modifiable by the largest possible audience of peer reviewers
and stakeholders.
DRAFT-DO NOT CITE OR QUOTE 106
------- |