DRAFT—DO NOT CITE OR QUOTE EPA/63 5/R-10/001 www.epa.gov/iris TOXICOLOGICAL REVIEW OF INORGANIC ARSENIC (CAS No. 7440-38-2) In Support of Summary Information on the Integrated Risk Information System (IRIS) February 2010 NOTICE This document is a final draft. This information is distributed solely for the purpose of pre-dissemination peer review under applicable information quality guidelines. It has not been formally disseminated by EPA. It does not represent and should not be construed to represent any Agency determination or policy. It is being circulated for review of its technical accuracy and science policy implications. U.S. Environmental Protection Agency Washington, DC ------- DISCLAIMER This document is a final draft for review purposes only. This information is distributed solely for the purpose of pre-dissemination peer review under applicable information quality guidelines. It has not been formally disseminated by EPA. It does not represent and should not be construed to represent any Agency determination or policy. Mention of trade names or commercial products does not constitute endorsement or recommendation for use. ii DRAFT—DO NOT CITE OR QUOTE ------- TABLE OF CONTENTS—TOXICOLOGICAL REVIEW of INORGANIC ARSENIC (CAS No. 7440-38-2) LIST OF TABLES VI LIST OF FIGURES VII LIST OF ABBREVIATIONS VIII FOREWORD XIII AUTHORS, CONTRIBUTORS, AND REVIEWERS XIV 1. INTRODUCTION 1 2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS 3 2.1. PROPERTIES 3 2.2. USES 3 2.3. OCCURRENCE 4 2.4. ENVIRONMENTAL FATE 5 3. TOXICOKINETICS 6 3.1. ABSORPTION 7 3.2. DISTRIBUTION 9 3.2.1. Transport in Blood 9 3.2.2. Tissue Distribution 10 3.2.3. Cellular Uptake, Distribution, and Transport 14 3.3. METABOLISM 15 3.3.1. Reduction 19 3.3.2. Arsenic Methylation 20 3.3.3. Species Differences in the Methylation of Arsenic 23 3.3.4. Thioarsenical Metabolites 24 3.4. ELIMINATION 26 3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS 27 4. HAZARD IDENTIFICATION 31 4.1. STUDIES IN HUMANS 31 4.1.1. Taiwan 32 4.1.2. Japan 46 4.1.3. South America 47 4.1.4. North America (United States and Mexico) 52 4.1.5. China 58 4.1.6. Finland 59 4.1.7. Denmark 61 4.1.8. Australia 61 4.2. PRECHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN ANIMALS—ORAL 62 4.2.1. Prechronic and Chronic Studies 62 4.2.2. Cancer Bioassays 62 4.2.2.1. Mice—Transplacental 62 iii DRAFT—DO NOT CITE OR QUOTE ------- 4.2.2.2. Rat—Oral 65 4.2.2.3. Other 66 4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL 67 4.4. OTHER STUDIES 67 4.4.1. Possible Modes of Action and Key Events of Possible Importance 67 4.4.1.1. In Vivo Human Studies 71 4.4.1.2. In Vivo Experiments Using Laboratory Animals 75 4.4.1.3. In Vitro Experiments 81 4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS 93 4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER CHARACTERIZATION 94 4.6.1. Summary of Overall Weight-of-Evidence 94 4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence 94 4.6.2.1. Skin Cancer 96 4.6.2.2. Lung Cancer 96 4.6.2.3. Kidney, Bladder, and Liver Cancer 97 4.6.2.4. In Utero Exposure 98 4.6.3. Mode of Action Information 98 4.6.3.1. General Comments onMOAs 98 4.6.3.2. Low-Dose Extrapolation 102 4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 102 4.7.1. Possible Childhood Susceptibility 102 4.7.2. Possible Gender Differences 105 4.7.3. Other 106 4.7.3.1. Genetic Polymorphism 106 4.7.3.2. Nutritional Status 109 4.7.3.3. Cigarette Smokers Ill 5. DOSE-RESPONSE ASSESSMENTS 112 5. LORAL REFERENCE DOSE (RfD) 112 5.2. INHALATION REFERENCE CONCENTRATION (RfC) 112 5.3. CANCER ASSESSMENT (ORAL EXPOSURE) 112 5.3.1. Background: History of Cancer Risk Assessments for Arsenic 112 5.3.2. Choice of Study/Data, Estimation Approach, and Input Assumptions 119 5.3.3. Dose-Response Model Selection for Cancer Mortality in Taiwan 120 5.3.4. Selection of Cancer Endpoints and Estimation of Risks for U.S. Populations 121 5.3.5. Nonwater Arsenic Intake and Drinking Water Consumption 123 5.3.6. Dose-Response Data 125 5.3.7. Risk Assessment Methodology 126 5.3.7.1. Dose-Response Estimation Based on Taiwan Cancer Mortality Data 127 5.3.7.2. Estimation of Confidence Limits on Cancer Slope Parameters 128 5.3.7.3. Estimation of LEDOT Values Using Relative Risk Models 129 5.3.7.4. Estimation of Unit Risks 129 5.3.8. Results 130 5.3.8.1. Ingestion Pathway Oral CSFs and Unit Risks 130 5.3.8.2. Comparison to Previous Cancer Risk Estimates 132 5.3.8.3. EDoi and LEDM Estimates From Chen et al. (1988a, 1992), Ferreccio iv DRAFT—DO NOT CITE OR QUOTE ------- et al. (2000), and Chiou et al. (2001) 133 5.3.8.4. Estimated Risk Associated With 10 ug/L Drinking Water Arsenic From NRC (2001) and U.S. EPA (2005c) 135 5.3.8.5. Sensitivity Analyses of Cancer Risk Estimates to Changes in Parameter Values 137 5.3.8.6. Sensitivity Analyses of Cancer Risk Estimates to Dose-Response Model Form 142 5.3.8.7. Significance of Cancer Risks at Low Arsenic Exposures 143 5.4. CANCER ASSESSMENT (INHALATION EXPOSURE) 145 6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE- RESPONSE 146 6.1. HUMAN HAZARD POTENTIAL 146 6.2. DOSE-RESPONSE 148 6.2.1. Choice of Models 150 6.2.2. Dose Metric 151 6.2.3. Human Population Variability 151 7. REFERENCES 153 APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC COMMENTS AND DISPOSITION A-l APPENDIX B. TABULAR DATA ON CANCER EPIDEMIOLOGY STUDIES B-l APPENDIX C. TABLES FOR STUDIES ON POSSIBLE MODE OF ACTION FOR INORGANIC ARSENIC C-l APPENDIX D. IMMUNOTOXICITY D-l APPENDIX E. QUANTITATIVE ISSUES IN THE CANCER RISK ASSESSMENT FOR INORGANIC ARSENIC E-l APPENDIX F. RISK ASSESSMENT FOR TOWNSHIPS AND LOW-EXPOSURE TAIWANESE POPULATIONS F-l DRAFT—DO NOT CITE OR QUOTE ------- LIST OF TABLES Table 2-1. Chemical and Physical Properties of Arsenic and Selected Inorganic Arsenic Compounds (ATSDR, 2000; Merck Index, 1989) 4 Table 4-1. Summary of Number of Rows Derived From Peer-Reviewed Publications for Different Hypothesized Key Eventsa 71 Table 5-1. Historical Summary of Arsenic Risk Assessment Efforts 114 Table 5-2. Cancer Mortality Data Used in the Arsenic Risk Assessment 126 Table 5-3. Cancer Incidence Risk Estimates for Lung and Bladder Cancers in Males and Females21 131 Table 5-4. Combined Lung and Bladder Cancer Incidence Risk Estimate for the U.S. Population (Males and Females) 132 Table 5-5. Comparison of EDoi and LEDMa Estimates From Past Studies'3 With Those From the Current Analysis 134 Table 5-6. Comparison of cancer risk assessment results with estimates from NRC (2001) and U.S. EPA(2005c) 135 Table 5-7. Drinking water intake and body weight assumptions in females in recent arsenic risk assessments 136 Table 5-8. Theoretical maximum likelihood estimates of excess lifetime risk (incidence per 10,000 people) of lung cancer and bladder cancer for US populations 137 Table 5-9. Arsenic oral CSFs (per mg/kg-d) for lung cancer and bladder cancer in US populations 137 Table 5-10. Sensitivity analysis of estimated cancer incidence risks associated with 10 ug/L to changes in modeling assumptions and inputs 139 Table 5-11. Proportional Changes in Cancer Risks at 10 ug/L Associated With Changes in Modeling Inputs and Assumptions 140 Table B-l. Taiwan Cancer Studies B-4 Table B-2. Japan Cancer Studies B-22 Table B-3. South America Cancer Studies B-23 Table B-4. North America cancer studies B-30 Table B-5. China cancer studies B-38 Table B-6. Finland cancer studies B-39 Table B-7. Denmark cancer studies B-41 Table B-8. Australia Cancer Studies B-42 Table C-l. In vivo human studies related to possible modes of action of arsenic in the development of cancer C-l9 Table C-2. In vivo experiments on laboratory animals related to possible modes of action of arsenic in the development of cancer—only oral exposures C-30 Table C-3. In vitro studies related to possible MOA of arsenic in the development of cancerC-57 Table D-l. Lymphocyte counts and labeling, mitotic, and replication indexes (mean ± se) in the peripheral blood lymphocytes in populations exposed to low (control) and high (exposed) levels of arsenic (Gonsebatt et al., 1994) D-2 Table F-l. Coefficients from linear regressions of age-adjusted cancer risk versus arsenic doses for townships identified by Lamm et al. (2006) F-5 Table F-2. Arsenic dose coefficients for study populations with median well water arsenic concentrations less than 127 ppb F-6 vi DRAFT—DO NOT CITE OR QUOTE ------- LIST OF FIGURES Figure 3-1. Traditional metabolic pathway for inorganic arsenic in humans 16 Figure 3-2. Alternative metabolic pathway for inorganic arsenic in humans proposed by Hayakawa et al. (2005) 17 Figure 3-3. Thioarsenical structures 25 Figure 4-1. Level of significant exposure of adult mice to sodium arsenite in drinking water in ppm As 76 Figure 5-1. Estimated oral CSFs for individual and combined cancer endpoints 132 Figure 5-2. Change in arsenic-related unit risk estimates associated with variations in input assumptions 140 Figure F-l. Lifetime crude total cancer risk (male + female) for the low- and high-exposure villages F-4 vii DRAFT—DO NOT CITE OR QUOTE ------- LIST OF ABBREVIATIONS 293 cells 2-AAAF 8-OHdG AG06 cells AGT AIC AMI AP APE As As111 Asv AS3MT AQP ATG ATO ATSDR B[or]P BBDR BCC BER BFD BMI BPDE BrdU BSO BW or bw CA Caco-2 CAE CASRN CAT CCA CCRIS cDNA cen+ cen- Chang cells CHO approximately (if before a listing of concentrations, it applies to all) a cell line derived from adenovirus-transformed human embryonic kidney epithelial cells 2-acetoxyacetylaminofluorene 8-hydroxydeoxyguanosine SV40-transformed human keratinocytes average generation time Akaike information criterion acute myocardial infraction activator protein or activating protein apurinic/apyrimidinic endonuclease arsenic arsenite arsenate arsenic(+3 oxidation state) methyltransferase aquaglycoporins arsenic triglutathione arsenic trioxide Agency for Toxic Substances and Disease Registry benzo[a]pyrene biologically based quantitative dose-response basal cell carcinoma base excision repair blackfoot disease body mass index benzo[a]pyrene diol epoxide, an active metabolite ofB[or]P bromodeoxyuridine L-buthionine-S,R-sulphoximine (depletes GSH, y- GCS inhibitor) body weight chromosome aberrations a human intestinal cell line cumulative arsenic exposure Chemical Abstracts Service Registry Number catalase (decomposes H2O2) chromate copper arsenate Chemical Carcinogenesis Research Information System complementary DNA centromere positive centromere negative a human cell line thought to be derived from HeLa cells Chinese hamster ovary viii DRAFT—DO NOT CITE OR QUOTE ------- Ill III CI c-Jun or c-jun CL3 cells COPD CSF DEB DBS dhfr gene DHLP DI-I or II or dL DMA DMA DMAV DMAG DMMTA111 DMMTAV DMPS DMSA DNA DNMT DTT DW E. coli ED EGFR-ECD EPA ER-a ERCC1 ERCC2 ERK FAK FPG G6PDH GAPDH GI GLM GM04312C cells GM-CSF confidence interval an AP-1 protein human lung adenocarcinoma cells (established from a non-small-cell lung carcinoma) chronic obstructive pulmonary disease cancer CSF diepoxybutane (DNA crosslinking agent) diethylstilbestrol dihydrofolate reductase gene dihydrolipoic acid iodothyronine deiodinase-I or II orIn (are 3 forms of this selenoenzyme) deciliter dimethyl arsenic (used when the oxidative state is unknown or not specified) dimethylarsenous acid dimethylarsinic acid dimethylarsinic glutathione dimethylmonothioarsinic acid dimethylmonothioarsonic acid 2,3-dimercaptopropane-l-sulfonic acid dimercaptosuccinic acid or meso 2,3- dimercaptosuccinic acid deoxyribonucleic acid DNA methyltransferase dithiothreitol drinking water Escherichia coli effective dose extracellular domain of the epidermal growth factor receptor Environmental Protection Agency estrogen receptor-alpha excision repair cross-complement 1 component excision repair cross-complementing rodent repair deficiency, complementation group 2 (also known as xeroderma pigmentosum group D or XPD) extracellular signal-regulated kinase focal adhesion kinase formamidopyrimidine-DNA glycosylase (digestion of DNA) glucose-6-phosphate dehydrogenase glyceraldehyde-3-phosphate dehydrogenase gastrointestinal generalized linear model a SV-40 transformed XPA human fibroblast NER- deficient cell line granulocyte-macrophage colony-stimulating factor ix DRAFT—DO NOT CITE OR QUOTE ------- GPx GSH GST GSTO1 GSTP1-1 H69AR H9c2 cells HAC HCC HEALS HELP cells HepG2 cells HGPRT hGST-Ol HMOX-1 hOGGl HPBM HSDB HXT IC50 IFN-y IL ILK IRIS IRR iv JAK LED LI LOEC LOEL MADG MAP MCF-7 cells M-CSF MDA mdm2 MEK MI MLE MMA MMAm MMAV MMS MN glutathione peroxidase glutathione glutathione-S-transferase glutathione-^-transferase omega 1 glutathione-^-transferase Pl-1 a multi-drug resistant human cancer cell line immortalized myoblast cell line derived from fetal rat hearts highest arsenic concentration hepatocellular carcinoma Health Effects of Arsenic Longitudinal Study human embryo lung fibroblast cell line human hepatocellular liver carcinoma cell line (Caucasian) hypoxanthine-guanine phosphoribosyltransferase human glutathione-S-transferase omega 1 heme oxygenase 1 human 8-oxoguanine DNA glycosylase human peripheral blood monocytes Hazardous Substances Data Bank hexose permease transporters concentration that is needed to cause 50% inhibition interferon-gamma interleukin integrin-linked kinase Integrated Risk Information System incidence rate ratio intravenous Janus kinase lowest effective dose labeling index lowest observed effect concentration lowest observed effect level monomethylarsonic diglutathione mitogen-activated protein human breast carcinoma cell line macrophage colony-stimulating factor malondialdehyde murine double minute 2 proto-oncogene MAP/ERK kinase mitotic index maximum likelihood monomethyl arsenic (used when oxidative state is unknown or not specified) monomethylarsonous acid monomethyl arsonic acid methyl methanesulfonate micronuclei x DRAFT—DO NOT CITE OR QUOTE ------- MNU MOA MPR2/cMOAT mRNA MRP MTHFR NAC NAD NADPH NCHS NCI NER NHEK cells NK NO NRC OATP-C ODC OGG1 OPP OR PARP PBPK model PBMC PCNA PCR PGK PHA PMI PNP POD ppb ppm PTEN PYR R15 RAGE RBCs RED RfC RfD RI RNS ROS RR RT N-methyl-N-nitrosourea mode of action multi-drug resistance associated protein 2 transporter messenger ribonucleic acid multidrug resistance protein methylene trihydrofolate reductase w-acetyl-cysteine nicotinamide adenine dinucleotide nicotinamide adenine dinucleotide phosphate- oxidase National Center for Health Statistics National Cancer Institute nucleotide excision repair primary normal human epidermal keratinocytes natural killer nitric oxide National Research Council organic anion transporting polypeptide-C ornithine decarboxylase 8-oxoguanine DNA glycosylase Office of Pesticide Programs odds ratio poly(adenosine diphosphate-ribose) polymerase physiologically based pharmacokinetic model peripheral blood mononuclear cells proliferating cell nuclear antigen polymerase chain reaction phosphoglyerate kinase phytohemagglutinin primary methylation indices purine nucleoside phosphorylase point of departure parts per billion parts per million phosphatase and tensin homolog person-years at risk arsenic-resistent cells receptor for advanced glycation end products red blood cells Reregi strati on Eligibility Decision inhalation reference concentration oral reference dose replication index reactive nitrogen species reactive oxygen species relative risk real time xi DRAFT—DO NOT CITE OR QUOTE ------- SAB SAM SBET SCC SCE SCGE Se SEER SHE cells SIR SMI SMR SOD STAT SV-HUC-1 cells T3 T4 TAT TCEP Tg.AC TGF-a TMAm TMAV TMAO TNF-a TPA Trx TrxR TWA UCL UROtsa UV V79 cells VEGF XRCC1 Science Advisory Board S-adenosylmethionine simplified bioaccessibility extraction test squamous cell carcinoma sister chromatid exchange single cell gel electrophoresis selenium surveillance epidemiology and end result Syrian hamster ovary cells standardized incidence ratio secondary methylation indices standard mortality ratio superoxide radical dismutase signal transducer and activator of transcription SV40 large T-transformed human urothelial cell line thyroid hormone triiodothyronine thyroid hormone thyroxine tyrosine aminotransferase tris(2-carboxylethyl)phospine a strain of transgenic mice that contains the fetal beta-globin promoter fused to the v-Ha-ras structural gene (with mutations at codons 12 and 59) and linked to a simian virus 40 polyadenylation/splice sequence transforming growth factor alpha trimethyl arsine trimethylarsinic acid trimethylarsine oxide tumor necrosis factor alpha 12-O-tetradecanoyl phorbol-13-acetate thioredoxin thioredoxin reductase time-weighted average upper confidence limits a SV40-immortalized human urothelium cell line ultraviolet radiation a cell line derived from lung fibroblasts of a male Chinese hamster vascular endothelial cell growth factor X-ray repair cross-complimentary group 1 xii DRAFT—DO NOT CITE OR QUOTE ------- FOREWORD 1 The purpose of this Toxicological Review is to provide scientific support and rationale 2 for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to inorganic 3 arsenic. It is not intended to be a comprehensive treatise on the chemical or toxicological nature 4 of inorganic arsenic. 5 The intent of Section 6, "Major Conclusions in the Characterization of Hazard and Dose 6 Response," is to present the major conclusions reached in the derivation of the reference dose, 7 reference concentration, and cancer assessment, where applicable, and to characterize the overall 8 confidence in the quantitative and qualitative aspects of hazard and dose-response by addressing 9 the quality of data and related uncertainties. The discussion is intended to convey the limitations 10 of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk 11 assessment process. 12 For other general information about this assessment or other questions relating to IRIS, 13 the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or 14 hotline.iris@epa.gov (email address). xiii DRAFT—DO NOT CITE OR QUOTE ------- AUTHORS, CONTRIBUTORS, AND REVIEWERS CHEMICAL MANAGER/AUTHOR Santhini Ramasamy, Ph.D., MPH, DABT Office of Science and Technology Office of Water U.S. Environmental Protection Agency Washington, DC OFFICE OF RESEARCH AND DEVELOPMENT CO-LEAD/AUTHOR Reeder Sams, Ph.D. National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Research Triangle Park, NC AUTHORS Robyn B. Blain, Ph.D. Gregory M. Blumenthal, Ph.D. William M. Mendez, Ph.D. Welford C. Roberts, Ph.D. ICF International Fairfax, VA Paul B Selby, Ph.D., DABT RiskMuTox Oak Ridge, TN Arthur W. Stange, Ph.D. Oak Ridge Associated Universities Arvada, CO 80005 Susan M. Wells, M.P.H. Oak Ridge Associated Universities Oak Ridge, TN 37831-0117 CONTRIBUTORS Elizabeth Doyle, Ph.D. Office of Science and Technology Office of Water U.S. Environmental Protection Agency Washington, DC xiv DRAFT—DO NOT CITE OR QUOTE ------- Jonathan Chen, Ph.D. Office of Pesticide Programs U.S. Environmental Protection Agency Washington, DC Andrew Schulman, Ph.D. Office of Enforcement and Compliance Assurance U.S. Environmental Protection Agency Washington, DC Chao Chen, Ph.D. National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Washington, DC Paul White, M.S. National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Washington, DC Irene Dooley Office of Water U.S. Environmental Protection Agency Washington, DC Brenda Foos, Ph.D. Office of Children's Health Protection U.S. Environmental Protection Agency Washington, DC Molly Rosett National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Research Triangle Park, NC xv DRAFT—DO NOT CITE OR QUOTE ------- REVIEWERS This document has been reviewed by EPA scientists, interagency reviewers from other federal agencies, and the public, and peer reviewed by independent scientists external to EPA. A summary and EPA's disposition of the comments received from the independent external peer reviewers and from the public is included in Appendix A. INTERNAL EPA REVIEWERS Ila Cote, Ph.D. National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Joyce Morrissey Donohue, Ph.D. Office of Science and Technology Office of Water U.S. Environmental Protection Agency Hisham El-Masri, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Nicole Hagan ORISE, National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Elaina Kenyon, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Kirk Kitchin, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Andrew Kligerman, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Danelle Lobdell, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency xvi DRAFT—DO NOT CITE OR QUOTE ------- Bob Hetes National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Stephen Nesnow, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Julian Preston, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency David Thomas, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency John Vandenberg, Ph.D. National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Tim Wade, Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency Debra Walsh National Center for Environmental Assessment Office of Research and Development U.S. Environmental Protection Agency Doug Wolf, D.V.M., Ph.D. National Health and Environmental Effects Research Laboratory Office of Research and Development U.S. Environmental Protection Agency xvii DRAFT—DO NOT CITE OR QUOTE ------- EXTERNAL PEER REVIEWERS Science Advisory Board Arsenic Review Panel CHAIR Genevieve Matanoski, M.D., Ph.D. Johns Hopkins University MEMBERS H. Vasken Aposhian, Ph.D. The University of Arizona Aaron Barchowsky, Ph.D. University of Pittsburgh David Brusick, Ph.D. Retired, Convance Labs Kenneth P. Cantor, Ph.D. National Cancer Institute John (Jack) Colford, Ph.D. University of California Yvonne P. Dragan, Ph.D. National Center for Toxicological Research, Food and Drug Administration Sidney Green, Ph.D. Howard University Sioban Harlow, Ph.D. University of Michigan Steven Heeringa, Ph.D. University of Michigan Claudia Maria Hopenhayn, Ph.D. University of Kentucky James E. Klaunig, Ph.D. Indiana University X. Chris Le, Ph.D. University of Alberta xviii DRAFT—DO NOT CITE OR QUOTE ------- Michele Medinsky, Ph.D. Toxcon Kenneth Portier, Ph.D. American Cancer Society Atlanta, GA Barry Rosen, Ph.D. Wayne State University Toby Rossman, Ph.D. New York University Miroslav Styblo, Ph.D. University of North Carolina Justin Teeguarden, Ph.D. Pacific Northwest National Laboratory Michael Waalkes, Ph.D. National Institute of Environmental Health Science Janice Yager, Ph.D. Electric Power Research Institute xix DRAFT—DO NOT CITE OR QUOTE ------- 1. INTRODUCTION 1 This document presents background information and justification for the Integrated Risk 2 Information System (IRIS) Summary of the hazard and dose-response assessment of inorganic 3 arsenic. The IRIS Summary may include oral reference dose (RfD) and inhalation reference 4 concentration (RfC) values for chronic and other exposure durations, as well as a carcinogenicity 5 assessment. 6 This document is based on EPA reviews of the reports Arsenic in Drinking Water and 7 Arsenic in Drinking Water, 2001 Update published by the National Research Council (NRC) in 8 1999 and 2001, respectively. In writing those reports, the NRC arsenic committee considered 9 presentations at the committee's public meetings, comments from the public, and the comments 10 made by technical experts on the draft NRC arsenic reports. The conclusions, recommendations, 11 and final content of the NRC (1999, 2001) reports rest entirely with the committee and the NRC. 12 This IRIS document—based on reviews of those reports—has undergone evaluation by EPA 13 health scientists from several program offices and regional offices, interagency review, and 14 external peer review by the Science Advisory Board (SAB). 15 Compared to the draft Toxicological Review submitted to the SAB in 2005, this 16 assessment is expanded: it provides a detailed review of epidemiological studies and the mode of 17 action (MOA) studies, as well as revisions to the dose-response analysis to address the 18 recommendations of the SAB (SAB, 2007). Specifically, it includes additional sensitivity 19 analyses on the effects of modeling assumptions on estimated cancer risk. 20 The RfD and RfC, if derived, provide quantitative information for use in risk assessments 21 for health effects known or assumed to be produced through a non-linear (presumed threshold) 22 MO A. The RfD (expressed in units of mg/kg-day) is defined as an estimate (with uncertainty 23 spanning perhaps an order of magnitude) of a daily exposure to the human population (including 24 sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a 25 lifetime. The inhalation RfC (expressed in units of mg/m3) is analogous to the oral RfD, but 26 provides a continuous inhalation exposure estimate. The inhalation RfC considers both toxic 27 effects on the respiratory system (portal of entry) and toxic effects peripheral to the respiratory 28 system (extrarespiratory or systemic effects). Reference values are generally derived for chronic 29 exposures (up to a lifetime), but may also be derived for acute (< 24 hours), short-term (>24 30 hours to 30 days), and subchronic (>30 days to 10% of lifetime) exposure durations, all of which 31 are derived based on an assumption of continuous exposure throughout the duration specified. 32 Unless specified otherwise, the RfD and RfC are derived for chronic exposure duration. 33 The carcinogenicity assessment provides information on the carcinogenic hazard 34 potential of the substance in question and quantitative estimates of risk from oral and inhalation 35 exposures may be derived. The information includes a weight-of-evidence judgment of the 36 likelihood that the agent is a human carcinogen and the conditions under which the carcinogenic 1 ------- 1 effects may be expressed. Quantitative risk estimates may be derived from the application of a 2 low-dose extrapolation procedure. If derived, the oral cancer CSF (CSF) is a plausible upper 3 bound on the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit risk 4 is a plausible upper bound on the estimate of risk per ug/m3 air breathed. 5 Development of these hazard identification and dose-response assessments for inorganic 6 arsenic has followed the general guidelines for risk assessment set forth by the National 7 Research Council (NRC, 1983). EPA Guidelines and Risk Assessment Forum Technical Panel 8 Reports that may have been used in the development of this assessment include the following: 9 Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines 10 for Mutagenicity Risk Assessment (U. S. EPA, 1986b), Recommendations for and Documentation 11 of Biological Values for Use in Risk Assessment (U.S. EPA, 1988a), Guidelines for 12 Developmental Toxicity Risk Assessment (U.S. EPA, 1991), Use of the Benchmark Dose 13 Approach in Health Risk Assessment (U.S. EPA, 1995), Guidelines for Reproductive Toxicity 14 Risk Assessment (U.S. EPA, 1996), Guidelines for Neurotoxicity Risk Assessment (U.S. EPA, 15 1998), Science Policy Council Handbook: Peer Review (U. S. EPA, 2000a), Science Policy 16 Council Handbook: Risk Characterization (U.S. EPA, 2000b), Benchmark Dose Technical 17 Guidance Document (U.S. EPA, 2000c), Supplementary Guidance for Conducting Health Risk 18 Assessment of Chemical Mixtures (U.S. EPA, 2000d), A Review of the Reference Dose and 19 Reference Concentration Processes (U.S. EPA, 2002), Guidelines for Carcinogen Risk 20 Assessment (U. S. EPA, 2005a), Supplemental Guidance for Assessing Susceptibility from Early- 21 Life Exposure to Carcinogens (U.S. EPA, 2005b), Science Policy Council Handbook: Peer 22 Review (U.S. EPA, 2006a), and A Framework for Assessing Health Risks of Environmental 23 Exposures to Children (U.S. EPA, 2006b). 24 The literature search strategy employed for this compound was based on the Chemical 25 Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent 26 scientific information submitted by the public to the IRIS Submission Desk was also considered 27 in the development of this document. The relevant literature was reviewed through December, 28 2007; however, a few references from 2008 have also been included. 29 ------- 2. CHEMICAL AND PHYSICAL INFORMATION RELEVANT TO ASSESSMENTS 2.1. PROPERTIES 1 Arsenic (As) is a metalloid that can exist in the -3, 0, +3, and +5 oxidation states.l The 2 arsenite (As111; +3) and arsenate (Asv ; +5) forms are the primary forms found in drinking water. 3 The chemical and physical properties of arsenic are listed in Table 2-1. 2.2. USES 4 The metalloid, arsenic, is used for hardening copper and lead alloys (HSDB, 2005). It 5 also is used in glass manufacturing as a decolorizing and refining agent, as a component of 6 electrical devices, in the semiconductor industry, and as a catalyst in the production of ethylene 7 oxide. Arsenic compounds are used as a mordant in the textile industry, for preserving hides, as 8 medicinals, pesticides, pigments, and wood preservatives. Production of chromate copper 9 arsenate (CCA), a wood preservative whose production is currently being phased out, accounts 10 for about 90% of the domestic consumption of arsenic (ATSDR, 2007). 11 1 Oxidation states for arsenic have been abbreviated differently by different organizations or authors. For example, arsenite can be abbreviated as either " As(m)' document uses the superscript abbreviation. arsenite can be abbreviated as either "As(m)" or "As111"; both refer to trivalent inorganic arsenic compounds. This 3 DRAFT—DO NOT CITE OR QUOTE ------- Table 2-1. Chemical and Physical Properties of Arsenic and Selected Inorganic Arsenic Compounds (ATSDR, 2000; Merck Index, 1989) CAS No. Oxidation State Molecular Weight Synonyms Physical State (25°C) Boiling Point (°C) Melting Point (°C) Density (g/cm3) Vapor Pressure (20°C) Taste Threshold Odor Threshold Conversion Factor Arsenic 7440-38-2 0 74.9 metallic arsenic, gray arsenic solid 613 (sublimes) 817@28atm 5.7 — — — — As2O3 1327-53-3 +3 197.8 arsenic trioxide, arsenolite, white arsenic (+3) solid 465 312 3.7 — — — — As2O5 1303-28-2 +5 229.8 arsenic pentoxide, arsenic acid anhydride (+5) solid — 315 (decompose) 4.3 — — — — NaAsO2 7784-46-5 +3 129.9 sodium arsenite (+3) solid — — 1.8 — — — — Na2HAsO4 7778-43-0 +5 185.9 disodium arsenate (+5) solid — 86.3 1.8 — — — — — No data available 2.3. OCCURRENCE 1 Arsenic naturally makes up about 3.4 parts per million (ppm) of the Earth's crust, where 2 it is the twentieth most abundant element (ATSDR, 2007; Merck Index, 1989). Arsenic leaches 3 from natural weathering of soil and rock into water, and low concentrations of arsenic are found 4 in water, food, soil, and air. However, industrial activities such as coal combustion and smelting 5 operations release higher concentrations of arsenic to the environment (Adams et al., 1994). The 6 highest background arsenic levels found in the environment are in soils, with concentrations 7 ranging from 1 to 40 ppm (ATSDR, 2007). Food typically contains arsenic concentrations of 20 8 to 140 parts per billion (ppb) (ATSDR, 2007). The majority of surface and ground waters 9 contain less than 10 ppb (although levels of 1,000-3,400 ppb have been reported, especially in 10 areas of the western United States). Average arsenic content in drinking water in the United 11 States is 2 ppb; 12% of water supplies from surface water in the central United States and 12% 12 of ground water sources in the western United States exceed 20 ppb (ATSDR, 2007). Mean 13 arsenic concentrations in ambient air have generally been found to range from 1 to 2,000 ng/m3 14 (ATSDR, 2007). DRAFT—DO NOT CITE OR QUOTE ------- 2.4. ENVIRONMENTAL FATE 1 Arsenic as a free element (0 oxidation state) is rarely encountered in the environment 2 (HSDB, 2005). Under normal conditions in water, arsenic is present as soluble inorganic Asv 3 because it is more thermodynamically stable in water than As111. In soil there are many biotic 4 and abiotic processes controlling arsenic's overall fate and environmental impact. Arsenic in 5 soil exists in various oxidation states and chemical species, depending upon soil pH and 6 oxidation-reduction potential (ATSDR, 2007). Arsenic is largely immobile in agricultural soils, 7 and tends to remain in upper soil layers (ATSDR, 2007). However, reducing conditions form 8 soluble mobile forms of arsenic and leaching is greater in sandy soil than in clay loam (ATSDR, 9 2007). The most influential parameter affecting arsenic mobility is the iron content of the soil. 10 DRAFT—DO NOT CITE OR QUOTE ------- 3. TOXICOKINETICS 1 This Toxicological Review discusses oral waterborne arsenic exposure. It does not 2 specifically address inhalation exposures, though they are also common. Dermal exposure and 3 exposure from food consumption, however, can be significant and may be confounding variables 4 in epidemiological studies. Therefore, this report's toxicokinetic information focuses on oral 5 exposure from water sources, but absorption from dermal exposure and arsenic in food is also 6 briefly addressed. 7 The behavior of arsenic in the body is very complex. After absorption, inorganic arsenic 8 can undergo a complicated series of enzymatic and non-enzymatic oxidation, reduction, and 9 conjugation reactions. Although all these reactions may occur throughout the body, the rate at 10 which they occur varies greatly from organ to organ. In addition, there are important differences 11 in arsenic metabolism across animal species, and these variations make it difficult to identify 12 suitable animal models for predicting human metabolic patterns. 13 Each metabolic transformation affects the subsequent biokinetic behavior (transport, 14 persistence, elimination) and toxicokinetics of the arsenic species. Thus, absorption, transport, 15 and metabolic processes are highly interdependent and cannot easily be discussed separately. 16 The general pattern described in this chapter involves the gastrointestinal (GI) absorption of 17 inorganic arsenic species, followed by a cascade of oxidation-reduction reactions and 18 methylation steps, resulting in the partial transformation of the inorganic species into mono- or 19 dimethylated species (collectively referred to as MMA and DMA, recognizing that there is often 20 ambiguity in characterizing the oxidation state of the methylarsenic compounds). Conjugated 21 arsenic species, either methylated or not (e.g., glutathione conjugates or other sulfur-containing 22 derivatives), also may be produced. 23 As discussed in Section 3.3, several metabolic schemes have been proposed that describe 24 the general pathway that converts inorganic arsenic to its primary metabolites MMA and DMA. 25 These pathways involve numerous enzymes and cofactors. Some of the proposed metabolic 26 pathways involve the cycling of arsenic species back and forth between the +3 (trivalent) and +5 27 (pentavalent) oxidation states, and there is evidence that key metabolic processes may be 28 saturable, so that metabolic patterns differ with exposure levels. MMA, DMA, and inorganic 29 arsenic levels in tissues, blood, and urine are the most easily and frequently measured 30 metabolites; the relative levels of these compounds in blood or urine are often the primary 31 evidence in support of one or another metabolic pathway. Genomic tools are being increasingly 32 employed to better characterize human arsenic metabolism and to identify individuals at higher 33 risk from arsenic exposures. 34 DRAFT—DO NOT CITE OR QUOTE ------- 3.1. ABSORPTION 1 Water-soluble forms of inorganic arsenic (both trivalent and pentavalent) are readily 2 absorbed from the GI tract in experimental animal models (about 80-90% 0.62 mg/kg of sodium 3 arsenate; Freeman et al., 1995) as well as humans (Pomroy et al., 1980, who recovered 62% of a 4 0.06 ng dose of arsenic in seven days). Monomethyl arsonic acid (MMAV) and dimethylarsinic 5 acid (DMAV) also appear to be well absorbed (75-85%) in humans and experimental animals 6 (Stevens et al., 1977; Buchet et al., 1981; Yamauchi and Yamamura, 1984; Hughes et al., 2005). 7 Using an in vivo swine test, however, Juhasz et al. (2006) determined that MMA (oxidation 8 state not specified) and DMA (oxidation state not specified) were poorly absorbed, with only 9 16.7% and 33.3%, respectively, bioavailable. 10 Laparra et al. (2006) used a Caco-2 permeability model, which measured transport 11 through a monolayer of human intestinal cells, to examine the intestinal permeability of As111. A 12 decrease in the apical to basolateral permeability with increasing dose was found, indicating the 13 presence of a saturable intestinal transport system. The data also indicated that Caco-2 cells 14 have a secretory system for As111. In an earlier study, Laparra et al. (2005a) demonstrated that 15 the retention and transport of As111 in Caco-2 cells was more efficient than that of Asv. However, 16 this could have been due to the presence of phosphate in the culture medium, which would 17 compete with arsenate for transport across the membrane. 18 Gastrointestinal absorption of low-solubility arsenic compounds such as arsenic 19 trisulfide, lead arsenate, arsenic selenide, gallium arsenide (Mappes, 1977; Webb et al., 1984; 20 Yamauchi et al., 1986), and arsenic-contaminated soil (Freeman et al., 1995) is much less 21 efficient than that of soluble inorganic arsenic compounds. The degree of absorption of arsenic 22 from soil was found to be dependent on the arsenic species present in the soil and on the type of 23 soil. Juhasz et al. (2007) performed in vivo bioavailability studies in swine and determined that 24 the bioavailability of total arsenic in soils was highly variable, with a range of 6.9% to 74.7% 25 depending on the soil type. They also determined that a simplified bioaccessibility extraction 26 test (SBET; a rapid in vitro chemical extraction method) had results highly correlated with the in 27 vivo results. Therefore, they concluded that the less expensive in vitro test was just as effective 28 for determining bioavailability. 29 There is little information concerning the bioavailability of inorganic arsenic from 30 various types of food (NRC, 1999, 2001). However, there have been recent studies examining 31 the bioaccessibility of arsenic from rice (Laparra et al., 2005b; Juhasz et al., 2006). Laparra et 32 al. (2005b) determined that while cooking rice (they tested several types, but did not specify 33 them) in deionized water caused no change in arsenic content compared to the raw form, cooking 34 in water contaminated with 0.5 ug/mL of Asv increased the inorganic arsenic content 5- to 17- 35 fold over the raw rice. Laparra et al. subjected the rice samples (10 grams) to an in vitro 36 simulated digestion process. They measured levels of soluble arsenic to determine 7 DRAFT—DO NOT CITE OR QUOTE ------- 1 bioaccessibility. The results demonstrated that large amounts of the arsenic (i.e., 63%-99%), 2 mainly in the pentavalent form, were bioaccessible for intestinal absorption. Ackerman et al. 3 (2005) also found 89%-105% bioaccessible arsenic in different samples of white and brown rice 4 cooked in water containing Asv. 5 Juhasz et al. (2006) examined the bioavailability of arsenic from rice (mainly white rice 6 samples) using an in vivo swine assay. Quest rice was grown in arsenic-contaminated water and 7 cooked in arsenic-free water. This caused the rice to contain arsenic, mainly in the form of 8 DMA. Administration of the cooked rice to swine demonstrated a bioavailability similar to that 9 observed after a single oral administration of DMA in water (i.e., 33.3%). Basmati white rice 10 cooked in water contaminated with 1,000 ppb of Asv, which contained entirely inorganic arsenic 11 as a result of the arsenate in the cooking water, had a bioavailability of 89.4%. 12 Although there have been no studies performed on the rate of inorganic arsenic 13 absorption through intact human skin, systemic toxicity due to high dermal occupational 14 exposure to aqueous inorganic arsenic solutions indicates that the skin may be a significant 15 exposure route (Hostynek et al., 1993). The systemic absorption via the skin from less 16 concentrated solutions, however, appears to be low (NRC, 1999). An in vivo study by Wester et 17 al. (1993) demonstrated that 2% to 6% of radiolabeled arsenate (as a water solution) was 18 absorbed by rhesus monkey skin over a 24-hour period. Results demonstrated that the lower 19 dose (0.000024 ug/cm2) was absorbed at a greater rate (6%) than the higher arsenic exposure 20 (2.1 ug/cm2; 2%), but the difference did not reach statistical significance. Wester et al. (2004) 21 performed another in vivo dermal absorption study using female rhesus monkeys. Using the 22 levels excreted in the urine and the applied dose, they calculated that 0.6% to 4.4% was absorbed 23 in the three monkeys tested, which was similar to their previous results. In vitro results on 24 human skin (from donors) demonstrated a 24-hour absorption of 1.9% (Wester et al., 1993). 25 Mouse dorsal skin was demonstrated to absorb 30% to 60% of applied arsenic (Rahman et al., 26 1994) using similar in vitro testing, with 60% to 90% of the absorbed arsenic being retained in 27 the skin. NRC (1999) suggests this indicates that inorganic arsenic binds significantly to skin 28 and hair. Lowney et al. (2007) found that dermal absorption of arsenic from soils was negligible 29 in an in vivo study in rhesus monkeys. 30 Harrington et al. (1978) compared arsenic metabolite levels in the urine from a group of 31 people in Fairbanks, Alaska, who had arsenic-contaminated water (345 ppb) in their home, but 32 drank only bottled water, with the levels measured in a group of people who drank home water 33 containing less than 50 ppb. The results demonstrated that the group with high arsenic in their 34 water had close to the same average concentration of total arsenic metabolites in their urine (i.e., 35 43 ug/L) as the group who drank home water with less than 50 ppb arsenic (i.e., 38 ug/L in 36 urine), indicating possible dermal absorption via bathing or other exposure sources. Levels of DRAFT—DO NOT CITE OR QUOTE ------- 1 arsenic in the bottled water, however, were not measured. Possible exposure through using 2 contaminated water for cooking also was not examined. 3.2. DISTRIBUTION 3 The retention and distribution patterns of arsenic species are strongly dependent on their 4 chemical properties. While both As111 and Asv bind to sulfhydryl groups, As111 has approximately 5 a 5- to 10-fold greater affinity for sulfhydryl groups than Asv (Jacobson-Kram and Montalbano, 6 1985). Cellular uptake rates and resulting tissue concentrations are substantially lower for the 7 pentavalent than for the trivalent forms of arsenic. DMA (an important metabolite of inorganic 8 arsenic) appears to be more readily excreted than MMA (NRC, 2001). Liu et al. (2002) found 9 arsenite to be transported into cells by aquaglycoporins (AQP7 and AQP9), whose usual 10 substrates are water and glycerol. Liu et al. (2006a) also detected transport of 11 monomethylarsonous acid (MMA111) by AQP9. MMA111 was transported at a rate nearly 3 times 12 faster than As111. A hydrophobic residue at position 64 was required for the transport of both 13 species, suggesting that both species are transported by AQP9 using the same translocation 14 pathway. Asv, however, has been suggested to be transported by the phosphate transporter 15 (Huang and Lee, 1996). Retention of arsenic can vary not only with its form, but also with tissue 16 (Thomas et al., 2001). Other factors that affect the retention and distribution of arsenic include 17 the chemical species, dose level, methylation capacity, valence state, and route of administration. 18 3.2.1. Transport in Blood 19 20 Once arsenic is absorbed, it is transported in the blood throughout the body. In the blood, 21 inorganic arsenic species are generally bound to sulfhydryl groups of proteins and low- 22 molecular-weight compounds such as glutathione (GSH) and cysteine (NRC, 1999). Binding of 23 As111 to GSH has been demonstrated by several investigators (Anundi et al., 1982; Scott et al., 24 1993; Delnomdedieu et al., 1994a,b). Because of the different binding and transport 25 characteristics of various arsenic compounds, the persistence in the blood varies across species. 26 Inorganic arsenic elimination in humans has been observed to be triphasic, with first-order half- 27 lives for elimination of 1 hour, 30 hours, and 200 hours (Mealey et al., 1959, used As111; Pomroy 28 et al., 1980, used Asv). A single intravenous (iv) dose of 5.8 ug As/kg body weight (in the form 29 of 73 Asv) administered to two male chimpanzees had a half-life plasma elimination rate of 30 1.2 hours and a half-life elimination rate from red blood cells (RBCs) of about 5 hours (Vahter et 31 al., 1995a). 32 Rats retain arsenic in the blood considerably longer than other species because 33 dimethylarsenous acid (DMA111) and DMAV accumulate in RBCs, apparently bound to 34 hemoglobin (Odanaka et al., 1980; Lerman and Clarkson, 1983; Vahter, 1983; Vahter et al., 9 DRAFT—DO NOT CITE OR QUOTE ------- 1 1984). Naranmandura et al. (2007) found that 75% of an oral dose of arsenite accumulated in rat 2 RBCs mainly in the form of DMA111; however, less than 0.8% of the same dose to hamsters was 3 found in their RBCs. Rats maintained this level in their RBCs for at least 7 days whereas the 4 treated hamsters had levels equivalent to those in controls by 3 days after the administered dose. 5 Stevens et al. (1977) calculated an elimination half-life for inorganic arsenic of 90 days in rat 6 whole blood after a single oral dose of 200 mg/kg. Lanz et al. (1950) also reported a high 7 retention of arsenic in the blood of cats, although less than in the rat. However, they did not 8 determine if the retained arsenic was in the form of DMA. 9 The relative concentration of arsenic in human plasma and RBCs apparently differs 10 depending on exposure levels and the health status of the exposed individuals. Heydorn (1970) 11 reported that healthy people in Denmark with low arsenic exposures had similar arsenic 12 concentrations in their plasma and RBCs (2.4 ug/L and 2.7 ug/L, respectively; the RBC:plasma 13 ratio was 1.1). However, normal healthy Taiwanese exposed to arsenic-contaminated water had 14 plasma levels of 15.4 ug/L and RBCs levels of 32.7 ug/L (RBC:plasma ratio 2.1). Blackfoot 15 disease (BFD) patients and their unaffected family members had 38.1 ug/L and 93 ug/L of 16 arsenic species in their plasma and RBCs, respectively (RBC:plasma ratio 2.4). These results 17 indicate a different distribution between the RBCs and the plasma depending on exposure levels. 18 However, examining the BFD patients and their families, who presumably have the same 19 exposure levels, demonstrates a different distribution, possibly due to disease state. BFD 20 patients had a ratio of 3.3 (106 ug/L in RBCs and 32.3 ug/L in plasma) compared to 1.8 (81 ug/L 21 in RBCs and 45.2 ug/L in plasma) in family members without BFD. This indicates that 22 accumulation of arsenic in the RBCs is greater as exposure increases and possibly even greater 23 when health is compromised. The ratio between plasma and RBC arsenic concentrations may 24 also depend on the exposure form of arsenic (NRC, 1999). 25 3.2.2. Tissue Distribution 26 27 Once arsenic compounds enter the blood, they are transported and taken up by other 28 tissues and organs, with a large proportion of ingested material being subject to "first pass" 29 processing through the liver. Uptake varies with arsenic species, dose, and organ. The observed 30 uptake of inorganic arsenic (mainly As111) in the skin, hair, oral mucosa, and esophagus is most 31 likely due to the binding of inorganic arsenic species with sulfhydryl groups of keratin in these 32 organs. In studies using rabbits and mice, where the transfer of methyl groups from 33 S-adenosylmethionine (SAM; a proposed major reaction during arsenic metabolism; see Section 34 3.3) was chemically inhibited, the concentration of arsenic in most tissues (especially the skin) 35 was found to be increased (Marafante and Vahter, 1984). The important role of chemical 10 DRAFT—DO NOT CITE OR QUOTE ------- 1 binding of arsenic species also is supported by the observed tissue distribution in the marmoset 2 monkey, which does not methylate inorganic arsenic (Vahter et al., 1982). 3 Human subjects also have demonstrated high concentrations of arsenic in tissues 4 containing a high content of cysteine-containing proteins, including the hair, nails, skin, and 5 lungs. Total arsenic concentrations in these tissues of human subjects exposed to background 6 levels of arsenic ranged from 0.01 to 1.0 mg/kg of dry weight (Liebscher and Smith, 1968; Cross 7 et al., 1979). Benign and malignant skin lesions from 14 patients, with a minimum of 4 years of 8 exposure to inorganic arsenical medication, had higher arsenic levels (0.8 to 8.9 ppm) than six 9 subjects with no history of arsenic intake (0.4 to 1.0 ppm; Scott, 1958). In West Bengal, India, 10 where the average arsenic concentration in the drinking water ranges from 193 to 737 ppb, 11 arsenic concentrations in the skin, hair, and nails were 1.6-5.5, 3.6-9.6, and 6.1-22.9 mg/kg dry 12 weight, respectively (Das et al., 1995). Mandal et al. (2004) measured different arsenic species 13 in the hair and fingernails of 41 subjects in West Bengal, India, who were drinking arsenic- 14 contaminated water and in blood from 25 individuals who had stopped drinking contaminated 15 water 2 years earlier. Results were: fingernail contained As111 (62.4%), Asv (20.2%), MMAV 16 (5.7%), DMA111 (8.9%), and DMAV (2.8%); hair contained As111 (58.9%), Asv (34.8%), MMAV 17 (2.9%), and DMAV (3.4%); RBCs contained arsenobetaine (22.5%) and DMAV (77.5%); and 18 blood plasma contained arsenobetaine (16.7%), As111 (21.1%), MMAV (27.1%), and DMAV 19 (35.1). However, the amount of arsenic in these tissues resulting from other exposure pathways 20 (e.g., dermal exposure) was not determined. 21 The longest retention of inorganic arsenic in mammalian tissues during experimental 22 studies has been observed in the skin (Marafante and Vahter, 1984), hair, squamous epithelium 23 of the upper GI tract (oral cavity, tongue, esophagus, and stomach wall), epididymis, thyroid, 24 skeleton, and the lens of the eye (Lindgren et al., 1982). Although the study authors measured 25 radioactive arsenic (74As) in the various tissues, they did not differentiate between the different 26 species of arsenic and could not determine if accumulation was due to the originally 27 administered compound or metabolites. Arsenic levels in all these tissues, with the exception of 28 the skeleton, were greater in mice administered As111 than in mice administered Asv. This could 29 indicate that As111 is taken up more efficiently than Asv and that less was found in the tissues of 30 AsV-treated mice due to the initial reduction to As111. The calcified areas of the skeleton in mice 31 administered As accumulated and retained more arsenic than mice administered As , most 32 likely due to the similarities between Asv and phosphate, causing a substitution of phosphate by 33 Asv in the apatite crystals in bone. Marmoset monkeys were found not to accumulate arsenic in 34 the ocular lens or the thyroid (Vahter et al., 1982); however, intravenous administration of 74As- 35 labelled DMA to mice resulted in accumulation of DMA in the ocular lens and the thyroid. 36 Marmoset monkeys do not methylate arsenic and DMA was found to accumulate in the ocular 37 lens and thyroid; this suggests that only the methylated species are retained in these organs. 11 DRAFT—DO NOT CITE OR QUOTE ------- 1 Mouse tissues with the largest retention of DMA were the lens of the eyes, thyroid, lungs, and 2 intestinal mucosa (Vahter et al., 1984). Methylated arsenic species (DMA), in general, have a 3 shorter tissue retention time in mice than rats (i.e., more than 99% of the administered dose was 4 eliminated in mice within 3 days as compared to 50% in rats due to accumulation in blood) 5 (Vahter etal., 1984). 6 Hughes et al. (2003) estimated that a steady-state, whole-body arsenic balance was 7 established after nine repeated oral daily doses of 0.5 mg As/kg as radioactive Asv in adult 8 female B6C3F1 mice. Twenty-four hours after the last dose, the whole-body burden of arsenic 9 was about twice that observed after a single dose. The rate of elimination was slower following 10 repeated doses. Accumulation of radioactivity was highest in the bladder, kidney, and skin, 11 while the loss of radioactivity was greatest from the lungs and slowest from the skin. Atomic 12 absorption spectrometry was used to characterize the organ distribution of arsenic species. 13 MMA was detected in all tissues except the bladder. DMA was found at the highest levels in the 14 bladder and lung after a single oral exposure, with increases after repeated exposures. Inorganic 15 arsenic was predominantly found in the kidney. After a single oral exposure of Asv (0.5 mg 16 As/kg), DMA was the predominant form of arsenic in the liver, but after nine repeat exposures, 17 the proportion of DMA decreased while the proportion of inorganic arsenic increased (this could 18 indicate metabolic saturation or GSH depletion; see Section 3.3 for more details). A 19 trimethylated form of arsenic also was detected in the liver. 20 Kenyon et al. (2005a) examined the time course of tissue distribution of different arsenic 21 species after a single oral dose of 0, 10, or 100 umole As/kg as sodium arsenate to adult female 22 B6C3F1 mice. The concentrations of all forms of arsenic were lower in the blood than in other 23 organs across all doses and time points. The concentration of inorganic arsenic measured in the 24 liver was similar to that measured in the kidney at both dose levels, with peak concentrations 25 observed 1 hour after dosing. For the first 1 to 2 hours, inorganic arsenic was the predominant 26 form in both the liver and kidney, regardless of dose. At the later times, DMA became the 27 predominant form. Kidney measurements 1 hour after dosing demonstrated that MMA levels 28 were 3 to 4 times higher than in other tissues. DMA concentrations in the kidney reached their 29 peak 2 hours after dosing. DMA was the predominant form measured in the lungs at all time 30 points following exposure to 10 umole As/kg as Asv. DMA concentrations in the lung were 31 greater than or equal to those of the other tissues beginning at four hours. The study did not 32 distinguish the different valence states of the MMA or DMA compounds. 33 In a follow-up study by Kenyon et al. (2008), adult female C57B1/6 mice were 34 administered 0, 0.5, 2, 10, or 50 ppm of arsenic as sodium arsenate in the drinking water for 35 12 weeks. The average daily intakes were estimated to be 0, 0.083, 0.35, 1.89, and 7.02 mg 36 As/kg/day, respectively. After 12 weeks of exposure, the tissue distributions were as follows: 37 kidney > lung > urinary bladder > skin > blood > liver. In the kidney, MMA was the 12 DRAFT—DO NOT CITE OR QUOTE ------- 1 predominant form measured, while DMA was more prominent in the lungs and blood. The skin 2 and urinary bladder had nearly equal levels of both inorganic arsenic and DMA and the liver had 3 equal proportions of all three species. 4 Naranmandura et al. (2007) characterized the tissue distribution in rats and hamsters 5 administered a single oral dose of As111 (5.0 mg As/kg body weight, or BW). In rats, the highest 6 concentrations were found in RBCs. Because hamsters did not accumulate arsenic species in 7 their RBCs, they exhibited a more uniform tissue distribution. While the quantity of arsenic in 8 the liver and kidneys of the hamster were significantly greater than those observed in the rat, 9 arsenic accumulated more and was retained longer in the kidneys than the liver in both species. 10 The hamster had greater levels of MMA111 bound to protein in the kidney than rats. 11 As111 and Asv, as well as methylated metabolites, cross the placenta at all stages of 12 gestation in mice, marmoset monkeys, and hamsters (Hanion and Perm, 1977; Lindgren et al., 13 1984; Hood et al., 1987; Jin et al., 2006a), with tissue distribution of arsenic similar between the 14 mother and the fetus in late gestation. Jin et al. (2006a) found increased levels of inorganic 15 arsenic and DMA in the livers and brains of newborn mice from dams administered either As111 16 or Asv in their drinking water throughout gestation and lactation. The levels of total arsenic in 17 the mothers' livers increased in a dose-dependent manner and were greater than those observed 18 in the mothers' brains or in the newborns' brains or livers. The levels of total arsenic in the 19 livers and brains of newborn mice, however, were greater than those observed in the mothers' 20 brains, suggesting easier passage through the placenta than through a mature blood-brain barrier. 21 Because the levels of inorganic arsenic in the newborn livers and brains were nearly identical, it 22 appears that there was no difficulty in passing through an immature blood-brain barrier. In 23 addition, the nearly 2:1 ratio of DMA in the brains compared to the livers of newborns indicates 24 either a preferential distribution of DMA in the newborns' brains or an increased distribution of 25 inorganic arsenic to the brain that is subsequently metabolized. The marmoset monkey (known 26 to not methylate arsenic) displayed somewhat less placental transfer after administration of As111 27 than was seen in mice (Lindgren et al., 1984). 28 The arsenic concentration in the cord blood (11 ug/L) was similar to that observed in 29 maternal blood (an average of 9 ug/L) in pregnant women living in a village in northwestern 30 Argentina, where the arsenic concentration in the drinking water was approximately 200 ppb 31 (Concha et al., 1998a). Hall et al. (2007) also found a strong association between maternal (11.9 32 ug/L) and cord blood levels (15.7 ug/L) in Matlab, Bangladesh (arsenic exposure ranged from 33 0.1 to 661 ppb in drinking water). They also measured arsenic metabolite levels and found that 34 the association also was observed for the metabolites MMA and DMA. Elevated arsenic 35 concentrations also were noted in pregnant women living in cities with low dust fall (i.e., low 36 arsenic inhalation exposures), where an average of 3 ug/L was measured in the maternal blood 37 and 2 ug/L in cord blood (Kagey et al., 1977). Women living near smelters also have been 13 DRAFT—DO NOT CITE OR QUOTE ------- 1 observed to have an increased concentration of placental arsenic (Tabacova et al., 1994). 2 Although the human fetus is exposed to arsenic, it may be more in the form of DMA (at least in 3 late gestation) because 90% or more of the arsenic in the urine and plasma of newborns and 4 mothers (at time of delivery) was DMA. 3.2.3. Cellular Uptake, Distribution, and Transport 5 Cellular uptake of inorganic arsenic compounds also depends on oxidation state, with 6 As111 generally being taken up at a much greater rate than arsenate (Cohen et al., 2006). In 7 Chinese hamster ovary (CHO) cells, the rate of uptake was DMA111 > MMA111 > As111 (Dopp et 8 al., 2004), with the pentavalent forms being taken up much more slowly than the trivalent forms. 9 Delnomdedieu et al. (1995) demonstrated that As111 is taken up more readily than Asv, MMAV, 10 or DMAV by RBCs in rabbits. Drobna et al. (2005) found that MMA111 and DMA111 were taken 11 up by modified UROtsa cells expressing arsenic methyltransferase (this is a human urothelial 12 cell line that normally does not methylate inorganic arsenic) at an order of magnitude faster than 13 As111. Because arsenate uptake is inhibited in a dose-dependent manner by phosphate (Huang 14 and Lee, 1996), it has been suggested that a common transport system is responsible for the 15 cellular uptake for both compounds. As111 uptake, however, is not affected by phosphate; 16 therefore, Huang and Lee (1996) suggested that cellular uptake of As111 occurs through simple 17 diffusion. Liu et al. (2002, 2006a), however, suggested that transport of As111 and MMA111 across 18 the cellular membrane may be mediated by AQP7 and AQP9 with MMA111 transported at a 19 higher rate. Lu et al. (2006) found that inorganic arsenic (both pentavalent and trivalent 20 oxidation states) can be transported by organic anion transporting polypeptide-C (OATP-C; 21 which was transfected into cells of a human embryonic kidney cell line), but not MMAV or 22 DMAV. In a cell line resistant to arsenic (R15), Lee et al. (2006a) found little AQP7 or AQP9 23 messenger RNA (mRNA) and only half the AQP3 mRNA expression compared to the parental 24 cell line (CL3, a human lung adenocarcinoma cell line). Suppressing the AQP3 expression in 25 CL3 cells caused less arsenic to accumulate in these cells. Over-expression of AQP3 in a 293 26 cell line (a human embryonic kidney cell line) resulted in an increase in arsenic accumulation in 27 the cells. Hexose permease transporters (HXT) also have been suggested as another influx 28 pathway for As111 (Thomas, 2007). 29 Shiobara et al. (2001) demonstrated that the uptake of DMA in RBCs was dependent on 30 not only the chemical form (or oxidation state), but animal species. DMA111 and DMAV were 31 incubated with rat, hamster, mouse, and human RBCs. DMAV was only minimally absorbed by 32 RBCs, and the cellular uptake was very slow in all animal species tested. DMA111, on the other 33 hand, was efficiently taken up by the RBCs in the following order: rats > hamsters > humans. 34 Mouse RBCs were less efficient at the uptake of DMA111 than any of the other species. Rat 35 RBCs retained the DMA111 throughout the 4 hours of the experiment, but hamster RBCs were 36 found to excrete the arsenic absorbed as DMA111 in the form of DMAV. Human RBCs also 14 DRAFT—DO NOT CITE OR QUOTE ------- 1 excreted DMA111 as DMAV, though the rate of uptake of DMA111 and efflux of DMAV was much 2 slower than in hamster RBCs. 3 Cellular excretion of arsenic species also depends on oxidation state and the degree of 4 methylation. Leslie et al. (2004), using membrane vesicles from a multi-drug resistant human 5 lung cancer cell line (H69AR), found that a multi-drug resistance protein (MRP) called MRP1 6 transports As111 in the presence of GSH but did not transport Asv under any conditions. This 7 suggests that Asv must be reduced to As111 before being excreted from the cell. Further, the 8 MRP1 transport was more efficient with arsenic triglutathione (ATG) as the substrate. This 9 finding, along with the observation that As111 transport is more efficient at neutral or low pH 10 where ATG is more readily formed and more stable, suggests that ATG is formed prior to 11 transport. Leslie et al. (2004) also suggest that the formation of the conjugate is catalyzed by the 12 glutathione-S-transferase Pl-1 (GSTP1-1) enzyme. MRP2 may also be involved in the efflux of 13 arsenic species from cells (Thomas, 2007). MRP2 expression was found to be five times higher 14 in arsenic-resistant (R15) cells compared to the parent cell line (CL3). However, expression 15 levels of MRP1 and MRP3 were similar to levels in parent cells (Lee et al., 2006a). Suppressing 16 the multi-drug resistant transporters reduced the efflux of arsenic from Rl 5 cells. 17 In a study of rabbits and mice exposed to radio-labeled arsenic (as As111), the majority of 18 the arsenic was found in the nuclear and soluble fractions of liver, kidney, and lung cells 19 (Marafante et al., 1981; Marafante and Vahter, 1984). The marmoset monkey had a different 20 intracellular distribution, with approximately 50% of the arsenic dose found in the microsomal 21 fraction in the liver (Vahter et al., 1982; Vahter and Marafante, 1985). Chemical inhibition of 22 arsenic methylation in rabbits did not alter the intracellular distribution of arsenic (Marafante and 23 Vahter, 1984; Marafante et al., 1985). 24 Increases in tissue arsenic concentration (especially in the liver) have been found to be 25 associated with increased arsenic concentrations in the microsomal fraction of the liver in rabbits 26 fed diets containing low concentrations of methionine, choline, or proteins, which leads to 27 decreased arsenic methylation (Vahter and Marafante, 1987). The levels of arsenic in the 28 microsomal fraction of the liver in these rabbits were similar to those observed in the marmoset 29 monkey (Vahter et al., 1982), indicating that nutritional factors may play a role in determining 30 the subcellular distribution of arsenic. 3.3. METABOLISM 31 After entering the body, Asv can be reduced to As111, which can then proceed through a 32 series of methylation and conjugation reactions, some of which involve re-oxidation of arsenic to 33 Asv. The traditional metabolic pathways proposed for arsenic are shown in Figure 3-1. In this 34 metabolic scheme, less toxic species (i.e., Asv, MMAV, and DMAV) can be converted to more 35 toxic species (i.e., As111, MMAm, and DMA111). The trivalent species have been found to be more 15 DRAFT—DO NOT CITE OR QUOTE ------- 1 cytotoxic, genotoxic, and more potent inhibitors of enzyme activity (Thomas et al., 2001). While 2 the final metabolite in humans is predominantly DMAV, as this is the form most highly excreted, 3 some animal species further metabolize DMAV through DMA111 to trimethylarsine oxide 4 (TMAO). Arsenatev \ <-- Arsenite111 Arsenate reductase, Glutathione S-transferase-(0, Glyceraldehyde phosphate dehydrogenase (GAPDH)?; Nonenzymatic pathway SAM Monomethylarsonic acid (MMAV) X •* Glutathione S transferase-co Monomethylarsonous acid (MMA111) VSAM ^___, SAH MethyltransferaseCASSMT) Dimethylarsink add (DMAV) \ J Dimethylarsenous acid (DMA111) •SAM Trimethyl arsine oxide (TMAOV) \ - Trimethyl arsine (TMA111) Source: Sams et al. (2007). Figure 3-1. Traditional metabolic pathway for inorganic arsenic in humans. 5 Hayakawa et al. (2005) suggested a possible alternate metabolic pathway for inorganic 6 arsenic (Figure 3-2). As in the previously described model, the first step involves reduction of 7 Asv to As111. A major difference, however, is that Hayakawa et al. (2005) suggest that arsenic- 8 glutathione complexes are important intermediates in the metabolism of arsenic and are the 9 primary substrates for arsenic methyltransferases. The proposed model was based on the 10 observation that more DMAV is produced from As111 than from MMAV. This should not be the 11 case if the reactions depicted in Figure 3-1 are the primary arsenic metabolic pathways. Their 12 data suggest that arsenite, in the presence of GSH, non-enzymatically reacts to form ATG. In 13 support of this mechanism, they observed a dose-dependent increase in concentration of ATG 14 with increasing doses of GSH, up to 4 mM. Monomethyl and dimethyl arsenic species were 15 generated by the transfer of a methyl group from SAM in the presence of human recombinant 16 DRAFT—DO NOT CITE OR QUOTE ------- 1 arsenic (+3 oxidation state) methyltransferase (AS3MT), and only occurred when ATG or 2 monomethylarsonic diglutathione (MADG) was present. At concentrations of glutathione of 3 2.0 mM or greater, there was a dose-dependent increase in DMAV levels, accompanied by a 4 dose-dependent decrease in Asv. o ii OH - Asv - OH I OH Arsenate O II OH - Asv - OH I CH3 MMA(V) OH-As"1-OH I OH Arsenite OH-As"1-OH I CH3 MMA(III) JGSH JGSH GS-As'"-SG I SG AsSMT SAM ATG (Arsenic triglutathione) GS-As1" I CH, SG AsSMT SAM MADG (Monomethylarsinic diglutathione) OH-Asv-CH3 I CH3 DMA(V) Hayakawa et al. 2005 OH-AsMI-CH3 I CH3 DMA(III) GS-Asm-CH3 I CH3 DMAG (Dimethylarsinic Glutathione) Arsenic 3 methyl transferase (As3MT); SAM -S-adenosyl methionine; GSH -Glutathione Source: Hayakawa et al. (2005). Figure 3-2. Alternative metabolic pathway for inorganic arsenic in humans proposed by Hayakawa et al. (2005). 17 DRAFT—DO NOT CITE OR QUOTE ------- 1 In summary, the proposed metabolic model of Hayakawa et al. (2005) suggests that Asv 2 is first reduced to As111, which then reacts (non-enzymatically) with GSH (producing ATG). In 3 the presence of AS3MT (specified as cyt!9 in the Hayakawa article),2 ATG is methylated to 4 MADG if the GSH concentration is sufficient, which then comes to equilibrium with MMA111 5 (GSH concentrations lower than 1 mM caused MADG to be unstable in solution and was readily 6 hydrolyzed and oxidized to MMAV). While some of the MMA111 is oxidized to MMAV, some of 7 the MADG is methylated by AS3MT to dimethylarsinic glutathione (DMAG), which, like 8 MADG, is in equilibrium with its trivalent form and can be oxidized to its pentavalent form. 9 This more recently proposed pathway leads to higher proportions of less toxic final species than 10 the original proposed metabolic pathway (Figure 3-1). 11 Results reported by Hughes et al. (2005) may provide support for the Hayakawa et al. 12 (2005) revised pathway. B6C3F1 mice administered MMAV per os demonstrated its rapid 13 absorption, distribution, and excretion, with 80% of the dose eliminated within 8 hours. Very 14 little of the absorbed dose, however, was methylated to DMA and/or TMAO. Less than 10% of 15 the dose excreted in urine and 25% or less of the dose measured in the tissues were in the form 16 of DMA. In contrast, in MMA111-treated mice, more than 90% of the excreted dose and more 17 than 75% of the arsenic measured in the tissues was identified as DMA. This discrepancy 18 between the two forms of MMA is not expected if the generally accepted metabolic pathway 19 (Figure 3-1) is followed. However, if MMA111 is the form methylated to DMA while MMAV is 20 an end product, as is suggested by Hayakawa et al. (2005), then it would be expected that a 21 greater proportion of MMA111 would be methylated to DMA than MMAV. There are, however, 22 factors that may limit the in vivo methylation of MMAV that are unrelated to the metabolic 23 pathway proposed by Hayakawa et al. (2005). First, MMAV does not appear to be taken up well 24 by the liver (Hughes et al., 2005), a major site of inorganic arsenic metabolism (Thomas et al., 25 2001). In fact, pentavalent species of arsenic are not taken up by cells as readily as trivalent 26 arsenicals (Dopp et al., 2004). In addition, in the generally accepted metabolic pathway (Figure 27 3-1), MMAV needs to be reduced to MMA111 in order to be methylated. Therefore, if very little is 28 taken up into cells, very little can be methylated. 29 Aposhian and Aposhian (2006) suggest that it is too early to accept AS3MT as the 30 primary methyltransferase responsible for arsenic methylation in humans because it has only 31 been observed in experiments involving deoxyribonucleic acid (DNA) recombinant technology 32 and because there is no indication that the enzyme is expressed in human liver. Although 33 AS3MT has been detected in human liver cell lines (Zakharyan et al., 1999), it has not been 2 Arsenic (+3 oxidative state) methyltransferase (AS3MT) has been referred to by many investigators as cyt!9 in their references. According to Thomas et al. (2007), the Human Genome Nomenclature Committee (http://www.gene.ucl.ac.uk/cgi-bin/nomenclature/searchgenes.pl) recommends that this protein be systematically named AS3MT. In this document, references to cyt!9 it has been changed to AS3MT to avoid confusion and for uniform consistency. 18 DRAFT—DO NOT CITE OR QUOTE ------- 1 isolated from surgically removed liver tissue. Thomas et al. (2007) also states the evidence 2 supports the conclusion that arsenic methylation catalyzed by AS3MT is not strictly dependent 3 on the presence of GSH, which would suggest that other pathways may be involved in addition 4 to those included in Hayakawa et al.'s (2005) model. GSH depletion would likely occur at high 5 arsenic exposures under Hayakawa et al.'s proposed pathway. Therefore, it is possible that both 6 pathways work in conjunction, or one is predominant over the other depending on the 7 concentration of arsenic. Hayakawa et al. (2005) found that levels of MMAV were not 8 dependent on GSH level (from 2 to 5 mM), suggesting that this indicated possible further 9 methylation to DMAV. Since this is not part of the proposed Hayakawa et al. (2005) pathway, at 10 least some of the MMAV may be methylated through the classic pathway. 3.3.1. Reduction 11 A substantial fraction of absorbed Asv is rapidly reduced to As111 in most species studied; 12 in mice, rabbits, and marmoset monkeys, the reduction apparently occurs mainly in the blood 13 (Vahter and Envall, 1983; Vahter and Marafante, 1985; Marafante et al., 1985). Reduction also 14 may occur in the stomach or intestines prior to absorption, but quantitative experimental data are 15 not available to determine the importance of this GI reduction. In addition to the reduction of 16 inorganic Asv, as shown in Figure 3-1, methylated Asv species also may be reduced, apparently 17 by different enzymes. 18 GSH may play a role in the reduction of Asv, but apparently is not the only cofactor, as 19 cysteine and dithiothreitol (DTT) also have been found to reduce Asv to As111 in vitro (Zakharyan 20 et al., 1995; NRC, 1999; Nemeti and Gregus, 2002). Inorganic phosphate inhibits the formation 21 of As111 from Asv in intact RBCs (Nemeti and Gregus, 2004), probably by competing with the 22 phosphate transporter for the uptake into cells. 23 Arsenate reductase enzymes have been detected in the human liver (Radabaugh and 24 Aposhian, 2000). At least one of these enzymes has been characterized as a purine nucleoside 25 phosphorylase (PNP) (Gregus and Nemeti, 2002; Radabaugh et al., 2002). This enzyme requires 26 a thiol and a heat-stable cofactor for activation. According to Radabaugh et al. (2002), 27 dihydrolipoic acid (DHLP) is the most active naturally occurring thiol in mammalian systems 28 and appears to be required for the enzymatic reduction of Asv to As111. PNP, however, did not 29 catalyze the reduction of MMAV to MMA111. An MMAV reductase has been detected in rabbit 30 liver (Zakharyan and Aposhian, 1999), hamster tissues (Sampayo-Reyes et al., 2000), and human 31 liver (Zakharyan et al., 2001). In humans, this reductase is human glutathione-S-transferase co 32 (hGST-Ol), which is a member of the glutathione-S-transferase (GST) superfamily (Aposhian 33 and Aposhian, 2006). 34 Although PNP has been determined to reduce Asv to As111, Nemeti et al. (2003) observed 35 this reduction only in vitro. PNP did not appear to be a major player in the reduction of Asv to 36 As111 in either human erythrocytes or in rats in vivo. Nemeti and Gregus (2004, 2005) further 19 DRAFT—DO NOT CITE OR QUOTE ------- 1 demonstrated that human erythrocytes exhibit a PNP-independent Asv-reducing pathway that 2 requires GSH, nicotinamide adenine dinucleotide (NAD), and a substrate for either one or both 3 of the following enzymes: glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or 4 phosphoglycerate kinase (PGK). This mechanism of reduction also was demonstrated in rat liver 5 cytosol (Nemeti and Gregus, 2005). In addition, another unidentified enzyme in the liver cytosol 6 had the capacity to reduce Asv. A further study (Gregus and Nemeti, 2005) demonstrated that 7 GAPDH exhibited Asv reductase activity, but that PGK served as an auxiliary enzyme when 8 3-phosphoglycerate was the glycolic substrate. 9 The reduction of pentavalent arsenicals also has been observed to be catalyzed by 10 AS3MT (Waters et al., 2004a). According to Waters et al. (2004b), AS3MT may possess both 11 As111 methyltransferase and Asv reductase activities. In the presence of an exogenous or 12 physiological reductant, AS3MT was found to catalyze the entire sequence converting arsenite to 13 all of its methylated metabolites through both methylation and reduction steps (Figure 3-1). 14 Thomas et al. (2007) also suggest that thioredoxin (Trx, isolated from E. coli) is necessary, 15 possibly reducing some critical cysteine residue in AS3MT as a step in the methyltransferase 16 reaction. Cohen et al. (2006) suggest that Trx, thioredoxin reductase (TrxR), and nicotinamide 17 adenine dinucleotide phosphate-oxidase (NADPH) are the primary reducing agents involved in 18 the conversion of MMAV to DMAV, but they are orders of magnitude less effective than the 19 arsenic methyltransferase isolated from rabbit liver (i.e., AS3MT). Zakharyan and Aposhian 20 (1999) found that MMAv-reductase was the rate-limiting enzyme in arsenic biotransformation in 21 rabbit livers. Jin et al. (2006a) also suggest that Asv reduction is possibly a rate-limiting step in 22 arsenic metabolism at low concentrations. At higher concentrations, saturation or methylation 23 inhibition may cause other reactions to become rate-limiting. 3.3.2. Arsenic Methylation 24 Methylation is an important factor affecting arsenic tissue distribution and excretion. 25 Humans and most experimental animal models methylate inorganic arsenic to MMA and DMA, 26 with the amounts differing across species, as determined by analysis of urinary metabolites. The 27 methylated metabolites in and of themselves have historically been considered less acutely toxic, 28 less reactive with tissue constituents, less cytotoxic, and more readily excreted in the urine than 29 inorganic arsenic (Vahter and Marafante, 1983; Vahter et al., 1984; Yamauchi and Yamamura, 30 1984; Marafante et al., 1987; Moore et al., 1997a; Rasmussen and Menzel, 1997; Hughes and 31 Kenyon, 1998; Sakurai et al., 1998). The trivalent species MMA111 and DMA111, however, have 32 been demonstrated to be more cytotoxic in a human liver cell line called Chang cells (Petrick et 33 al., 2000, 2001), CHO (Dopp et al., 2004), and cultured primary rat hepatocytes (Styblo et al., 34 1999a, 2000) than As111, Asv, MMAV, or DMAV. 35 Although the kinetics of arsenic methylation in vivo are not fully understood, it is 36 believed the liver may be the primary site of arsenic methylation. However, the testes, kidney, 20 DRAFT—DO NOT CITE OR QUOTE ------- 1 and lung also have been observed to have a high methylating capacity (Cohen et al., 2006). 2 Marafante et al. (1985) found that DMA appeared in the liver prior to any other tissue in rabbits 3 exposed to inorganic As. It also has been demonstrated oral administration of inorganic arsenic 4 favors methylation more than either subcutaneous or intravenous administration (Charbonneau et 5 al., 1979; Vahter, 1981; Buchet et al., 1984), presumably because the arsenic will pass through 6 the liver first after oral administration. However, liver disease (i.e., alcoholic, post-necrotic or 7 biliary cirrhosis, chronic hepatitis, hemochromatosis, and steatosis) can be associated with 8 increased ratios of DMA to MMA in the urine following a single injection of sodium arsenite 9 (Buchet et al., 1984; Geubel et al., 1988). This appears to indicate that efficient methylation of 10 arsenic continues in the presence of liver damage, possibly indicating that a different organ is 11 responsible for methylation under these circumstances. In addition, the site of methylation may 12 depend on the rate of reduction of Asv to As111. Isolated rat hepatocytes readily absorbed and 13 methylated As111, but not Asv (Lerman et al., 1983). Kidney slices, on the other hand, produced 14 five times more DMA from Asv than As111 (Lerman and Clarkson, 1983). Therefore, it is likely 15 that any Asv not initially reduced can be efficiently methylated in the kidney for subsequent 16 urinary excretion. 17 Identifying the main organs responsible for methylation of arsenic in vivo has not been 18 straightforward because in vitro results do not necessarily reflect in vivo methylation patterns 19 (NRC, 1999). Buchet and Lauwerys (1985) identified the rat liver as the main organ for 20 methylation, with the methylating capacities in the RBCs, brain, lung, intestine, and kidneys 21 being insignificant in comparison. Assays of arsenite methyltransferases from mouse tissues 22 demonstrated the testes had the highest methylating activity, followed by the kidney, lung, and 23 liver (Healy et al., 1998). Aposhian (1997) determined that the amount of methyltransferases 24 vary in the liver of different animal species. Arsenite bound to components of tissue can be 25 methylated and released (Marafante et al., 1981; Vahter and Marafante, 1983). This may explain 26 the initial rapid phase (immediate methylation and excretion) followed by a slow elimination 27 phase (continuous release of bound arsenite through methylation) (NRC, 1999), as described in 28 Section 3.4. 29 It has been demonstrated that inhibition of arsenic methylation results in increased tissue 30 concentrations of arsenic (Marafante and Vahter, 1984; Marafante et al., 1985). Loffredo et al. 31 (2003) suggest that the second methylation step is inducible and that the inducibility is possibly 32 polymorphic (i.e., more than one enzyme or enzyme form may be involved, depending on the 33 individual). This suggestion is based on observations that human urinary DMA concentrations 34 in high-exposure groups were higher and more variable than urinary MMA levels, and because 35 urinary DMA levels appeared to have a bimodal distribution in a population from Mexico, 36 regardless of exposure status. Others have suggested that the second methylation step may be 37 saturable, which would be consistent with the decreasing excretion of DMA with increasing 21 DRAFT—DO NOT CITE OR QUOTE ------- 1 arsenic exposures (Ahsan et al., 2007). Cysteine, GSH, and DTT have been shown to increase 2 the activity of arsenite methyltransferase and MMA methyltransferase (both later identified as 3 AS3MT; Lin et al., 2002) in purified rabbit liver enzyme preparations (Zakharyan et al., 1995). 4 Dithiols (e.g., reduced lipoic acid) have also been found to enhance arsenite methylation by 5 MMA111 methyltransferase (Zakharyan et al., 1999). Glutathione-S-transferase omega 1 6 (GSTO1) has also been associated with arsenic biotransformation (Meza et al., 2007). Although 7 humans have been observed to methylate arsenic, no arsenic methyltransferase has yet been 8 isolated from human tissues (Aposhian and Aposhian, 2006). 9 In vitro studies using rat liver preparations indicate that the methylating activity is 10 localized in the cytosol, with SAM being the main methyl donor for As111 methylation (Marafante 11 and Vahter, 1984; Buchet and Lauwerys, 1985; Marafante et al., 1985; Styblo et al., 1995, 1996; 12 Zakharyan et al., 1995). AS3MT catalyzes the transfer of the methyl group from SAM to the 13 arsenic substrates (Lin et al., 2002; Thomas, 2007). Expressing AS3MT in UROtsa (human 14 urothelial cells that do not normally methylate inorganic arsenic) caused the cells to effectively 15 methylate arsenite (Drobna et al., 2005). High concentrations of As111 or MMA111 in the culture 16 caused an inhibition in the formation of DMA, but had little effect on the formation of MMA. 17 The inhibition of DMA production resulted in MMA accumulation in cells. Drobna et al. (2006) 18 demonstrated that AS3MT was the major enzyme for arsenic methylation in human 19 hepatocellular carcinoma (HepG2) cells, but reducing it by 88% (protein levels) only accounted 20 for a 70% reduction in methylation capacity, suggesting that there is another methylation process 21 that is independent of AS3MT. 22 The addition of GSH has been found to increase the yield of mono- and dimethylated 23 arsenicals but suppressed the production of TMAO in the presence of rat AS3MT (Waters et al., 24 2004a), indicating that GSH suppresses the third methylation reaction but not the first two 25 (Thomas et al., 2007). Thomas et al. (2004) discovered a similar arsenic methyltransferase in the 26 rat liver, which they designated cyt!9 because an orthologous cyt!9 gene encodes an arsenic 27 methyltransferase in the mouse and human genome. It has subsequently been concluded that this 28 methyltransferase was the same as AS3MT. 29 GSH alone does not support recombinant rat AS3MT catalytic function, but when added 30 to a reaction mixture containing other reductants, the rate of arsenic methylation increases 31 (Waters et al., 2004b). GSH alone (5mM) does not support the catalytic activity of AS3MT, but 32 stimulates the methylation rate in the presence of the reductant tris(2-carboxylethyl)phosphine 33 (TCEP; 1 mM) (Thomas et al., 2007). GSH (5 mM) did not have any effect on DTT (1 mM)- 34 induced arsenic methylation. Drobna et al. (2004) linked the genetic polymorphism of AS3MT 35 with other cellular factors and to the inter-individual variability in the capacity of primary human 36 hepatocytes to retain and metabolize As111 (see Section 4.7). 22 DRAFT—DO NOT CITE OR QUOTE ------- 1 The main products of arsenic methylation in humans are MMAV and DMAV, which are 2 readily excreted in the urine (Marcus and Rispin, 1988). MMA111 and DMA111 have recently been 3 detected in human urine (NRC, 2001); however, most studies do not differentiate the valence 4 state of mono- or dimethylated arsenic species detected in urine or tissue samples. Le et al. 5 (2000a,b) and Del Razo et al. (2001) noted that the concentration of trivalent metabolites in the 6 urine may be underestimated because they are easily oxidized after collection. Le et al. (2000b) 7 found 43 to 227 ug/L of MMA111 in the urine of populations from Inner Mongolia, China, who 8 were exposed to 510-660 ppb (0.46 uM) of arsenic via the drinking water. 9 A small percent of DMA111 may further be methylated to TMAO in mice and hamsters 10 (see Kenyon and Hughes, 2001, for a review). A single human volunteer ingesting DMA 11 excreted 3.5% of the dose as TMAO (Kenyon and Hughes, 2001). TMAO can be detected in 12 urine following DMA exposure, but has not been detected in the blood or tissues of mice 13 exposed intravenously to DMA (Hughes et al., 2000) or in the urine of mammals orally exposed 14 to inorganic As. This may be due to rapid clearance of DMA and MMA from cells (Styblo et al., 15 1999b); however, most analytical methods are not optimized for the detection of TMAO that 16 could have been present but not detected. 3.3.3. Species Differences in the Methylation of Arsenic 17 There is considerable variation in the patterns of inorganic arsenic methylation among 18 mammalian species (NRC, 1999). Humans, rats, mice, dogs, rabbits, and hamsters have been 19 shown to efficiently methylate inorganic arsenic to MMA and/or DMA. Rats and hamsters 20 appear to methylate administered DMA into TMAO more efficiently than other species (NRC, 21 1999; Yamauchi and Yamamura, 1984). About 40% of urinary arsenic was present as TMAO 1 22 week after exposure to DMA in the drinking water, while 24% was present as TMAO after 7 23 months of exposure (100 mg/L) in male rats (Yoshida et al., 1998). 24 Humans (mainly exposed to background levels or exposed at work) have been estimated 25 through a number of studies to excrete 10% to 30% of the arsenic in its inorganic form, 10% to 26 20% as MMA, and 55% to 75% as DMA (see Vahter, 1999a, for a review). In contrast, a study 27 of urinary arsenic metabolites in a population from northern Argentina exposed to arsenic via 28 drinking water demonstrated an average of only 2% MMA in the urine (Vahter et al., 1995b; 29 Concha et al., 1998b). This may indicate variations in methylation activity depending on the 30 route of exposure, level of exposure, and possible nutritional or genetic factors. Although 31 humans are considered efficient at arsenic methylation, they are less efficient than many animal 32 models, as indicated by the larger proportion of MMAV excreted in the urine (Vahter, 1999a). 33 This is important because it may explain why humans are more susceptible to cancer from 34 arsenic exposures, and why no adult animal model for inorganic-arsenic-induced cancers has yet 35 been identified (Tseng et al., 2005). 23 DRAFT—DO NOT CITE OR QUOTE ------- 1 The rabbit (Marafante et al., 1981; Vahter and Marafante, 1983; Maiorino and Aposhian, 2 1985) and hamster (Charbonneau et al., 1980; Yamauchi and Yamamura, 1984; Marafante and 3 Vahter, 1987) appear to be more comparable to humans with respect to arsenic methylation than 4 other experimental animals (NRC, 1999). However, rabbits and hamsters, in general, excrete 5 more DMA and less MMA than humans. In contrast, Flemish giant rabbits (De Kimpe et al., 6 1996) excrete MMA in amounts similar to humans. Mice and dogs, efficient methylators of 7 arsenic, excrete more than 80% of a single arsenic dose administered as DMA within a few days 8 (Charbonneau et al., 1979; Vahter, 1981). Guinea pigs (Healy et al., 1997), marmoset monkeys 9 (Vahter et al., 1982; Vahter and Marafante, 1985), and chimpanzees (Vahter et al., 1995a), on 10 the other hand, do not appear to appreciably methylate inorganic arsenic. In addition, no 11 methyltransferase activity was detected in these species (Zakharyan et al., 1995, 1996; Healy et 12 al., 1997; Vahter, 1999a). Li et al. (2005) identified a frameshift mutation in the chimpanzee 13 AS3MT gene that resulted in the production of an inactive truncated protein, possibly explaining 14 the lack of methylation activity in that species. 15 AS3MT homolog proteins with five fully conserved cysteine residues have been 16 observed in the genome of numerous species (Thomas et al., 2007). Chimpanzees were found to 17 differ from other species studied in that their AS3MT protein was shorter and lacked the 5th 18 cysteine (Thomas et al., 2007). Healy et al. (1999) identified marked variations in the activity of 19 methyltransferases, while Vahter (1999b) characterized differences in methylation efficiency 20 among different human populations. The observed variations in methyltransferase activity and 21 methylation efficiency are probably the underlying reason for the cross-species variability in 22 methylation ability, as all the species had ample arsenate reductase activity (Vahter, 1999a; 23 NRC, 2001). 24 Although arsenic methylation is generally believed to take place in order to enhance 25 excretion, there are several species (guinea pigs, marmoset monkeys, and chimpanzees) that do 26 not methylate arsenic, but still efficiently excrete it. In fact, these animals do not retain arsenic 27 any longer than species that methylate arsenic (Cohen et al., 2006), indicating that factors other 28 than methylation also affect arsenic excretion rates. Supporting this is the fact that inorganic 29 arsenic is found in the urine of even the most efficient methylators (Vahter, 1994). 3.3.4. Thioarsenical Metabolites 30 In 2004, Hansen et al. reported the detection of unusual arsenic-containing metabolites in 31 the urine of sheep exposed to arsenic-contaminated vegetation. The metabolite was tentatively 32 identified as dimethylmonothioarsinic acid (DMMTA111), a sulfur-containing derivative of 33 DMA111 as shown in Figure 3-3. Because the exposed sheep consumed algae known to contain 34 arsenosugars, some of which contain sulfur, the relevance of this finding to human exposures 35 was not initially clear. Subsequently, Raml et al. (2006) detected the presence of DMMTA111 in 24 DRAFT—DO NOT CITE OR QUOTE ------- 1 the urine of Japanese men, but again, consumption of arsenosugars was suspected as a source of 2 the observed arsenic containing species. SH-Asm-CH3 OH-Asv-CH3 I I CH3 CH3 DMMTA111 DMMTAV Source: Hansen et al. (2004). Figure 3-3. Thioarsenical structures. 3 In experiments addressing this issue, Adair et al. (2007) and Naramandura et al. (2007) 4 found substantial concentrations of thioarsenical metabolites in arsenic-exposed experimental 5 animals. Adair et al. (2007) administered drinking water containing 100 ppm Asv or up to 200 6 ppm DMA111 to female Fisher 344 rats for 14 days. During analysis of the urine (collected during 7 the last 24 hours of exposure) for metabolites, they found high levels of DMMTA111 and 8 trimethylarsine sulfide (another sulfur-containing metabolite) in the urine of rats treated with 9 DMA111. Lower levels of the sulfur-containing metabolites were detected in the urine of 10 arsenate-treated animals. They proposed a mechanism whereby the reaction of DMA111 and 11 DMAV with hydrogen sulfide resulted in the observed metabolites. 12 Naranmandura et al. (2007) administered single doses of 5.0 mg/kg As111 to Syrian 13 hamsters and Wistar rats by gavage and measured the levels of sulfur-containing arsenic 14 metabolites in urine. Both DMMTA111 and dimethylmonothioarsonic acid (DMMTAV) were 15 found at appreciable levels in urine from hamsters, but only the latter metabolite was found in rat 16 urine. A previously uncharacterized metabolite, monomethylmonothioarsonic acid, was also 17 found in urine from both species. 18 These studies suggest that the generation of sulfur-containing arsenic metabolites does 19 not depend on exposures to arsenosugars, at least in rodents, but can occur during the 20 metabolism of inorganic arsenic compounds. In 2007, Raml et al. presented evidence that this 21 pathway was also significant in humans. DMMTA111 was detected in the urine of 44% (33 of 75) 22 women exposed to inorganic arsenic-contaminated drinking water in Bangladesh. The 23 metabolite was present in urine samples at concentrations between "trace" amounts and 24 ug/L, 24 with total arsenic concentrations ranging from 8 to 1034 ug/L. It was suggested that 25 DRAFT—DO NOT CITE OR QUOTE ------- 1 thioarsenical metabolites may have been present in urine from other epidemiological studies of 2 arsenic-exposed populations, but may have not been detected due to analytical difficulties. 3.4. ELIMINATION 3 The major route of excretion for most arsenic compounds by humans is via the urine 4 (Yamauchi and Yamamura 1979; Tarn et al., 1979; Pomroy et al., 1980; Buchet et al., 1981). Six 5 human subjects who ingested 0.01 ug of radio-labeled 74Asv excreted an average of 38% of the 6 administered dose in the urine within 48 hours and 58% within 5 days (Tarn et al., 1979). 7 Inorganic arsenic elimination in humans has been observed to be triphasic, with first-order half- 8 lives for elimination of 1 hour, 30 hours, and 200 hours (Mealey et al., 1959 used As111; Pomroy 9 etal., 1980 used Asv). 10 As mentioned in the preceding section, MMA and DMA are important metabolites 11 generated after exposure to inorganic As. These methylated metabolites are excreted in the 12 urine faster than the inorganic As. In humans orally exposed to MMA or DMA in aqueous 13 solution, about 78% of MMA and 75% of DMA were excreted in the urine within 4 days of 14 ingestion (Buchet et al., 1981). In mice, the half-time of MMA and DMA excretion was found 15 to be about 2 hours following iv administration (Hughes and Kenyon, 1998). 16 Kenyon et al. (2008) administered 0, 0.5, 2, 10, or 50 ppm of arsenic as sodium arsenate 17 to adult C57B1/6 female mice in the drinking water for 12 weeks. The average daily intakes 18 were estimated to be 0, 0.083, 0.35, 1.89, and 7.02 mg As/kg/day, respectively. Levels of 19 MMA111, DMA111, DMAV, and TMAO in the urine collected at the end of treatment increased in a 20 linear manner with dose, but Asv and MMAV did not. 21 Rats excrete DMA slowly compared to other species (Vahter et al., 1984), even though 22 they are efficient at methylating inorganic arsenic to DMA. The slow excretion is believed to be 23 associated with retention of a significant portion of the DMA in erythrocytes (Odanaka et al., 24 1980; Lerman and Clarkson, 1983; Vahter, 1983; Vahter et al., 1984). The biliary excretion of 25 inorganic arsenic by rats is about 800 times greater than observed in dogs and 37 times that of 26 rabbits, as proportion of administered dose. Hughes et al. (2005) found that in mice the level of 27 MMAV excreted in the urine compared to the bile was related to dose, with fecal excretion 28 increasing at higher doses. Cui et al. (2004a) also found that rat biliary excretion rates varied 29 with dose, but found it was also related to route of administration and chemical form. After oral 30 administration of inorganic arsenic (either form) to male Sprague-Dawley rats, MADG and 31 DMAV (likely present due to dissociation of DMAG) were the predominant forms in the bile. 32 MADG was found at a higher level after a higher (i.e., 100 ppm) dose, while DMAV was more 33 prevalent at the lower dose (i.e., 10 ppm). Kala et al. (2000) found that the secretion of arsenic 34 into the bile of rats was dependent on the multi-drug resistance-associated protein 2 transporter 26 DRAFT—DO NOT CITE OR QUOTE ------- 1 (MPR2/cMOAT) and that GSH is necessary for the transport, as arsenic-glutathione complexes 2 accounted for the majority of arsenic found in the bile. 3 Although absorbed arsenic is removed from the body mainly via the urine, small amounts 4 of arsenic are excreted through other routes (e.g., skin, sweat, hair, breast milk). While arsenic 5 has been detected at low levels in the breast milk of women in northwestern Argentina (i.e., 2 6 ug/kg), breastfeeding was associated with lower concentrations of arsenic in the urine of 7 newborn children (Concha et al., 1998c) than formula feeding, owing to the use of arsenic 8 contaminated water in formula preparation. Parr et al. (1991) measured arsenic (as well as other 9 elements) in the breast milk from three groups of mothers from four countries (Guatemala, 10 Hungary, Nigeria, and the Philippines), and one to two groups from Sweden and Zaire. The 11 breast milk was collected 3 months after birth. Levels of arsenic in the breast milk from women 12 in the Philippines were higher than other regions with levels about 19 ug/kg. Women from 13 Nigeria had levels similar to those observed by Concha et al. (1998c). Women from all the other 14 areas measured had levels of 0.24 to 0.55 ug/kg. 15 The average concentration of arsenic in sweat induced in a hot and humid environment 16 was 1.5 ug/L, with an hourly loss rate of 2.1 ug (Vellar, 1969). Based on an average arsenic 17 concentration in the skin of 0.18 mg/kg, Molin and Wester (1976) estimated that the daily loss of 18 arsenic through desquamation was 0.1 to 0.2 ug in males with no known exposure to arsenic. 3.5. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS 19 Physiologically based pharmacokinetic (PBPK) models for inorganic arsenic are 20 important for developing a biologically based dose-response (BBDR) model. The development 21 of useful BBDR models has proved to be challenging because inorganic arsenic appears to 22 mediate its toxicity through a range of metabolites, and their roles with regard to specific adverse 23 effects are not clear (Clewell et al., 2007). 24 A PBPK model for exposure to inorganic arsenic (orally, intravenously, and 25 intratracheally) was developed in hamsters and rabbits by Mann et al. (1996a). The model 26 includes tissue compartments for lung (nasopharynx, tracheobronchial, pulmonary), plasma, 27 RBCs, liver, GI tract, skin, kidney, keratin, and combined other tissues. Oral absorption of As111, 28 Asv, and DMA (pooled In and V oxidation states) was modeled as a first-order transport process 29 directly from the GI contents into the liver. Distribution to tissues was diffusion-limited, with 30 transfer rates estimated based upon literature values for capillary thickness and pore sizes for 31 each tissue. Reductive metabolism of Asv to As111 was modeled as a first-order process occurring 32 in the plasma. Oxidative metabolism of As111 to Asv was modeled as first-order processes in the 33 plasma and kidneys. Methylation of inorganic arsenic species to MMA (pooled In and V 34 oxidation states) and then to DMA were modeled as saturable Michaelis-Menten processes 35 taking place in the liver. Urinary, biliary, and fecal excretion of As111, Asv, MMA, and DMA 27 DRAFT—DO NOT CITE OR QUOTE ------- 1 also are modeled as first-order processes. Parameters for absorption, tissue partition, 2 metabolism, and biliary excretion were estimated by fitting the model to literature data on the 3 urinary and fecal excretion of total arsenic from rabbits and hamsters administered various 4 arsenic compounds by iv, oral gavage, or intratracheal instillation (Charbonneau et al., 1980; 5 Yamauchi and Yamamura, 1984; Marafante et al., 1985, 1987). The model was found to 6 accurately simulate the excretion of arsenic metabolites in the urine of rabbits and hamsters and 7 to produce reasonable fits to liver, kidney, and skin concentrations in rabbits and hamsters 8 (Yamauchi and Yamamura, 1984; Marafante et al., 1985; Marafante and Vahter, 1987). 9 Mann et al. (1996b) extended their PBPK model for use in humans by adjusting 10 physiological parameters (organ weights, blood flows) and re-estimating absorption and 11 metabolic rate constants. The model was fit to literature data on the urinary excretion of total 12 arsenic following a single oral dose of As111 or Asv in human volunteers (Tarn et al., 1979; 13 Buchet et al., 1981). The extended human model was further tested against empirical data on the 14 urinary excretion of the different metabolites of inorganic arsenic following oral intake of As111, 15 intake of inorganic arsenic via drinking water, and occupational exposure to arsenic trioxide 16 (ATO) (Harrington et al., 1978; Valentine et al., 1979; Buchet et al., 1981; Vahter et al., 1986). 17 The model predicted a slight decrease (about 10%) in the percentage of DMA in urine with 18 increasing single-dose exposure (highest dose of arsenic at 15 ug/kg of body weight), especially 19 following exposure to As111, and an almost corresponding increase in the percentage of MMA. 20 The model predicted that adults' drinking water containing 50 ppb would excrete more arsenic in 21 urine than an occupational inhalation exposure of 10 ug/m3 (Mann et al., 1996b). 22 Yu (1999a,b) also developed a PBPK model for arsenic in humans that includes tissue 23 compartments for lung, skin, fat, muscle, combined kidney and richly perfused tissues, liver, 24 intestine, GI and stomach contents, and bile. Oral absorption of As111, Asv, and DMA (pooledIn 25 and v oxidation states) was modeled as first-order transport from the GI contents into the 26 intestinal tissue. Distribution to tissues was modeled as perfusion-limited. Reductive 27 metabolism of Asv to As111 was modeled as a first-order, GSH-dependent process taking place in 28 the intestinal tissue, skin, liver, and kidney/rich tissues. Oxidative metabolism of As111 to Asv 29 was not modeled. Methylation of inorganic arsenic species to MMA (pooled In and v oxidation 30 states) and then to DMA was modeled as saturable Michaelis-Menten processes occurring in the 31 liver and kidney. Urinary, biliary, and fecal excretion of As111, Asv, MMA, and DMA were 32 modeled as first-order processes. Parameters for absorption, tissue partition, metabolism, and 33 biliary excretion were estimated by fitting the model to literature data on tissue concentrations of 34 total arsenic from a fatal human poisoning (Saady et al., 1989), and blood, urine, and fecal 35 elimination of total arsenic following oral administration (Odanaka et al., 1980; Pomroy et al., 36 1980). The model was not tested further against external data, and fits to the data sets used for 37 parameter estimation were not provided. 28 DRAFT—DO NOT CITE OR QUOTE ------- 1 Gentry et al. (2004) adapted the model proposed by Mann et al. (1996a) to different 2 mouse strains by adjusting physiological parameters (organ weights and perfusion rates). The 3 absorption, partition, and metabolic rate constants were re-estimated by fitting the model to 4 literature data on urinary excretion of various arsenic species following iv administration of 5 MMA to B6C3F1 mice (Hughes and Kenyon, 1998) or single oral administration of As111 or Asv 6 to mice (Kenyon et al., 1997; Hughes et al., 1999). Additionally, the description of methylation 7 in the model was refined to include the uncompetitive inhibition of the conversion of MMA to 8 DMA by As111. The PBPK model was then validated using data from a single oral administration 9 of Asv (Hughes et al., 1999) and a 26-week drinking water exposure of As111 to C57Black mice 10 (Moser et al., 2000). These data were found to adequately fit the model without further 11 parameter adjustment. Ng et al. (1999) had found arsenic-induced tumors in C57B1/6J mice, 12 while numerous other mouse strains (Swiss CR:Nffl[S], C57Bl/6p53 [+/-], C57Bl/6p53 [+/+], and 13 Swiss CD-I) had not experienced a significant increase in arsenic-induced tumors. The Gentry 14 et al. (2004) model was unable to explain the different outcomes in the mouse bioassay on the 15 basis of predicted target organ doses. 16 The Mann et al. (1996a,b) and Gentry et al. (2004) models are well documented, were 17 validated against external data, and appear to capture the salient features of arsenic 18 toxicokinetics in rodents and humans. The information provided by these models may help 19 explain the MO As involved in carcinogenesis along with possible reasons that humans are 20 apparently more susceptible to the carcinogenic effects of arsenic. 21 Clewell et al. (2007) noted that the then-available PBPK models did not incorporate the 22 most recent available information on arsenic methylation kinetics and suggested several steps for 23 improving the PBPK models. El-Masri and Kenyon (2008) have developed a PBPK model 24 incorporating some of the improvements suggested by Clewell et al. (2007) (although not the 25 simulation of changes in gene expression). The model predicts the levels of inorganic arsenic 26 and its metabolites in human tissues and urine following oral exposure of Asv, As111, and for oral 27 exposure to organoarsenical pesticides. The model consists of interconnecting submodels for 28 inorganic arsenic (As111 and Asv), MMAV, and DMAV. Reduction of MMAV and DMAV to their 29 trivalent forms is also modeled. The submodels include the GI tract (lumen and tissue), lung, 30 liver, kidney, muscle, skin, heart, and brain, with reduction of MMAV and DMAV to their 31 trivalent forms modeled as occurring in the lung, liver, and kidney. The model also incorporates 32 the inhibitory effects of As111 on the methylation of MMA111 to DMA and MMA111 on the 33 methylation of As111 to MMA into consideration, modeled as noncompetitive inhibition. This 34 model differs from the other models described above because it provides an updated description 35 of metabolism using recent biochemical data on the mechanism of arsenic methylation. In 36 addition, it uses in vitro studies to estimate most of the model parameters (statistically 37 optimizing those that are sensitive to urinary excretion levels to avoid problems with parameter 29 DRAFT—DO NOT CITE OR QUOTE ------- 1 identifiability), and can predict the formation and excretion of trivalent methylated arsenicals. 2 The partition coefficients estimated in the model are comparable to those developed by Yu 3 (1999a). The performance of the model was tested against limited human data on urinary 4 excretion; the model needs to be evaluated for its ability to predict the tissue and urinary 5 concentrations of arsenicals in large numbers of subjects. This model is an improvement over 6 previous models because it can quantitatively assess impacts of parameter variability arising 7 from genetic polymorphism. 30 DRAFT—DO NOT CITE OR QUOTE ------- 4. HAZARD IDENTIFICATION 4.1. STUDIES IN HUMANS 1 Numerous epidemiologic investigations have examined the association between 2 waterborne arsenic exposure and cancer outcome. These epidemiologic investigations used 3 many different study designs, each with their inherent limitations. Regardless of the study type, 4 the majority of these investigations found some level of association between arsenic exposure 5 and cancer outcome. This association is not new, since arsenic exposure has been linked with 6 cancer as far back as 1887 when Hutchinson reported an unusual number of skin tumors in 7 patients treated with arsenicals. Since 1887, the association between skin cancer and arsenic has 8 been reported in a number of studies (Tseng et al., 1968; Tseng, 1977; Chen et al., 1985, 9 1988a,b; Wu et al., 1989; Hinwood et al., 1999; NRC, 1999; Tsai et al., 1999; Karagas et al., 10 2001; Knobeloch et al., 2006; Lamm et al., 2007). 11 The SAB Arsenic Review Panel provided comments on key scientific issues associated 12 with arsenicals on cancer risk estimation in July 2007 (SAB, 2007). It was concluded that the 13 Taiwanese database is still the most appropriate source for estimating bladder and lung cancer 14 risk among humans (specifics provided in Section 5) because of: (1) the size and statistical 15 stability of the database relative to other studies; (2) the reliability of the population and 16 mortality counts; (3) the stability of residential patterns; and (4) the inclusion of long-term 17 exposures. However, SAB also noted considerable limitations within this data set (EPA-SAB- 18 07-008, http://www.epa.gov/sab). The Panel suggested that one way to mitigate the limitations 19 of the Taiwanese database would be to include other relevant epidemiological studies from 20 various countries. For example, SAB referenced other databases that contained studies of 21 populations also exposed to high levels of arsenic (e.g., Argentina and Chile), and recommended 22 that these alternate sources of data be used to compare the unit risks at the higher exposure levels 23 that have emerged from the Taiwan data. SAB also suggested that, along with the Taiwan data, 24 published epidemiology studies from the United States and other countries where the population 25 is chronically exposed to low levels of arsenic in drinking water (0.5 to 160 ppb) be critically 26 evaluated, using a uniform set of criteria presented in a narrative and tabular format. The 27 relative strengths and weaknesses of each study should be described in relation to each criterion. 28 The caveats and assumptions used should be presented so that they are apparent to anyone who 29 uses these data. The risk assessment background document should be a complete and transparent 30 treatment of variability within and among studies and how it affects risk estimates. Additionally, 31 SAB (2007) recommended considering the following issues when reviewing "low-level" and 32 "high-level" studies: (1) estimates of the level of exposure misclassification; (2) temporal 33 variability in assigning past arsenic levels from recent measurements; (3) the extent of reliance 34 on imputed exposure levels; (4) the number of persons exposed at various estimated levels of 31 DRAFT—DO NOT CITE OR QUOTE ------- 1 waterborne arsenic; (5) study response/participation rates; (6) estimates of exposure variability; 2 (7) control selection methods in case-control studies; and (8) the resulting influence of these 3 factors on the magnitude and statistical stability of cancer risk estimates. 4 In order to address these issues, this Toxicological Review provides a comprehensive 5 review of the significant epidemiologic investigations in the literature from 1968 to 2007 with 6 the focus on the more recent publications. The report includes data from all populations that 7 have been examined in regards to cancer from arsenic exposure via drinking water. Earlier 8 publications were reviewed and are included as needed to facilitate the understanding of results 9 from certain study populations. As recommended by SAB, studies were presented in both a 10 narrative (below) and tabular (Appendix B) format. Each publication was evaluated using a 11 uniform set of criteria, including the study type, the size of the study population and control 12 population, and the relative strengths and weaknesses of the study. While the information in the 13 tables mirrors the information in the narrative, the narrative may provide additional important 14 information concerning the investigation. The studies are presented by country of origin, then in 15 chronological order by publication year. In order to facilitate comparisons across the 16 epidemiological studies, the arsenic concentrations pertaining to water exposure levels have been 17 converted from milligrams (mg) per liter (or ppm) to parts per billion (ppb). This was not 18 applied when discussing animal or in vitro MOA studies because a wide range of concentrations 19 was employed; converting the arsenic levels or doses into ppb would not be reader-friendly. 4.1.1. Taiwan 20 More than 80 years ago (between 1910 and 1920), parts of southwestern Taiwan began 21 using artesian (ground water) wells to increase water supplies and decrease the salt content of 22 their drinking water. Some of these artesian wells were discovered to be contaminated with 23 naturally occurring arsenic, thus resulting in widespread arsenic exposure. As a result, the 24 Taiwanese population has been extensively studied. Due to the high arsenic content in the 25 artesian wells, water was piped into certain areas in Taiwan from the reservoir of the Chia-Nan 26 irrigation system in 1956. This water was reported to contain 10 ppb of arsenic (Tseng, 1977). 27 Almost 75% of the residences had tap water by the 1970s; however, a survey in 1988 noted that 28 artesian well water was still used for drinking, aquaculture, and agriculture in 1988, especially 29 during the dry season (Wu et al., 1989). 30 Tseng et al. (1968) conducted a general survey using an ecological study design of 31 40,421 inhabitants (21,152 females, 19,269 males) from the southwest coast of Taiwan in order 32 to determine the potential relationship between skin cancer and chronic arsenicism. The arsenic 33 content was measured in 142 samples from 114 wells (110 deep artesian and 4 shallow) and 34 ranged from 10 to 1,820 ppb. The authors noted, however, that the arsenic content varied 35 considerably over a 2-year period when measurements were taken. For example, in one well 36 measurements were 528 ppb in July, 1962; 530 ppb in June, 1963; and 1190 ppb in February, 32 DRAFT—DO NOT CITE OR QUOTE ------- 1 1964. These variations made dose-response relationships difficult to determine. Study subjects 2 were categorized by arsenic exposure into three groups (low: 0-290 ppb, medium: 300-590 ppb, 3 and high: 600 ppb or greater). The overall prevalence rate for skin cancer was 10.6 per 1,000. 4 The male-to-female ratio was 2.9:1 for skin cancer. The prevalence rate increased steadily with 5 age (recorded in 10-year increments), except for declining cancer prevalence rates for females 6 older than 69 years. Age-specific (plotted in 20-year intervals) and sex-specific prevalence rates 7 for skin cancer increased with arsenic concentration. The most common type of lesion was intra- 8 epidermal carcinoma (51.7%), and the body areas most frequently involved were unexposed 9 surfaces (74.5%). In addition, an extremely high percentage of cases with multiple skin cancer 10 (99.5%) was observed. The association between BFD and skin cancer was significantly higher 11 than expected. Strengths of the Tseng et al. (1968) study include the large number of 12 participants and the inclusion of dose-response information. Weaknesses include the lack of 13 individual exposure data (ecological study design) and the potential for recall bias among study 14 participants in determining the age of cancer onset and the length of residence in the area. In 15 addition, changes in water supply over time were not noted, information on smoking history was 16 not obtained, and the arsenic concentration from individual wells varied over time. 17 Tseng (1977) also used the general ecologic survey design discussed in Tseng et al. 18 (1968) to report skin cancer incidence among the 40,421 individuals and to follow up on 1,108 19 patients with BFD (identified between 1958 and 1975). By the end of the follow-up period, 528 20 of the BFD patients had died. Tseng (1977) identified 428 cases (prevalence of 10.6/1,000) of 21 skin cancer and 370 cases (prevalence of 9.0/1,000) of BFD, and analyzed the relationship 22 between the two. Skin cancer and BFD occurred in 61 cases (1.51/1,000), but only 4 cases 23 (0.09/1,000) were expected. The observed:expected ratio was 16.77. Tseng (1977) determined 24 that the patients with BFD consumed artesian water before the onset of the disease, and none of 25 the residents who had consumed only surface water or water from shallow wells developed BFD. 26 This finding illustrates that no cases were found among the inhabitants who were born after the 27 tap water supply was introduced, and supports the close association between the consumption of 28 arsenic contaminated water and the development of BFD. In addition, the study found that 29 patients with skin cancer or BFD had a greater incidence of death due to cancers of various sites 30 (28% and 19%, respectively) when compared to the general population of the endemic area 31 (13%) or to the entire population of Taiwan (8%). 32 Using similar arsenic exposure categories (low <300 ppb, medium 300-600 ppb, and 33 high >600 ppb) from the Tseng et al. (1968) investigation, the skin cancer and the BFD 34 prevalence rates showed an ascending gradient from low to high arsenic exposure for both sexes 35 (Tseng, 1977). Skin cancer prevalence rates by age and arsenic exposure group were as follows: 36 20-39 years (high exposure: 11.5; medium exposure 2.2; low exposure: 1.3); 40-59 years (high: 37 72.0; medium: 32.6; low: 4.9); and 60+ years (high: 192.0; medium: 106.2; low: 27.1). BFD 33 DRAFT—DO NOT CITE OR QUOTE ------- 1 prevalence rates by age and arsenic exposure group were as follows: 20-39 years (high: 14.2; 2 medium: 13.2; low: 4.5); 40-59 years (high: 46.9; medium: 32.0; low: 10.5); 60+ years (high: 3 61.4; medium: 32.2; low: 20.3). The common cause of death in the patients with skin cancer and 4 BFD was carcinoma of various sites, including lung, bladder, liver, and kidney. The Tseng 5 (1977) investigation observed that the prevalence of skin cancer increased steadily with age. It 6 was difficult to obtain the age at onset of cancer from patient interviews, as most of the patients 7 were unable to name a date. Strengths and weaknesses of this study are the same as Tseng et al. 8 (1968); however, this study also included adjusted analysis for age and gender. 9 The objective of the Chen et al. (1985) ecological study was to evaluate the possible 10 association between exposure to elevated levels of arsenic from artesian well water and cancer in 11 the BFD-endemic area of southwestern Taiwan (i.e., Peimen, Hsuechia, Putai, and Ichu 12 townships). The population of the BFD-endemic area in 1982 was 120,607 and consisted 13 primarily of individuals engaged in farming, fishing, and salt production operations. The 14 educational and socioeconomic status of the BFD-endemic area was below average for Taiwan. 15 Chen et al. (1985) cited arsenic measurements from 83,565 wells across Taiwan taken by Lo et 16 al. (1977), which showed that 29.1% of the wells in the study area had concentrations greater 17 than 50 ppb (with the highest concentration measuring 2500 ppb), while only 5.7% of wells in 18 other areas of Taiwan exceeded 50 ppb. A previous study by Chen et al. (1962) demonstrated a 19 range of 350 to 1,140 ppb, with a median of 780 ppb arsenic content in Taiwanese artesian wells 20 in BFD-endemic areas. As compared with the general population in Taiwan, both the 21 standardized mortality ratio (SMR) and cumulative mortality rate were significantly higher in 22 BFD-endemic areas. SMRs for males were significant for bladder (11.00, 95% confidence 23 interval [CI]: 9.33-12.87), kidney (7.72, 95% CI: 5.37-10.07), skin (5.34, 95% CI: 3.79-8.89), 24 lung (3.20, 95% CI: 2.86-3.54), liver (1.70, 95% CI: 1.51-1.89), and colon (1.80, 95% CI: 1.17- 25 2.03) cancers. SMRs for females also were significantly increased for bladder (20.09, 95% CI: 26 17.02-23.16), kidney (11.19, 95% CI: 8.38-14.00), skin (6.52, 95% CI: 4.69-8.35), lung (4.13, 27 3.60-4.66), liver (2.29, 95% CI: 1.92-2.66), and colon (1.68, 95% 1.26-2.10) cancers. Cancer 28 SMRs were greater in villages that used only artesian wells as the drinking water source, as 29 compared to villages that used both artesian and shallow wells. Villages and townships using 30 only shallow wells generally had the lowest SMRs. Strengths of the investigation include the 31 use of general population of Taiwan and world population for determining SMRs and potential 32 confounders of age and gender were controlled for in the analysis. Weaknesses were that arsenic 33 measurements were not linked to cancer mortality, death certificates list the main cause of death 34 (Yang et al., 2005) rather than all causes, and SMRs were only presented by township and by 35 well type. 36 To evaluate the association between high arsenic exposure from artesian well water and 37 cancer mortality in the BFD-endemic area of the southwest coast of Taiwan (i.e., the Peimen, 34 DRAFT—DO NOT CITE OR QUOTE ------- 1 Hsuechia, Putai, and Ichu townships), Chen et al. (1986) used a case-control study design to 2 evaluate 69 bladder cancer, 76 lung cancer, and 65 liver cancer deceased cases and 368 alive 3 community controls matched on age and gender. The study area was the same one Chen et al. 4 had used in 1985. Cases were selected from the Republic of China's National Health 5 Department between January 1980 and December 1982. The age distribution for cases was 6 significantly lower than the controls. Similar gender distributions were observed for bladder and 7 lung cancer cases and controls, though there was a slightly higher proportion of males in liver 8 cancer cases than in controls. Other sociodemographic factors (marital status, education, 9 occupation, and resident years) were comparable between cases and controls. Age and gender 10 differences were adjusted for in the analysis. The artesian well water arsenic content from the 11 BFD-endemic area ranged from 350 to 1,140 ppb (median 780 ppb), and the shallow well water 12 arsenic concentration ranged from below detection limits to 300 ppb (median 40 ppb). A 13 positive dose-response relationship was observed between the exposure to artesian well water 14 and cancers of bladder, lung, and liver. The age-gender-adjusted odds ratios (ORs) of bladder, 15 lung, and liver cancers for those who had used artesian well water for 40 or more years were 16 3.90, 3.39, and 2.67, respectively, when compared with those who never used artesian well 17 water. Regression analyses examined the associations between exposure to artesian well water 18 and bladder, lung, and liver cancers after adjusting for other variables including age, gender, and 19 cigarette smoking. Results showed a statistically significant association between exposure to 20 artesian well water and bladder and lung cancers (p < 0.01) when other variables were 21 controlled, but the association between the exposure to artesian well water and liver cancer was 22 not statistically significant (p < 0.05). (The text of the article specifies that liver cancers are not 23 significantly associated with arsenic, but the table that the text refers to illustrates a significant 24 association.) Strengths of the Chen et al. (1986) study include that most cases were confirmed 25 using histology or cytology findings, cancer cases and controls were from the same BFD 26 community, and potential confounders were adjusted for in the analysis (i.e., age, gender, 27 smoking, tea consumption, vegetable consumption, and fermented bean consumption). 28 Weaknesses include selection bias (control selection) and not controlling recall bias for the 29 following confounders: lifestyle, diet, daily water consumption, and source of water. 30 In a cohort study conducted by Chen et al. (1988a), cancer mortality associated with BFD 31 was analyzed in area residents (i.e., Peimen, Hsuechia, Putai, and Ichu townships, Taiwan) from 32 1973 to 1986. Arsenic levels in drinking water were measured between 1962 and 1964; these 33 levels were used to divide the study population into three groups: <300 ppb; 300-599 ppb; and 34 >600 ppb. Sociodemographic characteristics including lifestyle, diet, and living conditions were 35 comparable among study participants. Between 1974 and 1976, water from more than 83,000 36 wells in 313 villages throughout Taiwan was reanalyzed for arsenic content. The levels of 37 arsenic in the drinking water were consistent between the two measurement periods. Death 3 5 DRAFT—DO NOT CITE OR QUOTE ------- 1 certificates (n = 1031) were obtained from Taiwanese health care registration offices. Age- 2 adjusted cancer mortality rates were calculated using the 1976 world population as the standard. 3 Significantly elevated dose-response cancer mortality was observed among residents of the BFD 4 area (<300 ppb, SMRfemale=118.8, male=154.0; 300-599 ppb SMRfemale=182.6, 5 male=258.9; >600 ppb SMR female=369.1, male=434.7) as compared to the general population 6 of Taiwan (SMR female=85.5, male=128.1). For both genders, significantly elevated dose- 7 response mortality also was observed for cancers of the liver, lung, skin, bladder, and kidney in 8 comparison to the general population of Taiwan. A strength of the study is that data from 9 arsenic monitoring conducted in 1962-64 and 1974-76 revealed similar results. A weakness of 10 the study is that arsenic exposure levels are not individualized. 11 The objective of the Chen et al. (1988b) cohort (nested case-control) study was to 12 examine multiple risk factors and their correlation to malignant neoplasms related to BFD. A 13 total of 241 BFD cases, including 169 with spontaneous or surgical amputations of affected 14 extremities and 759 age-sex-residence-matched healthy community controls, were identified and 15 studied in the Peimen, Hsuechia, Putai, and Ichu townships of southwest Taiwan. Multiple 16 logistic regression analysis showed that artesian well water consumption, arsenic poisoning, 17 familial history of BFD, and undernourishment were significantly associated with the 18 development of BFD. In a nonconcurrent cohort, cancer mortality of 789 BFD patients followed 19 for 15 years also was examined using a life table. Results showed a significantly higher 20 mortality from cancers of the bladder (SMR=38.80, p < 0.001), skin (SMR=28.46, p < 0.01), 21 lung (SMR=10.49, p < 0.001), liver (SMR=4.66, p < 0.001), and colon (SMR=3.81, p < 0.05) as 22 compared with the general population in Taiwan. When non-BFD residents in the BFD-endemic 23 area were used as controls, significant differences in mortality rates were found for cancers of 24 the bladder (SMR=2.55, p < 0.01), skin (SMR=4.51, p < 0.05), lung (SMR=2.84, p < 0.01), and 25 liver (SMR=2.48, p < 0.01). The results strongly suggest carcinogenic effects from the artesian 26 well water in the BFD-endemic area. Study strengths include minimizing recall bias through 27 interview techniques, which identified the education, hours of occupational sunshine exposure, 28 artesian well use, family medical history, history of smoking and alcohol use, and frequency of 29 categories of food consumption. SMRs were calculated using both the national Taiwanese 30 population and the local endemic area population, and BFD cases were matched to healthy 31 community controls for age, sex, and residence. A weakness of the study was not providing the 32 individual arsenic dose levels. 33 Chiang et al. (1988) conducted a case-control prevalence study of bladder cancer in the 34 BFD-endemic and surrounding areas of the southwestern coast of Taiwan. Four groups (cases: 35 246 BFD patients; controls: 444 residents of the BFD-endemic area, 286 residents of the region 36 neighboring the endemic area, and 731 residents of the non-endemic area) were screened using a 37 detailed questionnaire and urinalysis. Three hundred and four subjects received urinary cytology 36 DRAFT—DO NOT CITE OR QUOTE ------- 1 examinations. The study revealed no difference in the prevalence of bladder cancer between the 2 BFD patients and non-BFD controls in the BFD-endemic area, indicating that individuals in the 3 BFD-endemic area were equally affected by a high prevalence of bladder cancer. A high 4 prevalence of bladder cancer in the BFD-endemic area was noted when compared with the 5 neighboring region and residents of the non-endemic area. However, sporadic cases of bladder 6 cancer were noted in the region neighboring the endemic area. This study also found that the 7 non-BFD-endemic areas, which had a high arsenic content in the well water, did not have a high 8 prevalence of bladder cancer, indicating other possible environmental factors. The histological 9 confirmation of bladder cancer diagnoses is a strength of the study; however, the lack of 10 individual arsenic exposure data is a limitation. 11 Wu et al. (1989) analyzed age-adjusted mortality rates using an ecological study design 12 to determine whether a dose-response relationship exists between ingested arsenic levels and the 13 risk of cancer among residents in the BFD endemic area. The study population consisted of a 14 cohort of individuals from the southwestern coast of Taiwan (27 villages from the townships of 15 Peimen, Hsuechia, Putai, and Ichu and 15 villages from the townships of Yensui and Hsiaying). 16 The arsenic levels in well water for the 42 villages were determined from 1964 to 1966, while 17 mortality and population data were obtained for the years of 1973 to 1986 from the local 18 registration offices and from the Taiwan Provincial Department of Health. Age-adjusted 19 mortality rates from various cancers by gender were calculated using the 1976 world population 20 as the standard population. A significant dose-response relationship was observed between 21 arsenic levels in well water and bladder, kidney, skin, and lung cancers in both males and 22 females. A similar relationship was observed for prostate and liver cancers in males. There was 23 no association for leukemia or cancers of the nasopharynx, esophagus, stomach, colon, and 24 uterine cervix. Strengths of the study include the fact that adjustments were made for age and 25 gender, and that lifestyle, access to medical care, and socioeconomic status were similar among 26 the study groups. The use of mortality data can be considered a weakness of the study, since 27 death certificates may not list all cancers. Additionally, associations observed at the local level 28 may not be accurate at the individual level. 29 The Chen and Wang (1990) ecological study was carried out to examine correlations 30 between the arsenic level in well water and mortality from various malignant neoplasms in 314 31 precincts and townships of Taiwan. The arsenic content of water from 83,656 wells was 32 available from measurements taken in 1974 through 1976. Mortality rates from 1972 to 1983 33 were derived from residents in study precincts and townships who displayed one or more of the 34 21 examined malignant neoplasms. Arsenic content in the water was available at the precinct or 35 township level. A statistically significant association with the arsenic level in well water was 36 observed for cancers of the liver, nasal cavity, lung, skin, bladder, and kidney in both males and 37 females, as well as for prostate cancer in males. These associations remained significant after 37 DRAFT—DO NOT CITE OR QUOTE ------- 1 adjusting for indices of urbanization and industrialization through multiple regression analyses. 2 No significant association was identified for the other 14 cancers examined. The multivariate- 3 adjusted regression coefficient showed an increase in age-adjusted mortality for cancers in males 4 and females for every 100 ppb increase in arsenic level in well water. Coefficients for males and 5 females, respectively, were as follows: 6.8 and 2.0 (liver), 0.7 and 0.4 (nasal cavity), 5.3 and 5.3 6 (lung), 0.9 and 1.0 (skin), 3.9 and 4.2 (bladder), and 1.1 and 1.7 (kidney). Results were 7 unchanged when 170 southwestern townships were included. Strengths of the study were that 8 potential confounders (including socioeconomic differences, i.e., urbanization and 9 industrialization) were controlled for, the study reported ecological correlations between arsenic 10 content in well water and mortality from various cancers, and cancer rates in endemic BFD 11 townships were compared with cancer rates in non-endemic townships of Taiwan. Potential 12 confounders not controlled for were gender, other potential well water exposure contaminants, 13 and individual arsenic exposures that were not available. 14 Using an ecologic investigation, Chen et al. (1992) showed a comparable excess risk of 15 cancer of liver, lung, bladder, and kidney cancers induced by arsenic in drinking water. The 16 study area and population were previously described by Wu et al. (1989). In order to compare 17 the risk of developing various cancers as the result of ingesting inorganic arsenic and to assess 18 the differences in risk between males and females, cancer potency indices were calculated with 19 the Armitage-Doll multistage model using mortality rates among residents of 42 villages in six 20 townships (Peimen, Hsuechia, Putai, Ichu, Yensui, and Hsiaying) located on the southwest coast 21 of Taiwan. Locations selected were considered to be chronic arsenicism endemic areas. Arsenic 22 exposure levels from drinking water in these villages were categorized into four groups: <100 23 ppb (13 villages), 100-299 ppb (8 villages), 300-599 ppb (15 villages), and 600 ppb or greater 24 (6 villages). Based on a total of 898,806 person-years during the study period from 1973 25 through 1986, a significant dose-response relationship was observed between the arsenic level in 26 drinking water and cancer mortality of the liver, lung, bladder, and kidney. The lifetime risk 27 (determined using the Armitage-Doll model) of developing cancer due to an intake of 10 ug/kg- 28 day of arsenic was estimated to be 4.3 x 10-3 (liver), 1.2 x 10-2 (lung), 1.2 x 10-2 (bladder), and 29 4.2 x 10-3 (kidney) for males and 3.6 x 10-3 (liver), 1.3 x 10-2 (lung), 1.7 x 10-2 (bladder), and 30 4.8 x 10-3 (kidney) for females. Strengths include that potential confounders including age, 31 gender, access to medical care, socioeconomic status, and lifestyle were all controlled for during 32 the analysis, and that villages shared similar socioeconomic status, living environments, 33 lifestyles, dietary patterns, and medical facilities. A weakness of the study is the assumption that 34 an individual's arsenic intake remained constant from birth to the end of the follow-up period; 35 this flaw possibly led to the underestimation of risk. Additional weaknesses include that the 36 Armitage-Doll model constrains risk estimates to be monotonically increasing function of age, 3 8 DRAFT—DO NOT CITE OR QUOTE ------- 1 that dietary sources of arsenic were not quantified, and that age stratification was for under 30, 2 over 70, and 20-year strata. 3 To determine whether a dose-response relationship exists between ingested inorganic 4 arsenic and cancer, Chiou et al. (1995) used a cohort study with a total of 263 BFD patients and 5 2293 healthy residents in the arseniasis-endemic area of southwestern coast of Taiwan (Peimen, 6 Hsuechia, Putai, and Ichu townships). Participants were followed for an average of 4.97 years 7 (range: 0.05-7.69 years). Data concerning the consumption of artesian well water containing 8 high levels of arsenic, sociodemographic characteristics, lifestyle and dietary habits, and cancer 9 histories were obtained through a standardized interview. Internal cancers were determined via 10 health examinations, personal interviews, household registration data checks, and Taiwan's 11 national death certification and cancer registry databases. Concentrations used in the assessment 12 were < 50 ppb, 50-70 ppb, 71+ ppb, and unknown. Disregarding the unknown category, a dose- 13 response relationship was observed between the long-term arsenic exposure from drinking 14 artesian well water and the incidence of lung cancer, bladder cancer, and cancers of all sites 15 combined after adjusting for age, sex, and cigarette smoking through a Cox's proportional 16 hazards regression analysis. BFD patients had a significantly increased incidence of bladder 17 cancer and for all sites combined after adjusting for age, gender, smoking history, and 18 cumulative arsenic exposure (CAE). Strengths include that the analysis adjusted for BFD status, 19 age, gender, and smoking; incidence data were reported; and the results of the study showed a 20 significant dose-response relationship. A weakness of the study is that well water artesian 21 arsenic concentrations were unknown for some study subjects; consequently, this was a 22 significant confounder. 23 To further evaluate the association between arsenic exposure in drinking water and 24 urinary cancers of various cell types, Guo et al. (1997) conducted an ecological study 25 encompassing 243 townships using Taiwanese National Cancer Registry data of patients 26 diagnosed with cancer between 1980 and 1987. Wells with known arsenic concentrations in 27 each township were used to separate people into the following exposures: <50 ppb, 50-80 ppb, 28 90-160 ppb, 170-320 ppb, 330-640 ppb, and >640 ppb. The effects of urbanization and 29 smoking were evaluated by an urbanization index based on 19 socioeconomic factors shown to 30 be good indicators of urbanization and the number of cigarettes sold per capita. For both 31 genders, Guo et al. observed associations between high arsenic levels in drinking water and 32 transitional cell carcinomas (bladder, kidney, ureter, and all urethral cancers combined). 33 Positive associations between the proportion of wells with arsenic levels above 640 ppb and the 34 incidence of transitional cell carcinomas of the bladder, kidney, ureter, and all urethral cancers 35 combined in both genders were identified after the model was adjusted for urbanization and age. 36 Arsenic exposure in males was associated with adenocarcinomas of the bladder, but not in 37 squamous cell carcinomas of the bladder or renal cell carcinomas or nephroblastomas of the 39 DRAFT—DO NOT CITE OR QUOTE ------- 1 kidney. Males also exhibited a positive association between the urbanization index and 2 transitional cell carcinomas of the ureter. The results support the case that the carcinogenicity of 3 arsenic may be cell-type specific. Analyses were adjusted for age, gender, urbanization, and 4 smoking; however, the ecologic study design was a limitation. 5 Tsai et al. (1999) conducted a cross-sectional study in BFD-endemic areas in the 6 southwest coastal region of Taiwan (Peimen, Hsuechia, Putai, and Ichu townships) to analyze 7 mortality from neglected cancers related to artesian well water containing high levels of arsenic. 8 The median artesian well water arsenic content was 780 ppb (range: 250-1,140 ppb). Local 9 endemic area residents' daily ingestion of arsenic was estimated to be < 1 mg. SMRs were 10 calculated for cancer diseases, by gender, during the period from 1971 to 1994. These SMRs 11 were compared to the local reference group (Chiayi-Tainan County population) and a national 12 reference group (Taiwanese population). The comparisons revealed significant differences 13 between SMRs of the three groups. Mortality increases (p < 0.05) were found in males and 14 females, respectively, for all cancers (SMR=2.19, 95% CI: 2.11-2.28; SMR=2.40, 95% CI: 15 2.30-2.51) when compared to the local reference population. Additionally, the following other 16 cancers showed mortality increases in males and females, respectively, when compared to the 17 local reference population: bladder (SMR=8.92, 95% CI: 7.96-9.96; SMR=14.07, 95% CI: 18 12.51-15.78); kidney (SMR=6.76, 95% CI: 5.46-8.27; SMR=8.89, 95% CI: 7.42-10.57); skin, 19 lung, nasal-cavity, bone, and liver (SMR=1.83, 95% CI: 1.69-1.98; SMR=1.88, 95% CI: 1.64- 20 2.14); and larynx, stomach, colon, intestine, rectum, lymphoma, and prostate cancer in males 21 only (SMR=2.52, 95% CI: 1.86-3.34). When compared to the national reference population, 22 significantly increased (p < 0.05) mortality was found in males and females, respectively, for all 23 cancers (SMR=1.94, 95% CI: 1.87-2.01; SMR=2.05, 95% CI: 1.96-2.14) and for the other 24 following cancers: bladder (SMR=10.50, 95% CI: 9.37-11.73; SMR=17.65, 95% CI: 5.70- 25 19.79) and lung (SMR=2.64, 95% CI: 2.45-2.84; SMR=3.50, 95% CI: 3.19-3.84). The results 26 of the Tsai et al. (1999) investigation indicate that the hazardous effect of arsenic may be 27 systemic. Key strengths of the study are that the exposed group and local reference group had 28 similar lifestyle factors; all cancers were pathologically confirmed; and the analysis controlled 29 for gender. Weaknesses of the study are that death certificates indicated only one underlying 30 cause of death (not multiple causes), resulting in possible distortion of association between 31 exposure and disease; individual exposure data were not provided; and certain potential 32 confounders were not controlled for (age, smoking history, alcohol consumption, and 33 occupational exposures). 34 The Morales et al. (2000) ecological investigation re-analyzed data originally reported by 35 Chen et al. (1988a, 1992) and Wu et al. (1989) from 42 villages in the arseniasis-endemic region 36 of southwestern Taiwan by considering the number of liver, lung, and bladder cancer deaths. 37 Morales et al. (2000) used a generalized linear model (i.e., Poisson distribution) and the 40 DRAFT—DO NOT CITE OR QUOTE ------- 1 multistage-Weibull models to determine lifetime cancer risk estimates. Liver, lung, and bladder 2 cancer mortality data were collected from death certificates of residents in 42 villages during 3 1973 through 1986. Drinking water samples had been collected from wells in the 42 villages 4 between 1964 and 1966. SMRs were used to summarize the observed patterns of mortality in the 5 collected data. Morales et al. (2000) selected two comparison populations (the Taiwanese 6 population as a whole and a population from a southwestern region of Taiwan) to account for 7 urban versus non-urban populations differences. Although a non-significant trend was observed 8 in the combined cancer analyses with respect to age, there was no observed tendency in liver, 9 lung, or bladder SMRs with respect to age. This suggests that there is no age dependency on the 10 risk ratio. Liver cancer mortality was higher than expected, although there was no strong 11 exposure-response relationship found. The Morales et al. (2000) investigation results showed 12 that exposure-response assessments were highly dependent on the choice of the analysis model 13 and whether or not a comparison population is used in the analysis. One possible explanation for 14 this observation is the inherent uncertainty associated with the limitations of an ecological study 15 design. Depending on the model used and the comparison population used in the analysis, the 16 effective dose at the 1% level (ED01) estimates ranged from 21 to 633 ppb for male bladder 17 cancer, and from 17 to 365 ppb for female bladder cancer. The lung cancer risk for males was 18 found to be slightly higher than the bladder cancer risk, with ED01 estimates ranging from 10 to 19 364 ppb. The risk for female cancer tended to be higher than that of males for each cancer type. 20 For lung cancer, female ED01 estimates ranged from 8 to 396 ppb. 21 In summary, the Morales et al. (2000) analysis of the Taiwan data suggests that excessive 22 cancer mortality may occur in many populations where the drinking water standard for arsenic is 23 set at 50 ppb, the drinking water standard for arsenic in the United States at the time of 24 publication. A strength of the study was that person-years at risk (PYR) were stratified by 5- 25 year age groups, gender, and median arsenic level for each village. Weaknesses include the 26 ecological study design (i.e., there were no individual monitoring data and individual exposures 27 were not available) and the fact that potential confounders such as smoking, dietary arsenic, and 28 the use of bottled water (U.S. population) were not controlled for in the analysis. 29 Between 1991 and 1994, Chiou et al. (2001) recruited a cohort of 8,102 residents aged 40 30 years or older from four townships (18 villages) in northeastern Taiwan (4 villages in Chiaohsi, 7 31 in Chuangwei, 3 in Wuchih, and 4 in Tungshan) and followed it until the end of 1996. The study 32 examined the risk of transitional cell carcinoma in relation to ingested arsenic. The Chiou et al. 33 (2001) findings were consistent with previously reported findings from the arsenic-endemic area 34 of southwestern Taiwan. Based on the arsenic concentration in well water, each study subject's 35 individual exposure to inorganic arsenic was estimated. Information concerning the duration of 36 consumption of the well water was obtained through standardized questionnaire interviews. 37 Urinary tract cancers were identified by follow-up interviews, community hospital records, the 41 DRAFT—DO NOT CITE OR QUOTE ------- 1 Taiwanese national death certification profile, and the cancer registry profile. A significantly 2 increased incidence of urinary tract cancers for the study cohort was observed (standardized 3 incidence ratio [SIR]=2.05; 95% CI: 1.22-3.24) when compared to the general population in 4 Taiwan. In addition, a dose-response relationship was observed between the risk of cancers of 5 the urinary organs, especially transitional cell carcinoma, and indices of arsenic exposure after 6 adjusting for age, sex, and cigarette smoking. The relative risks (RR) of developing transitional 7 cell carcinoma were 1.9, 8.2, and 15.3 for arsenic concentrations of 10.1-50.0 ppb, 50.1- 8 100.0 ppb, and >100.0 ppb, respectively, compared with the referent level of < 10.0 ppb. No 9 association was observed for the duration of well water drinking (<40 years compared to 10 > 40 years). The findings of this study suggest that arsenic ingestion may increase the risk of 11 urinary tract cancer at levels around 50 ppb. Strengths include adjustments for potential 12 confounders (age, gender, smoking history), individual arsenic exposure estimates, and a dose- 13 response relationship even with the low levels of arsenic. Weaknesses include possible 14 diagnostic bias as the result of medical data collection from various community hospitals and 15 recall bias from self-reported information. The short duration of follow-up also is a limitation 16 because it impacted: (1) the number of person-years of observation; and (2) only a few cases 17 were recorded. This study also has an apparent supralinear curve, which is likely due to dose 18 misclassification in the low-dose individuals. If food arsenic concentrations (estimated in NRC, 19 2001, to be approximately 50 ug/day) were included, the curve might not be supralinear. 20 Guo et al. (2001) conducted an ecological investigation of the 243 townships from their 21 1997 publication; however, this investigation focused on arsenic exposure through drinking 22 water and the potential association with skin cancers. Data regarding arsenic levels in drinking 23 water were available from the previous investigation, and cases of skin cancer were identified 24 using the Taiwanese National Cancer Registry. Data were analyzed with regression models 25 using multiple variables to describe exposures, including arsenic. To adjust for potential 26 confounding variables, an urbanization index based on 19 socioeconomic factors shown to be 27 good indicators of urbanization was developed. A total of 2,369 individuals with skin cancer 28 (954 females and 1,415 males) were registered with the Cancer Registry between January 1980 29 and December 1989. After age and urbanization adjustment, arsenic levels above 640 ppb 30 showed a statistically significant (p < 0.01) association with the incidence of basal cell 31 carcinoma (BCC) in males. Exposed females also exhibited an increased incidence in skin 32 cancer rates; however, this increase did not reach statistical significance (p = 0.20). For 33 squamous cell carcinomas (SCC), a significant (p < 0.01), positive association was found for 34 males exposed to 170-320 ppb and >640 ppb. However, a statistically significant (p < 0.01) 35 negative association was found for males exposed to 330-640 ppb. For females, a similar 36 statistically significant (p < 0.01) positive association was observed at >640 ppb, while a 37 statistically significant (p < 0.05) negative association was observed in 330-640 ppb females. 42 DRAFT—DO NOT CITE OR QUOTE ------- 1 For melanomas, no significant associations were identified in females or males at any exposure. 2 The results of the investigation suggest that skin cancers are cell-type-specific, as previously was 3 demonstrated for urinary tract cancers (Guo et al., 1997). Strengths of the study include that 4 cases were identified from a government operated National Cancer Registration Program, 5 pathological classifications were determined by board-certified pathologists, and potential 6 confounders (gender and age) were adjusted in the analysis. A limitation of the study is the 7 ecological study design. 8 Studies on cancers of the urinary system and skin showed that arsenic's carcinogenic 9 effect was cell-type-specific (Guo et al., 1997, 2001). Guo (2003) conducted an ecological 10 investigation in 243 townships in Taiwan, previously used in the Guo et al. (1997, 2001) 11 investigations for urinary and skin cancers, to determine if a similar relationship could be 12 identified for liver cancer. Many previous epidemiologic studies did not provide data on 13 pathological diagnoses; therefore, there was no information to support the hypothesis that 14 hepatocellular carcinoma (HCC) or cholangiocarcinoma of the liver were not associated with 15 arsenic ingestion. Liver cancers were identified through the Taiwanese National Cancer 16 Registry. The distribution of cancer cell-types between an arseniasis-endemic area and a 17 township outside the arseniasis area were compared. Between January 1980 and December 18 1999, 32,034 men and 8,798 women living in the study townships were diagnosed with liver 19 cancer. The distribution of two cancer cell-types (HCC and cholangiocarcinoma) did not appear 20 to be different between the arseniasis-endemic and non-arseniasis-endemic areas, and an 21 association between HCC and arsenic ingestion was not observed. The remainder of the cell- 22 types did not have enough cases to provide stable estimates. Identified strengths of the study 23 include the following: cases were identified from the government-operated National Cancer 24 Registration Program; pathological classifications were determined by board-certified 25 pathologists; and analyses were adjusted for gender and age. Weaknesses include the limitations 26 of ecological study design (no monitoring data were presented). 27 A cohort investigation of residents from two arsenic endemic areas were followed for 8 28 years by Chen et al. (2004a) to investigate the dose-response relationship between arsenic 29 exposure and lung cancer, as well as how cigarette smoking influenced the relationship between 30 arsenic and lung cancer. Arsenic-endemic areas included the southwestern coast (Peimen, 31 Hsuechia, Putai, and Ichu; n = 2,503) and the northeastern coast (Tungshan, Chuangwei, 32 Chiaohsi, and Wuchieh; n = 8,088) of Taiwan. The amount of arsenic in well water from these 33 areas ranged from less than 0.15 ppb to more than 3,000 ppb. The Taiwanese National Cancer 34 Registry was used to identify new cases of lung cancer diagnosed between January 1, 1985, and 35 December 31, 2000. For each participant, follow-up person-years were calculated using the time 36 from the initial interview date to the date of diagnosis, death, or December 31, 2000, whichever 37 came first. Arsenic concentration was arbitrarily divided into five categories: <10 ppb (referent), 43 DRAFT—DO NOT CITE OR QUOTE ------- 1 10-99.9 ppb, 100-299.9 ppb, 300-699.9 ppb, and >700 ppb. Smoking histories were obtained 2 from interviews. Cox proportional hazards regression models were used to estimate RR and 3 95% CI. The final model was adjusted for age, gender, years of schooling, study cohort (BFD 4 cases and matched controls of the southwestern coast, residents along the arseniasis- 5 hyperendemic southwestern coast villages, and residents living in the northeastern coastal 6 Lanyang Basin), smoking status, and alcohol consumption. During the study follow-up period, 7 there were 139 lung cancers diagnosed, resulting in an incidence rate of 165.9 per 100,000 8 person-years. When the highest level of arsenic exposure was compared to the lowest, the RR 9 was 3.29 (95% CI: 1.60-6.78). The risk of lung cancer was four times higher for past and 10 current smokers compared to non-smokers. A synergistic effect of ingested arsenic and cigarette 11 smoking on lung cancer was noted, with synergy indices ranging from 1.62 to 2.52. Strengths of 12 the study include controlling for confounders (age, gender, education, smoking history, and 13 alcohol consumption), having a long follow-up period, using a national computerized cancer 14 case registry, and pathologically confirming all lung cancer cases. Weaknesses include the lack 15 of historical monitoring data and possible misclassification bias (exposure measurements were 16 based on one survey). 17 Chiu et al. (2004), using a cohort study design, examined whether liver cancer mortality 18 rates were altered after the consumption of high-arsenic artesian well water ceased. SMRs for 19 liver cancer were calculated for the BFD-endemic area of the southwest coast of Taiwan (i.e., 20 Peimen, Hsuechia, Putai, and Ichu townships) for the years 1971 through 2000. Median well 21 water arsenic concentrations in the early 1960s were 780 ppb. Temporal changes in the SMRs 22 were monitored using cumulative-sum techniques and were reported for 3-year intervals between 23 1971 and 2000. Study results showed that female mortality from liver cancer started declining 9 24 years after consumption of high-arsenic artesian well water stopped. The SMR for liver cancer 25 in females was 2.041 during the 1983-1985 period (peak) and was 1.137 during 1998 through 26 2000. Data in males, however, showed fluctuations in liver cancer mortality rates. The SMR for 27 liver cancer in males from 1989 to 1991 was 1.868 and 1.242 during 1998 to 2000. Based on 28 analyses by Chiu et al. (2004), it was determined that the relationship between arsenic exposure 29 and liver cancer mortality was possibly causal in females, but not in males. Strengths of the 30 study are: (1) residents in the study area were similar in terms of socioeconomic status, living 31 environments, lifestyles, dietary patterns, and availability of health service facilities; and (2) the 32 study used an accurate death registration system. Weaknesses include the limitations of the 33 mortality data. 34 To obtain data on the potential dose-response relationship between lung cancer and the 35 level of arsenic in drinking water, Guo (2004) conducted an ecological investigation in 10 36 townships (138 villages) in Taiwan. Measurements of arsenic levels in drinking water were 37 available for the 138 villages from a census survey conducted by the Taiwanese government. 44 DRAFT—DO NOT CITE OR QUOTE ------- 1 Death certificates dated between January 1, 1971, and December 31, 1990, were reviewed, and 2 673 males and 405 females were identified as dying from lung cancer. Multivariate regression 3 models were applied to assess the relationship between arsenic levels in drinking water and lung 4 cancer mortality. After adjusting for age, arsenic levels above 640 ppb were associated with a 5 significant increase in lung cancer mortality for both genders; however, no significant effect was 6 observed at lower arsenic exposure levels. Regression analyses and stratified analyses 7 confirmed a dose-response relationship at >640 ppb. Guo (2004) noted that the results of this 8 investigation show a carcinogenic effect of high arsenic levels in drinking water on the lung. 9 Guo (2004), however, recommended that further studies with exposure data on individuals were 10 warranted to confirm these findings. As a result of the study's ecologic design, the association 11 observed on an aggregate level may not necessarily represent the association that exists at an 12 individual level. In addition, the study design may have contributed to biases introduced by the 13 effects of population mobility. Strengths of the study include that analyses adjusted for gender 14 and age, and cases were ascertained using information from household registry offices in each 15 township. Weaknesses of the investigation include the inherent limitations of ecological studies 16 and the fact that smoking was not controlled for in the analysis. 17 In a cross-sectional study, Yang et al. (2004) examined whether kidney cancer mortality 18 decreased in the southwest coast of Taiwan (Peimen, Hsuechia, Puta, and Ichu townships) after 19 the elimination of arsenic exposure in the 1970s. SMRs for kidney cancer were calculated for 20 the BFD-endemic area for the years 1971 through 2000. There were 308 kidney cancer deaths 21 (135 men and 173 women) in the BFD-endemic area between 1971 and 2000. The means of the 22 3-year SMRs for female and male kidney cancer were significantly higher than for Taiwan as a 23 whole. Time series plots for male SMRs showed decreasing mortality rates. The estimated 24 slope for male SMRs (rate of decrease per year) in the linear time trend analysis was -15.13 25 (p < 0.01). The time series plot for female SMRs also showed decreasing mortality rates. 26 Kidney cancer mortality rates among residents in the BFD-endemic area decreased after removal 27 of the arsenic source through tap water implementation. SMRs decreased each year, on average, 28 from 1971 to 2000 (p < 0.01). Study strengths include the adjustment of potential confounders 29 (gender and age); mandatory registering of all births, deaths, marriages, divorces, and migration 30 issues with the Household Registration Office in Taiwan, making it an accurate data source; and 31 a comparable study population (i.e., residents likely had similar socioeconomic status, living 32 environments, lifestyles, dietary patterns; they worked in farming, fisheries, or salt production) 33 that had comparable access to medical care (i.e., all kidney cancer cases likely had similar access 34 to medical care). Weaknesses of the study include cross-sectional mortality limitations and not 35 adequately controlling for smoking histories. 36 Tsai et al. (2005) used a cross-sectional study to compare primary urethral carcinomas 37 from the BFD-endemic area of Taiwan with those in the United States and explore the potential 45 DRAFT—DO NOT CITE OR QUOTE ------- 1 influence of chronic arsenic exposure. Cases were identified by the only medical center near the 2 BFD area. There were 21 pathologically proven primary urethral carcinomas diagnosed (7 3 females and 14 males) between 1988 and 2001. Seven of 14 male patients had reported an 4 average of 23 years of chronic arsenic exposure from drinking water. Tsai et al. (2005) 5 compared these cases to cases identified in three U.S. cancer centers (MD Anderson, Memorial 6 Sloan-Kettering, and Barbara Ann Karmanos; n = 79 females, n = 80 males), and analyzed for a 7 relationship with chronic arsenic exposure. In comparison to the three U.S. cancer centers, there 8 was a higher frequency of bulbomembranous adenocarcinoma (43% vs. 18%, 2%, and 0%, 9 respectively, p < 0.0001). In those with chronic arsenic exposure, there was an even greater 10 association with bulbomembranous adenocarcinoma compared to those without chronic arsenic 11 exposure (73% vs. 14%, p=0.031). Based on these results, Tsai et al. (2005) concluded that the 12 BFD-endemic area in Taiwan had a high frequency of bulbomembranous urethral 13 adenocarcinoma, which may be associated with chronic arsenic exposure. A strength of the 14 study is that cases were pathologically confirmed. The small number of cases and the lack of 15 arsenic exposure information are study weaknesses. 16 The objective of the Yang et al. (2005) cross-sectional study was to determine whether 17 bladder cancer mortality decreased after the implementation of the tap water system and the 18 subsequent elimination of arsenic exposure. SMRs for bladder cancer were calculated for the 19 BFD-endemic area for the years 1971-2000. The study showed that bladder cancer mortality 20 decreased gradually after the instillation of the tap water system, thereby eliminating exposure to 21 arsenic through artesian well water, (1971, male SMR=10.25, female SMR=14.89; 2000, male 22 SMR=2.15, female SMR=7.63). Strengths include similar access to medical care for bladder 23 cancer, the adjustment for age and gender, and the mandatory registering of all births, deaths, 24 marriages, divorces, and migration issues to the Household Registration Office in Taiwan, 25 making it an accurate data source. Limitations of the study include the cross-sectional mortality 26 study design and smoking history confounding. 4.1.2. Japan 27 Tsuda et al. (1995) used a cohort study to investigate the long-term effect of ingesting 28 arsenic in drinking water for an estimated exposure period of 5 years (1955-1959). Four 29 hundred and fifty-four residents identified in 1959 as living in an arsenic-polluted area of Niigata 30 Prefecture, Japan, were followed until 1992. The mortality of these residents between October 1, 31 1959, and February 29, 1992, was examined using death certificates. These individuals used 32 arsenic-contaminated well water, and none worked at a nearby factory that was the source of the 33 water contamination. Death certificates for the people who died between 1959 and 1992 were 34 examined and a total of 113 of the 454 residents were estimated to have consumed well water 35 containing a high concentration of arsenic (>1,000 ppb). The SMRs of these 113 residents were 36 15.69 for lung cancer (95% CI: 7.38-31.02) and 31.18 for urinary tract cancer (95% CI: 8.62- 46 DRAFT—DO NOT CITE OR QUOTE ------- 1 91.75). Cox's proportional hazard analyses demonstrated that the hazard ratios of the highest 2 exposure level group (> 1,000 ppb) versus the background exposure level group (1.0 ppb) were 3 1.74 (95% CI: 1.10-2.74) for all deaths, 1972.16 (95% CI: 4.34-895,385.11) for lung cancer, 4 and 4.82 (95% CI: 2.09-11.14) for all cancers. The study also analyzed skin signs of chronic 5 arsenicism, and results indicated that they were useful risk indicators for subsequent cancer 6 development. These results indicate a relationship between well water arsenic exposure and lung 7 and urinary tract cancers. The study also showed that arsenic-induced cancer could develop 8 years following the end of arsenic exposure. For lung cancer, there was evidence of synergistic 9 effects between arsenic exposure and smoking history. Strengths of this study include data on 10 smoking history, age, and gender, and an examination of the cohort by three arsenic exposure 11 categories. Weaknesses, however, include the lack of detailed arsenic intake information, a 12 small study population, as well as possible misclassification and recall bias pertaining to 13 smoking history. 4.1.3. South America 14 Hopenhayn-Rich et al. (1996a) used an ecological study design to investigate bladder 15 cancer mortality for the years 1986 through 1991 in the province of Cordoba, Argentina, using 16 rates for all of Argentina as the standard for comparison. The study compiled arsenic 17 measurements from a major water survey performed more than 50 years earlier. Using these 18 earlier arsenic data, a crude estimate of exposure was made. The data were matched to the 19 population listings from the national census bureau. This study grouped counties into three 20 defined arsenic exposure categories: low, medium, and high (groups were defined based on the 21 location of counties and the concentrations were only provided for the high group, which had a 22 mean arsenic level of 178 ppb). In the absence of smoking data for each county, mortality from 23 chronic obstructive pulmonary disease (COPD) was used as a surrogate. SMRs for bladder 24 cancer were higher in counties with known elevated levels of arsenic exposure through drinking 25 water. The SMRs (95% CI) for corresponding arsenic exposure categories were 0.80 (0.66- 26 0.96), 1.42(1.14-1.74), and 2.14 (1.78-2.53) for males, and 1.21 (0.85-1.64), 1.58(1.01-2.35), 27 and 1.82 (1.19-2.64) for females, respectively. Significant trends were noted in both males and 28 females. 29 Results of this study showed a dose-response relationship between arsenic exposure from 30 drinking water and bladder cancer in spite of the limitations inherent from the ecologic design. 31 Argentina has one of the world's highest rates of per capita beef consumption. The high-arsenic 32 region of Cordoba is an important agricultural and beef-producing area, and animal protein is 33 considered to be one of the basic foods of the population. This is important because protein 34 deficiency in the Taiwanese population has been suggested to diminish their capacity to detoxify 35 arsenic. The similar findings between the two populations, regardless of genetic and dietary 36 differences, strengthens the link between arsenic exposure and bladder cancer. Strengths of the 47 DRAFT—DO NOT CITE OR QUOTE ------- 1 study include the adjustment for age and gender, the use of stomach cancer as a non-arsenic- 2 induced comparison, and that the analysis was restricted to rural counties to limit confounders. 3 The lack of individual smoking history (mortality from COPD was used as a surrogate for 4 smoking), the lack of arsenic measurements in low and medium groups, and the lack of 5 individual arsenic exposure data (ecological study) are important potential weaknesses of this 6 study. 7 To investigate dose-response relationships between arsenic exposure from drinking water 8 and cancer mortality, Hopenhayn-Rich et al. (1998) conducted an ecological study in Cordoba, 9 Argentina. Cancer mortality from the lung, kidney, liver, and skin during the 1986-1991 period 10 in 26 counties of Cordoba were studied. This investigation expanded the analysis of the authors' 11 previous study (Hopenhayn-Rich et al., 1996a), which only examined bladder cancer in Cordoba. 12 Counties were grouped into low, medium, and high arsenic exposure categories based on arsenic 13 exposure data taken from Hopenhayn-Rich et al. (1996a). In the absence of smoking data for 14 each county, mortality from COPD was used as a surrogate. SMRs were calculated using all of 15 Argentina as the reference population. Hopenhayn-Rich et al. (1998) found increasing trends for 16 kidney and lung cancer mortality with increasing arsenic exposure (i.e., low, medium, high) as 17 follows: male kidney cancer SMRs=0.87 (95% CI: 0.66-1.10), 1.33 (95% CI: 1.02-1.68), and 18 1.57 (95% CI: 1.17-2.04); female kidney cancer SMRs=1.00 (95% CI: 0.71-1.37), 1.36 (95% CI: 19 0.94-1.89), and 1.81 (95% CI: 1.19-2.64); male lung cancer SMRs=0.92 (95% CI: 0.85-0.98), 20 1.54 (95% CI: 1.44-1.64), and 1.77 (95% CI: 1.63-1.90); and female lung cancer SMRs=1.24 21 (95% CI: 1.06-1.42), 1.34 (95% CI: 1.12-1.58), and 2.16 (95% CI: 1.83-2.52), respectively 22 (p < 0.001 in trend test). These findings were similar to the previously reported bladder cancer 23 results. Additionally, the Hopenhayn-Rich et al. (1998) study showed a weakly positive trend 24 for liver cancer, with SMRs being significantly increased even in the lowest exposure category. 25 Skin cancer mortality was elevated only for females in the highest arsenic exposure group, while 26 males showed an increase in mortality only in the lowest exposure group. The results add to the 27 evidence that arsenic ingestion through drinking water increases the risk of lung and kidney 28 cancers. The association between arsenic and mortality from liver and skin cancers was not as 29 clear. Risk analyses were restricted to rural Cordoba counties to limit confounders and to 30 account for cancer diagnosis and detection bias. Strengths and weaknesses are the same as those 31 observed for Hopenhayn-Rich et al. (1996a). 32 Smith et al. (1998), using an ecological design, studied cancer mortality in a population 33 of approximately 400,000 people exposed to high arsenic levels in drinking water in past years in 34 Region II of northern Chile. Arsenic concentrations in drinking water from 1950 to 1996 were 35 available. The population-weighted average arsenic levels reached 570 ppb between 1955 and 36 1969, but decreased to less than 100 ppb by 1980. SMRs were calculated for the years 1989 to 37 1993, and increased SMRs were identified for bladder, kidney, lung, and skin cancers. Bladder 48 DRAFT—DO NOT CITE OR QUOTE ------- 1 cancer mortality was the most elevated (female SMR=8.2, 95% CI: 6.3-10.5; male SMR=6.0, 2 95% CI: 4.8-7.4). Lung cancer mortality was likewise significantly elevated (female SMR=3.1, 3 95% CI: 2.7-3.7; male SMR=3.8, 95% CI: 3.5-4.1). Smoking survey data and mortality rates 4 from COPD provided evidence that smoking did not contribute to the increased mortality from 5 these cancers. These results provide additional evidence that ingestion of inorganic arsenic in 6 drinking water can lead to increases in cancers of the bladder and lung. Smith et al. (1998) 7 estimated that approximately 7% of all deaths in individuals more than 30 years old might be 8 attributable to arsenic exposure. Strengths of the study are the large size of the study population, 9 the adjustment of SMRs by age and gender, and the use of Chilean national data for comparison. 10 Weaknesses include that arsenic levels were not available at the individual source level, dose- 11 response information was not provided, and only limited individual smoking history information 12 was available (i.e., participants were asked if they had smoked cigarettes over a 1-month period 13 in 1990). 14 In a case-control study, Ferreccio et al. (2000) investigated the association between lung 15 cancer and arsenic in drinking water by comparing patients diagnosed with lung cancer (1994- 16 1996; 152 cases) with frequency-matched hospital controls (419 controls). Using a full-logistic 17 regression model, a clear trend in lung cancer ORs was observed with increasing concentration 18 of arsenic in drinking water: 10-29 ppb arsenic, OR: 1.6 (95% CI: 0.5-5.3), 30-49 ppb arsenic, 19 OR: 3.9 (95% CL1.2-12.3), 50-199 ppb arsenic, OR: 5.2 (95% CI: 2.3-11.7), and 200-400 ppb, 20 OR: 8.9 (95% CI: 4.0-19.6). Evidence of synergistic effects between arsenic in drinking water 21 and cigarette smoking history was much greater than expected, as the OR for lung cancer was 22 32.0 (95% CI: 7.2-198.0) among smokers exposed to more than 200 ppb. In comparison, an OR 23 of 8.0 was observed for those who never smoked but were in the highest arsenic category, and an 24 OR of 6.1 was observed for smokers in the lowest arsenic category. Based on these results, the 25 effect was considered synergistic because an OR of 13.1 was expected if the effect was additive. 26 This study provided strong evidence that ingestion of inorganic arsenic through drinking water 27 is associated with lung cancer. ORs for the full-analysis model were adjusted for age, gender, 28 cumulative lifetime cigarette smoking, working in copper smelting, and socioeconomic status; 29 this is considered a study strength. The fact that more controls were obtained from Antofagasta 30 than from the lower-exposure cities of Arica and Iquique, which could lead to an improper 31 (lower) estimation of risk, is considered a study limitation. 32 Bates et al. (2004) recognized that epidemiologic studies had found an association 33 between increased bladder cancer risk and high levels of arsenic in drinking water; however, 34 little information was found concerning cancer risks at lower concentrations. It also was 35 recognized that ecologic studies in Argentina had found increased bladder cancer mortality in 36 Cordoba Province, where some wells were contaminated with moderate arsenic concentrations. 37 Therefore, Bates et al. (2004) decided to use a population-based bladder cancer case-control 49 DRAFT—DO NOT CITE OR QUOTE ------- 1 study during 1996-2000 in two Cordoba counties and recruited 114 case-control pairs, matched 2 by age, sex, and county of residence over the past 40 years. Three arsenic exposure metrics 3 based on questionnaire and water sampling data were used: average arsenic concentration in 4 domestic water, arsenic concentration adjusted to fluid intake, and reported years of well water 5 consumption. Statistical analyses showed no evidence of an association of bladder cancer with 6 arsenic exposure estimates based on arsenic concentrations in drinking water. Additional time- 7 trend analyses, however, did suggest that the use of arsenic-contaminated well water at least 50 8 years prior to the study was associated with increased bladder cancer risk. This positive 9 association was limited to people who had ever smoked (OR=2.5, 95% CI: 1.1-5.5 for the time 10 period 51-70 years before the study interview). Bates et al. (2004) suggested that it could not be 11 excluded that these associations were based on chance. 12 The results of this study suggest a decreased bladder cancer risk for arsenic exposure than 13 had been predicted from other studies. The results of the Bates et al. (2004) study did add to the 14 evidence that the latency for arsenic-induced bladder cancers may be longer than previously 15 thought and that increased lengths of follow-up for studies may be required to accurately 16 measure the induced risk. Strengths include that potential confounders (age, gender, smoking 17 history, and residence county) were controlled for in the analysis. However, weaknesses related 18 to the lack of a cancer registry, arsenic exposure misclassification, and recall and selection bias 19 exist. Selection bias may have occurred, as the controls had a significantly lower rate of 20 participation than cases. Additional selection bias may have occurred with the selection of cases 21 from the tumor registry. An additional weakness is that other harmful exposures (including 22 arsenic exposure through food) were not measured. 23 Using a cohort study design, Smith et al. (2006) investigated lung cancer, bronchiectasis, 24 and COPD mortality rates in Antofagasta, Chile, from 1989 through 2000 and compared these 25 rates to the rest of Chile. Study subjects (30-49 years old at time of death) were selected 26 primarily from those born during or just prior to the peak in the arsenic exposure period. Results 27 show a lung cancer SMR of 7.0 (95% CI: 5.4-8.9, p < 0.001) for the cohort born just before the 28 peak exposure period (i.e., from 1950 through 1957), and, therefore, were exposed to arsenic 29 during their childhood. For those cases born between 1958 and 1971 (i.e., the high-exposure 30 period), a lung cancer SMR of 6.1 (95% CI: 3.5-9.9, p < 0.001) was estimated; this group was 31 probable exposed to arsenic in utero and early childhood. These findings suggest that exposure 32 to arsenic in drinking water during early childhood or in utero has pronounced pulmonary effects 33 greatly increasing subsequent mortality in young adults from malignant lung disease. The study 34 concluded that the observed effects are most probably due to arsenic in water, even though 35 possible effect-dilution occurred as the result of in-migration of those from other regions of 36 Chile. A strength of the study was the extensive documentation of drinking water arsenic levels 37 in the Antofagasta water system. Weaknesses include that place of residence was determined 50 DRAFT—DO NOT CITE OR QUOTE ------- 1 from the death certificates, which relates to residence at the time of death, and the reliance on 2 death certificates (potential diagnostic bias). Smoking, although considered unlikely by Smith et 3 al. (2006), is a potential confounder for this study. 4 Marshall et al. (2007) conducted an ecological study to investigate lung and bladder 5 cancer mortality from 1950 to 2000 in a region of Chile where drinking water was contaminated 6 with arsenic (Region II), and in another region of Chile where arsenic was not an issue (Region 7 V). Elevated arsenic exposure through drinking water began in Region II in 1958 and continued 8 into the early 1970s. Mortality data tapes and mortality data from death certificates for the two 9 regions for 1950 to 1970 identified 307,541 deaths from the two regions for 1971 to 2000. 10 Poisson regression models were used to compare Region II with Region V by identifying time 11 trends in rate ratios of mortality from lung and bladder cancers. Lung and bladder cancer 12 mortality rate ratios for Region II compared with Region V began to increase approximately 10 13 years after high arsenic exposures commenced and continued to rise, peaking between 1986 and 14 1997. The peak lung cancer mortality rate ratios for women and men were 3.26 (95% CI: 2.50- 15 4.23)and3.61 (95% CI: 3.13-4.16), respectively. The peak bladder cancer rate ratios for 16 women and men were 13.8 (95% CI: 7.74-24.5) and 6.10 (95% CI: 3.97-9.39), respectively. 17 Together, lung and bladder cancer mortality rates in Region II were highest from 1992 to 1994, 18 with mortality rates of 50/100,000 for women and 153/100,000 for men compared with 19 19/100,000 and 54/100,000, respectively, in Region V. The long latency for lung and bladder 20 cancer mortality continued to have a residual effect through the late 1990s, even though there 21 was a significant decrease in arsenic exposure through drinking water more than 25 years earlier. 22 Strengths of the investigation include the large study population, the availability of past 23 exposure data, and that potential confounders of age, gender, and smoking history were 24 controlled for in the analysis. However, weaknesses include the inability to account for 25 migration, the ecologic design (i.e., lack of individual exposure data) and lack of information 26 concerning occupation. 27 Yuan et al. (2007) investigated mortality from 1950 to 2000 using an ecological study 28 design in the arsenic-exposed Region II of Chile and the unexposed population from Region V. 29 Before 1958, the drinking water in Region II contained approximately 90 ppb of arsenic. In 30 1958, it became necessary to supplement the Region II water supply using rivers that had an 31 average arsenic concentration of 870 ppb. After the installation of an improved water treatment 32 operation in the early 1970s, the arsenic concentrations in the Region II water supply dropped 33 sharply (<10 ppb). While acute myocardial infarction (AMI) mortality was the predominant 34 cause of excess deaths during and immediately after the high-exposure period, due to the longer 35 latency of cancer, excess deaths from lung and bladder cancer became predominated years later. 36 Yuan et al. (2007) concluded that after a 15- to 20-year lag period following initial exposure to 37 significantly elevated levels of arsenic from drinking water (1958-1970), mortality from bladder 51 DRAFT—DO NOT CITE OR QUOTE ------- 1 and lung cancer surpassed other causes of mortality. Strengths of the study included known 2 arsenic concentrations and the large study population. In addition, to ensure appropriate 3 selection of a control population, preliminary investigations were conducted to compare regional 4 income, smoking history, and availability and quality of death certificate information. The major 5 weakness of the study was its ecological study design (i.e., lack of individual arsenic exposure). 6 In addition, potential confounders (i.e., smoking histories, diet, and exercise) were not examined 7 on an individual basis, but were compared on a regional basis. 4.1.4. North America (United States and Mexico) 8 Bates et al. (1995), in a case-control study, used data obtained from Utah respondents for 9 the 1978 National Bladder Cancer Study to examine the potential relationship between bladder 10 cancer in a U.S. population exposed to measurable levels of arsenic in drinking water. Arsenic 11 levels in drinking water were lower than those in Asian and South American studies. A total of 12 117 cases and 266 controls were selected as participants for this study. Restricting subjects to 13 those who had lived in study areas for at least half of their lives, the number of subjects still 14 eligible was 71 cases and 160 controls. Arsenic exposures ranged from 0.5 to 160 ppb (mean, 15 5.0 ppb). Two measurements of arsenic exposure were used. One measure used was the total 16 CAE and the other was the arsenic concentration ingested adjusted for individual water 17 consumption. Bates et al. (1995) found no association between bladder cancer and either arsenic 18 exposure measure. However, among smokers, positive trends in cancer risk were found for 19 arsenic exposures between 30 to 39 years prior to cancer diagnosis. The risk estimates were 20 stronger for the drinking water measure that estimated the ingested arsenic concentration than 21 the CAE. The risk estimates obtained, however, were higher than predicted based on the results 22 of the Taiwanese studies, which raised concerns by Bates et al. (1995) regarding confounders, 23 bias, and chance. 24 The data from this study raised the potential that smoking contributes to the increased 25 effect of arsenic on the risk of bladder cancer. Potential confounders included in the logistic 26 models were gender, age, smoking status, years of exposure to chlorinated water, history of 27 bladder infection, and the highest educational level attained. Strengths of the Bates et al. (1995) 28 investigation are that these confounders were controlled for; occupation, population size of 29 geographic area, and urbanization were addressed in the analysis; and cases were histologically 30 confirmed. Potential weaknesses of the study are the small size of the study population, the fact 31 that the subjects were mostly male and the data on females were inadequate, and that arsenic 32 exposure levels were based on measurements close to the time that cases were diagnosed. Due 33 to the low concentration in the water, the lack of measurement of arsenic in the food was a 34 limitation of this study. Although the purpose of the Bates et al. (1995) study was to compare 35 low-level arsenic exposure and bladder cancer with the results from the Taiwanese population, 52 DRAFT—DO NOT CITE OR QUOTE ------- 1 the results cannot be interpreted without consideration of potential confounders and bias 2 resulting from the retrospective study design. 3 Employing a retrospective cohort mortality investigation of residents from Millard 4 County, Utah, Lewis et al. (1999) examined the relationship between arsenic exposure from 5 drinking water and mortality outcome. Median drinking water arsenic concentrations for 6 selected study areas ranged from 14 to 166 ppb. Drinking water samples were obtained from 7 public and private sources and were collected and analyzed under supervision of the State of 8 Utah Department of Environmental Quality, Division of Drinking Water. Cohort members were 9 assembled using historical documents made available by the Church of Jesus Christ of Latter- 10 Day Saints. Residential histories and median drinking water arsenic concentration were used to 11 construct a matrix for CAE. Previous drinking water arsenic concentrations (from 1964 forward) 12 were obtained from historical records of arsenic measurements maintained by the state of Utah. 13 Without regard to specific exposure levels, statistically significant increases in mortality from 14 prostate cancer (SMR=1.45, 95% CI: 1.07-1.91) among cohort males was observed. Non- 15 significant increases in mortality for males were observed in cancer of the kidney (SMR=1.75, 16 95% CI: 0.80-3.32). There was no increased risk for cancer of the bladder and other urinary 17 organs (SMR=0.42, 95% CI: 0.08-1.22) in males. Among cohort females, no statistically 18 significant increase in mortality was observed. Females did, however, exhibit non-significant 19 increases in mortality from kidney cancer (SMR=1.60, 95% CI: 0.44-4.11) and melanoma of the 20 skin (SMR=1.82, 95% CI: 0.50-4.66). Female cancer of the bladder and other urinary organs 21 (SMR=0.81, 95% CI: 0.10-2.93) was not increased. Risk analysis using low-, medium-, and 22 high-arsenic exposure groups did not provide any clear indication of a dose-response for prostate 23 cancer. Confounding was not considered to be a significant concern by Lewis et al. (1999). 24 Exposure to other arsenic sources (food- or airborne), however, may have contributed to the total 25 exposure potential of this population. Strengths of the study included the cohort study design. 26 In this design type, the exposure precedes the effect being measured so a variety of effects from 27 a single type of exposure can be considered. The study population was mostly rural and 28 Mormon (low tobacco and alcohol use). In addition, NRC (2001) and EPA (U.S. EPA, 2001) 29 identified that the Lewis et al. (1999) study was not powerful enough to estimate risk. 30 To address the association between skin cancer and arsenic exposure in drinking water, 31 Karagas et al. (2001) used data collected on 587 basal cell and 284 squamous cell skin cancer 32 cases and 524 controls. Cases and controls were interviewed as part of a case-control study 33 conducted in New Hampshire (and bordering regions) between 1993 and 1996. Arsenic 34 exposure levels were determined using toenail clippings. The ORs for SCC (range 0.93-1.10) 35 and BCC (range 0.72-1.06) were not significant and near unity (1.0) in all but the highest 36 category (0.345-0.81 ug/g). For cases with significantly elevated toenail arsenic concentrations, 37 the adjusted ORs were 2.07 (95% CI 0.92-4.66) for SCC and 1.44 (95% CI: 0.74-2.81) for BCC, 53 DRAFT—DO NOT CITE OR QUOTE ------- 1 compared with those with concentrations at or below the median. Since the risks of SCC and/or 2 BCC were not elevated in the range of toenail arsenic concentrations detected in most study 3 subjects, the authors did not exclude the possibility of a dose-related increase at the highest 4 levels of exposure. Strengths include evaluating the effects of potential confounders such as age, 5 gender, race, educational attainment, smoking status, skin reaction to first exposure to the sun, 6 and history of radiotherapy. Toenail arsenic concentrations can be considered a strength and a 7 weakness. They are a strength because they individualize the dose and could account for arsenic 8 exposure from other sources (e.g., food), but they also could be considered a weakness because 9 toenail arsenic is a biomarker of recent past exposure (covering a period of about one year 10 according to Cantor and Lubin, 2007). Some confounding variables were not controlled for and 11 may have influenced the results. The latency of arsenic-induced skin cancer is unknown and, as 12 a result, the follow-up period for this study may have been inadequate. 13 The identification of a potential leukemia cluster in Churchill County, Nevada, where 14 arsenic levels in water supplies are relatively high, prompted a study by Moore et al. (2002). 15 Using an ecological study design, Moore et al. examined the incidence of childhood cancer 16 between 1979 and 1999 in all 17 Nevada counties. For analysis, arsenic exposures were grouped 17 into low (<10 ppb), medium (10-25 ppb), and high (35-90 ppb) population-weighted arsenic 18 levels based on the levels obtained from public drinking water. SIRs for all childhood cancers 19 combined were 1.00 (95% CI: 0.94-1.06) for low-exposure, 0.72 (95% CI: 0.43-1.12) for 20 medium, and 1.25 (95% CI: 0.91-1.69) for high-exposure counties. Moore et al. (2002) found 21 no apparent relationship between the three arsenic levels and childhood leukemia with SIRs of 22 1.02 (95% CI: 0.90-1.15), 0.61 (95% CI: 0.12-1.79), and 0.86 (95% CI: 0.37-1.70) in the low, 23 medium, and high exposure categories, respectively. No association was found for all childhood 24 cancers, excluding leukemia, with SIRs of 0.99 (95% CI: 0.92-1.07), 0.82 (95% CI: 0.47-1.33), 25 and 1.37 (95% CI: 0.96-1.91), respectively. There was, however, an excess for bone cancers in 26 5- to 9-year-olds and 10- to 14-year-olds and an excess in cancer (primarily lymphomas) in 15- 27 to 19-year-old young adults in the high-exposure group. The findings in this study showed no 28 increase in leukemia risk at the concentrations of arsenic identified and categorized in the water. 29 Although the results did not eliminate the possibility for increased risks for non-leukemia 30 childhood cancers, there is no reason to suspect that the exposures to low levels of arsenic in the 31 small study group are responsible. Strengths of the study are that the analysis of the data was 32 stratified by age, the study was a low-level arsenic exposure study, and the findings were 33 reported at different arsenic concentrations. Weaknesses of the study include the small study 34 size, the potential for exposure misclassification, and the limitations of the ecological study 35 design. 36 Steinmaus et al. (2003) used a case-control study to evaluate the effects of arsenic 37 ingestion on bladder cancer risk in seven counties in the western United States. These counties 54 DRAFT—DO NOT CITE OR QUOTE ------- 1 contain the largest populations historically exposed to arsenic via drinking water at levels of 2 approximately 100 ppb. These populations gave Steinmaus et al. the opportunity to critically 3 evaluate the effects of relatively low-level arsenic exposure on bladder cancer incidence. 4 Incident bladder cancer cases diagnosed between 1994 and 2000 were recruited based on 5 information obtained from the Nevada Cancer Registry and the Cancer Registry of Central 6 California. Arsenic measurements for community-supplied drinking water within the study were 7 provided by the Nevada State Health Division and the California Department of Health Services. 8 Over 7000 arsenic measurements were obtained. Individuals' data on water sources, water 9 consumption patterns, smoking history, and other sociodemographic factors were obtained for 10 181 bladder cancer cases and 328 matched controls. There was no observed increased risk for 11 bladder cancer associated with intakes greater than 80 ug/day (OR=0.94, 95% CI: 0.56-1.57; 12 linear trend, p=0.48). This observed OR was below the risk predicted based on higher arsenic 13 concentrations in drinking water studies from Taiwan. However, when the analysis focused 14 solely on previous smokers who had arsenic exposures greater than 80 ug/day (median 177 15 ug/day) for more than 40 years, the risk was significantly increased (OR=3.67, 95% CI: 1.43- 16 9.42; linear trend, p< 0.01). These data provide evidence that smoking and ingesting arsenic at 17 elevated concentrations (i.e., greater than 100 ug/day) may result in an increased risk of bladder 18 cancer. A strength of the Steinmaus et al. (2003) study is the use of individual exposure level 19 data to examine low-dose drinking water arsenic exposure; however, the lack of arsenic exposure 20 from food is a study weakness due to the low levels of exposure through drinking water. In 21 addition, the use of cancer registries allowed for improved case identification. Potential 22 confounders adjusted for in the analysis included gender, age, smoking history, education, 23 occupation associated with elevated rates of bladder cancer, and income. However, bias as the 24 result of next-of-kin interviews may have influenced the exposure assessment. Arsenic 25 exposures from outside the study area also may have influenced the exposure assessment. In the 26 arsenic-exposed areas, the percentage of non-participants was 5% higher among cases than 27 controls. This difference probably means that more exposed cases were missed in analyses of 28 recent exposure, biasing the OR toward the null. 29 There has been little research investigating the link between arsenic and cutaneous 30 melanoma, although arsenic has been associated with increased risk of non-melanoma skin 31 cancer. Beane-Freeman et al. (2004) performed a case-control study to examine the potential 32 relationship between melanoma and environmental arsenic exposure in a cohort from Iowa. 33 Study participants included 368 cutaneous melanoma cases (selected from 645 eligible cases) 34 and 373 colorectal cancer controls (selected from 732 eligible controls) diagnosed in 1999 or 35 2000, frequency-matched on gender and age. Participants completed a mailed survey and 36 submitted toenail clippings (obtained from 355 cases and 353 controls) for analysis of arsenic 37 content. The authors identified an increased risk of melanoma in study cases with elevated 55 DRAFT—DO NOT CITE OR QUOTE ------- 1 toenail arsenic concentrations (OR=2.1, 95% CI: 1.4-3.3; p-trend=0.001) and an increased risk 2 of melanoma with previous diagnosis of skin cancer and elevated toenail arsenic concentrations 3 (OR=6.6, 95% CI: 2.0-21.9). There was a greater association between the toenail arsenic and 4 melanoma when subjects reported a previous diagnosis of melanoma. Strengths of this 5 investigation include the fact that the potential confounders (age, gender, skin color/skin type, 6 prior history of sunburn, education, and occupational exposure) were controlled for in the 7 analysis. Ascertainment of cases and controls was accomplished by using the Iowa Cancer 8 Registry, a Surveillance, Epidemiology, and End Results Program registry. This allowed newly 9 diagnosed melanoma cases to be identified for a specific period and assured a greater degree of 10 certainty regarding the accuracy of diagnosis. Another strength is that toenail arsenic 11 concentrations individualize the exposure and account for arsenic exposure from other sources. 12 A limitation of this study was that toenail samples were collected 2 to 3 years after diagnosis and 13 therefore do not measure arsenic concentrations prior to diagnosis, resulting in possible exposure 14 misclassification. 15 Karagas et al. (2004) used a case-control study design to examine the effects of low-level 16 arsenic exposure on the incidence of bladder cancer in New Hampshire (and bordering regions), 17 where levels above 10 ppb are commonly found in private wells. The authors studied 383 cases 18 of transitional cell carcinoma of the bladder, diagnosed between July 1, 1994, and June 30, 1998, 19 and 641 general population controls. Individual exposure to arsenic was determined through the 20 use of toenail clippings. Karagas et al. (2004) found arsenic concentrations ranged from 0.014 to 21 2.484 ug/g among bladder cancer cases and 0.009 to 1.077 ug/g among controls. When stratified 22 by smoking history, toenail arsenic concentrations were not associated with the risk of bladder 23 cancer. However, among smokers in the uppermost category of arsenic exposure, an elevated 24 OR for bladder cancer was observed (OR: 2.17, 95% CI: 0.92-5.11 for >0.330 ug/g compared to 25 <0.06 ug/g). When Karagas et al. (2004) stratified their analysis by duration of current water 26 system usage (<15 years and >15 years), an increased bladder cancer OR for people who ever 27 smoked with the highest category of arsenic exposure with less than 15 years of use was 28 identified (<15 years, OR=3.09, 95% CI: 0.80-11.96; >15 years, OR=1.86, 95% CI: 0.57-6.03). 29 These data suggest that ingestion of low to moderate arsenic levels may affect bladder cancer 30 incidence and that cigarette smoking may act as a co-carcinogen. Strengths of the study include 31 its use of a stratified analysis to evaluate the potential that an extended latency period was 32 required for bladder cancer development and its minimizing of misclassification by using 33 biomarkers. The following potential confounders were adjusted for: age, gender, race, 34 educational attainment, smoking status, family history of bladder cancer, study period, and 35 average number of glasses of tap water consumed per day. Toenail clippings were used in an 36 attempt to minimize misclassification. This, however, is a limitation because it only measures 37 recent past exposures. Limitations of the study were that misclassification at the lower 56 DRAFT—DO NOT CITE OR QUOTE ------- 1 exposures was possible and that lifetime exposure could not be calculated since data from 2 previous residences could not be determined. In addition, there was limited data at extreme ends 3 of exposure. 4 The Lamm et al. (2004) ecological study investigated the association between arsenic 5 exposure from drinking water and bladder cancer mortality in 133 counties in the United States. 6 Caucasian male county-specific bladder cancer mortality data between 1950 and 1979 and 7 county-specific ground water arsenic concentration data were obtained for counties solely 8 dependent on ground water for their public drinking water supply. Arsenic exposure was based 9 on measurements for at least 5 wells for each county. No arsenic-related increase in bladder 10 cancer mortality (SMR=0.94, 95% CI: 0.90-0.98) was identified (arsenic exposure range: 3-60 11 ppb) using stratified analysis and regression analyses. These findings are consistent with other 12 previously published U.S. studies. Strengths of the study include the large nationwide study 13 population, which included more than 75 million person-years of observation. Weaknesses, 14 however, are the lack of available individual exposure data, the assumption that study 15 participants consumed only local drinking water, the assumption that available data were 16 representative of actual arsenic content in the water, that arsenic contribution from food sources 17 were not analyzed, and that the analysis did not directly adjust for smoking, urbanization, or 18 industrialization. 19 The Wisconsin Division of Public Health, in July 2000 through January 2002, conducted 20 a cross-sectional study in 19 rural Wisconsin townships concerning private drinking-water wells 21 and arsenic exposure (Knobeloch et al., 2006). Residents in these townships were asked to 22 collect well-water samples and complete a questionnaire regarding residential history, 23 consumption of drinking water, and family health. In Wisconsin, skin cancer is not reportable; 24 therefore, no skin cancer registry data were available. During the study, 2,233 private wells 25 were tested, and 6,669 residents provided information on water consumption and health. Water 26 arsenic levels ranged from less than 1.0 to 3,100 ppb. The median arsenic level was 2.0 ppb. 27 Eighty percent of the wells had arsenic levels below 10 ppb, but 11% had an arsenic level of 28 above 20 ppb. Age-, gender-, and smoking-adjusted ORs of residents 35 years of age and older 29 who had consumed water with arsenic levels greater than 1.0 ppb for at least 10 years showed a 30 significant increase in individuals who reported skin cancer compared to those whose water 31 arsenic levels were less than 1.0 ppb (arsenic 1.0-9.9 ppb OR=1.81, 95% CI: 1.10-3.14). 32 Similarly, adults whose well-water reportedly contained arsenic concentrations greater than 10 33 ppb were significantly more likely to report skin cancer than those whose water arsenic levels 34 were less than 1.0 ppb (OR=1.92, 95% CI: 1.01-3.68). Tobacco use also was associated with 35 higher rates of skin cancer and may—synergistically with arsenic exposure—affect the 36 development of skin cancer. Strengths of the study include: the large sample size, a history of 37 individual tobacco use, arsenic well water analysis for each household, an exposure duration of 57 DRAFT—DO NOT CITE OR QUOTE ------- 1 at least 10 years in participants who consumed water from the tested wells, and the fact that the 2 analysis controlled for age, gender, and tobacco use. Weaknesses include the following: skin 3 cancers were self-reported and not confirmed by a medical records review, few people could 4 provide information about specific types of cancer, potential bias could have resulted from the 5 participating families being concerned about arsenic exposure, sun exposure and occupation 6 were not controlled for in the analysis, and food sources of arsenic were not considered. 7 8 Meliker et al. (2007) performed an ecological study in a contiguous six-county study area 9 of southeastern Michigan to investigate the relationship between moderate arsenic levels (10- 10 100 ppb) and selected disease outcomes. This region of southeastern Michigan was chosen 11 because it had moderately high arsenic concentrations in the ground water and low rates of 12 migration. The six counties had a population-weighted mean arsenic concentration of 11.00 ppb 13 and a population-weighted median of 7.58 ppb. In comparison, the remainder of Michigan has a 14 population-weighted mean of 2.98 ppb with a median of 1.27 ppb. SMRs for cancers were not 15 significantly different from the age- and race-adjusted expected values for males or females for 16 the state of Michigan (SMR skin melanoma female=0.97, 95% CI: 0.73-1.27, melanoma 17 male=0.99, 95% CI: 0.79-1.22; SMR bladder female=0.98, 95% CI: 0.80-1.19, bladder 18 male=0.94, 0.82-1.08; SMR kidney female=1.00, 95% CI: 0.80-1.20, kidney male=1.06, 95% 19 CI: 0.91-1.22; SMR trachea, lung, bronchus female=1.02, 95% CI: 0.96-1.07, trachea, lung, 20 bronchus male=1.02, 95% CI: 0.98-1.06). The only exception was cancer of the female 21 reproductive organs (SMR=1.11, 95% CI: 1.03-1.19). The potential explanations for the lack of 22 significant cancer findings were the relatively low level of arsenic in the ground water of 23 southeastern Michigan, which may be below the threshold for cancer induction and other 24 moderating factors that were not considered by this study (i.e., food as a source of arsenic 25 exposure). Strengths include that mortality rates, which were gathered from Michigan Resident 26 Death Files for a 20-year period, were stratified by gender, age, and race. Weaknesses include 27 the following: the ecological study design did not provide individual arsenic exposure data and 28 may not permit the detection of significant risk, there may have been differences in reporting and 29 classification of underlying causes of death, case migration occurred, preferential sampling was 30 conducted based on home owners' request, arsenic contribution from food was not measured, 31 and there was a lack of information concerning smoking history and obesity. 4.1.5. China 32 Using an ecological study design, Lamm et al. (2007) conducted dermatological 33 examinations for 3,179 of the 3,228 (98.5%) residents of three villages (Zhi Ji Liang, Tie Men 34 Geng, and Hei He) in Huhhot, Inner Mongolia, with well water arsenic levels that ranged from 35 undetectable (<10 ppb) to 2,000 ppb. Individual water consumption histories were obtained for 36 this population, and arsenic levels were measured for 184 wells. Arsenic exposures were 58 DRAFT—DO NOT CITE OR QUOTE ------- 1 summarized as the highest arsenic concentration (HAC) and CAE. Thirty-five percent of the 2 study population had HAC of less than 50 ppb, 86% had HAC less than 150 ppb, and only 1% of 3 the participants had HAC greater than 500 ppb. The proportion of females to males was similar 4 in each of the three villages (female range 49%-50% and male range 50%-51%), and almost all 5 study subjects identified themselves as being of Chinese (99.8%) rather than Mongolian (0.2%) 6 origin. The median age for all participants was 29 years; however, participants from Hei He 7 tended to be older than those from the other two villages (55.0% older than 30 in Hei He, 42.4% 8 in Zhi Ji Liang and Tie Men Geng). Participants (female or male) who reported occupations 9 listed "student" or "farmer." None of the examinations revealed any evidence of BFD. Analyses 10 included frequency-weighted, simple linear regression, and most likely estimate models. Eight 11 people were found to have skin cancer. In addition to skin cancer, these eight cases also had 12 both hyperkeratoses and dyspigmentation. Skin cancer cases were only identified in those 13 participants with HAC exposures >150 ppb or whose CAE was less than 1,000 ppb-years. The 14 models showed a threshold of 122-150 ppb. Lamm et al. (2007) identified a general exposure- 15 prevalence pattern (higher prevalence for HAC exposure group) for skin disorders 16 (hyperkeratosis, dyspigmentation, and skin cancers). Duration of water usage (arsenic 17 exposure), age, latency, and misclassification did not appear to markedly affect the analysis. 18 Strengths of the study include the large study population, the fact that HAC and CAE were used 19 in the analyses, and the fact that arsenic concentrations were measured in 184 wells. 20 Confounders that were controlled for included age, differences in cumulative arsenic dose, and 21 duration of exposure. A confounder not adjusted for in the analysis was sun exposure. 22 Additional weaknesses are the ecological study design and the potential for recall or 23 misclassification bias resulting from the collection of arsenic exposure histories through 24 interviews. 4.1.6. Finland 25 In a case-cohort study, Kurttio et al. (1999) examined the levels of arsenic in Finnish 26 water wells and their relationship to the risk of bladder and kidney cancers. The study 27 population consisted of 61 bladder cancer cases and 49 kidney cancer cases diagnosed between 28 1981 and 1995, and a randomly selected age- and gender-adjusted reference cohort of 275 29 subjects. Arsenic exposure was estimated for cancer cases and for the reference cohort for two 30 periods. The first period was from the third to ninth calendar years (the shorter latency period) 31 prior to either the cancer diagnosis or the respective year for referent cohort, while the other was 32 from the tenth or earlier calendar years (the longer latency period). Water specimens were 33 obtained from the wells used by the study cohort from 1967 to 1980. The arsenic concentrations 34 in the wells of the control population were low, with approximately 1% exceeding 10 ppb. 35 Bladder cancer was associated with arsenic concentration and daily dose during the third to ninth 36 calendar years prior to the cancer diagnosis. The risk ratios for arsenic exposure concentration 59 DRAFT—DO NOT CITE OR QUOTE ------- 1 categories 0.1-0.5 and >0.5 ppb relative to the category with <0.1 ppb were 1.53 (95% CI: 0.75- 2 3.09) and 2.44 (95% CI: 1.11-5.37), respectively. In spite of low levels of arsenic exposure, 3 Kurttio et al. (1999) found evidence of a relationship between exposure to arsenic at the higher 4 exposure level and bladder cancer risk. No association, however, was observed between arsenic 5 exposure level and kidney cancer risk. Strengths include the following: Finnish Cancer Registry 6 records were accessible; Statistics Finland's 1985 Population Census file was used to identify 7 areas in which less than 10% of the population used the municipal water supply; and age, gender, 8 and smoking histories were accounted for in the risk ratio calculations. Possible weaknesses 9 include misclassification and recall bias resulting from the study choosing to use water 10 consumption from the 1970s. In addition, because of the low arsenic concentrations, arsenic 11 exposure from other sources (e.g., food) could bias the results. 12 Michaud et al. (2004) used a cohort (nested case-control) study design to investigate the 13 relationship between arsenic levels in toenail and bladder cancer risk among Finnish male 14 smokers aged 50-69 years who were participating in the Alpha-Tocopherol, Beta-Carotene 15 Cancer Prevention Study. Data for 280 incident bladder cancer cases, identified between 1985 16 and 1988 as well as April 1999, were available for analysis. Controls (n = 293) were matched to 17 each case on the basis of age, toenail collection date, intervention group, and duration of 18 smoking. Logistic regression analyses were performed to estimate ORs. Arsenic toenail 19 concentrations in this Finnish study (cases and controls) ranged between 0.01 and 2.11 ug/g, 20 with one control outlier at 17.5 ug/g. Arsenic toenail concentrations were similar to those 21 reported in the United States (range: 0.02-17.7 ug/g). Men were categorized into quartiles based 22 on the distribution of arsenic among the controls (O.050, 0.050-0.105, 0.106-0.161, and 23 >0.161). The study observed no significant relationship between arsenic concentration and 24 bladder cancer risk (OR=1.13, 95% CI: 0.70-1.81 for the highest vs. lowest quartile). Strengths 25 of the Michaud et al. (2004) study were that the authors excluded toenail samples with non- 26 detectable arsenic levels greater than 0.09 ug/g, in an attempt to avoid potential misclassification 27 of samples with high detection limits, and that they controlled for potential confounders in the 28 analysis (i.e., smoking history, beverage intake, place of residence, toenail weight, smoking 29 cessation, smoking inhalation, educational level, beverage intake, and place of residence). Cases 30 and controls were matched according to age, toenail collection date, intervention group (alpha 31 tocopherol and beta carotene), and smoking duration. Toenail arsenic concentrations are a 32 strength because they individualize the dose and could account for arsenic exposure from other 33 sources, but they also could be considered a weakness because toenail arsenic is a biomarker of 34 recent past exposure (covering about 1 year according to Cantor and Lubin, 2007). Another 35 weakness of the study was that water consumption was not included in the total beverage intake 36 variable. 60 DRAFT—DO NOT CITE OR QUOTE ------- 4.1.7. Denmark 1 The Baastrup et al. (2008) cohort study was designed to determine whether exposure to 2 low levels of arsenic in drinking-water in Denmark is associated with an increased risk for 3 cancer. The study population was selected from participants in the prospective Danish cohort 4 Diet, Cancer, and Health. A cohort of 56,378 people (39,378 from Copenhagen and 17,000 from 5 Aarhus) accepted an invitation to participate in the study. Cancer cases were identified in the 6 Danish Cancer Registry, and the Danish civil registration system was used to trace residential 7 addresses of the cohort members. The study used a geographic information system to link 8 residential addresses with water supply areas and using this information estimated arsenic 9 exposure by addresses. The average arsenic exposure for the cohort ranged between 0.05 and 10 25.3 ppb (mean =1.2 ppb) and was based on 4,954 measurements reported between 1987 and 11 2004 (the majority between 2002 and 2004). The exposure was generally higher among Aarhus 12 participants than those enrolled in the Copenhagen area (Aarhus mean = 2.3 ppb, min = 0.09 ppb 13 and max=25.3 ppb; Copenhagen mean = 0.7 ppb, min = 0.05 ppb, and max=15.8 ppb). 14 Regression models were used to analyze possible relationships between arsenic and cancer. The 15 study found no significant association between arsenic exposure and risk for cancers of the lung, 16 bladder, liver, kidney, prostate, colon, or melanoma skin cancer. The incidence rate ratio (IRR) 17 for non-melanoma skin cancer (0.88, 95% CI: 0.84-0.94) decreased with per ppb increases in the 18 time-weighted average exposure to arsenic. The study did identify a significant increased risk 19 for breast cancer in association with time-weighted average exposure to arsenic (IRR=1.05, 95% 20 CI: 1.01-1.10). Strengths of the study include the large study population, the 21 socioeconomic/demographic similarities of the cohort, and the adjustment for potential 22 confounders (smoking, alcohol consumption, education, body mass index [BMI], daily intake of 23 fruits/vegetables, red meat, fat and dietary fiber, skin reaction to the sun, hormone replacement 24 therapy use, reproduction, occupation, and enrollment area). Weaknesses of the study include 25 the low arsenic levels in Danish drinking water, the lack of information on other sources of 26 arsenic exposure, and the inability to assess arsenic exposures before 1970, all resulting in 27 possible misclassification bias. 4.1.8. Australia 28 Hinwood et al. (1999) conducted an ecological study that investigated areas of Victoria, 29 Australia, with elevated environmental arsenic concentrations, areas with arsenic concentrations 30 in the soil of more than 100 mg/kg and/or drinking water arsenic concentrations greater than 10 31 ppb, and the relationship with cancer incidence. SIRs for cancer were generated for 22 areas 32 between 1982 and 1991 using cancer registry data. In addition, SIRs for combined areas 33 according to environmental exposure (high soil and/or high water arsenic concentrations, etc.) 34 were generated. The SIRs (females and males together) for the combined 22 areas were 35 significantly elevated for all cancers (1.06, 95% CI: 1.03-1.09), melanoma (1.36, 95% CI: 1.24- 61 DRAFT—DO NOT CITE OR QUOTE ------- 1 1.48), chronic myeloid leukemia (1.54, 95% CI 1.13-2.10), breast cancer in females (1.10, 95% 2 CI: 1.03-1.18), and prostate cancer in males (1.14, 95% CI: 1.05-1.23). The SIR for kidney 3 cancer (females and males combined) was 1.16 (95% CI: 0.98-1.37), and although elevated was 4 not statistically significant. When stratified by exposure category, the SIR for prostate cancer 5 was significant at 1.20 (95% CI: 1.06-1.36) for the high soil/high water category only. This 6 result was likely confounded by misclassification (level of population exposure) and limited by 7 low statistical power. There was no significant dose-response relationship observed between 8 drinking water and any individual cancer. Strengths of the study include that water and soil 9 arsenic levels were provided and a large area was examined. Hinwood et al. (1999) recognized 10 that the results of this study were potentially confounded by a number of factors, including the 11 ecological study design, socioeconomic status, race, occupation, and urban versus rural status. 12 Due to the low concentrations in the drinking water, the lack of arsenic exposure from food 13 could cause exposure misclassification. 4.2. PRECHRONIC AND CHRONIC STUDIES AND CANCER BIOASSAYS IN ANIMALS—ORAL 4.2.1. Prechronic and Chronic Studies 14 Wei et al. (1999, 2002) demonstrated that 10-week-old male F344/DuCrj rats (36/group) 15 administered 12.5, 50, or 200 ppm DMAV (a major metabolite of inorganic arsenic) in their 16 drinking water for 104 weeks had no effect on the morbidity, mortality, body weights, 17 hematology, or serum biochemistry. Reductions in electrolyte concentrations in the urine were 18 related to an increase in urinary volume resulting from increased water consumption in the 50- 19 and 200-ppm groups. There was no difference in the urinary pH between control and treated 20 rats. 4.2.2. Cancer Bioassays 21 Cancer bioassays with inorganic arsenic have generally obtained negative results with 22 mice, rats, hamsters, rabbits, beagles, and cynomologus monkeys (for review see Kitchin, 2001; 23 NRC, 1999). However, the following studies have observed increases in tumors in animals 24 exposed to arsenic species. 4.2.2.1. Mice—Transplacental 25 Timed pregnant female C3H mice (10/group) were administered 0 (control), 42.5, or 85 26 ppm As111 in their drinking water ad libitum from day 8 to day 18 of gestation (Waalkes et al., 27 2003). Strain and doses used in the experiment were determined through preliminary short-term 28 testing that determined C3H mice to be the most sensitive to arsenic toxicity of the three strains 29 tested (i.e., C3H, C57BL/6NCr, and B6C3Fl/NCr), and the preliminary test indicated that a dose 30 of 100 ppm was unpalatable and resulted in approximately 10% reduced growth in the offspring. 62 DRAFT—DO NOT CITE OR QUOTE ------- 1 The doses used in this study did not affect maternal water consumption or body weight in the 2 dams. It was estimated that the pregnant females consumed 9.55 to 19.13 mg arsenic/kg-day, for 3 a total dose of 95.6 to 191.3 mg arsenic/kg. 4 Offspring were weaned at 4 weeks and received no additional exposure to arsenic. Male 5 and female offspring (25/sex/group) were observed for the next 74 or 90 weeks, respectively. 6 Males were sacrificed at 74 weeks due to high mortality in the high-dose group beginning at 52 7 weeks. Both the 42.5- and 85-ppm males had a significant increase in the incidence of HCC 8 (12.5% in the control group versus 38.1% in the 42.5-ppm group and 60.9% in the 85-ppm 9 group) and adrenal cortical tumors (37.5% in the control group versus 66.6% in the 42.5-ppm 10 group and 91.3% in the 85-ppm group), which followed a significant (p<0.001), dose-related 11 trend. In addition, the 85-ppm group had a significant increase in the multiplicity (tumor/mouse) 12 for both HCC (0.13, 0.42, and 1.30, respectively) and adrenal tumors (0.71, 1.10, and 1.57, 13 respectively), which also had a significant (p<0.02), dose-related trend. Although there were no 14 differences in the incidence of hepatocellular adenomas in males, the multiplicity of 15 hepatocellular adenomas (0.71, 1.43, and 3.61, respectively) followed a significant (p < 0.0001), 16 dose-related trend. 17 Males and females had an increase in lung tumors (8.0%, 13.0%, and 25.0%, 18 respectively, in females; 0%, 0%, and 13.0%, respectively, in males), which followed a 19 significant (p<0.03), dose-response trend. In addition, females had increases in the incidence of 20 benign ovarian tumors, which reached statistical significance in the 85-ppm group. Although a 21 significant increase was not observed in malignant ovarian tumors, the total incidence (benign 22 plus malignant) of ovarian tumors was significant in the 85-ppm group and followed a 23 significant (p=0.015), dose-related trend (8% in the control group versus 26% in the 42.5-ppm 24 group and 37.5% in the 85-ppm group). There was an increase in uterine tumors that was not 25 significant and did not follow a dose-response trend, but was accompanied by a significant 26 (p=0.0019), dose-related increase in hyperplasia occurring at both doses. Females also had a 27 dose-related increase in hyperplasia of the oviduct. The number of both tumor-bearing and 28 malignant tumor-bearing males was significantly increased in both dose groups and followed a 29 significant (p=0.0006 and 0.0001, respectively), dose-related trend. Female animals had a slight 30 increase in the number of tumors, which did not reach statistical significance and did not appear 31 to be dose-related. The number of females bearing malignant tumors was significantly increased 32 for both dose groups, but not in a dose-dependent manner. 33 Waalkes et al. (2004a) followed the same procedure (except that offspring were observed 34 for 104 weeks), but exposed 25 male and 25 female offspring from each exposure group (0, 42.5, 35 or 85 ppm in the drinking water from gestational days 8 to 18 with no additional exposure after 36 birth) to acetone or 12-O-tetradecanoyl phorbol-13-acetate (TPA; 2 ug/0.1 mL in acetone) twice 37 a week—via a shaved area of dorsal skin—for 21 weeks after weaning in an attempt to promote 63 DRAFT—DO NOT CITE OR QUOTE ------- 1 skin tumors. However, very few skin lesions occurred and were not associated with arsenic 2 exposure either in the absence or presence of TPA. As was noted in Waalkes et al. (2003), there 3 was a dose-dependent increase in the incidence and/or multiplicity of hepatocellular adenomas 4 and carcinomas in treated males, both in the absence and presence of TPA. In the absence of 5 TPA, the incidence of adenomas was 41.7%, 52.2%, and 90.5% for the 0-, 42.5-, and 85-ppm 6 exposure groups, respectively; the incidence of carcinomas was 12.5%, 34.8%, and 47.6%, 7 respectively; total incidence was 50%, 60.9%, and 90.5%, respectively; and multiplicity was 8 0.75, 1.87, and 2.14, respectively. In the presence of TPA, the incidence of adenomas was 9 34.8%, 52.2%, and 76.2% for the 0-, 42.5-, and 85-ppm exposure groups, respectively; the 10 incidence of carcinomas was 8.7%, 26.0%, and 33.3%, respectively; total incidence was 39.1%, 11 65.2%, and 85.7%, respectively; and multiplicity was 0.61, 1.44, and 2.14, respectively. A 12 statistically significant increase was noted at 85 ppm. Arsenic only caused a dose-dependent 13 increase in hepatocellular adenomas and carcinomas in the presence of TPA in females 14 (adenomas: 8.3%, 18.2%, and 28.6% for the 0-, 42.5-, and 85-ppm exposure groups with TPA 15 exposure, respectively; carcinomas: 4.2%, 9.1%, and 19.0%, respectively; total incidence: 12.5, 16 27.3, and 38.1%, respectively; multiplicity: 0.13, 0.32, and 0.71, respectively), with a 17 statistically significant increase in total incidence and multiplicity for the 85-ppm group. 18 There also was an increase in ovarian adenomas in treated female offspring regardless of 19 whether they were treated with TPA (0%, 22.7%, 19.0%, respectively) or acetone (0%, 17.4%, 20 and 19.0%, respectively). There was no effect on the incidence of ovarian carcinomas. This was 21 accompanied by increases in the incidence of uterine epithelial hyperplasia (cystic) and total 22 uterine proliferative lesions, which increased in severity with dose. There also was a dose- 23 dependent increase in oviduct hyperplasia. Male offspring exposed to arsenic had an increase in 24 the incidence and multiplicity of cortical adenomas of the adrenal glands. The increases were 25 statistically significant for both arsenic exposure groups, but were only related to dose in the 26 absence of TPA (p=0.020). Incidences were as follows: 37.5%, 65.2%, and 71.4% for the 0-, 27 42.5-, and 85-ppm dose groups, respectively, in the absence of TPA and 30.4%, 65.2%, and 28 57.1%, respectively, with TPA treatment. Multiplicities also were statistically significantly 29 increased in arsenic-exposed male offspring with a significant dose-dependent trend both in the 30 absence (0.58, 2.13, and 2.19, respectively; p=0.0014) or presence (0.54, 1.65, and 1.62, 31 respectively; p=0.016) of TPA. 32 Lung adenomas were increased in a dose-dependent manner in females exposed to TPA 33 (4.2%, 9.1%, and 28.%, respectively; p=0.018), but not in the absence of TPA (4.2%, 8.7%, and 34 9.5%, respectively; not significant). Males only had a statistically significant increase (5-fold 35 increase) in lung adenomas in the 42.5-ppm group exposed to TPA. 36 A statistically significant increase in the multiplicity of all tumors in males (with or 37 without TPA) was observed after arsenic exposure, but was not dependent on dose. Although 64 DRAFT—DO NOT CITE OR QUOTE ------- 1 females also had an increase in the multiplicity of all tumors, the only statistically significant 2 increase occurred in the 85-ppm group exposed to TPA. The increase in females exposed to 3 TPA also appeared to be dose-dependent. The statistically significant increase observed in the 4 multiplicity of malignant tumors in males was greater in the absence of TPA, but was dose- 5 dependent in the presence of TPA. In females, there was also an increase in the multiplicity of 6 malignant tumors in arsenic treated mice (regardless of TPA exposure), but the results did not 7 reach statistical significance, nor were they dose-dependent. 8 Waalkes et al. (2006a) used female CD1 mice, which have a low rate of spontaneous 9 tumors. Thirty-five percent (12/34) of female offspring receiving 85 ppm of As111 via the dams' 10 drinking water on gestational days 8 to 18 developed urogenital tumors, with 9% being 11 malignant compared to 0% in the controls. 4.2.2.2. Rat—Oral 12 Soffritti et al. (2006) administered male and female Sprague-Dawley rats 0, 50, 100, or 13 200 mg/L (i.e., ppm) of sodium arsenite via the drinking water for 104 weeks. There was a 14 consistent dose-dependent decrease in water and food consumption accompanied by a dose- 15 related decrease in body weight (there was no difference in body weight in females administered 16 50 mg/L). There was only a slight decrease in survival in male rats administered 100 or 200 17 mg/L beginning at 40 weeks of age. Females only had a decrease in survival rate after 104 18 weeks of age. Males and females administered 100 mg/L had an increase in the number of 19 tumor-bearing animals and in the number of tumors. Although there is no dose-related trends in 20 tumors, there were sporadic benign and malignant tumors of the lung, kidney, and bladder 21 observed in treated rats that are extremely rare in the authors' extensive historical controls. 22 These tumors consisted of adenomas and carcinomas of the lung, adenomas and carcinomas of 23 the kidney, papillomas and one carcinoma of the renal pelvis transitional cell epithelium, and one 24 carcinoma of the bladder transitional cell epithelium. 25 Wei et al. (1999 and 2002) demonstrated that 10-week-old male F344/DuCrj rats 26 (36/group) administered 50 or 200 ppm DMAV in their drinking water for 104 weeks developed 27 bladder tumors (mainly carcinomas) and papillary or nodular hyperplasia in a dose-dependent 28 manner. Controls and rats administered 12.5 ppm did not develop any bladder tumors or 29 hyperplasia. There was a significant (p < 0.05) increase in bromodeoxyuridine (BrdU) labeling 30 of morphologically normal epithelium of the bladder in the 50- and 200-ppm groups (Wei et al., 31 2002). There was no significant increase in any other tumor type related to DMAV treatment. 32 There appeared to be a dose-related increase in subcutis fibromas (i.e., 4% in controls, 12% in 33 the 12.5-ppm group, and 16% in both the 50- and 200-ppm groups). Data indicate that multiple 34 genes are involved in the stages of DMAv-induced urinary bladder tumors. Wei et al. (2002) 35 further indicate that reactive oxygen species (ROS) may play an important role during the early 3 6 stages of DMA carcinogenesi s. 65 DRAFT—DO NOT CITE OR QUOTE ------- 1 Shen et al. (2003) administered TMAO, an organic metabolite of inorganic As, to male 2 F344 rats for 2 years via their drinking water at concentrations of 0, 50, or 200 ppm. Total 3 intakes were estimated to be 0, 638, and 2475 mg/kg, respectively. From 87 weeks of treatment 4 on, there was an increase in the incidence and multiplicity of hepatocellular adenomas in rats 5 sacrificed or dead. Incidences of 14.3%, 23.8%, and 35.6%, respectively, were reported. The 6 respective multiplicities were 0.21, 0.33, and 0.53. The results were statistically significant in 7 the 200-ppm dose group. 4.2.2.3. Other 8 Transgenic models also have been developed to examine arsenic carcinogenesis. Arsenic 9 exposure (200 ppm sodium arsenite in drinking water for 4 weeks) in Tg. AC transgenic mice 10 containing activated H-ras did not induce skin tumors alone; however, the group of mice that 11 were administered arsenic and a subsequent skin painting with TPA showed an increase in the 12 number of papillomas compared to mice treated with TPA alone. Thus, it was suggested that 13 arsenite may be a "tumor enhancer" in skin carcinogenesis (Germolec et al., 1997; Luster et al., 14 1995). 15 Ten ppm of either sodium arsenite or DMAV (cacodylic acid) administered for 5 months 16 in the drinking water of K6/ODC transgenic mice induced a small number of skin papillomas 17 (Chen et al., 2000a). K6/ODC transgenic mice have hair follicle keratinocytes (likely targets for 18 skin carcinogens), which over express ornithine decarboxylase (ODC). ODC is involved in 19 polyamine synthesis, which is needed in S phase. Over expression of ODC is sufficient to 20 promote papilloma formation without administration of TPA, which has been demonstrated to 21 induce ODC (O'Brien et al., 1997). 22 Rossman et al. (2001) administered sodium arsenite (10 ppm) in the drinking water of 23 hairless Skh 1 mice for 26 weeks. Mice were also administered 1.7 kJ/m2 solar ultraviolet 24 radiation (UV), which is considered a low, nonerythemic dose, 3 times weekly, either with or 25 without sodium arsenite exposure. Results demonstrated a 2.4-fold increase in the yield of skin 26 tumors for mice exposed to both sodium arsenite and UV than in mice administered UV alone. 27 A second experiment by the same group (Burns et al., 2004), demonstrated a 5-fold increase in 28 skin tumors using 5 mg/L As111 with 1 kJ/m2 solar UV, but also observed a significant increase 29 with 1.25 mg/L As111 with 1 kJ/m2 solar UV. The skin tumors (mainly SCCs) occurred earlier, 30 were larger, and were more invasive in mice administered As111. Arsenite alone did not induce 31 skin tumors. Rossman (2003) concluded that this demonstrates that arsenite enhances the onset 32 and growth of malignant skin tumors induced by a genotoxic carcinogen in mice. Rossman 33 (2003) also suggested that the increased tumor incidence observed by Waalkes et al. (2003) may 34 be due to the same enhancement as C3H mice have a high background of spontaneous tumors 35 and suggests the need for examining the transgenic effects in another strain of mice with a lower 36 b ackground tumorgeni city. 66 DRAFT—DO NOT CITE OR QUOTE ------- 1 A critical review of the inhalation data was not conducted as part of the evaluation 2 discussed in this report. 4.3. REPRODUCTIVE/DEVELOPMENTAL STUDIES—ORAL 3 Not addressed in this document. 4.4. OTHER STUDIES 4.4.1. Possible Modes of Action and Key Events of Possible Importance 4 As discussed in Section 3.3, the metabolism of inorganic arsenic in humans occurs 5 through alternating steps of reduction and oxidative methylation mostly to DMAV. Many of the 6 metabolites have been subjected to a variety of toxicological tests in vivo and in vitro, and they 7 often differ considerably in their toxicological responses. The relative contributions of the many 8 different forms of arsenic to the toxicity and carcinogenicity of inorganic arsenic are uncertain. 9 Each of the arsenical metabolites exhibits its own pattern of toxicity, possibly via similar and/or 10 separate MO As that together are responsible for inorganic arsenic toxicity and tumor formation 11 (Kitchin, 2001). 12 The biotransformation and pharmacodynamics of inorganic arsenic are complex in 13 mammals, with inorganic arsenic being biotransformed through a complex cycle of reduction, 14 oxidation, and methylation steps to form the trimethylated TMAO metabolite, and possibly its 15 reduced form, trimethylarsine, which may not be of consequence in humans. Arsenical forms of 16 greater instability (i.e., trivalent forms) are produced within each step, and those forms have 17 greater reactivity toward biological and biochemical intermediates and biological 18 macromolecules. The trivalent species MMA111 and DMA111 have been identified as the most 19 toxic and genotoxic forms in several assay systems (Thomas et al., 2001). Each intermediate 20 arsenical form, however, has the potential to induce cancer or to affect the promotion and 21 progression of cancer, such as by disrupting signal transduction pathways and gene expression. 22 Many of these forms have been detected in the urine of humans exposed to inorganic arsenic and 23 in rodents exposed to inorganic and organoarsenicals. Through the process of metabolizing 24 arsenic, cells and organs are exposed to mixtures of these intermediates, which bring to the 25 forefront potential synergistic interactions between them that could enhance the tumorigenesis 26 process. 27 Inorganic arsenic has been demonstrated to cause tumors in humans at multiple sites 28 (bladder, lung, skin, liver, and possibly kidney). Rodents are generally much less sensitive to the 29 tumorigenic effects of inorganic arsenic, except for a few recent transplacental mouse studies in 30 which As111 caused liver, lung, ovarian, and/or adrenal cortical tumors (Waalkes et al., 2003, 31 2004a, and 2006a). Currently, there is insufficient information to fully explain the differences 32 between human and rodent sensitivity to arsenic carcinogenicity. 67 DRAFT—DO NOT CITE OR QUOTE ------- 1 Based on its extensive review of health consequences of inorganic arsenic in drinking 2 water, NRC (1999) concluded that O 4 • "The mode of action for arsenic carcinogenicity has not been established. Inorganic 5 arsenic and its metabolites have been shown to induce deletion mutations and 6 chromosomal alterations (aberrations, aneuploidy, and SCE [sister chromatid exchange]), 7 but not point mutations. Other genotoxic responses that can be pertinent to the mode of 8 action for arsenic carcinogenicity are co-mutagenicity, DNA methylation, oxidative 9 stress, and cell proliferation; however, data on those genotoxic responses are insufficient 10 to draw firm conclusions. The most plausible and generalized mode of action for arsenic 11 carcinogenicity is that it induces structural and numerical chromosomal abnormalities 12 without acting directly with DNA." 13 14 • "For arsenic carcinogenicity, the mode of action has not been established, but the several 15 modes of action that are considered most plausible (namely, indirect mechanisms of 16 mutagenicity) lead to a sublinear dose-response at some point below the level at which a 17 significant increase in tumors is observed. However, because a specific mode (or modes) 18 of action has not been identified at this time, it is prudent not to rule out the possibility of 19 a linear response." 20 21 Several of the report's other concluding statements drew attention to the possible 22 importance of ROS to several health effects caused by arsenic and suggested that "intracellular 23 production of ROS might play an initiating role in the carcinogenic process by producing DNA 24 damage" (NRC, 1999). At the time of the NRC report, the prevailing view was that metabolism 25 of inorganic arsenic through several methylated forms represented a detoxification pathway. 26 One of the fundamental changes in thinking about the effects of inorganic arsenic since the NRC 27 report has been the growing awareness that some of those metabolites (specifically, MMAm and 28 DMA111) can have especially high levels of toxicity. Thus, metabolism also represents a 29 toxification pathway. Regardless, when there is a steady influx of inorganic arsenic into the 30 body as through continual exposure from drinking water, metabolism is essential to eliminate 31 that arsenic, including the highly reactive As111, from the body. 32 In 2001, NRC produced an update to its major review on inorganic arsenic in drinking 33 water. It summarized, in tabular format, the mechanistic studies completed since 1998 and 34 included a discussion of them. It focused on experiments that appeared to induce biochemical 35 effects at moderate to relatively low concentrations of arsenic in vitro (e.g., less than 10 uM); 36 however, some studies that used higher concentrations were included for comparative purposes. 37 The focus was on moderate- to relatively low-dose studies because it was felt that studies that 38 required arsenic concentrations greater than 10 uM to produce a biological response in vitro 39 would be less likely to be relevant to the health effects related to chronic ingestion of arsenic in 40 drinking water. NRC (2001) concluded that "The mechanistic studies reviewed herein and those 41 reviewed previously in the 1999 NRC report suggest that trivalent arsenic species (primarily 68 DRAFT—DO NOT CITE OR QUOTE ------- 1 As111, MMAm, and, possibly, DMA111) are the forms of arsenic of greatest toxicological concern." 2 They estimated concentrations of arsenic that could be expected in human urine from the known 3 human experience and concluded that "Arsenite concentrations in excess of 10 uM generally 4 exceed concentrations that can occur in the urine of individuals chronically exposed to arsenic in 5 drinking water and have less direct relevance to understanding the modes of action responsible 6 for human cancer induced by this route of exposure." They also stated that: 7 8 • "Experiments in animals and in vitro have demonstrated that arsenic has many 9 biochemical and cytotoxic effects at low doses and concentrations that are potentially 10 attainable in human tissues following ingestion of arsenic in drinking water. Those 11 effects include induction of oxidative damage to DNA; altered DNA methylation and 12 gene expression; changes in intracellular levels of murine double minute 2 proto- 13 oncogene (mdm2) protein and p53 protein; inhibition of thioredoxin reductase (TrxR; 14 MMA111 but not As111); inhibition of pyruvate dehydrogenase; altered colony-forming 15 efficiency; induction of protein-DNA cross-links; induction of apoptosis; altered 16 regulation of DNA-repair genes, thioredoxin, glutathione reductase, and other stress- 17 response pathways; stimulation or inhibition of normal human keratinocyte cell 18 proliferation, depending on the concentration; and altered function of the glucocorticoid 19 receptor." 20 21 Despite the extensive research on MO A up to that time, NRC stated that "the 22 experimental evidence does not allow confidence in distinguishing between various shapes 23 (sublinear, linear, or supralinear) of the dose-response curve for tumorigenesis at low doses." 24 The present review uses the terms "mode of action" and "key event" as they are 25 described in the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a). According to 26 EPA, '"mode of action' is defined as a sequence of key events and processes, starting with 27 interaction of an agent with a cell, proceeding through operational and anatomical changes, and 28 resulting in cancer formation. A 'key event' is an empirically observable precursor step that is 29 itself a necessary element of the mode of action or is a biologically based marker for such an 30 element. Mode of action is contrasted with 'mechanism of action', which implies a more 31 detailed understanding and description of events, often at the molecular level, than is meant by 32 mode of action. The toxicokinetic processes that lead to formation or distribution of the active 33 agent to the target tissue are considered in estimating dose, but are not part of the mode of action 34 as the term is used here. There are many examples of hypothesized modes of carcinogenic 35 action, such as mutagenicity, mitogenesis, inhibition of cell death, cytotoxicity with reparative 36 cell proliferation, and immune suppression." 37 In this review, tables have been compiled in order to make a large amount of information 38 on the biological effects of inorganic arsenic readily available. Appendix C contains tables that 39 deal with in vivo human studies (Table C-l), in vivo experiments on laboratory animals (Table 40 C-2), and in vitro studies (Table C-3). These tables include as many experiments published from 69 DRAFT—DO NOT CITE OR QUOTE ------- 1 2005 through August 2007 as possible. Numerous earlier experiments have been included as 2 well, based on various selection criteria: being mentioned in the SAB Arsenic Review Panel 3 comments of July 2007 (SAB, 2007) or in NRC's update (NRC, 2001), or inclusion in an earlier 4 draft that lacked tables (U.S. EPA, 2005c). The tables provide information on: (1) the arsenic 5 species tested; (2) the cell types, tissues, or species tested; (3) all concentrations or doses tested; 6 (4) all durations of exposure; (5) estimates of the LOEC or LOEL (i.e., lowest observed effect 7 concentration or level); (6) a summary of the most important results of each study; and (7) the 8 citations. The 22 categories into which the hypothesized key events are grouped in those tables 9 are listed in column 1 of Table 4-1, and the number of data rows under each category provide an 10 estimate of the amount of available data pertaining to each category topic. Data from a single 11 publication are sometimes entered under multiple event categories. For example, the results in 12 Wang et al. (1996) are summarized in rows under Apoptosis, Cytotoxicity, and Effects Related 13 to Oxidative Stress (ROS). 14 When judging the possible relevance of in vitro experiments or in vivo laboratory animal 15 experiments on human health, it is useful to keep in mind that the total concentration of As111 and 16 Asv in drinking water pumped from tube wells in Bangladesh (as an example of one country 17 with high exposures to inorganic arsenic in drinking water) ranges from 20 to over 2,000 ppb 18 arsenic (i.e., 0.3 to 27 uM). In people exposed at those high levels, total blood arsenic levels 19 range from 0.5 to 1.2 uM (Snow et al., 2005), and total arsenic concentrations in urine would 20 probably not exceed 10 uM (NRC, 2001). 21 70 DRAFT—DO NOT CITE OR QUOTE ------- Table 4-1. Summary of Number of Rows Derived From Peer-Reviewed Publications for Different Hypothesized Key Events3 Hypothesized Key Events Aberrant Gene or Protein Expression13 Apoptosis Cancer Promotion Cell Cycle Arrest or Reduced Proliferation Cell Proliferation Stimulation Chromosomal Aberrations and/or Genetic Instability Co-carcinogenesis Co-mutagenesis Cytotoxicity DNA Damage DNA Repair Inhibition or Stimulation Effects Related to Oxidative Stress (ROS) Enzyme Activity Inhibition Gene Amplification Gene Mutations Hypermethylation of DNA Hypomethylation of DNA Immune System Response Inhibition of Differentiation Interference With Hormone Function Malignant Rransformation or Morphological Transformation Signal Transduction Number of Rows in Tables In Vivo Human Studies (Table C-l) 6 1 0 0 0 13 0 0 0 5 2 2 0 0 1 2 1 1 0 0 0 1 In Vivo Experiments Using Laboratory Animals (Table C-2) 32 6 3 1 18 3 2 1 2 6 0 30 0 0 2 1 2 0 0 1 0 2 In Vitro Experiments (Table C-3) 124 78 3 29 21 83 3 21 118 35 11 69 5 5 7 2 7 46 13 7 13 51 a Details of the studies are presented in Appendix C. b Some hypothesized key events are shown in boldface to emphasize that in at least one of the tables they contain much more data than the other categories. 4.4.1.1. In Vivo Human Studies 1 Table C-l summarizes in vivo human studies. Here and elsewhere in the consideration of 2 human studies there was particular interest in the subset of people who develop skin lesions 3 (usually keratoses, which are often considered premalignant, or hyperpigmentation) following 4 long-term exposure to inorganic arsenic in drinking water. Indeed, four of the six studies related 5 to Aberrant Gene or Protein Expression compared groups of people with and without arsenic- 6 related skin lesions following similar exposures to high levels of inorganic arsenic in drinking 7 water, and in three cases, they also compared them to groups of people with much lower 71 DRAFT—DO NOT CITE OR QUOTE ------- 1 inorganic arsenic exposure levels. The genomics study by Argos et al. (2006) showed that 312 2 more genes were down-regulated in the group with skin lesions than in the inorganic arsenic- 3 exposed group without such lesions. No genes were shown to be up-regulated. Other studies 4 showed increased levels of the EGFR-ECD protein (i.e., extracellular domain of the epidermal 5 growth factor receptor) in serum (Li et al., 2007), increased levels of transforming growth factor 6 alpha (TGF-a) protein in bladder urothelial cells (Valenzuela et al., 2007), and decreased levels 7 of three integrins in and around skin lesions following exposures to inorganic arsenic in drinking 8 water (Lee et al., 2006b). Integrins are important in the control of differentiation and 9 proliferation of the epidermis. Many skin diseases, including arsenical keratosis, show altered 10 patterns of integrin distribution and expression. In the first two instances, there were bigger 11 increases in the group with skin lesions. The study on integrins only made comparisons to a 12 control group. One of the other studies showed a decrease in the concentration of the receptor 13 for advanced glycation end products (RAGE) protein in sputum when there was a higher 14 concentration of inorganic arsenic in the urine (Lantz et al., 2007). Changes in that biomarker 15 are related to several chronic inflammatory diseases in the lung, including lung cancer. The 16 remaining study showed that two oncogenes were up-regulated in tumor tissues in patients with 17 arsenic-related urothelial cancer, but not in those from patients with non-arsenic-related 18 urothelial cancer (Hour et al., 2006). 19 The Chromosomal Aberrations and/or Genetic Instability category has the most entries in 20 the table on human studies. Although some of the studies found no effects (usually on SCE 21 induction) in people exposed to inorganic As, most of the studies included in the table showed 22 clear increases of chromosomal aberrations (C A) in lymphocytes, micronuclei (MN; in various 23 cell types), or both CA and MN in people who had been exposed to high levels of inorganic 24 arsenic in drinking water or to Fowler's solution (i.e., a solution containing 1% arsenic that was 25 commonly used as a medicine in the 1800s and early 1900s). Arsenic was shown to increase the 26 incidence of MN specifically in bladder cells (Warner et al., 1994; Moore et al., 1996, 1997b). 27 There also was suggestive evidence that some arsenic-induced MN (a minority of them) result 28 from aneuploidy (Moore et al., 1996). There was some evidence for induction of SCE. Three of 29 the papers showed that those persons with arsenic-induced skin lesions had higher frequencies of 30 induced chromosomal damage seen either as CA or MN than those without lesions (Gonsebatt et 31 al., 1997; Ghosh et al., 2006; Banerjee et al., 2007). It is intriguing that one of the studies 32 demonstrated an apparent predisposition to both skin lesions and CA that was correlated with 33 (and was thus perhaps caused by) a single polymorphism of the ERCC2 (excision repair cross- 34 complementing rodent repair deficiency gene, complementation group 2) gene, which plays a 35 key role in the nucleotide excision repair (NER) pathway. The polymorphism resulted from an 36 A—>C mutation at codon 751 that caused a change from lysine to glutamine, and the allele 37 conferring the higher predisposition in homozygotes had the remarkably high gene frequency of 72 DRAFT—DO NOT CITE OR QUOTE ------- 1 0.40 in that population (Banerjee et al., 2007). Although only some of the homozygotes heavily 2 exposed to inorganic arsenic in drinking water developed skin lesions or had chromosomal 3 aberrations, those that were affected had both endpoints. 4 Table C-l also provides data showing that oral inorganic arsenic exposure increases 5 DNA damage. Two papers reported oxidative damage to DNA revealed by increases in the 6 concentration of 8-hydroxydeoxyguanosine (8-OHdG) in the urine. Both studies were in Japan, 7 with the first showing a positive correlation between urinary concentrations of arsenic and 8- 8 OHdG after analyzing samples from 248 people in the general population (Kimura et al., 2006). 9 The other study (Yamauchi et al., 2004) involved clinical examination of 52 patients following 10 an incident in which 63 people (four of whom died within about 12 hours of being poisoned) 11 were poisoned by eating food contaminated with ATO. Those 52 patients were followed up for 12 various effects including levels of 8-OHdG in urine. Maximal levels of-150% compared to 13 normal Japanese levels were reached 30 days after the exposure, and by 180 days the levels had 14 returned to normal. The same paper reported that people in Inner Mongolia, China, who drank 15 water contaminated with about 130 ppb arsenic had a significant increase in urinary 8-OHdG, 16 which returned to normal after they drank "low-arsenic" water for one year. 17 Table C-l includes data that demonstrate DNA damage (i.e., single-strand breaks) 18 detected by the single cell gel electrophoresis (SCGE) comet assay. One of those studies, in 19 which the high-exposure group drank water containing about 247 ppb As, also included a comet 20 assay combined with formamidopyrimidine-DNA glycosylase (FPG) digestion and thereby 21 showed that arsenic also induced oxidative base damage. (Digestion with the FPG enzyme 22 breaks the DNA at the sites of oxidative damage so that those sites are seen in this modified 23 comet assay.) Besides looking at baseline DNA damage, the other comet study investigated the 24 capacity of the lymphocytes of subjects who used drinking water containing 13-93 ppb arsenic 25 to repair damage induced by an in vitro challenge with the mutagen 2- 26 acetoxyacetylaminofluorene (2-AAAF). Adducts formed following treatment with 2-AAAF are 27 primarily repaired through the NER pathway and lymphocytes from arsenic-exposed individuals 28 had more adducts. The lymphocytes from the people with high-arsenic exposure had reduced 29 NER ability (Basu et al., 2005). The remaining DNA damage study (Mo et al., 2006) used 8- 30 oxoguanine DNA glycosylase (OGG1) expression as an indicator of oxidative-induced DNA 31 damage. The OGG1 gene codes for an enzyme involved in base excision repair (BER) of 32 residues that result from oxidative damage to DNA. OGG1 expression was found to be closely 33 linked to the levels of arsenic in drinking water and in toenails, thereby indicating a link between 34 ROS damage to DNA and inorganic arsenic exposure. An inverse relationship between OGG1 35 expression and selenium (Se) levels in toenails was found, which suggests a possible protective 36 effect of Se against arsenic-induced oxidative stress. As was often the case when populating the 37 MO A tables in Appendix C, some studies could equally well be placed into one or another 73 DRAFT—DO NOT CITE OR QUOTE ------- 1 hypothesized key event category, and clearly some studies listed under DNA Damage also relate 2 to the hypothesized key events of DNA Repair Inhibition or Stimulation and Effects Related to 3 Oxidative Stress (ROS). 4 In another polymorphism study, homozygotes for two different alleles of the p53 gene 5 were shown to be at higher risk (than those carrying other alleles) of developing arsenic-induced 6 keratosis among individuals who used drinking water that contained roughly 180 ppb arsenic 7 (De Chaudhuri et al., 2006). Because that gene is so important in controlling apoptosis, that 8 study was listed under Apoptosis. It is unclear, however, why mutations at that gene would 9 predispose those who consume high levels of arsenic to develop skin lesions. Two studies 10 described under DNA Repair Inhibition or Stimulation demonstrated reduced expression of three 11 nucleotide excision repair (NER) genes in a population that used drinking water that contained 12 10-75 ppb arsenic (Andrew et al., 2003, 2006). Still more evidence that arsenic causes Effects 13 Related to Oxidative Stress (ROS) comes from school children in Taiwan who showed a positive 14 correlation between urinary concentrations of arsenic and 8-OHdG; no information was provided 15 regarding the level of arsenic in their drinking water (Wong et al., 2005). Subjects with arsenic- 16 related skin lesions from a population in Inner Mongolia, China, that used drinking water with a 17 mean of 158 ppb arsenic showed a statistically significant positive correlation between 8-OHdG 18 adducts in their urine and individual urinary concentrations of inorganic As, MMA, and DMA. 19 In contrast, those without skin lesions showed no correlation (Fujino et al., 2005). 20 Evidence is presented under Hypermethylation of DNA that arsenic exposure causes 21 hypermethylation of the promoter sequence in the DNA for four tumor suppressor genes. For 22 two of the genes, p53 and pi 6, there was a positive dose-response between arsenic 23 contamination of drinking water and the level of effect; however, this was only seen in 24 individuals with skin lesions (Chanda et al., 2006). For the other two genes, RASSF1A and 25 PRSS3, the association was demonstrated with regard to the level of arsenic consumption 26 estimated from toenail clippings (Marsit et al., 2006). Because the Marsit et al. (2006) study was 27 done on bladder cancer patients, it provides a potential link between arsenic exposure and 28 epigenetic alterations in patients with bladder cancer. The Chanda et al. (2006) study also 29 demonstrated hypomethylation in a few individuals, but it was found only in persons having 30 prolonged arsenic exposure at high doses. 31 Regarding the hypothesized key event category Immune System Response, there was 32 suggestive evidence of an association between changes in sensitive markers of lung 33 inflammation (i.e., metalloproteinase concentrations in induced sputum) and levels of only about 34 20 ppb of arsenic in drinking water. The initial comparison between the high- and low-level 35 exposure towns showed no difference with regard to these biomarkers, but a significant 36 association appeared when the analysis was adjusted for possible confounding factors (Josyula et 37 al., 2006). Islam et al. (2007) found that IgG and IgE levels were significantly elevated in 74 DRAFT—DO NOT CITE OR QUOTE ------- 1 arsenic-exposed individual with skin lesions. More details about that experiment, including 2 clinical findings possibly related to inflammatory reactions, are found in Appendix D. Appendix 3 D discusses several other studies (including in vitro experiments and experiments on laboratory 4 animals) related to immunotoxicity, including some that are not included in any of the tables in 5 Appendix C. 6 The only study listed under Gene Mutations gave no more than a hint of an effect 7 (Ostrosky-Wegman et al., 1991). Regarding Signal Transduction, a study in Taiwan showed that 8 both the levels of plasma TGF-a and the proportion of individuals with TGF-a over-expression 9 were significantly higher in the high CAE group than in the control group (Hsu et al., 2006). 10 Only limited information from the cited experiments has been included in this discussion. 11 Much more detail on these studies can be found in Table C-l of Appendix C as well as in Table 12 C-2 for in vivo experiments using laboratory animals and Table C-3 for in vitro experiments. 13 Brief discussions of the information in Table C-2 and C-3 are found in Sections 4.4.1.2 and 14 4.4.1.3, respectively. 4.4.1.2. In Vivo Experiments Using Laboratory Animals 15 Table C-2 summarizes in vivo experiments using laboratory animals. All doses given in 16 this section are stated in terms of the amount of arsenic in the dose. Twenty-four of the 112 rows 17 in Table C-2 involve studies of nine key event categories in mice that drank water containing 18 arsenic for several to many weeks. Results are of particular interest because they involved most 19 of the lowest dose levels tested, and As111 is the most toxic oxidation state of inorganic As. 20 Figure 4-1 summarizes the results according to key events by showing, for each endpoint, the 21 concentration of arsenic in the water that was the LOEL, the period of treatment, and the organ 22 or tissue in which the effect was seen. Because the result for gene mutations was a negative 23 finding, it is not shown in the figure. Sometimes more than one entry in Table C-2 corresponds 24 to a single item in the figure, and sometimes a single entry in the table deals with separate groups 25 of animals. Consequently, there may be multiple LOELs shown in the figure. It should also be 26 kept in mind that sometimes only one dose was tested in an experiment, and, of course, if an 27 effect was found, that dose became the LOEL (even though a much lower dose might have been 28 effective). One benefit of the detailed descriptions found in Table C-2 is that all doses tested are 29 listed. As Figure 4-1 shows, roughly half the dose levels used exceed 2,000 ppb and are thus 30 much higher than levels ever found in drinking water used for human consumption. While all of 31 the experiments summarized in Table C-2 are useful in terms of showing their effects in mice, 32 this discussion gives more attention to doses that overlap higher levels of exposure to humans 33 from drinking water. A better understanding of the pharmacokinetic characteristics in different 34 species may aid in determining the relevance of the high-dose animal studies to human subjects 35 exposed to arsenic in drinking water at lower concentrations for a longer period. 75 DRAFT—DO NOT CITE OR QUOTE ------- ppm A 1000 100 10 1 0.1 0.01 0.001 0.0001 Cancer Promotion = CP Co-carcinogenesis = CC Aberrant Gene or or Cell Cycle Arrest or Cell Proliferation or Effects Related to Protein Expression Hypomethylation of DNA Reduced Proliferation Stimulation Co-mulagenesis Oxidative Stress (ROS) • 4w, S CP«14w, S • 10w, S • 48w,Li »48w.Li »4w'B •'6w'B • Sw.Bl.K.and • 8w, Lu .., „ •13w,Bl,Br, • 16w,B an(] u • 10 w, S CC»29w, S • 26 w, Li • 20w,P •20w,VH CC»23w,S • 5w, H - • 4w, Lu;10w, H;20w, H •Sw, V • 9w, TS »8w, TS • 5», V _ • LOEL . B-bladder Br-brain It-kidney Lu-lung S-skin V-blood vessels Effects seen in: w-Week(s) Bl-blood H-heart Li-liver P-blood plasma T-tumor tissue K, Figure 4-1. Level of significant exposure of adult mice to sodium arsenite in drinking water in ppm As. 2 The Aberrant Gene or Protein Expression effects seen at those lower levels included 3 increases in levels of several proteins and in mRNA levels of a few genes that are important in 4 angiogenesis and remodeling. For example, vascular endothelial cell growth factor [VEGF] and 5 its receptors VEGFR1 and VEGFR2 were measured in hearts, and increases were sometimes 6 restricted to areas around blood vessels (Kamat et al., 2005; Soucy et al., 2005). However, 7 increases in dose (up to 0.5 ppm in drinking water) and duration (up to 20 weeks) actually 8 caused decreases in the protein and mRNA levels for VEGFR1 and VEGFR2, suggesting that 9 chronic exposure at these higher levels was toxic to the cardiac vasculature in mice. Consistent 10 with the decreased mRNA levels seen for VEGFR1 and VEGFR2 following 20-week chronic 11 exposures to 0.5 ppm, the same treatment regimen produced evidence of reduced cell 12 proliferation, which was represented as a decrease in the density of microvessels of less than 12 76 DRAFT—DO NOT CITE OR QUOTE ------- 1 um in the heart (Soucy et al., 2005). These data thus provide an interesting example of the 2 concentration and time-dependent effects of arsenic exposure that might be important in the 3 etiology of some of the diseases that it causes. In contrast, stimulation of cell proliferation at 4 low-dose levels involved increases in (1) blood vessel number in Matrigel implants (Soucy et al., 5 2005), (2) tumor growth rates after implantation of tumor cells (Kamat et al., 2005), and (3) 6 number of metastases to the lungs after implantation of those tumor cells (Kamat et al., 2005). 7 Proteomic analysis of bronchoalveolar lavage fluid from lungs of mice that drank 0.05 8 ppm (i.e., 50 ppb) arsenic in water for 4 weeks showed an increase in peroxiredoxin-6 and 9 enolase 1 levels and a decrease in GSTO1, RAGE, contraspin, and apolipoproteins A-I and A-IV 10 (Lantz et al., 2007). That same paper had demonstrated a decrease in the level of RAGE protein 11 in human sputum that was associated with arsenic exposure. Two microarray experiments at 12 much higher dose levels of 28.8 and 45 ppm showed changes in expression of dozens of genes 13 (Chen et al., 2004b; Lantz and Hays, 2006). In each experiment, the LOEL was the only dose 14 tested, which leaves open the possibility that such high doses might not have been necessary to 15 obtain these changes. 16 Mice that were exposed for 23 weeks to 0.7-5.8 ppm arsenic in drinking water developed 17 no skin tumors; however, when they were also exposed to UV thrice weekly for most of that 18 time, they showed a strong dose-related increase up through 2.9 ppm As, thus providing strong 19 evidence of co-carcinogenesis (Burns et al., 2004). Another part of the same study (reported in 20 Uddin et al., 2005) demonstrated that at 2.9 ppm there was oxidative DNA damage caused by the 21 co-treatment. Effects Related to Oxidative Stress (ROS) following 26 weeks of exposure at 1.8 22 ppm included decreases in GSH content, and in the activities of glucose-6-phosphate 23 dehydrogenase (G6PDH), glutathione peroxidase (GPx), and plasma membrane Na+/K+ 24 ATPase. Additional changes suggestive of such damage, such as an increase in the 25 concentration of malondialdehyde (MDA), were apparent after 9, 12, or 15 months at the same 26 dose level (Mazumder, 2005). 27 Eighteen of the 112 rows in Table C-2 involved rats that drank water containing sodium 28 arsenite for several to many weeks, but those studies are distributed among only two key event 29 categories and do not extend down to nearly as many effects at low exposure levels. Most 30 experiments cited in the 18 rows involved drinking water containing 57.7 ppm arsenic for 31 several to many weeks and showed findings of numerous changes indicative of oxidative damage 32 in several organs. A few experiments show differing levels of oxidative damage in different 33 regions of the brain (Samuel et al., 2005; Shila et al., 2005a,b). By far the lowest dose tested 34 among these experiments was 0.03 ppm As, and it was found to be effective in decreasing the 35 GSH level and superoxide radical dismutase (SOD) activity in the liver. The other two dose 36 levels tested, 1.4 and 2.9 ppm, caused bigger changes in these two variables, as well as 37 additional changes indicative of oxidative stress. It is of interest that the changes per unit dose 77 DRAFT—DO NOT CITE OR QUOTE ------- 1 were much higher for GSH and SOD at 0.03 ppb than they were at the two much higher doses 2 tested (Bashir et al., 2006a). In experiments using 5.8 ppm As, which rats drank for 4, 8, or 12 3 weeks, activities of catalase (CAT) and SOD in kidney, liver, and RBCs were found to be 4 elevated at 4 weeks, but they decreased to baseline levels or lower by 12 weeks; MDA levels 5 were always elevated (Nandi et al., 2006). Consumption of water containing 1.4 ppm arsenic for 6 60 days led to a demonstrable increase in apoptosis in liver cells (Bashir et al., 2006a). 7 Twenty-six of the 112 rows in Table C-2 involve rats or mice that consumed pentavalent 8 arsenicals (Asv, MMAV, DMAV, or TMAV) for several to many weeks, and in all but three rows 9 they were delivered in drinking water instead of food. As would be expected for these less 10 potent forms of arsenic, LOELs were typically high and usually above 50 ppm. Only a few 11 results occurred at much lower concentrations, and are mentioned in this discussion. After rats 12 were exposed for 28 days to 0.35 ppm arsenic in drinking water in the form of DMAV, 13 microarray analysis demonstrated significant effects on the expression of 503 genes (i.e., 11% of 14 the genes tested with that microarray) in urothelial cells. Even more genes were affected at the 15 three higher doses tested (i.e., 1.4, 14, and 35 ppm As). Most of the effected genes related to the 16 functional categories of apoptosis, cell cycle regulation, adhesion, signal transduction, stress 17 response, or growth factor and hormone receptors. There was a change in the types of genes 18 affected at the different doses, particularly when comparing the higher two doses (both 19 cytotoxic) with the two non-cytotoxic doses (Sen et al., 2005). When rats were exposed to 0.24 20 ppm Asv for 1 or 4 months in drinking water, changes in signal transduction were increased 21 expression of integrin-linked kinase (ILK) and decreased expression of phosphatase and tensin 22 homolog (PTEN) in the liver. At higher doses, the expression of these genes and additional 23 cancer-related genes was affected (Cui et al., 2004b). 24 DNA damage (both fragmentation and oxidative) was demonstrated in peripheral blood 25 leukocytes of mice using the comet assay following exposure of 50, 200, or 500 ppb arsenic in 26 drinking water in the form of Asv for 3 months with and without a low-Se diet. Arsenic caused 27 increased DNA fragmentation only in mice consuming the low-Se diet, and induced oxidative 28 damage only in mice consuming the normal-Se diet. Neither case showed a positive dose- 29 response (Palus et al., 2006). In lung adenocarcinomas from mice exposed for 18 months to 30 0.24, 2.4, or 24 ppm Asv in drinking water, there was an increase in the extent of 31 hypermethylation of promoter regions of tumor suppressor genes p!6INK4a and RASSF1A 32 (genes frequently found inactivated in many types of cancer including lung cancer), based on 33 methylation-specific polymerase chain reaction (PCR). All doses had an effect, and there was a 34 positive dose-response. Reduced expression or lack of expression of these two genes was 35 correlated with the extent of hypermethylation. Mice without tumors, whether control or 36 arsenic-treated, had normal (i.e., not reduced or eliminated) expression of these genes in their 78 DRAFT—DO NOT CITE OR QUOTE ------- 1 lungs. The authors concluded that epigenetic changes of tumor suppressor genes are involved in 2 inorganic arsenic-induced lung carcinogenesis (Cui et al., 2006). 3 Of the experiments described in Table C-2 in which arsenic exposure occurred through 4 consumption of arsenic in drinking water or food, the only group not yet discussed consists of 5 the series of experiments in which pregnant female mice drank water containing 42.5 or 85 ppm 6 arsenic in the form of sodium arsenite for 10 days on gestation days 8 to 18. These studies 7 follow up on the interesting observation that arsenic seems to be a complete carcinogen in mice 8 following such a treatment. The offspring were observed for effects (sometimes only after they 9 had grown to be adults), and results are categorized in Table C-2 under Aberrant Gene or Protein 10 Expression, Cell Proliferation or Stimulation, Hypomethylation of DNA, and Signal 11 Transduction. Some of the more noteworthy findings were as follows. Numerous microchip 12 analyses were conducted, often with some of the findings confirmed by real-time (RT) PCR. 13 Microarrays containing from 588 to 22,000 genes were used. It was not unusual to find changes 14 in the expression of scores of genes (sometimes even of thousands) in the different studies. 15 Changes (often many-fold) included both increases and decreases of expression, occurring at 16 both dose levels. Some of the many types of genes often altered included oncogenes, HCC 17 biomarkers, cell proliferation-related genes, stress proteins, insulin-like growth factors, estrogen- 18 linked genes, and genes involved in cell-cell communication. Tissues in which gene expression 19 changes were found in offspring that had been exposed to arsenic in utero included: (1) arsenic- 20 induced HCC tumors that developed in adult males, (2) normal-appearing cells in livers of adult 21 males, (3) fetal livers of males right at the end of treatment, (4) livers of newborn males, (5) fetal 22 lungs of females right at the end of treatment, and (6) arsenic-induced adenomas and 23 adenocarcinomas that developed in lungs of adult females. 24 The expression of three estrogen-related genes was shown to increase synergistically in 25 the uteri of females (at 11 days of age) that had been exposed in utero to arsenic and also 26 subcutaneously injected with diethylstilbestrol (DBS) on the first 5 days after birth. These and 27 other results showed that inorganic arsenic acts with estrogens to enhance production of 28 urogenital cancers in female mice (Waalkes et al., 2006a). Females that had been exposed to 29 arsenic in utero and then received a 21-week post-weaning treatment with TPA showed changes 30 in gene expression that were similar to those seen in liver samples from males that had received 31 only the arsenic treatment in utero. This is interesting because it parallels another situation in 32 which TPA-treated females showed a response similar to males without TPA treatment. 33 Specifically, female mice exposed in utero to arsenic develop HCC only after TPA treatment 34 (Liu et al., 2006b); however, male mice exposed in utero to arsenic develop those tumors without 35 receiving any TPA treatment. Observed changes in estrogen-related genes sometimes seemed 36 especially important in the interpretation of results, and fetal lungs of females exposed to arsenic 37 in utero showed a large increase in estrogen receptor-alpha (ER-a), as well as several other 79 DRAFT—DO NOT CITE OR QUOTE ------- 1 estrogen-related genes and numerous other genes, including some associated with lung cancer. 2 There also was a large increase in nuclear ER-a in adenomas and adenocarcinomas that 3 developed in the lungs of adult females that had been exposed to arsenic in utero (Shen et al., 4 2007). 5 Stimulation of cell proliferation during treatment of males while in utero at 85 ppm 6 induced kidney cystic tubular hyperplasia in 23% of the animals, and although males did not 7 develop bladder hyperplasia from the arsenic treatment alone, they often did if treated in 8 conjunction with DBS or tamoxifen on the first 5 days after birth because of a synergistic 9 interaction that occurred with those chemicals. Although females exposed while in utero showed 10 bladder hyperplasia similar to the males, arsenic exposure in utero alone caused no hyperplasia 11 in their kidneys (Waalkes et al., 2006a,b). Global hypomethylation of GC-rich regions was 12 demonstrated in livers of newborn males that received 85 ppm in utero (Xie et al., 2007). 13 Almost all remaining experiments summarized in Table C-2 involved treatments of mice 14 or rats by gavage, and those results are summarized under Aberrant Gene or Protein Expression, 15 Apoptosis, Chromosomal Aberrations and/or Genetic Instability, Effects Related to Oxidative 16 Stress (ROS), and Interference With Hormone Function. In all rows where As111 was 17 administered, it was usually as sodium arsenite, but sometimes as arsenic trioxide (ATO). One 18 study also included treatment with pentavalent arsenicals. By using gavage, the amount of the 19 arsenical administered to each animal was controlled precisely, and it was given as a certain 20 weight of arsenic per animal, often with adjustment to the individual weight of each animal (i.e., 21 ug/animal or mg/kg bw, respectively). Most treatments were administered repeatedly, with 22 treatment regimens in one case lasting an entire year. As in all other studies on experimental 23 animals, there was an attempt here to state all doses in terms of the amount of arsenic. Because 24 it was unclear from the reporting of a few experiments whether doses were expressed as arsenic 25 compound or as As, Table C-2 always makes it clear whether or not such a correction was made. 26 In a gavage study with one of the smallest amounts of arsenic per dose (equivalent to 36 27 ug/mouse if a mouse weighed 25 g), Patra et al. (2005) found induction of chromosomal 28 aberrations in mice that received 1.44 mg As/kg bw given as sodium arsenite by gavage once- 29 per-week for 4 weeks. Induction of chromosomal aberrations also was seen after 5 and 6 30 treatments; however, 7 and 8 treatments were lethal to the mice. A 25 g mouse in that study 31 would have received the same amount of arsenic in that one day if it had drunk water that 32 contained 6 ppm arsenic (assuming that it drank 6 mL of water, which would be a reasonable 33 amount for a mouse). 34 In the only gavage study with in utero treatments, 9 daily treatments of 4.35 mg As/kg 35 bw was shown to increase the activity of the selenoprotein iodothyronine deiodinase-II (DI-II) in 36 fetal brains and to decrease the activity of the selenoprotein TrxR in fetal livers. In both cases, 37 these results were observed only if the mice were on a Se-deficient diet (Miyazaki et al., 2005). 80 DRAFT—DO NOT CITE OR QUOTE ------- 1 In a gavage study lasting a full year (Das et al., 2005), mice were administered 50, 100, or 150 2 ug/mouse, 6 days a week for 3, 6, 9, or 12 months; it took 9 months before substantial increases 3 were seen in the activities of tumor necrosis factor alpha (TNF-a) and interleukin (IL)-6 at any 4 dose, but by then all doses had an effect and there was a positive dose-response. Three months 5 later, both effects had increased substantially at all doses, still with a positive dose-response. A 6 similar response was seen for the concentration of total collagen, although increases were not as 7 large in comparison to the control group. That same study examined six components of the 8 antioxidant defense system and found numerous interesting changes over time. While all of the 9 affected components had a LOEL of 50 ug at the 3-, 9-, and 12-month test periods, all five 10 affected components had a LOEL of 100 ug at 6 months. GSH levels and activities of GPx and 11 CAT increased by 3 months, but decreased by 9 and 12 months. In another experiment with 12 single, large doses of As111 or Asv given to mice by gavage, there were large increases in heme 13 oxygenase 1 (HMOX-1) activity within 6 hours in liver and kidney but not in the brain. The 14 effect was somewhat higher with As111, but DMAV had no effect. This study also tested some 15 much smaller doses, and a dose as high as 2.25 mg/kg bw had no effect on this endpoint in 16 kidneys (Kenyon et al., 2005b). 17 Various biochemical indicators of apoptosis were seen in brain and liver 24 hours after 18 giving rats a single high dose of sodium arsenite by gavage (Bashir et al., 2006b). The same 19 paper showed that single, large doses of sodium arsenite given to rats by gavage affected many 20 biochemical indicators of oxidative stress in liver and brain 24 hours after treatment. Some 21 studies on Effects Related to Oxidative Stress (ROS) included co-treatments with antioxidants 22 that were shown to reduce the level of effects seen (Modi et al., 2006; Sohini and Rana, 2007). 23 With regard to Interference With Hormone Function, rats given 30.3 mg Asin/kg bw as ATO by 24 gavage every other day for 30 days were shown to have a large increase in the levels of thyroid 25 hormones triiodothyronine (T3) and thyroxine (T4) in their blood serum (Rana and Allen, 2006). 4.4.1.3. In Vitro Experiments 26 Table C-3 summarizes a large number of in vitro experiments; and some highlights are 27 discussed below. The potencies of many arsenicals, including both trivalent and pentavalent 28 forms, have been compared in several series of experiments, with the obvious conclusion that the 29 pentavalent forms almost always have much higher LOECs (e.g., Moore et al., 1997a; Sakurai et 30 al., 1998; Petrick et al., 2000; Drobna et al., 2002; Kligerman et al., 2003). Consequently, the 31 discussion below does not focus on the studies that analyzed pentavalent arsenicals. 32 Three chemical properties of arsenic likely to account for its biological activity are: 33 (1) the soft acid/soft base principle (which is related to trivalent arsenicals and sulfhydryl 34 binding); (2) the nucleophilicity of trivalent arsenicals; and (3) the formation of free radicals, 35 ROS, or both by arsenicals (Kitchin et al., 2003). As noted by Kitchin et al. (2003): 81 DRAFT—DO NOT CITE OR QUOTE ------- 1 • "If trivalent arsenicals acting as soft acids are causally important, then the likely modes 2 of action of arsenic carcinogenesis may include altered DNA repair, altered growth 3 factors, cell proliferation, altered DNA methylation patterns and promotion of 4 carcinogenesis." 5 6 Arsenic is readily absorbed from the GI tract in humans and is primarily transported in 7 the blood bound to sulfhydryl groups in proteins and low-molecular-weight compounds, such as 8 amino acids and peptides (NRC, 1999). At any given time, about 99% of absorbed As111 is bound 9 to tissue sulfhydryls, mostly to monothiol sites (Kitchin and Wallace, 2006). Based on the 10 results of their peptide binding studies, Kitchin and Wallace (2006) suggested that dithiol- and 11 trithiol-binding sites would be "the most likely causal triggers of biological effects because of 12 their stronger affinity and because the bi- and tri-dentate complexes last so much longer than the 13 rapidly dissociating and reforming binding of arsenite to monothiol sites." While the As111 14 attachment to the monothiol-binding sites are short-lived, a substantial part of the total As111 15 attaches to those sites because of their great abundance in mammals. Because the functional 16 group of the amino acid cysteine in a protein or peptide is a thiol group, any proteins that contain 17 cysteine are of importance for interactions with As111. Although Table C-3 includes large 18 amounts of data under Effects Related to Oxidative Stress (ROS), arsenic's action as a soft acid 19 and its nucleophilicity are not included as key events. It is obvious, nonetheless, that those 20 chemical properties play important roles in the interactions of inorganic arsenic with organisms 21 at early stages in multiple key event(s) leading to tumor development. 22 Table C-3 summarizes a great deal of data under Aberrant Gene or Protein Expression. 23 Abundant evidence is presented showing that changes can easily occur at concentrations of As111 24 (as either sodium arsenite or arsenic trioxide) of less than 10 uM and often with durations of 25 exposure of 24 hours or less. Results from 10 microarray analyses are found in this category, 26 and they all demonstrated changes in expression of large numbers of genes, often numbering in 27 the hundreds. Two studies with longer exposures to especially low concentrations are of special 28 interest. In one study, NB4 cells were exposed to 0.5 uM ATO for periods up to 72 hours for 29 transcriptome analysis and up to 48 hours for proteomic analysis. The regulation of 487 genes 30 was affected at the transcriptome level; however, at the proteome level, 982 protein spots were 31 affected. The finding of more significant changes at the proteomic level, in comparison with the 32 relatively minor changes found at many of the corresponding genes at the transcriptome level, 33 suggests that ATO particularly enhances mechanisms of post-transcriptional/translational 34 modification (Zheng et al., 2005). In the second experiment, which was a cDNA 35 (complementary DNA) microarray analysis of about 2,000 genes, the LOECs for SV40 large T- 36 transformed human urothelial cells (SV-HUC-1) exposed to As111, MMAm, or DMA111 for 25 37 passages (with subculturing twice weekly) were found to be 0.5, 0.05, and 0.2 uM, respectively. 38 DMA111 was shown to have a substantially different gene profile from the other two arsenicals. 82 DRAFT—DO NOT CITE OR QUOTE ------- 1 Most genes were down-regulated by these arsenicals, and evidence suggested that the 2 suppression of two of these genes resulted from epigenetic hypermethylation (Su et al., 2006). 3 Since each finding is presented only one time in Table C-3, subjectivity was often involved in 4 the placement of data into the different key event categories. As a result, the densities of data in 5 the different categories presented in Table 4-1 are only approximate estimates. This situation 6 was especially common for several key event categories that have large densities of data: 7 Aberrant Gene or Protein Expression, Signal Transduction, and Effects Related to Oxidative 8 Stress (ROS). 9 Table C-3 also presents details on the genes and proteins affected and changes related to 10 dose and time. It also provides the possible significance of such changes, when available. A few 11 examples follow. When primary normal human epidermal keratinocytes (NHEK) cells were 12 exposed to 1 uM sodium arsenite for 24, 48, and 72 hours, there was an increase in focal 13 adhesion kinase (FAK) protein at 24 hours followed by a decrease to below the background level 14 at later times, with almost none being present at 72 hours (Lee et al., 2006b). The concentration 15 of some enzymes increased after exposures to 0.5 uM for 24 hours, but the concentrations 16 decreased at higher levels of exposure up to 25 uM (Snow et al., 2001). DuMond and Singh 17 (2007) demonstrated the same relationship for proliferating cell nuclear antigen (PCNA) with 18 exposures to sodium arsenite lasting 70 days. The expression of PCNA increased at 0.008 uM, 19 but decreased at 0.77 and 7.7 uM. Similar results have been observed for telomerase activity 20 (Zhang et al., 2003). Numerous studies investigated effects of various modulators or inhibitors 21 or of different genetic conditions (e.g., knockout mutations or transfections). Cell type can have 22 a major influence on the effect of arsenic on protein expression, as was shown for p53 23 expression, with some cells having no response to 50 uM sodium arsenite for 24 hours while 24 other cells showed an increase after exposure to only 1 uM sodium arsenite (Salazar et al., 1997). 25 Clearly, small levels of arsenic exposure can have large effects on many genes and proteins, and 26 the relationships involving time and dose can be complicated and subject to many influences. 27 Results found in the Apoptosis category show that ATO and sodium arsenite can often 28 induce apoptosis in cells with exposures to less than 10 uM (often much less) for a few days or 29 less. Zhang et al. (2003) demonstrated a large difference in the sensitivity of cell lines to 30 arsenic-induced apoptosis. The authors found a positive association between telomerase activity 31 in cell lines and their susceptibility to induction of apoptosis by exposure to sodium arsenite. 32 Exposure to extremely low concentrations of sodium arsenite (i.e., 0.1-1 uM in HaCaT cells and 33 0.1-0.5 uM in HL-60 cells) for 5 days increased telomerase activity, maintained or elongated 34 telomere length, and promoted cell proliferation. At higher concentrations, exposure of these 35 cell lines to sodium arsenite for 5 days decreased telomerase activity, decreased telomere length, 36 and induced apoptosis. The positive association noted earlier means that cell lines that innately 37 have more telomerase activity are more likely to be affected by sodium arsenite in inducing 83 DRAFT—DO NOT CITE OR QUOTE ------- 1 apoptosis. Many experiments tested effects of modulators on the arsenic-induced apoptosis. For 2 example, Chen et al. (2006) demonstrated that co-treatment with L-buthionine-S,R-sulphoximine 3 (BSO) markedly increased induction of apoptosis, presumably because of its effect in decreasing 4 GSH levels. Other experiments looked at the effects of inhibitors of various proteins involved in 5 signal transduction pathways. For example, Lunghi et al. (2005) showed that use of MAP/ERK 6 kinase (MEK) 1 inhibitors greatly increased ATO-induced apoptosis. Other studies showed that 7 different genetic conditions established using knockout mutations or transfections could 8 markedly affect the extent of arsenic-induced apoptosis (e.g., Bustamante et al., 2005; Poonepalli 9 et al., 2005; Ouyang et al., 2007). Many of the experiments related to apoptosis were motivated 10 by the desire to improve methods for using ATO in cancer therapy, but in the process they have 11 provided much additional information about the complex pathways by which arsenic can affect 12 apoptosis. 13 In the hypothesized key event category Cancer Promotion, Tsuchiya et al. (2005) tested 14 sodium arsenite and three pentavalent arsenicals in a two-stage transformation assay in BALB/c 15 3T3 A31-1-1 cells. Sodium arsenite caused cancer promotion at aLOEC of 0.5 uM when the 16 initiating treatment was exposure to 0.2 ug/mL 20-methylcholanthrene for 3 days before the 18- 17 day post-treatment with sodium arsenite. Sodium arsenite caused promotion at a LOEC of 1 uM 18 when the initiating treatment was exposure to 10 uM sodium arsenite for 3 days before the 18- 19 day post treatment with sodium arsenite. When Asv was tested in the same way with the same 20 initiating treatments, it was somewhat less potent, with LOECs of 1 and 5 uM respectively. The 21 two methylated arsenicals had little or no effect. Paralleling their cancer promotion effects, the 22 same study demonstrated LOECs for As111 and Asv of 0.7 and 5 uM, respectively, for inhibition 23 of gap-junctional intercellular communication, which is a mechanism linked to many tumor 24 promoters. 25 The Cell Cycle Arrest or Reduced Proliferation category includes many experiments that 26 showed that levels of exposure to ATO and sodium arsenite of less than 10 uM (often much less) 27 for a few days or less can often increase the numbers of cells in mitosis and otherwise disrupt 28 mitosis, so as to reduce cell proliferation. In the Drobna et al. (2002) experiment, the LOECs for 29 reduced cell proliferation were 1,1, and 5 uM for 24-hour exposures to As111, MMA111, and 30 DMA111, respectively; no effects were seen following exposures to the pentavalent forms of these 31 arsenicals at 200 uM. By testing cells enriched in different phases of the cell cycle using 32 centrifugal elutriation, McCollum et al. (2005) showed that As111 slowed cell growth in every 33 phase of the cell cycle. Cell passage from any cell cycle phase to the next was inhibited by 5 uM 34 sodium arsenite. By looking at caspase activity, they showed that Asin-induced apoptosis 35 specifically in cell populations delayed in the G2/M phase. Tests with knockout mutations 36 showed that poly(adenosine diphosphate-ribose) polymerase-1 (PARP-1) (Poonepalli et al., 37 2005) and securin (Chao et al., 2006a) protect against arsenic-induced cell cycle disruption. Yih 84 DRAFT—DO NOT CITE OR QUOTE ------- 1 et al. (2005) provided evidence that 1 uM sodium arsenite appears to inhibit activation of the G2 2 DNA damage checkpoint and thereby allows cells with damaged DNA to proceed from G2 into 3 mitosis. 4 Extremely small concentrations of As111 can stimulate cell proliferation. For example, 5 0.005 uM sodium arsenite exposure for 24 hours stimulated cell proliferation in NHEK; 6 however, concentrations of 0.05 uM or higher inhibited it (Vega et al., 2001). In other studies, 7 stimulation occurred at much higher concentrations. Mudipalli et al. (2005) exposed NHEK 8 cells to many exposure levels of As111, MMAm, and DMA111 for 24 hours. The LOECs were 2, 9 0.5, and 0.6 uM, respectively. There was increased stimulation of cell proliferation up to doses 10 of 6, 0.8, and 0.6 uM, respectively, and in all cases significant cytotoxicity was observed at 11 higher doses. Proliferation was often stimulated to a considerable extent. Yang et al. (2007) 12 showed that human embryo lung fibroblast (HELP) cells exposed to 0.5 uM sodium arsenite for 13 24 hours had 175% of the cell proliferation efficiency of control cells. When the concentration 14 of As111 was increased to 5 uM, however, the cell proliferation efficiency decreased to 60% that 15 of the control. The increased proliferation rates can extend over long periods, as shown by 16 Bredfeldt et al. (2006), who exposed UROtsa cells to 0.05 uM MMA111 for 12, 24, or 52 weeks. 17 Cell population doubling times were 27, 25, and 21 hours, respectively, in comparison to the 42 18 hours observed in the control. 19 Mutations can play an important part in initiating carcinogenesis or in the development of 20 cancers, and they range from gene mutations that involve a single base-pair change to 21 chromosomal aberrations (CAs). Much evidence is presented in Table C-3 under Chromosomal 22 Aberrations and/or Genetic Instability to show that inorganic arsenic can induce CAs, SCEs, 23 MN, multilocus deletions, and several other endpoints such as changes in the length of 24 telomeres. Arsenic appears to be ineffective in inducing gene (point) mutations, but mutations at 25 some genes tend to be deletions that are so large that they extend over several genes (termed 26 multilocus deletions). These multilocus deletions have been grouped with CA in Table C-3. 27 CD59 mutations (Liu et al., 2005) and gpt mutations (Klein et al., 2007) provide examples of 28 such mutations. Numerous experiments are summarized in Table C-3 that show that CAs can be 29 induced by exposure to 10 uM or less of sodium arsenite for periods of 24 hours or less. 30 Following exposures of human primary peripheral blood lymphocytes for 24 hours, LOECs for 31 As111, MMA111, and DMA111 were 2.5, 0.6, and 1.35 uM, respectively (Kligerman et al., 2003). 32 Examination of data shown in the table for the few other experiments on MMA111 and DMA111 are 33 consistent with this experiment in suggesting that both of those methylated arsenicals tend to be 34 more effective in inducing CAs than As111. The table includes estimates of about 15 LOECs for 35 induction of SCEs and about 20 LOECs for induction of MN following exposure to As111, and it 36 appears that CAs, SCEs, and MN are all induced to roughly the same extent by As111. Some 37 experiments fail to show a dose-response, which makes them difficult to interpret. 85 DRAFT—DO NOT CITE OR QUOTE ------- 1 Several of the experiments on CAs provided evidence of arsenic-induced changes in 2 chromosome number (e.g., Barrett et al., 1989; Ochi et al., 2004). In the Ochi et al. (2004) 3 experiment, DMA111 was much more potent than As111, and it induced mitotic spindle, 4 centrosome, and microtubule elongation abnormalities. Experiments on induction of MN were 5 conducted in such a way as to distinguish between MN caused by aneuploidy and those caused 6 by chromosomal breakage; these experiments provided evidence that both mechanisms may be 7 important (e.g., Colognato et al., 2007; Ramirez et al., 2007). Chou et al. (2001) showed that 8 exposure to 0.25 uM ATO for 4 weeks caused a decrease in telomere length. Mouse embryo 9 fibroblasts that are homozygous for the PARP knockout mutation were shown to be much more 10 sensitive to both arsenite-induced telomere attrition and induction of MN by As111 (Poonepalli et 11 al., 2005). Many experiments investigated the effects of various modulators on induction of 12 arsenic-induced chromosomal damage. For example, Jan et al. (2006) found that co-treatment 13 with low concentrations of dimercaptosuccinic acid, meso 2,3-dimercaptosuccinic acid (DMSA), 14 or 2,3-dimercaptopropane-l-sulfonic acid (DMPS) markedly increased the induction of MN by 15 sodium arsenite, ATO, MMA111, and DMA111, while co-treatment with high concentrations of the 16 same chemicals decreased the ability of arsenic to induce MN. Although the authors stated that 17 the reasons are obscure why these dithiol compounds effectively enhanced the toxic effects of 18 arsenic when they were at micromolar concentrations, they speculated that the observed results 19 might be related to the influence of dithiols on retention of arsenite in cells, with low 20 concentrations of dithiols increasing arsenite levels and high concentrations of dithiol decreasing 21 them. Ramirez et al. (2007) also showed that co-treatment with SAM blocked As111 induction of 22 centromere positive (cen+) MN without having any effect on its induction of centromere 23 negative (cen-) MN. The authors suggested that the reason for this might be that SAM in some 24 way influences some components (probably microtubules) of the mitotic spindle. As the main 25 methyl group donor, SAM plays a major role in chromatin methylation and condensation, and it 26 might stop the lagging of chromosomes by in some way correcting the cell's methylation status. 27 Alternatively, they suggested that SAM might interfere with the effects of ROS in causing 28 aneuploidy. Whatever SAM does to block induction of cen+ MN, it does not appear to affect 29 induction of double strand DNA breaks that would lead to cen- MN. 30 The results from the Co-Carcinogenesis category all relate to promotion of 31 benzo[a]pyrene (B[a]P)-mediated carcinogenesis via exposure to 1.5 uM sodium arsenite for 12 32 weeks. Transformation (i.e., anchorage-independent growth in soft agar) of a rat lung epithelial 33 cell line occurred because of the arsenite treatment alone, and the transformed cells were shown 34 by proteomic analysis to have changes in the amounts present of many proteins. When the 35 arsenite treatment was preceded by exposure to 100 nM B[a]P for 24 hours, there was a 36 synergistic interaction. Results indicate that the transformation rate increased more than 500 and 37 200 times when compared to arsenite and B[a]P treatments alone, respectively. The findings in 86 DRAFT—DO NOT CITE OR QUOTE ------- 1 the proteomic analysis also showed synergistic interactions (Lau and Chiu, 2006). BPDE 2 (benzo[a]pyrene diol epoxide) is an active metabolite of B[a]P. Shen et al. (2006) showed that a 3 24-hour pretreatment of GM04312C cells, a SV-40 transformed XPA human fibroblast NER- 4 deficient cell line, with 10 or 50 uM As111 markedly increased the cellular uptake of BPDE in a 5 dose-dependent manner. 6 The results found under Co-Mutagenesis showed that As111 affected the induction of 7 mutations (using different assays) when there was also a treatment with UV, diepoxybutane 8 (DEB), methyl methanesulfonate (MMS), X-radiation, gamma-radiation, or N-methyl-N- 9 nitrosourea (MNU). Many of the types of mutations affected were gene mutations (i.e., point 10 mutations and numerous other changes in the DNA of single genes, such as small deficiencies), 11 which are not normally induced by arsenic alone. Arsenic treatment also caused co-mutagenesis 12 regarding CAs and MN. Sometimes the timing of the As111 treatment relative to the treatment 13 with the other agent was of importance to the result observed. For example, a 24-hour 14 pretreatment with 10 uM sodium arsenite reduced the frequency of induction of hypoxanthine- 15 guanine phosphoribosyltransferase (HGPRT) mutations by MMS, but a 24-hour post-treatment 16 with the same concentration of sodium arsenite caused a synergistic interaction with MMS in 17 induction of HGPRT gene mutations (Lee et al., 1986). 18 The data found in Table C-3 under Cytotoxicity are sometimes important to help 19 determine the possible relevance to human health of findings related to other key events. For 20 example, a large arsenic-induced increase in the expression of some protein that is important in 21 signal transduction is much more likely to have such relevance if it occurs at concentrations 22 having little or no cytotoxicity than if it occurs only when most cells are dying. Table C-3 shows 23 that large differences in LOECs for cytotoxicity can result from a change in any of the following 24 variables: species of arsenic, duration of treatment, cell line, and particular assay used. As 25 another example, LOECs of As111 were 0.1 and 50 uM after 24-hour exposures in Jurkat cells and 26 HeLa cells, respectively (Salazar et al., 1997). Petrick et al. (2000) showed that three different 27 cytotoxicity assays yielded substantially different 24-hour LCSOs for each of five different 28 arsenic species. Sometimes the different assays yield more similar results when treatments last 29 at least 48 hours (Komissarova et al., 2005). Overall it appears that in comparison to As111, 30 MMA111 has substantially higher cytotoxicity, DMA111 has higher cytotoxicity, and Asv has 31 sub stanti ally 1 ower cytotoxi city. 32 Effects of modulators on arsenic-induced cytotoxicity were tested in many experiments. 33 Snow et al. (1999) showed that pretreatment with BSO, to decrease GSH levels, markedly 34 increased cytotoxicity of sodium arsenite following a 48-hour exposure. Jan et al. (2006) found 35 that co-treatment with low concentrations of DMSA or DMPS (dithiols that are currently used to 36 treat arsenic poisoning) markedly increased the cytotoxicity of ATO, while co-treatment with 37 high concentrations of DMSA or DMPS had the opposite effect. Probably the most important 87 DRAFT—DO NOT CITE OR QUOTE ------- 1 observation related to cytotoxicity from perusal of Table C-3 is that exposure of a large number 2 of different cell lines to trivalent arsenicals results in significant cytotoxicity at molarities 3 smaller than what would be found in urine, or even in the blood streams, of individuals exposed 4 to high levels of inorganic arsenic in drinking water in places like Bangladesh. In some cell 5 lines, even the pentavalent arsenicals destroyed more than 50% of the cells following a 7-day 6 exposure with concentrations such as those observed in Bangladesh; As111 and MMA111 would do 7 the same at concentrations far below such levels (Wang et al., 2007). Also, from the numerous 8 dose-response curves published in those papers, it is apparent that cytotoxicity generally has a 9 threshold below which there is no apparent effect. 10 DNA Damage is another key event category for which many experimental data are 11 summarized in Table C-3. Evidence showed induction of oxidative DNA damage, DNA single- 12 strand breaks, and DNA-protein crosslinks by exposures at 10 uM (and often much less) of As111 13 for periods of often much less than one day. MMA111 is especially effective in inducing damage 14 detected by the comet assay (Gomez et al., 2005). Much more DNA damage was detected in the 15 comet assay by using enzyme treatments to reveal oxidative DNA adducts and DNA protein 16 crosslinks, and DNA damage was induced at levels of sodium arsenite that caused no 17 cytotoxicity in two different cell types (Wang et al., 2001). In a third cell type, no DNA damage 18 was observed up to the maximum concentration tested (2 uM), even though in each of the other 19 two cell types the LOEC was 0.25 uM. Jan et al. (2006) found that co-treatment with low 20 concentrations of DMSA or DMPS markedly increased the DNA damage detected by the comet 21 assay following treatment with ATO, while co-treatment with high concentrations of DMSA or 22 DMPS had the opposite effect. Several experiments looked at induction of 8-OHdG formation 23 as a measure of oxidative DNA damage. In one such experiment, sodium arsenite was shown to 24 be effective. However, MMA111 was shown to be about 200 times more effective than As111 (with 25 an LOEC of 0.05 uM) following a 1-hour treatment (Eblin et al., 2006). Pre-incubation with 26 SOD or catalase to reduce effects of ROS almost completely blocked induction of 8-OHdG 27 formation by a 24-hour treatment with sodium arsenite (Ding et al., 2005). Tests with a cell line 28 containing a knockout mutation of the PARP-1 gene showed that the PARP-1 protein protects 29 against arsenic-induced DNA damage detected by the comet assay at pH >13 in the version of 30 the assay that does not include further digestion to detect additional types of DNA damage 31 (Poonepalli et al., 2005). 32 The DNA Repair Inhibition or Stimulation category includes rather few experiments in 33 Table C-3. A microarray experiment that showed decreased expression of DNA repair genes 34 involved exposure to only 0.77 uM of sodium arsenite for 70 days (DuMond and Singh, 2007). 35 Arsenic does not always have the effect of decreasing repair. Snow et al. (2005) found that 36 W138 cells exposed to 0.1 uM sodium arsenite for 24 hours showed increased DNA ligase 37 activity. Increasing the As111 concentration to 1 uM further increased the activity, but 5 uM 88 DRAFT—DO NOT CITE OR QUOTE ------- 1 decreased DNA ligase activity to below normal levels. The same paper demonstrated a rather 2 similar reversal-of-direction effect for DNA polymerase p. In another experiment, when CHO 3 Kl cells were treated with MMS followed by 5 uM sodium arsenite for 6 hours, there was a 4 decrease in repair of MMS-induced single-strand breaks in DNA (Lee-Chen et al., 1993). 5 Andrew et al. (2006) demonstrated that in Jurkat cells the LOEC for sodium arsenite was 0.01 6 uM for reduction of expression of NER gene ERCC1 (excision repair cross-complement 1 7 component). The decrease in expression was 45% at that concentration and 60% at 8 concentrations of 0.1 and 1 uM. The functional effect of this decrease in expression was shown 9 by reduced repair following a challenge with the mutagen 2-AAAF immediately after the sodium 10 arsenite treatment. Clearly, exposure to inorganic arsenic at low concentrations can modify the 11 level of DNA repair. 12 The Effects Related to Oxidative Stress (ROS) category in Table C-3 includes many 13 experiments in which antioxidants or radical scavengers were used as modulators. When a 14 reduction in the effects was seen, it was taken as evidence that oxidative stress was the cause of 15 the original effects observed, as, for example, in the study by Sasaki et al. (2007). Results from a 16 series of experiments by Lynn et al. (2000) led to the conclusion that As111 activates NADH 17 oxidase to produce superoxide, which then causes oxidative damage to DNA. Experiments by 18 Liu et al. (2005) dealt with the effects of various modulators on induction of CD59- mutations 19 and lead to the conclusion that peroxynitrites, which are formed as a result of ROS and reactive 20 nitrogen species, have an important role in the induction by As111 of such mutations. Wang et al. 21 (2007) measured formation of oxidative damage to lipids, proteins, and DNA (comet assay) by 22 three trivalent arsenicals and three pentavalent arsenicals in two different cell lines. For As111, 23 Asv, MMAm, and DMA111, the LOECs were all 0.2 uM for a 24-hour exposure for all three types 24 of damage. The order of effectiveness of the different arsenicals differed in the two cell lines 25 used and for the different types of damage. Consistent with these effects, increased levels of 26 nitric oxide, superoxide ions, hydrogen peroxide, and the cellular free iron pool were 27 consistently detected in both cell lines after treatments by all three trivalent arsenicals. A 28 microarray analysis in which genes were identified for which the response to ATO and hydrogen 29 peroxide was reversed by n-acetyl-cysteine (NAC) suggested that 26% of the genes significantly 30 responsive to ATO were directly altered by ROS (Chou et al., 2005). Further evidence that ROS 31 is likely involved in arsenite-induced DNA damage comes from comet assays done on splenic 32 lymphocytes from SOD knockout mice (Kligerman and Tennant, 2007). Results showed 33 homozygotes exhibiting a large decrease in splenic SOD levels and a large increase in arsenite- 34 induced DNA damage, while heterozygotes had intermediate changes in SOD levels and DNA 35 damage. 36 Table C-3 includes little information on Enzyme Activity Inhibition. Hu et al. (1998) and 37 Snow et al. (1999) tested the effect of sodium arsenite on the activity of several purified enzymes 89 DRAFT—DO NOT CITE OR QUOTE ------- 1 in vitro, including enzymes required for DNA repair and some related to GSH metabolism. The 2 purpose of the study was to examine whether As111 binding to sulfhydryls caused protein 3 denaturation and inhibited enzyme activity. In almost all cases, the purified enzymes were not 4 inhibited by physiologically relevant concentration of As111. The concentrations that are needed 5 to cause 50% inhibition (ICSOs) for the rate of the reaction (over 6 minutes for many of those 6 enzymes) ranged from 6.3 to 381 mM. The one exception was purified pyruvate dehydrogenase 7 for which the IC50 was 5.6 uM. Table C-3 also lists ICSOs for GSH peroxidase and ligase when 8 tested in extracts of AG06 (SV40-transformed human keratinocyte) cells that were pretreated for 9 24 hours with an unspecified concentration of sodium arsenite; these ICSOs were both low, i.e., 10 2.0 and 14.5 uM, respectively. 11 Table C-3, under Gene Amplification, shows that As111 caused amplification of 12 dihydrofolate reductase (dhfr) genes in three different experiments with LOECs ranging from 13 0.0125 to 6 uM (Barrett et al., 1989; Rossman and Wolosin, 1992; Mure et al., 2003). Takahashi 14 et al. (2002) showed that several neoplastic transformed cell lines produced by 48-hour 15 treatments with either < 8 uM As111 or < 150 uM Asv contained gene amplification of either the 16 c-Ha-ras or the c-myc oncogene. Almost all of the data in Table C-3 for Gene Mutations show 17 no induction of mutations by arsenic. 18 Hypermethylation of DNA was demonstrated in a number of specific DNA sequences in 19 two human kidney carcinoma cell lines and in one human lung carcinoma cell line. In the lung 20 cell line, the LOEC for As111 was 0.08 uM for a 7-day exposure, and there was a positive dose- 21 response extending over the two higher doses tested (0.4 and 2.0 uM). Hypermethylation in this 22 cell line was demonstrated within a 341-base-pair fragment of the promoter region of p53 (Mass 23 and Wang, 1997; Zhong and Mass, 2001). 24 Hypomethylation of DNA has been demonstrated globally and for a number of specific 25 DNA sequences. In one instance, exposure of HaCaT cells to 0.2 uM sodium arsenite for 10 26 serial passages in folic-acid depleted media caused genomic hypomethylation. Sodium arsenite 27 repressed the expression of the DNA methyltransferase (DNMT) genes DNMT1 and DNMT3A 28 and caused depletion of SAM, the main cellular methyl donor. It is thought that long-term 29 exposure to sodium arsenite may have resulted in DNA hypomethylation as a consequence of 30 those two complementary mechanisms (Reichard et al., 2007). Singh and DuMond (2007) 31 demonstrated methylation changes in DNA at 18 genetic loci in TM3 cells, with some showing 32 hypomethylation and others hypermethylation, following sodium arsenite exposures ranging 33 from 0.008-7.7 uM that lasted for either 25 or 75 days. The LOEC was the lowest dose. Some 34 loci were affected only after 25 days of exposure, while others were affected after 75 days of 35 exposure. In one of several other demonstrations of hypomethylation, a 19-week exposure of 36 TRL 1215 cells to 0.125 uM sodium arsenite was sufficient to cause global hypomethylation 37 (Zhao etal., 1997). 90 DRAFT—DO NOT CITE OR QUOTE ------- 1 Under Immune System Response, Table C-3 describes a wide-range of effects on the 2 immune system. This discussion provides highlights from that table and Appendix D, which is 3 devoted entirely to the immunotoxicity of inorganic arsenic. Appendix D discusses some aspects 4 of the immunotoxicity of inorganic arsenic in much more detail, including more emphasis on 5 human studies and in vivo experiments on laboratory animals, as well as on some older in vitro 6 studies. It overlaps very little with data found in Table C-3. Effects thought to be related to 7 Immune System Response were grouped under that heading in Table C-3 even if they dealt 8 mainly with other key events. For example, several findings related to Apoptosis, Cytotoxicity, 9 or Signal Transduction are included in this section of Table C-3. 10 Exposures to low concentrations of As111 over 1-2 weeks inhibited maturation of human 11 peripheral blood monocytes (HPBMs) into the following types of cells: M-type and GM-type 12 macrophages, immature dendritic cells, and multinucleated giant cells (Sakurai et al., 2006). The 13 ICSOs for this inhibition ranged from 0.06 to 0.70 uM. Lemarie et al. (2006a) showed that ATO 14 inhibited macrophage differentiation of peripheral blood mononuclear cells (PBMCs) and that 15 concentrations as low as 0.125 uM over 6 days induced apoptosis and necrosis in PBMCs co- 16 treated with granulocyte-macrophage colony-stimulating factor (GM-CSF) or macrophage 17 colony-stimulating factor (M-CSF). Differentiated macrophages developed from PBMCs treated 18 with GM-CSF for 6 days were exposed to 0.25 uM ATO for 6 days. The ATO treatment caused 19 major alterations in morphology, adhesion, and actin organization, giving the impression that the 20 ATO "de-differentiated" the macrophages back into monocytic cells (Lemarie et al., 2006b). 21 The same series of experiments showed that macrophages exposed to 1 uM ATO for 6 days also 22 caused a reduction in several surface markers, markedly decreased endocytosis and 23 phagocytosis, and increased the secretion of inflammatory cytokines in response to a co- 24 treatment with lipopolysaccharide. 25 Exposure of PBMCs that had been stimulated with phytohemagglutinin (PHA) after 26 exposure to 1-5 uM sodium arsenite for 120 hours caused a marked dose-related decrease in 27 both cell proliferation and the percentage of divided cells (Tenorio and Saavedra, 2005). Even at 28 the higher doses, most of the cells were viable but unable to divide. The treatments also 29 modified the expression of CD4 and CDS molecules. Judging from evaluation of blast 30 transformation, CD4+ and CD8+ T cells appear to have different sensitivities to As111. As the 31 concentration of the sodium arsenite increased from 1 to 5 uM in the 120-hour treatment, there 32 was an accumulation of resting CD8+ cells with a positive dose-response, but there was not an 33 accumulation of CD4+ cells. The Janus kinase (JAK)-signal transducer and activator of 34 transcription (STAT) pathway is an essential cascade for mediating normal functions of different 35 cytokines in the development of the hematopoietic and immune systems. Huang et al. (2007a) 36 showed that exposure of SV-HUC-1 cells to sodium arsenite for 48 hours caused changes in 37 levels of proteins that are part of that cascade, and the LOEC was 2 uM. Sometimes there was a 91 DRAFT—DO NOT CITE OR QUOTE ------- 1 dose-response, and sometimes the direction of the change reversed. Cheng et al. (2004) showed 2 that a 48-hour pretreatment of HepG2 cells with 4 uM sodium arsenite was sufficient to block 3 induction of STATS activity by an IL-6 treatment. Other experiments showed that As111 acted 4 directly on the JAK1 protein to cause JAK-STAT inactivation. Di Gioacchino et al. (2007) 5 studied the effects of several arsenicals on PBMC proliferation and cytokine release. At a 6 concentration of 100 uM, sodium arsenite was effective in decreasing PHA-induced cell 7 proliferation and in reducing interferon-gamma (IFN-y) and TNF-a release. However, at a 8 concentration of 0.1 uM, As111 significantly increased cell proliferation. More details about that 9 experiment are found in Appendix D. 10 Regarding Inhibition of Differentiation, in experiments done on spontaneously 11 immortalized human keratinocytes and on normal human epidermal cells derived from foreskin, 12 sodium arsenite was shown to delay differentiation and preserve the proliferative potential of 13 keratinocytes (Patterson et al., 2005; Patterson and Rice, 2007). A concentration of sodium 14 arsenite as low as 0.1 uM over 4 days had a noticeable effect, but most experiments were done 15 using 2 uM sodium arsenite over 4-14 days, which yielded a much larger effect. Treatment of 16 C3H 10T1/2 cells with 6 uM sodium arsenite for 8 weeks completely inhibited their 17 differentiation into adipocytes following dexamethasone/insulin treatment, and treatment with 18 3 uM sodium arsenite for only 48 hours was the LOEC for that effect (Trouba et al., 2000). 19 Interference With Hormone Function was demonstrated in experiments by Bodwell et al. 20 (2004, 2006). Some effects were observed at approximately 0.09 uM of sodium arsenite; 21 however, the increases found in glucocorticoid-receptor-mediated gene transcription of reporter 22 genes that contained tyrosine aminotransferase (TAT) response elements were highly dependent 23 on, and inversely related to, the amount of activated steroid receptor within cells. More detailed 24 information on interference with hormone function can be found in Table C-3. 25 Under Malignant Transformation or Morphological Transformation, Table C-3 shows 26 that concentrations of less than 1 uM of As111, MMA111, or DMA111 are capable of causing 27 transformation. HaCaT cells exposed to 0.5 uM As111 for 20 passages caused the cells to become 28 tumorigenic, as shown by production of tumors 2 months after injection into Balb/c nude mice 29 (Chien et al., 2004). Zhao et al. (1997) found similar results with another cell line after 18 weeks 30 of exposure to 0.25 uM As111. UROtsa cells exposed to 0.05 uM MMA111 for 52 weeks caused 31 anchorage-independent growth as detected by colony formation in soft agar, and cells from those 32 colonies showed enhanced tumorigenicity in SCID mouse xenographs (Bredfeldt et al., 2006). 33 After 26 weeks, this experiment showed much anchorage-independent growth but not yet 34 enhanced tumorigenicity. Syrian hamster ovary (SHE) cells exposed to DMA111 for 48 hours 35 showed morphological transformation at a concentration of only 0.1 uM, and at the highest dose 36 tested of 1.0 uM, 3.35% of the surviving colonies had become transformed (Ochi et al., 2004). 92 DRAFT—DO NOT CITE OR QUOTE ------- 1 In contrast, at a dose of 10 uM after the same exposure duration of 48 hours, As111 had only 2 transformed 0.48% of the surviving cells. 3 Table C-3 summarizes many findings related to the Signal Transduction category, even 4 though considerable data found under Aberrant Gene or Protein Expression could have been 5 placed into this category. Most of the data in this category are for sodium arsenite or ATO. In 6 addition, there are numerous LOECs smaller than 10 uM (often much less), and they are often 7 for treatments that lasted much less than one day. Drobna et al. (2002) evaluated 8 phosphorylation of extracellular signal-regulated kinase (ERK)-2, activator protein (AP)-l 9 binding activity, and phosphorylation of c-Jun (an AP-1 protein) by six arsenicals in treatments 10 lasting up to 2 hours. Asv, MMAV, and DMAV were all tested at concentrations up to 100 uM 11 and had no effect. As111, MMA111, and DMA111 each had an LOEC of 0.1 for at least one endpoint. 12 Details presented in Table C-3 show that the responses of those three arsenicals were different 13 and that, in some cases, the direction of the response reversed as the concentration increased. In 14 some cases a reduction from an increase was observed, which is interesting because various 15 responses for some endpoints described above showed a reversal in which the lowest doses 16 caused a bigger effect. Another experiment showing a reversal in response (from a decrease to 17 an increase) was for phosphorylation of Akt Thr308 in JB6 C141 cells (P+ mouse epidermal cell 18 line) (Ouyang et al., 2006). Following 1-hour exposures to sodium arsenite, there was slight 19 decrease at 0.1 uM, a larger decrease at 0.5 uM, increases above the control level at 1 and 5 uM, 20 and a much larger increase at 10 uM. Additionally, several experiments in this category related 21 to different ways in which arsenic affects signal transduction to either increase or decrease 22 apoptosis. For example, MCF-7 cells exposed to 2 uM ATO for 1 hour activated the pro- 23 survival MEK/ERK pathway (Ye et al., 2005). By decreasing apoptosis, such an effect might 24 permit the survival of cells containing damage that could eventually lead to a cancer. Yancy et 25 al. (2005) did a series of experiments on H9c2 cells (an immortalized myoblast cell line derived 26 from fetal rat hearts) and concluded that sodium arsenite exposure decreases cell migration 27 through an effect on focal adhesions and by disrupting cell interactions with the extra-cellular 28 matrix. Focal adhesions are involved in integrin signaling. Florea et al. (2007) showed that 29 ATO triggered three different kinds of Ca2+ signals (i.e., steady state increases, transient 30 elevations, and calcium spikes). The Ca2+ concentration in cells was substantially increased (and 31 by rather similar amounts) by exposure to either 0.1 or 1 uM ATO for about 1 hour in two 32 different cell lines (i.e., the human neuroblastoma cell line SY-5Y and the human embryonic 33 kidney cell line HEK 293). 4.5. SYNTHESIS AND EVALUATION OF MAJOR NONCANCER EFFECTS 34 Not addressed in this document. 93 DRAFT—DO NOT CITE OR QUOTE ------- 4.6. WEIGHT-OF-EVIDENCE EVALUATION AND CANCER CHARACTERIZATION 4.6.1. Summary of Overall Weight-of-Evidence 1 Based upon the EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) 2 inorganic arsenic is categorized as "carcinogenic to humans" due to convincing epidemiological 3 evidence of a causal relationship between oral exposure of humans to inorganic arsenic and 4 cancer. Arsenic is a multisite carcinogen, with numerous studies finding an association between 5 arsenic and increased incidences of a number of different types of cancers. The carcinogenic 6 effect of arsenic has been reported for populations in many different countries. While the studies 7 detailed in this document provide evidence for cancer after oral exposure to arsenic, arsenic also 8 has been associated with cancer after inhalation exposure (U.S. EPA, 1994). 4.6.2. Synthesis of Human, Animal, and Other Supporting Evidence 9 Numerous epidemiologic investigations, each conducted differently and containing its 10 own biases (e.g., lack of confounding variables, possible recall bias), provide support for an 11 association between oral exposure to inorganic arsenic and cancer including skin, bladder, 12 kidney, lung, liver, and prostate. The most extensively studied population is from southwest 13 Taiwan. This is because between 1910 and 1920, water supplies were changed from shallow 14 surface water wells to artesian wells, which were subsequently found to contain high levels of 15 arsenic in various regions. Studies in these arsenic-endemic regions of Taiwan have found 16 increases in all of the aforementioned cancer types. The link between these cancers and arsenic 17 exposure in drinking water also have been observed in other parts of the world, including Japan, 18 Chile, and Argentina. Therefore, it is unlikely that any single environmental factor (e.g., 19 nutritional habits) associated with a single population is entirely responsible for the increased 20 cancer rates. Although many studies did not account for confounding variables (e.g., cigarette 21 smoking in association with lung cancer), the positive associations between arsenic intake and 22 cancer risk were still observed in studies that did account for confounding variables (e.g., 23 lifestyle habits, age, and socioeconomic status). 24 Most of the epidemiology studies examining the relationship between arsenic exposure 25 from drinking water and cancers are ecological in nature and are therefore subject to the 26 limitations inherent in such studies (e.g., lack of measured individual exposure). For a number 27 of reasons, the southwest Taiwanese database remains the most appropriate source for estimating 28 bladder and lung cancer risk among humans (NRC, 1999, 2001; SAB, 2000, 2007), despite 29 lacking individual water consumption and nonwater arsenic intake. Strengths of the data include 30 the size of the population, the reliability of the population and mortality counts, the stability of 31 residential patterns, the homogenous lifestyle as confirmed by surveys, the long-term exposures, 32 the extensive follow-up (almost 900,000 person-years), the large number of exposed villages 33 (42), and the large number of cancer deaths (1152 recorded from 1973 to 1986). Population 94 DRAFT—DO NOT CITE OR QUOTE ------- 1 records in Taiwan have been well kept since 1905, and death certificates include all primary 2 cancers. In addition, cancer cases were pathologically confirmed in some of the Taiwanese 3 studies. 4 Although dose-response relationships have been observed for the majority of cancers 5 noted in areas with high levels of arsenic in their drinking water, results for low-level arsenic 6 epidemiologic investigations (primarily from the United States and Europe) have been equivocal 7 with regard to the relationship between these cancers and arsenic exposure. This could be due to 8 the fact that none of the studies accounted for arsenic exposure through food sources. Kile et al. 9 (2007) found that as the level of arsenic in the water decreased for women in Bangladesh, the 10 contribution of arsenic from dietary sources became of greater importance. Uchino et al. (2006) 11 found that with concentrations of 50 ppb or less of arsenic in the drinking water in a population 12 in West Bengal, India, the contribution of arsenic from food was the main source of arsenic 13 exposure (i.e., contribution from water with less than 50 ppb was less than 27% of the total 14 arsenic consumed). Therefore, as the exposure of arsenic from drinking water decreases and the 15 relative contribution from food increases, misclassification of exposure groups can become 16 significant. The average estimate of inorganic arsenic consumption in food ranges from 1.34 17 ug/day in infants to 18 ug/day in adults, for a total arsenic average of 62 ug/day for people in the 18 United States (NRC, 1999). At the lower concentrations, dietary intake could easily create total 19 arsenic intake levels to be similar between the referent group and what is considered the 20 exposure group. 21 Cantor and Lubin (2007) also conclude that misclassification occurs because exposure is 22 not necessarily assessed during disease-relevant exposure periods. In regards to cancer, there is 23 a long latency period, which appears to vary depending on the type of cancer and exposure. This 24 means that exposure to arsenic sources during the decades prior to cancer outcome is necessary. 25 Therefore, studies with low levels of exposure that are ecological in nature (no individual 26 exposure) are more prone to misclassification, which means they are biased toward the null 27 hypothesis. In addition, studies that attempted to individualize exposure by examining toenail 28 arsenic levels are looking at only the prior year of exposure (Cantor and Lubin, 2007) and may 29 miss the important exposure period. Despite all these numerous limitations in low-level 30 exposure studies, significant associations have been observed for cancers of the prostate 31 (Hinwood et al., 1999; Lewis et al., 1999), skin (Hinwood et al., 1999; Karagas et al., 2001; 32 Beane-Freeman et al., 2004; Knobeloch et al., 2006), and bladder (Kurttio et al., 1999; 33 Steinmaus et al., 2003; Karagas et al., 2004). In most cases, however, there is no dose-response 34 with increases observed at the highest concentrations only and in many cases significant results 35 occurred in smokers only. 36 There are very few animal data demonstrating the carcinogenic potential of arsenic. This 37 is likely due to the fact that rodents, which are the most likely animal model, are better 95 DRAFT—DO NOT CITE OR QUOTE ------- 1 methylators of arsenic than humans (Vahter, 1999a). Since it has been noted that humans who 2 are better methylators are at lower risk (Yu et al., 2000; Chen et al., 2005a; Steinmaus et al., 3 2005; Valenzuela et al., 2005; Ahsan et al., 2007; Huang et al., 2007b; McCarthy et al., 2007a), 4 it is not surprising that animals that are better methylators are at even lower risk. As stated 5 before, arsenic has been associated with cancers of the skin, lung, kidney, bladder, and liver. 6 Below is a summary these different types of cancers and their association with arsenic exposure 7 in drinking water. 4.6.2.1. Skin Cancer 8 Epidemiologic investigations of populations in the arseniasis-endemic areas of Taiwan 9 have shown that exposure to arsenic from drinking water is associated with skin cancer (Tseng et 10 al., 1968; Tseng, 1977; Chen et al., 1985, 1988a,b; Wu et al., 1989; Chen and Wang, 1990; Tsai 11 et al., 1999). The prevalence rate for skin cancer showed an increasing gradient according to the 12 arsenic content of the well water. Guo et al. (2001) found significant increases in SCCs at the 13 highest dose only (>640 ppb) with results at lower doses variable, suggesting that skin cancers 14 may be cell-type specific. Contrastingly, Karagas et al. (2001) found increases in both SCC and 15 BCC in the highest toenail arsenic concentration in a population in the United States. Beane- 16 Freeman et al. (2004) also found an increase in the risk of melanoma with elevated toenail 17 arsenic concentrations. Therefore, these results demonstrate that skin cancers may not be cell- 18 type-specific. Although Taiwan has been the area most associated with skin cancers in relation 19 to arsenic exposure, the association has been made in other populations as well. Arsenic has also 20 been associated with skin cancers in Argentina, where signs of arsenicism also have been 21 observed (Smith et al., 1998). Hopenhayn-Rich et al. (1998), however, found a significant 22 association in women in the highest category and surprisingly in males in the lowest category 23 only. Skin cancer has also been found in China with drinking water concentrations of 150 ppb or 24 greater (Lamm et al., 2007). Skin cancer was not found associated with arsenic in Denmark 25 (Baastrup et al., 2008) or in the United States (Meliker et al., 2007), but these studies were at 26 lower concentrations of arsenic. 27 Skin tumors have only been induced in transgenic mice or with subsequent TPA or UV 28 exposure (indicating co-carcinogenesis) in mice. Because co-carcinogenesis has been 29 demonstrated in animal models, it is possible that the same occurs in humans. Sun exposure 30 would likely be high and the use of sunblock is less likely in the areas where skin cancer has 31 been noted (i.e., Taiwan and Argentina). Therefore, a possible co-carcinogenic effect also may 32 be contributing to the association. 4.6.2.2. Lung Cancer 33 Lung cancer has been associated with arsenic in populations that were exposed to 34 exceedingly high arsenic levels in Taiwan, Chile, and Argentina. Studies of populations with 96 DRAFT—DO NOT CITE OR QUOTE ------- 1 lower arsenic exposure, especially <50 ppb, have not conclusively found an association between 2 arsenic and lung cancer. Lung cancer was not associated with arsenic exposure in the United 3 States (Lewis et al., 1999 and Meliker et al., 2007), Denmark (Baastrup et al., 2008), or Australia 4 (Hinwood et al., 1999). Yang et al. (2004) found that lung cancer incidence in endemic areas of 5 Taiwan remained elevated even after the use of the arsenic-containing well water ceased. Yuan 6 et al. (2007) also found that mortality from lung cancers exceeded that observed in regions with 7 consistently low arsenic exposure even after a 10- to 20-year lag period after removal of the 8 arsenic source. These were likely due to the long latency for cancer. Many of the studies have 9 not controlled for smoking history, which is a potential confounder for lung cancer. 4.6.2.3. Kidney, Bladder, and Liver Cancer 10 Significant increases in mortality rates for cancers of the kidney, bladder, and liver have 11 been identified in populations from Taiwan, Argentina, and Chile. These three regions all have 12 elevated levels of arsenic exposure through drinking water. Yang et al. (2004) found that arsenic 13 was associated with kidney cancers in Taiwan. Unlike lung cancer, the mortality associated with 14 kidney cancer decreased after reducing arsenic exposure. Yang et al. (2005) also found a 15 reduction in bladder cancer after removal of arsenic exposure (through tap water instillation), but 16 the decline was gradual. In Chile, supplementation of drinking water with water from rivers 17 caused exposure to high levels of arsenic, but after the installation of improved water treatment 18 in the early 1970s, arsenic exposure dropped dramatically. Yuan et al. (2007), however, found 19 that even after a 10- to 20-year lag period after removal of the arsenic source, mortality from 20 bladder cancers still exceeded that observed in regions with consistently low arsenic exposure. 21 While high levels of arsenic have been found to be related to bladder, kidney, and liver 22 cancers, low-dose exposures from the United States, Europe, and Australia have been less clear. 23 Lewis et al. (1999) observed increased SMRs in kidney cancer for both males (SMR=1.75) and 24 females (SMR=1.60), but the results were not significant. Because the highest concentration in 25 this population was 166 ppb, the results are still noteworthy. Kurttio et al. (1999) found that 26 despite the low levels of arsenic (median = 0.1 ppb; max=64 ppb) there was evidence of a 27 relationship between exposure to arsenic at levels above 0.5 ppb and bladder cancer risk. No 28 association was observed for kidney cancer risk. Hinwood et al. (1999), Meliker et al. (2007), 29 and Baastrup et al. (2008) did not find associations between these cancers and the low levels of 30 exposure in Australia, the United States, and Denmark. 31 Although inorganic arsenic exposure in rodents has not been observed to cause increases 32 in cancer, long-term (104 weeks) exposure to DMAV in rats has been found to increase bladder 33 tumors with doses of 50 ppm or greater. These concentrations are quite high in comparison to 34 the amount of inorganic arsenic exposure in humans. 97 DRAFT—DO NOT CITE OR QUOTE ------- 4.6.2.4. In Utero Exposure 1 There is no adult animal model available to study the relationship between arsenic 2 exposure via drinking water and cancer outcome; however, lung and liver tumors have been 3 induced by inorganic arsenic in mice when exposed during gestation. Pregnant dams were 4 exposed for 10 days during gestation only; this increases the evidence that lung and liver cancers 5 are associated with oral exposure to inorganic arsenic. Reproductive and adrenal tumors also 6 have been observed with transplacental exposure in mice. 7 There is very little epidemiology information specifically linking in utero arsenic 8 exposure to cancer outcome. Although the available epidemiological studies conducted in 9 Taiwan and other countries included women of reproductive age, the cancer outcomes from adult 10 exposures were not differentiated from in utero exposures. Recently, Smith et al. (2006) 11 examined lung cancer rates (and other respiratory diseases) in cohorts born just before the peak 12 exposure period in Antofagasta, Chile (meaning that they were not exposed in utero to high 13 levels of arsenic, but were exposed during childhood) and cohorts born during the high-exposure 14 period (indicating likely in utero exposure). Results demonstrated that exposure during either 15 period of development caused increased risk of lung cancer; however, the results from early 16 childhood exposures and/or in utero exposures were not compared to exposures during adulthood 17 to determine the possible cancer sensitivity effects in humans. 18 Because both in utero studies in mice and a study in humans by Smith et al. (2006) 19 indicate that lung cancer development may be associated with transplacental arsenic exposure, 20 there is an opportunity to examine the similarities in mechanistic effects mediating lung cancers 21 between the two species. Several PBPK models exist for humans (Yu, 1999a,b; El-Masri and 22 Kenyon, 2008) and mice (Gentry et al., 2004). However, these studies are inadequate in 23 interpreting the findings from the in utero studies in mice and relating them to human exposure 24 concentrations. 4.6.3. Mode of Action Information 4.6.3.1. General Comments on MOAs 25 The carcinogenic MOA for inorganic arsenic is unknown. Multiple MOAs for inorganic 26 As seem likely in view of the numerous ways in which arsenic acts upon living organisms and 27 the several metabolites produced before it is excreted from the body. While this review focuses 28 on inorganic As, the methylated species produced during its metabolism, especially the highly 29 reactive MMA111 and DMA111, probably play an important role in the carcinogenesis of inorganic 30 arsenic consumed in drinking water. Each successive product in the metabolic pathway has its 31 own toxicity and carcinogenic potential, with possible differential transport into and out of 32 different organs. In comparison to laboratory animals, humans excrete more MMA in urine and 33 are more prone to arsenic-induced carcinogenesis. These findings suggest that MMA (probably 98 DRAFT—DO NOT CITE OR QUOTE ------- 1 in the trivalent form) may be of special importance to arsenic-induced carcinogenesis in humans. 2 The finding of numerous different tumor types associated with arsenic exposure both in humans 3 and transplacental animal models also supports the view that multiple MOAs are likely. Due to 4 the complexities of the available data related to MO A, including the range of possible toxicities 5 of the different arsenic species, the different levels of each arsenic compound in target tissues, 6 multiple hypothesized key events, and multiple tissue tumor effects in humans, there is a need 7 for improved PBPK models to assist in understanding the MO A. Although there are several 8 PBPK models available (see Section 3.5), none have sufficiently addressed the complex nature 9 of the kinetics associated with arsenic carcinogenesis; therefore, this is an ongoing effort along 10 with BBDR modeling. 11 It seems useful to describe a few MOAs for cancer to use as a frame of reference when 12 considering arsenic specifically. Although inorganic arsenic and its metabolites have not been 13 found to induce gene (point) mutations, the key events involved in mutagenesis—i.e., (1) 14 exposure of target or stem cells; (2) reaction with DNA to produce DNA damage; (3) 15 misreplication of a damaged DNA template or misrepair of DNA damage leading to a mutation 16 in a critical gene in the replicating target cell; (4) replication forming a clone of mutated cells; 17 (5) DNA replication, possibly leading to additional mutations in critical genes; (6) unbalanced 18 and uncontrolled clonal growth of mutant cells, possibly leading to pre-neoplastic lesions; (7) 19 progression of pre-neoplastic cells in those lesions, resulting in emergence of overt neoplasms, 20 solid tumors (which require neoangiogenesis), or leukemia; (8) additional mutations in critical 21 genes occurring as a result of uncontrolled cell division; and (9) cancer occurring due to 22 malignant behavior (adapted from Preston and Williams, 2005)—may contribute to one or more 23 arsenic-mediated MOA(s) for carcinogenesis. A mutagen with the above MOA would likely be 24 thought to have a linear dose-response. It is unclear what the shape of the dose-response curve is 25 for any specific key event that might be involved in the MOA for arsenic and its metabolites. 26 Therefore, a linear dose-response is the prudent choice unless the dose-response of the identified 27 key events mediating the carcinogenesis is fully understood. 28 A second example of a MOA is the one hypothesized for arsenical-induced urinary 29 bladder carcinogenesis as follows: after the requisite arsenical ingestion, absorption, and 30 metabolism, (1) DMA111 is excreted into urine above a critical concentration, (2) it reacts with 31 urothelial critical sulfhydryl groups, (3) urothelial cytotoxicity and necrosis results, (4) urothelial 32 regenerative cell proliferation (hyperplasia) results, and (5) urothelial cancer develops; oxidative 33 damage might possibly stimulate both steps 3 and 4 (adapted from Cohen et al., 2007). 34 Obviously this MOA directly relates to the topic of this review, and any combination of factors 35 in which consumption of inorganic arsenic would lead to more than the critical (threshold) 36 concentration of DMA111 for a particular individual for a sufficient time could result in bladder 37 cancer. 99 DRAFT—DO NOT CITE OR QUOTE ------- 1 Section 4.4.1 provided abundant evidence that many potential key events can occur at 2 levels of exposure that would be encountered in populations exposed to high levels of inorganic 3 arsenic in drinking water. It seems possible that those key events could fit together in many 4 ways to result in a MOA for carcinogenesis. For example, some known mutagen and/or 5 carcinogen commonly encountered in the environment might cause the initiation step, and then 6 various arsenic-induced key events would provide the later steps necessary to result in a cancer. 7 Alternatively, oxidative damage to DNA (or other types of DNA damage caused by arsenic) 8 would make the DNA more prone to be acted upon by some other agent to produce a mutation 9 that fulfills the initiation step. Although arsenic exposure does not induce gene mutations, 10 evidence from all three tables in Appendix C shows that chromosomal aberrations can be 11 induced, and if a chromosome happened to break, for example in a tumor suppressor gene, that 12 mutation might provide an important step in a MOA. After the steps in a MOA resulted in cell 13 proliferation and genomic instability, cancer would result when changes occurred that provided 14 evasion of apoptosis, self-sufficiency of growth signals and insensitivity to anti-growth signals, 15 and limitless replicative potential (Hanahan and Weinberg, 2000). Vascularization would also 16 be needed to help the tumors grow larger. 17 Many detailed reviews in the past decade have discussed possible MO As for arsenic 18 carcinogenesis. Numerous ideas expressed in these reviews agree that exposure to inorganic 19 arsenic may be able to cause cancer by many alternative MOAs. For example, Kitchin (2001) 20 discussed nine possible MOAs for arsenic carcinogenesis, suggesting that the three with the most 21 positive evidence in both animals and human cells are chromosomal abnormalities, oxidative 22 stress, and a continuum of altered growth factors leading to increased cell proliferation and then 23 the promotion of carcinogenesis. Florea et al. (2005) suggested that genomic damage, apoptosis, 24 and changes in gene expression associated with arsenic exposure are related to arsenic-induced 25 intracellular calcium disruption. Rossman (2003), Huang et al. (2004), and Simeonova and 26 Luster (2000) also provided noteworthy reviews related to MOAs of arsenic carcinogenesis. 27 Snow et al. (2005) reviewed effects of arsenic at low concentrations and suggested that hormesis 28 (i.e., a biphasic response) occurs in regard to cell proliferation and/or viability, base excision 29 DNA repair, and telomerase activity. While some low-dose effects (e.g., increased DNA repair) 30 may be protective of carcinogenesis, other effects (e.g., cell proliferation or telomerase 31 activation) may be protective and thus permit mutant cells to survive by preventing cellular 32 senescence and death and may thereby be involved in arsenic's cancer-promoting capacity. 33 Kitchin and Ahmad (2003) provided an in-depth review on oxidative stress. They did not 34 reach a definitive conclusion on the role of oxidative stress in arsenic carcinogenesis, but rather 35 stated, 36 100 DRAFT—DO NOT CITE OR QUOTE ------- 1 • " .. .it may eventually be found that many arsenic species act through several modes of 2 carcinogenic action at many stages of multistage carcinogenesis and that the concept of a 3 single cause of arsenic carcinogenesis simply does not fit the existing facts." 4 5 Oxidative stress seems particularly attractive as an important early step for some of the 6 following reasons. Some ROS can interconvert between themselves or react with nitric oxide 7 (NO) to become reactive nitrogen species (RNS). RNS have their own spectra of biological 8 reactivity. High-energy ROS can convert to lower-energy forms and in the process can damage 9 biological molecules. ROS and related species can be inactivated by cellular defenses. 10 Extended, high-level exposure to reactive arsenic species might result in the depletion of 11 generalized cellular defense mechanisms against oxidative damage. ROS have been postulated 12 to be involved in both the initiation and promotional stages of carcinogenesis (Zhong et al., 13 1997; Bolton et al., 1998, 2000; Shackelford et al., 2000; Chen et al., 2000b). Low levels of 14 ROS can modulate gene expression by acting as a secondary messenger, while high doses of 15 ROS can cause oxidative injury leading to cell death (Perkins et al., 2000). It has also been 16 demonstrated or suggested that ROS can (or does) damage cells by the following mechanisms: 17 lipid peroxidation; DNA and protein-modification; structural alterations in DNA including base- 18 pair mutations, rearrangements, deletions, insertions, and sequence amplifications (but not point 19 mutations); involvement in the signaling of the cell transformation response; affecting 20 cytoplasmic and nuclear signal transduction pathways that regulate gene expression; and 21 increasing the expression of certain genes (e.g., MDM2 protein, a key regulator of the tumor 22 suppression gene p53) (Li et al., 1998; Sen and Parker, 1996; Lander, 1997). Activation of 23 signal transduction pathways that enhance cell proliferation, reduce antiproliferative signaling, 24 and override checkpoints controlling cell division after genotoxic insult also have been 25 considered as possible mechanisms of arsenic's co-carcinogenic properties (Rossman, 2003). 26 Luster and Simeonova (2004) cited the results of in vitro studies suggesting that arsenic 27 stimulates cell proliferation through specific signal transduction pathways that are similar to 28 other classic tumor promotors. There has been much research in the last few years on the 29 effectiveness of As111, especially ATO, on apoptosis, with much of it aimed at improving cancer 30 therapy. Those results reveal the extreme complexity of the signal transduction cascades 31 involved in controlling apoptosis. Regarding causation of cancer, any effects that inorganic 32 arsenic ingestion might have on signal transduction pathways that inhibit apoptosis could result 33 in proliferation of damaged cells and thereby lead to cancer. 34 The few animal studies (Waalkes et al., 2006a, 2006b, 2004a, 2004b, 2003a, 2003b) that 35 suggest inorganic arsenic is a complete carcinogen are those of Waalkes and his group that 36 involved treatments in utero. Doses received by the pregnant dams were large compared to 37 human exposures, but tissue levels in the fetuses were reported as being comparable to levels 38 sometimes seen in humans. Almost all of the categories of key events discussed in this 101 DRAFT—DO NOT CITE OR QUOTE ------- 1 document can be caused by inorganic arsenic at exposure levels comparable to, or lower than, 2 those that would be present in large population groups presently. The experiments also indicate 3 that typically when a treatment is extended over a longer period of time, the concentration of 4 inorganic arsenic necessary to cause an effect decreases. This indicates that the impact in 5 humans suggested by the in vitro findings might be substantially greater than might be expected 6 by just comparing the concentrations found in humans and in those used in experiments. Due to 7 the complexities of the possible MO As of inorganic-arsenic-mediated carcinogenesis, various 8 scientific tools (e.g., genomic tools, human pharmacokinetic and biologically based dose- 9 response models) may be needed in order to interpret the data for the hypothesized key events 10 qualitatively and quantitatively in a meaningful way. 4.6.3.2. Low-Dose Extrapolation 11 According to the 2005 Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), a 12 linear extrapolation to low doses is to be used either when there are MOA data to indicate that 13 the dose-response curve is expected to have a linear component below the point of departure 14 (e.g., DNA-reactivity or direct mutagenic activity) or when the available data are insufficient to 15 establish the MOA for a tumor site. Since the MOA of inorganic arsenic is unknown, a linear 16 low-dose extrapolation was applied as a default option. 4.7. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 17 Several studies (Yu et al., 2000; Chen et al., 2005a; Steinmaus et al., 2005; Valenzuela et 18 al., 2005; Ahsan et al., 2007; Huang et al., 2007b; McCarthy et al., 2007a) have observed a 19 correlation between increased disease risk and low urinary DMA and/or high urinary MMA, 20 indicating a slower secondary methylation. Valenzuela et al. (2005) measured the levels of 21 MMA111 in the urine of the residents of the Zimapan region of central Mexico. They found that 22 individuals exposed chronically to arsenic who also had arsenic-related skin lesions had 23 significantly greater concentrations and proportions of MMA111 in their urine than exposed 24 individuals without skin lesions. These findings support the hypothesis that any factor (e.g., 25 genetic variability in metabolic enzymes) associated with reduced secondary methylation (i.e., 26 the conversion of MMA to DMA) may also be correlated with increase susceptibility to arsenic- 27 induced disease. In the following sections, factors affecting DMA and/or MMA ratios and level 28 in the urine or secondary methylation will be evaluated with regard to how they may affect 29 individual susceptibility. 4.7.1. Possible Childhood Susceptibility 30 Although children are exposed to arsenic through generally the same sources as adults 31 (i.e., air, water, food, and soil), their behaviors and physiology may result in them receiving 32 higher absorbed doses in relation to their body weight than adults for a given set of exposure 102 DRAFT—DO NOT CITE OR QUOTE ------- 1 conditions. Because children tend to eat less varied foods than adults, exposure to contaminated 2 food, juice, or infant formula prepared with contaminated water may result in higher doses than 3 adults. In addition, children are more likely to ingest arsenic-contaminated soil, either 4 intentionally or by putting dirty hands in their mouths. 5 There are few data on the relative efficiency of absorption of arsenic from the 6 gastrointestinal tract of children compared to adults, but measurement of urinary arsenic levels in 7 children indicate that absorption does occur. ATSDR (2007) suggests that there is some 8 evidence that children may be less efficient at methylating arsenic. A decreased methylation 9 capacity could lead to different tissue distribution and longer retention times that might possibly 10 increase their susceptibility relative to adults. Adults have been demonstrated to excrete 40% to 11 60% of the arsenic as DMA, 20% to 25% as inorganic As, and 15% to 25% as MMA. Concha et 12 al. (1998b), however, determined that children ingesting 200 ppb (ug/L) arsenic in their drinking 13 water excreted about 49% as inorganic arsenic and 47% as DMA. Women in the same study 14 were found to excrete 66% of the arsenic as DMA and 32% as inorganic arsenic. In contrast, 15 others (Chowdhury et al., 2003; Meza et al., 2005, 2007; Sun et al., 2007) have found that 16 children have a higher urinary DMA:MMA ratio than adults, suggesting increased capacity for 17 secondary methylation. Lindberg et al. (2008) also concluded that children and adolescents (i.e., 18 <20 years of age) are more efficient methylators than adults (i.e., >20 years of age). Studying a 19 population in Bangladesh exposed to high levels of arsenic in drinking water, Sun et al. (2007) 20 found increased secondary methylation indices (SMI) in children exposed to 90 or 160 ppb of 21 arsenic in drinking water, but not in controls. Chowdhury et al. (2003) also found that the 22 increased methylation in children was only observed in exposed individuals (average 23 concentration in drinking water 382 ppb) and not in the controls (<3 ppb in drinking water). 24 This could indicate a lower saturation point for secondary methylation in adults than in children. 25 Primary methylation indices (PMI) were not age-dependent in any case. 26 Epidemiological studies provide only limited data on whether childhood exposures to 27 arsenic may result in increased cancer risk later in life. Because a significant dose-response 28 relationship has been found between cancer mortality and increased years of exposure to the 29 high-arsenic artesian well water of southwestern Taiwan (Chen et al., 1986), it is important to 30 consider the extent to which childhood exposures contributed to lifetime arsenic intake. The 31 analysis of cancer risks in the same population (Chen et al., 1992) included "only residents who 32 had lived in the study area after birth," and assumed that the arsenic intake of each person 33 continued from birth to the end of the follow-up period (1973 to 1986)3. No information was 34 provided on the exposure of pregnant women in this population to the artesian well water. The artesian wells were introduced in 1910 to 1920; prior sources of fresh water included ponds, streams, and rainwater (Tseng, 1968). 103 DRAFT—DO NOT CITE OR QUOTE ------- 1 Arsenic has been found to pass through the placenta (Hanion and Perm, 1977; Lindgren et al., 2 1984; Hood et al., 1987; Concha et al., 1998a; Jin et al., 2006a). 3 Chen et al. (1992) stated that their cancer study results may somewhat underestimate 4 arsenic-related risks in this population because tap water with lower arsenic concentrations was 5 introduced into the study area in 1956 and was available to almost 75% of the residents in the 6 1970s. Thus, the actual lifetime arsenic ingestion may be lower than estimated as residents 7 switched from the high-arsenic artesian wells to alternate water sources. Also, because this 8 study is based on mortality records (1973 to 1986) from the study region, it would not capture 9 cancer incidence among individuals exposed during childhood and early adulthood who then 10 migrated from the region. Chen et al. (1986) reported that the 1982 migration rate for this area 11 was 27%, with primarily the youths and young adults leaving the area to move to cities and those 12 45+ years old emigrating at a rate less than 6%. There is limited migration into this region, and 13 it has been reported that more than 90% of the local residents lived in the study area all their 14 lives (Wuetal., 1989). 15 There is very little epidemiology information specifically linking in utero arsenic 16 exposure to cancer outcome. Although the available epidemiological studies conducted in 17 Taiwan and other countries included women of reproductive age, the cancer outcomes from adult 18 exposures were not differentiated from in utero exposures. Recently, Smith et al. (2006), 19 examined lung cancer rates (and other respiratory diseases) in cohorts born just before the peak 20 exposure period in Antofagasta, Chile (meaning that they were not exposed in utero to high 21 levels of arsenic, but were exposed during childhood) and cohorts born during the high-exposure 22 period (indicating likely in utero exposure). Results demonstrated that exposure during either 23 period of development caused increased risk of lung cancer; however, the results from early 24 childhood exposures and/or in utero exposures were not compared to exposures during 25 adulthood to determine the possible cancer sensitivity effects in humans. 26 Although there is no adult animal model available for arsenic carcinogenesis, 27 administering inorganic arsenic to mice for 10 days during gestation has been found to increase 28 the incidence of lung, liver, reproductive, and adrenal tumors (Waalkes et al., 2003, 2004a, 29 2006a). This demonstrates that, at least in animals, embryos are more sensitive to the 30 carcinogenic effects of arsenic. 31 The Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to 32 Carcinogens (U.S. EPA, 2005b) indicates that age-dependent adjustment factors should be 33 applied to the CSF and combined with early-life exposure estimates when estimating cancer 34 risks from exposures to carcinogens with a mutagenic MOA. A mutagenic MOA for inorganic 35 arsenic has not been determined; therefore, the application of age-dependent adjustment factors 36 is not recommended. 104 DRAFT—DO NOT CITE OR QUOTE ------- 4.7.2. Possible Gender Differences 1 Differences in methylation patterns have been noted between men and women in a 2 number of studies. Higher MMA:DMA ratios have been observed in men than in women in a 3 variety of populations tested, including in the United States (Hopenhayn-Rich et al., 1996b; 4 Steinmaus et al., 2005, 2006, 2007), Taiwan (Tseng et al., 2005), and Bangladesh (Ahsan et al., 5 2007). In contrast, Loffredo et al. (2003) found that gender differences in arsenic methylation 6 varied across populations studied in Mexico, China, and Chile, sometimes by exposure level. 7 Based on mean urinary metabolite levels, they found no difference in the MMA:DMA ratio 8 between males and females in China in the group with the highest arsenic levels in their drinking 9 water (i.e., 405 ppb). Low-exposure Chinese males (i.e., those exposed to 18 ppb in drinking 10 water) had MMA:DMA ratios similar to both the high-dose males and females (0.31 to 0.32), but 11 low-dose females had a much lower (i.e., 0.22) MMA:DMA ratio. In Mexico, there was a 12 difference between the sexes at high concentrations (408 ppb in the drinking water) of arsenic 13 (i.e., the MMA:DMA ratio was 0.23 in males vs. 0.18 in females), but there was no differences 14 in the MMA:DMA ratio (0.11) at low concentrations (i.e., 30 ppb in the drinking water). In 15 Chile, a completely different pattern was observed, with females exposed to high concentrations 16 (600 ppb in the drinking water) demonstrating a higher MMA:DMA ratio (0.27) than males 17 (0.20), while the opposite pattern was seen at low concentrations (30 ppb in the drinking water; 18 0.18 in males vs. 0.13 in females). Studying a population in Bangladesh exposed to high levels 19 of arsenic in drinking water, Heck et al. (2007) found a higher percentage of urinary MMA in 20 men and a higher proportion of urinary DMA in women. 21 Age and reproductive status also may affect the male-female differences in arsenic 22 methylation patterns. Concha et al. (1998a) demonstrated that pregnant women in their third 23 trimester excrete approximately 90% of arsenic as DMA. Engstrom et al. (2007) also found 24 pregnant women to have an increased proportion of DMA in their urine compared to non- 25 pregnant women in the same population, with increases occurring with gestational age. This 26 indicates possible hormonal effects on arsenic methylation. Lindberg et al. (2007) also found 27 possible hormonal effect on arsenic methylation, noting that females younger than 60 (i.e., likely 28 pre-menopausal) generally had a more efficient methylation than men of the same age, while the 29 difference narrowed considerably in males and females over 60. Lindberg et al. (2008) found 30 that although females of all ages generally were better at methylating arsenic than males, the 31 greatest disparity between the sexes occurred between the ages of 20 and 55 (childbearing age in 32 women). Lindberg et al. (2007) also found that selenium, BMI, and AS3MT polymorphism 33 affected the observed proportions of methylated urinary arsenic metabolites in males only. The 34 pattern of arsenic methylation was also altered in males with mutations in one allele of the 35 methylenetetrahydrofolate reductase (MTHFR) gene, but in females variants in both alleles were 36 required. 105 DRAFT—DO NOT CITE OR QUOTE ------- 1 Brenton et al. (2006) used a case-control study with 900 case-control pairs to examine the 2 effect of hemoglobin levels on skin lesion prevalence in Pabna, Bangladesh. A 1.0 g/dL increase 3 in hemoglobin was found to be associated with a 21% decrease in the odds for having skin 4 lesions even after adjusting for toenail arsenic levels, BMI, education, biri or cigarette smoking, 5 chewing tobacco, and betel nut chewing. However, when the data was examined further, it was 6 discovered that the hemoglobin levels were correlated with decreased skin lesion prevalence 7 only in males (40% reduction), but not in females. Females, however, were more likely to have 8 anemia than males (18.2% vs. 8.2%; p < 0.0001). A subsequent cohort study (Brenton et al., 9 2006) found that hemoglobin levels were not associated with changes in urinary arsenic levels or 10 MMA/DMA ratios. 4.7.3. Other 4.7.3.1. Genetic Polymorphism 11 Despite the observed differences in methylation related to age and sex, data from 12 Bangladesh analyzed by Lindberg et al. (2008) suggest that genetic polymorphism is the most 13 important factor affecting the methylation of inorganic arsenic, with only 30% of variation in 14 methylation patterns attributable to level of arsenic exposure, gender, and age. Most humans 15 excrete 10% to 30% of absorbed inorganic arsenic as unchanged in urine, 10% to 20% as MMA, 16 and 60% to 80% as DMA. Excretion patterns vary across populations, however. A study of 17 urinary arsenic in a population in northern Argentina exposed to arsenic via drinking water 18 demonstrated an average of only 2% MMA in the urine (Vahter et al., 1995b; Concha et al., 19 1998b). Studies on populations in San Pedro and Toconao in northern Chile demonstrated 20 differences in the ratio of MMA:DMA excretion between the two populations (Hopenhayn-Rich 21 et al., 1996b). Chiou et al. (1997) found that in a population in northeastern Taiwan, 27% of the 22 arsenic consumed was excreted as MMA. Although these variations have not been 23 unequivocally linked with genetic factors, as opposed to environmental or nutritional factors, 24 human genetic polymorphism has been reported for methyltransferases believed to be involved 25 in arsenic metabolism (e.g., thiopurine S-methyltransferase; Yates et al., 1997). 26 Chung et al. (2002) studied the association of familial relationships with urinary arsenic 27 methylation patterns in 11 families (father, mother, and two children studied from each family) 28 from Chile where drinking water concentrations were 735-762 ppb. Their results indicate that 29 13-52% of the variation in methylation patterns could be explained by being a member of a 30 specific family. There was a high and significant correlation in the methylation patterns between 31 siblings and a much lower correlation between parent and child, which could be attributed to 32 inherent differences in methylation patterns between children and adults. Adjusting for 33 nutritional factors (blood levels of methionine, homocysteine, folate, vitamin Be, selenium, and 34 vitamin 812) did not notably alter the correlation. As might be expected, the correlation between 106 DRAFT—DO NOT CITE OR QUOTE ------- 1 father and mother was relatively low, even when adjusted for age and gender. However, the 2 correlation became stronger when adjusting for homocysteine levels as well. 3 Meza et al. (2005) found a strong association between the variations in the DNA 4 sequence of AS3MT and urinary DMA:MMA ratios in native populations in Yaqui Valley in 5 Sonora, Mexico. Three polymorphic sites were found to be associated with increased 6 DMA:MMA levels in the study population, but site 30585 was most strongly associated with 7 urinary arsenic metabolite patterns. Using a stepwise linear regression model with DMA:MMA 8 as the dependent variable and 30585 genotype, age, sex, and log-converted daily arsenic dose as 9 independent variables, only the 30585 genotype and age were found to have a highly significant 10 association with DMA:MMA levels. Further investigation determined that there was no 11 significant genetic association observed in adults, but there was a highly significant effect in 12 children aged 7 to 11 years. There was no difference in the allele frequencies at the 23 sites 13 examined between the adults and children. 14 Engstrom et al. (2007) also found a strong association between the presence of three 15 intronic single nucleotide polymorphisms in AS3MT (i.e., G12390C, C14215T, and A35991G) 16 and increased DMA levels. The study population consisted of adult women living in San 17 Antonio de los Cobres (a village in the northern Argentinean Andes) who were exposed to 18 approximately 200 ppb of arsenic in their drinking water. This group provided a rather uniform 19 genetic background against which to examine the impact of polymorphism alone as a variant. 20 Subjects who were homozygous for one or more of the variant alleles had lower MMA and 21 higher DMA levels than heterozygotes, who in turn had lower MMA:DMA ratios than 22 individuals lacking the alleles. Because the proportion of ingested inorganic arsenic that was 23 excreted was relatively constant across the groups, the effects of the variants were attributed 24 primarily to increased secondary methylation. Individuals homogenous for all three variant 25 alleles were found to have the lowest proportions of urinary MMA and the highest proportions of 26 DMA among all the groups studied. 27 A case-referent study in Bangladesh evaluated arsenic metabolite patterns in 594 28 individuals with arsenic-related skin lesions compared to 1,041 controls (Ahsan et al., 2007). A 29 correlation was found between increased arsenic concentrations in the drinking water, increased 30 proportions of MMA in the urine, and the risk of skin lesions, suggesting that variations in 31 secondary methylation could increase the risk of developing such lesions. Individuals with 32 variants in MTHFR (677TT/1298AA and 677CT/1298AA diplotypes) also had slightly increased 33 skin lesion risk (OR 1.66 and 1.77, respectively). However, the risk for developing skin lesions 34 in relation to all at-risk alleles for the GSTO1 diplotype was 3.91. Additivity of effect was 35 observed when the genotypes were analyzed jointly with water arsenic concentrations and 36 proportion of urinary MMA. 107 DRAFT—DO NOT CITE OR QUOTE ------- 1 Steinmaus et al. (2007) examined the association between genetic polymorphisms in 2 MTHFR and GST and urinary arsenic metabolites in 170 subjects from Argentina. Subjects with 3 the TT/AA variant of MTHFR 677/1298 were found to have higher urinary proportions of 4 inorganic arsenic and MMA (not statistically significant) and lower levels of DMA, with the 5 results being more pronounced in males. A null genotype of GSTM1 in women was 6 significantly associated with lower proportions of urinary MMA and higher proportions of 7 urinary DMA compared to women with the active genotype. While the same trend was observed 8 in males, it was weaker and did not achieve statistical significance. Polymorphism in the GSTT1 9 gene was not associated with differences in arsenic methylation. Lindberg et al. (2007) also 10 found that carriers of the variant allele of the M287T (C—>T) polymorphism of the AS3MT gene 11 or the A222V (C—>T) polymorphism in the MTHFR gene had higher proportions of urinary 12 MMA. 13 McCarthy et al. (2007a,b) examined the effect of GST polymorphisms on skin lesion risk 14 in a case-control (600 pairs) study in Pabna, Bangladesh. In one study (2007a), they found that a 15 10-fold increase in MMA/inorganic arsenic ratio was associated with a 1.5-fold increase in risk 16 of skin lesions. There was a significant interactive effect between GSTT1 wild-type and 17 secondary methylation on skin lesions, but no interactive effects with the GSTM1 or GSTP1 18 genotypes or any of the genotypes with primary methylation. In their second study (2007b), 19 however, they found a greater risk for skin lesions in GSTT1 wild-type (OR=1.56, 95% CI 1.10- 20 2.19) compared to GSTT1 null status (referent group). The presence of the GSTP1 GG genotype 21 was associated with a 1.86-fold increase (95% CI: 1.15-3.00) in risk of skin lesions over the AA 22 genotype. However, none of the polymorphisms examined (i.e., GSTT1, GSTM1, and GSTP1) 23 were found to modify the association between arsenic exposure and skin lesion risk. 24 Banerjee et al. (2007) also found a significant correlation between genetic polymorphism 25 and skin lesions in a population in West Bengal, India. This population was selected because 26 even though over 6 million people are exposed to high arsenic levels, only 15% to 20% 27 developed skin lesions. Polymorphisms in ERCC2, which is a NER pathway gene, was 28 examined. Specifically, the relationship between the ERCC2 codon 751 A—>C polymorphism 29 (lysine to glutamine) and skin lesion risk. Subjects exposed to arsenic-contaminated drinking 30 water with hyperkeratosis (n = 165) were compared to those without skin lesions (n = 153). 31 Occurrence of hyperkeratosis was strongly associated with the Lys/Lys genotype in the ERCC2 32 codon 751, with an OR of 4.77 (95% CI: 2.75-8.23). A significant increase in chromosomal 33 aberrations in individuals with the AA genotype compared to either the AC or CC genotypes 34 combined was also observed. 35 Brenton et al. (2007a) observed a positive association between total urinary arsenic and 36 oxidative stress (as measured by 8-OHdG) in healthy women (only females were studied) from 37 Pabna, Bangladesh, with the GSTM1 null genotype. No such association was found in GSTM1 108 DRAFT—DO NOT CITE OR QUOTE ------- 1 positive women. APE1 (apurinic/apyrimidinic endonuclease) was found to be a predictor of 8- 2 OHdG levels with the variant allele associated with a decrease in 8-OHdG. Other factors that 3 also were predictive of 8-OHdG levels included creatinine, betel nut chewing, presence of 4 environmental tobacco smoke in the home (even though none of the women reportedly smoked 5 themselves), and education. 6 In a case-control study with 792 pairs with and without skin lesions in Pabna, 7 Bangladesh, Brenton et al. (2007b) studied the association between genetic polymorphisms in 8 the base excision DNA repair pathway and arsenic-induced skin lesions. Four common base 9 excision repair (BER) genetic polymorphisms (X-ray repair cross-complimentary group 1 10 [XRCC1] Arg399Gln, XRCC1 Argl94Trp, human 8-oxoguanine DNA glycosylase [hOGGl] 11 Ser326Cys, and APE1 Aspl48Glu) were examined. APE1 148 Glu/Glu individuals were twice 12 as likely to have skin lesions as APE1 148 Asp/Asp individuals, even after adjusting for toenail 13 arsenic concentration, BMI, education, smoking, and betel nut use. Presence of the Glu/Glu 14 variant of APE1 Aspl48 Glu was associated with a 2- to 2.5-fold increased OR for skin lesions 15 compared to the Asp/Asp variant, in the low and middle tertiles, but no increase was observed in 16 risk at the highest tertile of exposure. XRCC1 Argl94 Trp genotypes, however, were not 17 associated with skin lesion risk in the low and middle tertiles, but were associated with a 3-fold 18 difference in the highest exposure tertile (i.e., OR of 2.9 for Trp/Trp compared to 8.4 for 19 Arg/Arg where Arg/Arg at the lowest tertile is the referent group). No association was observed 20 between skin lesions and genetic polymorphisms in XRCC1 Arg399Gln or hOGGl Ser326Cys 21 alleles. 4.7.3.2. Nutritional Status 22 In many of the epidemiological studies discussed above (e.g., southwestern Taiwan and 23 Bangladesh), the study subjects were relatively poor and had poor nutritional status. Mazumder 24 et al. (1998) demonstrated that people in and around West Bengal who had body weights below 25 80% for their age and sex had an increased RR (2.1 for females and 1.5 for males) in the 26 prevalence of arsenic-associated keratosis. Lindberg et al. (2008), however, found that women 27 in Bangladesh were better at methylating arsenic than men even though they were less likely to 28 eat nutritious food (e.g., meat and fresh vegetables) than men, indicating that gender was a better 29 predictor of methylation capacity than nutritional status in this group. 30 Selenium has been demonstrated to reduce the teratogenic, clastogenic, and cytogenic 31 effects of arsenic (ATSDR, 1993). Chen et al. (2007) found that individuals in the Health 32 Effects of Arsenic Longitudinal Study (HEALS; population from Araihazar, Bangladesh) with 33 low selenium intake were at a greater risk for developing pre-malignant skin lesions than those 34 with adequate intake. In 93 pregnant women from Antofagasta, Christian et al. (2006) found that 35 increases in urinary selenium levels were associated with increased urinary arsenic excretion, 36 and with a greater percent excreted as DMA and less excreted as inorganic arsenic. The 109 DRAFT—DO NOT CITE OR QUOTE ------- 1 proportion of urinary MMA was fairly consistent in the study population. Using four quartiles of 2 increasing urinary selenium levels, results showed that the total arsenic excretion increased 3 steadily across quartiles of selenium intake. The proportion of DMA excreted increased, and the 4 proportion of inorganic arsenic excreted decreased with increasing selenium intake, but only in 5 the first two quartiles. Although different gestational stages of pregnancy have been associated 6 with differences in urinary arsenic excretion patterns, this was controlled for in the analysis. 7 Gamble et al. (2005) suggest that adequate folate is necessary for both primary and 8 secondary arsenic methylation and that adequate folate intake is associated with increased 9 urinary DMA. Gamble et al. (2006) found that providing folate supplements to individuals from 10 Araihazar, Bangladesh, with a diet low in folate significantly increased the proportion of arsenic 11 excreted as DMA in the urine. Heck et al. (2007), however, found that levels of folate 12 consumption (measured by levels in the food) were directly related to percentages of urinary 13 MMA, but not to changes in urinary DMA in a population from Bangladesh (participants of the 14 HEALS study) exposed to arsenic in drinking water. Heck et al. found no correlation between 15 intake of folate-related nutrients and urinary DMA levels, but found that increases in methionine, 16 vitamin B12, calcium, protein, and riboflavin were associated with decreases in the proportion of 17 urinary inorganic arsenic and increases in the percent of urinary MMA. Niacin and choline were 18 found to be the better predictors of secondary methylation (as measured by DMA/MMA). 19 Although high levels of plasma homocysteine were not associated with urinary MMA levels, 20 they were associated with a decrease in DMA levels (Gamble et al., 2005). 21 Mitra et al. (2004) studied whether nutritional deficiencies increased the susceptibility of 22 individuals to arsenic-related health effects as measured by skin lesions. In West Bengal, India, 23 where exposures were <500 ppb, nutritional assessments were based on a 24-hour recall for 24 major dietary constituents and a 1-week recall for less common constituents. Increases in risk 25 were associated with low intake of animal protein (OR=1.94, 95% CI: 1.05-3.59), calcium 26 (OR=1.89, 95% CI: 1.04-3.43), fiber (OR=2.20, 95% CI: 1.15-4.21), and folate (OR=1.67, 95% 27 CI: 0.87-3.2). Nutrient intake was not related to arsenic exposure. The authors concluded that 28 the potential protective effects of these nutrients were small in comparison to eliminating the 29 exposure to arsenic. 30 Steinmaus et al. (2005) found an association between low dietary protein, iron, zinc, and 31 niacin, and decreased production of urinary DMA accompanied by increased levels of urinary 32 MMA in arsenic-exposed individuals from a U.S. population. An associations between arsenic 33 methylation patterns and dietary intake of thiamine, vitamin B6, lutein, and a-carotene were 34 found, but the links were not as clear when adjusted for confounding variables (i.e., age, sex, 35 smoking, and total urinary arsenic levels). The authors suggest, however, that the effect of 36 specific nutrient intake levels on methylation patterns was small in comparison with the known 110 DRAFT—DO NOT CITE OR QUOTE ------- 1 magnitude of inter-individual variability associated with genetic polymorphisms. Kreppel et al. 2 (1994) found that dietary zinc protects mice against acute arsenic toxicity. 4.7.3.3. Cigarette Smokers 3 Cigarette smokers (current or former) were found to have a decreased secondary 4 methylation capacity, resulting in increased urinary MMA and decreased DMA concentrations 5 (Huang et al., 2007b). Tseng et al. (2005) reported a decrease in secondary metabolism in 6 cigarette smokers exposed to arsenic-contaminated drinking water, resulting in a significant 7 increase in the secreted MMA as a fraction of total metabolites. Steinmaus et al. (2005) found 8 that current smokers in a U. S. population had lower proportion of arsenic excreted as DMA than 9 either former or never-smokers (although the difference was not statistically significant). 10 Steinmaus et al. (2006) found that in a population in Argentina the proportion of excreted MMA 11 was associated with bladder cancer risk in former smokers, but not in individuals who had never 12 smoked. Subjects who had ever smoked and had proportions of MMA in the upper tertile had a 13 2-fold elevated risk of bladder cancer compared to subjects with proportions of MMA in the 14 lower two tertiles. Therefore, it was concluded that individuals who smoke had an increased 15 susceptibility to arsenic toxicity. Steinmaus et al. (2006) also studied a population in the United 16 States. Although the results indicated increased MMA was associated with increased cancer 17 risk, the number of cases was too small to estimate separate ORs for never-smokers and ever- 18 smokers. 19 111 DRAFT—DO NOT CITE OR QUOTE ------- 5. DOSE-RESPONSE ASSESSMENTS 5.1. ORAL REFERENCE DOSE (RfD) 1 An RfD was developed for inorganic arsenic and posted on the IRIS database in 2 1991. An oral noncancer dose-response estimation is not addressed in this document. However, 3 the Agency is currently reviewing the literature and will develop an updated RfD at a later date. 5.2. INHALATION REFERENCE CONCENTRATION (RfC) 4 An inhalation noncancer dose-response estimation is not addressed in this document. An 5 RfC is not developed for inorganic arsenic, nor does a current value exist on the IRIS database. 5.3. CANCER ASSESSMENT (ORAL EXPOSURE) 5.3.1. Background: History of Cancer Risk Assessments for Arsenic 6 This assessment is unusual in that it builds on a long history of previous efforts by EPA 7 and others to evaluate potential risks from oral exposure to arsenic via drinking water. Table 5-1 8 summarizes previous assessments and expert reviews of arsenic carcinogenicity. 9 The table starts (chronologically) with EPA's 1988 risk assessment for skin cancer (U.S. 10 EPA, 1988b). The scope of the 1988 assessment was to review the applicability of EPA's 1984 11 assessment (U.S. EPA, 1984) on skin cancer risk from the Taiwanese population to the U.S. 12 population. The skin cancer risk from oral exposure was estimated based on two studies (Tseng 13 et al., 1968; Tseng, 1977) of age-specific prevalence rates for skin cancer in a large cohort of 14 Taiwanese (40,241 subjects in 37 villages) in an "arseniasis-endemic" area, where arsenic 15 concentrations in water supply wells ranged from less than 10 ug/L (ppb) to 1,820 ug/L. The 16 occurrence of skin cancer was estimated in a survey lasting approximately 2 years (U.S. EPA, 17 1988b). Preliminary data from the same cohort suggested that risks of internal cancers (lung, 18 liver, and bladder) were also elevated, but U.S. EPA (1988b) concluded that insufficient data 19 were available to support a dose-response assessment for these effects. 20 The second entry in the table is the National Research Council's 1999 review (NRC, 21 1999) of EPA's 1988 risk assessment. EPA commissioned NRC to review the U.S. EPA (1988b) 22 assessment and also the qualitative and quantitative evidence on arsenic and health effects for 23 reassessment of human health risks from arsenic in drinking water. One of the major 24 recommendations of NRC's 1999 review was that studies from the arsenic-endemic area of 25 Taiwan (Wu et al., 1989; Chen et al., 1988a, 1992) provide the best available empirical human 26 data for assessing the risks of arsenic-induced cancer. The report explored quantitative modeling 27 approaches for the male bladder cancer data, but did not provide a formal risk assessment; 112 DRAFT—DO NOT CITE OR QUOTE ------- 1 additional modeling analyses were recommended. NRC 1999 applied absolute Taiwan risks to 2 the U.S. populations. 3 NRC (1999) published the arsenic concentration in village wells, person-years of males 4 and females by village and the village-specific lung, bladder, and liver deaths for the Wu et al. 5 (1989) and Chen et al. (1992) studies. Additional raw data were obtained from study authors by 6 Morales and Ryan during reanalysis and these data were subsequently provided to EPA 7 (personal communications). All of the succeeding assessments summarized in Table 5-1 derive 8 dose-response estimates based on the internal cancer data. 9 In the first of these efforts, Morales et al. (2000) gathered data on lung, bladder, and liver 10 cancer, as well as detailed exposure data (well arsenic concentrations) from the three 11 epidemiological studies (Wu 1989; Chen et al., 1988a, 1992), and evaluated a range of statistical 12 models for estimating potential arsenic-related cancer risks in the Taiwanese population and for 13 extrapolating these risks to the U.S. population. In promulgating the Primary Drinking Water 14 Standard for Arsenic, U.S. EPA (2001) adopted one of Morales et al.'s models, with adjustments 15 of some exposure assumptions, for estimating the health benefits of regulatory alternatives. The 16 Office of Pesticide Programs (OPP) also recently applied oral CSFs based on the U.S. EPA 17 (2001) assessment in their Reregi strati on Eligibility Decision (RED) Documents for organic 18 arsenic pesticides (U.S. EPA, 2006c) and for Inorganic Arsenicals and/or Chromium Based 19 Wood Preservatives (U.S. EPA, 2008). 20 In response to continued public concern over arsenic-related cancer risks, EPA asked 21 NRC to update its 1999 recommendations in light of new scientific evidence, and to review the 22 risk assessment in support of the 2001 drinking water standard. NRC (2001) reviewed the 23 methodology used in EPA's arsenic risk assessment (U.S. EPA, 2001) and provided a systematic 24 analysis of and recommendations for applying the Taiwanese epidemiological data for assessing 25 cancer risks from arsenic exposure in U.S. populations. Recommendations included the 26 inclusion of a reference population in the dose-response assessment, the form of the dose- 27 response model, exposure assumptions, and approaches for extrapolating risks to the U.S. 28 population. As the committee noted, the cancer risk estimates that it developed were higher than 29 those reported by U.S. EPA (2001), and reasons for those differences were reviewed. EPA 30 examined and applied the NRC (2001) statistical methodology and submitted its revised analysis 31 (U.S. EPA, 2005c) to SAB for review and comment. SAB (2007) provided additional discussion 32 related to the treatment of arsenic exposure, and recommended expanded sensitivity analyses of 33 other exposure-related assumptions. EPA adopted these recommendations, along with responses 34 to comments from interagency reviewers, into the current assessment. The current quantitative 35 risk assessment can thus be described as EPA's reimplementation of the technical cancer risk 36 modeling recommendations in NRC (2001), with additional examination of arsenic exposure 113 DRAFT—DO NOT CITE OR QUOTE ------- 1 assumptions and taking into account SAB's (2007) advice for the expansion of sensitivity 2 analyses of modeling methods and choices. Table 5-1. Historical Summary of Arsenic Risk Assessment Efforts Assumption/ Method Goals/Scope of Assessment Critical Study Critical Study Endpoint(s) Dose-Response Model Reference Population Arsenic Concentration U.S. EPA (1988b) Revise EPA's 1984 risk assessment for skin cancer, evaluate evidence of arsenic essentiality Taiwan skin cancer prevalence studies (Tseng etal., 1968; Tseng, 1977) Skin cancer incidence Linear multistage Taiwanese outside arseniasis- endemic area Stratified: 0- 300, 300- 600, 600-900 ug/L in well water, unknown exposure NRC (1999) Review EPA's 1988b risk assessment, suggest alternative approaches; was "not a risk assessment" Taiwan epidemiologic al studies (Wu etal., 1989; Chen et al., 1988a, 1992) Bladder cancer mortality Weibull, Poisson regression With and without all- Taiwan Median well arsenic concentrations Morales et al. (2000) Test dose- response models, modeling assumptions U.S. EPA (2001) Estimate U.S. cancer risks in support of drinking water standard NRC (2001) Review EPA's 2001 methods and results U.S. EPA (2005c) Incorporate NRC (2001) recommenda- tions for SAB Review Taiwan epidemiological studies (Wu et al., 1989; Chen et al., 1988a, 1992) Bladder, lung, liver cancer mortality Nine Poisson forms with varying age, dose representation s; one multistage Weibull None, southwest Taiwan, all- Taiwan Median well arsenic concentration s Bladder, lung cancer mortality Morales et al. "Model 1" (multiplicative linear dose, quadratic age) None Median well arsenic concentrations Additive Poisson, linear dose, quadratic age All-Taiwan, southwest Taiwan Median; sensitivity analysis of other values Additive Poisson, linear dose, quadratic age; UCLs on dose coefficients estimated by Bayesian simulation Southwest Taiwan Median well arsenic concentrations 114 DRAFT—DO NOT CITE OR QUOTE ------- Assumption/ Method Taiwanese Water Intake Taiwanese Body Weight Nonwater arsenic Intake Risk Model for U.S. Population U.S. Incidence, Mortality Data U.S. Water Intake U.S. Body Weight U.S. EPA (1988b) 3.5 L/day (M), 2.0 L/day (F) 55 kg (M), 50 kg(F) None (0 ug/day) Simple life table Not specified 2.0 L/day (approximate 90th percentile value) 70 kg (M and F) NRC (1999) 3.5 L/day (M), 2.0 L/day (F) 55 kg (M), 50 kg(F) Not explored Simple life table NCHS 1994 mortality data 2.0 L/day (approximate 90th percentile value) 70 kg (M and F) Morales et al. (2000) Water intakes not specified Body weights not specified None (0 ug/day) Life table, 5- year age strata U.S. EPA (2001) 3. 5 L/day (M), 2.0 L/day (F) + 1.0 L/day cooking 55 kg (M), 50 kg (F) 50 ug/day (exposed population) Life table, 5 -year age strata NRC (2001) Recommenda tions based on approx. 2 L/day; sensitivity analysis of U.S./Taiwan intake ratios is presented 55 kg (M), 50 kg (F) None (0 ug/day) in baseline assessment; sensitivity analysis showed little effect of adding 30 or 50 ug/day to study village exposure estimates U.S. EPA (2005c) 2.0 L/day 50 kg (M and F) 30 ug/day exposed population only, sensitivity analyses of 0-50 ug/day BEIR IV survival model (relative risk) NCHS 1996 mortality Average U.S. water intake Average U.S. body weights 1.0-1. 2 L/day used as central tendency values; 2. 1-2.3 L for 90th percentile risk in Monte Carlo model 1.0 L/day with sensitivity analyses 1.0 L/day 70 kg (M and F) 115 DRAFT—DO NOT CITE OR QUOTE ------- Assumption/ Method Endpoints Calculated U.S. EPA (1988b) Unit risk = 3xlQ-5per Hg/L (females), 7xlO'5per ug/L (males); CSFs = 1 to 2 per mg/kg- day (incidence) NRC (1999) Lifetime bladder cancer risk at 10 ug/L = 3xl(T3 (males), 9xlO"3 (females); EDM = 404- 443 ug/L, LEDoi = 323-407 ug/L Morales et al. (2000) "Model 1," no reference pop. Males (ug/L) EDoi LEDoi Lung 364 294 Bladder 3 95 326 Females (iig/L) EDoi LEDoi Lung 258 213 Bladder 252 211 Many other results presented U.S. EPA (2001) CSFs derived from Morales et al. (2000) EDoi, LED 01 values Unit risk, per ug/L: Male bladder= 2.5xlO'5 (MLE), 3.1xlO'5(UCL) Male lung = 2.8xlO"5 (MLE), 3.4xlO'5 (UCL) Female bladder = 4.0xlO"5 (MLE), 4.7xlO"5 (UCL) Female lung= 3.9xlO'5, (MLE), 4.7xlO"5 (UCL) NRC (2001) Lifetime cancer risk incidence from 10 ug/L: Male lung = l.SxlO"3 bladder = 2.3 xlO'3 Female lung = 1.4xlO"3 bladder = 1.2xlO"3 U.S. EPA (2005c) Female lung + bladder incidence: unit risk = 1.6xlO'4per ug/L Incidence at 10 ug/L in drinking water = 1.6xlO"3 Drinking water concentration for 10'4 incidence risk = 0.63 ug/L 1 The techniques and assumptions used in arsenic risk assessment have evolved and 2 changed over time, and it is not possible to do justice to all of the changes and innovations in 3 each assessment in this chapter. Table 5-1 provides a general summary of the important data 4 sources, techniques, and assumptions employed in each assessment. Where cells in the table are 5 merged across the columns, it indicates that the same assumptions were used in more than one 6 assessment, implying a solidification of a technical consensus. The major issues addressed in 7 each study include: 8 9 • Scope and goals. Some of the efforts in Table 5-1 (the NRC studies most importantly) 10 were not intended to be comprehensive risk assessment, but to provide recommendations 11 for EPA and other agencies. Some were pure modeling studies (Morales et al., 2000), 12 and some were employed to derive quantitative risk estimates for regulatory support 13 purposes (U.S. EPA, 2001) or for health criteria development (U.S. EPA, 2005c). 14 15 • Selection of critical studies for use in the risk assessment. As noted above, the U.S. 16 EPA (1988b) assessment was based on skin cancer prevalence data (Tseng et al., 1968; 17 Tseng, 1977). All of the subsequent assessments in the table use data from later 116 DRAFT—DO NOT CITE OR QUOTE ------- 1 epidemiological studies (Wu et al., 1989; Chen et al., 1988a, 1992), which provide 2 information on PYR and cancer mortality in narrowly defined age strata, and exposure 3 concentrations from individual water supply wells. 4 5 • Critical study endpoints. Over time, assessments have moved from evaluating skin 6 cancer (U.S. EPA, 1984, 1988b) to internal cancers (lung and bladder). As discussed 7 below, the change in endpoint is the major reason that the cancer potency estimated in the 8 current assessment is so different from that derived in 1988. Wu et al. (1989) and Chen 9 et al. (1988a, 1992) also reported data on liver cancer, but in response to concerns related 10 to a high incidence of viral hepatitis in Taiwan (U.S. EPA, 2001), liver cancer has not 11 been included as an endpoint in recent assessments. 12 13 • Dose-response models. The form of the dose-response models used to assess risks in the 14 Taiwanese population has evolved over time as different investigators explored the 15 performance of various models under a wide range of exposure assumptions. In the early 16 models, linear regression and multistage models were used for dose-response assessment 17 in the Taiwanese population. In the more recent analyses, Poisson regression with linear 18 dose terms and quadratic age terms have been employed, as recommended by NRC 19 (2001), to derive primary risk estimates. In addition, sensitivity analyses of other Poisson 20 models (different transformations of dose) have been conducted, as recommended by 21 SAB (2007). Changes in the modeling approaches, like changes in the endpoints 22 modeled, have resulted in changes in estimated cancer potency. 23 24 • Inclusion/exclusion of a reference population. EPA's 2001 risk assessment was based 25 on a dose-response model for the Taiwanese population that did not include a reference 26 population (i.e., a group with similar characteristics not exposed to arsenic in drinking 27 water). In keeping with NRC (2001) and SAB (2007) comments, the primary estimates 28 in this chapter are derived based on the inclusion of a reference population from 29 southwest Taiwan; sensitivity analyses are provided for risk estimates with the reference 30 population excluded and with a reference population from all regions of Taiwan (i.e., 31 "all-Taiwan"). 32 33 • Arsenic concentration used in the dose-response model. The available exposure data 34 (Wu et al., 1989; Chen et al., 1992) consist of measurements from 155 village drinking 35 water wells taken between 1964 and 1966 for 42 exposed villages. Most of the 36 assessments in Table 5-1 employed the median exposure concentrations for each group. 37 That approach also is followed in this assessment; however, following SAB (2007) 38 recommendations, a sensitivity analyses on the impacts of using minimum and maximum 39 village arsenic concentrations in the risk assessment has been conducted. 40 41 • Water intake and body weight of the exposed population. As discussed in Section 42 5.3.5, there are few precise data available concerning the distribution of daily drinking 43 water intake volumes in the exposed populations. As shown in Table 5-1, past 44 assessments have employed a range of assumptions; the basic consensus is that 45 Taiwanese men appear to consume more water than men in the U.S. owing to the hotter 46 climate, and because a large proportion of them engage in vigorous outdoor activity as 47 part of their livelihood. Consistent with the limited information, the current analysis has 117 DRAFT—DO NOT CITE OR QUOTE ------- 1 followed this consensus. Following other analyses, this assessment assumes an average 2 body weight of 50 kg for both Taiwanese men and women. 3 4 • Nonwater arsenic intake. Because the risk modeling for the Taiwanese population is 5 based on estimated daily arsenic dosage, it is important to include reasonable 6 assumptions about the amount of arsenic intake coming from non-drinking water sources. 7 This is an area where there is relatively little data, and considerable confusion about, for 8 example, whether and how to include a contribution from cooking water, reasonable 9 estimates of arsenic concentrations in food, and whether the arsenic-exposed and 10 reference populations should be assumed to receive the same nonwater arsenic intake. 11 The various assumptions used in previous analyses are summarized in Table 5-1, and the 12 basis for nonwater arsenic intake estimates used in this assessment is discussed in Section 13 5.3.5. As is the case for many other assumptions, the approach to dealing with 14 uncertainty in nonwater arsenic intake is to conduct sensitivity analyses based on a 15 reasonable range of values. 16 17 • Risk model for the U.S. population. The outputs of the dose-response modeling for the 18 Taiwanese population were arsenic dose-response coefficients that described the 19 relationship between estimated arsenic intake in the Taiwanese population and 20 proportional increases in age-specific lung and bladder cancer mortality risk. Consistent 21 with NRC (2001) recommendations, lifetime cancer incidence in U.S. populations was 22 then estimated by using a modified version of the "BEIRIV" relative risk model, as 23 described in Appendix E. A key assumption underlying this model is that the risk of 24 arsenic-related cancer mortality or incidence for the U.S. population is a constant 25 multiplicative function of the current "background" age profile of cancer risks in the 26 same U.S. population. 27 28 • U.S. mortality and cancer incidence data. Models for extrapolating cancer risks for the 29 U.S population require data on overall mortality, and the BEIR IV model requires non- 30 arsenic related cancer incidence data for the U.S. population. One source of variation in 31 the cancer risk estimates over time has been the use of more recent mortality and cancer 32 incidence data in the most recent assessments. 33 34 • U.S. water intake and body weight. Estimates of the drinking water intake and typical 35 body weight of the exposed population are also needed to predict cancer risks in the U.S. 36 population. All of the recent assessments assume body weight of 70 kg for males and 37 females. For the primary risk estimates, the current assessment assumes a water intake of 38 2.0 L/day, as discussed in Section 5.3.5, with sensitivity analyses of other values. Adult 39 water intake of 2.0 L/day is used as a standard factor in EPA IRIS assessments, and 40 represents approximately the 90th percentile of intake of community water in the U. S. 41 population. Other intake assumptions (e.g., mean versus upper percentile) can be used in 42 risk assessments, depending on target population characteristics and assessment needs. 43 44 • Endpoints calculated. As can be seen in Table 5-1, different assessments have 45 calculated a range of risk endpoints, including EDoiS, LEDMs, lifetime cancer risks, CSFs, 46 and drinking water concentrations corresponding to various cancer risk levels. As 47 discussed in Section 5.3.8.2, this can create some difficulty in comparing the results 48 across assessments, since converting from one measure to another can require 118 DRAFT—DO NOT CITE OR QUOTE ------- 1 assumptions related to exposure that may not have been clearly specified. Where they 2 have been calculated, the most commonly used and easily comparable endpoints are 3 provided, including drinking water unit risks (lifetime cancer incidence associated with 4 1 ug/L exposure), estimated cancer risk at 10 ug/L in drinking water, and the drinking 5 water concentration associated with a lifetime cancer risk of 10"4. 6 7 Given the many features of the risk assessment for arsenic that have changed over time, it 8 is not surprising that the magnitude of the risk estimates has also varied from assessment to 9 assessment. As discussed above, the CSF from U.S. EPA's (1988b) assessment, which is 10 derived based on skin cancer prevalence, is not directly comparable to CSFs derived from 11 internal cancer data in the later assessments. Section 5.3.8.2 discusses modeling methods and 12 assumptions used in the current assessment, describing precisely how they differ from previous 13 analyses. 5.3.2. Choice of Study/Data, Estimation Approach, and Input Assumptions 14 As discussed in Section 4.2, the few animal carcinogenicity bioassays that have been 15 conducted on inorganic arsenic compounds do not provide data of high enough quality to use in 16 human dose-response modeling (NRC, 2001; SAB, 2000, 2007). There are, however, several 17 epidemiologic studies that relate human exposures to arsenic in drinking water to cancer risk. 18 NRC (2001) and SAB (2007) concluded that the epidemiological studies by Chen et al. (1988a, 19 1992) and Wu et al. (1989) that use the southwestern Taiwanese population provide the best 20 available data for conducting a quantitative risk assessment for exposure to arsenic in drinking 21 water. SAB (2007) cited the important strengths of the data, including the large population, 22 extensive follow-up (almost 900,000 person-years), large number of exposed villages (42), large 23 number of lung and bladder cancer deaths (441), reliability of the population and mortality 24 counts, and stability of residential patterns, stating that: 25 • ".. .in view of the size and statistical stability of the database relative to other studies, the 26 reliability of the population and mortality counts, the stability of residential patterns, and 27 the inclusion of long-term exposures, it is the Panel's view that this [the Taiwanese] 28 database remains, at this time, the most appropriate choice for estimating cancer risk 29 among humans. Supporting this view is the fact that the datasets from Taiwan have been 30 subjected to many years of peer review as part of published studies." 31 In keeping with SAB's recommendations, epidemiological studies by Smith et al. (1998) 32 and Ferreccio et al. (2000) on arsenic-related lung cancer in Chile, as well as studies by Chiou et 33 al. (2001) and Chen et al. (2004a), were evaluated (see Section 4.1 and Appendix B); however, 34 these studies were not considered to be of comparable quality to the Taiwanese data set for use 35 in the quantitative assessment. The dose-response estimation discussed below, like previous 36 analyses, is based on the southwest Taiwanese data and incorporates the NRC and SAB 37 recommendations for modeling approaches and sensitivity analyses. 119 DRAFT—DO NOT CITE OR QUOTE ------- 5.3.3. Dose-Response Model Selection for Cancer Mortality in Taiwan 1 Despite the high quality of the data set, estimation of dose-response relationships based 2 on the Taiwanese data is challenging for a number of reasons. First, owing to the "ecological" 3 nature of the study, drinking water exposure information is not available for individual study 4 subjects. Instead, drinking water arsenic exposure must be estimated based on measured arsenic 5 concentrations in wells serving the 42 population groups ("villages") that constitute the study 6 population. For 20 of the 42 villages, water was supplied by a single well at the time of 7 sampling. For another 10 villages, water was supplied by two wells; the remaining villages used 8 more than two wells. Data provided are related to all the arsenic measurements for each well in 9 each village, but no information is available concerning the time variability of arsenic levels in 10 individual wells. 11 In addition to villages where drinking water arsenic concentrations were measured, the 12 epidemiological data used in this assessment include information on the cancer mortality in two 13 reference populations (southwest Taiwan and all of Taiwan) for the same period covered by the 14 Chen et al. (1988a, 1992) studies. Drinking water concentrations for the reference populations 15 were not measured, but are assumed to be lower than those seen in the 42 arsenic-exposed 16 villages (zero drinking water arsenic intake was assumed for the reference populations). As 17 discussed below, the data on the nonwater arsenic intakes available for both the exposed and 18 reference populations are very limited (Schoof et al., 1998), so the impacts of different 19 assumptions are explored through a sensitivity analysis. 20 It is clear that cancer mortality in the reference population and in the arsenic-exposed 21 villages is strongly age-dependent, with the older study subjects generally exhibiting higher 22 mortality. The age-dependence does not appear to be monotonic, however, but rather peaks 23 around age 60 and declines thereafter. This non-linear age-dependence complicates the 24 estimation of dose-response relationships because it requires the estimation of models using non- 25 standard methods. 26 Chen et al. (1992) used an Armitage-Doll time-to-tumor model to estimate cancer risks as 27 a function of dose in this population for 20-year age strata, but the model they used assumed 28 monotonically increasing cancer risk with age. As discussed below, using narrower age strata (5 29 years), the non-monotonic dependence of cancer risk on age becomes more apparent. Morales et 30 al. (2000) used a variety of non-linear models to fit dose-response functions to data derived from 31 the Chen et al. (1988a, 1992) and Wu et al. (1989) studies. They derived cancer slope estimates 32 for arsenic-associated cancers of the bladder, liver, and lung by using Poisson regression with a 33 number of different methods for expressing the dependence of risks on age and arsenic intake. 34 When no reference population was included in the data, the best-fitting model included a 35 quadratic function of age and a linear exponential term for dose. When the southwest Taiwan 36 reference population was included in the risk modeling, the best-fitting model again included a 120 DRAFT—DO NOT CITE OR QUOTE ------- 1 quadratic age model, but an exponential function of log-transformed dose. A number of other 2 models with different age and dose terms were found to fit nearly as well as judged by the 3 Akaike Information criterion (AIC). Many of the models also were very sensitive to changes in 4 input assumptions. 5 NRC (2001) reviewed the U.S. EPA (2001) cancer assessment including application of 6 the model from the Morales et al. (2000) study and conducted independent analyses of the data 7 in order to systematically evaluate the effects of different modeling approaches, assumptions 8 related to background cancer rates, and individual variability in exposures. As noted above, they 9 recommended two specific changes to EPA's modeling approach; the inclusion of a reference 10 population, and the use of an additive (rather than multiplicative) linear dose term in the Poisson 11 regression. SAB (2007) also reviewed EPA's modeling procedures. Given the NRC 12 recommendations and results of the SAB review, the current model (see Section 5.3.7) employs 13 the following approaches: 14 15 • Poisson regression (of cancer mortality against age and dose) fit by maximum likelihood 16 estimation (MLE). 17 18 • A quadratic age model. 19 20 • Additive linear dose term. 21 22 • Confidence limits on the dose terms estimated by profile likelihood. 23 24 • Primary risk estimates derived for the data set that includes the southwest Taiwan 25 reference population. 26 27 As recommended by SAB, sensitivity analyses were conducted to evaluate the impacts of 28 different modeling assumptions (nonwater arsenic intake, daily water intake, and reference 29 population) on risk estimates. Several different model forms (quadratic, exponential linear, and 30 exponential quadratic dose transformations) also were evaluated (see Section 5.3.8.4 for further 31 detail). 5.3.4. Selection of Cancer Endpoints and Estimation of Risks for U.S. Populations 32 Lung and bladder cancer mortality in the Taiwanese population have been chosen as 33 endpoints in the dose-response modeling because they are the internal cancers most consistently 34 observed and best characterized in epidemiological studies of arsenic exposure (U.S. EPA, 2001; 35 NRC, 2001). Oral CSFs and other risk metrics were calculated separately for each endpoint and 36 gender. 37 Although liver cancer risks also were examined by Morales et al. (2000), they were not 38 included in the quantitative risk assessment because the observed liver cancer mortality in the 121 DRAFT—DO NOT CITE OR QUOTE ------- 1 southwest Taiwanese population was thought to be affected by a high incidence of viral 2 hepatitis, which made attribution of risks to arsenic problematic. As noted in Section 4.1, 3 arsenic-related skin cancer also has been noted in the Taiwanese population (and in other 4 arsenic-exposed groups), but this endpoint was not included in the cancer risk assessment for 5 several reasons. The high mortality rates for internal cancers, compared to skin cancers which 6 are rarely fatal, makes the internal cancers an appropriate critical health endpoints for the cancer 7 risk assessment. In addition, the internal cancers were identified as the critical endpoints 8 because the estimated cancer potency of arsenic for lung and bladder cancers was much greater 9 than the potency estimated for skin cancers (see Section 5.3.8.1). The development of pre- 10 cancerous skin lesions (as reported by Ahsan et al., 2006) is being addressed separately in EPA's 11 noncancer risk assessment. 12 The current risk model includes multiplicative terms for age and dose. Therefore, the 13 risk calculated for a target population (e.g., a U.S. population exposed to arsenic in drinking 14 water) depends on the "background" cancer risk, i.e., the expected age-specific cancer risk in the 15 U.S. population in the absence of arsenic exposure. Morales et al. (2000) calculated lifetime 16 arsenic-related mortality risks for the U.S. population exposed to different drinking water 17 concentrations by applying age-specific hazard functions (derived from the dose-response 18 models estimated for the Taiwanese population) to a "life table" of age-specific probabilities of 19 death for the U.S. population. These calculations were based on data from 1996. 20 In response to comments from NRC and SAB, a slightly different approach to estimate 21 cancer risks for U.S. populations is being used. In the following analysis, arsenic concentrations 22 corresponding to an additional 1% lifetime cancer incidence (effective dose; ED01 values) above 23 "background" are derived for each endpoint. Also derived are lowest effective dose (LED01) 24 values, which represent the lower confidence limits on the dose corresponding to a one percent 25 lifetime incidence risk in the U.S. population. Consistent with EPA's Guidelines for Carcinogen 26 Risk Assessment (U.S. EPA, 2005a) and the NRC (2001) cancer assessment, risk estimates are 27 derived based on a linear extrapolation from the points of departure (LEDOTs for lung, bladder, 28 and combined cancers) because the MOA for inorganic arsenic is unknown. 29 The ED01 and LEDM values are estimated using a variation on the "BEIRIV" model 30 derived for use in estimating population cancer risks for radionuclide exposures (NRC, 2001). 31 This method, which is described further in Section 5.3.7.3 and Appendix E.2, includes the 32 application of relative cancer risk estimate derived from the Taiwanese dose-response 33 assessment multiplicatively to age-specific cancer risks for the United States. In this model, the 34 background hazard consists of age-specific cancer incidence data for bladder and lung cancer 35 from the United States for the years 2000 to 2003 (NCI, 2006). The ratios of cancer mortality to 36 incidence for arsenic-related cancers are assumed to be the same in the U.S. and Taiwanese 37 populations. 122 DRAFT—DO NOT CITE OR QUOTE ------- 5.3.5. Nonwater Arsenic Intake and Drinking Water Consumption 1 It is important to clarify that the nonwater arsenic intake value corresponds to the arsenic 2 amount from dietary sources (rice and yams, the dietary staples for the Taiwanese population in 3 the endemic area) only. It does not include the arsenic intake value from water used for cooking 4 rice or produce, which was addressed separately via sensitivity analysis modeling with higher 5 water intake values. 6 For the baseline risk calculations, the nonwater arsenic intake was assumed to be 10 7 ug/day for the reference and exposed populations. Although the data supporting this value are 8 scarce, it appears to be a reasonable intake estimate for the reference populations based on the 9 available information. U.S. EPA (1989) estimated the arsenic intake based on soil arsenic level 10 and rice consumption in Taiwan to be between 2 and 16 ug/day. The higher value was presumed 11 to result from possible soil contamination by organic arsenical herbicides applications. U.S. 12 EPA (1989) found no reliable data to estimate arsenic intake from sweet potato (yam) 13 consumption by the southwest Taiwanese population. In a separate study, Schoof et al. (1998) 14 estimated that the total inorganic arsenic intake from food sources in the endemic area in Taiwan 15 ranged between 15 and 211 ug/day, with the average intake value as 50 ug/day. This arsenic 16 intake value is based on analysis of limited rice and yam samples collected in the endemic area 17 of Taiwan during 1993 and 1995 (Schoof et al., 1998). It is likely that the arsenic intake in the 18 non-endemic area (background arsenic intake value for reference population) is lower than that 19 reported in the endemic area. 20 EPA also examined the arsenic intake value from food sources in countries where the 21 arsenic exposures are much lower than in Taiwan. The average nonwater inorganic arsenic 22 intake from food consumption is reported to range from 8.3 to 14 ug/day in the United States and 23 from 4.8 to 12.7 ug/day in Canada, with variation across age groups (Yost et al., 1998). Based 24 on the available information, EPA selected 10 ug/day as the best estimate for nonwater arsenic 25 intake (food sources) in baseline calculations. Alternate values of nonwater arsenic intake were 26 also explored in the sensitivity analysis (Section 5.3.8.3). 27 NRC (1999) reported the background arsenic intake of 50 ug/day in endemic areas based 28 on the Schoof et al. (1998) findings. It is not clear if this value was ever used for dose-response 29 modeling in estimating bladder cancer risk. However, NRC (2001) included the background 30 intake of 30 ug/day in the dose-response modeling; the basis for the latter value is not clear. 31 NRC (2001) also reported that there was no difference in the lung and bladder cancer risk 32 estimates when 30 or 50 ug/day were used as the nonwater intake values in the exposed 33 populations. It is not clear if NRC (2001) assumed any nonwater arsenic intake value for the 34 reference populations. In the draft Toxicological Review submitted to SAB in 2005 (U.S. EPA, 35 2005c), nonwater arsenic intake values of 0, 30, and 50 ug/day were assumed for the exposed 36 populations only, and the background inorganic arsenic intake was assumed to be zero for the 123 DRAFT—DO NOT CITE OR QUOTE ------- 1 reference populations. SAB (2007) recommended that the background arsenic intake for 2 reference (control) populations should not be assumed to be zero. However, SAB did not specify 3 a nonwater inorganic arsenic intake value for the reference population. 4 Given the state of the available data and the recommendations from SAB, EPA has 5 assumed 10 ug/day nonwater arsenic intake for the current assessment for both reference and 6 exposed populations in the baseline risk calculations. EPA also evaluated 0, 30, and 50 ug/day 7 for dietary arsenic intake assumption for reference populations, and up to 200 ug/day for 8 exposed populations. The high-end background arsenic intake value was recommended by SAB 9 in 2007 (i.e., the background arsenic intake value in the exposed populations as high as 200 10 ug/day should be included to assess the impact in lung and bladder cancer risk estimates) 11 (Section 5.3.8.3). 12 In the current assessment, the drinking water consumptions for Taiwanese males and 13 females are assumed to be 3.5 L/day and 2.0 L/day, respectively, in the baseline risk 14 calculations. These values are consistent with the assumptions applied by U.S. EPA (1988b), 15 Chen et al. (1992), and NRC (1999 and 2001) for cancer risk estimations. There is conflicting 16 information concerning the extent to which these values include both direct drinking water 17 consumption and water used for cooking. To examine the impact of additional water 18 consumption in cancer risk estimations, NRC (2001) also examined different ratios of water 19 intake-rates between Taiwanese and U.S. populations (up to ratio of 3.0). 20 In the U.S. EPA (1989) report, the arsenic workgroup estimated that the total water 21 consumption for the Taiwanese men, including the water used for cooking rice and yams (the 22 dietary staples in the southwest Taiwanese population), was 4.5 L/day since Taiwanese workers 23 could drink 3.0 to 4.0 L/day of water and the 3.5 L/day seemed to be a reasonable estimate for 24 direct water consumption. Indirect water consumption from cooking rice and yams was 25 estimated to be 1.0 L/day. The basis for the derivation of the drinking water values in the U.S. 26 EPA (1989) report is approximate and gathered from very limited populations (three or four 27 residents were surveyed). In the Arsenic Rule (U.S. EPA, 2001), the total water Taiwanese 28 consumption rates (including water used for cooking) were assumed to be 4.5 L/day for males 29 and 3.5 L/day for females. 30 SAB (2007) did not recommend specific water intake values to be used for cancer risk 31 modeling in the Taiwanese populations. Therefore, in the current assessment, the baseline water 32 intake values modeled are 3.5 L/day for males and 2.0 L/day for females, to be consistent with 33 NRC (1999) recommendations. In addition, a range of water consumption values (up to 5.1 34 L/day in males and 4.1 L/day in females) were evaluated in the sensitivity analysis to study the 35 impact of alternate water consumption in the cancer risk estimates. The water consumption 36 values modeled in the baseline calculations for Taiwanese populations are also close to the 37 average estimates provided for populations in West Bengal, India (Chowdhury et al., 2001), 124 DRAFT—DO NOT CITE OR QUOTE ------- 1 where the climate is close to Taiwan. The average drinking water intake values for children, 2 adult females, and adult males were reported as 2.0, 3.0, and 4.0 L/day, respectively. 3 The drinking water consumption for the U.S. reference population is estimated to be 2.0 4 L/day for both men and women. This is approximately equal to the 90th percentile estimate 5 (2.014 L/day) from the 1994-1996 and 1998 data gathered as part of the Continuing Survey of 6 Food Intake by Individuals (U.S. EPA, 2004), and is consistent with upper percentile estimates 7 from previous surveys. Alternative assumptions about U.S. drinking water consumption result in 8 simple reciprocal adjustments to CSF estimates (discussed further in Section 5.3.8.3). Within 9 the range analyzed, changes in the assumptions about Taiwanese drinking water consumption 10 also result in nearly linear effects on estimated dose-response slope estimates. 5.3.6. Dose-Response Data 11 Table 5-2 summarizes the cancer mortality data from the Morales et al. (2000) study. For 12 this assessment, the original data set containing age-specific PYR, mortality statistics, and 13 village water concentration data was obtained from Dr. Morales (Morales et al., 2000). 14 Water arsenic concentration data were provided for each village. Single concentration 15 measurements were provided for each well. For 20 of the 42 villages only data for one well was 16 reported. However, for the remaining 22 villages, multiple well concentrations were available 17 (range between 2 and 47 measurements) (NRC, 1999). For dose-response estimation, models 18 were fit to the median well concentration for each village. As part of the sensitivity analysis, the 19 reported maximum or minimum well arsenic concentrations were also applied to the models. 125 DRAFT—DO NOT CITE OR QUOTE ------- Table 5-2. Cancer Mortality Data Used in the Arsenic Risk Assessment Gender Male Female Village Water Concentration, jig/L <100 100-299 300-599 >600 Total <100 100-299 300-599 >600 Total Age PYRa Deaths'3 PYR Deaths PYR Deaths PYR Deaths PYR Deaths PYR Deaths PYR Deaths PYR Deaths PYR Deaths PYR Deaths 20-30 35,818 (0, 0, 0) 18,578 (0, 0, 0) 27,556 (0, 3, 0) 16,609 (0, 0, 1) 98,561 (0, 3, 1) 27,901 (0, 0, 0) 13,381 (0, 0, 0) 19,831 (0, 0, 0) 12,988 (0, 0, 0) 74,101 (0, 0, 1) 30-49 34,196 (1, 10, 2) 16,301 (0, 4, 3) 25,544 (5, 7, 9) 15,773 (4, 12, 3) 91,814 (10, 33, 17) 32,471 (3, 1, 5) 15,514 (0, 3, 4) 24,343 (0, 5, 6) 15,540 (0, 4, 6) 87,868 (3, 13,21) 50-69 21,040 (6, 17, 12) 10,223 (7, 15, 14) 15,747 (15,23,30) 8,573 (15, 15,23) 55,583 (43, 70, 79) 21,556 (9, 6, 18) 11,357 (9, 6, 10) 16,881 (19, 6, 20) 9,084 (21,7, 28) 58,878 (58, 25, 76) >70 4,401 (10, 4, 14) 2,166 (2, 4, 13) 3,221 (12, 6, 14) 1,224 (8, 2, 6) 11,012 (32, 16, 47) 5,047 (9, 5, 5) 2,960 (2, 5, 5) 3,848 (11,2, 10) 1,257 (7, 1, 4) 13,112 (29, 13, 24) Total 95,455 (17,31,28) 47,268 (9, 23, 30) 72,068 (32, 39, 53) 42,179 (27, 29, 33) 256,970 (85, 122, 144) 86,975 (21, 12, 29) 43,212 (11, 14, 19) 64,903 (30, 13, 36) 38,869 (28, 12, 38) 233,959 (90, 51, 122) 1 PYR = person-years at risk ' Numbers in parentheses = number of cancer deaths due to bladder, liver, and lung cancer, respectively. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 5.3.7. Risk Assessment Methodology The cancer risk assessment for U.S. population exposure to arsenic in drinking water was conducted in four steps: • Models were fit to the data using mg/kg-day intake metrics calculated from the estimated water consumption values for the Taiwanese population and village water arsenic concentrations, assuming a 10 ug/day nonwater dietary intake in the baseline analysis. Dose-response models were fit to the Morales et al. (2000) data for bladder and lung cancer in both genders using maximum likelihood methods (see Section 5.3.7.1). • Upper confidence limits (UCLs) on the dose coefficients from the fitted models were estimated using the profile likelihood method (see Section 5.3.7.2). values for U.S. populations were calculated for each endpoint and gender based on the dose coefficient UCLs calculated for the Taiwanese populations in the previous step. Using the "BEIR IV" methodology, U.S. bladder and lung cancer incidence data for the 126 DRAFT—DO NOT CITE OR QUOTE ------- 1 years 2000 to 2003 (NCI, 2006) were used as the reference values for calculating U.S. 2 lifetime cancer risks. Thus, the LEDM values are expressed in terms of lifetime cancer 3 incidence for the U.S. population (see Section 5.3.7.3). 4 5 The LEDoi values were used to calculate ingestion drinking water unit risks for lung and 6 bladder cancer for arsenic-exposed men and women in the United States. This step involved 7 linear extrapolation from the LED01 values to zero dose and risk, yielding estimates of low-dose 8 CSFs. Unit risk and CSF calculations were adjusted for differences between body weights and 9 drinking water ingestion rates in Taiwan and the United States. Other risk metrics (estimated 10 lifetime incidence risk per mg/kg-day arsenic intake and corresponding to specific drinking 11 water concentrations) were calculated for each endpoint from the LEDM values (see Section 12 5.3.7.4). 5.3.7.1. Dose-Response Estimation Based on Taiwan Cancer Mortality Data 13 A "Poisson model" was used to fit the cancer mortality data for the Taiwanese 14 population. The general form of the Poisson model is: 15 16 h(x,t) = h0(t) x g(x) (Equation 5-1) 17 18 where: h(x,t) = cancer mortality risk at dose "x" and age "t" 19 h0(t) = cancer mortality risk in the reference population at age "t" 20 g(x) = risk attributable to arsenic exposure at dose "x" (mg/kg-day) 21 22 Taiwanese cancer mortality and PYR data were available for 5-year ranges for ages 20 to 23 84. Cancer mortality data for the southwest Taiwan reference groups also were included in the 24 preferred version of the model; estimates were derived without the reference population and with 25 cancer mortality statistics from all regions of Taiwan. In the Poisson model, which is widely 26 applied in the analysis of epidemiology data, cancer deaths are assumed to be "rare" events and 27 Poisson-distributed within each age-dose group. When hO(t) and/or g(x) are non-linear 28 functions, as is the case for arsenic, the model cannot be fit using conventional least-squares 29 regression methods or general linear models (GLM). Based on recommendations from NRC 30 (2001) and after testing a number of different models, the following model form was selected for 31 primary risk estimates based on goodness-of-fit and parsimony criteria:4 32 33 h(x,t) = exp(ai + a2 x age + as x age2) x (1 + b x dose) (Equation 5-2) 34 35 where: ai, a2, as = age coefficients; b = dose coefficient 36 Results obtained using alternative model forms are discussed in Section 5.3.8.4. 127 DRAFT—DO NOT CITE OR QUOTE ------- 1 Specifically, the model parameters in h(x,t) in Equation 5-2 were obtained by assuming 2 that the number of cases in each exposure-age category has a Poisson distribution with parameter 3 ^(x,t), Cases ~ Poisson (Py x X(x,t)), where Py is person-years, and X is the intensity of Poisson 4 parameter at the exposure-age,(x,t), category. Because data are given in 5-year age intervals, the 5 parameter X is related to hazard rate h which is equal to X/5. 6 In this model, the exponential term represents "hO(t)"in Equation 5-1, the age-dependent 7 risk of cancer at the "background" doses of arsenic (zero from drinking water and 10 ug/day 8 from diet in the preferred model). The last term in the equation captures the dependency of risk 9 on the daily ingestion dose of arsenic. 10 Cancer mortality data were stratified across 13 5-year age groups and 43 villages (42 11 exposed villages plus the reference population). This stratification yielded 559 data points per 12 cancer endpoint for model fitting. Mid-range values for the age ranges were standardized to 13 their mean values and treated as nuisance parameters. 14 The unit of dose used in the modeling was mg/kg-day. In the primary (baseline) risk 15 model, the estimated nonwater arsenic intake was 10 ug/day for both the exposed and reference 16 populations. The total arsenic dose received by the population of any village was estimated as 17 the sum of the nonwater dietary intake plus the median arsenic well water concentration for the 18 village (baseline model), multiplied by the estimated water Taiwanese consumption rates (3.5 19 L/day for men, 2.0 L/day for women) and divided by estimated average body weights for 20 Taiwanese men and women (50 kg for both genders; Chen et al., 1992). The southwest 21 Taiwanese population outside of the arseniasis-endemic area (Morales et al., 2000) served as the 22 reference population in the baseline model. 23 Values for the coefficients al, a2, a3, and b were fit using MLE methods. Likelihood 24 maximization was performed using the Solver add-in of Excel®. The MLE fits for the baseline 25 model were replicated using the Non-Linear Estimation module of Statistica®. Replicated 26 results (estimated age and dose coefficients) were identical to Solver estimates to the third 27 decimal place for all endpoints. 5.3.7.2. Estimation of Confidence Limits on Cancer Slope Parameters 28 The LEDoi values were derived based on estimated upper confidence limits on the 29 estimated dose coefficients ("b") for each endpoint and gender. The confidence limits were 30 calculated using the likelihood profile method (Venson and Moolgavkar, 1988). In this 31 approach, the value of the dose parameter, b, was varied from its estimated mean value. The 32 ratio of the log likelihood for the best-fit model to the log likelihood for other values of "b" is 33 known to follow an approximate chi-squared distribution with one degree of freedom. Thus, the 34 5th and 95th confidence limits on the dose coefficient "b" correspond to the values where the 35 likelihood ratio is equal to 1.92. Upper and lower confidence limits were calculated using 36 Solver®. The fact that the profile likelihood method ignores the likelihood impact of the age 128 DRAFT—DO NOT CITE OR QUOTE ------- 1 "nuisance parameters" implies that the calculated confidence limits are only approximate. 2 Confidence limit calculations using other methods (empirical Bayesian simulations and 3 "bootstrap-t") gave comparable results (within a few percent of the values estimated by profile 4 likelihood). 5.3.7.3. Estimation ofLED01 Values Using Relative Risk Models 5 Once the dose coefficients were calculated, they were used to estimate arsenic-associated 6 lifetime risks in the U.S. population. In this analysis, LEDM values for the U.S. population were 7 calculated using a variant of the "BEIRIV" relative risk model recommended by NRC (2001). 8 The method applied the relative risk estimated as (1 + bUCL x dose) to the age profile of cancer 9 incidence for the reference (U.S. male or female) population, where bUCL is the 95% upper 10 confidence limit on "b" (the arsenic coefficient from the dose-response model for the Taiwanese 11 population, estimated as explained in Section 5.3.7.2). The BEIR IV model also takes into 12 account the effect of noncancer mortality, cancer mortality, and previous cancer incidence on the 13 number of individuals in the exposed population who survive to the start of each 5-year age 14 stratum. To estimate cancer risks in the U.S. population, incidence risks are calculated for each 15 5-year age stratum and summed to give an estimate of lifetime incidence. The dose is then 16 adjusted until the estimated extra incidence risk from arsenic-associated cancer risk equals 0.01 17 (1%) for the U.S. reference population. The dose (in mg/kg-day) that fulfills this condition is the 18 LEDoi, which becomes the point of departure (POD) for estimating the CSF. 19 The BEIR IV model takes as its input age-specific mortality data and lung and bladder 20 cancer incidence for the U.S. reference population.6 U.S. cancer incidence was estimated in this 21 analysis based on mortality data for the year 2000 (NCHS, 2000). Lung and bladder incidence 22 data for the years 2000 to 2003 were obtained from the National Cancer Institute's SEER 23 (surveillance epidemiology and end result) program (NCI, 2006). Arsenic intakes resulting in 24 10"4 lifetime risks were estimated using Solver®. Details of the relative risk methodology are 25 provided in Appendix E.2. 5.3.7.4. Estimation of Unit Risks 26 For each endpoint and gender, the slope of a line from the LED0i dose through the 27 intercept (water-related arsenic dose = 0, water-related arsenic risk = 0) was calculated. The 28 slopes of these lines represent the oral CSF for the endpoint: 29 The empirical Bayes modeling involved taking random samples within the neighborhoods of the MLE coefficient values, calculating the log likelihood, and after many iterations, building up an estimate of the posterior distribution of the "b" coefficient (mean and standard error). Confidence limits were then estimated assuming the posterior probability of b was normally distributed. Note that the age dependence estimated for the Taiwanese population—represented by the parameters a1; a2, and a3—is specific to that population, and is not carried over to the United States. 129 DRAFT—DO NOT CITE OR QUOTE ------- 1 oral CSF (per mg/kg-day) = 0.01/LEDM (Equation 5-3) 2 3 Linear low-dose extrapolation was employed consistent with EPA's finding that 4 insufficient mode of action data are available to justify the use of non-linear, low-dose models 5 (Section 4.6.3.2). Unit risks (cancer risk per ug/L arsenic in drinking water) also were 6 estimated: 7 8 unit risk (per ug/L) = CSF (per mg/kg-day) x 0.001 x DW/BW (Equation 5-4) 9 10 where: 0.001 = conversion from milligrams to micrograms 11 BW = body weight for exposed population in kilograms (U.S. male and female) 12 DW = daily drinking water consumption for exposed population in liters (U.S. male 13 and female) 14 As discussed previously, the estimated drinking water consumption for the U.S. adult 15 population is 2.0 L/day for both males and females. U.S. male and female body weights are 16 estimated to be 70 kg. The 2.0 L/day is a standard factor used in EPA IRIS assessments, and 17 represents approximately the 90th percentile of intake of community water in the U.S. 18 population. Other intake assumptions (e.g., mean versus upper percentile) can be used in risk 19 assessments, depending on target population characteristics and assessment needs. 5.3.8. Results 5.3.8.1. Ingestion Pathway Oral CSFs and Unit Risks 20 Table 5-3 presents the estimated risk metrics for lung and bladder cancers in males and 21 females under baseline assumptions (see Footnote "a" to the table for baseline modeling 22 assumptions). 23 The estimated oral CSF for female lung cancer (16.6 per mg/kg-day) is higher than that 24 for males (6.7 per mg/kg-day), but the bladder cancer oral CSFs for males and females are 25 comparable (11.2 and 10.5 per mg/kg-day, respectively). Drinking water unit risks for lung 26 cancer are 1.9 x 10"4 and 4.8 x 10"4 per ug/L, respectively, for males and females while the 27 drinking water unit risks for bladder cancer are 3.2 x 10"4 and 3.0 x 10"4 per ug/L, respectively. 28 Estimated lifetime incidence risks corresponding to 10 ug/L arsenic in drinking water follow 29 similar patterns for the various endpoints. Estimated drinking water concentrations associated 30 with 10"4 lifetime incidence range from 0.21 ug/L (female lung cancer) to 0.52 ug/L (male lung 31 cancer). 130 DRAFT—DO NOT CITE OR QUOTE ------- Table 5-3. Cancer Incidence Risk Estimates for Lung and Bladder Cancers in Males and Females3 Metric Lung Cancer Bladder Cancer Males ED0i, mg/kg-day LEDoi, mg/kg-day Oral CSF, per mg/kg-day Unit risk, per ug/L drinking water Lifetime incidence risk at 10 ug/L in drinking water Water concentration for 10~4 risk, ug/L 1.9E-03 1.5E-03 6.7 1.9E-04 1.9E-03 0.52 1.1E-03 8.9E-04 11.2 3.2E-04 3.2E-03 0.31 Females ED0i, mg/kg-day LEDoi, mg/kg-day Oral CSF, per mg/kg-day Unit risk, per ug/L drinking water Lifetime incidence risk at 10 ug/L in drinking water Water concentration for 10"4 risk, ug/L 7.5E-04 6.0E-04 16.6 4.8E-04 4.8E-03 0.21 1.2E-03 9.5E-04 10.5 3.0E-04 3.0E-03 0.33 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 aBaseline assumptions: reference population = southwest Taiwan; Taiwanese male and female body weight = 50 kg, Taiwanese male water intake = 3.5 L/day, Taiwanese female water intake = 2.0 L/day; reference and exposed population nonwater arsenic intake = 10 ug/day. Male and female U.S. body weights are assumed to be 70 kg, and U.S. water intake for both males and females is assumed to be 2.0 L/day. Arsenic-related cancer risks also are calculated for the population as a whole, that is, for combined bladder and lung cancer incidence in a population composed of both men and women. In this analysis, total cancer risk (lung plus bladder) for males and females is calculated by combining the risk for the individual tumor types. Upper confidence limits on the combined cancer risks can be calculated based in the assumption that the uncertainties in the two CSFs are both normally distributed. If this is the case, the 95% upper bound, U, for the combined cancer potency can be calculated as: U = +(u2 -m2)2 (Equation 5-5) where mi and ui, i = 1,2, are respectively mean and 95% upper bound cancer potency for the two tumor types. The results of these calculations are summarized in Table 5-4. Using this approach, the combined cancer potency factor estimate for males is 16.9 per mg/kg-day for males and 25.7 per mg/kg-day for females. The estimated drinking water unit risk for combined male lung and bladder cancer is 4.8 x 10"4 per ug/L; for females, the estimated value is 7.3 x 10"4 per ug/L. The drinking water concentrations corresponding to 10-4 combined cancer risks for males and females are 0.21 and 0.14 ug/L, respectively. 131 DRAFT—DO NOT CITE OR QUOTE ------- Table 5-4. Combined Lung and Bladder Cancer Incidence Risk Estimate for the U.S. Population (Males and Females) Metric Oral CSF, per mg/kg-day Unit risk, per ug/L drinking water Lifetime incidence risk at 10 ug/L in drinking water Water concentration for 10~4 risk, Mg/L Male Combined Lung+Bladder 16.9 4.8E-04 4.8E-03 0.21 Female Combined Lung+Bladder 25.7 7.3E-04 7.3E-03 0.14 1 Figure 5-1 shows the estimated oral CSFs for each of the endpoints separately, along 2 with oral CSF estimates for the combined cancers in males and females. In keeping with EPA 3 policy, the combined oral CSF for women (25.7 per mg/kg-day) is appropriate for use in 4 establishing health criteria, since, based on the available data, women appear to be the 5 more sensitive group. OJ CJ C « U Male Lung Male Bladder Male Female Female Female Combined Lung Bladder Combined Figure 5-1. Estimated oral CSFs for individual and combined cancer endpoints. 5.3.8.2. Comparison to Previous Cancer Risk Estimates 6 As discussed in Section 5.3.1, a number of risk assessments have been conducted by EPA 7 and others. Results of the present dose-response assessment were compared to cancer risk 8 estimates derived from the same and other data sets in previous studies (NRC, 2001; U.S. EPA, 132 DRAFT—DO NOT CITE OR QUOTE ------- 1 2005c). Note that the results of the U.S. EPA (1988b) analysis, which estimated a CSF of 1.0- 2 2.0 per mg/kg-day, are not comparable to the results of the current assessment (CSF 25.7 per 3 mg/kg-day), because the former was based on skin cancer, while all of the more recent analyses 4 estimate risks of internal (lung and bladder) cancers. Thus, the detailed comparisons in this 5 section are limited to assessments that also address lung and bladder cancer. The drinking water 6 standard (U.S. EPA, 2001) also provides numerical risk estimates for exposures to arsenic in 7 drinking water. However, Tables III.D-2(a) and (b) of the rule (U.S. EPA, 2001) display ranges 8 of cancer risks for populations exposed to distributions of arsenic concentrations in drinking 9 water at and above the proposed MCL options. Thus, the numerical risk results of that analysis 10 are also not directly comparable to the NRC (2001), U.S. EPA (2005c), and current assessments, 11 which apply to populations exposed to single concentrations. In the analyses that follow, some 12 of the risk comparisons are based on mortality estimates that have been converted to incidence 13 using recent U.S. incidence-mortality ratios. This conversion introduces additional uncertainty 14 into the comparisons; different results would have been obtained had the incidence been 15 modeled directly rather than estimated after the fact. 5.3.8.3. EDoi andLED01 Estimates From Chen et aL (1988a, 1992), Ferreccio et al (2000), and Chiou et al. (2001) 16 Consistent with SAB (2007) recommendations, Table 5-5 presents risk estimates from 17 previous studies and compares them to estimates derived in this analysis. The estimates in Table 18 5-5 come from Table 5-3 of NRC (2001), and include EDOI and LEDM estimates (expressed as 19 ug/L arsenic in drinking water) from a number of studies of arsenic-related cancer risks in Chile 20 (Ferreccio et al., 2000) and Taiwan (Chiou et al., 2001; Chen et al., 1988a, 1992). 21 NRC calculated EDOI and LEDM values for lung and bladder cancer mortality from the 22 same Taiwanese cohort used in the current assessment, based on the results presented in Chen et 23 al. (1988a, 1992), but without a reference population. In addition, these values do not account 24 for differences in drinking water consumption between the U.S. and Taiwanese populations, and 25 did not apply life-table adjustments. 133 DRAFT—DO NOT CITE OR QUOTE ------- Table 5-5. Comparison of ED0i and LED,/ Estimates From Past Studies'1 With Those From the Current Analysis Study Chenetal. (1988a, 1992), Taiwan Ferreccio et al. (2000), Chile Chiouetal. (2001), Taiwan Current analysis Male Lung EDoi 38-84 5-17 — 66 LED01 37-72 3-14 — 52 Female Lung EDoi 33- 94 7-27 — 26 LED01 31-84 5-21 — 21 Male Bladder EDoi 102-317 — 129-500+ 40 LED01 94-286 — 42- 500+ 31 Female Bladder EDoi 138-443 — 231-500+ 41 LED01 125-406 — 88-500+ 33 a Units = ug/L arsenic in drinking water b Source of estimates: NRC (2001) 1 NRC also estimated EDOI and LEDM values based on data from the Ferreccio et al. (2000) 2 case-control study of male and female lung cancer data from a Chilean population that included 3 151 lung cancer cases and 419 controls. The EDOI and LED01 derived by NRC were obtained by 4 linear regression of mortality odds ratio estimates on exposures, with the intercept forced to a 5 value of 1.0 at zero exposure. These estimates are shown in the second row of Table 5-5. 6 Multiplicative linear dose and log dose models were used to derive EDOI and LED0i estimates 7 from the study by Chiou et al. (2001) of urinary tract cancer incidence over a 4-year period in 8 8,000 Taiwanese exposed to arsenic in drinking water. These results are presented in the third 9 row of Table 5-5. Where ranges are given in the table, the minimum and maximum values 10 represent the lowest and highest EDOI or LEDM estimates that were derived when different 11 models were used. 12 The bottom row of the table shows the EDOI and LEDM values for cancer incidence 13 derived in this analysis using the Poisson regression and BEIRIV models. The EDOI and LEDM 14 values for lung cancer derived in the current assessment fall within, or are close to, the ranges 15 estimated from the Chen et al. (1988a, 1992) data. This finding is not surprising because the 16 results are estimated for the same cohort in both cases, and because the case mortality for lung 17 cancer is so high (nearly 100%). The EDOI and LED0i values derived in the current assessment 18 are, however, higher than those estimated by Ferreccio et al. (2000). One possible explanation 19 involves differences in modeling methods; to estimate EDOI and LED0i values from the Ferreccio 20 study, NRC applied linear regression to the odds ratio estimates, forcing the intercept through 21 1.0 at zero dose. Thus, these values must be considered highly uncertain. The differences also 22 may be due to differences in exposure conditions (e.g., NRC did not account for differences in 23 drinking water intake between the Chilean and U.S. populations) or other covariates (e.g., 24 smoking) between the two studies. 134 DRAFT—DO NOT CITE OR QUOTE ------- 1 For bladder cancer, the ED01 and LEDM values estimated in this analysis are lower (2.5- 2 to 10-fold) than those derived from the Chen et al. studies (1988a, 1992). In addition to the 3 differences in modeling approaches outlined above, another possible reason for this difference is 4 that the Chen et al. (1988a, 1992) studies are based on bladder cancer mortality, while the ED01 5 and LEDM values in this analysis are for bladder cancer incidence. Adjustment for bladder cancer 6 case mortality (in the order of 16-20%) would make EPA's current results much more similar to 7 those of Chen et al. (1988a, 1992). 8 Finally, the ED01 and LEDOT values from the current analysis are below the lower end of 9 the ranges estimated by Chiou et al. (2001). Reasons for this finding are not entirely clear. The 10 sensitivity of the Chiou et al. study may have been limited by the short follow-up period (NRC, 11 2001), and only 18 total urinary tract cancers were identified in the study. Only four exposure 12 categories were analyzed (less than 10 ug/L, 10-50, 50-100, and more than 100 ug/L in water; 13 nonwater exposures were not evaluated). The low sensitivity could have caused the ED01 and 14 LEDoi estimates derived by Chiou et al. (2001) to be biased upward from what would have been 15 seen with a more extended follow-up period. 5.3.8.4. Estimated Risk Associated With 10 fig/L Drinking Water Arsenic From NRC (2001) 16 Table 5-6 provides an additional set of comparisons between the current risk estimates 17 and the results from a previous analysis by NRC (2001). Lifetime incidence risks are presented 18 for a hypothetical U.S. population exposed to 10 ug/L arsenic in drinking water. NRC (2001) 19 estimated arsenic-associated risks using an "additive Poisson model with dose entered as a linear 20 term and using the BEIRIV formula" (p. 201). Table 5-6. Comparison of cancer risk assessment results with estimates from NRC (2001) Source of Estimate NRC (2001), Taiwan Current analysis Estimated Cancer Incidence at 10 ug/L Arsenic in Drinking Water (per 10,000 Exposed Population) Bladder Male 23 32 Female 12 30 Lung Male 14 19 Female 18 48 a The original mortality risk estimates from U.S. EPA (2005c) were multiplied by incidence- mortality ratios for the various endpoints to obtain incidence estimates. For the Taiwanese populations, case mortality for lung cancer was assumed to be 100% and mortality for bladder cancer was assumed to be 80% (NRC, 2001). 21 The incidence risks derived in the current analysis, however, are reasonably close, but not 22 identical, to the NRC (2001) estimates. Differences in the calculated cancer potency relate to 23 several factors. Changes in the assumed drinking water intake in females in the current 13 5 DRAFT—DO NOT CITE OR QUOTE ------- 1 assessment compared to the NRC (2001) and U.S. EPA (2005c) analyses are summarized in 2 Table 5-7. In particular, the change in the assumed ratios of Taiwanese/U.S. female water intake 3 from 2.8 in the earlier assessments to 1.4 in the current analysis are relevant to the differences in 4 risk shown in Table 5-6. The lower ratio in the current analysis translates into a slightly greater 5 than 2-fold greater estimated risk for females in the current assessment than in the NRC (2001) 6 and current analyses. Table 5-7. Drinking water intake and body weight assumptions in females in recent arsenic risk assessments Assessment NRC (2001) U.S. EPA (2005c) Current analysis Body Weight, kg Taiwan 50 50 50 U.S. 70 70 70 Water Intake, L/day Taiwan 2 2 2 U.S. 1 1 2 Ratio of Taiwan/U.S. Drinking Water Intake 2.8 2.8 1.4 7 In addition, the NRC (2001) risk estimates are based on maximum likelihood estimates 8 (MLE) of the arsenic slope parameters in the Poisson regression, while U.S. EPA (2005c) and 9 the current assessment derive risks based on the statistical upper confidence bounds on these 10 parameters. As shown in Table 5-3, the difference between the MLE estimates (ED01 values) 11 compared to the upper confidence limit (LEDM) is on the order of 20%. This would translate into 12 approximately 20% greater risks calculated based on the upper confidence limit values compared 13 to the MLE estimates. 14 The use of more recent cancer incidence and mortality data in the BEIRIV model than 15 in the previous risk assessments also probably contributes to the differences in risks in Table 5-6. 16 Also, the current assessment includes a modification to the BEIR IV model suggested by Gail et 17 al. (1999) for obtaining more accurate estimates of incidence within multi-year age strata. The 18 modifications to the model are described in detail in Appendix E.2. 19 Changes in the assumptions related to nonwater arsenic intake also would be expected to 20 have small to moderate effects on the results within the range in question. In this assessment, 21 both the reference and exposed populations are assumed to receive 10 ug/day nonwater arsenic 22 intake (see Section 5.3.5). Section 5.3.8.3 presents the results of uncertainty analyses that 23 explore the effects of changes in selected modeling assumptions, including nonwater arsenic 24 intake, on the risk estimates. 25 The cancer risk estimates presented in Table 5-8 for consumption of drinking water with 26 specified arsenic concentrations provide information that is scientifically equivalent to estimates 27 of CSFs. The NRC's (200l)recommended risk models provide estimates that consumption of 28 drinking water containing 10 ug/L arsenic is associated with the site specific cancer risks below. 136 DRAFT—DO NOT CITE OR QUOTE ------- 1 Note that the same CSF values, other than small differences due to rounding error, would be 2 obtained starting with any of the water concentrations presented in the NRC (2001) Table S-l. 3 4 5 6 7 8 9 10 11 12 Table 5-8. Theoretical maximum likelihood estimates of excess lifetime risk (incidence per 10,000 people) of lung cancer and bladder cancer for US populations Arsenic concentration (HS/L) 10 Bladder Male 23 Female 12 Lung Male 14 Female 18 The equivalent CSFs can be calculated as follows: Using the exposure factors for US populations applied in NRC (2001), consumption of 10 ug/L arsenic in drinking water results in a daily exposure of (10 ug/L) x (1 L/d) x (1 mg/1,000 ug) x (1/70 kg) = 0.000143 mg/kg-d of inorganic arsenic. As the NRC risk estimates are linear (proportional to dose) for these exposures, equivalent CSF values come from the equation: Risk = CSF (per mg/kg-d) x dose (mg/kg-d) As an example, applying this equation to bladder cancers in females: 12 x 10'4 = CSF x 0.000143 mg/kg-d, or CSF = 8.4 per mg/kg-d Thus the CSF estimates resulting from Table 5-8 are shown below in Table 5-9. Table 5-9. Arsenic oral CSFs (per mg/kg-d) for lung cancer and bladder cancer in US populations Bladder Male 16 Female 8 Lung Male 10 Female 13 13 As these are maximum likelihood estimates, it is appropriate to add risks across the two 14 sites resulting in combined CSFs for lung and bladder cancer of 21 and 26 per mg/kg-d in 15 females and males respectively. 5.3.8.5. Sensitivity Analyses of Cancer Risk Estimates to Changes in Parameter Values 16 NRC (2001) and SAB (2007) recommended that the impacts of different modeling 17 assumptions and input parameter values be investigated in the risk assessment for arsenic in 18 drinking water. EPA, therefore, examined several aspects of the cancer risk modeling through 19 single-value sensitivity analysis. The Agency felt that the currently available data were 20 insufficient to support detailed probabilistic uncertainty and variability estimation. In response 21 to SAB comments, EPA evaluated the impacts of: 137 DRAFT—DO NOT CITE OR QUOTE ------- 1 • Varying the assumed daily nonwater arsenic intake of the exposed and reference 2 populations. Sensitivity cases were run in which the nonwater arsenic intake in the 3 exposed populations was varied from its default value of 10 ug/day to 0, 100, and 200 4 ug/day. An additional case was run in which both the exposed and reference populations 5 were assumed to receive 0, 30, and 50 ug/day nonwater arsenic exposure. Because the 6 Poisson risk model for female bladder cancer is particularly sensitive to changes in 7 assumptions related to nonwater arsenic intakes (see below), nonwater arsenic intake was 8 limited to below 50 ug/day in reference populations. 9 10 • Varying assumptions related to drinking water intake by the exposed Taiwanese 11 population. Cases were run in which the male drinking water consumption was varied 12 from its baseline value of 3.5 L/day to 5.1 L/day, 3.0 L/day, and 2.75 L/day. Female 13 drinking water intake in the Taiwanese population was varied from its baseline value of 14 2.0 L/day to 2.75 and 4.1 L/day. 15 16 • Varying the arsenic well concentrations used to fit the dose-response model for the 17 Taiwanese population. The baseline risk model used the median village arsenic 18 concentrations as the exposure metric. In the sensitivity analysis, cases also were run 19 using the minimum and maximum well concentrations in each village. 20 21 • Including different Taiwanese reference populations in the dose-response assessment. 22 The baseline (southwest Taiwan) reference population was replaced by data from all 23 Taiwan. The model also was run without any distinct reference population. 24 25 Tables 5-10 and 5-11 summarize the results of the sensitivity analysis runs. Table 5-10 26 shows the estimated (incidence) risks associated with a drinking water concentration of 10 ug/L 27 for the U.S. population estimated when calculated using the assumptions specified in the left- 28 hand column of the table. Table 5-11 shows the proportional changes in estimated risks in 29 relations to the baseline estimate. Figure 5-2 summarizes the impact of alternative modeling 30 assumptions, showing the ratios of estimated cancer risks to the base case estimates for changes 31 in input variables having a substantial (>20%) effect on the risk estimates. 13 8 DRAFT—DO NOT CITE OR QUOTE ------- Table 5-10. Sensitivity analysis of estimated cancer incidence risks associated with 10 ug/L to changes in modeling assumptions and inputs Estimated Cancer Risk at 10 |j,g/L Baseline (all default values)3 Nonwater arsenic intake = 0 ug/day (reference and exposed populations) Nonwater arsenic intake = 30 ug/day (reference and exposed populations) Nonwater arsenic intake = 50 ug/day (reference and exposed populations) Nonwater arsenic intake (exposed population) = 0 ug/day Nonwater arsenic intake (exposed population) = 100 ug/day Nonwater arsenic intake (exposed population) = 200 ug/day Taiwan water consumption =3.0 L/day (M), 2.0 L/day (F) Taiwan water consumption =5.1 L/day (M), 4.1 L/day (F) Taiwan water consumption = 2.75 L/day (M, F) Village water arsenic concentrations = minimum values Village water arsenic concentrations = maximum values Reference population = none Reference population = all Taiwan Male Lung 1.9E-03 1.9E-03 2.0E-03 2.0E-03 1.9E-03 1.8E-03 1.7E-03 2.3E-03 1.3E-03 2.5E-03 2.5E-03 1.4E-03 1.2E-03 2.4E-03 Female Lung 4.8E-03 4.6E-03 5.1E-03 5.5E-03 4.8E-03 4.4E-03 3.9E-03 4.8E-03 2.3E-03 3.4E-03 5.7E-03 3.5E-03 1.5E-03 3.9E-03 Male Bladder 3.2E-03 3.0E-03 3.5E-03 3.9E-03 3.2E-03 3.0E-03 2.8E-03 3.8E-03 2.2E-03 4.1E-03 4.0E-03 2.3E-03 8.3E-04 4.8E-03 Female Bladder 3.0E-03 2.6E-03 4.5E-03 1.1E-02 3.0E-03 2.8E-03 2.4E-03 3.0E-03 1.4E-03 2.1E-03 4.0E-03 2.1E-03 3.5E-04 6.2E-03 aBaseline inputs: reference population = southwest Taiwan; male and female body weight = 50 kg, male water intake = 3.5 L/day, female water intake = 2.0 L/day, reference and exposed population nonwater arsenic intake = 10 |ag/day. U.S. population male and female body weights = 70 kg, male and female water consumption = 2.0 L/day. 139 DRAFT—DO NOT CITE OR QUOTE ------- Table 5-11. Proportional Changes in Cancer Risks at 10 ug/L Associated With Changes in Modeling Inputs and Assumptions Modeling Assumptions/Input Values Baseline (all default values)3 Nonwater arsenic intake = 0 ug/day (reference and exposed populations) Nonwater arsenic intake = 30 ug/day (reference and exposed populations) Nonwater arsenic intake = 50 ug/day (reference and exposed populations) Nonwater arsenic intake (exposed population) = 0 ug/day Nonwater arsenic intake (exposed population) = 100 ug/day Nonwater arsenic intake (exposed population) = 200 ug/day Taiwan water consumption =3.0 L/day (M), 2.0 L/day (F) Taiwan water consumption =5.1 L/day (M), 4. 1 L/day (F) Taiwan water consumption = 2.75 L/day (M, F) Village water arsenic concentrations = minimum values Village water arsenic concentrations = maximum values Reference population = none Reference population = all Taiwan Male Lung 0% 0% 5% 5% 0% -5% -11% 21% -32% 32% 32% -26% -37% 26% Female Lung 0% -4% 6% 15% 0% -8% -19% 0% -52% -29% 19% -27% -69% -19% Male Bladder 0% -6% 9% 22% 0% -6% -13% 19% -31% 28% 25% -28% -74% 50% Female Bladder 0% -13% 50% 267% 0% -7% -20% 0% -53% -30% 33% -30% -88% 107% a Baseline inputs as described in footnote to Table 5-8. 4.0 -i Q) ^ n .1 0) ro ? n CQ ^u o 0) ^ -in _> 1 .U "ro 0) °^ 0.0 ^^ 1 | | (/) c JS in \lon-water A - 55 0) (0 c - M § 1 1-1 - |f ru~u n M " •— > i >•» ^* c -§ "o 1 — - -i ^ >s ro ^^ R i- CD -- i- in - i- ,-g rv n n in llage H20 A = minimum > DMale Lung • Female Lung D Male Bladder D Female Bladder (/) < $ S 3 ^C (0 (D > = maximum rt (D C ° "5 Q. IV O Q. — 0) Referenc - 1 TO C Q- ro Figure 5-2. Change in arsenic-related unit risk estimates associated with variations in input assumptions. 140 DRAFT—DO NOT CITE OR QUOTE ------- 1 These results indicate that varying most of the risk modeling inputs within the tested 2 ranges have a small or moderate effect on risk estimates for most endpoints. For all of the 3 endpoints except female bladder cancer, changing assumptions related to nonwater arsenic intake 4 for the reference and/or exposed populations results in small changes (<25%) in the estimated 5 oral CSF and cancer risks at 10 ug/L in drinking water. Risk estimates for female bladder 6 cancer, in contrast, are quite sensitive to changes in nonwater arsenic intake in the range from 0 7 to 50 ug/day. When nonwater arsenic intake is assumed to be 30 ug/day (rather than 10 ug/day 8 in the baseline estimate), estimated female bladder cancer risks are approximately 50% higher 9 than under baseline assumptions. When nonwater arsenic intake increases to 50 ug/day, female 10 bladder cancer risk increases by 267% compared to baseline. The sensitivity of the risk 11 estimates is greater for changes in reference population arsenic intake; when nonwater intake 12 increases to 100 and 200 ug/day for the exposed populations alone, the impacts on female 13 bladder cancer risks are much less (7% and 20%, respectively). 14 As expected, the risk estimates obtained when making different assumptions concerning 15 Taiwanese drinking water consumption are very nearly inversely proportional to the assumed 16 water intake. For example, when male drinking water consumption is assumed to be 5.1 L/day, 17 rather than 3.5 L/day in the baseline case, estimated cancer risks for male lung and bladder 18 cancer are both approximately 0.69 (= 3.5/5.1) times the values derived using baseline 19 assumptions. Similar results are seen for the other endpoints. 20 Using different exposure concentration metrics also shows relatively limited impacts on 21 the estimated cancer risks. When the village minimum water concentrations are used as inputs to 22 the Poisson risk model, the estimated cancer risks increase slightly (32%, 19%, 25%, and 33% 23 over baseline) for male and female lung and male and female bladder cancer, respectively. 24 When village maximum water concentrations are used as model inputs, the estimated cancer 25 incidence risks decrease between 26 and 30% relative to baseline. These changes are roughly 26 reciprocal to the changes in average exposure concentrations, as expected. 27 The final two rows of Tables 5-8 and 5-9 illustrate the impact of alternative assumptions 28 about which reference populations are included in the Taiwanese risk assessment model. When 29 no reference population is included (the Poisson model is fit only to the data from the 42 30 exposed villages), the estimated risks for all four endpoints are considerably lower than under 31 the baseline case, which included the southwest Taiwan population. This finding is not 32 unexpected, because the addition of the relatively large reference population serves to "anchor" 33 the low-exposure end of the model and decrease the impact of the high variability ("noise") in 34 the exposed population data. When the reference population is excluded from the assessment, 35 estimated cancer risks are reduced between 37% (male lung) and 88% (female bladder cancer) 36 compared to the baseline model that included the southwest Taiwan reference populations. All 37 of the exposure-response "b" parameters retain statistical significance, however, even when the 141 DRAFT—DO NOT CITE OR QUOTE ------- 1 reference population is excluded. Finally, including the "all Taiwan" reference population, 2 rather than southwest Taiwan, has smaller and variable effects on the risk estimates. Predicted 3 risks for male lung and bladder cancer are increased (decreased) by approximately 26% and 4 19%, respectively, while risks for female lung and bladder cancer are increased by 50% and 5 107%, respectively, compared to baseline. 6 Based on these outcomes, it appears that the risk model results are relatively stable and 7 react predictably to reasonable changes in exposure assumptions. The exception is female 8 bladder cancer, for which the dose-response parameter estimated in the Poisson model is very 9 sensitive to the assumed nonwater arsenic intake by the reference population in the range 10 between 0 and 50 ug/day. In addition, risk estimates for all endpoints are strongly affected by 11 the inclusion or exclusion of a low-dose reference population in the Poisson risk model. 5.3.8.6. Sensitivity Analyses of Cancer Risk Estimates to Dose-Response Model Form 12 In the course of this analysis, EPA has investigated the impact of alternative model forms 13 on the cancer risks estimated for the Taiwanese and U.S. populations for individual endpoints 14 (lung and bladder cancer). Based on the past experience of Morales et al. (2000) and modeling 15 results presented by NRC (2001), this effort was limited to exploring alternative forms for the 16 dose dependence of risks. Equation 5-5 shows EPA's baseline model, which is "linear Poisson" 17 with the form: 18 19 h(x,t) = exp(ai + a2 x age + a3 x age2) x (l + b x dose) (Equation 5-5) 20 21 In addition to the linear model, three other models were evaluated. First, the quadratic form of 22 dose dependence: 23 24 h(x,t) = exp(ai + a2 x age + as x age2) x (l + bl x dose +b2 x dose2) (Equation 5-6) 25 26 Next, two models in which the dose dependence was exponential, one linear and one quadratic: 27 28 h(x,t) = exp(ai + a2 x age + as x age2) x Exp(bO + bi x dose) (Equation 5-7) 29 30 h(x,t) = exp(ai + a2 x age + a3 x age2) x Exp(bO + bi x dose + b2 x dose2) (Equation 5-8) 31 32 The last model (Equation 5-8) was specifically recommended by SAB (2007) for 33 evaluation. In the discussion that follows, these four models are referred to, respectively, as the 34 "linear" (baseline) model (Equation 5-5), quadratic model (Equation 5-6), linear exponential 35 model (Equation 5-7), and quadratic exponential model (Equation 5-8).7 "Absolute risk" models (models in which arsenic exposure was assumed to result in additive, rather than multiplicative, increments in risks) were found to fit the data much less well than the multiplicative forms shown in Equations 5-6 to 5-8 and are not discussed further. 142 DRAFT—DO NOT CITE OR QUOTE ------- 1 All four models were fit to lung cancer data from the Taiwanese population, using the 2 baseline exposure parameter values and including the southwest Taiwanese reference population. 3 Models were fit using the Non-Linear Estimation module of Statistica®. For males, the 4 quadratic and quadratic exponential models curve sharply downward at high doses, whereas the 5 linear exponential model curves sharply upward. Over the dose range from 0 to 0.05 mg/kg-day 6 in males, which corresponds to an arsenic drinking water concentration range of 0 to 710 ug/L 7 (which covers approximately 95% of the exposed population years at risk), predictions from the 8 non-linear models are never more than 22% higher or 24% lower than the predictions from the 9 linear (baseline) model. As noted previously, these differences are relatively small compared to 10 the degree of statistical uncertainty in the estimates of the dose-response coefficients. 11 For females, two of the models (quadratic and quadratic exponential) predict lung cancer 12 risks for 60- to 65-year-olds that are very close to those predicted by the linear model. The 13 linear exponential model, however, curves strongly upward at high doses. Over the dose range 14 from 0 to 0.03 mg/kg-day in females (corresponding to 0 to 750 ug/L arsenic in drinking water, 15 about 95% of the exposed population years at risk), the cancer risks predicted by the non-linear 16 models are never more than 9% above or 37% below the risks predicted by the linear (baseline) 17 model. 18 These analyses indicate that, within the range of exposures covered by the 19 epidemiological data, the alternative model forms predict very similar risks (i.e., variations in 20 risk estimates across models are well within the estimated statistical uncertainty of the models). 21 The behavior to the various models at the extremes of the data (high and low exposures) depends 22 to a large extent on the model specification; models with non-linear dose specifications will 23 predict risks that increase more or less rapidly in the extremes than the linear additive Poisson 24 regression, depending on the form of the dose term. As discussed in Section 4.6.3, given the 25 limitations in data related to mode of action, there is no compelling reason to prefer non-linear 26 models, and the additive Poisson model is the simplest, best-fitting, and most parsimonious 27 model currently available for establishing a point of departure for establishing health criteria. 5.3.8.7. Significance of Cancer Risks at Low Arsenic Exposures 28 Several recently published studies have called into question the strength and significance 29 of the exposure-response relationship for arsenic in the Taiwanese population studied by Chen et 30 al. (1988a, 1992) and Wu et al. (1989) that have been used by EPA for estimating cancer risk. 31 Based on "graphical and regression analysis," Lamm et al. (2003) found no significant dose- 32 response relationship for arsenic-related bladder cancer in the subset of the Taiwanese 33 population with median drinking water well concentrations less than 400 ug/L. Kayajanian 34 (2003) found that combined male and female lung, bladder, and liver cancers were relatively 35 elevated at low arsenic exposures, then decreased to minimums for villages with water arsenic 36 concentrations in the range between 42 and 60 ug/L, and then again increased with increasing 143 DRAFT—DO NOT CITE OR QUOTE ------- 1 arsenic exposure. In a more recent analysis, Lamm et al. (2006) found that (1) dummy variables 2 related to "township" location were significant (along with arsenic well concentration) when all 3 the townships were included in the analysis and (2) the dose-response parameter for arsenic 4 exposure became insignificant for arsenic well concentrations less than 151 ug/L when only a 5 subset of the data was included in the regression. 6 The studies by Lamm et al. (2003, 2006) and Kayajanian (2003) have severe limitations. 7 In evaluating the findings of these studies, it is important to recognize the complexity and 8 limitations of the Taiwanese data set. Cancer mortality and person-years at risk observations are 9 provided for a large number (n = 559) of relatively small age- and village-stratified populations 10 (median person-years at risk ~ 340 for both males and females). Most population groups have 11 zero cancer deaths, and the data are very "noisy." Cancer mortality is strongly age-dependent, 12 and simultaneously evaluating the age- and dose-dependence of cancer mortality based on a data 13 set in which cancer deaths are "rare events" requires appropriately structured models. All of 14 these features of the data drove the selection of the Poisson regression methods described in 15 Section 5, and the use of simpler models (linear regression, for example) can (and did) produce 16 misleading results. 17 With regard to the Lamm et al. (2003) paper, it is likely that the use of linear regression 18 and the failure to correctly account for the age-dependency of bladder cancer risks combined to 19 make it impossible to detect a significant exposure-response relationship in villages with water 20 arsenic levels less than 400 ug/L. U.S. EPA (2005d) evaluated this study and noted the 21 following weaknesses: 22 • Classification of wells as artesian or shallow was based solely on arsenic concentration. 23 24 • Age was not included as a variable in the regression analysis, despite the clear strong 25 dependence of cancer risks on age. 26 27 • Previous studies have found little evidence for the presence of other potential carcinogens 28 in the sampled wells. 29 30 The major limitation of Kayanjaian's (2003) analysis of the Taiwanese data is that it 31 breaks the data into strata that are too small to be used to calculate reliable mortality risks, and 32 that it is very sensitive to the specific way that the data are stratified. The observed trend in 33 cancer mortality versus arsenic dose would be very different if only few cancer deaths were 34 misclassified, or if the pattern of cancer deaths had been slightly different by chance. Lamm et 35 al.'s (2006) failure to find a significant exposure-response relationship in villages with arsenic 36 water concentrations below 151 ug/L can also be explained by (1) the use of linear regression 37 without age-adjustment; and (2) the omission of data from three of the six townships from the 38 regression. 144 DRAFT—DO NOT CITE OR QUOTE ------- 1 Appendix F provides additional analyses supporting the significance and robustness of 2 the dose-response relationship for arsenic at low doses and in the defined subsets of the 3 population studied by Lamm et al. (2006). 5.4. CANCER ASSESSMENT (INHALATION EXPOSURE) 4 An inhalation unit risk was developed for inorganic arsenic and posted on the IRIS 5 database in 1988. This document does not present a re-assessment of the cancer dose-response 6 estimation for inhalation exposure to inorganic arsenic. 145 DRAFT—DO NOT CITE OR QUOTE ------- 6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE- RESPONSE 6.1. HUMAN HAZARD POTENTIAL 1 Arsenic is readily absorbed from the GI tract, either from drinking water or food sources. 2 Although dermal absorption is not significant compared to absorption from oral exposure, it 3 may have contributed to the total arsenic exposures and health effects reported in many 4 epidemiological studies in the literature. There appears, however, to be little if any dermal 5 absorption (NRC, 1999) except at high occupational exposures (Hostynek et al., 1993). 6 Inhalation is not being addressed in this document. 7 After absorption, inorganic arsenic can undergo a complicated series of enzymatic and 8 non-enzymatic reduction, enzymatic oxidative methylation, and conjugation reactions. Although 9 these reactions occur throughout the body, the rate at which they occur varies greatly from organ 10 to organ, with major metabolism occurring in the liver. While there are two proposed pathways 11 (Figures 3-1 and 3-2) for arsenic metabolism—with each pathway likely to occur depending on 12 exposure level and/or individual—the main urinary excretion products in humans are MMA and 13 DMA and the parent compound. Arsenic metabolism (mainly methylation) varies greatly across 14 different species (Vahter, 1994, 1999a), which may explain why there has been no adult animal 15 model for the carcinogenic potential of arsenic. Although a few animal bioassays have been 16 conducted, they have all been negative. Arsenic-induced cancers have been observed with 17 transplacental exposure in mice. Transplacental exposure to arsenic in mice has found increases 18 in the development of lung, liver, reproductive, and adrenal tumors. Skin tumors in animals have 19 only been induced in transgenic models or in co-carcinogenesis studies. 20 Despite the lack of a good animal model for arsenic carcinogenesis, numerous 21 epidemiological studies have examined the carcinogenic potential of inorganic arsenic via oral 22 exposure. Although each of the investigations has its own inherent strengths and weaknesses, 23 the combination of all the study results supports an association between oral exposure to 24 inorganic arsenic and cancer including bladder, kidney, skin, lung, liver, and prostate. Because 25 the association between arsenic and these cancers has been found in different populations, it is 26 unlikely that any single attribute (e.g., nutritional habits) associated with a single population is 27 responsible for the increased cancer rates. However, genetic polymorphisms have been found to 28 be an important factor in the methylation of arsenic. Evidence suggests that people who have a 29 greater capacity to methylate arsenic completely to DMA are at a lower risk for developing 30 arsenic-related cancers. Nutritional and personal habits including smoking also affect the 31 methylation rate. Therefore, genetic, nutritional, and lifestyle factors contribute to the inter- 32 individual variations. 146 DRAFT—DO NOT CITE OR QUOTE ------- 1 Although dose-response relationships have been observed for the majority of cancers 2 noted in areas with high levels of arsenic in their drinking water, results for low-level arsenic 3 epidemiologic investigations (primarily from the United States and Europe) have been equivocal 4 in the relationship between these cancers and arsenic exposure. This could be due to the fact that 5 none of the studies accounted for arsenic exposure through food sources, which would be a 6 significant source as the levels in the drinking water decreased (Uchino et al., 2006; Kile et al., 7 2007). Because cancer has a long latency period, misclassification also occurs due to lack of 8 data on disease-relevant exposures (Cantor and Lubin, 2007), which would be more significant 9 in studies examining lower exposures. Therefore, studies with low levels of exposure that are 10 ecological in nature (no individual exposure) are more prone to exposure misclassification, 11 which means they are biased toward the null hypothesis. Despite all these numerous limitations 12 in low-level exposure studies, positive associations have been observed for cancers of the 13 prostate (Hinwood et al., 1999; Lewis et al., 1999), skin (Hinwood et al., 1999; Karagas et al., 14 2001; Beane-Freeman et al., 2004; Knobeloch et al., 2006), and bladder (Kurttio et al., 1999; 15 Steinmaus et al., 2003; Karagas et al., 2004). In most cases, however, there is no dose-response 16 with increases observed at the highest concentrations only and in many cases significant results 17 occurred in smokers only. 18 Based upon current EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), 19 inorganic arsenic is determined to be "carcinogenic to humans" due to convincing 20 epidemiological evidence of a causal relationship between oral exposure of humans to inorganic 21 arsenic and cancer. 22 The available evidence is inadequate to establish a MOA by which arsenic induces 23 tumors. The genotoxicity data for arsenic are equivocal. Chromosomal aberrations have been 24 observed in humans and animals exposed to arsenic, but arsenic has been generally negative in 25 bacterial mutagenicity tests and has only been observed to be a weak mutagen at the hprt locus in 26 Chinese hamster V79 cells at toxic concentrations (Li and Rossman, 1989a). In addition, even 27 though it appears genotoxic in animal models, it does not generally induce tumors in animal 28 models. Arsenic does not appear to cause point mutations in standard assays, but instead causes 29 large deletion mutations (Rossman, 1998). These large deletions can cause lethality when 30 closely linked to essential genes. Therefore, the mutations are not easily observed in standard 31 bacterial and mammalian cell mutation assays. However, even in transgenic cell lines, which 32 were tolerant of large deletions, arsenic was still only weakly mutagenic at doses causing overt 33 cytotoxicity (Rossman, 2003). It has been suggested that arsenic acts as an aneugen (affects the 34 number of chromosomes) at low doses, but as a clastogen (causes chromosomal breaks) at high 35 doses (Rossman, 2003). However, arsenic has also been demonstrated to affect other processes 36 possibly involved with carcinogenesis, including aberrant gene/protein expression, ROS, DNA 37 repair inhibition, signal transduction, and cancer promotion. Therefore, it is likely that arsenic 147 DRAFT—DO NOT CITE OR QUOTE ------- 1 acts via multiple MO As, which would explain the number of different internal cancers associated 2 with arsenic. 6.2. DOSE-RESPONSE 3 Only the oral cancer assessment is addressed in this document. Lung and bladder cancer 4 mortality in the Taiwanese population were selected as endpoints in the dose-response modeling 5 because they are the internal cancers with the most consistent results and are best characterized 6 in epidemiology studies of arsenic exposure (NRC, 1999, 2001; SAB 2000, 2007). Dose- 7 response models were estimated for the Taiwanese population using additive Poisson regression 8 with linear dose terms and quadratic age terms. 9 ED0i values were derived from the MLE dose-response parameter estimates. LED01 10 estimates were derived from the 95% upper confidence limits on the dose-response parameters, 11 as described in Appendix E. The analysis was done in two phases. The first phase consisted of 12 the derivation and fitting of dose-response models using the Taiwanese epidemiology data from 13 Chen et al. (1988a, 1992) and Wu et al. (1989). The outputs of this phase of the analysis were 14 arsenic dose-response coefficients that described the relationship between estimated arsenic 15 intake in the Taiwanese population and proportional increases in age-specific lung and bladder 16 cancer mortality risk. Lifetime cancer incidence in U.S. populations was then estimated by using 17 a modified version of the "BEIRIV" relative risk model. A key assumption underlying this 18 model is that the risk of arsenic-related cancer is a constant multiplicative function of the 19 "background" age profile of cancer risks in the target U.S. population. Estimates of arsenic- 20 related cancer risks in a (hypothetical) U.S. population exposed to arsenic at varying levels in 21 drinking water were then derived. 22 The oral CSFs for lung and bladder cancers in U.S. males and females were derived using 23 the following assumptions: nonwater arsenic intake for the reference and exposed populations 24 was 10 ug/day; drinking water consumption was 3.5 and 2.0 L/day in Taiwanese men and 25 women, respectively; 50 kg was the average Taiwanese body weight; and a 70 kg individual in 26 the United States consumes 2.0 L/day of water (Section 5.3.5). The oral CSF is dependent on 27 assumptions related to the volume of contaminated water consumed over the course of a day and 28 the amount of arsenic consumed through the diet. Changes in these assumptions would result in 29 different cancer potency estimates (as discussed in Section 5.3.8.3), and corresponding changes 30 in the other risk criteria (drinking water unit risk, drinking water concentration associated with 31 lOLEDoi lifetime cancer risk, etc.). Sensitivity analyses were performed to test the effects of 32 differences in drinking water intake assumptions, nonwater arsenic intake assumptions, using 33 median well water values compared to minimum and maximum values, and including different 34 Taiwanese reference populations on the estimates (Section 5.3.8.3). Based on the results of the 35 sensitivity analyses, the risk model results, with the exception of female bladder cancer, appear 148 DRAFT—DO NOT CITE OR QUOTE ------- 1 to be relatively stable and react predictably to reasonable changes in exposure assumptions. 2 Female bladder cancer estimates were particularly sensitive to variations in nonwater arsenic 3 intake. 4 Estimated cancer potency factors for lifetime U.S. male lung and bladder cancer 5 incidence were 6.7 and 11.2 per mg/kg-day, respectively. The corresponding values for females 6 were 16.6 and 10.5 per mg/kg-day (Table 5-3). Cancer potency for combined lung and bladder 7 cancer risks were estimated for males and females, as described in Section 5.3.8.1. The 8 estimated cancer potency factors for combined (lung plus bladder) cancer incidence were 16.9 9 and 25.7 per mg/kg-day, respectively. The potency factor estimate for women (25.7 per 10 mg/kg-day) was identified as the recommended point of departure for derivation of health 11 criteria, with women being the more sensitive population. 12 The cancer potency estimates derived in this analysis are not directly comparable to those 13 estimated in EPA's 1988 assessment (U.S. EPA, 1988b). That analysis derived a much lower 14 potency factor estimate (1.0-2.0 per mg/kg-day) based on an analysis of skin cancer incidence in 15 the Taiwanese population studied by Tseng et al. (1968; Tseng, 1977). Since the exposure- 16 response data on internal cancers has become available, all the subsequent assessments 17 (including this one) have been based on internal (bladder and/or lung) cancer (see Section 5.3.1). 18 The difference in endpoints (skin versus internal cancers) is the main reason for the relatively 19 large difference in estimated cancer potency in the more recent assessment compared to the 1988 20 assessment. 21 As discussed in Section 5.3.8.2, the lifetime risk estimates for male and female lung and 22 bladder cancer calculated in this assessment are generally consistent with the risk estimates from 23 previous analyses that used the internal cancers (NRC, 2001). The bulk of the difference 24 between the cancer potency estimates in this assessment and those from previous analyses can be 25 explained by differences in dose-response models, changes in the assumptions related to the 26 relative drinking water consumption by women in Taiwan and the United States, and the use of 27 more recent data on U.S. population mortality and cancer incidence in the BEIRIV relative risk 28 model. 29 The Supplemental Guidance for Assessing Susceptibility From Early-Life Exposure to 30 Carcinogens (U.S. EPA, 2005b) indicates that age-dependent adjustment factors should be 31 applied to the CSF and combined with early-life exposure estimates when estimating cancer 32 risks from exposures to carcinogens with a mutagenic MO A. As discussed in Section 4.6.3, 33 insufficient data are available to adequately demonstrate a mutagenic mode of action for 34 inorganic arsenic. Therefore, the application of age-dependent adjustment factors is not 35 recommended. 36 The overall level of confidence in the data is high. The data used in the dose-response 37 assessment come from human epidemiology rather than animal bioassays. The Taiwanese 149 DRAFT—DO NOT CITE OR QUOTE ------- 1 studies characterize the cancer risks of an extremely large, well-characterized population with a 2 wide range of exposure concentrations. Reliability and accuracy of mortality records, 3 verification of endpoints with histological examinations, several decades of exposure to arsenic 4 in drinking water to detect internal cancer outcomes, apparent similarities in lifestyle habits 5 (similar urbanization in the endemic area versus the rest of southwestern Taiwan) between 6 exposed and reference populations, and the residential stability of the population (i.e., little 7 migration or emigration) are high. The data demonstrate a statistically significant dose-related 8 effect in humans, across the entire range of exposures (i.e., 10-934 ppb median levels) evaluated. 9 The currently used BEIRIV model is an improvement over previous models because it contains 10 a quadratic age model, an additive linear dose term, and a reference population, and adjusts for 11 differences between the exposed and target (i.e., U.S.) populations. 12 Despite all their strengths, the Chen et al. (1988a, 1992) and Wu et al. (1989) studies are 13 "ecological"; data on individual exposure (which are a function of both water consumption rates 14 and concentrations) are not available. In addition, smoking information was not provided in the 15 critical studies (however, it appears comparable—40% vs. 32% in endemic area vs. the rest of 16 Taiwan according to Chen et al., 1985). Lacking this information introduces an unquantifiable 17 degree of uncertainty into the risk estimates. In EP A's judgment, these factors are equally likely 18 to have resulted in overestimates or underestimates of risks. 6.2.1. Choice of Models 19 As discussed in Section 5.3.1, the Taiwanese data have been used as the basis for 20 quantitative risk assessment by a number of investigators. In this current analysis, EPA is 21 building on the experience of previous efforts by itself and others, and has incorporated 22 comments and recommendations by NRC (2001) and SAB (SAB, 2007) in the selection of 23 statistical methods for use in the risk assessment. As discussed in Section 5.3.7.1, the current 24 assessment employs a Poisson regression model with additive linear dose terms and quadratic 25 age terms for dose-response model fitting in the Taiwanese population. This model was found to 26 be the simplest, best-fitting model among a number of alternatives tested. Sensitivity analyses of 27 other models (quadratic, exponential linear, and exponential quadratic dose transformation) were 28 also conducted (see Section 5.3.8.4 for further details). 29 To extrapolate arsenic-related cancer risks to the U.S. population, the current assessment 30 employs a variant of the "BEIR IV" relative risk model (Section 5.3.7.3). This model takes as its 31 inputs the dose-response coefficients from the Poisson regressions and "background" cancer 32 incidence and population mortality data from the target (U.S.) population. Population mortality 33 data for the year 2000 (NCHS, 2000) and background lung and bladder cancer incidence for 34 2000-2003 (NCI, 2006) were used as inputs to the BEIR IV model. 150 DRAFT—DO NOT CITE OR QUOTE ------- 6.2.2. Dose Metric 1 Inorganic arsenic is metabolized in vivo, with some of the known metabolites being more 2 toxic than the parent compound. However, it is not known whether it is a metabolite, the parent 3 compound, or a combination of the two that is responsible for the observed carcinogenic 4 potential. An increase in MMA or decreased DMA in the urine has been associated with an 5 increase in disease risk (Yu et al., 2000; Chen et al., 2005a; Steinmaus et al., 2005; Valenzuela et 6 al., 2005; Ahsan et al., 2007; Huang et al., 2007b; McCarthy et al., 2007a); therefore, the actual 7 carcinogenic moiety may not be proportional to administered exposure and use of administered 8 exposure may produce a bias in the model. However, the exposure assessment for the model is 9 ecological in nature and produces its own inherent bias. Detailed arsenic speciation data are not 10 available for the Taiwanese population used in the risk assessment. Therefore, estimated total 11 daily arsenic dose (water + other dietary) has been used as the dose metric in the risk assessment. 12 Arsenic dose is estimated based on well water concentration data, and it is assumed that the 13 arsenic concentrations have been constant over the period of exposure. Since there are no data 14 related to the temporal variability in the well water concentrations, this introduces uncertainty 15 into the dose estimates for the 43 villages. Sensitivity analyses were conducted to investigate the 16 impact of using alternative exposure indices, as discussed in Section 5.3.8.3. 6.2.3. Human Population Variability 17 Although the extent of inter-individual variability in arsenic metabolism has not been 18 adequately characterized, genetic polymorphism, nutritional status, and personal habits (e.g., 19 smoking) have all been associated with differences in arsenic methylation. Data exploring 20 whether there is a differential sensitivity to arsenic carcinogenicity across life stages is limited. 21 Data by Waalkes et al. (2003, 2004a) indicate that transplacental exposure in mice is a sensitive 22 stage for carcinogenic potential. These are the only studies in which inorganic arsenic exposure 23 has been associated with cancer in rodents. Lung, liver, reproductive, and adrenal tumors were 24 associated with arsenic administration during gestation (10 days only). A single epidemiological 25 study by Smith et al. (2006) examined lung cancer rates (and other respiratory diseases) in 26 cohorts exposed during childhood and cohorts likely exposed in utero to arsenic concentrations 27 of 860 ppb that subsequently dropped to 100 ppb. Results demonstrated that exposure during 28 either period of development caused increased risk of lung cancer in females aged 40 to 49 born 29 between 1950 and 1957 and in males aged 30 to 49 born between 1950 and 1970. However, the 30 risks associated with early childhood exposures and/or in utero exposures were not compared to 31 risks from exposures during adulthood. Thus, the available data do not allow for a quantitative 32 assessment of the relative sensitivity to arsenic exposures between the Taiwanese population 33 used in the dose-response assessment and U.S. populations exposed to arsenic in drinking water. 34 SAB (2007) acknowledged "the possible issue of compromised nutrition among 35 segments of the exposed population" in the Taiwanese study population, along with the lack of 151 DRAFT—DO NOT CITE OR QUOTE ------- 1 data related to smoking history. However, data are not available that would allow quantitative 2 evaluation of these factors. Therefore, this risk assessment assumes that the observed 3 carcinogenic potency in the Taiwanese population, with suitable corrections for differences in 4 drinking water intake and background cancer incidence, is an appropriate predictor of the 5 potential for human cancer risk in the U.S. population. 152 DRAFT—DO NOT CITE OR QUOTE ------- 7. REFERENCES Achanzar, WE; Brambila, EM; Diwan, BA; et al. (2002) Inorganic arsenite induced malignant transformation of human prostate epithelial cells. JNatl Cancer Inst 94:1888-1891. Ackerman, AH; Creed, PA; Parks, AN; et al. (2005) Comparison of a chemical and enzymatic extraction of arsenic from rice and an assessment of the arsenic absorption from contaminated water by cooked rice. Environ Sci Technol 39:5241-5246. Adair, BM; Moore, T; Conklin, SD; et al. (2007) Tissue distribution and urinary excretion of dimethylated arsenic and its metabolites in dimethyl arsenic acid- or arsenate-treated rats. Toxicol Appl Pharmacol 222:235-242. Adams, MA; Bolger, PM; Gunderson, EL. (1994) Dietary intake and hazards of arsenic. In: Chappell, WR; Abernathy, CO; and Cothern, CR; eds. Arsenic: exposure and health. Northwood, UK: Sci Tech Lett; pp. 41-49. Ahsan, H; Chen, Y; Parvez, F; et al. (2006) Arsenic exposure from drinking water and risk of premalignant skin lesions in Bangladesh: baseline results from the health effects of arsenic longitudinal study. Am J Epidemiol 163(12):1138-1148. Ahsan, H; Chen, Y; Kibriya, MG; et al. (2007) Arsenic metabolism, genetic susceptibility, and risk of premalignant skin lesions in Bangladesh. Cancer Epidemiol Biomarkers Prev 16(6): 1270-1278. An, Y; Kato, K; Nakano, M; et al. (2005) Specific induction of oxidative stress in terminal bronchiolar Clara cells during dimethylarsenic-induced lung tumor promoting process in mice. Cancer Lett 230:57-64. Andersen, O. (1983) Effects of coal combustion products and metal compounds on sister chromatid exchange (SCE) in a macrophage like cell line. Environ Health Perspect 47:329-253. Andrew, AS; Karagas, MR; Hamilton, JW. (2003) Decreased DNA repair gene expression among individuals exposed to arsenic in United States drinking water. Int J Cancer 104:263-268. Andrew, AS; Burgess, JL; Meza, MM; et al. (2006) Arsenic exposure is associated with decreased DNA repair in vitro and in individuals exposed to drinking water arsenic. Environ Health Perspect 114:1193-1198. Andrewes, P; Kitchin, KT; Wallace, K. (2003) Dimethylarsine and trimethylarsine are potent genotoxins in vitro. ChemRes Toxicol 16:994-1003. Anundi, I; Hogberg, J; Vahter, M. (1982) GSH release in bile as influenced by arsenite. FEES Lett 145:285-288. Aposhian, HV. (1997) Enzymatic methylation of arsenic species and other new approaches to arsenic toxicity. Annu Rev Pharmacol Toxicol 37:397-419. Aposhian, HV; Aposhian, MM. (2006) Arsenic toxicology: five questions. ChemRes Toxicol 19(1): 1-15. Argos, M; Kibriya, MG; Parvez, F; et al. (2006) Gene expression profiles in peripheral lymphocytes by arsenic exposure and skin lesion status in a Bangladeshi population. Cancer Epidemiol Biomarkers Prev 15:1367-1375. Arkusz, J; Stariczyk, M; Lewiriska, D; St^pnik, M. (2005) Modulation of murine peritoneal macrophage function by chronic exposure to arsenate in drinking water. Immunopharmacol Immunotoxic 27:315-330. Askar, N; Cirpan, T; Toprak, E; et al. (2006) Arsenic trioxide exposure to ovarian carcinoma cells leads to decreased level of topoisomerase II and cytotoxicity. Int J Gynecol Cancer 16:1552-1556. ATSDR (Agency for Toxic Substances and Disease Registry). (1993) Toxicological profile for arsenic (update). Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA. 153 DRAFT—DO NOT CITE OR QUOTE ------- ATSDR (Agency for Toxic Substances and Disease Registry). (2000) Toxicological profile for arsenic (update). Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA. ATSDR (Agency for Toxic Substances and Disease Registry). (2007) Toxicological profile for arsenic (update). Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA. Avani, G; Rao, MG. (2007) Genotoxic effects in human lymphocytes exposed to arsenic and vitamin A. Toxicol in Vitro 21:626-631. Baastrup, R; Sorensen, M; Balstrom, T; et al. (2008) Arsenic in drinking-water and risk for cancer in Denmark. Environ Health Prospect 116:231-237. Banerjee, M; Sarkar, J; Das, JK; et al. (2007) Polymorphism in the ERCC2 codon 751 is associated with arsenic- induced premalignant hyperkeratosis and significant chromosome aberrations. Carcinogenesis 28(3):672-676. Barchowsky, A; Dudek, EJ; Treadwell, MD; Wetterhahn, KE. (1996) Arsenic induces oxidant stress and NF-KB activation in cultured aortic endothelial cells. Free Radic Biol Med 21:783-790. Barchowsky, A; Roussel, RR; Klei, LR; et al. (1999a) Low levels of arsenic trioxide stimulate proliferative signals in primary vascular cells without activating stress effector pathways. Toxicol Appl Pharmacol 159:65-75. Barchowsky, A; Klei, LR; Dudek, EJ; et al. (1999b) Stimulation of reaction oxygen, but not reactive nitrogen species, in vascular endothelial cells exposed to low levels of arsenite. Free Radic Biol Med 27:1405-1412. Barrett, JC; Lamb, PW; Wang, TC; Lee, TC (1989) Mechanism of arsenic-induced cell transformation. Biol Trace Element Res 21:421^29. Bashir, S; Sharma, Y; Irshad, M; et al. (2006a). Arsenic induced apoptosis in rat liver following repeated 60 days exposure. Toxicol 217:63-70. Bashir, S; Sharma, Y; Irshad, M; et al. (2006b) Arsenic-induced cell death in liver and brain of experimental rats. Basic ClinPharmacol Toxicol 98:38-43. Basu, A; Mahata, J; Roy, AK; et al. (2002) Enhanced frequency of micronuclei in individuals exposed to arsenic through drinking water in West Bengal, India. Mutat Res 516:29^0. Basu, A; Som, A; Ghoshal, S; et al. (2005) Assessment of DNA damage in peripheral blood lymphocytes of individuals susceptible to arsenic induced toxicity in West Bengal, India. Toxicol Lett 159:100-112. Bates, MN; Smith, AH; Cantor, KP. (1995) Case-control study of bladder cancer and arsenic in drinking water. Am J Epidemiol 141:523-530. Bates, MN; Rey, OA; Biggs, ML; et al. (2004) Case-control study of bladder cancer and exposure to arsenic in Argentina. Am J Epidemiol 159(4):381-389. Beane-Freeman, LE; Dennis, LK; Lynch, CF; et al. (2004) Toenail arsenic content and cutaneous melanoma in Iowa. Am J Epidemiol 160(7):679-687. Beckman, G; Beckman, L; Nordenson, I. (1977) Chromosome aberrations in workers exposed to arsenic. Environ Health Perspect 19:145-146. Benbrahim-Tallaa, L; Waterland, RA; Styblo, M; et al. (2005) Molecular events associated with arsenic-induced malignant transformation of human prostatic epithelial cells: aberrant genomic DNA methylation and K-ras oncogene activation. Toxicol Appl Pharmacol 206:288-298. Bodwell, JE; Kingsley, LA; Hamilton, JW. (2004) Arsenic at very low concentrations alters glucocorticoid receptor (GR)-mediated gene activation but not GR-mediated gene repression: complex dose-response effects are closely 154 DRAFT—DO NOT CITE OR QUOTE ------- correlated with levels of activated GR and require a functional GR DNA binding domain. Chem Res Toxicol 17:1064-1076. Bodwell, JE; Gosse, JA; Nomikos, AP; Hamilton, JW. (2006) Arsenic disruption of steroid receptor gene activation: complex dose-response effects are snared by several steroid receptors. Chem Res Toxicol 19(12): 1619-1629. Bolton, JL; Pisha, E; Zhang, F; Qiu, S. (1998) Role of quinoids in estrogen carcinogenesis. Chem Res Toxicol 11:1113-1127. Bolton, JL; Trash, MA; Penning, TM; et al. (2000) Role of quinones in toxicology. Chem Res Toxicol 13:135-160. Bredfeldt, TG; Kopplin, MJ; Gandolfi, AJ. (2004) Effects of arsenite onUROtsa cells: low-level arsenite causes accumulation of ubiquitinated proteins that is enhanced by reduction in cellular glutathione levels. Toxicol Appl Pharmacol 198:412^118. Bredfeldt, TG; Jagadish, B; Eblin, KE; et al. (2006) Monomethylarsonous acid induces transformation of human bladder cells. Toxicol Appl Pharmacol 216:69-79. Brenton, CV; Houseman, EA; Kile, ML; et al. (2006) Gender-specific protective effects of hemoglobin on arsenic- induced skin lesions. Cancer Epidemiol Biomarkers Prev 15(5):902-907. Brenton, CV; Kile, ML; Catalano, PJ; et al. (2007a) GSTM1 and APE1 genotypes affect arsenic-induced oxidative stress: a repeated measures study. Environ Health 6:39. Available online at http://www.ehjournal.net/content/6/l/39. Brenton, CV; Zhou, W; Kile, ML; et al. (2007b) Susceptibility to arsenic-induced skin lesions from polymorphisms in base pair excision repair genes. Carcinogenesis 28(7): 1520-1525. Buchet, JP; Lauwerys, R; Roels, H. (1981) Comparison of the urinary excretion of arsenic metabolites after a single dose of sodium arsenite, monomethylarsonate or dimethylarsinate in man. Int Arch Occup Environ Health 48:71-79. Buchet, JP; Geubel, A; Pauwels, S; et al. (1984) The influence of liver disease on the methylation of arsenite in humans. Arch Toxicol 55:151-154. Buchet, JP; Lauwerys, R. (1985) Study of inorganic arsenic methylation by rat liver in vitro: relevance for the interpretation of observations in man. Arch Toxicol 57:125-129. Bunderson, M; Pereira, F; Schneider, MC; et al. (2006) Manganese enhances peroxynitrite and leukotriene E4 formation in bovine aortic endothelial cells exposed to arsenic. Cardiovasc Toxicol 6:15-23. Burgdorf, W; Kurvink, K; Cervenka, J. (1977) Elevated sister chromatid exchange rate in lymphocytes of subjects treated with arsenic. Hum Genet 36(l):69-72. Bums, FJ; Uddin, AN; Wu, F; et al. (2004) Arsenic-induced enhancement of ultraviolet radiation carcinogenesis in mouse skin: a dose-response study. Environ Health Perspect 112(5):599-603. Bustamante, J; Nutt, L; Orrenius, S; Gogvadze, V. (2005) Arsenic stimulates release of cytochrome c from isolated mitochondria via induction of mitochondrial permeability transition. Toxicol Appl Pharmacol 207(2 Suppl): S110- S116. Cantoni, O; Hussain, S; Guidarelli, A; et al. (1994) Cross-resistance to heavy metals in hydrogen peroxide-resistant CHO cell variants. Mutat Res 324:1-6. Cantor, KP; Lubin, JH. (2007) Arsenic, internal cancers, and issues in inference from studies of low-level exposures in human populations. Toxicol Appl Pharmacol 222:252-257. Chanda, S; Dasgupta, UB; Guha Mazumder, D; et al. (2006) DNA hypermethylation of promoter of gene p53 and p!6 in arsenic-exposed people with and without malignancy. Toxicol Sci 89:431-437. 155 DRAFT—DO NOT CITE OR QUOTE ------- Chao, J-I; Hsu, SH; Tsou, T-C. (2006a) Depletion of securin increases arsenite-induced chromosome instability and apoptosis via a p53-independent pathway. Toxicol Sci 90:73-86. Chao, HR; Tsou, TC; Li, LA; et al. (2006b) Arsenic inhibits induction of cytochrome P450 1A1 by 2,3,7,8- tetrachlorodibenzo-p-dioxin in human hepatoma cells. J Hazard Mater 137:716-722. Charbonneau, SM; Tarn, GKH; Bryce, F; et al. (1979) Metabolism of orally administered inorganic arsenic in the dog. Toxicol Lett 3:107-114. Charbonneau, SM; Hollins, JG; Tarn, GKH; et al. (1980) Whole-body retention, excretion, and metabolism of [ As]arsenic acid in the hamster. Toxicol Lett 5:175-182. Chen, CJ; Wang, CJ. (1990) Ecological correlation between arsenic level in well water and age-adjusted mortality from malignant neoplasms. Cancer Res 50:5470-5474. Chen, KP; Wu, HY; Wu, TC. (1962) Epidemiologic studies onblackfoot disease in Taiwan. 3. Physicochemical characteristics of drinking water in endemic blackfoot disease areas. In: Memoirs, College of Medicine, National Taiwan University, Vol. 8, pp. 115-129. Taipei: National Taiwan University College of Medicine. Chen, C-J; Chuang, Y-C; Lin, TM; Wu, HY. (1985) Malignant neoplasms among residents of a blackfoot disease- endemic area in Taiwan: high-arsenic artesian well water and cancers. Cancer Res 45:5895-5899. Chen, C-J; Chuang, Y-C; You, S-L; et al. (1986) A retrospective study on malignant neoplasms of bladder, lung, and liver in blackfoot disease endemic area in Taiwan. Br J Cancer 53:399-405. Chen, C-J; Kuo, T-L; Wu, M-M. (1988a) Arsenic and cancers. Lancet: letter to the editor. February 20, 1988. Chen, C-J; Wu, M-M; Lee, S-S; et al. (1988b) Atherogenicity and carcinogenicity of high-arsenic artesian well water. Arteriosclerosis 8:452^60. Chen, C-J; Chen, C-W; Wu, M-M; Kuo, T-L. (1992) Cancer potential in liver, lung, bladder, and kidney due to ingested inorganic arsenic in drinking water. Br J Cancer 66(5):888-892. Chen, Y; Megosh, LC; Gilmour, SK; et al. (2000a) K6/ODC transgenic mice as a sensitive model for carcinogenic information. Toxicol Lett 116:27-35. Chen, Y; Liu, X; Pisha, E; et al. (2000b) A metabolite of equine estrogens, 4-hydroxyequilenin, induces DNA damage and apoptosis in breast cancer cell lines. ChemRes Toxicol 13:342-350. Chen, H; Liu, J; Merrick, BA; Waalkes, MP. (2001) Genetic events associated with arsenic-induced malignant transformation: applications of cDNA microarray technology. Mol Carcinog 30:79-87. Chen, CL; Hsu LI; Chiou HY; et al. (2004a) Ingested arsenic, cigarette smoking, and lung cancer risk: a follow-up study in arseniasis-endemic areas in Taiwan. JAMA 292:2984-90. Chen, H; Li, S; Liu, J; et al. (2004b) Chronic inorganic arsenic exposure induces hepatic global and individual gene hypomethylation: implications for arsenic hepatocarcinogenesis. Carcinogenesis 25(9): 1779-1786. Chen, Y-C; Su, H-JJ; Guo, Y-LL; et al. (2005a) Interaction between environmental tobacco smoke and arsenic methylation ability on the risk of bladder cancer. Cancer Causes Control 16:75-81. Chen, P-H; Lan, C-CE; Chiou, M-H; et al. (2005b) Effects of arsenic and UVB on normal human cultured keratinocytes: impact on apoptosis and implication on photocarcinogenesis. Chem Res Toxicol 18:139-144. 156 DRAFT—DO NOT CITE OR QUOTE ------- Chen, D; Chan, R; Waxman, S; Jing, Y. (2006) Buthionine sulfoximine enhancement of arsenic trioxide-induced apoptosis in leukemia and lymphoma cells is mediated via activation of c-Jun NH2-terminal kinase and up- regulation of death receptors. Cancer Res 66:11416-11423. Chen, Y; Hall, M; Graziano, JH; et al. (2007) A prospective study of blood selenium levels and the risk of arsenic- related premalignant skin lesions. Cancer Epidemiol Biomarkers Prev 16(2):207-213. Cheng, HY; Li, P; David, M; et al. (2004) Arsenic inhibition of the JAK-STAT pathway. Oncogene 23:3603-3612. Cheng, Y; Chang, LW; Tsou, TC. (2006) Mitogen-activated protein kinases mediate arsenic-induced down- regulation of survivin in human lung adenocarcinoma cells. Arch Toxicol 80:310-318. Cheng, TJ; Wang, YJ; Kao, WW; et al. (2007) Protection against arsenic trioxide-induced autophagic cell death in U118 human glioma cells by use of lipoic acid. Food Chem Toxicol 45:1027-1038. Chiang, HS; Hong, CL; Guo, H-R; et al. (1988) Comparative study on the high prevalence of bladder cancer in the blackfoot disease endemic area in Taiwan. J Formosan Med Assoc 83(11): 1074-1080. Chien, C-W; Chiang, M-C; Ho, I-C; Lee, T-C. (2004) Association of chromosomal alterations with arsenite-induced tumorigenicity of human HaCaT keratinocytes in nude mice. Environ Health Perspect 112(17): 1704-1710. Chiou, HY; Hsueh, YM; Liaw, KF; et al. (1995) Incidence of internal cancers and ingested inorganic arsenic: a seven-year follow-up study in Taiwan. Cancer Res 55:1296-1300. Chiou, HY; Hsueh, YM; Hsieh, LL; et al. (1997) Arsenic methylation capacity, body retention, and null genotypes of glutathione S-transferase Ml and Tl among current arsenic-exposed residents in Taiwan. Mutat Res 386:197- 207. Chiou, H-Y; Chiou, S-T; Hsu, Y-H; et al. (2001) Incidence of transitional cell carcinoma and arsenic in drinking water: a follow-up study of 8,102 residents in an arseniasis-endemic area in northeastern Taiwan. Am J Epidemiol 153:411-418. Chiu, HF; Ho, S-C; Wang, L-Y; et al. (2004) Does arsenic exposure increase the risk for liver cancer? J Toxicol Environ Health A 67:1491-1500. Chou, W-C; Hawkins, AL; Barrett, JF; et al. (2001) Arsenic inhibition of telomerase transcription leads to genetic instability. J Clinlnves 108:1541-1547. Chou, WC; Chen, HY; Yu, SL; et al. (2005) Arsenic suppresses gene expression in promyelocytic leukemia cells partly through Spl oxidation. Blood 106:304-310. Chowdhury, UK; Rahman, MM; Mondal, BK; et al. (2001) Groundwater arsenic contamination and human suffering in West Bengal - India and Bangladesh. Environ Sci 8(5):393^U5. Chowdhury, UK; Rahman, MM; Sengupta, MK; et al. (2003) Pattern of excretion of arsenic compounds [arsenite, arsenate, MMA(V), DMA (V)] in urine of children compared to adults from an arsenic exposed area in Bangladesh. J Environ Sci Health (Part A) A38(l):87-113. Christian, WJ; Hopenhayn, C; Centeno, JA; Todorov, T. (2006) Distribution of urinary selenium and arsenic among pregnant women exposed to arsenic in drinking water. Environ Res 100:115-122. Chung, JS; Kalman, DA; Moore, LE; et al. (2002) Family correlations of arsenic methylation patterns in children and parents exposed to high concentrations of arsenic in drinking water. Environ Health Perspect 110:729-733. Clewell, HJ; Thomas, RS; Gentry, PR; et al. (2007) Research toward the development of a biologically based dose response assessment for inorganic arsenic carcinogenicity: a progress report. Toxicol Appl Pharmacol 222:388-398. 157 DRAFT—DO NOT CITE OR QUOTE ------- Cohen, SM; Arnold, LL; Uzvolgyi, E; et al. (2002) Possible role of dimethylarsinous acid in dimethylarsinic acid- induced urothelial toxicity and regeneration in the rat. Chem Res Toxicol 15(9): 1150-1157. Cohen, SM; Arnold, LL; Eldan, M; et al. (2006) Methylated arsenicals: the implications of metabolism and carcinogenicity studies in rodents to human risk assessment. Crit Rev Toxicol 36:99-133. Cohen, SM; Ohnishi, T; Arnold, LL; Lee, XC. (2007) Arsenic-induced bladder cancer in an animal model. Toxicol Appl Pharmacol 222:258-263. Colognato R; Coppede, F; Ponti, J; et al. (2007) Genotoxicity induced by arsenic compounds in peripheral human lymphocytes analysed by cytokinesis-block micronucleus assay. Mutagenesis 22:255-261. Concha, G; Vogler, G; Lezeano, D; et al. (1998a) Exposure to inorganic arsenic metabolites during early human development. Toxicol Sci 44:185-190. Concha, G; Nermell, B; Vahter, M. (1998b) Metabolism of inorganic arsenic in children with chronic high arsenic exposure in northern Argentina. Environ Health Perspect 106:355-359. Concha, G; Vogler, G; Nermell, B; Vahter, M. (1998c) Low-level arsenic excretion in breast milk of native Andean women exposed to high levels of arsenic in the drinking water. Int Arch Occup Environ Health 71:42-46. Cross, JD; Dale, IM; Leslie, ACD; Smith, H. (1979) Industrial exposure to arsenic. J Radioanal Chem 48:197-208. Crossen, PE. (1983) Arsenic and SCE in human lymphocytes. Mut Res 119:415-419. Cui, X; Kobayashi, Y; Hayakawa, T; Hirano, S. (2004a) Arsenic speciation in bile and urine following oral and intravenous exposure to inorganic and organic arsenics in rats. Toxicol Sci 82:478-487. Cui, X; Li, S; Shraim, A; et al. (2004b) Subchronic exposure to arsenic through drinking water alters expression of cancer-related genes in rat liver. Toxicol Pathol 32:64-72. Cui, X; Wakai, T; Shirai, Y. (2006) Chronic oral exposure to inorganic arsenate interferes with methylation status of p!6INK4a and RASSF1A and induces lung cancer in A/J mice. Toxicol Sci 91:372-381. Das, D; Chatterjee, A; Mandal, BK; et al. (1995) Arsenic in ground water in six districts of West Bengal, India: the biggest arsenic calamity in the world. Part 2. Arsenic concentration in drinking water, hair, nails, urine, skin-scale, and liver tissue (biopsy) of the affected people. Analyst 120:917-924. Das, S; Santra, A; Lahiri, S; Guha Mazumder, DN. (2005) Implications of oxidative stress and hepatic cytokine (TNF-a and IL-6) response in the pathogenesis of hepatic collagenesis in chronic arsenic toxicity. Toxicol Appl Pharmacol 204:18-26. Datta, S; Talukder, G; Sharma, A. (1986) Cytotoxic effects of arsenic in dietary oil primed rats. Sci Culture 52:196- 198. Deaglio, S; Canella, D; Baj, G; et al. (2001) Evidence of an immunologic mechanism behind the therapeutic effects of arsenic trioxide (As2O3) on myeloma cells. Leuk Res 25:227-235. De Chaudhuri, S; Mahata, J; Das, JK; et al. (2006) Association of specific p53 polymorphisms with keratosis in individuals exposed to arsenic through drinking water in West Bengal, India. Mutat Res 601:102-112. De Kimpe, J; Cornells, R; Mees, L; Vanholder, R. (1996) Basal metabolism of intraperitoneally injected carrier-free As-labeled arsenate in rabbits. Fundam Appl Toxicol 34:240-248. De La Fuente, H; Portales-Perez, D; Baranda, L; et al. (2002) Effect of arsenic, cadmium, and lead on the induction of apoptosis of normal human mononuclear cells. Clin Exp Immunol 129:69-77. 158 DRAFT—DO NOT CITE OR QUOTE ------- Del Razo, LM; Styblo, M; Cullen, WR; Thomas DJ. (2001) Determination of trivalent methylated arsenicals in biological matrices. Toxicol Appl Pharmacol 174:283-293. Delnomdedieu, M; Basti, MM; Otvos, JD; Thomas, DJ. (1994a) Reduction and binding of arsenate and dimethylarsinate by glutathione: a magnetic resonance study. Chern Biol Interact 90:139-155. Delnomdedieu, M; Basti, MM; Styblo, M; et al. (1994b) Complexation of arsenic species in rabbit erythrocytes. ChemRes Toxicol 7:621-627. Delnomdedieu, M; Styblo, M; Thomas, DJ. (1995) Time dependence of accumulation and binding of inorganic and organic arsenic species in rabbit erythrocytes. ChemBiol Interact 98:69-83. Diaz, Z; Colombo, M; Mann, KK; et al. (2005) Trolox selectively enhances arsenic-mediated oxidative stress and apoptosis in APL and other malignant cell lines. Blood 105:1237-1245. Di Gioacchino, M; Verna, N; Di Giampaolo, L; et al. (2007) Immunotoxicity and sensitizing capacity of metal compounds depends on speciation. Int J Immunopathol Pharmacol 20(2): 15-22. Ding, W; Hudson, LG; Liu, KJ. (2005) Inorganic arsenic compounds cause oxidative damage to DNA and protein by inducing ROS and RNS generation in human keratinocytes. Mol Cell Biochem 279:105-112. DiPaolo, JA; Casto, BC. (1979) Quantitative studies of in vitro morphological transformation of Syrian hamster cells by inorganic metal salts. Cancer Res 39:1008-1013. Dong, J-T; Luo, X-M. (1993) Arsenic-induced DNA-strand breaks associated with DNA-protein crosslinks in human fetal lung fibroblasts. Mutat Res 302(2):97-102. Dopp, E; Hartmann, LM; Florea, A-M; et al. (2004) Uptake of inorganic and organic derivatives of arsenic associated with induced cytotoxic and genotoxic effects in Chinese hamster ovary (CHO) cells. Toxicol Appl Pharmacol 201:156-165. Drobna, Z; Jaspers, I; Thomas, DJ; Styblo, M. (2002) Differential activation of AP-1 in human bladder epithelial cells by inorganic and methylated arsenicals. FASEB J 17(l):67-69. Drobna, Z; Waters, SB; Walton, FS; et al. (2004) Interindividual variation in the metabolism of arsenic in cultured primary human hepatocytes. Toxicol Appl Pharm 201(2): 166-177. Drobna, Z; Waters, SB; Devesa, V; et al. (2005) Metabolism and toxicity of arsenic in human urothelial cells expressing rat arsenic (+3 oxidation state)-methyltransferase. Toxicol Appl Pharmacol 207:147-159. Drobna, Z; Xing, W; Thomas, DJ; Styblo, M. (2006) shRNA silencing of AS3MT expression minimizes arsenic methylation capacity of HepG2 cells. ChemRes Toxicol 19:894-898. DuMond, JW, Jr; Singh, KP. (2007) Gene expression changes and induction of cell proliferation by chronic exposure to arsenic of mouse testicular Leydig cells. J Toxicol Environ Health A 70:1150-1154. Eblin, KE; Bowen, ME; Cromey, DW; et al. (2006) Arsenite and monomethylarsonous acid generate oxidative stress response in human bladder cell culture. Toxicol Appl Pharmacol 217:7-14. Eblin, KE; Bredfeldt, TG; Buffmgton, S; Gandolfi, AJ. (2007) Mitogenic signal transduction caused by monomethylarsonous acid in human bladder cells: role in arsenic-induced carcinogenesis. Toxicol Sci 95:321-330. El-Masri, H; Kenyon, EM. (2008) Development of a human physiologically based pharmacokinetic (PBPK) model for inorganic arsenic and its mono- and di-methylated metabolites. J Pharmacokinet Pharmacodyn 35:31-68. Engstrom, KS; Broberg, K; Concha, G; et al. (2007) Genetic polymorphism influencing arsenic metabolism: evidence from Argentina. Environ Health Perspect 115(4):599-605. 159 DRAFT—DO NOT CITE OR QUOTE ------- Fan, S-R; Ho, I-C; Yeoh, FL-F; et al. (1996) Squalene inhibits sodium arsenite-induced sister chromatid exchanges and micronuclei in Chinese hamster ovary-Kl cells. Mutat Res 368:165-169. Felix, K; Manna, SK; Wise, K; Barr, J; Ramesh, GT. (2005) Low levels of arsenite activates nuclear factor-KB and activator protein-1 in immortalized mesencephalic cells. J Biochem Mol Toxicol 19:67-77. Ferreccio, C; Gonzalez, C; Milosavjlevic; et al. (2000) Lung cancer and arsenic concentrations in drinking water in Chile. Epidem 1:673-679. Fischer, JM; Robbins, SB; Al-Zoughool, M; et al. (2005) Co-mutagenic activity of arsenic andbenzo[a]pyrene in mouse skin. MutRes 588:35-46. Fischer, JM; Robbins, SB; Kannamkumarath, SS; et al. (2006) Exposure of mice to arsenic and/or benzo[a]pyrene does not increase the frequency of Aprt-deficient cells recovered from explanted skin of Aprt heterozygous mice. Environ Mol Mutagen 47:334-344. Flora, SJ. (1999) Arsenic-induced oxidative stress and its reversibility following combined administration of N- acetylcysteine and meso 2,3-dimercaptosuccinic acid in rats. ClinExp Pharmacol Physiol 26:865-869. Flora, SJS; Bhadauria, S; Pant, SC; Dhaked, RK. (2005) Arsenic induced blood and brain oxidative stress and its response to some thiol chelators in rats. Life Sci 77:2324-2337. Florea, A-M; Yamoah, EN; Dopp, E. (2005) Intracellular calcium disturbances induced by arsenic and its methylated derivatives in relation to genomic damage and apoptosis induction. Environ Health Perspect 113(6):659-664. Florea, A-M; Splettstoesser, F; Btisselberg, D. (2007) Arsenic trioxide (As2O3) induced calcium signals and cytotoxicity in two human cell lines: SY-5Y neuroblastoma and 293 embryonic kidney (HEK). Toxicol Appl Pharmacol 220:292-301. Freeman, GB; Schoof, RA; Ruby, MV; et al. (1995) Bioavailability of arsenic in soil and house dust impacted by smelter activities following oral administration in Cynomologus monkeys. Fund Appl Toxicol 28:215-222. Fujino, Y; Guo, X; Liu, J; et al. (2005) Chronic arsenic exposure and urinary 8-hydroxy-2'-deoxyguanosine in an arsenic-affected area in Inner Mongolia, China. JExpo Anal Environ Epidemiol 15:147-152. Gail MH; Kessler M; Midthune D; Scoppa S. (1999) Two approaches for estimating disease prevalence from population-based registries of incidence and total mortality. Biometrics 55:1137-1144. Galicia, G; Leyva, R; Tenorio, EP; et al. (2003) Sodium arsenite retards proliferation of PHA-activated T cells by delaying the production and secretion of IL-2. Int Immunopharmacol 3:671-682. Gamble, MV; Liu, X; Ahsan, H; et al. (2005) Folate, homocysteine, and arsenic metabolism in arsenic-exposed individuals in Bangladesh. Environ Health Perspect 113(12): 1683-1688. Gamble, MV; Liu, X; Ahsan, H; et al. (2006) Folate and arsenic metabolism: a double-blind, placebo-controlled folic acid-supplementation trial in Bangladesh. Am J Clin Nutr 84:1093-1101. Ganyc, D; Talbot, S; Konate, F; et al. (2007) Impact of trivalent arsenicals on selenoprotein synthesis. Environ Health Perspect 115:346-353. Gentry, PR; Covington, TR; Mann, S; et al. (2004) Physiologically based pharmacokinetic modeling of arsenic in the mouse. J Toxicol Environ Health A 67(1):43-71. Germolec, DR; Spalding, J; Boorman, GA; et al. (1997) Arsenic can mediate skin neoplasia by chronic stimulation of keratinocyte-derived growth factors. Mutat Res 386:209-218. 160 DRAFT—DO NOT CITE OR QUOTE ------- Germolec, DR; Spalding, J; Yu, HS; et al. (1998) Arsenic enhancement of skin neoplasia by chronic stimulation of growth factors. AmJPathol 153:1775-1785. Geubel, AP; Mairlot, MC; Buchet, JP; Lauwerys, R. (1988) Abnormal methylation capacity in human liver cirrhosis. Int J ClinPharmacol Res 8:117-122. Ghosh, P; Basu, A; Mahata, J; et al. (2006) Cytogenetic damage and genetic variants in the individuals susceptibile to arsenic-induced cancer through drinking water. Int J Cancer 118:2470-2478. Gomez, SE; del Razo, LM; Sanchez, JLM. (2005) Induction of DNA damage by free radicals generated either by organic or inorganic arsenic (As ,MMA , and DMA ) in cultures of B and T lymphocytes. Bio Trace Elem Res 108:115-126. Gonsebatt, ME; Vega, L; Montero, R; et al. (1994) Lymphocyte replicating ability in individuals exposed to arsenic via drinking water. Mutat Res 313:293-299. Gonsebatt, ME; Vega, L; Salazar, AM. (1997). Cytogenic effects in human exposure to arsenic. Mutat Res 386:219- 228. Gonzalez-Rangel, Y; Portales-Perez, DP; Galicia-Cruz, O; Escudero-Lourdes, C. (2005) Chronic exposure to arsenic sensitizes CD3+ and CD56+ human cells to sodium arsenite-mediated apoptosis. Proc West Pharmacol Soc 48:89- 91. Gottschalg, E; Moore, NE; Ryan, AC; et al. (2006) Phenotypic anchoring of arsenic and cadmium toxicity in three hepatic-related cell systems reveals compound- and cell-specific selective up-regulation of stress protein expression: implications for fingerprint profiling of cytotoxicity. Chem Biol Interact 161:251-261. Graham-Evans, B; Cohly, HH; Yu, H; Tchounwou, PB. (2004) Arsenic-induced genotoxic and cytotoxic effects in human keratinocytes, melanocytes and dendritic cells. Int J Environ Res Public Health 1:83-89. Gregus, Z; Nemeti, B. (2002) Purine nucleoside phosphorylase as a cytosolic arsenate reductase. Toxicol Sci 70:13- 19. Gregus, Z; Nemeti, B. (2005) The glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase works as an arsenate reductase in human red blood cells and rat liver cytosol. Toxicol Sci 85(2):859-869. Griffin, RJ; Monzen, H; Williams, BW; et al. (2003) Arsenic trioxide induces selective tumor vascular damage via oxidative stress and increases thermosensitivity of tumours. Int J Hypertherm 19:575-589. Guo, HR; Chiang, HS; Hu, H; et al. (1997) Arsenic in drinking water and incidence of urinary cancers. Epidem 8:545-550. Guo, HR; Yu, H-S; Hu, H; Monson, RR. (2001) Arsenic in drinking water and skin cancers: cell-type specificity (Taiwan, R.O.C.). Cancer Causes Control 12:909-916. Guo, HR. (2003) The lack of a specific association between arsenic in drinking water and hepatocellular carcinoma. JHepatol 39:383-388. Guo, HR. (2004) Arsenic level in drinking water and mortality of lung cancer (Taiwan). Cancer Causes Control 15(2):171-177. Gupta, R; Flora, SJ. (2005) Therapeutic value of Hippophae rhamnoides L. against subchronic arsenic toxicity in mice. J Med Food 8:353-361. Gupta, R; Flora, SJ. (2006) Effect of Centella asiatica on arsenic induced oxidative stress and metal distribution in rats. J Appl Toxicol 26:213-222. 161 DRAFT—DO NOT CITE OR QUOTE ------- Guyton, KZ; Xu, QB; Holbrook, NJ. (1996) Induction of the mammalian stress response gene GADD153 by oxidative stress: role of AP-1 element. Biochem J 314:547-554. Hagiwara, M; Watanabe, E; Barrett, JC; Tsutsui, T. (2006) Assessment of genotoxicity of 14 chemical agents used in dental practice: ability to induce chromosome aberrations in Syrian hamster embryo cells. Mutat Res 603:111- 120. Hall, M; Gamble, M; Slakovich, V; et al. (2007) Determinants of arsenic metabolism: blood arsenic metabolites, plasma folate, cobalamine, and homocysteine concentrations in maternal-newborn pairs. Environ Health Perspect 115:1503-1509. Hamadeh, HK; Vargas, M; Lee, M; et al. (1999) Arsenic disrupts cellular levels of p53 and mdm2: a potential mechanism of carcinogenesis. Biochem Biophys Res Commun 263:446-449. Han, SG; Castranova, V; Vallyathan, V. (2005) Heat shock protein 70 as an indicator of early lung injury caused by exposure to arsenic. Mol Cell Biochem 277:153-164. Hanahan, D; Weinberg, RA. (2000) The hallmarks of cancer. Cell 100:57-70. Hanlon, DP; Perm, VH. (1977) Placenta! permeability of arsenate ion during early embryogenesis in the hamster. Experientia 33(9):1221-1222. Hansen, HR- Raab, A; Jaspars, M; et al. (2004) Sulfur-containing arsenicals mistaken for dimethylarsinous acid[DMA( )] and identified as a natural metabolite in urine: major implication for studies on arsenic metabolism andtoxicity. Chem Res Toxicol 17:1086-1091. Hansen, JM; Zhang, H; Jones, DP. (2006) Differential oxidation of thioredoxin-1, thioredoxin-2, and glutathione by metal ions. Free Radic Biol Med 40:138-145. Harrington, JM; Middaugh, JP; Morse, DL; Housworth, J. (1978) A survey of a population exposed to high concentrations of arsenic in well water in Fairbanks, Alaska. Am J Epidemiol 108:377-385. Harrison, MT; McCoy, KL. (2001) Immunosuppression by arsenic: a comparison of cathepsin L inhibition and apoptosis. Int Immunopharmacol 1:647-656. Hasgekar, N; Beck, JP; Dunkelberg, H; et al. (2006) Influence of antimonite, selenite, and mercury on the toxicity of arsenite in primary rat hepatocytes. Biol Trace Elem Res 111:167-183. Hayakawa, T; Kobayashi, Y; Cui, X; Hirano, S. (2005) A new metabolic pathway of arsenite: arsenic-glutathione complexes are substrates for human arsenic methyltransferase Cytl9. Arch Toxicol 79:183-191. He, X; Chen, MG; Lin, GX; Ma, Q. (2006) Arsenic induces NAD(P)H-quinone oxidoreductase I by disrupting the Nrf2 x Keapl x Cul3 complex and recruiting Nrf2 x Maf to the antioxidant response element enhancer. J Biol Chem 281:23620-23631 Healy, SM; Zakharyan, RA; Aposhian, HV. (1997) Enzymatic methylation of arsenic compounds: IV. In vitro and in vivo deficiency of methylation of arsenite and monomethylarsonic acid in the guinea pig. Mutat Res 386:229-239. Healy, SM; Casarez, EA; Ayala-Fierro, F; Aposhian, HV. (1998) Enzymatic methylation of arsenic compounds: V. Arsenite methyltransferase activity in tissues of mice. Toxicol Appl Pharmacol 148:65-70. Healy, SM; Wildfang, E; Zakharyan, RA; Aposhian, HV. (1999) Diversity of inorganic arsenite biotransformation. Biol Trace Elem Res 68(3):249-266. Heck, JE; Gamble, MV; Chen, Y; et al. (2007) Consumption of folate-related nutrients and metabolism of arsenic in Bangladesh. Am J Clin Nutr 85:1367-1374. 162 DRAFT—DO NOT CITE OR QUOTE ------- Hei, TK; Liu, SX; Waldren, C. (1998) Mutagenicity of arsenic in mammalian cells: role of reactive oxygen species. Proc Natl Acad Sci USA 95:8103-8107. Hernandez-Zavala, A; Cordova, E; Del Razo, LM; et al. (2005) Effects of arsenite on cell cycle progression in a human bladder cancer cell line. Toxicology 207:49-57. Heydorn, K. (1970) Environmental variation of arsenic levels in human blood determined by neutron activation analysis. Clin Chem Acta 28:349-357. Hinwood, AL; Jolley, DJ; Sim, MR. (1999) Cancer incidence and high environmental arsenic concentrations in rural populations: results of an ecological study. Int J Environ Health Res 9(2):131-141. Hirano, S; Kobayashi, Y; Cui, X; et al. (2004) The accumulation and toxicity of methylated arsenicals in endothelial cells: important roles of thiol compounds. Toxicol Appl Pharmacol 198:458^67. Hood, RD; Vedel-Macrander, GC; Zaworotko, MJ; et al. (1987) Distribution, metabolism, and fetal uptake of pentavalent arsenic in pregnant mice following oral or intraperitoneal administration. Teratology 35:19-25. Hopenhayn-Rich, C; Biggs, ML; Fuchs, A; et al. (1996a) Bladder cancer mortality associated with arsenic in drinking water in Argentina. Epidemiology 7:117-124. Hopenhayn-Rich, C; Biggs, ML; Kalman, DA; et al. (1996b) Methylation study of a population environmentally exposed to arsenic in drinking water. Environ Health Perspect 104:620-628. Hopenhayn-Rich, C; Biggs, ML; Smith, AH. (1998) Lung and kidney cancer mortality associated with arsenic in drinking water in Cordoba, Argentina. Int J Epidemiol 27:561-569. Hornhardt, S; Gomolka, M; Walsh, L; Jung, T. (2006) Comparative investigations of sodium arsenite, arsenic trioxide and cadmium sulphate in combination with gamma-radiation on apoptosis, micronuclei induction and DNA damage in a human lymphoblastoid cell line. Mutat Res 600:165-176. Hostynek, JJ; Hinz, RS; Lorence, CR; et al. (1993) Metals and the skin. Crit Rev Toxicol 23:171-235. Hou, YC; Hsu, CS; Yeh, CL; et al. (2005) Effects of glutamine on adhesion molecule expression and leukocyte transmigration in endothelial cells exposed to arsenic. J Nutr Biochem 16:700-704. Hour, T-C; Pu, Y-C; Lin, C-C; et al. (2006) Differential expression of molecular markers in arsenic- and non- arsenic-related urothelial cancer. Anticancer Res 26:375-378. HSDB (Hazardous Substance Data Bank). (2005) National Library of Medicine. Records for element arsenic and arsenic compounds. Bethesda, MD: NLM, HSDB. Hsu, S-H; Tsou, T-C; Chiu, S-J; Chao, J-I. (2005) Inhibition of a7-nicotinic acetylcholine receptor expression by arsenite in the vascular endothelial cells. Toxicol Lett 159:47-59. Hsu, K-H; Brandt-Rauf, P; Lin, T-M; et al. (2006) Plasma-transforming growth factor-alpha expression in residents of an arseniasis area in Taiwan. Biomarkers 11:538-546. Hu, Y; Su, L; Snow, ET. (1998) Arsenic toxicity is enzyme specific and its affects on ligation are not caused by the direct inhibition of DNA repair enzymes. Mutat Res 408:203-218. Huang, RN; Lee, TC. (1996) Cellular uptake of trivalent arsenite and pentavalent arsenate in KB cells cultured in phosphate-free medium. Toxicol Appl Pharmacol 136(2):243-249. Huang, SC; Lee, TC. (1998) Arsenite inhibits mitotic division and perturbs spindle dynamics in HeLa S3 cells. Carcinogenesis 19:889-896. 163 DRAFT—DO NOT CITE OR QUOTE ------- Huang, H; Huang, CF; Wu, DR; et al. (1993) Glutathione as a cellular defense against arsenite toxicity in cultured Chinese hamster ovary cells. Toxicology 79:195-204. Huang, C; Ma, WY; Li, J; et al. (1999a) Requirement of Erk, but not INK, for arsenite-induced cell transformation. J Biol Chem 274:14595-14601. Huang, C; Ma, WY; Li, J; et al. (1999b) Arsenic induces apoptosis through a c-Jun NH2-terminal kinase-dependent, p-53-independent pathway. Cancer Res 59:3053-3058. Huang, C; Ke, Q; Costa, M; Shi, X. (2004) Molecular mechanisms of arsenic carcinogenesis. Mol Cell Biochem 255:57-66. Huang, HS; Liu, ZM; Ding, L; et al. (2006) Opposite effect of ERK1/2 and JNK on p53-independent p21 WAF1/CIP1 activation involved in the arsenic trioxide-induced human epidermoid carcinoma A431 cellular cytotoxicity. J Biomed Sci 13:113-125. Huang, Y-C; Hung, W-C; Kang, W.Y; et al. (2007a) Expression of STATS and Bcl-6 oncoprotein in sodium arsenite-treated SV-40 immortalized human uroepithelial cells. Toxicol Lett 173:57-65. Huang, Y-K; Tseng, C-H; Huang, Y-L; et al. (2007b) Arsenic methylation capability and hypertension risk in subjects living in arseniasis-hyperendemic areas in southwestern Taiwan. Toxicol Appl Pharmacol 218:135-142. Hughes, MF; Kenyon, EM. (1998) Dose-dependent effects on the disposition of monomethylarsonic acid and dimethylarsinic acid in the mouse after intravenous administration. J Toxicol Environ Health 53:95-l 12. Hughes, M; Kenyon, E; Edwards, B; et al. (1999) Strain-dependent disposition of inorganic arsenic in the mouse. Toxicology 137:95-108. Hughes, MF; Del Razo, LM; Kenyon, EM. (2000) Dose-dependent effects on tissue distribution and metabolism of dimethylarsinic acid in the mouse after intravenous administration. Toxicology 143(2): 155-166. Hughes, MF; Kenyon, EM; Edwards, BC; et al. (2003) Accumulation and metabolism of arsenic in mice after repeated oral administration of arsenate. Toxicol Appl Pharmacol 191(3):202-210. Hughes, MF; Devesa, V; Adair, BM; et al. (2005) Tissue dosimetry, metabolism and excretion of pentavalent and trivalent monomethylated arsenic in mice after oral administration. Toxicol Appl Pharmacol 208:186-197. Hutchinson, J. (1887) Arsenic cancer. BrMed J 2:1280-1281. Hwang, BJ; Utti, C; Steinberg, M. (2006) Induction of cyclinDl by submicromolar concentrations of arsenite in human epidermal keratinocytes. Toxicol Appl Pharmacol 217:161-167. Islam, LN; Nabi, AHM; Rahman, MM; Zahid, MSH. (2007) Association of respiratory complications and elevated serum immunoglobulins with drinking water arsenic toxicity in humans. J Environ Sci Health A 42:1807-1814. Ivanov, VN; Hei, TK. (2006) Sodium arsenite accelerates TRAIL-mediated apoptosis in melanoma cells through upregulation of TRAIL-R1/R2 surface levels and downregulation of cFLIP expression. Exp Cell Res 312:4120- 4138. Jacobson-Kram, D; Montalbano, D. (1985) The reproductive effects assessment group's report on the mutagenicity of inorganic arsenic. Environ Mutagen 7:787-804. Jan, KY; Wang, TC; Ramanathan, B; Gurr, JR. (2006) Dithiol compounds at low concentrations increase arsenite toxicity. Toxicol Sci 90:432-439. Jha, AN; Noditi, M; Nilsson, R; Natarajan, AT. (1992) Genotoxic effects of sodium arsenite on human cells. Mutat Res284(2):215-221. 164 DRAFT—DO NOT CITE OR QUOTE ------- Jiang, HY; Jiang, L; Wek, RC. (2007) The eukaryotic initiation factor-2 kinase pathway facilitates differential GADD45a expression in response to environmental stress. J Biol Chem 282:3755-3765. Jin, Y; Xi, S; Li, X; et al. (2006a) Arsenic speciation transported through the placenta from mother mice to their newborn pups. Environ Res 101:349-355. Jin, HO; Yoon, SI; Seo, SK; et al. (2006b) Synergistic induction of apoptosis by sulindac and arsenic trioxide in human lung cancer A549 cells via reactive oxygen species-dependent down-regulation of survivin. Biochem Pharmacol 72:1228-1236. Jing, Y; Dai, J; Chalmers-Redman, RM; et al. (1999) Arsenic trioxide selectively induces acute promyelocytic leukemia cell apoptosis via a hydrogen peroxide-dependent pathway. Blood 94:2101-2 111. Josyula, AB; Poplin, GS; Kurzius-Spencer, M; et al. (2006) Environmental arsenic exposure and sputum metalloproteinase concentrations. Environ Res 102:283-290. Juhasz, AL; Smith, E; Weber, J; et al. (2006) In vivo assessment of arsenic bioavailability rice and its significance for human health risk assessment. Environ Health Perspect 114:1826-1831. Juhasz, AL; Smith, E; Weber, J; et al. (2007) Comparison of in vivo and in vitro methodologies for the assessment of arsenic bioavailability in contaminated soils. Chemosphere 69:961-966. Kagey, BY; Bumgarner, JE; Creason, JP. (1977) Arsenic levels in maternal-fetal tissue sets. In: Hemphill, DD; ed. Trace substances in environmental health-XI, proceedings of the University of Missouri's llth Annual Conference on Trace Substances in Environmental Health. Columbia, MO: University of Missouri Press. Pp. 252-256. Kala, SV; Neely, MW; Kala, G; et al. (2000) The MRP2/cMOAT transporter and arsenic-glutathione complex formation are required for biliary excretion of arsenic. J Biol Chem 275(43):33404-33408. Kalia, K; Narula, GI; Kannan, GM; Flora, SJ. (2007) Effects of combined administration of captopril and DMSA on arsenite induced oxidative stress and blood and tissue arsenic concentration in rats. Comp Biochem Physiol C Toxicol Pharmacol 144:372-379. Kaltreider, RC; Pesce, CA; Ihnat, MA; et al. (1999) Differential effects of arsenic( ) and chromium(VI) on nuclear transcription factor binding. Mol Carcinog 25:219-229. Kaltreider, RC; Davis, AM; Lariviere, JP; Hamilton, JW. (2001) Arsenic alters the function of the glucocorticoid receptor as a transcription factor. Environ Health Perspect 109:245-251. Kamat, CD; Green, DE; Curilla, S; et al. (2005) Role of HIF signaling on tumorigenesis in response to chronic low- dose arsenic administration. Toxicolog Sci 86(2):248-257. Kann, S; Huang, MY; Estes, C; et al. (2005a) Arsenite-induced aryl hydrocarbon receptor nuclear translocation results in additive induction of phase I genes and synergistic induction of phase II genes. Mol Pharmacol 68:336- 346. Kann, S; Estes, C; Reichard, JF; et al. (2005b) Butylhydroquinone protects cells genetically deficient in glutathione biosynthesis from arsenite-induced apoptosis without significantly changing their prooxidant status. Toxicol Sci 87:365-384. Karagas, MR; Stukel, TA; Moriss, JS; et al. (2001) Skin cancer risk in relation to toenail arsenic concentrations in a U.S. population-based case-control study. Am JEpidemiol 153:559-565. Karagas, MR; Tosteson, TD; Morris, JS; et al. (2004) Incidence of transitional cell carcinoma of the bladder an arsenic exposure in New Hampshire. Cancer Causes Control 15:465-472. 165 DRAFT—DO NOT CITE OR QUOTE ------- Karasawas, N; Carcamo, JM; Stratis, G; Golde, DW. (2005) Vitamin C protects HL60 and U266 cells from arsenic toxicity. Blood 105:4004-4012. Kawaguchi, K; Oku, N; Rin, K; et al. (1996) Dimethylarsenics reveal DNA damage induced by superoxide anion radicals. Biol Pharm Bull 19:551-553. Kawata, K, Yokoo, H, Shimazaki, R and Okabe, S. (2007) Classification of heavy-metal toxicity by human DNA microarray analysis. Environ Sci Technol 41:3769-3774. Kayajanian, G. (2003) Arsenic, cancer, and thoughtless policy. EcotoxEnv Safety 55:139-142. Kenyon, E; Hughes, M; Levander, O. (1997) Influence of dietary selenium on the disposition of arsenate in female B6C3F1 mouse. J Toxicol Environ Health 51:279-299. Kenyon, EM; Hughes, MF. (2001) A concise review of the toxicity and carcinogenicity of dimethylarsinic acid. Toxicology 160(l-3):227-236. Kenyon, EM; Del Razo, LM; Hughes, MF. (2005a) Tissue distribution and urinary excretion of inorganic arsenic and its methylated metabolites in mice following acute oral administration of arsenate. Toxicol Sci 85(1):468^175. Kenyon, EM; Del Razo, LM; Hughes, MF; Kitchin, KT. (2005b) An integrated pharmacokinetic and pharmacodynamic study of arsenite action. 2. Heme oxygenase induction in mice. Toxicology 206:389-401. Kenyon, EM; Huges, MF; Adair, BM; et al. (2008) Tissue distribution and urinary excretion of inorganic arsenic and its methylated metabolites in C57BL6 mice following subchronic exposure to arsenate in drinking water. Toxicol Appl Pharmacol 232:448-455. Kessel, M; Liu, SX; Xu, A; et al. (2002) Arsenic induces oxidative DNA damage in mammalian cells. Mol Cell Biochem 234-235:301-308. Khalil, S; Luciano, J; Chen, W; Liu, AY. (2006) Dynamic regulation and involvement of the heat shock transcriptional response in arsenic carcinogenesis. J Cell Physiol 207:562-569. Kile, ML; Houseman, EA; Brenton, CV; et al. (2007) Dietary arsenic exposure in Bangladesh. Environ Health Perspect 115:889-893. Kimura, S; Yamauchi, H; Hibino, Y; et al. (2006) Evaluation of urinary 8-hydroxydeoxyguanine in healthy Japanese people. Basic Clin Pharmacol Toxicol 98:496-502. Kinoshita, A; Wanibuchi, H; Wei, M; et al. (2007a) Elevation of 8-hydroxydeoxyguanosine and cell proliferation via generation of oxidative stress by organic arsenicals contributes to their carcinogenicity in the rat liver and bladder. Toxicol Appl Pharmacol 221:295-305. Kinoshita, A; Wanibuchi, H; Morimura, K; et al. (2007b) Carcinogenicity of dimethylarsinic acid in Oggl-deficient mice. Cancer Sci 98:803-814. Kitchin, KT. (2001) Review. Recent advances in arsenic carcinogenesis: modes of action, animal model systems, and methylated arsenic metabolites. Toxicol Appl Pharmacol 172:249-261. Kitchin, KT; Ahmad, S. (2003) Oxidative stress as a possible mode of action for arsenic carcinogenesis. Toxicol Lett 137:3-13. Kitchin, KT; Wallace, R. (2006) Dissociation of arsenite-peptide complexes: triphasic nature, rate constants, half- lives, and biological importance. J Biochem Mol Toxicol 20(l):48-56. 166 DRAFT—DO NOT CITE OR QUOTE ------- Kitchin, KT; Wallace, K; Andrewes, P. (2003) Some chemical properties underlying arsenic's biological activity. In: Chappell, WR; Abernathy, CO; Calderon, RL; Thomas, DL; eds. Chapter 26 in arsenic exposure and health effects. Elserier Science B.V. Publishers. Klein, CB; Leszczynska, J; Hickey, C; Rossman, TG. (2007) Further evidence against a direct genotoxic mode of action for arsenic-induced cancer. Toxicol Appl Pharmacol 222:289-297. Kligerman, AD; Doerr, CL; Tennant, AH; et al. (2003) Methylated trivalent arsenicals as candidate ultimate genotoxic forms of arsenic: induction of chromosomal mutations but not gene mutations. Environ Mol Mutagen 42:192-205. Kligerman, AD; Doerr, CL; Tennant, AH. (2005) Oxidation and methylation status determine the effects of arsenic on the mitotic apparatus. Mol Cell Biochem 279:113-121. Kligerman, AD; Tennant, AH. (2007) Insights into the carcinogenic mode of action of arsenic. Toxicol Appl Pharmacol 222(3):281-288. Knobeloch, LM; Zierold, KM; Anderson, HA. (2006) Association of arsenic-contaminated drinking-water with prevalence of skin cancer in Wisconsin's Fox River Valley. J Health Popul Nutr 24(2):206-213. Kochhar, TS; Howard, W; Hoffman, S; Brammer-Carleton, L. (1996) Effect of trivalent and pentavalent arsenic in causing chromosome alterations in cultured Chinese hamster ovary (CHO) cells. Toxicol Lett 84:37-42. Kojima, C; Qu, W; Waalkes, MP; et al. (2006) Chronic exposure to methylated arsenicals stimulates arsenic excretion pathways and induces arsenic tolerance in rat liver cells. Toxicol Sci 91:70-81. Kokilavani, V; Devi, MA; Sivarajan, K; Panneerselvam, C. (2005) Combined efficacies of DL-a-lipoic acid and meso 2,3 dimercaptosuccinic acid against arsenic induced toxicity in antioxidant systems of rats. Toxicol Lett 160:1- 7. Komissarova, ET; Saha, SK; Rossman, TG. (2005) Dead or dying: the importance of time in cytotoxicity assays using arsenite as an example. Toxicol Appl Pharmacol 202:99-107. Konig, H; Hartel, N; Schultheis, B; et al. (2007) Enhanced Bcr-Abl-specific antileukemic activity of arsenic trioxide (Trisenox) through glutathione-depletion in imatinib-resistant cells. Haematologica. 92:838-841. Kreppel, H; Liu, J; Liu, Y; et al. (1994) Zinc-induced arsenite tolerance in mice. Fundam Appl Toxicol 23:32-37. Kurttio, P; Pukkala, E; Kahelin, H; et al. (1999). Arsenic concentration in well water and risk of bladder and kidney cancer in Finland. Environ Health Perspect 107:705-710. Lamm, SH; Byrd, DM; Kruse, MB; et al. (2003) Bladder cancer and arsenic exposure: Differences in the two populations enrolled in a study in southwest Taiwan. Biomed Environ Sci 16:355-368. Lamm, SH; Engel, A; Kruse, MB; et al. (2004) Arsenic in drinking water and bladder cancer mortality in the U.S.: an analysis based on 133 U.S. counties and thirty years of observation. J Occupation Environ Med 46:298-306 Lamm, SH; Engel, A; Perm, CA; et al. (2006) Arsenic cancer risk confounder in southwest Taiwan data set. Env Health Persp 114(7): 1077-1082. Lamm, S; Luo, Z-D; Bo, F-B; et al. (2007) An epidemiologic study of arsenic-related skin disorders and skin cancer and the consumption of arsenic contaminated well waters in Huhhot, Inner Mongolia, China. Hum Ecol Risk Assess 13:713-746. Lander, HM. (1997) An essential role for free radicals and derived species in signal transduction. FASEB J 11:118- 124. 167 DRAFT—DO NOT CITE OR QUOTE ------- Lantz, RC; Hays, AM. (2006) Role of oxidative stress in arsenic-induced toxicity. Drug Metab Rev 38:791-804. Lantz, RC; Lynch, BJ; Boitano, S; et al. (2007) Pulmonary biomarkers based on alterations in protein expression following exposure to arsenic. Environ Health Perspect 115:586-591. Lanz, H Jr; Wallace, PC; Hamilton, JG. (1950) The metabolism of arsenic in laboratory animals using As74 as a tracer. Univ Calif Publ Pharmacol 2:263-282. Laparra, JM, Velez, D, Barbera, R, et al. (2005a) An approach to As( ) and As(V) bioavailability studies with Caco-2 cells. Toxicol In Vitro 19:1071-1078. Laparra, JM; Velez, D; Barbera, R; et al. (2005b) Bioavailability of inorganic arsenic in cooked rice: practical aspects for human health risk assessments. J Agric Food Chem 53:8829-8833. Laparra, JM; Velez, D; Barbera, R; et al. (2006) Cytotoxic effect of As( ) in Caco-2 cells and evaluation of its human intestinal permeability. Toxicol iln Vitro 20:658-663. Larramendy, ML; Popescu, NC; DiPaolo, JA. (1981) Induction by inorganic metal salts of sister chromatid exchanges and chromosome aberrations in human and Syrian hamster strains. Environ Mutagen 3:597-606. Lau, ATY; Chiu, J-F. (2006) Proteomic and biochemical analyses of in vitro carcinogen-induced lung cell transformation: synergism between arsenic andbenzo[a]pyrene. Proteomics 6:1619-1630. Le, XC; Ma, M; Cullen WR; et al. (2000a) Determination of monomethylarsonous acid, a key arsenic methylation intermediate, in human urine. Environ Health Perspect 108(11): 1015-1018. Le, XC; Lu, X; Ma, M; et al. (2000b) Speciation of key arsenic metabolic intermediates in human urine. Anal Chem 72:5172-5177. Lee, T-C; Ho, I-C. (1995) Modulation of cellular antioxidant defense activities by sodium arsenite in human fibroblasts. Arch Toxicol 69:498-504. Lee, T-C; Huang, RY; Jan, KY. (1985) Sodium arsenite enhances the cytotoxicity clastogenicity and 6-thioguanine- resistant mutagenicity of ultraviolet light in Chinese hamster ovary cells. Mutat Res 148:83-89. Lee, T-C; Wang-Wuu, S; Huang, RY; et al. (1986) Differential effects of the pre- post-treatment of sodium arsenite on the genotoxicity of methylmethanesulfonate in Chinese hamster ovary cells. Cancer Res 46:1854-1857. Lee, C; Lee, YM; Rice, RH. (2005) Human epidermal cell protein responses to arsenite treatment in culture. Chem- Bio Interactions 155:43-54. Lee, T-C; Ho, I-C; Lu, W-J; Huang, J-D. (2006a) Enhanced expression of multidrug resistance-associated protein 2 and reduced expression of aquaglyceroporin 3 in an arsenic-resistant human cell line. J Biologic Chem 281(27): 18401-1840. Lee, C-H; Chen, J-S; Sun, Y-L; et al. (2006b) Defective pl-integrins expression in arsenical keratosis and arsenic- treated cultured human keratinocytes. J Cutan Pathol 33:129-138. Lee, M; Hwang, JT; Yun, H; et al. (2006c) Critical roles of AMP-activated protein kinase in the carcinogenic metal- induced expression of VEGF and HIF-1 proteins inDU145 prostate carcinoma. Biochem Pharmacol 72:91-103. Lee-Chen, SF; Gurr, JR; Lin, IB; Jan, KY. (1993) Arsenite exchanges DNA double-strand breaks and cell killing of methyl methanesulfonate-treated cells by inhibiting the excision of alkali-labile sites. Mutat Res 294(l):21-28. Lemarie, A; Morzadec, C; Merino, D; et al. (2006a). Arsenic trioxide induces apoptosis of human monocytes during macrophagic differentiation through nuclear factor-KB-related survival pathway down-regulation. J Pharmacol Exp Ther 316:304-314. 168 DRAFT—DO NOT CITE OR QUOTE ------- Lemarie, A; Morzadec, C; Bourdonnay, E; et al. (2006b). Human macrophages constitute targets for immunotoxic inorganic arsenic. J Immunol 177:3019-3027. Lerman, SA; Clarkson, TW. (1983) The metabolism of arsenite and arsenate by the rat. Fundam Appl Toxicol 3(4):309-314. Lerman, SA; Clarkson, TW; Gerson, RJ. (1983) Arsenic uptake and metabolism by liver cells is dependent on arsenic oxidation state. ChemBiol Interact 45:401-406. Leslie, EM; Haimeur, A; Waalkes, MP. (2004) Arsenic transport by the human multidrug resistance protein 1 (MRP1/ABCC1). JBiol Chem 279(31):32700-32708. Lewis, DR; Southwick, JW; Ouellet-Hellstrom, R; et al. (1999) Drinking water arsenic in Utah: a cohort-mortality study. Environ Health Perspect 107:359-365. Li, J-H; Rossman, TG. (1989a) Mechanism of comutagenesis of sodium arsenite with N-methyl-N-nitrosourea. Biol Trace Element Res 21:373-381. Li, J-H; Rossman, TG. (1989b) Inhibition of DNA ligase activity by arsenite: a possible mechanism of its comutagenesis. Mol Toxicol 2:1-9. Li, J-H; Rossman, TG. (1991) Comutagenesis of sodium arsenite with ultraviolet radiation in Chinese hamster V79 cells. Biol Met 4:197-200. Li, W; Chou, I-N. (1992) Effects of sodium arsenite on the cytoskeleton and cellular glutathione levels in cultured cells. Toxicol Appl Pharmacol 114:132-139. Li, YM; Broome, JD. (1999) Arsenic targets tubulins to induce apoptosis in myeloid leukemia cells. Cancer Res 59:776-780. Li, JJ; Oberley, LW; Fan, M; et al. (1998) Inhibition of AP-1 andNF-kappaB by manganese-containing superoxide dismutase in human breast cancer cells. FASEB J 12:1713-1723. Li, D; Morimoto, K; Takeshita, T; Lu, Y. (2001) Formamidopyrimidine-DNA glycosylase enhances arsenic-induced DNA strand breaks in PHA-stimulated and unstimulated human lymphocytes. Environ Health Perspect 109:523- 526. Li, J; Waters, SB; Drobna, Z; et al. (2005) Arsenic (+3oxidation state) methyltransferase and the inorganic arsenic methylation phenotype. Toxicol Appl Pharmacol 204(2): 164-169. Li, J-P; Lin, J-C; Yang, J-L. (2006a) ERK activation in arsenite-treated Gl-enriched CL3 cells contributes to survival, DNA repair inhibition, and micronucleus formation. Toxicol Sci 89:164-172. Li, Q; Chen, H; Huang, X; Costa, M. (2006b) Effects of 12 metal ions on iron regulatory protein 1 (IRP-1) and hypoxia-inducible factor-1 alpha (HIF-la) and HIF-regulated genes. Toxicol Appl Pharmacol 213:245-255. Li, Y; Chen, Y; Slavkovic, V; et al. (2007) Serum levels of the extracellular domain of the epidermal growth factor receptor in individuals exposed to arsenic in drinking water in Bangladesh. Biomarkers 12(3):256-265. Liao, WT; Chang, KL; Yu, CL; et al. (2004) Arsenic induces human keratinocyte apoptosis by the FAS/FAS ligand pathway, which correlates with alterations in nuclear factor-KB and activator protein-1 activity. J Invest Dermatol 122:125-129. Liebscher, K; Smith, H. (1968) Essential and nonessential trace elements. A method of determining whether an element is essential or nonessential in human tissue. Arch Environ Health 17:881-890. 169 DRAFT—DO NOT CITE OR QUOTE ------- Lin, S; Cullen, WR; Thomas, DJ. (1999) Methylarsenicals and arsinothiols are potent inhibitors of mouse liver thioredoxin reductase. ChemRes Toxicol 12:924-930. Lin, S; Del Razo, LM; Styblo, M; et al. (2001) Arsenicals inhibit thioredoxin reductase in cultured rat hepatocytes. ChemRes Toxicol 14(3):305-311. Lin, S; Shi, Q; Brent, F; et al. (2002) A novel S-adenosyl-L-methionine: arsenic( ) methyltransferase from rat liver cytosol. J Biol Chem 277(13): 10795-10803. Lin, YL; Ho, 1C; Su, PF; Lee, TC. (2006) Arsenite pretreatment enhances the cytotoxicity of mitomycin C in human cancer cell lines via increased NAD(P)H quinone oxidoreductase 1 expression. Toxicol Appl Pharmacol 214:309- 317. Lindberg, A-L; Kumar, R; Goessler, W; et al. (2007) Metabolism of low-dose inorganic arsenic in a central European population: influence of sex and genetic polymorphism. Environ Health Perspect 115:1081-1086. Lindberg, A-L; Ekstrom, E-C; Nerwell, B; et al. (2008) Gender and age differences in the metabolism of inorganic arsenic in a highly exposed population in Bangladesh. Environ Res 106:110-120. Lindgren, A; Vahter, M; Dencker, L. (1982) Autoradiographic studies on the distribution of arsenic in mice and hamster administered As-arsenite or -arsenate. Acta Pharmacol Toxicol 51:253-365. Lindgren, A; Danielsson, BRG; Dencker, L; Vahter, M. (1984) Embryotoxicity of arsenite and arsenate: distribution in pregnant mice and monkeys and effects on embryonic cells in vitro. Acta Pharmacol Toxicol 54:311-320. Lindskog, M; Gleissman, H; Ponthan, F; et al. (2006) Neuroblastoma cell death in response to docosahexaenoic acid: sensitization to chemotherapy and arsenic-induced oxidative stress. Int J Cancer 118:2584-2593. Ling, Y-H; Jiang, J-D; Holland, JF; Perez-Soler, R. (2002) Arsenic trioxide produces polymerization of microtubules and mitotic arrest before apoptosis in human tumor cell lines. Mol Pharmacol 62:529-538. Liou, SH; Lung, J-C; Chen, YH; et al. (1999) Increased chromosome-type chromosome aberration frequencies as biomarkers of cancer risk in a blackfoot endemic area. Cancer Res 59:1481-1484. Liu, SX; Athar, M; Lippai, I; etal. (2001) Induction of oxyradicals by arsenic: Implication for mechanism of genotoxicity. Proc Natl Acad Sci USA 98:1643-1648. Liu, J; Xie, Y; Ward, JM; et al. (2004) Toxicogenomic analysis of aberrant gene expression in liver tumors and nontumorous livers of adult mice exposed inutero to inorganic arsenic. Toxicol Sci 77:249-257. Liu, SX; Davidson, MM; Tang, X; et al. (2005) Mitochondrial damage mediates genotoxicity of arsenic in mammalian cells. Cancer Res 65:3236-3242. Liu, Z; Styblo, M; Rosen, BP. (2006a) Methylarsonous acid transport by aquaglyceroporins. Environ Health Persp 114(4):527-531. Liu, J; Xie, Y; Merrick, BA; et al. (2006b). Transplacental arsenic plus postnatal 12-O-teradecanoyl phorbol-13- acetate exposures associated with hepatocarcinogenesis induce similar aberrant gene expression patterns in male and female mouse liver. Toxicol Appl Pharmacol 213:216-223. Liu, J; Xie, Y; Ducharme, DMK; et al. (2006c) Global gene expression associated with hepatocarcinogenesis in adult male mice induced by inutero arsenic exposure. Environ Health Persp 114(3):404-411. Liu, J; Benbrahim-Tallaa, L; Qian, X; et al. (2006d) Further studies on aberrant gene expression associated with arsenic-induced malignant transformation in rat liver TRL1215 cells. Toxicol Appl Pharmacol 216:407-415. 170 DRAFT—DO NOT CITE OR QUOTE ------- Liu, B; Pan, S; Dong, X; et al. (2006e) Opposing effects of arsenic trioxide on hepatocellular carcinomas in mice. Cancer Sci 97:675-681. Liu, J; Xie, Y; Cooper, R; et al. (2007a) Transplacental exposure to inorganic arsenic at a hepatocarcinogenic dose induces fetal gene expression changes in mice indicative of aberrant estrogen signaling and disrupted steroid metabolism. Toxicol Appl Pharmacol 220:284-291. Liu, QZ; Jiang, GF; He, Y; et al. (2007b) Arsenite-induced alterations in Ku70-deficient cells: a model to study genotoxic effects. J Toxicol Environ Health A 70:938-946. Lo, MC; Hsen, YC; Lin, BK. (1977) The second report on the investigation of arsenic content in underground water in Taiwan Providence, Talchung. Provincial Institute of Environmental Sanitation. Loffredo, CA; Aposhian, HV; Cebrian, ME; et al. (2003) Variability in human metabolism of arsenic. Environ Res 92:85-91. Lowney, YW; Wester, RC; Schoof, RA; et al. (2007) Dermal absorption of arsenic from soils as measured in the rhesus monkey. Toxicol Sci 100(2):381-392. Lu, WJ; Tamai, I; Nezu, J; et al. (2006) Organic anion transporting polypeptide-C mediates arsenic uptake in HEK- 293 cells. J Biomed Sci 13(4):525-533. Lunghi, P; Tabilio, A; Lo-Coco, F; et al. (2005) Arsenic trioxide (ATO) and MEK1 inhibition synergize to induce apoptosis in acute promyelocytic leukemia cells. Leukemia 19:234-244. Lunghi, P; Costanzo, A; Salvatore, L; et al. (2006) MEK1 inhibition sensitizes primary acute myelogenous leukemia to arsenic trioxide-induced apoptosis. Blood 107:4549-4553. Luster, MI; Simeonova, PP. (2004) Review. Arsenic and urinary bladder cell proliferation. Toxicol Appl Pharmacol 198:419-423. Luster, MI; Wilmer, JL; Germolec, DR; et al. (1995) Role of keratinocyte-derived cytokines in chemical toxicology. Toxicol Lett 82-83:471-476. Lynn, S; Gurr, JR; Lai HT; Jan, KY. (2000) NADH oxidase activation is involved in arsenite-induced oxidative DNA damage in human vascular smooth muscle cells. Circ Res 86:514-519. Mahata, J; Ghosh, P; Sarkar, JN; et al. (2004) Effect of sodium arsenite on peripheral lymphocytes in vitro: individual susceptibility among a population exposed to arsenic through the drinking water. Mutagenesis 19(3):223- 229. Maier, A; Dalton TP; Puga, A. (2000) Disruption of dioxin-inducible phase I and phase II gene expression patterns by cadmium, chromium, and arsenic. Mol Carcinog 28:225-235. Maiorino, RM; Aposhian, HV. (1985) Dimercaptan metal-binding agents influence the biotransformation of arsenite in the rabbit. Toxicol Appl Pharmacol 77:240-250. Mandal, BK; Ogra, Y; Anzai, K; Suzuki, KT. (2004) Speciation of arsenic in biological samples. Toxicol Appl Pharmacol 198:307-318. Mann, S; Droz, PO; Vahter, M. (1996a) A physiologically based pharmacokinetic model for arsenic exposure: I. development in hamster and rabbits. Toxicol Appl Pharmacol 137:8-22. Mann, S; Droz, PO; Vahter, M. (1996b) A physiologically based pharmacokinetic model for arsenic exposure: II. validation and application in humans. Toxicol Appl Pharmacol 140:471^86. Mappes, R. (1977) Experiments on the excretion of arsenic in urine. Int Arch Occup Environ Health 40:267-272. 171 DRAFT—DO NOT CITE OR QUOTE ------- Marafante, E; Vahter, M. (1984) The effect of methyltransferase inhibition on the metabolism of ( As)arsenite in mice and rabbits. Chem-Biol Interact 50:49-57. Marafante, E; Vahter, M. (1987) Solubility, retention, and metabolism of intratracheally and orally administered inorganic arsenic compounds in the hamster. Environ Res 42:72-82. Marafante, E; Rade, J; Sabbioni, E; et al. (1981) Intracellular interaction and metabolic fate of arsenite in the rabbit. ClinToxicol 18:1335-1341. Marafante, E; Vahter, M; Envall, J. (1985) The role of methylation in the detoxication of arsenate in the rabbit. Chem-Biol Interact 56:225-238. Marafante, E; Vahter, M; Norin, H; et al. (1987). Biotransformation of dimethylarsinic acid in mouse, hamster, and man. J Appl Toxicol 7:111-117. Marcus, WL; Rispin, AS. (1988) Threshold carcinogenicity using arsenic as an example. In: Cothern, CR; Mehlman, MA; eds. Advances in modern environmental toxicology. Vol. XV. Risk assessment and risk management of industrial and environmental chemicals. Princeton, NJ: Princeton Scientific Publishing Co., Inc.; pp. 133-158. Marshall, G; Ferreccio, C; Yuan, Y; et al. (2007) Fifty-year study of lung and bladder cancer mortality in Chile related to arsenic in drinking water. J Natl Can Inst 99(12):920-928. Marsit, CJ; Karagas, MR; Danaee, H; et al. (2006) Carcinogen exposure and gene promoter hypermethylation in bladder cancer. Carcinogenesis 27:112-116. Mass, MJ; Wang, L. (1997) Arsenic alters cytosine methylation patterns of the promoter of the tumor suppressor gene p53 in human lung cells: a model for mechanism of carcinogenesis. Mutat Res 386:263-277. Mass, MJ; Tennant, A; Roop, BC; et al. (2001) Methylated trivalent arsenic species are genotoxic. Chem Res Toxicol 14:355-361. Massrieh, W; Derjuga, A; Blank, V. (2006) Induction of endogenous Nrf2/small Maf heterodimers by arsenic- mediated stress in placental choriocarcinoma cells. Antioxid Redox Signal 8:53-59. Mazumder, DNG; Haque, R; Ghosh, N; et al. (1998) Arsenic levels in drinking water and the prevalence of skin lesions in West Bengal, India. Int J Epidemiol 27:871-877. Mazumder, DNG. (2005) Effect of chronic intake of arsenic-contaminated water on liver. Toxicol Appl Pharmacol 206:169-175. McCarthy, KM; Chen, Y-C; Quamruzzaman, Q; et al. (2007a) Arsenic methylation, GSSTT1, GSTM1, GSTP1 polymorphisms, and skin lesions. Environ Health Perspect 115:341-345. McCarthy, KM; Ryan, L; Houseman, EA; et al. (2007b) A case-control study of GST polymorphisms and arsenic related skin lesions. Environ Health 6:5. McCollum, G; Keng, PC; States, JC; McCabe, MC Jr. (2005) Arsenite delays progression through each cell cycle phase and induces apoptosis following G2/M arrest in U937 myeloid leukemia cells. J Pharmacol Exp Ther 313:877-887. McNeely, SC; Xu, X; Taylor, BF; et al. (2006) Exit from arsenite-induced mitotic arrest is p53 dependent. Environ Health Perspect 114:1401-1406. Mealey, J; Brownwell, GL; Sweet, WH. (1959) Radioarsenic in plasma, urine, normal tissues, and intracranial neoplasms. Distribution and turnover after intravenous injection in man. AMA Arch Neurol Psychiatry 81(3):310- 320. 172 DRAFT—DO NOT CITE OR QUOTE ------- Meliker, JR; Wahl, RL; Cameron, LL; Nriagu, J. (2007) Arsenic in drinking water and cerebrovascular disease, diabetes mellitus, and kidney disease in Michigan: a standardized mortality ratio analysis. Environ Health 6:4. Page number not for citation purposes. Meng, Z; Hsie, AW. (1996) Polymerase chain reaction-based deletion analysis of spontaneous and arsenite- enhanced gpt mutants in CHO-AS52 cells. Mutat Res 356:255-259. Merck Index (1989) Budavari, S., M.J. O'Neil, A. Smith and P.E. Heckelman, eds. The Merck Index—an encyclopedia of chemicals, drugs, and biologicals, llth Ed. Rahway, NJ: Merck & Co., Inc.; pp.126-127, 8515. Meza, MM; Yu, L; Rodriguez, YY; et al. (2005) Developmentally restricted genetic determinants of human arsenic metabolism: Association between urinary methylated arsenic and CYT19 polymorphisms in children. Environ Health Persp 113(6):775-781. Meza, M; Gandolfi, AJ; Klimecki, WT. (2007) Developmental and genetic modulation of arsenic biotransformation. A gene by environment interaction? Toxicol Appl Pharmacol 222:381-387. Michaud, DS; Wright, ME; Cantor, KP; et al. (2004) Arsenic concentrations in prediagnostic toenails and the risk of bladder cancer in a cohort study of male smokers. Am J Epidemiol 160(9):853-859. Mitra, SR; Guha Mazumder, DN; Basu, A; et al. (2004) Nutritional factors and susceptibility to arsenic-caused skin lesions in West Bengal, India. Environ Health Perspect 112:1104-1109. Mittal, M; Flora, SJS. (2006) Effects of individual and combined exposure to sodium arsenite and sodium fluoride on tissue oxidative stress, arsenic and fluoride levels in male mice. Chem Biol Interact 162:128-139. Miyazaki, K; Watanabe, C; Mori, K; et al. (2005) The effects of gestational arsenic exposure and dietary selenium deficiency on selenium and selenoenzymes in maternal and fetal tissues in mice. Toxicology 208:357-365. Mizoi, M; Takabayashi, F; Nakano, M; et al. (2005) The role of trivalent dimethylated arsenic in dimethylarsinic acid-promoted skin and lung tumorigenesis in mice: tumor-promoting action through the induction of oxidative stress. Toxicol Lett 158:87-94. Mo, J; Xia, Y; Wade, TJ; et al. (2006) Chronic arsenic exposure and oxidative stress: OGG1 expression and arsenic exposure, nail selenium, and skin hyperkeratosis in inner Mongolia. Environ Health Persp 114(6):835-841. Modi, M; Kaul, RK; Kannan, GM; Flora, SJ. (2006) Co-administration of zinc and n-acetylcysteine prevents arsenic-induced tissue oxidative stress in male rats. J Trace Elem Med Biol 20:197-204. Molin, L; Wester, PO. (1976) The estimated daily loss of trace elements from normal skin by desquamation. Scand J ClinLab Invest 36:679-682. Mondal, TK; Li, D; Swami, K; et al. (2005) Mercury impairment of mouse thymocyte survival in vitro: involvement of cellular thiols. J Toxicol Environ Health A 68:535-556. Moore, LE; Warner, ML; Smith, AH; et al. (1996). Use of the fluorescent micronucleus assay to detect the genotoxic effects of radiation and arsenic exposure in exfoliated human epithelial cells. Environ Mol Mutagen 27:176-184. Moore, MM; Harrington-Brock, K; Doerr, CL. (1997a) Relative genotoxic potency of arsenic and its methylated metabolites. Mutat Res 386:279-290. Moore, LE; Smith, AH; Hopenhayn-Rich, C; et al. (1997b) Micronuclei in exfoliated bladder cells among individuals chronically exposed to arsenic in drinking water. Cancer Epidemiol Biomarkers Prev 6(1):31-36. Moore, LE; Lu, M; Smith, AH. (2002) Childhood cancer incidence and arsenic exposure in drinking water in Nevada. Arch Environ Health 57(3):201-206. 173 DRAFT—DO NOT CITE OR QUOTE ------- Morales, KH; Ryan, L; Kuo, T-L; et al. (2000) Risk of internal cancers from arsenic in drinking water. Environ Health Perspect 108:655-661. Moser, G; Goldsworthy, T; Tice, R. (2000). Sodium arsenite studies inp53 +/- mice. Denver, CO; AWWA Research Foundation and the American Water Works Association. Motiwale, L; Ingle, AD; Rao, KVK. (2005) Mouse skin tumor promotion by sodium arsenate is associated with enhanced PCNA expression. Cancer Lett 223:27-35. Mouron, SA; Grille, CA; Dulout, FN; Golijow, CD. (2005) DNA-protein cross-links and sister chromatid exchanges induced by dimethylarsinic acid in human fibroblasts cells. MutatRes 581:83-90. Mouron, SA; Grillo, CA; Dulout, FN; Golijow, CD. (2006) Induction of DNA strand breaks, DNA-protein crosslinks and sister chromatid exchanges by arsenite in a human lung cell line. Toxicol In Vitro 20:279-285. Mudipalli, A; Owen, RD; Preston, RJ. (2005) The effect of arsenicals on ultraviolet-radiation-induced growth arrest related signaling events in human keratinocytes. Int J Oncol 27:769-778. Mure, K; Uddin, A; Lopez, LC; et al. (2003) Arsenite induces delayed mutagenesis and transformation in human osteosarcoma cells at extremely low concentrations. Environ MolMutagen 41:322-331. Namgung, U; Xia, Z. (2001) Arsenic induces apoptosis in rat cerebellar neurons via activation of JNK3 and p38 MAP kinases. Toxicol Appl Pharmacol 174:130-138. Nandi, D; Patra, RC; Swamp, D. (2005) Effect of cysteine, methionine, ascorbic acid, and thiamine on arsenic- induced oxidative stress and biochemical alterations in rats. Toxicology 211:26-3 5. Nandi, D; Patra, RC; Swamp, D. (2006) Oxidative stress indices and plasma biochemical parameters during oral exposure to arsenic in rats. Food Chem Toxicol 44:1579-1584. Naranmandura, H; Suzuki, N; Iwata, K; et al. (2007) Arsenic metabolism and thioarsenicals in hamsters and rats. Chem Res Toxicol 20:616-624. NCI (National Cancer Institute). (2006) U.S. population data 2000-2003. Surveillance, epidemiology, and end results (SEER) program populations (1969-2003) DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2006. Available online at www.seer.cancer.gov/popdata. NCHS (National Center for Health Statistics). (2000) Vital statistics of the United States. NRC (National Research Council). (1983) Risk assessment in the federal government: managing the process. National Academy Press, Washington, DC. Available online at http://www.nap.edu:80/openbook.php?isbn=0309033497. NRC (National Research Council). (1999) Arsenic in drinking water. National Academy Press, Washington, DC. Available online at http://www.nap.edu/openbook/0309063337/html/Rl.html. NRC (National Research Council). (2001) Arsenic in drinking water (2001 update). National Academy Press, Washington, DC. Available online at http://www.nap.edu/openbook/0309076293/html/Rl.html. Nayak, AS; Lage, CR; Kim, CH. (2007) Effects of low concentrations of arsenic on the innate immune system of the zebrafish (Danio rerio). Toxicol Sci 98(1): 118-124. Nemeti, B; Gregus, Z. (2002) Reduction of arsenate to arsenite in hepatic cytosol. Toxicol Sci 70:4-12. Nemeti, B; Gregus, Z. (2004) Glutathione-dependent reduction of arsenate in human erythrocytes: a process independent of purine nucleoside phosphorylase. Toxicol Sci 82(2):419-428. 174 DRAFT—DO NOT CITE OR QUOTE ------- Nemeti, B; Gregus, Z. (2005) Reduction of arsenate to arsenite by human erythrocyte lysate and rat liver cytosol: characterization of a glutathione- and NAD-dependent arsenate reduction linked to glycolysis. Toxicol Sci 85(2):847-858. Nemeti, B; Csanaky, I; Gregus, Z. (2003) Arsenate reduction in human erythrocytes and rats: testing the role of purine nucleoside phosphorylase. Toxicol Sci 74:22-31. Nesnow, S; Roop, BC; Lambert, G; et al. (2002) DNA damage induced by methylated trivalent arsenicals is mediated by reactive oxygen species. ChemRes Toxicol 15:1627-1634. Ng, J; Seawright, S; Wi, L; et al. (1999) Tumours in mice induced by exposure to sodium arsenate in drinking water. In: Chappell, WR; Abernathy, CO; Cothern, CR; eds. Arsenic: exposure and health. Northwood, UK: Sci Tech Lett; pp. 217-223. Nordenson, I; Salmonsson, S; Bran, E; et al. (1979) Chromosome aberrations inpsoriatic patients treated with arsenic. Hum Genet 48:1-6. Nordenson, I; Beckman, L. (1991) Is the genotoxic effect of arsenic mediated by oxygen free radicals? Hum Hered 41:71-73. Nutt, LK; Gogvadze, V; Uthaisang, W; et al. (2005) Indirect effects of Bax and Bak initiate the mitochondria! alterations that lead to cytochrome c release during arsenic trioxide-induced apoptosis. Cancer Biol Ther 4:459-467. O'Brien, TG; Megosh, LC; Gillard, G; Soler, AP. (1997) Ornithine decarboxylase over expression is a sufficient condition for tumor promotion in mouse skin. Cancer Res 57:2630-2637. Ochi, T; Kaise, T; Oya-Ohta, Y. (1994) Glutathione plays different roles in the induction of the cytotoxic effects of inorganic and organic arsenic compounds in cultured BALB/c 3T3 cells. Experientia 50:115-120. Ochi, T; Nakajima, F; Nasui, M. (1999a) Distribution of gamma-tubulin in multipolar spindles and multinucleated cells induced by dimethylarsinic acid, a methylated derivative of inorganic arsenics, in Chinese hamster V79 cells. Toxicology 136:79-88. Ochi, T; Nakajima, F; Shimizu, A; Harada, M. (1999b) Induction of multinucleated cells in V79 Chinese hamster cells exposed to dimethylarsenic acid, a methylated derivative of inorganic arsenics: Mechanism associated with aberrant mitotic spindles. Toxicol In Vitro 13:11-25. Ochi, T; Suzuki, T; Barrett, JC; Tsutsui, T. (2004) A trivalent dimethylarsenic compound, dimethylarsine iodide, induces cellular transformation, aneuploidy, centrosome abnormality, and multipolar spindle formation in Syrian hamster embryo cells. Toxicol 203:155-163. Odanaka, Y; Matano, O; Goto, S. (1980) Biomethylation of inorganic arsenic by the rat and some laboratory animals. Bull Environ Contam Toxicol 24:452-459 O'Hara, KA; Nemec, AA; Alam, J; et al. (2006) Chromium (VI) inhibits heme oxygenase-1 expression in vivo and in arsenic-exposed human airway epithelial cells. J Cell Physiol 209:113-121. Ohno, H; Hanaoka, F; Yamada, M. (1982) Inducibility of sister chromatid exchanges by heavy-metal ions. Mutat Res 104(1-3): 141-146. Ostrosky-Wegman, P; Gonsebatt, ME; Montero, R; et al. (1991) Lymphocyte proliferation kinetics and genotoxic findings in a pilot study on individuals chronically exposed to arsenic in Mexico. Mutat Res 250:477-482. Othumpangat, S; Kashon, M; Joseph, P. (2005) Sodium arsenite-induced inhibition of eukaryotic translation initiation factor 4E (eIF4E) results in cytotoxicity and cell death. Mol Cell Biochem 279:123-131. 175 DRAFT—DO NOT CITE OR QUOTE ------- Ouyang, W; Ma, Q; Li, J; et al. (2005) Cyclin Dl induction through M3 Kinase P/Nuclear Factor-KB pathway is responsible for arsenite-induced increased cell cycle Gl-S phase transition in human keratinocytes. Cancer Res 65(20):9287-9293. Ouyang, W; Li, J; Ma, Q; Huang, C. (2006) Essential roles of PI-SK/Akt/IKKp/NFicB pathway in cyclinDl induction by arsenite in JB6 CI41 cells. Carcinogenesis 27:864-873. Ouyang, W; Zhang, D; Ma, Q; et al. (2007) Cyclooxygenase-2 induction by arsenite through the IKK(3/NFKB pathway exerts an antiapoptotic effect in mouse epidermal C141 cells. Environ Health Perspect 115:513-518. Oya-Ohta, Y; Kaise, T; Ochi, T. (1996) Induction of chromosomal aberrations in cultured human fibroblasts by inorganic and organic arsenic compounds and the different roles of glutathione in such induction. Mutat Res 375:123-129. Palus, J; Lewinska, D; Dziubaltowska, E; et al. (2006) Genotoxic effects in C57B1/6J mice chronically exposed to arsenate in drinking water and modulation of the effects by low-selenium diet. J Toxicol Environ Health A 69:1843- 1860. Park, WH; Seol, JG; Kim, ES; et al. (2000) Arsenic trioxide-mediated growth inhibition in MC/CAR myeloma cells via cell cycle arrest in association with induction of cyclin-dependent kinase inhibitor, p21, and apoptosis. Cancer Res 60:3065-3071. Parr, RM; De Maeyer, EM; lyengar, VG; et al. (1991) Minor and trace elements in human milk from Guatemala, Hungary, Nigeria, Philippines, Sweden, and Zaire: results from a WHO/IAEA joint project. Biol Trace Element Res 29:51-75. Partridge, MA; Huang, SX; Hernandez-Rosa, E; et al. (2007) Arsenic induced mitochondrial DNA damage and altered mitochondrial oxidative function: implications for genotoxic mechanisms in mammalian cells. Cancer Res 67:5239-5247. Patra, M; Haider, A; Bhowmik, N; De, M. (2005) Use of black tea in modulating clastogenic effects of arsenic in mice in vivo. J Environ Pathol Toxicol Oncol 24:201-210. Patterson, TJ; Rice, RH. (2007) Arsenite and insulin exhibit opposing effects on epidermal growth factor receptor and keratinocyte proliferative potential. Toxicol Appl Pharmacol 221:119-128. Patterson, TJ; Reznikova, TV; Phillips, MA; Rice, RH. (2005) Arsenite maintains germinative state in cultured human epidermal cells. Toxicol Appl Pharmacol 207(l):69-77. Peraza, MA; Cromey, DW; Carolus, B; et al. (2006) Morphological and functional alterations in human proximal tubular cell line induced by low level inorganic arsenic: evidence for targeting of mitochondria and initiated apoptosis. J Appl Toxicol 26:356-367. Perez-Pasten, R; Martinez-Galero, E; Garduno-Siciliano, L; et al. (2006) Effects of dimethylsulphoxide on mice arsenite-induced dysmorphogenesis in embryo culture and cytotoxicity in embryo cells. Toxicol Lett 161:167-173. Perkins, C; Kim, CN; Fang, G; Bhalla, KN. (2000) Arsenic induces apoptosis of multi-drug-resistant human myeloid leukemia cells that express Ber-Abl or over express MDR, Bel-2, or Bel-x(L). Blood 95:1014-1022. Petrick, JS; Ayala-Fierro, F; Cullen, WR; et al. (2000) Monomethylarsonous acid (MMA ) is more toxic than arsenite in Chang human hepatocytes. Toxicol Appl Pharmacol 163:203-207. Petrick, JS; Jagadish, B; Mash, EA; Aposhian, HV. (2001). Monomethylarsonous acid (MMA ) and arsenite: LD50 in hamsters and in vitro inhibition of pyruvate dehydrogenase. Chem Res Toxicol 14:651-656. 176 DRAFT—DO NOT CITE OR QUOTE ------- Pi, J; He, Y; Bornter, C; et al. (2005) Low level, long- term inorganic arsenite exposure causes generalized resistance to apoptosis in cultured human keratinocytes: potential role in skin co-carcinogenesis. Int J Cancer 116:20-26. Piga, R; Saito, Y; Yoshida, Y; Niki, E. (2007) Cytotoxic effects of various stressors onPC12 cells: involvement of oxidative stress and effect of antioxidants. Neurotoxicology 28:67-75. Pomroy, C.; Charbonneau, SM; McCullough, RS; Tarn, GKH (1980) Human retention studies with 74As. Toxicol Appl Pharmacol 53:550-556. Poonepalli, A; Balakrishnan, L; Khaw, AK; et al. (2005) Lack of Poly(ADP-Ribose) Polymerase-1 gene product enhances cellular sensitivity to arsenite. Cancer Res 65:10977-10983. Preston, PJ; Williams, GM. (2005) DNA-reactive carcinogens: mode of action and human cancer hazard. Crit Rev Toxicol 35:673-683. Pu, YS; Jan, KY; Wang, TC; et al. (2007) 8-Oxoguanine DNA glycosylase and MutY homolog are involved in the incision of arsenite-induced DNA adducts. Toxicol Sci 95:376-382. Qian, Y; Liu, KJ; Chen, Y; et al. (2005) Cdc42 regulates arsenic-induced NADPH oxidase activation and cell migration through actin filament reorganization. J Biol Chem 280:3875-3884. Radabaugh, TR; Aposhian, HV. (2000) Enzymatic reduction of arsenic compounds in mammalian systems: reduction of arsenate to arsenite by human liver arsenate reductase. Chem Res Toxicol 13(1)26-30. Radabaugh, TR; Sampayo-Reyes, A; Zakharyan, RA; Aposhian, HV. (2002) Arsenate reductase II. Purine nucleoside phosphorylase in the presence of dihydrolipoic acid as a route for reduction of arsenate to arsenite in mammalian systems. Chem Res Toxicol 15(5):692-698. Rahman, MS; Hall, LL; Hughes, MF. (1994) In vitro percutaneous absorption of sodium arsenate in B6C3F1 mice. Toxicol In Vitro 8:441-448. Ramanathan, K; Anusuyadevi, M; Shila, S; Panneerselvam, C. (2005) Ascorbic acid and a-tocopherol as potent modulators of apoptosis on arsenic induced toxicity in rats. Toxicol Lett 156:297-306. Ramirez, P; Eastmond, DA; Laclette, JP; Ostrosky-Wegman, P. (1997) Disruption of microtubule assembly and spindle formation for the induction of aneuploid cells by sodium arsenite and vanadium pentoxide. Mutat Res 386:291-298. Ramirez, P; Del Razo, LM; Gutierrez-Ruiz, MC; Gonsebatt, ME. (2000) Arsenite induces DNA-protein crosslinks and cytokeratin expression in the WRL-68 human hepatic cell line. Carcinogenesis 21:701-706. Ramirez, T; Stopper, H; Hock, R; Herrera, LA. (2007) Prevention of aneuploidy by S-adenosyl-methionine in human cells treated with sodium arsenite. Mutat Res 617:16-22. Raml, A; Goessler, W; Francesconi, KA. (2006) Improved chromatographic separation of thio-arsenic compounds by reverse-phase high performance liquid chromatography-inductively coupled plasma mass spectrometry (HPLC- ICPMS). J Chromatogr A 1128:164-170. Raml, R; Rumpler, A; Goessler, W; et al. (2007) Thio-dimethylarsenite is a common metabolite in urine samples from arsenic-exposed women in Bangladesh. Toxicol Appl Pharmacol 222:374-380. Ramos, AM; Fernandez, C; Amran, D; et al. (2005) Pharmacologic inhibitors of PI3K/Akt potentiate the apoptotic action of the antileukemic drug arsenic trioxide via glutathione depletion and increased peroxide accumulation in myeloid leukemia cells. Blood 105:4013-4020. 177 DRAFT—DO NOT CITE OR QUOTE ------- Rana, SVS; Allen, T. (2006) Influence of thyroxine and n-propylthiouracil on nephro-toxicity of inorganic arsenic in rat. Toxicol Ind Health 22:137-145. Rasmussen, RE; Menzel, DB. (1997) Variation in arsenic-induced sister chromatid exchange in human lymphocytes and lymphoblastoid cell lines. Mutat Res 386:299-306. Reichard, JF; Schnekenburger, M; Puga, A. (2007) Long-term low-dose arsenic exposure induces loss of DNA methylation. Biochem Biophys Res Commun 352:188-192. Rizki, M; Kossatz, E; Velazquez, A; et al. (2006) Metabolism of arsenic in Drosophila melanogaster and the genotoxicity of dimethylarsinic acid in the Drosophila wing spot test. Environ Mol Mutagen 47:162-168. Rossman, TG. (1981) Enhancement of UV-mutagenesis by low concentrations of arsenite inE. coli. Mutat Res 91:207-211. Rossman, TG. (1998) Molecular and genetic toxicology of arsenic. In: Rose, J, ed. Environmental toxicology: current developments. Amsterdam, NL: Gorden and Breach Science Publishers; pp. 171-187. Rossman, TG. (2003) Review: Mechanism of arsenic carcinogenesis: an integrated approach. Mutat Res 533:37-65. Rossman, TG; Wolosin, D. (1992) Differential susceptibility to carcinogen-induced amplification of SV40 and dhfr sequences in SV40-transformed humankeratinocytes. Mol Carcinog 6:203-213. Rossman, TG; Stone, D; Molina, M; Troll, W. (1980) Absence of arsenite mutagenicity inE. coli and Chinese hamster cells. Environ Mutagen 2:371-379. Rossman, TG; Molina, M; Meyer, LW. (1984) The genetic toxicology of metal compounds: I. Induction of I prophage inE.coli WP2s (I). Environ Mutagen 6:59-69. Rossman, TG; Uddin, AN; Burns, FJ; et al. (2001) Arsenic is a co-carcinogen with solar ultraviolet radiation for mouse skin: an animal model for arsenic carcinogenesis. Toxicol Appl Pharmacol 176:64-71. Rousselot, P; Labaume, S; Marolleau, J; et al. (1999) Arsenic trioxide and melarsoprol induce apoptosis in plasma cell lines and in plasma cells from myeloma patients. Cancer Res 59:1041-1048. Saady, JJ; Blanke, RV; Poklis, A. (1989) Estimation of the body burden of arsenic in a child fatally poisoned by arsenite weedkiller. J Anal Toxicol 13:310-312. SAB (Science Advisory Board). (2000) Arsenic proposed drinking water regulation: a U.S. EPA Science Advisory Board review of certain elements of the proposal: a report by the U.S. EPA Science Advisory Board. December 2000. EPA/SAB/DWC/01/001. SAB (Science Advisory Board). (2007) Advisory onEPA's asssessments of carcinogenic effects of organic and inorganic arsenic: a report of the U.S. EPA Science Advisory Board. June 2007. Available online at http://yosemite.epa.gov/sab/SABPRODUCT.NSF/EAD ABBF40DED2A0885257308006741EF/$File/sab-07- 008.pdf. Sahu, GR; Jena, RK. (2005) Significance of intracellular arsenic trioxide for therapeutic response in acute promyelocytic leukemia. Am J Hematol 78:113-116. Sakurai, T; Kaise, T; Matsubara, C. (1998) Inorganic and methylated arsenic compounds induce cell death in murine macrophages via different mechanisms. Chem Res Toxicol 11:273-283. Sakurai, T; Ochiai, M; Kojima, C; et al. (2005a). Preventive mechanism of cellular glutathione in monomethylarsonic acid-induced cytolethality. Toxicol Appl Pharmacol 206:54-65. 178 DRAFT—DO NOT CITE OR QUOTE ------- Sakurai, T; Ohta, T; Fujiwara, K. (2005b) Inorganic arsenite alters macrophage generation from human peripheral blood monocytes. Toxicol Appl Pharmacol 203:145-153. Sakurai, T; Ohta, T; Tomita, N; et al. (2006) Evaluation of immunotoxic and immunodisruptive effects of inorganic arsenite on human monocytes/macrophages. Int Immunopharmacol 6:304-315. Salazar, AM; Ostrosky, WP; Menendez, D; et al. (1997) Induction of p53 protein expression by sodium arsenite. Mutat Res 381:259-265. Sampayo-Reyes, A; Zakharyan, RA; Healy, SM; Aposhian, HV. (2000) Monomethylarsonic acid reductase and monomethylarsonic acid in hamster tissue. Chem Res Toxicol 13(11): 1181-1186. Sams, RII; Wolf, DC; Ramasamy S; Ohanian E; Chen J; Lowit A. (2007) Workshop overview: Arsenic research and risk assessment. Toxicol Appl Pharmacol 222: 245-251. Samuel, S; Kathirvel, R; Jayavelu, T; Chinnakkannu, P. (2005) Protein oxidative damage in arsenic induced rat brain: influence of DL-a-lipoic acid. Toxicol Lett 155:27-34. Sasaki, A; Oshima, Y; Fujimura, A. (2007) An approach to elucidate potential mechanism of renal toxicity of arsenic trioxide. Exp Hematol 35:252-262. Schoof, RA; Yost, LJ; Crecelius, C; et al. (1998) Dietary arsenic intake in Taiwanese districts with elevated arsenic in drinking water. Hum Ecol Risk Assess 4:117-135. Sciandrello, G; Barbara, R; Caradonna, F; et al. (2002) Early induction of genetic instability and apoptosis by arsenic in cultured Chinese hamster cells. Mutagenesis 17:99-103. Sciandrello, G; Caradonna, F; Mauro, M; Barbata, G. (2004) Arsenite-induced DNA hypomethylation affects chromosomal instability in mammalian cells. Carcinogenesis 25(3):413^U7. Scott, N; Hatlelid, KM; MacKenzie, NE; Carter, DE. (1993) Reactions of arsenic( ) and arsenic(V) species with glutathione. Chem Res Toxicol 6:102-106. Scott, A. (1958) The retention of arsenic in the late cutaneous complications of its administration. Br J Dermatol 70:195-200. Sen, CK; Parker, L. (1996) Antioxidant and redox regulation of gene transcription. FASEB J 10:709-720. Sen, B; Wang, A; Hester, SD; et al. (2005) Gene expression profiling of responses to dimethylarsinic acid in female F344 rat urothelium. Toxicology 215:214-226. Seol, JG; Park, WH; Kim, ES; et al. (1999) Effect of arsenic trioxide on cell cycle arrest in head and neck cancer cell line PCI-1. Biochem Biophys Res Commun 265:400-404. Shackelford, RE; Kaufmann, WK; Paules, RS. (2000) Oxidative stress and cell cycle checkpoint function. Free Radic Biol Med 28:1387-1404. Shen, J; Wanibuchi, H; Salim, El; et al. (2003) Liver tumorigenicity of trimethylarsine oxide in male Fischer 344 rats-association with oxidative DNA damage and enhanced cell proliferation. Carcinogen 24(11): 1827-1835. Shen, S; Lee, J; Sun, X; et al. (2006) Elevation of cellular BPDE uptake by human cells: a possible factor contributing to co-carcinogenicity by arsenite. Environ Health Perspect 114:1832-1837. Shen, J; Liu, J; Xie, Y; et al. (2007) Fetal onset of aberrant gene expression relevant to pulmonary carcinogenesis in lung adenocarcinoma development induced by inutero arsenic exposure. Toxicol Sci 95:313-320. 179 DRAFT—DO NOT CITE OR QUOTE ------- Shila, S; Subathra, M; Devi, MA; Panneerselvam, C. (2005a) Arsenic intoxication-induced reduction of glutathione level and of the activity of related enzymes in rat brain regions: reversal by DL-a-lipoic acid. Arch Toxicol 79:140- 146. Shila, S; Kokilavani, V; Subathra, M; Panneerselvam, C. (2005b). Brain regional responses in antioxidant system to a-lipoic acid in arsenic intoxicated rat. Toxicol 210:25-36. Shinkai, Y; Sumi, D; Fukami, I; et al. (2006) Sulforaphane, an activator of Nrf2, suppresses cellular accumulation of arsenic and its cytotoxicity in primary mouse hepatocytes. FEES Lett 580:1771-1774. Shiobara Y; Ogra, Y; Suzuki, KT. (2001) Animal species difference in the uptake of dimethylarsinous acid (DMA( )) by red blood cells. ChemRes Toxicol 14(10): 1446-1452. Simeonova, PP; Wang, S; Toriuma, W; et al. (2000) Arsenic mediates cell proliferation and gene expression in the bladder epithelium: association with activating protein-1 transactivation. Cancer Res 60:3445-3453. Simeonova, PP; Luster MI. (2000) Mechanisms of arsenic carcinogenicity: genetic orepigenetic mechanisms? Environ Pathol Toxicol Oncol 19:281-286. Singh, KP; DuMond, JW, Jr. (2007) Genetic and epigenetic changes induced by chronic low dose exposure to arsenic of mouse testicular Leydig cells. January 2007. Int J Oncol 30(1):253-260. Sinha, D; Bhattacharya, RK; Siddiqi, M; Roy, M. (2005a) Amelioration of sodium arsenite-induced clastogenicity by tea extracts in Chinese hamster V79 cells. J Environ Pathol Toxicol Oncol 24:129-140. Sinha, D; Roy, M; Siddiqi, M; Bhattacharya, RK. (2005b) Arsenic-induced micronuclei formation in mammalian cells and its counteraction by tea. J Environ Pathol Toxicol Oncol 24:45-56. Smith, AH; Goycolea, M; Haque, R; Biggs, ML. (1998) Marked increase in bladder and lung cancer mortality in a region of Northern Chile due to arsenic in drinking water. Am J Epidemiol 147(7):660-669. Smith, KR; Klei, LR; Barchowsky, A. (2001) Arsenite stimulates plasma membrane NADPH oxidase in vascular endothelial cells. Am J Physiol Lung Cell Mol Physiol 280:L442-L449. Smith, AH; Marshall, G; Yuan, T; et al. (2006) Increased mortality from lung cancer and bronchiectasis in young adults following exposure to arsenic in utero and early childhood. Environ Health Perspect 114(8): 1-4. Snow, ET; Hu, Y; Yan, CC; Chouchane, S. (1999) Modulation of DNA repair and glutathione levels in human keratinocytes by micromolar arsenite. In: Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic exposure and health effects. Oxford, UK: Elsevier; pp. 243-251. Snow, ET; Schuliga, M; Chouchane, S; Hu, Y. (2001) Sub-toxic arsenite induces a multi-component protective response against oxidative stress in human cells. Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic exposure and health effects. Amsterdam, NL: Elsevier; pp. 265-276. Snow, ET; Sykora, P; Durham, TR; Klein, CB. (2005) Arsenic, mode of action at biologically plausible low doses: what are the implications for low dose cancer risk? Toxicol Appl Pharmacol 207:8557-8564. Soffritti, M; Belpoggi, F; Degli, D; et al. (2006) Results of a long-term carcinogenicity bioassay on Sprague-Dawley rats exposed to sodium arsenite administered in drinking water. Ann NY Acad Sci 1076:578-591. Sohini; Rana, SV. (2007) Protective effect of ascorbic acid against oxidative stress induced by inorganic arsenic in liver and kidney of rat. Indian J Exp Biol 45:371-375. Soto-Reyes, E; del Razo, LM; Valverde, M; Rojas, E. (2005) Role of the alkali labile sites, reactive oxygen species and antioxidants in DNA damage induced by methylated trivalent metabolites of inorganic arsenic. BioMetals 18:493-506. 180 DRAFT—DO NOT CITE OR QUOTE ------- Soucy, NV; Ihnat, MA; Kamat, CD; et al. (2003) Arsenic stimulates angiogenesis and tumorigenesis in vivo. Toxicol Sci 76:271-279. Soucy, NV; Mayka, D; Klei, LR; et al. (2005) Neovascularization and angiogenic gene expression following chronic arsenic exposure in mice. Cardiovasc Toxicol 5:29-41. Steinmaus, C; Yuan, Y; Bates, MN; Smith AH. (2003) Case-control study of bladder cancer and drinking water arsenic in the western United States. Am J Epidemiol 158(12): 1193-1201. Steinmaus, C; Carrigan, K; Kalman, D; et al. (2005) Dietary intake and arsenic methylation in a U.S. population. Environ Health Persp 113(9): 1153-1159. Steinmaus, C; Bates, MN; Yuan, Y; et al. (2006) Arsenic methylation and bladder cancer risk in case-control studies in Argentina and the United States. J Occup Environ Med 48:478-488. Steinmaus, C; Moore, LE; Shipp, M; et al. (2007) Genetic polymorphisms inMTHFR 677 and 1298, GSTM1 and Tl, and metabolism of arsenic. J Toxicol Environ Health A 70:159-170. Stepnik, M; Staficzyk, M; Arkusz, J; Lewifiska, D. (2005) Assessment of apoptosis inthymocytes and splenocytes from mice exposed to arsenate in drinking water: cytotoxic effects of arsenate on the cells in vitro. J Environ Sci Health A40:369-384. Stevens, JT; Hall, LL; Farmer, JD; et al. (1977) Disposition of 14C and/or As-cacodylic acid in rats after intravenous, intratracheal orperoral administration. Environ Health Perspect 19:151-157. Styblo, M; Yamauchi, H; Thomas, DJ. (1995) Comparative in vitro methylation of trivalent and pentavalent arsenicals. Toxicol Appl Pharmacol 135:172-178. Styblo, M; Delnomdedieu, M; Thomas, DJ. (1996) Mono- and dimethylation of arsenic in rat liver cytosol in vitro. Chem Biol Interact 99:147-164. Styblo, M; Vega, L; Germolec, DR; et al. (1999a) Metabolism and toxicity of arsenicals in cultured cells. In: Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic exposure and health effects. Oxford, UK: Elsevier; pp. 311-323. Styblo, M; Del Razo, LM; LeCluyse, EL; et al. (1999b) Metabolism of arsenic in primary cultures of human and rat hepatocytes. Chem Res Toxicol 12(7):560-565. Styblo, M; Del Razo, LM; Vega, L; et al. (2000) Comparative toxicity of trivalent and pentavalent inorganic and methylated arsenicals in rat and human cells. Arch Toxicol 74(6):289-299 Su, P-F; Hu, Y-J; Ho, I-C; et al. (2006) Distinct gene expression profiles in immortalized human urothelial cells exposed to inorganic arsenite and its methylated trivalent metabolites. Environ Health Perspect 114:394-403. Sumi, D; Taguchi, K ; Sun, Y ; et al. (2005) Monomethylarsonous acid inhibits endothelial nitric oxide synthase activity. J Health Science 51:728-730. Sun, G; Xu, Y; Li, X; et al. (2007) Urinary arsenic metabolites in children and adults exposed to arsenic in drinking water in Inner Mongolia, China. Environ Health Persp 115(4):648-652. Szymczyk, KH; Kerr, BA; Freeman, TA; et al. (2006) Involvement of hydrogen peroxide in the differentiation and apoptosis of preosteoclastic cells exposed to arsenite. Biochem Pharmacol 72:761-769. Tabacova, S; Baird, DD; Balabaeva, L; et al. (1994) Placental arsenic and cadmium in relation to lipid peroxides and glutathione levels in maternal-infant pairs from a copper smelter area. Placenta 15:873-881. 181 DRAFT—DO NOT CITE OR QUOTE ------- Takahashi, M; Barrett, JC; Tsutsui, T. (2002) Transformation by inorganic arsenic compounds of normal Syrian hamster embryo cells into a neoplastic state in which they become anchorage-independent and cause tumors in newborn hamsters. Int J Cancer 99:629-634. Tarn, GKH; Charbonneau, SM; Bryce, F; et al. (1979) Metabolism of inorganic arsenic ( As) in humans following oral ingestion. Toxicol Appl Pharmacol 50:319-322. Tang, F; Liu, G; He, Z; et al. (2006) Arsenite inhibits p53 phosphorylation, DNA binding activity, and p53 target gene p21 expression in mouse epidermal JB6 cells. Mol Carcinog 45:861-870. Tenorio, EP; Saavedra, R. (2005) Differential effect of sodium arsenite during the activation of human CD4+ and CD8+ T lymphocytes. Int Immunopharmacol 5:1853-1869. Terek, MC; Karabulut, B; Selvi, N; et al. (2006) Arsenic trioxide-loaded, microemulsion-enhanced cytotoxicity on MDAH 2774 ovarian carcinoma cell line. Int J Gynecol Cancer 16(2):532-537. Tezuka, M; Hanioka, K; Yamanaka, K; Okada, S. (1993) Gene damage induced in human alveolar type II (L-132) cells by exposure to dimethylarsenic acid. Biochem Biophys Res Commun 191:1178-1183. Thomas, DJ. (2007) Molecular processes in cellular arsenic metabolism. Toxicol Appl Pharmacol 222:365-373. Thomas, DJ; Styblo, M; Lin, S. (2001) The cellular metabolism and systemic toxicity of arsenic. Toxicol Appl Pharmacol 176:127-144. Thomas, DJ; Waters, SB; Styblo, M. (2004) Elucidating the pathway for arsenic methylation. Toxicol Appl Pharmacol 198(3):319-326. Thomas, DJ; Li, J; Waters, SB; et al. (2007) Arsenic(+3 oxidative state) methyltransferase and methylation of arsenicals. Exp Biol Med 232:3-13. Trouba, KJ; Wauson, EM; Vorce, RL. (2000) Sodium arsenite inhibits terminal differentiation of murine C3H 10T1/2 preadipocytes. Toxicol Appl Pharmacol 168:25-35. Trouba, KJ; Germolec, DR. (2004) Micromolar concentrations of sodium arsenite induce cyclooxygenase-2 expression and stimulate p42/44 mitogen-activated protein kinase phosphorylation in normal human epidermal keratinocytes. Toxicol Sci 79:248-257. Tsai, S-M; Wang, T-N; Ko, Y-C. (1999) Mortality for certain diseases in areas with high levels of arsenic. Arch Environ Health 54(3): 186-193. Tsai, YS; Yang, W-H; Tong, T-C; et al. (2005) Experience with primary urethral carcinoma from the blackfoot disease-endemic area of south Taiwan: increased frequency of bulbomembranous adenocarcinoma? Urol Int 74:229- 234. Tseng, WP. (1977) Effects and dose-response relationships of skin cancer and blackfoot disease with arsenic. Environ Health Perspect 19:109-119. Tseng, WP; Chu, HM; How, SW; et al. (1968) Prevalence of skin cancer in an endemic area of chronic arsenicism in Taiwan. J Natl Cancer Inst 40(3):453-463. Tseng, C-H; Huang, Y-K; Huang, Y-L; et al. (2005) Arsenic exposure, urinary arsenic speciation, and peripheral vascular disease in blackfoot-disease-hyperendemic villages in Taiwan. Toxicol Appl Pharmacol 206:299-308. Tseng, YS; Tzeng, CC; Huang, CY; et al. (2006) Aurora-A overexpression associates with Ha-ras codon-12 mutation and blackfoot disease endemic area in bladder cancer. Cancer Lett 241:93-101. 182 DRAFT—DO NOT CITE OR QUOTE ------- Tsuchiya, T; Tanaka-Kagawa, T; Jinno, H; et al. (2005) Inorganic arsenic compounds and methylated metabolites induce morphological transformation in two-stage B ALB/c 3T3 cell assay and inhibit metabolic cooperation in V79 cell assay. Toxicolog Sci 84:344-351. Tsuda, T; Babazono, A; Yamamoto, E; et al. (1995) Ingested arsenic and internal cancer: a historical cohort study followed for 33 years. Am J Epidemiol 141:198-209. Uchino, T; Roychowdhury, T; Ando, M; Tokunaga, H. (2006) Intake of arsenic from water, food composites and excretion through urine, hair from a studied population in West Bengal, India. Food Chem Toxicol 44(4):455-61. Uddin, AN; Burns, FJ; Rossman, TG. (2005) Vitamin E and organoselenium prevent the cocarcinogenic activity of arsenite with solar UVR in mouse skin. Carcinogenesis 26:2179-2186. Upreti, RK; Kannan, A; Pant, AB. (2007) Experimental exposure of arsenic in cultured rat intestinal epithelial cells and cell line: lexicological consequences. Toxicol In Vitro 21:32^0. U.S. EPA (Environmental Protection Agency). (1984) Health assessment document for inorganic arsenic. Final report. Office of Health and Environmental Assessment; EPA/600/8-83-021. U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk assessment of chemical mixtures. FedRegist 51(185):34014-34025. U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk assessment. Fed Regist 51(185):34006-34012. U.S. EPA (Environmental Protection Agency). (1988a) Recommendations for and documentation of biological values for use in risk assessment. EPA/600/6-87/008. Available from the National Technical Information Service, Springfield, VA; PB88-179874/AS, February 1988. U.S. EPA (Environmental Protection Agency). (1988b) Special report on ingested inorganic arsenic: skin cancer; nutritional essentiality. Risk Assessment Forum, Washington, DC; EPA/625/3-87/013. U.S. EPA (1989) Report on Arsenic (As) Work Group meetings. Memorandum. From: Abernathy, CO; Marcus, W; Office of Drinking Water; and Chen, C; Gibb, H; White, P; Office of Research and Development; to Cook, P; Office of Drinking Water; and Preuss, P; Office of Regulatory Support and Scientific Management. February 23. U.S. EPA (Environmental Protection Agency). (1991) Guidelines for developmental toxicity risk assessment. Fed Regist 56(234):63798-63826. Available online at http://www.epa.gov/ncea/raf/rafguid.htm. U.S. EPA (Environmental Protection Agency). (1994) Integrated Risk Information System (IRIS) on inorganic arsenic. Verification date February 3, 1994.Cincinnati, OH. Accessed on June 26, 2008. U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose approach in health risk assessment. EPA/630/R-94/007. U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive toxicity risk assessment. Fed Regist 61(212):56274-56322. Available online at http://www.epa.gov/ncea/raf/rafguid.htm. U.S. EPA (Environmental Protection Agency). (1998) Guidelines for neurotoxicity risk assessment. Fed Regist 63(93):26926-26954. Available online at http://www.epa.gov/ncea/raf/rafguid.htm. U.S. EPA (Environmental Protection Agency). (2000a) Science policy council handbook: peer review. 2nd ed. Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-00/001. U.S. EPA (Environmental Protection Agency). (2000b) Science policy council handbook: risk characterization. Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-00/002. 183 DRAFT—DO NOT CITE OR QUOTE ------- U.S. EPA (Environmental Protection Agency). (2000c) Benchmark dose technical support document. External review draft. Office of Research and Development, Risk Assessment Forum, Washington, DC; EPA/630/R-00/001, October 2000. U.S. EPA (Environmental Protection Agency). (2000d) Supplementary guidance for conducting health risk assessment of chemical mixtures. Office of Research and Development, Risk Assessment Forum, Washington, DC; EPA /630/R-00/002. U.S. EPA (Environmental Protection Agency). (2001) National primary drinking water regulations: arsenic and clarifications to compliance and new source contaminants monitoring. Fed Regist 66(14):6975-7022, January 2001. U.S. EPA (Environmental Protection Agency). (2002) A review of the reference dose concentration and reference concentration processes. Risk Assessment Forum, Washington, DC; EPA/630/P-02/002F. Available online at http ://www. epa. gov/ncea/raf. U.S. EPA (Environmental Protection Agency). (2004) Estimated per capita water ingestion and body weight in the United States—an update based on data collected by the United States Department of Agriculture's 1994-1996 and 1998 continuing survey of food intakes by individuals. EPA Office of Water; EPA/822/R-00/001, October 2004. U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk assessment. Office of Research and Development, Risk Assessment Forum, Washington, DC; EPA/630/P-03/001F. U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for assessing susceptibility from early-life exposure to carcinogens. Risk Assessment Forum; EPA/630/R-03/003F, March 2005. U.S. EPA (Environmental Protection Agency). (2005c) Toxicological review of inorganic arsenic internal review draft. Health Effects Criteria Division. Available online at http://epa.gov/waterscience/criteria/arsenic/sab. U.S. EPA (Environmental Protection Agency). (2005d). Issue paper: inorganic arsenic cancer CSF. Arsenic Cancer CSF Workgroup. (Submitted to SAB) U.S. EPA (Environmental Protection Agency). (2006a) Science policy council handbook: peer review. 3rd ed. Office of Science Policy, Office of Research and Development, Washington, DC; EPA/100/B-06/002. Available online at http ://www. epa. gov/OS A/spc/2peerrev. htm. U.S. EPA (Environmental Protection Agency). (2006b) A framework for assessing health risk of environmental exposures to children. National Center for Environmental Assessment, Washington, DC; EPA/600/R-05/093F. Available online at http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=158363. U.S. EPA (Environmental Protection Agency). (2006c). Revised re-registration eligibility decision document for MSMA, DSMA, CAMA, and cacodylic acid. Office of Pesticides, Prevention, and Toxic Substances; EPA/738-R- 06-021. U.S. EPA (Environmental Protection Agency). (2008). Human health risk assessment and ecological risk assessment for the reregistration eligibility. Decision document for inorganic arsenicals and chromium-based wood preservatives. Office of Pesticides, Prevention, and Toxic Substances; EPA/739/R-08/006. Vahter, M. (1981) Biotransformation of trivalent and pentavalent inorganic arsenic in mice and rats. Environ Res 25:286-293. Vahter, M. (1983) Metabolism of inorganic arsenic in relation to chemical form and animal species. Doctoral Thesis. Karolinska Institutet, Stockholm. Vahter, M. (1994) Species differences in the metabolism of arsenic compounds. Appl Organomet Chem 8:175-182. Vahter, M. (1999a) Methylation of inorganic arsenic in different mammalian species and population groups. Sci Prog82(Pt.l):69-88. 184 DRAFT—DO NOT CITE OR QUOTE ------- Vahter, M. (1999b) Variation in human metabolism of arsenic. In: Chappell, WR; Abernathy, CO; Calderon, RL; eds. Arsenic exposure and health effects. Oxford, UK: Elsevier; pp. 267-279. Vahter, M; Envall, J. (1983) In vivo reduction of arsenate in mice and rabbits. Environ Res 32:14-24. Vahter, M; Marafante, E. (1983) Intracellular interaction and metabolic fate of arsenite and arsenate in mice and rabbits. ChemBiol Interact 47:29^4. Vahter, M; Marafante, E. (1985) Reduction and binding of arsenate in marmoset monkeys. Arch Toxicol 57:119- 124. Vahter, M; Marafante, E. (1987) Effects of low dietary intake of methionine, choline or proteins on the biotransformation of arsenite in the rabbit. Toxicol Lett 37:41^6. Vahter, M; Marafante, E; Lindgren, A; Dencker, L. (1982) Tissue distribution and subcellular binding of arsenic in marmoset monkeys after injection of As-arsenite. Arch Toxicol 51:65-77. Vahter, M; Marafante, E; Dencker, L. (1984) Tissue distribution and retention of As-dimethylarsinic acid in mice and rats. Arch Environ Contam Toxicol 13:259-264. Vahter, M; Friberg L; Rahnster B; et al. (1986) Airborne arsenic and urinary excretion of metabolites of inorganic arsenic among smelter workers. Int Arch Occup Environ Health 57:79-91. Vahter, M; Couch, R; Nermell, B; Nilsson, R. (1995a) Lack of methylation of inorganic arsenic in the chimpanzee. Toxicol Appl Pharmacol 133:262-268. Vahter, M; Concha, G; Nermell, B; et al. (1995b) A unique metabolism of inorganic arsenic in native Andean women. Eur J Pharmacol 293:455-462. Valentine, JL; Kang, HK; Spivey, G. (1979) Arsenic levels in human blood, urine, and hair in response to exposure via drinking water. Environ Res 20:24-32. Valenzuela, OL; Borja-Aburto, VH; Garcia-Vargas, GG; et al. (2005) Urinary trivalent methylated arsenic species in a population chronically exposed to inorganic arsenic. Environ Health Persp 113(3):250-254. Valenzuela, OL; Germolec, DR; Borja-Aburto, VH; et al. (2007) Chronic arsenic exposure increases TGF-alpha concentration in bladder urothelial cells of Mexican populations environmentally exposed to inorganic arsenic. Toxicol Appl Pharmacol 222:264-270. Vega, L; Gonsebatt, ME; Ostrosky-Wegman, P. (1995) Aneugenic effect of sodium arsenite on human lymphocytes in vitro: an individual susceptibility effect detected. Mutat Res 334:365-373. Vega, L; Styblo, M; Patterson, R; et al. (2001) Differential effects of trivalent and pentavalent arsenicals on cell proliferation and cytokine secretion in normal human epidermal keratinocytes. Toxicol Appl Pharmacol 172:225- 232. Vellar, OD. (1969) Nutrient losses through sweating. Thesis. Oslo University, Oslo, Norway. Venson DJ; Moolgavkar AH. (1988). A method for computing profile-likelihood based confidence intervals. Appl Stat 37:87-94. Vig, BK; Figueroa, ML; Cornforth, MN; Jenkins, SH. (1984) Chromosome studies in human subjects chronically exposed to arsenic in drinking water. Am J Ind Med 6(5):325-338. Vogt, BL; Rossman, TG. (2001) Effects of arsenite onp53, p21 and cyclinD expression in normal human fibroblasts—a possible mechanism for arsenite's comutagenicity. Mutat Res 478:159-168. 185 DRAFT—DO NOT CITE OR QUOTE ------- Waalkes, MP; Ward, JM; Liu, J; Diwan, BA. (2003) Transplacental carcinogenicity of inorganic arsenic in the drinking water: induction of hepatic, ovarian, pulmonary, and adrenal tumors in mice. Toxicol Appl Pharmacol 186:7-17. Waalkes, MP; Ward, JM; Diwan, BA. (2004a) Induction of tumors of the liver, lung, ovary and adrenal in adult mice after brief maternal gestational exposure to inorganic arsenic: promotional effects of postnatal phorbol ester exposure on hepatic and pulmonary, but not dermal cancers. Carcinogenesis 25(1): 133-141. Waalkes, MP; Liu, J; Ward, JM; Diwan, BA. (2004b) Animal models for carcinogenesis: inorganic arsenic is a transplacental carcinogen in mice. Toxicol Appl Pharmacol 198:377-384. Waalkes, MP; Liu, J; Ward, J; et al. (2006a) Urogenital carcinogenesis in female GDI mice induced by in utero arsenic exposure is exacerbated by postnatal diethylstilbestrol treatment. Cancer Res 66:1337-1345. Waalkes, MP; Liu, J; Ward, JM; Diwan, BA. (2006b). Enhanced urinary bladder and liver carcinogenesis in male GDI mice exposed to transplacental inorganic arsenic and postnatal diethylstilbestrol or tamoxifen. Toxicol Appl Pharmacol 215:295-305. Walter, I; Schwerdtle, T; Thuy, C; et al. (2007) Impact of arsenite and its methylated metabolites on PARP-1 activity, PARP-1 gene expression and poly(ADP-ribosyl)ation in cultured human cells. DNA Repair (Amst) 6:61- 70. Wan, B; Christian, RT; Soukup, SW. (1982) Studies of cytogenetic effects of sodium arsenicals on mammalian cells in vitro. Environ Mutag 4:493-498. Wang, TS; Huang, H. (1994) Active oxygen species are involved in the induction of micronuclei by arsenite in XRS- 5 cells. Mutagenesis 9:253-257. Wang, TS; Kuo, CF; Jan, KY; Huang, H. (1996) Arsenite induces apoptosis in Chinese hamster ovary cells by generation of reactive oxygen species. J Cell Physiol 169:256-268. Wang, TS; Shu, YF; Liu, YC; et al. (1997) Glutathione peroxidase and catalase modulate the genotoxicity of arsenite. Toxicol 121:229-237. Wang, TS; Hsu, TY; Chung, CH; et al. (2001) Arsenite induces oxidative DNA adducts and DNA-protein crosslinks in mammalian cells. Free Radical Biol Med 31:321-330. Wang, XJ; Yang, J; Gang, H; et al. (2005) Gene expression alteration during redox-dependent enhancement of arsenic cytotoxicity by emodin inHeLa cells. Cell Res 15:511-522. Wang, TC; Jan, KY; Wang, AS; Gurr, JR. (2007) Trivalent arsenicals induce lipid peroxidation, protein carbonylation, and oxidative DNA damage in human urothelial cells. Mutat Res 615:75-86. Warner, ML; Moore, LE; Smith, MT; et al. (1994) Increased micronuclei in exfoliated bladder cells of individuals who chronically ingest arsenic-contaminated water in Nevada. Cancer Epidemiol Biomarkers Prev 3:583-590. Waters, SB; Devesa, V; Fricke, MW; et al. (2004a) Glutathione modulates recombinant rat arsenic (+3 oxidation state) methyltransferase-catalyzed formation of trimethylarsine oxide and trimethylarsine. Chem Res Toxicol 17(12):1621-1629. Waters, SB; Devesa, V; Del Razo, LM; et al. (2004b) Endogenous reductants support the catalytic function of recombinant rat cyt!9, an arsenic methyltransferase. Chem Res Toxicol 17:404-409. Webb, DR; Sipes, IG; Carter, DE (1984) In vitro solubility and in vivo toxicity of gallium arsenide. Toxicol Appl Pharmacol 76:96-104. 186 DRAFT—DO NOT CITE OR QUOTE ------- Wei, M; Wanibuchi, H; Morimura, K; et al. (1999) Urinary bladder carcinogenicity of dimethylarsinic acid in male F344 rats. Carcinogenesis 20(9): 1873-1876. Wei, M; Wanibuchi, H; Morimura, K; et al. (2002) Carcinogenicity of dimethylarsinic acid in male F344 rats and genetic alterations in induced urinary bladder tumors. Carcinogenesis 23(8): 1387-1397. Wei, M; Arnold, L; Cano, M; Cohen, SM. (2005) Effects of co-administration of antioxidants and arsenicals on the rat urinary bladder epithelium. Toxicol Sci 83:237-245. Wester, RC; Maibach, HI; Sedik, L; et al. (1993) In vitro and in vivo percutaneous absorption and skin decontamination of arsenic from water and soil. Fundam Appl Toxicol 20:336-340. Wester, RC; Hui, X; Barbadillo, S; et al. (2004) In vivo percutaneous absorption of arsenic from water and CCA- treated wood residue. Toxicol Sci 79:287-295. Wetzler, M; Earp, JC; Brady, MT; et al. (2007) Synergism between arsenic trioxide and heat shock protein 90 inhibitors on signal transducer and activator of transcription protein 3 activity—pharmacodynamic drug-drug interaction modeling. Clin Cancer Res 13:2261-2270. Wiencke, JK; Yager, JW. (1992) Specificity of arsenite in potentiating cytogenic damage induced by the DNA crosslinking agent diepoxybutane. Environ Mol Mutagen 19:195-200. Wong, R-H; Kuo, C-Y; Hsu, M-L; et al. (2005) Increased levels of 8-hydroxy-2'-deoxyguanosine attributable to carcinogenic metal exposure among schoolchildren. Environ Health Perspect 113:1386-1390. Wu, MM; Kuo, TL; Hwang, YH; Chen CJ. (1989) Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases. Am J Epidemiol 130:1123-1132. Wu, F; Burns, FJ; Zhang, R; et al. (2005) Arsenite-induced alterations of DNA photodamage repair and apoptosis after solar-simulation UVR in mouse keratinocytes in vitro. Environ Health Perspect 113:983-986. Xie, Y; Liu, J; Benbrahim-Tallaa, L; et al. (2007) Aberrant DNA methylation and gene expression in livers of newborn mice transplacentally exposed to a hepatocarcinogenic dose of inorganic arsenic. Toxicology 236:7-15. Yamanaka, K; Hayashi, H; Tachikawa, M; et al. (1997) Metabolic methylation is a possible genotoxicity-enhancing process of inorganic arsenics. MutatRes 394(13):95-101. Yamanaka, K; Mizoi, M; Tachikawa, M; et al. (2003) Oxidative DNA damage following exposure to dimethylarsinous iodide: the formation of cis-thymine glycol. Toxicol Lett 143:145-153. Yamauchi, H; Yamamura, Y. (1979) Dynamic change of inorganic arsenic and methylarsenic compounds in human urine after oral intake as arsenic trioxide. Ind Health 17:79-84. Yamauchi, H; Yamamura, Y. (1984) Metabolism and excretion of orally administered dimethylarsinic acid in the hamster. Toxicol Appl Pharmacol 74:134-140. Yamauchi, H; Takahashi, K; Yamamura, Y. (1986) Metabolism and excretion of orally and intraperitoneally administered gallium arsenide in the hamster. Toxicology 40:237-246. Yamauchi, H; Aminaka, Y; Yoshida, K; et al. (2004) Evaluation of DNA damage in patients with arsenic poisoning: urinary 8-hydroxydeoxyguanine. Toxicol Appl Pharmacol 198:291-296. Yancy, SL; Shelden, EA; Gilmont, RR; Welsh, MJ. (2005) Sodium arsenite exposure alters cell migration, focal adhesion localization and decreases tyrosine phosphorylation of focal adhesion kinase in H9C2 myoblasts. Toxicol Sci 84:278-286. 187 DRAFT—DO NOT CITE OR QUOTE ------- Yang, JL; Chen, MF; Wu, CW; Lee, TC. (1992) Posttreatment with sodium arsenite alters the mutation spectrum induced by ultraviolet light irradiation in Chinese hamster ovary cells. Environ Mol Mutagen 20:156-164. Yang, CH; Kuo, ML; Chen, JC; Chen, YC. (1999) Arsenic trioxide sensitivity is associated with low level of glutathione in cancer cells. Br J Cancer 81:796-799. Yang, CY; Chiu, H-F; Wu, T-N; et al. (2004) Reduction in kidney cancer mortality following installation of a tap water supply system in an arsenic-endemic area of Taiwan. Arch Environ Health 59(9):484-488. Yang, CY; Chiu, HF; Chang, C-C; et al. (2005) Bladder cancer mortality reduction after installation of a tap-water supply system in an arsenious-endemic area in southwestern Taiwan. Environ Res 98:127-132. Yang, P; He, XQ; Peng, L; et al. (2007) The role of oxidative stress in hormesis induced by sodium arsenite in human embryo lung fibroblast (HELP) cellular proliferation model. J Toxicol Environ Health A 70:976-983. Yates, CR; Krynetski, EY; Loennechen, T; et al. (1997) Molecular diagnosis of thiopurine S-methyltransferase deficiency: genetic basis for azathioprine and mercaptopurine intolerance. Ann Intern Med 126:608-614. Ye, J; Li, A; Liu, Q; et al. (2005) Inhibition of mitogen-activated protein kinase kinase enhances apoptosis induced by arsenic trioxide in human breast cancer MCF-7 cells. Clin Exp Pharmacol Physiol 32:1042-1048. Yedjou, CG; Tchounwou, PB. (2007) In-vitro cytotoxic and genotoxic effects of arsenic trioxide on human leukemia (HL-60) cells using the MTT and alkaline single cell gel electrophoresis (comet) assays. Mol Cell Biochem 301:123-130. Yi, J; Yang, J; He, R; et al. (2004) Emodin enhances arsenic trioxide-induced apoptosis via generation of reactive oxygen species and inhibition of survival signaling. Cancer Res 64:108-116. Yih, J-H; Lee, T-C. (1999) Effects of exposure protocols on induction of kinetochore-plus and -minus micronuclei by arsenite in diploid human fibroblasts. Mutat Res 440:75-82. Yih, LH; Hsueh, SW; Luu, WS; et al. (2005) Arsenite induces prominent mitotic arrest via inhibition of G2 checkpoint activation in CGL-2 cells. Carcinogenesis 26:53-63. Yoshida, K; Inoue, Y; Kuroda, K; et al. (1998) Urinary excretion of arsenic metabolites after long-term oral administration of various arsenic compounds to rats. J Toxicol Environ Health 54(3): 179-192. Yost, LJ; Schoof, RA; Aucoin, R. (1998) Intake of inorganic arsenic in the North American diet. Hum Ecol Risk Assess 4:137-152. Yu, D. (1999a) A physiologically based pharmacokinetic model of inorganic arsenic. Regul Toxicol Pharmacol 29(2 Ptl):128-141. Yu, D. (1999b) A pharmacokinetic modeling of inorganic arsenic: a short-term oral exposure model for humans. Chemosphere 39(15):2737-2747. Yu, H-S; Chang, K-L; Yu, C-L; et al. (1998) Defective IL-2 receptor expression in lymphocytes of patients with arsenic-induced Bowen's disease. ArchDermatol Res 290:681-687. Yu, RC; Hsu, K-H; Chen, C-J; Froines, JR. (2000) Arsenic methylation capacity and skin cancer. Cancer Epidemiol, Biomarkers Prev 9:1259-1262. Yuan, Y; Marshall, G; Ferreccio, C; et al. (2007) Acute myocardial infarction mortality in comparison with lung and bladder cancer mortality in arsenic-exposed Region II of Chile from 1950 to 2000. Am J Epidemiol Advance Access 1-11. 188 DRAFT—DO NOT CITE OR QUOTE ------- Zakharyan, RA; Aposhian, HV. (1999) Enzymatic reduction of arsenic compounds in mammalian systems: the rate- limitin 1283. limiting enzyme of rabbit liver arsenic biotransformation is MMA reductase. Chem Res Toxicol 12(12): 1278- Zakharyan, RA; Wu, Y; Bogdan, GM; Aposhian, HV. (1995) Enzymatic methylation of arsenic compounds: assay, partial purification, and properties of arsenite methyltransferase and monomethylarsonic acid methyltransferase of rabbit liver. Chem Res Toxicol 8:1029-1038. Zakharyan, RA; Wildfang, E; Aposhian, HV. (1996) Enzymatic methylation of arsenic compounds: . The marmoset and tamarin, but not the rhesus, monkeys are deficient in methyltransferases that methylate inorganic arsenic. Toxicol Appl Pharmacol 140:77-84. Zakharyan, RA; Ayala-Fierro, F; Cullen, WR; et al. (1999) Enzymatic methylation of arsenic compounds. VII. Monomethylarsonous acid (MMA ) is the substrate for MMA methyltransferase of rabbit liver and human hepatocytes. Toxicol Appl Pharmacol 158(1):9-15. Zakharyan, RA; Sampayo-Reyes, A; Healy, SM; et al. (2001) Human monomethylarsonic acid (MMA(V)) reductase is a member of the glutathione-S-transferase superfamily. Chem Res Toxicol 14(8): 1051-1057. Zhang, T-C; Cao, EH; Li, JF; et al. (1999) Induction of apoptosis and inhibition of human gastric cancer MGC-803 cell growth by arsenic trioxide. Eur J Cancer 35:1258-1263. Zhang, T-C; Schmitt, MT; Mumford, JL. (2003) Effects of arsenic on telomerase and telomeres in relation to cell proliferation and apoptosis in human keratinocytes and leukemia cells in vitro. Carcinogenesis 24(11): 1811-1817. Zhang, Y; Lu, Y; Ding, M; et al. (2005) Deficiency in Ikk(3 gene enhances arsenic-induced gadd45a expression. Mol Cell Biochem 279:163-168. Zhang, Y; Bhatia, D; Xia, H; et al. (2006) Nucleolin links to arsenic-induced stabilization of GADD45a mRNA. Nucleic Acids Res 34:485^95. Zhao, CQ; Young, MR; Diwan, BA; et al. (1997) Association of arsenic-induced malignant transformation with DNA hypomethylation and aberrant gene expression. Proc Natl Acad Sci USA 94:10907-10912. Zheng, PZ ; Wang, K-K; Zhang, Q-Y; et al. (2005) Systems analysis of transcriptome and proteome in retinoic acid / arsenic trioxide-induced cell differentiation / apoptosis of promyelocytic leukemia. Proc Natl Acad Sci USA 102:7653-7658. Zhong, W; Oberley, LW; Oberley, TD; Clair, DK. (1997) Suppression of the malignant phenotype of human glioma cells by overexpression of manganese superoxide dismutase. Oncogene 14:481^190. Zhong, CX; Mass, MJ. (2001) Both hypomethylation and hypermethylation of DNA associated with arsenite exposure in cultures of human cells identified by methylation-sensitive arbitrary-primed PCR. Toxicol Lett 122:223- 234. 189 DRAFT—DO NOT CITE OR QUOTE ------- APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC COMMENTS AND DISPOSITION 1 The Toxicological Review of Inorganic Arsenic has been formally reviewed by scientists 2 outside EPA—i.e., the SAB Arsenic Review Panel—in accordance with EPA guidance on peer 3 review (U.S. EPA, 2000a). The reviewers on the Panel were tasked with providing written 4 answers to general questions on the overall assessment and on chemical-specific charge 5 questions, addressing key scientific issues of the assessment. While the Panel was supplied with 6 questions regarding both DMAV and inorganic arsenic, this appendix addresses only questions 7 and responses pertaining to inorganic arsenic. Charge question B3 asked SAB to comment on 8 EPA's hypothesis that inorganic arsenic acts via different modes of action for carcinogenicity. 9 SAB agreed with EPA's conclusion, but during a discussion on the mode of action of DMAV, a 10 member of the Panel stated that the description for inorganic arsenic's mode of action could be 11 strengthened. In addition to strengthening the mode of action discussion, studies on the mode of 12 action for inorganic arsenic have been placed in a table in Appendix C. Section 4.4.1 provides a 13 summary of the specifics in the tables instead of detailed write-ups for all the studies. A 14 summary of significant comments made by the external reviewers and EPA's responses to these 15 comments arranged by charge question follow. Public comments were submitted to SAB and 16 were taken into consideration by the Panel during their review. The summary of significant 17 comments and responses below is inclusive of the major issues raised by public commenters 18 which specifically focused on the choice of study for cancer quantitation and the nature of the 19 dose-response. Editorial comments were considered and incorporated into the document as 20 appropriate and are not discussed further. 21 22 Charge Question B3 23 24 EPA concluded that inorganic arsenic mostly likely causes human cancer by many 25 different modes of action. This is based on the observed findings that inorganic arsenic 26 undergoes successive methylation steps in humans and results in the production of a number of 27 intermediate metabolic products and that each has its own toxicity. EPA asked SAB to comment 28 on the soundness of its conclusion. 29 30 SAB Comments 31 32 The Panel concluded that: 33 1) Multiple modes of action may operate in carcinogenesis induced by inorganic 34 arsenic because there is simultaneous exposure to multiple metabolic products A-1 DRAFT—DO NOT CITE OR QUOTE ------- 1 as well as multiple target organs and the composition of metabolites can differ 2 in different organs. 3 2) Each arsenic metabolite has its own cytotoxic and genotoxic capability. 4 3) Inorganic arsenic (iAs111) and its metabolites are not direct genotoxicants because 5 these compounds do not directly react with DNA. However, iAs111 and some of its 6 metabolites can exhibit indirect genotoxicity, induce aneuploidy, cause changes in 7 DNA methylation, and alter signaling and hormone action. In addition, inorganic 8 arsenic can act as a transplacental carcinogen and a cocarcinogen. 9 4) Studies of indirect genotoxicity strongly suggest the possibility of a threshold for 10 arsenic carcinogenicity. However, the studies discussed herein do not show 11 where such a threshold might be, nor do they show the shape of the dose-response 12 curve at these low levels. In addition, a threshold has not been confirmed by 13 epidemiological studies. This issue is an extremely important area for research 14 attention, and it is an issue that should be evaluated in EPA's continuing risk 15 assessment for inorganic arsenic. 16 5) Arsenic essentiality and the possibility of hormetic effects are in need of 17 additional research to determine how they would influence the determination of 18 a threshold for specific arsenic-associated health endpoints. 19 EPA Response 20 21 EPA agrees that the available data potentially support multiple modes of action for 22 inorganic arsenic. The Agency believes that, at this point, the data concerning mode of action 23 are not well-enough understood to support their use in quantitative risk assessment. 24 25 Charge Question C2 26 27 EPA reviewed the available epidemiologic studies, including those published since the 28 NRC 2001 review, for U.S. populations exposed to inorganic arsenic via drinking water. EPA 29 concluded that the Taiwanese data set remains the most appropriate choice for estimating cancer 30 risk in humans. SAB was asked to comment on the soundness of this conclusion and also on 31 whether these data provide adequate characterization of the impact of childhood exposure to 32 inorganic arsenic. 33 34 SAB Comments 35 36 The Panel concluded that: 37 38 1) Because of various factors (e.g., the size and statistical stability of the 39 Taiwanese database relative to other studies, the reliability of the population 40 and mortality counts, the stability of residential patterns, and the inclusion of 41 long-term exposures), this database remains, at this time, the most appropriate 42 choice for estimating bladder cancer risk among humans, though the data have 43 considerable limitations that should be described qualitatively or quantitatively 44 to help inform risk managers about the strength of the conclusions. A-2 DRAFT—DO NOT CITE OR QUOTE ------- 1 2) There are other epidemiologic databases from studies of populations also 2 exposed at high levels of arsenic, and the panel recommends that these be used 3 to compare the unit risks at the higher exposure levels that have emerged from 4 the Taiwan data. 5 3) The panel also suggests that published epidemiology studies of U.S. and other 6 populations chronically exposed from 0.5 to 160 ug/L inorganic arsenic in 7 drinking water be critically evaluated, using a uniform set of criteria, and that the 8 results from these evaluations be transparently documented in EPA's assessment 9 documents. If, after this evaluation, one or more of these studies are shown to be 10 of potential utility, the low-level studies and Taiwan data may be compared for 11 concordance. Comparative analyses could lead to further insights into the 12 possible influence of these differences on population responses to arsenic in 13 drinking water. 14 4) Regarding childhood exposure to inorganic arsenic, it was the Panel's view that, 15 based on available data, it is not clear whether children differ from adults with 16 regard to their sensitivity to the carcinogenic effects of arsenic in drinking water. 17 However, the possibility of a different response in degree or kind should not be 18 ignored and needs to be investigated. 19 20 EPA Response 21 22 After considering additional studies, EPA agreed with SAB that the Taiwanese data were 23 the best available for quantitative analysis. Studies assessed, but not used in the analysis, are 24 summarized in Section 4.1 of the document. The studies were systematically evaluated for their 25 suitability in risk assessment based on a uniform set of criteria including the study type, the size 26 of the study population and control population, and the relative strengths and weaknesses of the 27 study based on SAB-recommended criteria (i.e., estimates of the level of exposure 28 misclassification; temporal variability in assigning past arsenic levels from recent measurements; 29 the extent of reliance on imputed exposure levels; the number of persons exposed at various 30 estimated levels of waterborne arsenic; study response/participation rates; estimates of exposure 31 variability; control selection methods in case-control studies; and the resulting influence of these 32 factors on the magnitude and statistical stability of cancer risk estimates). Study summaries are 33 also provided in tabular form in Appendix B for ease of comparison. Studies are arranged 34 geographically and include other areas of high arsenic exposure (e.g., South America) as well as 35 areas of low exposure (e.g., U.S. and Europe). Studies examining children were evaluated and 36 are discussed in Section 4.7.1 of the document, but EPA believes that the available data do not 37 yet allow a definitive conclusion on children's differential susceptibility to arsenic exposure. 38 EPA notes that recent animal studies demonstrating the potential for cancer after in utero arsenic 39 exposures give rise to additional concerns regarding exposures early in development. 40 A-3 DRAFT—DO NOT CITE OR QUOTE ------- 1 Charge Question D2 2 3 EPA determined that the most prudent approach for modeling cancer risk from inorganic 4 arsenic is to use a linear model because of the remaining uncertainties regarding the ultimate 5 carcinogenic metabolites and whether mixtures of toxic metabolites interact at the site(s) of 6 action. EPA asked SAB if it concurred with the selection of a linear model following the 7 recommendations of the NRC (2001) to estimate cancer risk in light of the multiple modes of 8 carcinogenic action for inorganic arsenic. 9 SAB Comments 10 11 The Panel concluded that: 12 1) Inorganic arsenic has the potential for a highly complex mode of action. 13 2) Until more is learned about the complex PK and PD properties of inorganic 14 arsenic and its metabolites, there is not sufficient justification for the choice of a 15 specific nonlinear form of the dose-response relationship. 16 3) The NRC (2001) recommendation to base risk assessments on a linear dose- 17 response model that includes the southwestern Taiwan population as a 18 comparison group seems the most appropriate approach. 19 4) The Panel also recommends that EPA perform a sensitivity analysis of the 20 Taiwanese data with different exposure metrics, with the subgroup of villages 21 with more than one well measurement, and using a multiplicative model that 22 includes a quadratic term for dose. 23 EPA Response 24 25 As discussed in Section 5.3, EPA investigated a range of model forms for use in the risk 26 assessment, building on previous efforts, including U.S. EPA (2001) and Morales et al. (2000). 27 The model used in the derivation of the preferred risk assessments (see Section 5.3.3) employs: 28 29 • Poisson regression (of cancer mortality against age and dose) fit by maximum likelihood 30 estimation (MLE). 31 • A quadratic age model. 32 • A linear multiplicative dose term. 33 • Confidence limits on the dose term estimated by profile likelihood. 34 • Estimates derived for the data set that includes the southwest Taiwan reference 35 population. 36 37 A range of alternative model forms were investigated, as discussed in Section 5.3.8.4, 38 and the impacts of alternative assumptions about nonwater arsenic intake, drinking water 39 consumption, and other exposure factors were investigated through sensitivity analyses, as 40 described in Section 5.3.8.3. EPA also investigated the properties of the dose-response 41 relationship in the low-dose range of the Taiwanese data, and found that arsenic slope 42 coefficients were positive and statistically significant even when high-exposure groups were A-4 DRAFT—DO NOT CITE OR QUOTE ------- 1 excluded from the analysis. EPA's dose-response modeling found no indication of the existence 2 of a threshold arsenic exposure below which cancer risks are not elevated. As discussed in 3 Section 4.6.3, EPA believes that the available mode of action data do not justify the use of non- 4 linear low-dose extrapolation from the point of departure (POD). 5 6 Charge Question D3 1 8 EPA re-implemented the model presented in the NRC (2001) in the language R as well as 9 in an Excel spreadsheet format. In addition, extensive testing of the resulting code was 10 conducted. Please comment upon precision and accuracy of the re-implementation of the model. 11 SAB Comments 12 13 The Panel concluded: 14 1) That the EPA program conformed to the NRC (2001) recommendation for 15 modeling cancer hazard as a function of age and the average daily dose of 16 exposure to arsenic through drinking water sources. 17 2) The Panel did, however, identify and report to the EPA on two potential 18 discrepancies in the data inputs and one computational error in the portion of the 19 program that employs the BEIR-IV formula to evaluate excess lifetime cancer 20 risk from arsenic exposure. 21 3) The Panel made several suggestions for improvements in the model's 22 programming and documentation conventions, as well as recommendations for 23 specific sensitivity analyses designed to test the robustness of the model to 24 alternative formulations of the hazard function and aggregate population data 25 inputs. 26 EPA Response 27 28 EPA made a number of changes to the model implementation in response to the SAB 29 comments. As in the previous analyses, the linear Poisson dose-response models were estimated 30 using maximum likelihood methods; models were implemented in Excel® and replicated using 31 Statistica®. In the latest analyses, confidence limits on the arsenic dose-response coefficients 32 were estimated using profile likelihood, rather than Bayesian simulation. The confidence limit 33 estimates derived using profile likelihood were very similar to those obtained using Bayesian 34 simulation and estimates derived by "bootstrap" methods. 35 In this latest analysis, the BEIRIV formula for estimating lifetime cancer incidence risks 36 was modified in response to SAB and internal EPA comments. The revised model estimates 37 lifetime cancer incidence data based on "background" cancer incidence and mortality data from 38 the NCI SEER program (see Section 5.3.7.3). The revised approach is discussed in detail in 39 Appendix E.2. A-5 DRAFT—DO NOT CITE OR QUOTE ------- 1 As discussed in the previous response, EPA conducted sensitivity analyses on a number 2 of model parameters. These analyses are described in Section 5.3.8.3. O 4 Charge Question D4 5 6 In calculating estimated cancer risk to the U.S. general population from drinking water 7 exposure to inorganic arsenic, the EPA used epidemiologic data from Taiwan. EPA followed the 8 NRC (2001) recommendations to account for the differences in the drinking water consumption 9 rates for the Taiwanese population and U.S. populations. On the basis of more recent data 10 (noted in U.S. EPA, 2005b), EPA used water intake adjustments for 2 to 3.5 liters/day. EPA 11 asked SAB to recommend a drinking water value. 12 13 SAB Comments 14 15 The Panel agreed that water consumption (via drinking as water, in beverages, or in 16 cooking water) assumptions have a substantial impact on the assessment of arsenic's risk. 17 However, the Panel did not recommend specific values for EPA to use in evaluating dose- 18 response in the Taiwanese study nor for levels of exposure in the U.S. population risk estimates. 19 It did recommend that uncertainty in this parameter be evaluated for both the Taiwanese study 20 population and the U.S. populations at risk. The Panel recommended that EPA should: 21 22 1) Evaluate the impact of drinking water consumption rates associated with more 23 highly exposed population groups with differing exposures and susceptibilities 24 (e-g-, children, pregnant women). 25 2) Incorporate variability parameters for individual water consumption into their 26 analysis for dose-response in the Taiwanese population, as they have done for 27 the U.S. population. 28 3) Conduct sensitivity analyses of the impact of using a range of 29 consumption values for the Taiwanese population. 30 4) Provide a better justification for assuming different consumption levels by gender 31 or, in the absence of such a justification, conduct additional sensitivity analyses to 32 examine the impact of equalizing the gender-specific consumption level. 33 5) More fully articulate and document how different sources of water intake, as 34 well as variability, are incorporated into the risk model (e.g., data for intake from 35 beverages and cooking water). 36 EPA Response 37 38 Data are not available regarding individual water consumption rates and background 39 (nonwater) arsenic intake in the Taiwanese study populations. EPA, therefore, conducted a 40 series of sensitivity analyses involving ranges of drinking water consumption and "background" 41 (nonwater) arsenic consumption that the Agency believes spans a reasonable range of values for 42 these parameters. Arsenic dose-response models were fit assuming nonwater arsenic intakes of A-6 DRAFT—DO NOT CITE OR QUOTE ------- 1 0, 10, 30, 50, 100, and 200 ug/day in the exposed populations, nonwater arsenic intake of 0, 30, 2 and 50 ug/day in the reference population, and daily water consumption ranging from 2.75 to 5.1 3 L/day for (Taiwanese) males and water consumption ranging from 2.0 to 4.1 L/day for females. 4 Risk models also were fit using three different sets of village arsenic drinking water 5 concentrations (median, minimum, and maximum), and three sets of assumptions related to 6 reference (unexposed) populations (southwest Taiwan, all Taiwan, and none). The results of 7 these analyses are summarized in Tables 5-8 and 5-9. Overall, EPA found that cancer slope 8 estimates for male and female lung cancer and male bladder cancer were relatively insensitive to 9 assumptions related to nonwater arsenic intake and varied more or less inversely with the 10 assumed daily water consumption, and with drinking water arsenic concentration estimates. 11 When alternative reference populations were assumed (all Taiwan or none), cancer slope 12 coefficients were lower than when the southwest Taiwan comparison group was included in the 13 analysis. The cancer slope estimates for female bladder cancer were generally more sensitive to 14 changes in exposure assumptions than the other endpoints. 15 16 Charge Question D5 17 18 As recommended by NRC (2001), EPA considered the background dietary intake of 19 inorganic arsenic and incorporated adjustment values of 0, 10, 30, and 50 ug per day into the 20 cancer modeling based on available new data. SAB was asked to recommend a value for the 21 background dietary intake of inorganic arsenic for both the control population and study 22 population of southwestern Taiwan. 23 24 SAB Comments 25 26 The Panel agreed that arsenic levels in food are important considerations for EPA's 27 assessment of lung and bladder cancer risk associated with exposures to arsenic in drinking 28 water. However, the Panel did not recommend a specific value for EPA to use in its base risk 29 assessment. It did recommend a range of values for consideration by EPA in its sensitivity 30 analysis and the Panel offered suggestions to EPA for additional analytical steps to clarify the 31 impact of food levels of arsenic on dose-response and exposure as it revises its risk estimates. 32 These Panel recommendations include that EPA should: 33 34 1) Conduct sensitivity analyses using a range of total arsenic food intake values 35 from at least 50 to 100 ug /day to perhaps as high as 200 ug/day to assess the 36 impact of this range of dietary intakes on risk of lung and bladder cancer from 37 exposure via drinking water in the Taiwan cohort. 38 2) Not assume that the control population has an intake value of zero arsenic from 39 food. 40 3) Apply greater rigor in their discussions of data used in these assessments (e.g., 41 sources, methodological and analytical issues, bioavailability). A-7 DRAFT—DO NOT CITE OR QUOTE ------- 1 4) Give immediate research attention to the issue of arsenic bioavailability. 2 3 EPA Response 4 As discussed in the previous response, EPA conducted sensitivity analyses that 5 assumed nonwater arsenic intakes (doses) for the exposed populations ranging from 0 to 200 6 ug/day and ranging from 0 to 50 ug/day in the reference population. EPA did not specifically 7 conduct sensitivity analyses related to arsenic bioavailability. The Agency notes, however, that 8 the range of absorbed dose that was evaluated implicitly addresses potential bioavailability 9 differences. For example, assuming 50 ug arsenic intake absorbed dose is equivalent to 10 assuming 50% of absorption of a 100 ug/day dose, etc. The Agency believes that the range of 11 arsenic intake that was considered covers the plausible ranges of nonwater dietary arsenic and 12 bioavailability thereof. A-8 DRAFT—DO NOT CITE OR QUOTE ------- APPENDIX B. TABULAR DATA ON CANCER EPIDEMIOLOGY STUDIES 1 The SAB Arsenic Review Panel provided comments on key scientific issues associated 2 with arsenicals on cancer risk estimation in July 2007 (SAB, 2007). It was concluded that the 3 Taiwanese database is still the most appropriate source for estimating bladder and lung cancer 4 risk among humans (specifics provided in Section 5) because of: (1) the size and statistical 5 stability of the database relative to other studies; (2) the reliability of the population and 6 mortality counts; (3) the stability of residential patterns; and (4) the inclusion of long-term 7 exposures. However, SAB also noted considerable limitations within this data set (SAB, 2007). 8 The Panel suggested that one way to mitigate the limitations of the Taiwanese database would be 9 to include other relevant epidemiological studies from various countries. For example, SAB 10 referenced other databases that contained studies of populations also exposed to high levels of 11 arsenic (e.g., Argentina and Chile), and recommended that these alternate sources of data be used 12 to compare the unit risks at the higher exposure levels that have emerged from the Taiwan data. 13 SAB also suggested that, along with the Taiwan data, published epidemiology studies from the 14 United States and other countries where the population is chronically exposed to low levels of 15 arsenic in drinking water (0.5 to 160 ppb) be critically evaluated, using a uniform set of criteria 16 presented in a narrative and tabular format. The relative strengths and weaknesses of each study 17 should be described in relation to each criterion. Additionally, SAB (2007) recommended 18 considering the following issues when reviewing "low-level" and "high-level" studies: (1) 19 estimates of the level of exposure misclassification, (2) temporal variability in assigning past 20 arsenic levels from recent measurements, (3) the extent of reliance on imputed exposure levels, 21 (4) the number of persons exposed at various estimated levels of waterborne arsenic, (5) study 22 response/participation rates, (6) estimates of exposure variability, (7) control selection methods 23 in case-control studies, and (8) the resulting influence of these factors on the magnitude and 24 statistical stability of cancer risk estimates. 25 In light of the SAB recommendations, epidemiological studies in the literature from 1968 26 to 2007 have been reviewed. The report includes data from all populations that have been 27 examined in regard to cancer from arsenic exposure via drinking water. Earlier publications 28 were reviewed and are included as needed to facilitate the understanding of results from certain 29 study populations. As recommended by SAB, studies were presented in both a narrative 30 (Section 4.1) and tabular (below) format. Each publication was evaluated using a uniform set of 31 criteria, including the study type, the size of the study population and control population, and the 32 relative strengths and weaknesses of the study, focusing on the major strengths and weaknesses. 33 While the information in the tables mirrors the information in the narrative, the narrative may 34 provide additional important information concerning the investigation. The studies are presented B-1 DRAFT—DO NOT CITE OR QUOTE ------- 1 by country of origin, then in chronological order by publication year. Below also are definitions 2 of terms that are used in the tables (and the narratives in Section 4.1). 3 Cross-sectional studies have inherent limitations including: (1) difficulty in making 4 causal inference; (2) the fact that data are collected for only one point in time, so that different 5 results may be found if another time-frame had been chosen; and (3) prevalence-incidence bias 6 (also called Neyman bias), which is especially prevalent for longer-lasting diseases, where any 7 risk factor that results in death will be under-represented among those with the disease. 8 Ecological studies provide low cost, convenience, simplicity of analysis, and ease of 9 exposure measurement at population or group level rather than at the individual level; therefore, 10 a wider range of exposures can often be obtained. Concerns about the methodological weakness 11 of ecological studies arise from three facts: estimates of effect do not equate to estimates of 12 biological effect obtained from individual level analysis, exposure data from this design cannot 13 be used to obtain direct estimates of the rate of injury in exposed and unexposed populations, 14 existing data sources are often flawed, and it is difficult to control confounding. 15 Cohort studies are research studies in which the medical records of groups of individuals, 16 who are alike in many ways, but differ by a certain characteristic (for example, individuals who 17 smoke and those who do not smoke) are compared for a particular outcome (such as lung 18 cancer). Cohort studies are generally used to follow large groups over a long period to study rare 19 or long-latency diseases. 20 A case-control study is a retrospective study that compares two groups of people: those 21 with the disease or condition under study (cases) and a very similar group of people (matched 22 controls) who do not have the disease or condition. Researchers study the medical and lifestyle 23 histories of the people in each group to determine which factors may be associated with the 24 disease or condition under investigation. An example is where one group may have been 25 exposed to a particular substance that the other was not. 26 In a nested case-control study, cases of a disease that occur in a defined cohort are 27 identified and, for each, a specified number of matched controls is selected from among those in 28 the cohort who have not developed the disease by the time of disease occurrence in the case. 29 The nested case-control design can potentially offer a lower cost and effort for data collection 30 and analysis compared with the full cohort approach, with relatively minor loss in statistical 31 efficiency. The nested case-control design is particularly advantageous for studies of biologic 32 precursors of disease. 33 Recall bias is a type of systematic bias that occurs when the way a survey respondent 34 answers a question is affected not just by the correct answer, but also by the respondent's 35 memory. 36 Selection bias is the error of distorting a statistical analysis due to the methodology of 37 how the samples were collected. As an example, sample selection may involve pre- or post- B-2 DRAFT—DO NOT CITE OR QUOTE ------- 1 selecting the samples that may preferentially include or exclude certain kinds of results. 2 Selection bias is possible whenever the group of people being studied has any form of control 3 over whether to participate making the participants a non-representative sample. Selection bias 4 may also occur when investigators preferentially select individuals to be included as cases or 5 controls based on prior knowledge of study hypotheses or outcomes. Selection bias in 6 epidemiology is a distortion of data that arises from the way that the data have been collected. If 7 the selection bias is not taken into account, conclusions drawn from the results obtained may be 8 wrong. Self-selection bias is when individuals who make up the study population have any 9 control over whether or not they are allowed to participate. An individual's decision to 10 participate in a study may be associated with other factors that affect the study, which results in 11 the participants being a non-representative sample. 12 The standardized mortality ratio (SMR) in epidemiology is the ratio of observed deaths 13 to expected deaths in a population for a specific health outcome. The SMR also serves as an 14 indirect means for adjusting a rate. The number of observed deaths is obtained for a particular 15 sample of a population that is under investigation, and the number of expected deaths reflects the 16 number of deaths for a larger population from which the study sample has been taken. The 17 calculation used to determine the SMR is simply the number of observed deaths divided by the 18 number of expected deaths. The SMR may be displayed as either a ratio or sometimes as a 19 percentage. If the SMR is shown as a ratio and is equal to 1.0, this means the number of 20 observed deaths equals that of expected cases. If the SMR is greater than 1.0 there is a higher 21 number of deaths than expected, and if the SMR is less than 1.0 there is a lower number of 22 observed than expected deaths. 23 The standardized incidence ratio (SIR) is a common tool for monitoring disease rates. 24 Incidence is the number of newly diagnosed cases in a given location during a given time period. 25 An SIR compares the actual number of cases for a given place and time to the number that 26 would be expected based on disease rates in some comparison area. 27 In statistics and epidemiology, relative risk (RR) is the risk of an event (or of developing 28 a disease) relative to exposure. Relative risk is a ratio of the probability of the event occurring in 29 the exposed group versus the control (non-exposed) group. 30 Time-weighted average (TWA) is the average exposure to a contaminant or condition 31 (such as noise) to which workers are exposed over a period, such as in an 8-hour work day. B-3 DRAFT—DO NOT CITE OR QUOTE ------- Table B-l. Taiwan Cancer Studies Study Period Not indicated Subjects/ Controls 19,269 males 2 1,1 52 females (40,421 total) Exposure Assessment Arsenic concentration in well water (ppb): low (L) = 0- 290 mid (M) = 300-590 high(H) = >600 undetermined (U) Study Outcome Age-/gender-specific skin cancer prevalence rate (1/1000) by arsenic concentration (L, M, H, U): Males, 20-39 yrs.— L=1.5,M = 4.3,H = 22.4, U= 1.7 Males, 40-59 yrs.— L = 6.5, M = 47.7, H = 98.3, U = 51.7 Males, 60 yrs. and over — L = 48. 1,M= 163.4, H = 255. 3, U= 148.2 Total all males combined — L = 4.0, M= 14.4, H = 31.0, U= 16.5 Females, 20-3 9 yrs.— L = 0.1,M = 0.7,H = 3.5, U = 0.9 Females, 40-59 yrs. — L = 3.6, M= 19.7, H = 48.0, U = 9.2 Females, 60 yrs. and over — L = 9.1, M = 62.0, H = 110.1,U = 62.9 Total all females combined — L= 1.3,M = 6.3,H = 12.1, U = 4.7 Both genders, 20-39 yrs. — L=1.3,M = 2.2,H = 11.5,U=1.2 Both genders, 40-59 yrs.— L = 4.9, M= 32.6, H = 72.0, U = 28.3 Both genders, 60 yrs. and over — L = 27. 1,M= 106.2, H = 192.0, U= 107.9 Total both genders combined — L = 2.6, M= 10.1, H = 21.4, U= 10.4 Observed rate/1000: hyperpigmentation = Strengths/ Weaknesses Strengths: -Large number of participants. -Dose-response information provided. Weaknesses: -No individual exposure data. -Possible recall bias among study participants in determining the age of cancer onset and length of residence in the study area. -Water supply changes over time were not collected, nor was information on smoking histories; the arsenic concentration from individual wells varied over time. Reference/ Type of Study Tseng et al., 1968 Ecological B-4 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1958- 1975 Subjects/ Controls 40,421 individuals Exposure Assessment Arsenic concentration in well water (ppb): <300 = low (L) 300-600 = mid(M) >600 = high (H) Study Outcome 183.52 keratosis = 70.95 skin cancer = 10.59 blackfoot disease = 8.91 Age-specific prevalence (per 1000): Skin cancer 20-3 9 y ears— L=1.3,M = 2.2,H = 11.5 40-59 years— L = 4.9, M= 32.6, H = 72.0 60+ years — L = 27. 1,M= 106.2, H = 192.0 Blackfoot disease 20-3 9 years— L = 4.5, M= 13.2, H = 14.2 40-59 years— L = 10.5, M = 32.0, H = 46.9 60+ years — L = 20.3, M = 32.2, H = 61.4 Skin cancer and BFD combined: observed — 61 cases, 1.51/1000 expected — 4 cases, 0.09/1000 observed to expected ratio = 16.77 Strengths/ Weaknesses Strengths: -Large study population. -Adjusted for age and gender. Weaknesses: -No individual monitoring data. -Possible recall bias among study participants (interviews and mailed surveys) in determining the age of cancer onset and the length of residence in the area. Reference/ Type of Study Tseng, 1977 Ecological B-5 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1968- 1982 Subjects/ Controls Subjects from BFD-endemic area Exposure Assessment Median arsenic concentration (ppb): artesian well water — 780 (range: 350- 1140) shallow well water — 40 (range: 0-300) Study Outcome Cancer SMRs (95% CI,p value O.05): Males — bladder =11.00 (9.33- 12.67) kidney = 7.72 (5.37- 10.07) skin = 5. 34 (3.79-8.89) lung =3.20 (2.86-3. 54) liver = 1.70(1.51-1.89) colon =1.60 (1.17- 2.03) Females — bladder = 20.09 (17.02- 23.16) kidney =11. 19 (8.38- 14.00) skin = 6.52 (4.69-8.35) lung = 4.13 (3.60-4.66) liver =2.29 (1.92-2.66) colon =1.68 (1.26- 2.10) Strengths/ Weaknesses Strengths: -The SMRs for the study cohort taken from BFD endemic area in Taiwan were determined using the general population of Taiwan and world population. -Controlled for the potential confounders age and gender. Weakness: -Arsenic measurements not linked to cancer mortality. - Death certificates list the main cause of death rather than all causes - SMRs were only presented by township and villages. Reference/ Type of Study Chenetal., 1985 ecological B-6 DRAFT—DO NOT CITE OR QUOTE ------- Study Period January 1980- December 1982 Subjects/ Controls Deceased cancer cases: 69 bladder 76 lung 59 liver Controls: 368 (community matched) Exposure Assessment Median arsenic concentration: artesian well water — 780 ppb (range: 350-1140) shallow well water — 40 ppb (range: 0-300) Study Outcome Age-/sex-adjusted odds ratios, well water use >40 years: bladder cancer =3.90 lung cancer = 3.39 liver cancer = 2.67 Mantel-Haenszel x2: bladder cancer = 13.74* lung cancer = 8.49* liver cancer = 9.01* *p<0.01 Multivariate logistic regression: improvement x2 value — bladder cancer = 11.45* lung cancer = 9.04* liver cancer = 6.34* *p<0.01 Strengths/ Weaknesses Strengths: -Cases confirmed using histology or cytology findings. -Cancer cases and controls were from the same BFD community. -Potential confounders adjusted for in the analysis included age, gender, smoking, tea drinking, vegetable consumption, and fermented bean consumption. Weaknesses: -Confounders not controlled for included recall bias from case and control interviews regarding lifestyle, diet, and daily water consumption and source of water. -Selection bias (control selection). Reference/ Type of Study Chenetal., 1986 Case- control B-7 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1973- 1986 Subjects/ Controls Blackfoot- endemic area residents Population of Taiwan as reference population World population as reference population Exposure Assessment Three exposure categories (ppb): <300 300-590 >600 Study Outcome Age-standardized mortality per 100,000 for various cancers: World population: <300 ppb Males — all sites = 154.0, liver = 32.6, lung = 35.1, skin =1.6, prostate = 0.5, bladder = 15.7, kidney = 5.4 Females — all sites = 118.8, liver = 14.2, lung = 26.5, skin =1.6, bladder = 16.7, kidney = 3.6 300-590 ppb Males — all sites = 258.9, liver =42.7, lung = 64.7, skin =10.7, prostate = 5.8, bladder = 37.8, kidney =13.1 Females — all sites = 182.6, liver= 18.8, lung = 40.9, skin = 10.0, bladder =35.1, kidney = 12.5 >600 ppb Males — all sites = 434.7, liver =68.8, lung = 87.9, skin =28.0, prostate = 8.4, bladder = 89.1, kidney = 2 1.6 Females — all sites = 369.4, liver =3 1.8, lung = 83.8, skin =15.1, bladder =91. 5, kidney = 35.3 Taiwan: Males — all sites = 128.1, liver =28.0, lung = 19.4, skin =0.8, prostate = 1.5, bladder = 3.1, kidney = 1.1 Females — all sites = 85.5, liver =8.9, lung = 9.5, skin = 0.8, bladder = 1.4, kidney = 0.9 Strengths/ Weaknesses Strengths: -Data from arsenic monitoring conducted in 1962-64 and 1974-76 found similar results. Weaknesses: -Individual arsenic exposure levels were not presented. Reference/ Type of Study Chenetal., 1988a Cohort B-8 DRAFT—DO NOT CITE OR QUOTE ------- Study Period January 1968- December 1983 August 1983- February 1987 Subjects/ Controls 241 cases 759 controls General population of Taiwan Local endemic area population 246 BFD bladder cancer cases 444 BFD- endemic area residents 286 residents neighboring the endemic area 731 non- endemic area residents Exposure Assessment Arsenic concentration (ppb): artesian well water — median = 780 range = 350- 1140 shallow well water — median = 40 range = 0-300 Percent of area well water with arsenic content of >50 ppb: Pei-men= 81 Hsueh-Chia = 27 Pu-Tai = 58 Jinag-Jium = 24 Tai-Pao = 45 Pao-Chung = 54 >350 ppb: Pei-men = 62 Hsueh-Chia = 7 Pu-Tai = 8 Jinag-Jium = 0 Tai-Pao = 6 Pao-Chung = 0 Study Outcome Significant SMRs (p values) (compared to population of Taiwan): Cancers — bladder =3 8. 80 (0.001) skin = 28.46 (O.01) lung= 10.49(0.001) liver =4.66(0.001) colon =3.81(0.05) Significant SMRs (p values) (compared to population of BFD- endemic area): Cancers — bladder = 2.55 (0.01) skin = 4.51 (O.05) lung = 2.84(0.01) liver =2.48(0.01) Positive cytology (bladder cancer/atypia) prevalence rate (%): BFD cases = 4.5 endemic area = 2.5 neighboring area = 0.7 non-endemic area = 0.13 Strengths/ Weaknesses Strengths: -Cases consisted of blackfoot disease cases, matched to healthy community controls for age, sex, and residence. -Recall bias was minimized through interview techniques. -SMRs were determined using both the national Taiwanese population and the local endemic area population. Weakness: -Arsenic dose levels were not provided. Strengths: -Histological confirmation of bladder cancer diagnoses. Weaknesses: -Lack of individual exposure data. Reference/ Type of Study Chenetal., 1988b Cohort/ nested case- control Chiang et al., 1988 Case- control B-9 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1973- 1986 1972- 1983 Subjects/ Controls Residents of 42 villages 1976 world population used as comparison Arsenic- exposed subjects from 314 townships and precincts Exposure Assessment Three exposure categories (ppb): <300 300-590 >600 Total wells tested = 83,656, >50 ppb in 15,649 wells (18.7%), > 350 ppb in 2,224 wells (2.7%) Concentrations in the remainder of the wells were not given Study Outcome Trend test of the extension of the Mantel-Haenszel Chi square test: Cancers — Both genders: bladder, skin, lung — p< 0.001 Males only: kidney, liver, prostate— p < 0.05 Females only: kidney— p< 0.001 Multivariate adjusted regression coefficient for cancers (SE): Males — liver =6.8 (1.3), nasal cavity = 0.7(0.2), lung = 5.3(0.9), skin = 0.9 (0.2), bladder =3. 9 (0.5), kidney = 1.1 (0.2), prostate = 0.5 (0.2) Females — liver = 2.0 (0.5), nasal cavity = 0.4 (0.1), lung = 5.3(0.7), skin = 1.0 (0.2), bladder = 4.2 (0.5), kidney = 1.7(0.2) No p values indicated. Strengths/ Weaknesses Strengths: -Adjustments made for age and gender. -Lifestyle, access to medical care, and socioeconomic status were similar among the study groups. Weaknesses: -Limitations of mortality data. -Associations observed at the local level may not be accurate at the individual level (ecological fallacy). Strengths: -Potential confounders controlled for included socioeconomic differences, i.e., urbanization and industrialization. -Cancer rates in endemic BFD townships were compared with cancer rates in non-endemic townships of Taiwan. -Ecological correlations reported between arsenic content in well water and mortality from various cancers. Weaknesses: -Potential confounders not controlled for were gender and other potential well water contaminants. -No individual arsenic exposures. Reference/ Type of Study Wu et al., 1989 Ecological Chen and Wang, 1990 Ecological B-10 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1973- 1986 Followed up for 0.05-7.69 years (4.97 ±1.72 [SD] Subjects/ Controls Arsenic- exposed subjects from 42 villages 263 BFD cases 2,293 healthy residents Exposure Assessment Well water arsenic exposure categories (ppb): <100 100-290 300-590 >600 Overall range: 10-1,752 Artesian well water median arsenic level = 780 ppb Shallow well water median Study Outcome Cancer development potency index (daily arsenic intake of 10 |ig/kg): Males — liver = 4.3 x 10'3 lung= 1.2 x 10'2 bladder =1. 2 x io~2 kidney = 4.2 x 10'3 Females — liver =3. 6 x 10'3 lung= 1.3 x 10'2 bladder =1. 7 x io~2 kidney = 4.8 x 10'3 Multivariate adjusted RR (95% CI), cancer: All sites — Age: every -1-yr increment = 1.05 (1.03-1.06)* Sex: men = 1.00, Strengths/ Weaknesses Strengths: -Potential confounders included age, gender, access to medical care, socioeconomic status, and lifestyle and were all controlled for in the analysis. -Villages share similar socioeconomic status, living environments, lifestyles, dietary patterns, and even medical facilities. Weaknesses: -Armitage-Doll model constrains risk estimates to be monotonically increasing function of age. -Age stratification only available for 20- year strata. -Possible underestimation of risk because it was assumed that an individual's arsenic intake remained constant from birth to the end of the follow- up period. -Assumption that an individual's arsenic intake remained constant from birth to the end of the follow- up period and the possible underestimation of risk because other sources of arsenic exposure were not considered. Strengths: -Showed a significant dose-response relationship with increasing concentrations of arsenic. Reference/ Type of Study Chenetal., 1992 Ecological Chiou et al., 1995 Cohort B-11 DRAFT—DO NOT CITE OR QUOTE ------- Study Period years) until January 1993 Subjects/ Controls Exposure Assessment arsenic level = 40ppb Study Outcome women = 0.72 (0.43-1.18)* Cigarette smoking: no= 1.00, yes = 1.52 (1.00-2.48)* Status of blackfoot disease: no= 1.00, yes = 2.69 (1.80-4.01)* Cumulative arsenic exposure (mg/liter x yr): 0= 1.00 0.1-19.9= 1.39(0.82- 2.37) 20+ =1.76 (1.01- 3.06)* unknown = 0.72 (0.42- 1.22) Lung- Age: every -1-yr increment = 1.06(1.02- 1.10)* Sex: men= 1.00, women = 1.79(0.44- 7.32)* Cigarette smoking: no= 1.00, yes = 4.31 (1.08-17.20)* Status of blackfoot disease: no= 1.00, yes = 2.45 (1.07-0.57)* Cumulative arsenic exposure (mg/liter x yr): 0=1.00 0.1-9.9 = 2.74(0.69- 11.0) 20+ = 4.01 (1.00- 16.12)* unknown =2.01 (0.55- 7.36) Bladder- Age: every 1-yr increment = 1.04 (1.05- 1.08)* Sex: men= 1.00, women =0.45 (0.18- 1.16) Cigarette smoking: no= 1.00, yes = 1.00 (0.37-2.31) Status of blackfoot disease: Strengths/ Weaknesses -Analysis adjusted for BFD status, age, sex, and smoking. -Reported incidence data. Weaknesses: -Artesian well water arsenic concentration was unknown for some study subjects. Reference/ Type of Study B-12 DRAFT—DO NOT CITE OR QUOTE ------- Study Period January 1980- December 1987 Subjects/ Controls 2,915 urinary cancer cases Exposure Assessment 6 categories of arsenic exposure (ppb): <50 50-80 90-160 170-320 330-640 >640 Study Outcome no= 1.00, yes = 4.41 (2.06-9.45)* Cumulative arsenic exposure (mg/liter x yr): 0= 1.00 0.1-19.9=1.57(0.44- 5.55) 20+ = 3.58 (1.05- 12.19)* unknown = 1.25(0.38- 4.12) *p < 0.05 Rate differences (SE)* with positive associations: Males — Bladder cancer: transitional cell >640 ppb = 0.57(0.07), adenocarinoma >640 ppb = 0.027(0.008) Kidney cancer: transitional cell 330-640 ppb = 0.05(0.02) Females — Urethral cancer, all cell types combined >640 ppb = 0.027(0.007) *Estimates for 1 unit increase (1%) in predictor (exposure category) Strengths/ Weaknesses Strengths: -Adjusted for age, gender, urbanization, and smoking. Weaknesses: - Limitations of ecological study design. Reference/ Type of Study Guo et al., 1997 Ecological B-13 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1971- 1994 Subjects/ Controls 11,193 mortalities from all causes of disease Local reference population National reference population Exposure Assessment Median artesian wells water arsenic content: 780 ppb (range = 250- 1140 ppb) Individual exposure data not available Study Outcome Males — BFD area compared to local reference — SMR (95% CI): all cancers = 2.19 (2.11-2.28) BFD area compared to national reference — SMR (95% CI): all cancers =1.94 (1.87-2.01) Females — BFD area compared to local reference — SMR (95% CI): all cancers = 2.40 (2.30-2.51) BFD area compared to national reference — SMR (95% CI): all cancers = 2.05 (1.96-2.14) p<0.05 Strengths/ Weaknesses Strengths: -Exposed group and local reference group had similar lifestyle factors. -All cancers were pathologically confirmed. -Controlled for gender, a potential confounder. Weaknesses: -Only one underlying cause of death (not multiple causes) was indicated on death certificate, resulting in possible distortion of association between exposure and disease. -Lack of individual exposure data. -Potential confounders not controlled for were age, smoking, alcohol consumption, and occupational exposures. Reference/ Type of Study Tsaietal., 1999 Cross- sectional B-14 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1973- 1986 Subjects/ Controls 42 arseniasis- endemic villages Population of Taiwan Exposure Assessment Arsenic exposure categories (ppb) = 0-50 50-100 100-200 200-300 300^00 400-500 500-600 600+ Study Outcome SMRs (male and female combined.) Bladder cancer SMRs:* 0-50 ppb = 10.02 50-100 ppb = 4. 15 100-200 ppb = 10.47 200-300 ppb = 7.66 300-400 ppb = 7.44 400-500 ppb = 29.68 500-600 ppb = 14.90 600+ ppb = 32.71 Lung cancer SMRs:* 0-50 ppb = 1.56 50-100 ppb = 1.43 100-200 ppb = 2.43 200-300 ppb = 3. 08 300-400 ppb = 1.97 400-500 ppb = 3. 65 500-600 ppb = 3. 32 600+ ppb = 5. 14 Liver cancer SMRs:* 0-50 ppb = 1.18 50-100 ppb = 0.65 100-200 ppb = 1.74 200-300 ppb = 1.44 300-400 ppb = 0.77 400-500 ppb =1.60 500-600 ppb =1.59 600+ ppb = 2. 17 Bladder, lung, and liver combined cancer SMRs:* 0-50 ppb = 1.83 50-100 ppb =1.16 100-200 ppb = 2.51 200-300 ppb = 2.47 300-400 ppb = 1.63 400-500 ppb = 3. 93 500-600 ppb = 3. 06 600+ ppb = 4.86 *No significance levels presented. Strengths/ Weaknesses Strengths: -Person-years at risk stratified by age, gender, and arsenic level. -Individual well concentrations were available for each village. Weaknesses: -Ecological study design (no individual monitoring data, individual exposures not available). -Potential confounding by smoking, use of bottled water, and dietary intake, since this information was not available. Reference/ Type of Study Morales et al., 2000 Ecological B-15 DRAFT—DO NOT CITE OR QUOTE ------- Study Period October 1991- September 1994 with follow-up through the end of 1996 Subjects/ Controls 8, 102 residents (4,056 men and 4,046 women) General population of Taiwan used as comparison Exposure Assessment Exposure categories (ppb) : <10.00 10.1-50.0 50.1-100.0 >100.0 Study Outcome Standardized incidence ratio (95% CI): urinary cancer = 2.05 (1.22, 3.24) bladder =1.96 (0.94- 3.61) kidney = 2.82 (1.29- 5.36) p<0.05 Multivariate adjusted RR (95% CI): Well water arsenic concentration (ppb): Urinary organs — 10.1-50.0=1.5(0.3- 8.0) 50.1-100.0 = 2.2(0.4- 13.7) >100.0 = 4.8 (1.2-19.4) TCC 10.1-50.0=1.9(0.1- 32.5) 50.1-100.0 = 8.2(0.7- 99.1) >100.0= 15.3(1.7- 139.9) Strengths/ Weaknesses Strengths: - Showed a significant dose- response relationship with increasing concentrations of arsenic. -Potential confounders adjusted for included age, gender, and smoking. -Individual exposure estimates were available. Weaknesses: -Possible diagnosis bias, since data were collected from various community hospitals. -Possible recall bias resulting from self- reported information. - Short duration of follow-up, which limited the number of person-years of observation. -Possible misclassification, especially in the low- dose region due to lack of arsenic exposure information in the food. Reference/ Type of Study Chiou et al., 2001 Cohort B-16 DRAFT—DO NOT CITE OR QUOTE ------- Study Period January 1980- December 1989 January 1980- December 1999 Subjects/ Controls 2,369 skin cancer cases (1,415 men and 954 women) 40,832 liver cancer patients (32,034 men and 8,798 women) Exposure Assessment 6 categories of arsenic exposure (ppb): <50, 50-80, 90-160, 170-320, 330-640, >640 BFD area average arsenic concentration = 220 ppb Non-BFD area average arsenic concentration = 20 ppb Study Outcome Statistically significant rate differences per 100,000 person-years (SE):* Males — Basal cell carcinoma >640 ppb = 0.128(0.025)** Squamous cell carcinoma 170-320 ppb = 0.073(0.024)** 330-640 ppb= - 0.10(0.031)** >640 ppb = 0.155(0.028)** Females — Squamous cell carcinoma 330-640 ppb = - 0.064(0.027)* >640 ppb = 0.212(0.024)** *p < 0.05 **p<0.01 No statistically significant (P > 0.05) differences were noted for cell types of liver cancer between the BFD area and the other areas. Strengths/ Weaknesses Strengths: -Cases were identified from government operated National Cancer Registration Program. -Pathological classifications determined by board- certified pathologists. -Potential confounders adjusted for in the analysis included gender and age. Weaknesses: -Limitations of ecological study design. (No monitoring data were presented.) Strengths: -Cases identified from government operated National Cancer Registration Program. -Pathological classifications were determined by board- certified pathologists. -Potential confounders adjusted for included gender and age. Weaknesses: -Limitations of ecological study design. (No monitoring data were presented). Reference/ Type of Study Guo et al., 2001 Ecological Guo, 2003 Ecological B-17 DRAFT—DO NOT CITE OR QUOTE ------- Study Period January 1985- December 2000; average follow-up of 8 years 1971- 2000 Subjects/ Controls 2,503 residents in southwestern area 8,088 residents in northeastern area Residents of 4 BFD-endemic townships Exposure Assessment Southwestern area average arsenic exposure categories (ppb): <10 10-99.9 100-299.9 300-699.9 >700 Unknown Median well water arsenic level, early 1960s = 780 ppb Study Outcome Multivariate-adjusted RR of lung cancer for average arsenic level in well water (ppb): <10= 1.00 (referent) 10-99.9=1.09(0.63- 1.91) 100-299.9 =2.28 (1.22- 4.27) 300-699.9 = 3.03 (1.62-5.69) >700 = 3.29(1.60- 6.78) Unknown = 1.10(0.60- 2.03) SMR liver cancer: Males — 1989-1991 = 1.868 1998-2000=1.242 Females — 1983-1985 = 2.041 1998-2000= 1.137 Strengths/ Weaknesses Strengths: -Confounders controlled for were age, gender, education, and alcohol consumption. -Long follow-up period and the use of a national computerized cancer case registry. -All lung cancer cases were pathologically confirmed. Weaknesses: -Historical monitoring data not available. -Possible misclassification bias because exposure measurements were based on one survey. Strengths: -Residents in the study area were similar in terms of socioeconomic status, living environments, lifestyles, dietary patterns, and health service facilities. -Accurate death registration system. Weaknesses: -Limitations of mortality data. Reference/ Type of Study Chenetal., 2004a Cohort Chiuetal., 2004 Cohort B-18 DRAFT—DO NOT CITE OR QUOTE ------- Study Period January 1971- December 1990 Subjects/ Controls 1,078 lung cancer mortality cases Exposure Assessment Arsenic exposure levels (ppb): <050 50-80 90-160 170-320 330-640 >640 Study Outcome Lung cancer mortality increase with 1,000 ppb increase in mean arsenic level (p=0.01): Men — 27.45/100,000 person- years Women — 18.93/100,00 person- years Strengths/ Weaknesses Strengths: -Adjusted for gender and age. -Cases were ascertained using information from household registry offices in each township. Taiwanese law requires timely reporting of deaths to these offices. Weaknesses: -Limitations of ecological studies. -Smoking was not controlled for in the analysis. Reference/ Type of Study Quo, 2004 Ecological B-19 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1971- 2000 1988- 2001 Subjects/ Controls Residents of 4 BFD-endemic townships 7 females 14 males Exposure Assessment Median arsenic level (ppb), early 1960s = 780 (range: 350- 1140) No exposure data Study Outcome Kidney cancer SMR (observed vs. expected): 1971— Men= 19.04 (4 vs. 0.21) Women = 23. 52 (8 vs. 0.34) 2000— Men = 4.46 (8 vs. 1.79) Women = 6. 52 (9 vs. 1.38) Chi square (Taiwan case series compared to 3 U.S. case series studies): Males — urethra! adenocarcinoma: p< 0.0001 Strengths/ Weaknesses Strengths: -Adjusted for gender and age. -Mandatory registering of all births, deaths, marriages, divorces, and migration to the Household Registration Office in Taiwan, making it an accurate data source. -Most residents had similar socioeconomic status, living environments, lifestyles, dietary patterns, and health service facilities and worked in farming, fisheries, or salt production. -All kidney cancer cases in the area probably had similar access to medical care. Weaknesses: -Mortality data limitations. -Cross-sectional study limitations. -Smoking may possibly have been a confounder not adequately controlled for. Strengths: -Cases were pathologically confirmed. Weaknesses: -Limited number of cases. -No exposure information. Reference/ Type of Study Yangetal., 2004 Cross- sectional Tsaietal., 2005 Cross- sectional B-20 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1971- 2000 Subjects/ Controls Residents in 4 BFD-endemic area townships Exposure Assessment Median arsenic level, early 1960s = 780 ppb Study Outcome Bladder cancer SMRs (observed vs. expected): 1971— Males = 10.25 (8 vs. 0.78) Females = 14.89 (7 vs. 0.47) 2000— Males = 2. 15 (5 vs. 2.32) Females = 7.63 (10 vs. 1.31) Strengths/ Weaknesses Strengths: -All bladder cancer cases in the area probably had similar access to medical care. -Adjusted for age and gender. -Mandatory registering of all births, deaths, marriages, divorces, and migration to the Household Registration Office in Taiwan, making it an accurate data source. Weaknesses: -Limitations of a cross-sectional mortality study. -Smoking may possibly have been a confounder. Reference/ Type of Study Yangetal., 2005 Cross- sectional B-21 DRAFT—DO NOT CITE OR QUOTE ------- Table B-2. Japan Cancer Studies Study Period 1959- 1992 Subjects/ Controls 454 residents Exposure Assessment Well arsenic concentration (ppb): <50 50-990 >1000 Study Outcome >1000 ppb SMRs (95% CI): Males — all deaths =1.88 (1.17-2.96) all cancers = 4.19 (2.20-7.56) lung cancer = 19.08 (8.88-38.76) urinary cancer = 33.16 (5.92-121.58) all cancers except lung = 2.22 (0.87-5.22) Females — all deaths = 1.31 (0.76-2.18) all cancers = 3.00 (1.40-6.13) lung cancer = 7.15 (0.36-41.11) urinary cancer = 27.85 (1.42-159.89) all cancers except lung 2.73 (1.19-6.04) Cox's proportional hazard analysis (95% CI), highest group vs. background: concentration categories (ppb) >1 000 vs. 1 all deaths = 1.74 (1.10-2.74) all cancers = 4.82 (2.09-11.14) lung cancer = 1,972.16(4.34- 895,385.11) Strengths/ Weaknesses Strengths: -Cohort examined by 3 exposure categories. -Included information on smoking, age and gender. Weaknesses: -Lacking detailed arsenic intake information. -Small study population. -Possible misclassification bias. -Recall bias (smoking history).. Reference/ Type of Study Tsuda et al., 1995 Cohort B-22 DRAFT—DO NOT CITE OR QUOTE ------- Table B-3. South America Cancer Studies Study Period 1986- 1991 Subjects/ Controls Bladder cancer deaths in 26 Cordoba counties Population of Argentina Exposure Assessment Exposure categories: low medium high (crude average estimate of 178 ppb) Two counties in high-exposure group Study Outcome Bladder cancer SMR (95% CI) by exposure category: Men — low = 0.80 (0.66- 0.96) medium = 1.42 (1.14- 1.74) high =2.14 (1.78- 2.53) test for trend: p=0.001 Women — low = 1.21 (0.85- 1.64) medium = 1.58(1.01- 2.35) high= 1.82(1.19- 2.64) test for trend: p=0.04 Strengths/ Weaknesses Strengths: -Adjusted for age and gender. -Analysis restricted to rural counties to limit confounders. -To account for cancer diagnosis and detection bias, stomach cancer, which is known not to be related to arsenic exposure, was used as a comparison cancer. Weaknesses: -Limitations of ecological studies. -Lack of comprehensive, systematic monitoring data. -No arsenic exposure levels in low and medium groups reported. -Lack of individual smoking history. Reference/ Type of Study Hopenhayn- Richetal., 1996a Ecological B-23 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1986- 1991 Subjects/ Controls Population from 26 counties in Cordoba Population of Argentina Exposure Assessment Exposure categories: low medium high (crude average estimate of 178 ppb) Study Outcome SMRs (95% CI) by exposure categories: Kidney cancer — Men low = 0.87 (0.66- 1.10) medium = 1.33 (1.02- 1.68) high =1.57 (1.17- 2.05) Women low= 1.00(0.71- 1.37) medium = 1.36(0.94- 1.89) high =1.81 (1.19- 2.64) Lung cancer Men low = 0.92 (0.85- 0.98) medium = 1.54 (1.44- 1.64) high =1.77 (1.63- 1.90) Women low = 1.24 (1.06- 1.42) medium = 1.34(1.12- 1.58) high =2.16 (1.83- 2.52) p< 0.001 in trend test Strengths/ Weaknesses Strengths: -Adjusted for age and gender. -Analysis restricted to rural counties to limit confounders. -To account for cancer diagnosis and detection bias, stomach cancer, that is known not to be related to arsenic exposure, as a comparison cancer. Weaknesses: -Limitations of ecological studies. -Lack of comprehensive, systematic monitoring data. -No arsenic exposure levels in low and medium groups reported. -Lack of individual smoking history. Reference/ Type of Study Hopenhayn- Richetal., 1998 Ecological B-24 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1989- 1993 Subjects/ Controls 390,340 residents national mortality data from 1991 Population of Chile used as reference group Exposure Assessment Region II average water arsenic level (ppb): 1950-1954 = 123 1955-1959 = 569 1960-1964 = 568 1965-1969 = 568 1970-1974 = 272 1975-1979 = 176 1980-1984 = 94 1985-1989 = 71 1990-1994 = 43 Study Outcome SMRs (95% CI, p value) >30 years old: Men — bladder =6.0 (4.8- 7.4, 0.001) kidney =1.6 (1.1-2.1, 0.012) liver =1.1 (0.8-1. 5, 0.392) lung =3.8 (3. 5-4.1, O.001) skin = 7.7 (4.7-11. 9, 0.001) Women — bladder =8.2 (6.3- 10.5,0.001) kidney = 2.7 (1.9-3. 8, O.001) liver =1.1 (0.8-1. 5, 0.377) lung =3. 1(2.7-3. 7, 0.001) skin =3.2 (1.3-6.6, 0.016) Strengths/ Weaknesses Strengths: -Large study size. -Used national data for comparison. No other major populations in Chile were exposed to arsenic in drinking water. -SMRs adjusted for age and gender. Weaknesses: -Arsenic levels in drinking water available only by city or town. -Deaths were not linked to town so individual exposure is not known. -Limited smoking data. -No dose-response information provided. Reference/ Type of Study Smith etal., 1998 Ecological B-25 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1994- 1996 Subjects/ Controls 152 lung cancer cases 419 controls Exposure Assessment Average water arsenic concentration (ppb) during peak exposure years: 0-10 10-29 30-59 60-89 90-199 200-399 400-699 700-999 Study Outcome Lung cancer odds ratio (95% CI): Age/gender adjusted — 0-10 ppb = 1 (referent) 10-29 ppb = 0.4 (0.1- 0.5) 30-59 ppb = 0.0 (0.6- 7.2) 60-89 ppb = 0.1 (1.8- 9.2) 90-199 ppb = 0.8 (1.1-7.0) 200-399 ppb = 0.4 (2.0-10.0) 400-699 ppb = 0.9 (2.4-19.8) 700-999 ppb = 0.3 (3.1-12.8) Male vs. female = 0.7 (1.1-2.7) Full model (95% CI) (included smoking and copper smelting): 0-10 ppb = 1 (referent) 10-29 ppb = 0.3 (0.1- 1.2) 30-59 ppb = 1.8(0.5- 6.9) 60-89 ppb = 4.1 (1.8- 9.6) 90-199 ppb = 2.7 (1.0-7.1) 200-399 ppb = 4.7 (2.0-11.0) 400-699 ppb = 5.7 (1.9-16.9) 700-999 ppb = 7.1 (3.4-14.8) Male vs. female =1.1 (0.6-1.8) Ever vs. never smoked = 4.3 (2.6-7.3) SES medium vs. low = 1.3 (0.7-2.5) SES high vs. low = 2.3(0.5-12.1) Copper smelting (ever/never) =1.7 (0.7-4.4) Strengths/ Weaknesses Strengths: -Odds ratios adjusted for age, gender, cumulative lifetime cigarette smoking, working in copper smelting, and socioeconomic status. -Because the control group selection was complex, several validity checks were completed. Weaknesses: -Relatively more controls were chosen from the highly exposed city of Antofagasta than from the lower exposure cities of Arica and Iquique resulting in possible underestimation of risk. Reference/ Type of Study Ferreccio et al., 2000 Case- control B-26 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1996- 2000 1989- 2000 Subjects/ Controls 114 bladder cancer cases 114 individuals without bladder cancer -200,000 residents Exposure Assessment Average arsenic concentration (ppb) of 5 years of highest exposure during the period of 6-40 years prior to interview: 0-50 51-100 101-200 >200 (mean: 164 ppb) Water arsenic levels: prior to 1958, -90 ppb; in the late 1950s, water supplementation from a nearby river where arsenic levels approached 1000 ppb was added to the existing city water supply Study Outcome Bladder cancer Odds ratio (95% CI) — ever smokers by time before interview: 5 1-60 years earlier = 2.65 (1.2-5.8) 61-70 years earlier = 2.54 (1.0-6.4) periods combined = 2.5(1.1-5.5) SMRs (95% CI): 1950-1957 birth cohort (early childhood exposure): lung cancer = 7.0 (5.4-8.9, p< 0.001) High exposure period (1958-1971) with probable exposure in utero and early childhood: lung cancer = 6.1 (3.5-9.9, p< 0.001) Strengths/ Weaknesses Strength: -Potential confounders controlled included age, gender, smoking, and county of residence. Weaknesses: -Lack of a cancer registry, arsenic exposure misclassification (use of current water source arsenic measurements possibly causing underestimation of exposure), and recall bias. -Possible selection bias since controls had a significantly reduced rate of participation than cases and cases were selected from the tumor registry. -Other harmful exposures not measured. Strengths: -Extensive documentation of arsenic in drinking water in the Antofagasta water system. Weaknesses: -Residence was determined from death certificates and relates to residence at the time at death. -Reliance on death certificates resulting in potential diagnostic bias. -Information bias (smoking history). Reference/ Type of Study Bates et al., 2004 Case- control Smith etal., 2006 Cohort B-27 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1950- 2000 Subjects/ Controls Region II residents Region V residents as comparison group Population of Chile Exposure Assessment Average arsenic concentration (ppb): Region II 1950-1954 = 123 1955-1959 = 569 1960-1964 = 568 1965-1969 = 568 1970-1974 = 272 1975-1979 = 176 1980-1984 = 94 1985-1989 = 71 1990-1994 = 43 Region V unexposed Study Outcome Peak rate ratios (95% CI) compared to Region V and Chile: Lung Cancer 1992-1994 Men 3.61 (3.13-4.16) (Region V) 4.20 (3.76-4.70) (Chile) 1989-1991 Women 3.26 (2.50-4.23) (Region V) 3.41 (2.76-4.22) (Chile) Bladder Cancer 1986-1988 Men 6.10(3.97-9.39) (Region V) 5.99(4.41-8.14) (Chile) 1992-1994 Women 13.8 (7.74-24.5) (Region V) 9.32 (6.67-13.0) (Chile) Strengths/ Weaknesses Strengths: -Large population size. -Accurate past exposure data. -Known exposure pattern. -Controlled for potential confounding by age, gender, and smoking. Weaknesses: -Could not account for migration. -No individual exposure data or data on other risk factors (smoking and occupation). Reference/ Type of Study Marshall et al., 2007 Ecological B-28 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1950- 2000 Subjects/ Controls 314,807 exposed 1,230,498 unexposed Exposure Assessment Average water concentration (ppb) in Region II: Before arsenic removal plant — 1950-1957 = 90 1958-1970 = 870 After arsenic removal plant — 1971-1985=110 1986-2000 = 40 Present = 10 Study Outcome Excess deaths as percentage of total deaths (%) due to acute myocardial infarction, lung cancer, and bladder cancer combined: Males — 1950-1957=1.00 1958-1964 = 4.19 1965-1970 = 6.03 1971-1979* = 6.48 1980-1985 = 8.94 1986-1990 = 10.07 1991-1995 = 10.87 1996-2000 = 7.92 Total = 6.93 Females — 1950-1957 = 0.48 1958-1964= 1.59 1965-1970 = 3.11 1971-1979* = 3.78 1980-1985 = 2.75 1986-1990 = 3.85 1991-1995 = 4.00 1996-2000 = 3.36 Total = 2.94 *No data available for 1976 Strengths/ Weaknesses Strengths: -Almost all drinking water came from a few municipal water sources, which had known arsenic concentrations. -The study involved a large population that experienced a rapid increase in arsenic exposure followed by a rapid decrease in arsenic exposure. -To ensure that an appropriate comparison population was chosen, preliminary investigations were conducted to compare income, smoking, and quality of death certificate information. Weaknesses: -Possible biases resulting from a lack of individual exposure data and confounders. Reference/ Type of Study Yuan et al., 2007 Ecological B-29 DRAFT—DO NOT CITE OR QUOTE ------- Table B-4. North America cancer studies Study Period 3 9 years (endpoint- 1978 diagnosis) Subjects/ Controls 71 National Bladder Cancer Study participants 160 National Bladder Cancer Study participants without bladder cancer Exposure Assessment Mean arsenic level (ppb) = 5.0 (range = 0.5-160 ) Exposure indices: Index 1 — cumulative dose (<19, 19to<33, 33 to <53, >53 mg) Index 2 — intake concentration adjusted to fluid intake (<33, 33 to <53, 53 to <74, >74 mg- years) Study Outcome Odds ratio for bladder cancer and arsenic exposure: no association of bladder cancer with Index 1 or Index 2. Among smokers, positive trend in 10 year intervals. Strengths/ Weaknesses Strengths: -Age, gender, smoking status, years of chlorinated surface water exposure, history of bladder infection, education, occupation, population size of geographic area, and urbanization were addressed. -Cases were histologically confirmed. Weaknesses: -Small size of study population. -Absence of historical monitoring data and data on arsenic levels in public water supplies were collected in 1978- 1979. -The subjects were mostly males and the data on females were inadequate. -Arsenic exposure levels were based on measurements close to the time that cases were diagnosed. -Arsenic from food was not considered. Reference/ Type of Study Bates et al., 1995 Case- control B-30 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1996 1993- 1996 Subjects/ Controls 2,203 deceased individuals from Millard County General Utah population used as comparison 587 BCC cases 284 SCC cases 524 controls Exposure Assessment Arsenic exposure index (ppb-years): low = <1000 medium = 1000- 4999 high = >5000 Toenail arsenic level (ug/g): BCC cases = 0.01-2.03 SCC cases = 0.01-2.57 controls = 0.01- 0.81 Study Outcome Cancer SMRs (95% CI): kidney — males = 1.75(0.80- 3.32) females =1.60 (0.44-4.11) bladder and other urinary organs — males = 0.42 (0.08- 1.22) females = 0.81 (0.10-2.93) melanoma of the skin — females = 1.82 (0.50-4.66) prostate = 1.45* (1.07-1.91) *p<0.05 OR (95% CI), toenail arsenic concentrations above the 97th percentile: SCC = 2.07 (0.92- 4.66) BCC = 1.44 (0.74- 2.81) Strengths/ Weaknesses Strengths: -A major strength of the study is that it measured the effects of chronic arsenic exposure in U.S. population. -Advantages of cohort design include the fact that the exposure precedes the effect being measured and that the cohort design has the ability to measure a variety of effects from a single type of exposure. Weaknesses: -Exposure assessment. -Study power. -Exposure to atmospheric arsenic and arsenic from food were potential confounder. Strengths: -Evaluated the effects of age, gender, race, educational attainment, smoking status, skin reaction to first exposure to the sun, history of radiotherapy (potential confounders). -Toenail concentrations individualize exposure and account for arsenic from other sources. Weaknesses: -Latency of arsenic- induced skin cancer unknown, follow-up period may have been inadequate. -Toenail arsenic measurements only account for recent past exposure. Reference/ Type of Study Lewis etal., 1999 Cohort Karagas et al., 2001 Case- control B-31 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1979- 1999 1994- 2000 Subjects/ Controls Not applicable 181 cases 328 controls Exposure Assessment Arsenic exposure categories (ppb): low = <10 medium = 10-25 high = 35-90 Exposure categories (ppb): 0-19 20-79 80-120 >120 Arsenic exposure indices: (1) highest average daily arsenic intake for any one year, (2) highest average daily arsenic intake averaged over any contiguous 5 years, (3) highest average daily arsenic intake averaged over any contiguous 20 years, and (4) total lifetime cumulative exposure Study Outcome SIR (95% CI), childhood leukemia and all childhood cancers excluding leukemia: Low-exposure group- leukemia = 1.02 (0.90-1.15) all cancers = 0.99 (0.92-1.07) Medium-exposure group: leukemia = 0.61 (0.12-1.79) all cancers = 0.82 (0.47-1.33) High-exposure group: leukemia = 0.86 (0.37-1.70) all cancers = 1.37 (0.96-1.91) Bladder cancer OR (95% CI): >80 ug/day = 0.94 (0.56-1.57) linear trend, p = 0.48 >80 jig/day, >40 years ago — smokers = 3.67 (1.43-9.42) linear trend, p <0.01 Strengths/ Weaknesses Strengths: -The analysis was stratified by age. -Low arsenic exposure study. -Findings were reported for different concentration ranges. Weaknesses: -Small study size. -Limitations of ecological study design. -Arsenic from food was not measured, leading to possible exposure misclassification. Strengths: -Potential confounders adjusted included gender, age, smoking history, education, occupation associated with elevated rates of bladder cancer, and income. -Use of cancer registry. -Individual exposure levels. Weaknesses: -Information bias (next- of-kin interviews). -Arsenic exposures outside the study area were not incorporated. -In the arsenic-exposed areas, the percentage of nonparticipants was 5% higher among cases than controls. This difference would probably mean that more exposed cases were missed in analyses of recent exposure, biasing the odds ratio toward the null. -Arsenic exposure from food was not considered. Reference/ Type of Study Moore et al, 2002 Ecological Steinmaus et al., 2003 Case- control B-32 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1999- 2000 Subjects/ Controls 368 cutaneous melanoma cases 373 colorectal cancer controls Exposure Assessment Median toenail arsenic concentration: cases = 0.06 ug/g, controls = 0.04 ug/g Study Outcome OR = 2.1(95%CI = 1.4-3.3, p-trend = 0.001) for increased risk of melanoma with elevated toenail arsenic concentrations OR = 6.6 (CI = 2.0- 2 1.9) for increased risk of melanoma with previous diagnosis of skin cancer and elevated toenail arsenic concentrations Strengths/ Weaknesses Strengths: -Potential confounders controlled for were age, gender, skin color/skin type, prior history of sunburn, education, and occupational exposure(s). -Ascertainment of cases and controls was accomplished by using the Iowa Cancer Registry, a Surveillance, Epidemiology, and End Results Program registry. This allowed newly diagnosed melanoma cases to be identified for a specific period and ensured a greater degree of certainty regarding the accuracy of diagnosis. -Toenail arsenic measurements individualize exposure and account for arsenic exposure from other sources. Weaknesses: -A limitation was that toenail samples were collected 2-3 years after diagnosis, resulting in possible exposure misclassification. Reference/ Type of Study Beane- Freeman et al., 2004 Case- control B-33 DRAFT—DO NOT CITE OR QUOTE ------- Study Period July 1, 1994 and June 30, 1998 Subjects/ Controls 383 transitional cell bladder cancer cases 641 controls Exposure Assessment Toenail arsenic level (ug/g): cases = 0.014- 2.484 controls = 0.009- 1.077 Study Outcome Odds ratio (95% CI)- bladder cancer among smokers: >0.330ug/g = 2.17 (0.92-5.11) Strengths/ Weaknesses Strengths: -Evaluated the following potential confounders: age, gender, race, educational attainment, smoking status, family history of bladder cancer, study period and average number of glasses of tap water consumed per day. -Conducted stratified analyses according to how long subjects used their current water system (<15 years, >15 years) to evaluate the possibility that an extended latency period is required for bladder cancer development. -Attempted to minimize misclassification by using biomarker (toenails). Weaknesses: -Possible misclassification at lower end of exposure range. -Limited data at extreme ends of exposure. -Lifetime exposure could not be calculated since data from previous residences could not be determined. Reference/ Type of Study Karagas et al., 2004 Case- control B-34 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1950- 1979 Subjects/ Controls 2,498,185 white males 1970 U.S. standard population Exposure Assessment Median water arsenic concentration (ppb): 3.0-3.9 4.0-4.9 5.0-7.4 7.5-9.9 10.0-19.9 20.0-49.9 50.0-59.9 Study Outcome Bladder cancer SMRs (95% CI), white males by median arsenic concentration in ground water (ppb): 3.0-3.9 = 0.95 (0.89-1.01) 4.0-4.9 = 0.95 ( 0.88-1.02) 5.0-7.4 = 0.97 (0.85-1.12) 7.5-9.9 = 0.89 (0.75-1.06) 10.0-19.9 = 0.90 (0.78-1.04) 20.0-49.9 = 0.80 (0.54-1.17) 50.0-59.9 = 0.73 ( 0.41-1.27) All levels combined = 0.94 (0.90-0.98) Strengths/ Weaknesses Strengths: -Large study population. -Study was nationwide. -Included over 75 million person-years of observation. Weaknesses: -No individual exposure data. -Assumed that study participants consumed local drinking water. -Available data assumed to represent actual arsenic content of water. -Analysis did not directly adjust for smoking, urbanization, and industrialization. -Arsenic contribution from food was not measured. Reference/ Type of Study Lammet al., 2004 Ecological B-3 5 DRAFT—DO NOT CITE OR QUOTE ------- Study Period July 2000- January 2002 Subjects/ Controls 6,669 residents Exposure Assessment Three arsenic exposure categories (ppb): <1.0 1.0-9.0 >10 Study Outcome Skin cancer adjusted odds ratio (95% CI): Arsenic level (ppb)— <1.0 = referent 1-9.9=1.81(1.10- 3.41) > 10 =1.92 (1.10- 3.68) Age (years) — 35-64 = referent > 65 = 4.53 (2.79- 7.38) Gender — female = referent males = 2.25 (1.33- 3.79) Cigarette use — no = referent yes= 1.37(0.84- 2.24) Strengths/ Weaknesses Strengths: -Large sample size. -History of individual tobacco use. -Arsenic well water analysis for each household. -Participants consumed water from the tested wells for at least 10 years. -Analysis controlled for age, gender, and tobacco use. Weaknesses: -Skin cancers were serf- reported and not confirmed by a medical records review. -Few people could provide information about specific types of cancer. -Families that participated may have been especially concerned about arsenic exposure or family members may have had existing health conditions. -Not controlled for sun exposure or occupation. -Arsenic contribution from food was not measured. Reference/ Type of Study Knobeloch et al., 2006 Cross- sectional B-36 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1979- 1997 Subjects/ Controls Residents of six Michigan counties Remainder of Michigan population as comparison Exposure Assessment Population- weighted mean arsenic concentration (ppb): exposed counties = 11.00 remainder of Michigan = 2.98 Study Outcome Elevated cancer SMRs (95% CI): Males — liver/biliary = 0.85 (0.72-1.00) trachea, bronchus, lung= 1.02 (0.98-1.06) melanoma = 0.99 (0.79-1.22) other skin cancer = 1.24 (0.86-1.72) bladder =0.94 (0.82-1.08) kidney /urinary = 1.06 (0.91-1.22) Females — liver/biliary =1.04 (0.89-1.20) trachea, bronchus, lung = 1.02 (0.96-1.07) melanoma = 0.97 (0.73-1.27) other skin cancer = 1.06 (0.60-1.72) female reproductive organs = 1.11* (1.03-1.19) bladder =0.98 (0.80-1.19) kidney /urinary organs = 1.00 (0.82-1.20) *p<0.01 Strengths/ Weaknesses Strengths: -Mortality data gathered from Michigan Resident Death Files for 20-year period. -Mortality rates stratified by gender, age, and race. Weaknesses: -Possible differences in reporting and classification of underlying causes of death. -No assessment of individual exposures and case migration. -Smoking and obesity, possible confounders, were not included in the analysis. -Preferential sampling based on home owners' request. -Arsenic contribution from food was not measured. Reference/ Type of Study Meliker et al., 2007 Ecological B-37 DRAFT—DO NOT CITE OR QUOTE ------- Table B-5. China cancer studies Study Period 1990 Subjects/ Controls 3,179 residents Exposure Assessment HAC (ppb): <10 10- 30- 50- 60- 100- 150- 500+ CAE (ppb-year): <10 10- 32- 100- 316- 1000- 3162- 10000+ Study Outcome Grade and (age- adjusted) skin cancer prevalence rates by HAC: <10 = 0.0 (0.0) 10- = 0.0 (0.0) 50- = 0.0 (0.0) 150-= 1.2(1.0) 500+ = 7. 1(5.9) Grade and (age- adjusted) skin cancer rates by CAE: <10 = 0.0 (0.0) 10- = 0.0 (0.0) 32- = 0.0 (0.0) 100- = 0.0 (0.0) 3 16- = 0.0 (0.0) 1000- = 0.4 (0.3) 3162- =0.8(0.2) 10000+ = 2.7 (2.0) Strengths/ Weaknesses Strengths: -Large study population. -Used both HAC and CAE in the analyses. -Arsenic concentrations measured in 184 wells. -Controlled for age and differences in cumulative arsenic exposure dose and duration of exposure. Weaknesses: -Possible recall and misclassification bias resulting from the collection of exposure histories through interviews. -Inherent limitations of ecological study design. -Did not control for sun exposure. Reference/ Type of Study Lammet al., 2007 Ecological B-3 8 DRAFT—DO NOT CITE OR QUOTE ------- Table B-6. Finland cancer studies Study Period 1981- 1995 Subjects/ Controls 61 bladder cancer cases and 49 kidney cancer cases 275 referents Exposure Assessment Water arsenic concentration (ppb): <0.1 0.1-0.5 >0.5 Arsenic daily dose (ug/day): 0.2 0.2-1.0 >1.0 Cumulative dose (V&): <500 500-2000 >2000 Study Outcome Bladder cancer risk ratios (95% CI): Shorter latency — Water arsenic concentration (ppb): 0.1-0.5= 1.53(0.75- 3.09) >0.5 = 2.44 (1.11- 5.37) Daily arsenic dose (ug/day): 0.2-1.0= 1.34(0.66- 2.69) >1.0= 1.84(0.84- 4.03) Cumulative dose (ug): 500-2000=1.61 (0.74-3.54) >2000= 1.50(0.71- 3.15) Longer latency — Water arsenic concentration (ppb): 0.1-0.5 = 0.81(0.41- 1.63) >0.5= 1.51(0.67- 3.38) Daily arsenic dose (ug/day): 0.2-1.0 = 0.76(0.38- 1.52) >1.0= 1.07 (0.48- 2.38) Cumulative dose (ug): 500-2000 = 0.81 (0.39-1.69) >2000 = 0.53 (0.25- 1.10) Strengths/ Weaknesses Strengths: -Cases were identified through the Finnish Cancer Registry. -The 1985 Population Census file of Statistics Finland was used to identify areas in which less than 10% of the population used the municipal water supply. -Risk ratios adjusted for age, gender, and smoking. Weaknesses: -Possible misclassification and possible recall bias resulting from the study choosing to use water consumption from the 1970. -Lacks other sources of arsenic exposure. Reference/ Type of Study Kurttio etal., 1999 Case-cohort B-39 DRAFT—DO NOT CITE OR QUOTE ------- Study Period 1985- 1988 and April 1999 Subjects/ Controls 280 incident bladder cancer cases 293 controls Exposure Assessment Arsenic exposure quartiles (ug/g)— 1:0.050 2:0.050-0.105 3:0.106-0.161 4:>0.161 Study Outcome Bladder cancer odds ratio (95% CI): highest vs. lowest quartile of toenail arsenic = 1.13, (0.70, 1.81) p trend = 0.65 for the highest vs. lowest quartile) Strengths/ Weaknesses Strengths: -Study used toenail arsenic as biomarkers of exposure. -Cases and controls matched according to age, toenail collection date, intervention group (alpha tocopherol and beta carotene), and smoking duration. -Study adjusted for matching factors, smoking, educational level, beverage intake, and place of residence. -Cut point of >0.09 ug/g used to avoid sample misclassification. -Potential confounders, including smoking cessation, smoking inhalation, educational level, beverage intake, and place of residence, were controlled for in the study analysis. Weaknesses: -Water intake was not included in the total beverage variable. -Toenail arsenic measures recent past exposures. Reference/ Type of Study Michaud et al., 2004 Cohort/nested case-control B-40 DRAFT—DO NOT CITE OR QUOTE ------- Table B-7. Denmark cancer studies Study Period 1970- 2003 Subjects/ Controls 39,378 Copenhagen residents 17,000 Aarhus residents Exposure Assessment TWA arsenic exposure (ppb) from 41 years old to date of enrollment: Copenhagen: min = 0.05 max= 15.8 Aarhus: min = 0.09 max = 25.3 Entire cohort: min = 0.05 max = 25.3 Study Outcome Cancer incidence rate ratios (95% CI): Time-weighted average exposure: Copenhagen — melanoma = 0.73 (0.46-1.14) non-melanoma = 1.09 (0.95-1.24) breast = 1.04(0.88- 1.22) Aarhus — melanoma = 0.85 (0.61-1.20) non-melanoma = 0.97 (0.90-1.05) breast =1.06 (1.01- 1.11) Cumulative exposure: Copenhagen — melanoma = 0.94 (0.81-1.08) non-melanoma = 1.01 (0.97-1.06) breast =1.0 1(0.95- 1.06) Aarhus — melanoma = 0.97 (0.90-1.05) non-melanoma = 0.98 (0.95-1.01) breast =1.01 (0.99- 1.03) Strengths/ Weaknesses Strengths: -Large study population. - Socioeconomic/demographic similarities of the cohorts. -Potential confounders adjusted were smoking, alcohol consumption, education, body mass index, daily intake of fruits/vegetables, red meat, fat and dietary fiber, skin reaction to the sun, hormone replacement therapy use, reproduction, occupation, and enrollment area. Weaknesses: -Possible misclassification bias. -Overall low arsenic concentration in drinking water in Denmark. -Lack of data regarding other sources of arsenic. Reference/ Type of Study Baastrup et al., 2008 Cohort B-41 DRAFT—DO NOT CITE OR QUOTE ------- Table B-8. Australia Cancer Studies Study Period 1982- 1991 Subjects/Control Victoria Cancer Registry cancer data Australian Bureau of Statistics denominator data Exposure Assessment Water/soil exposure groups: High water/high soil — >10ppb/>100 mg/kg High water/low soil — >10ppb/<100 mg/kg High soil/low water — <10ppb/>100 mg/kg Study Outcome Cancer SIRs (95% CI): Males and females — all cancers = 1.06 (1.03-1.09) prostate =1.14 (1.05-1.23) kidney =1.16 (0.98-1.37) melanoma = 1.36 (1.24-1.48) chronic myeloid leukemia = 1.54 (1.13-2.10) Females — breast = 1.10 (1.03-1.18) Strengths/ Weaknesses Strengths: -Study included both water and soil in exposure categories. -Twenty-two areas included in the study. Weaknesses: -Socioeconomic status, race, occupation and living in a rural area were possible confounders. -Possible exposure misclassification. -Ecological study limitations. Reference/ Type of Study Hinwood et al., 1999 Ecological B-42 DRAFT—DO NOT CITE OR QUOTE ------- APPENDIX C. TABLES FOR STUDIES ON POSSIBLE MODE OF ACTION FOR INORGANIC ARSENIC 1 This appendix contains three tables that deal with possible MOAs of arsenic in the 2 development of cancer based on in vivo human studies (Table C-l), in vivo experiments on 3 laboratory animals (Table C-2), and in vitro studies (Table C-3). They describe numerous 4 experiments published from 2005 through August 2007, as well as earlier experiments that were 5 mentioned in the Science Advisory Board Arsenic Review Panel comments of July 2007 (SAB, 6 2007), 2001 NRC document on arsenic (NRC, 2001), or a detailed early draft of this document 7 that lacked MOA tables. The data from these studies are distributed among 22 key-event 8 categories, with the data from different experiments from a single publication often being 9 summarized under different key-event categories. For example, the results in Wang et al. (1996) 10 are summarized by rows under Apoptosis, Cytotoxicity, and Effects Related to Oxidative Stress 11 (ROS). The advantage of distributing the data in this way is that it helped to focus on a 12 particular key event for each set of data. The disadvantage of using this approach is that it 13 spatially separated the different parts of each experiment. An exception to this procedure is the 14 category Immune System Response, in which results from different parts of each experiment are 15 presented in successive rows. 16 A brief discussion of the approaches and conventions used in preparing the tables is 17 included here. Abbreviations are used liberally in an attempt to reduce the size of the table. An 18 attempt was made to provide a summary of the main findings of each experiment, with the 19 expectation that any reader wanting more detail would read the publication. A search for any 20 specific citation should make it easy to pull together the information from the numerous parts of 21 some studies that related to different categories. Although, for example, cytotoxicity data are 22 generally summarized in the Cytotoxicity category, exceptions sometimes were made in an 23 attempt to decrease the size of the table. For example, if data presented on apoptosis contained 24 only slight, but interesting, data on cytotoxicity, a brief summary of those cytotoxicity findings 25 was sometimes added at the end of the results column in the row that described the results on 26 apoptosis. When an experiment that tested only one concentration yielded interesting results, the 27 results column is sometimes merged with one or more columns to its left in that same row so the 28 long description of results did not drastically increase the height of the table. In such a case, the 29 only dose tested was obviously the LOEC or LOEL. 30 In vivo experiments on laboratory animals were almost always restricted to experiments 31 in which the route of exposure was oral. In most cases this meant that the arsenical was 32 administered in drinking water or was given by gavage. A few experiments had the arsenical in 33 the feed. Two experiments on chicken embryos had a solution (with concentration in uM) put 34 onto the embryo, and one genetic assay done on Drosophila melanogaster had the concentration C-1 DRAFT—DO NOT CITE OR QUOTE ------- 1 (given in mM) reported for the media. All other in vivo experiments were done on mice or rats. 2 Numerous studies were excluded on other non-mammalian species, including, for example, fish, 3 nematodes, and algae. 4 Tables C-2 and C-3 list all doses or concentrations tested as well as the duration of 5 testing. It was often necessary to estimate the concentrations or doses tested from figures. For 6 brevity, the control dose of 0 is not listed as a concentration tested. In the rare instances in 7 which there was no zero-dose control group, this omission is mentioned in the results section. In 8 many cases the papers themselves did not specify the LOECs or LOELs, and those values were 9 estimated from tables or figures. Because of the large variation in the way that papers presented 10 data and variability in their findings, and because of the rather common failure to clearly define 11 the error bars around data points in figures, there was often subjectivity involved in selecting the 12 LOEC or LOEL. There was no strict requirement that the LOEC or LOEL declared for each 13 experiment had to be shown to be statistically significantly higher than the control, although it 14 was not uncommon for that to be the case. The wording in the results column often helps to 15 clarify this situation. If six concentrations were tested, for example, and if the second from the 16 lowest concentration had error bars that did not overlap those of the control, and if the third from 17 the lowest concentration was identified as being statistically significantly higher then the control, 18 then the second from the lowest concentration tested would have been declared the LOEC. The 19 LOEC, for example, should be viewed as the lowest concentration that was "quite likely" to have 20 caused an effect—without any specific statistical interpretation being attached to it. As long as 21 this was made clear, it was felt that this approach would be most useful to readers who want to 22 know the lowest concentration level at which a particular effect would probably occur. 23 Arrows are used to indicate changes that were increases or decreases from the control. If 24 the change was relative to some other group, it was clearly indicated as such. In most cases, the 25 changes in magnitude of effects relative to the control were described as, for example, "2.34x" or 26 "0.46x"—2.34 times higher than the control or only 46% as high as the control. When those 27 ratios were based on estimates made from a graph, they are generally preceded by a "~" mark; if 28 they were calculated from tabulated values, they are generally presented without that mark. 29 In Table C-2 the doses are presented in terms of the amount of arsenic. When doses were 30 reported in mg arsenic/L or in ppm As, it was assumed that the doses included adjustment to 31 determine the amount of arsenic administered. In a few publications it was unclear if the 32 reported doses were for the compound or for the amount of arsenic administered. Partly because 33 of this uncertainty, all doses shown in the table that were corrected to the amount of arsenic from 34 values that were clearly reported as concentrations of some arsenical compound (or for which 35 that was assumed to be the case) are preceded by an asterisk. Species of arsenic are shown in 36 Tables C-2 and C-3, and Asv is almost always sodium arsenate. 37 C-2 DRAFT—DO NOT CITE OR QUOTE ------- Abbreviations for Tables in Appendix C ft 1RB3AN27 cells 1T1 cells 293 cells 2-AAAF 2BS cells 3-NT 4HNE 4NQO 5-aza-dC 6-4 PPs 7-AAD 8-OHdG 8-oxoG A2780 cells A431 cells A5/SG assays A549 cells AA AB assay ABTS AC ADM ADSB AFP AG06 cells AGT Ahr+l+ MEFs Aktl ALAD ALAS increase decrease approximately (if before a listing of concentrations, it applies to all) approximately equal an immortalized dopamine-producing rat mesencephalic cell line a human epithelial cell line a cell line derived from adenovirus-transformed human embryonic kidney epithelial cells 2-acetoxyacetylaminofluorene human fetal lung fibroblasts 3-nitrotyrosine 4-hydroxy-2-nonenal 4-nitroquinoline 1-oxide 5-aza-deoxycytidine, a demethylating agent 6-4 photoproducts (UV-induced DNA photoproduct) 7-aminoactinomycin D 8-hydroxy-2'-deoxyguanosine or 8- hydroxydeoxyguanosine (synonym) 7,8-dihydro-8-oxoguanine human ovarian carcinoma cell line human epidermoid carcinoma cell line A5 (Annexin V-Alexa568) and SG (a green fluorescent DNA dye) staining assays; A5+/SG- cells are apoptotic human non-small cell lung cancer (NSCLC) cell line (alveolar basal epithelial cell line) ascorbic acid (vitamin C) AlamarBlue assay 2,2'-azinobis(3-ethylbenzothiazoline)-6-sulfonic acid arsenic chloride adriamycin apparent DNA strand break a-fetoprotein SV40-transformed human keratinocytes average generation time mouse embryo fibroblasts of genotype Ahr+ + from C57BL/6J mice, which are cells known to respond to a B(a)P or TCCD challenge by activation of the AhR V-akt murine thymoma viral oncogene homolog 1 (a human gene) 5-aminolevulinic acid dehydratase 5-aminolevulinic acid synthetase C-3 DRAFT—DO NOT CITE OR QUOTE ------- AL hybrid cells AMs AML AMPK AO APE/Ref-1 AP-PCR Aprt AP sites AR230 cells AR230-r cells AR230-S cells ARE AS52 cells in As As: Asv ASK1 ATO B0653 B16-F10 cells BAEC BALF B[or]P BCS BEAS-2B cells BER BFTC905 cells BFU BHMT BHT Bid BPDE BrdU BSO BUG a cell line that contains structural set of CHO-K1 chromosomes and one copy of human chromosome 11 alveolar macrophages acute myelogenous leukemia adenosine monophosphate-activated protein kinase acridine orange apurinic/apyrimidinic endonuclease (hAPEl) arbitrarily primed polymerase chain reaction adenosine phosphoribosyl transferase sites of base loss (apurinic/apyrimidinic [AP] sites) a CML cell line that expresses large amounts of Bcr- Abl AR230 cells that are resistant to the Bcr-Abl inhibitor imatinib mesylate AR230 cells that are sensitive to the Bcr-Abl inhibitor imatinib mesylate antioxidant response element a pSV2 gpt-transformed Chinese hamster ovary cell line; cells in this line carry a single copy of a transfected E. coli gpt gene arsenic arsenite arsenate apoptosis signal-regulating kinase 1 arsenic trioxide 2,3-dihydro-5-hydroxy-2,2-dipentyl-4,6-di-tert- butylbenzofuran mouse melanoma cells bovine aortic endothelial cells bronchoalveolar lavage fluid benzo[a]pyrene bathocuproinedisulphonic acid human bronchial (pulmonary) epithelial cell line base excision repair a human urothelial carcinoma cell line burst-forming units betaine-homocysteine methyltransferase butylated hydroxytoluene a BH3 domain-containing proapoptotic Bcl2 family member that is a specific proximal substrate of CaspS in the Fas apoptotic signaling pathway benzo[a]pyrene diol epoxide bromodeoxyuridine L-buthionine-S,R- sulphoximine (depletes GSH, y- GCS inhibitor) bladder urothelial cells C-4 DRAFT—DO NOT CITE OR QUOTE ------- C-33A cells CAM CAM assay CAs CAT Cdc Cdc42 cen+ cen- CFE c-Fos CFSE CFU CGL-2 cells cGpx Chang cells ChAT CHO CI c-Jun or c-jun CK8 CL3 cells CL3R15 cells c-met c-Mos CM-H2DCFDA CML Cone Contraspin COS-7 cells CoTr COX COX-2 CPDs Cpp32 CREEP a transformed human non-differentiated carcinoma cell line cell adhesion molecule chorioallantoic membrane assay of angiogenesis chromosome aberrations catalase (decomposes H^C^) cell division cycle a small GTPase in the Rho/Rac subfamily of Ras- like GTPases centromere positive (micronuclei) centromere negative (micronuclei) colony-forming efficiency an AP-1 protein 5,6-carboxyfluorescein diacetate succinimidyl ester colony-forming units a cell line derived from a hybrid (ESH5) of the HeLa variant, D98/AH2, and a normal human fibroblast strain, GM77 cellular glutathione peroxidase a human cell line thought to be derived from HeLa cells choline acetyltransferase Chinese hamster ovary confidence interval an AP-1 protein cytokeratin 8 human lung adenocarcinoma cells (established from a non-small-cell lung carcinoma) cell line derived from CL3 cells that were maintained in 4 uM arsenic SA the oncogene that encodes HGF (hepatocyte growth factor) receptor proto-oncogene 5-(and-6)-carboxy-2',7'-dichlorofluorescein diacetate chronic myeloid leukemia concentration a serine—or cysteine—proteinase inhibitor isoform African green monkey kidney fibroblast cell line containing 10,000 glucocorticoid receptors per cell that are transcriptionally inactive co-treatment cytochrome c oxidase; its activity is a measure of mitochondrial function cyclooxygenase-2 cyclobutane pyrimidine dimers (UV-induced DNA photoproduct) caspase-3 cAMP response element binding protein C-5 DRAFT—DO NOT CITE OR QUOTE ------- CRL1675 cells CRL-1609 cells cRNA CSTP Cul3 CV assay CYP1A1 CYP7B1 DA DAP DCF assay DCFH-DA DCHA DEB DENA DES Dex DHA dhfr gene DHR123 DIG DI-I or II or m DKO dL DMA in DMA DMAmI DMBA DMN DMNQ DMPO DMPS DMSA DMSO DNA DNA-PK D-NMMA DNMT a human melanocyte cell line chimpanzee transformed skin fibroblast cells RNA derived from complimentary DNA through standard RNA synthesis clonal survival treat and plate Cullin 3, an Nrf2-fmding protein crystal violet assay; it measures cellular protein, which is related to cell number cytochrome P450 1 Al cytochrome P450 family 7, subfamily b polypeptide 1 disodium arsenate 2,6-diaminopurine dichlorofluorescein assay 2',7'-dichlorofluorescein diacetate docosahexaenoic acid, a co-3 polyunsaturated fatty acid vital for the developing nervous system diepoxybutane (DNA crosslinking agent) diethylnitrosamine diethylstilbestrol dexamethasone (synthetic glucocorticoid) dehydroascorbic acid dihydrofolate reductase gene dihydrorhodamine 123 dicumarol, andNqol inhibitor iodothyronine deiodinase-I or II or m (are 3 forms of this selenoenzyme) double knock out deciliter dimethylarsenous acid dimethylarsinic acid dimethyl arsenic (used when the oxidative state is unknown or not specified) dimethylarsinous iodide dimethylbenzanthracene dimethylnitrosamine 2,3-dimethoxy-l,4-naphthoquinone 5,5'-dimethyl-1-pyrroline TV-oxide (a spin-trap agent) 2,3-dimercaptopropane-l-sulfonic acid dimercaptosuccinic acid or meso 2,3- dimercaptosuccinic acid dimethyl sulfoxide deoxyribonucleic acid DNA-dependent protein kinase, which has 3 subunits, of which the Ku70 protein is one NG-methyl-D-arginine, the inactive enantiomer of a nitric oxide synthase inhibitor DNA methyltransferase C-6 DRAFT—DO NOT CITE OR QUOTE ------- DPC DPI DPIC DR DRE-CALUX DSB DTNB DTT DU145 cells DW E2N E7 cells EA EB E. coli EDR3 cells EGCG EOF EGFR EGFRECD EGR elF eIF4E ELISA Emodin EMSA Enm eNOS ER-a ERCC1 ERCC2 Erk or ERK EROD ESR ETU FACS FADD FAK FBS DNA protein crosslinks diphenyleneiodonium diphenylene iodonium chloride, an NADPH-oxidase inhibitor death receptor dioxin-responsive element (DRE)-mediated Chemical Activated LUciferase expression double strand break (in DNA) 5,5'-dithiobis(2-nitrobenzoicacid) dithiothreitol a human prostate carcinoma cell line drinking water ubiquitin-conjugating enzyme an immortalized human bladder cell line ethacrynic acid (a GST inhibitor) ethidium bromide Escherichia coli a rat hepatoma cell line (glucocorticoid receptor negative, with neither protein nor mRNA detectable) (-)-epigallocatechin gallate epidermal growth factor epidermal growth factor receptor extracellular domain of the epidermal growth factor receptor early growth response eukaryotic initiation factor eukaryotic translation initiation factor 4E, which is the mRNA cap binding and rate-limiting factor required for translation enzyme-linked immunosorbent assay (l,3,8-trihydroxy-6-methylanthraquinone) electrophoretic mobility shift assays endonuclease m endothelial nitric acid synthase estrogen receptor-a excision repair cross-complement 1 component excision repair cross-complementing rodent repair deficiency, complementation group 2 (also known as xeroderma pigmentosum group D or XPD) extracellular signal-regulated kinase ethoxyresorufin-O-deethylase electron spin resonance S-ethylisothiourea, a NOS inhibitor fluorescence-activated cell sorting Fas-associated death domain protein focal adhesion kinase fetal bovine serum C-7 DRAFT—DO NOT CITE OR QUOTE ------- FeTMPyP 5,10,15,20-tetrakis (jV-methyl-4'-pyridyl) porphinato iron(in) chloride (ONOCT decomposition catalyst) FGC4 cells rat hepatoma cells FGF-2 fibroblast growth factor -2 FGFR1 fibroblast growth factor receptor 1 FISH fluorescent in situ hybridization FITC fluorescein isothiocyanate FLIP FLICE-inhibitory protein, an antiapoptotic protein controlled by NF-KB FLIPL long-splice variant of FLIP Fox O3a an oxidative stress inducible forkhead transcription factor FPG formamidopyrimidine-DNA glycosylase (digestion ofDNA) G12 cells a pSV2gpt-transformed Chinese hamster V79 (hprf) cell line G6PDH glucose-6-phosphate dehydrogenase G-6-P glucose-6-phosphatase; the paper that presented data on this chemical called it G-6-PD in the discussion GADD growth arrest and DNA damage-inducible GCLM glutamate cysteine ligase modifier, GCLM knockout mice (-/-) have only 9%-16% of GSH level of wt littermates GCR glucocorticoid receptor GFP green fluorescent protein (GFP expressing tumor cells) GLN glutamine GlycoA glycophorin A GM043 1 2C a S V-40 transformed XP A human fibroblast NER- cells deficient cell line GM847 cells a SV-40-transformed human lung fibroblast cell line GM-CSF granulocyte-macrophage colony-stimulating factor GM-Mp GM-type macrophage gpt guanine phosphoribosyltransferase GPx glutathione peroxidase GR glutathione reductase GRE glucocorticoid response elements GSH glutathione GSSG glutathione disulfide GST glutathione-S-transferase GTP guanosine-5'-triphosphate Gy gray (unit of ionizing radiation) HI 3 55 cells a human lung adenocarcinoma cell line H2C>2 hydrogen peroxide H22 cells a hepatocellular carcinoma cell line H41 IE cells a rat hepatoma cell line H460 cells a human non-small-cell lung cancer cell line (also called human lung large cell carcinoma cells) C-8 DRAFT— DO NOT CITE OR QUOTE ------- H9c2 cells HaCaT cells Hb HCC HCT116 cells HCT15 cells HEC HEK 293 cells HEK293T cells Hepa-lclc? cells HepG2 cells HeLa cells HeLa S3 cells HELP cells HEL cells hEp cells HFF cells HFW cells HGF HGPRT HIF HK-2 cells HL-60 cells HLA HLA-DR HLF cells HLFC cells HLFK cells HMEC-1 cells HMOX-1 HO- HOS cells an immortalized myoblast cell line derived from fetal rat hearts a human epidermal keratinocyte cell line hemoglobin hepatocellular carcinoma a human colorectal cancer cell line (available in securin-wild-type and securin-null forms) a human colon adenocarcinoma cell line hamster embryo cells an adenovirus-transformed human embryonic kidney epithelial cell line (non-tumor cells), also called HEK293 cells human embryonic kidney cells a mouse hepatoma cell line known to respond to a B[a]P or TCCD challenge by activation of the AhR a human hepatocellular liver carcinoma cell line (Caucasian) a human cervical adenocarcinoma cell line a human cervical carcinoma cell line, derived from the parent HeLa cell line; adapted to grow in suspension (spinner) culture and has the same virus susceptibility as the parent line a human embryo lung fibroblast cell line an AML cell line that is a cytokine-independent human erythroleukemia cell line that has constitutive STAT3 activity normal human epidermal cells derived from foreskin a human foreskin fibroblasts cell line a diploid human fibroblast cell line hepatocyte growth factor hypoxanthine-guanine phosphoribosyltransferase hypoxia inducible factor a human proximal tubular cell line human promyelocytic leukemia cells human leukocyte antigen human leukocyte antigen DR, which is a major histocompatibility complex class-II antigen human embryo lung fibroblasts an HLF subline that is not Ku70 deficient; it has the null pEGFP-Cl vector transferred into it an HLF subline that is Ku70 deficient; it has a recombinant plasmid of Ku70 gene antisense RNA transferred into it; it had 38% as much Ku70 protein content as the HLFC cell line human microvascular endothelial cells heme oxygenase 1 hydroxyl radicals a human osteogenic sarcoma cell line C-9 DRAFT—DO NOT CITE OR QUOTE ------- Hpall or HP All HPBM HPLC HPRT HRE Hr HSF1 HSP HT1080 cells hTER hTERT HT1197 cells HU Huh? cells HuR HUVEC cells IAP iAs icAA ICAM-1 ICE IC50 ID1 IEC cells IEC-6 cells IGF IGFBP-1 IKKP IL ILK Imatinib IM9 cells IRE IRP-1 J82 cells JAK JAR cells JB6C141 cells JB6C141PG13 cells Haemophilusparainfluenzae (restriction endonucleases) human peripheral blood monocytes high-performance liquid chromatography hypoxanthine phosphoribosyl transferase hypoxia response element, the DNA binding element of HIF-mediated transactivation hour(s) heat shock transcription factor 1 heat shock protein a human sarcoma cell line RNA component of telomerase human telomerase reverse transcriptase a human (Caucasian) epithelial bladder cancer cell line hydroxyurea a human hepatoma cell line RNA binding protein a human umbilical vein endothelial cell line (or HUVECs) inhibitor of apoptosis protein family inorganic arsenic intracellular ascorbic acid, which is accumulated at up to high concentrations by culturing cells in DHA inter-cellular adhesion molecule-1 interleukin-lp-converting enzyme concentration that causes 50% inhibition of activity inhibitor of DNA binding-1 a primary culture of rat intestinal epithelial cells a rat intestinal epithelial cell line insulin growth factor (system) insulin-like growth factor binding protein 1 inhibitor of kappa light polypeptide gene enhancer in B-cells, kinase beta; also called IkappaB kinase beta subunit interleukin integrin-linked kinase imatinib mesylate a human multiple myeloma cell line iron responsive element iron regulatory protein 1 human bladder tumor cells Janus kinase a human placental choriocarcinoma cell line a P+ mouse epidermal cell line (sometimes called JB6 Cl 41 cells) stable p53 luciferase reporter plasmid transfectant of cell line JB6C141 C-10 DRAFT—DO NOT CITE OR QUOTE ------- JB6C141P+1-1 cells JC-1 JNK Jurkat cells K1735-SW1 cells K562 cells KCL22 cells KCL22-r cells KCL22-S cells kDa Keapl KMS12BM cells Ku70 L-132 cells LAK cells LCL-EBV cells LDH LD50 LI LOEC LOEL LOH LPO Luc LU1205 cells Z-NAME L-NMMA LPS LTE4 Lys Maf stable activator protein-1 (AP-1) transfectant of cell lineJB6C141 voltage-sensitive lipophilic cationic fluorescence probe 5,5',6,6'-tetrachloro-l,l',3,3'- tetraethylbenzimidazolcarbocyanine iodide c-Jun N-terminal kinase a transformed human T-lymphocyte cell line (also called lymphoblast cells) a mouse melanoma cell line a human immortalized myelogenous leukemia cell line that is a bcrabl positive erythroleukemia line derived from a 53-year-old female CML patient in blast crisis a Bcr-Abl positive CML cell line KCL22 cells that are resistant to the Bcr-Abl inhibitor imatinib mesylate KCL22 cells that are sensitive to the Bcr-Abl inhibitor imatinib mesylate kilodalton, a unit of mass the cytoplasmic Nrf2-binding protein a human multiple myeloma cell line one of the three subunits of DNA-dependent protein kinase human alveolar type II cells lymphokine activated killers (effector cells) mononuclear cells obtained from healthy donors and transformed by Epstein-Barr virus 50% lethal concentration lactate dehydrogenase 50% lethal dose labeling index lowest observed effect concentration lowest observed effect level loss of heterozygosity lipid peroxidation the PEPCK-luciferase construct a human melanoma cell line 7Va>-nitro-L-arginine methyl ester (an inhibitor of NOS) NG-methyl-L-arginine, the active enantiomer of a nitric oxide synthase inhibitor lipopolysaccharide leukotriene, a proinflammatory mediator lysine musculoaponeurotic fibrosarcoma (transcription factor) C-11 DRAFT—DO NOT CITE OR QUOTE ------- MAP MAPK MCA MC/CAR cells MCF-7 cells MCR M-CSF MDA MDAH 2774 cells MDA-MB-231 cells MDA-MB-435 mdm2 MDR MED MEF MEF cells MEK MGC-803 cells MI MiADMSA min MK-571 MKP-1 MMA MMA111 MMAinO MMAV MMC MMP MMP-2 MMP-9 MMP-13 MMS MN MNNG MnTMPyP MNU MRC-5 cells mRNA MRP mitogen-activated protein mitogen-activated protein kinase 20-methylcholanthrene a human multiple myeloma cell line human breast carcinoma cell line mineralocorticoid receptor macrophage colony-stimulating factor malondialdehyde (the thiobarbituric acid-reactive substance in the brain that reflects extensive lipid peroxidation) human ovarian carcinoma cells a human breast cancer cell line (an invasive estrogen unresponsive cell line) a human metastatic breast cancer cell line murine double minute 2 proto-oncogene multidrug resistance gene minimal erythemic dose mouse embryo fibroblasts a mouse embryonic fibroblast cell line MAP/ERK kinase (also, a family of related serine- threonine protein kinases that regulate mitogen- activated protein kinase) a human gastric cancer cell line mitotic index monoisoamyl meso 2,3- dimercaptosuccinic acid minutes(s) MRP antagonist MAP kinase phosphatase 1 monomethyl arsenic (used when oxidative state is unknown or not specified) monomethylarsonous acid methylarsine oxide monomethyl arsonic acid mitomycin C mitochondrial membrane potential matrix metalloproteinase-2 matrix metalloproteinase-9 matrix metalloproteinase-13 methyl methanesulfonate micronuclei 1 -methyl-3 -nitro-1 -nitrosoguanidine Mn(m)tetrakis(l-methyl-4-pyridyl)porphyrin pentachloride (a cell permeable SOD mimic) N-methyl-N-nitrosourea a human lung fibroblast cell line messenger ribonucleic acid multidrug resistance-associated protein C-12 DRAFT—DO NOT CITE OR QUOTE ------- Mrps MS MT mtDNA MTOC MTS assay MTT MIX MT-1 MT2A MW MYH MYP3 cells N-18 cells NAC NADH NADPH Namalwa cells NB4 cells NB4-AsR NB4-M-AsR2 cells NCE NCI cells NE NER NF-KB NHEK cells NIH3T3 cells NO' NOS Nqol NR Nrf2 efflux transporters encoded by MRP genes mass spectrometer or mass spectrometry metallothionein mitochondrial DNA microtubule-organizing center 3-(4,5-dimethylthiazol-2-yl)-5-(3- carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H- tetrazolium, inner salt assay; in Yi et al. (2004) study this was referred to as the CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (MTS) Kit (Promega, Madison, WI) 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide methotrexate metallothionein-1 gene symbol for metallothionein 2A molecular weight MutY homolog, an endonuclease rat epithelial cells line (urinary bladder cells) a mouse neuroblastoma cell line w-acetyl-cysteine (precursor of GSH; it elevates cellular GSH levels, also an antioxidant), also N- acetyl-Z-cysteine reduced form of nicotinamide adenine dinucleotide nicotinamide adenine dinucleotide phosphate- oxidase a human Burkitt' s lymphoma cell line a human acute promyelocytic leukemia cell line an arsenic-resistant subline of NB4 that was made by culturing and maintaining cells in luM As2C>3 an arsenic-resistant human acute promyelocytic leukemia cell line, which is routinely grown in RPMI 1640 media containing 2 uM As2C>3 normochromatic erythrocytes a human myeloma cell line nuclear extract nucleotide excision repair (pathway) nuclear factor-kappa B primary normal human epidermal keratinocytes a mouse fibroblast cell line nitric oxide nitric oxide synthase nicotinamide adenine dinucleotide phosphate- quinone oxidoreductase (or NAD(P)H-quinone oxidoreductase) neutral red cap 'n' collar basic leucine zipper transcription factor (nuclear factor erythroid 2-related factor 2) C-13 DRAFT—DO NOT CITE OR QUOTE ------- NSAID NSE NTUB1 cells NuF OATP-C ODA OGG1 OM431 cells ONOO OR p21 PAEC cells PAI-1 PARP PBMC PC PC12 cells PCE PCI-1 cells PCNA PCR PDH PDT PD98059 PEG PEPCK pEpREpgeo PGE2 P-gP PHA PHEN PI PI3K PK FLAP PLC/PR/5 cells PMA PMN non-steroidal anti-inflammatory drug no significant effect (often not based on a statistical test but on whether an effect appears likely to be real based on examination of graphs) a human urothelial carcinoma cell line nuclear fragmentation organic anion transporting polypeptide-C oxidative DNA adducts 8-oxoguanine DNA glycosylase a human melanoma cell line peroxynitrite odds ratio a cyclin-dependent kinase inhibitor porcine aortic endothelial cells plasminogen activator inhibitor-1 poly(adenosine diphosphate-ribose) polymerase peripheral blood mononuclear cell (human) protein carbonyl (form of protein oxidation) a rat sympathetic (neuronal) pheochromocytoma cell line polychromatic erythrocyte a human head and neck squamous cell carcinoma cell line proliferating cell nuclear antigen polymerase chain reaction pyruvate dehydrogenase population doubling time inhibitor of MEK1/2, which are ERK upstream kinases (structurally unrelated to U0126) monomethoxypolyethylene glycol (covalent attachment of PEG to CAT or SOD extends their plasma half-lives) phosphoenolpyruvate carboxykinase gene (a hormone-inducible gene) p-galactosidase-neomycin-resistance reporter plasmid prostaglandin E2 P-glycoprotein, the efflux transporter encoded by MDR phytohemagglutinin o-phenanthroline (an iron chelator) propidium iodide phosphatidylinositol 3-kinase proteinase K placental alkaline phosphatase a human hepatocellular carcinoma cell line phorbol 12-myristate 13-acetate polymorphonuclear neutrophils (or PMNs) C-14 DRAFT—DO NOT CITE OR QUOTE ------- PMs PNA ppb P-PKB ppm PQ PR PRCC PSH P-STAT3 pt PTEN p-XSC R-3T3 cells Rac RACs Raf RAGE RANKL RAPD-PCR Ras RAW264.7 cells RBC RFU RHMVE cells RI RKO cells ROCK RNA RNS ROS RPMI-8226 cells RT-PCR RWPE-1 cells SA SACs peritoneal macrophages peptide nucleic acid parts per billion phosphorylated protein kinase B parts per million paraquat (a generator of 02") progesterone receptor primary renal cortical cell protein thiol phosphorylated-STAT3 pretreatment phosphatase and tensin homolog (mutated in multiple advanced cancers 1) l,4-phenylenebis(methylene)selenocyanate Ras-transformed NIH 3T3 cells, a mouse fibroblast cell line a subfamily of the Rho family of GTPases, which are small (-21 kDa) signaling G proteins (more specifically GTPases). rapidly adhering cells; epidermal cells with the highest proliferative potential and with properties of stem cells a proto-oncogene receptor for advanced glycation end products receptor activator of NFicB ligand random(ly) amplified polymorphic DNA polymerase chain reaction a name of a proto-oncogene a mouse macrophage cell line (another source described it as mouse macrophage-like cells) red blood cell, erythrocyte relative fluorescence units (units of ROS) rat heart microvessel endothelial cells replicative index a human colorectal carcinoma cell line that expresses wild-type p53 proteins Rho/kinase, and effector molecule of RhoA ribonucleic acid reactive nitrogen species reactive oxygen species a human myeloma cell line reverse transcription-polymerase chain reaction human prostate epithelial cell line sodium arsenite slowly adhering cells; epidermal cell fraction that contains cells undergoing terminal differentiation, with little ability to form colonies C-15 DRAFT—DO NOT CITE OR QUOTE ------- SAH SAM SCC SCE SCGE Se SE SEM Se-Met Ser SF SFN SHE cells SIK cells siRNA SLC30A1 SMART SMC cells socs SOD SP SRB assay Src SSB STAT StRE site SU5416 SVEC4-10 cells SV-HUC-1 cells SV-40 SW13 cells SW480 cells SY-5Y cells T3 T4 T47D cells TAM S-adenosylhomocysteine S-adenosylmethionine squamous cell carcinoma sister chromatid exchange single cell gel electrophoresis (assay) selenium standard error of the mean scanning electron microscopy selenomethionine serine, an amino acid sodium formate, an 'OH radical scavenger sulforaphanem, an activator of transcription factor Nrf2, which plays a critical role in metabolism and excretion of xenobiotics Syrian hamster ovary cells spontaneously immortalized human keratinocytes (or epidermal cells) small interfering RNA (ribonucleic acid) gene symbol for the zinc transporter, solute carrier family 30, member 1 somatic mutation and recombination test human bladder smooth muscle cells suppressors of cytokine signaling superoxide dismutase (an antioxidant to (V") shock protein sulforhodamine B colorimetric assay first oncogene discovered, the transforming protein of the chicken retrovirus, Rous sarcoma virus single strand break (in DNA) signal transducer and activator of transcription stress response element recognition site inhibitor of VEGF receptor-2 kinase a C3H/HeN mouse vascular endothelium cell line (also called immortalized mouse endothelial cell line) an SV40 large T-transformed human urothelial cell line (non-tumor cells, derived from urethra, immortalized) simian virus 40 a human adrenal carcinoma cell line a colorectal adenocarcinoma cell line derived from a Caucasian male that has two base-pair substitution mutations in the p53 gene a human neuroblastoma cell line thyroid hormone triiodothyronine thyroid hormone thyroxine a human mammary adenocarcinoma cell line tamoxifen C-16 DRAFT—DO NOT CITE OR QUOTE ------- TAT TEARS tBHQ TCDD TF Tg.AC TGF THP-1 + A23187 cells TIG-112 cells TIMP-1 Tiron TK6 cells TM TMAVO TM3 cells TNF-a TPA TR9-7 cells TRAIL TRAIL-R TRAP TRF TRL 1215 cells Trolox Trx TrxR TrxRl Trxl Trx2 TUNEL assay tyrosine aminotransferase thiobarbituric acid reactive substances (a measure of tissue lipid peroxidation) ^-butylhydroquinone 2,3,7,8-tetrachlorodibenzo-p-dioxin theaflavin strain of transgenic mice that contains the fetal beta- globin promoter fused to the v-Ha-ras structural gene (with mutations at codons 12 and 59) and linked to a simian virus 40 polyadenylation/splice sequence transforming growth factor a human dendritic cell line; THP-1 cells acquire the characteristics of dendritic cells in the presence of the calcium ionophore A23187 human normal skin diploid cells tissue inhibitor of metalloproteinase-1 4,5-dihydroxy-w-benzenedisulfonic acid, disodium salt human lymphoblastoid cells tail moment trimethylarsine oxide immortalized Ley dig cells derived from normal mouse testis tumor necrosis factor a (an inflammatory cytokine) 12-O-tetradecanoylphorbol-13-acetate a spontaneously immortalized human fibroblast cell line, derived from a Li-Fraumeni patient, and subsequently stably transfected with a tetracycline- regulated p53 expression vector TNF-related apoptosis-inducing ligand TRAIL receptor tartrate resistant acid phosphatase (RAW264.7 cells can undergo osteoclast differentiation, which is accompanied by an increase in the number of multinucleate cells expressing TRAP) terminal restriction fragment nontumorigenic adhesive rat epithelial liver cells originally derived from the liver of 10-day-old Fisher F344 rats 6-hydroxy-2,5,7,8- tetramethylchroman-2-carboxylic acid thioredoxin thioredoxin reductase cytosolic thioredoxin reductase cytoplasmic thioredoxin-1 mitochondrial thioredoxin-2 terminal deoxynucleotidyl transferase-mediated deoxyuridine nick-end labeling assay C-17 DRAFT—DO NOT CITE OR QUOTE ------- U0126 U118MG cells U266 cells U937 cells U-937 cells U-2OS cells Ub UROtsa cells UV UVA UVB UVC V79 cells VEGF VEGFR1 VEGFR2 V-FITC VH16 vs. VSMC W138 wk wt WM9 cells WRL-68 WT-1 XIAP XPA (B or F) XRS XTT YC-1 yptlocus Z-DEVD-FMK ZPP inhibitor of MEK1/2, which are ERK upstream kinases (structurally unrelated to PD98059) a human glioblastoma cell line, also called Ul 18MG (ATCCHTB-lS)cells a human multiple myeloma cell line a human leukemic monocyte lymphoma cell line (also described as a human promonocytic cell line or as a human myeloid leukemia cell line) human diffuse histiocytic lymphoma cells, perhaps the same as U937 cells a human osteogenic sarcoma cell line ubiquitin an SV40-immortalized human urothelium cell line ultraviolet radiation ultraviolet radiation A ultraviolet radiation B ultraviolet radiation C a cell line derived from lung fibroblasts of a male Chinese hamster vascular endothelial growth factor or vascular endothelial cell growth factor a vascular endothelial cell growth factor receptor (flt-1) a vascular endothelial cell growth factor receptor (Flk-1, KDR) V-fluorescein isothiocyanate human primary fibroblasts versus vascular smooth muscle cells a human diploid lung fibroblast cell line week(s) wild-type a human melanoma cell line a human hepatic cell line Wilm's tumor protein-1 X-linked inhibitor of apoptosis protein, an antiapoptotic protein controlled by NF-KB xeroderma pigmentosum, complementation group A (B or F) X-ray sensitive 2,3-bis[2-methyloxy-4-nitro-5-sulfophenyl]-2H- tetrazolium-5-carboxanilide a small molecule inhibitor of HIF signaling xanthine-guanine phosphoribosyltransferase locus benzyloxycarbonyl-L-Asp-Glu-Val-Asp- fluoromethyl ketone, a caspase 3 inhibitor zinc protoporphyrin C-18 DRAFT—DO NOT CITE OR QUOTE ------- Z-VAD-FMK a7-nAChR a-Toc yGCS yH2A.X p° cells Z-Val-Ala-DL-Asp-fluoromethylketone, a general caspase inhibitor a7-nicotinic acetylcholine receptor a-tocopherol, an antioxidant y-glutamylcysteine synthetase phosphorylated histone variant H2A.X that is indicative of DNA double strand breaks AL hybrid cells made highly deficient in mitochondrial DNA by long-term treatment with ditercalinium Table C-l. In vivo human studies related to possible modes of action of arsenic in the development of cancer Topic(s) Population Sampled Information on Exposure Levels and Durations and on Biomarkers Results Reference Aberrant Gene or Protein Expression Effect of inorganic arsenic exposure from DWon concentration of RAGE protein in sputum Effect of inorganic arsenic exposure from DW on serum levels of extracellular domain of EGFR (i.e., EGFR BCD) Effect of inorganic arsenic exposure from DW on levels of TGF-a in bladder urothelial cells (BUC) People in Ajo (high dose) and Tucson (low dose), Arizona, USA Araihazar area of Bangladesh 3 towns in central Mexico Compared subjects from Ajo (-20 ppb of arsenic in DW) with subjects from Tucson (~5 ppb of arsenic inDW). They also determined total inorganic arsenic concentrations in urine in individuals. Estimates of inorganic arsenic exposure level were based on well water arsenic (ranged from 0. 1 to 768 ppb), urinary arsenic, and cumulative arsenic index. Such estimates and EGFR BCD protein levels were compared in 574 people. Estimates of inorganic arsenic exposure level were based on levels of different metabolites of arsenic in urine from 72 women who used drinking water that contained 2-378 ppb As. No difference was seen in concentration of RAGE protein in sputum between cities. Since there was much overlap of total inorganic arsenic concentrations in urine in individuals in those cities, a comparison was also made using inorganic arsenic levels in urine. The regression analysis yielded a significant negative association between urinary total inorganic arsenic concentrations and RAGE concentrations in sputum. Thus inorganic arsenic exposure caused U in RAGE level as was seen in mice. Found significant positive correlation between EGFR BCD protein levels in serum and all of these measures of inorganic arsenic exposure, with the association being strongest among individuals with As-induced skin lesions. Found significant positive correlation between TGF- a protein levels in exfoliated BUC and each of 6 arsenic species present in urine. Women from areas with high arsenic exposures had significantly higher TGF-a protein levels in BUC than those from areas of low arsenic exposure. BUC cells from people with As-induced skin lesions contained significantly more TGF-a. Lantz et al., 2007 Li et al., 2007 Valenzuela etal.,2007 C-19 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Microarray- based gene expression study comparing groups with and without arsenical skin lesions, both of which were exposed to inorganic arsenic in DW but to different extents Comparison of expression of several genes between patients with As-related urothelial cancer and non- As- related urothelial cancer Comparison of expression of several integrins between people with arsenic- related keratosis and people with normal skin Population Sampled Bangladesh Taiwan, patients with urothelial cancer Taiwan, patients with arsenical keratosis Information on Exposure Levels and Durations and on Biomarkers Compared subjects with cutaneous signs of arsenicism (mean of 343+258 ppb of arsenic in DW) with asymptomatic individuals (mean of 40+50 ppb of arsenic in D W in one set, and 95+91 ppb in another). All 33 patients with arsenic- related urothelial cancer had been living in the arseniasis- endemic area of southwest Taiwan, where people had drunk the As-contaminated artesian well water for at least 10 years. They were compared with 25 patients who had nonarsenic -related urothelial cancer. All 25 arsenical keratosis patients were from arseniasis-endemic areas of southwest Taiwan, where water is contaminated by high concentrations of inorganic arsenic. Control specimens were obtained from the non-sun-exposed skin of 8 age-comparable patients who did not live in the endemic areas. Results Looked at expression of -22,000 transcripts in RNA from peripheral blood lymphocytes. When the comparison was restricted to female never-smokers, 3 12 differentially expressed genes were identified between those with and without As-induced skin lesions, with all of them being down-regulated in the skin-lesion group. Signal transduction through the IL-1 receptor was identified as a significant pathway of differentially expressed genes between the arsenical skin lesion (n = 11) and nonlesion (n = 2) groups. It discriminated between the 2 groups. Comparisons were made of protein expression of GST-Ti, p53, Bcl-2, and c-Fos by Western blotting oi tumor tissues. A significantly higher proportion of the patients with the arsenic-induced cancer had the proteins present for Bcl-2 (33/33 vs. 19/25) and for c-Fos (30/33 vs 16/25), suggesting that up- regulation of these 2 oncoproteins may play important roles in arsenic -mediated urothelial carcinogenesis. Cellular GSH content was down- regulated in both types of tumors, but to a greater extent in the arsenic -induced ones. Immunohistochemical staining patterns of integrin Pi, a2pi, and a3pi were observed. The various patterns of staining among the patients in comparison to the controls showed decreased expression of all 3 integrins in both arsenical keratosis and in perilesional skin. None showed the normal expression pattern of all 3 integrins. However, there was no association with the occurrence of basal cell carcinoma or squamous cell carcinoma and the expression pattern of any of the 3 integrins. Reference Argos etal., 2006 Hour etal., 2006 Lee et al., 2006b Apoptosis Possible association of specific p53 polymorphisms with arsenic- related keratosis in individuals exposed to arsenic in DW West Bengal, India Compared 177 arsenic- exposed subjects with keratosis (mean of 177 ppb of arsenic inDW) with 189 arsenic -exposed subjects without such skin lesions (mean of 180 ppb of arsenic in DW), and looked for association of keratosis with 3 specific p53 polymorphisms. Used arsenic concentration comparisons in DW, urine, nails, and hair. Homozygotes for alleles at 2 of the polymorphisms were significantly over represented in the individuals with keratosis. Results suggest that individuals carrying the arginine homozygous genotype at codon 72 and/or the no duplication homozygous genotype at intron 3 are at higher risk for the development of arsenic -induced keratosis. In both cases the OR was 2.086 and the 95% CI did not overlap 1. Urinary excretion of arsenic was slightly lower (NSE) in the group with keratosis suggesting higher retention of arsenic in the body, which was reflected in significantly higher arsenic content in nails and hair. De Chaudhuri etal., 2006 C-20 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Population Sampled Information on Exposure Levels and Durations and on Biomarkers Results Reference Chromosomal Aberrations and/or Genetic Instability Nested case- control study/ CAs and/or SCEs as biomarkers for the prediction of cancer development Induction of MN Induction of MN and CAs (relationship to presence of arsenicism and GST polymorphisms) Blackfoot- endemic area in Taiwan West Bengal, India West Bengal, India Looked at CAs and SCEs in lymphocytes from venous blood samples Compared subjects with cutaneous signs of arsenicism (368 ppb of arsenic in DW) with asymptomatic individuals (5.5 ppb of arsenic inDW). Also used arsenic concentration comparisons in urine, nails, and hair. Compared arsenic-exposed subjects with cutaneous signs of arsenicism (mean of 242 ppb of arsenic in DW), arsenic -exposed subjects without cutaneous signs of arsenicism (mean of 202 ppb of arsenic in DW), and arsenic- unexposed subjects (mean of 7.2 ppb of arsenic in DW), and looked for association of effects with different GSTT1 and GSTM1 genotypes. Used arsenic concentration comparisons in DW, urine, nails, and hair. Chromosome-type CAs, but not chromatid-type CAs or SCEs, were significantly higher in the cases than in the controls. The cancer risk OR for subjects with >0 chromosome-type breaks was 5.0 (95% CI = 1 .09-22.82). The OR became even higher with more refinements. Thus chromosome-type CAs (but not chromatid-type CAs or SCEs) can serve as useful biomarkers for prediction of cancer development. In the exposed group, the frequencies of MN per 1,000 cells were highly elevated over those of the control group (# per 1000 cells): 5.15 vs 0.77 in the oral mucosa, 5.74 vs 0.56 in urothelial cells, and 6.39 vs. 0.53 in peripheral lymphocytes, respectively. arsenic-exposed groups showed ft in MN in the lymphocytes, oral mucosa, and urothelial cells and ft in frequencies of CAs in lymphocytes. The symptomatic (i.e., with cutaneous signs of arsenicism) exposed group had more of all types of cytogenetic damage than the asymptomatic exposed group, and the asymptomatic exposed group had more of all types of cytogenetic damage than the unexposed group. Asymptomatic and symptomatic exposed groups demonstrated rather similar concentrations in the urine, nails, and hair. Individuals carrying at least one GSTM1 -positive allele had a significantly higher risk of developing cutaneous signs of arsenicism. Liou et al., 1999 Basuetal., 2002 Ghosh et al., 2006 C-21 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Association between a polymorphism in ERCC2 codon 751 that probably improves NER and (1) the incidence of CAs and (2) the presence of inorganic arsenic-induced hyperkeratosis Induction of MN (bladder cells) Induction of SCEs (Fowler's solution, (lymphocytes) Induction of CAs and SCEs (Fowler's solution, lymphocytes) Population Sampled West Bengal, India Chile, men 6 patients treated with Fowler's solution who developed arsenicism and biopsy- proven skin cancers 8 psoriasis patients treated with Fowler's solution were compared with 8 psoriasis patients not treated with inorganic arsenic (7 men in each group) Information on Exposure Levels and Durations and on Biomarkers Comparisons were made between people with hyperkeratosis and individuals with no skin lesions who were drinking similar inorganic arsenic- contaminated water. Groups with and without hyperkeratosis had means ol 195 and 185 ppb arsenic in DW, respectively, with large standard deviations. Compared subjects having high (average 600 ppb of arsenic in DW) and low (average 15 ppb of arsenic in DW) exposures. Nothing is known about doses; duration of treatment with inorganic arsenic ranged from 4 months to 27 years, and in most cases treatment ceased decades before this cytogenetic analysis. The total doses of inorganic arsenic were from 300 to 1 200 mg for the 7 with known doses. Inorganic arsenic treatments ceased many years before this study. Comparisons were also made to 30 apparently healthy untreated males. Results The polymorphism resulted from a base pair change from A to C at codon 75 1 that resulted in an amino acid substitution from lysine to glutamine. The A/A (i.e., Lys/Lys) genotype was compared with the A/C and C/C genotypes combined. In the study population, the allele frequencies of A and C were 0.4 and 0.6, respectively. A/A individuals were shown to be at significantly higher risk of having hyperkeratosis and also to have a higher frequency of CAs in their lymphocytes, as follows: A/A individuals were over-represented among individuals with inorganic arsenic-induced hyperkeratosis (OR = 4.77, 95% CI = 2.75-8.23). There was a higher percentage of cells with CAs in A/A individuals than in (A/C and C/C) individuals: 43% more in those exposed to inorganic arsenic but not having hyperkeratosis, 18% more in those exposed to inorganic arsenic and having hyperkeratosis, and 3 1% in both groups combined. Also, CAs were significantly more frequent in inorganic arsenic-exposed people with hyperkeratosis. Used a fluorescent version of exfoliated bladder cell MN assay to identify presence or absence of whole chromosomes within MN. Significant ft in induction of MN by arsenic was found, and chromosome breakage appeared to be its major cause. 4th highest quintile of exposure groups gave the highest response, but there was a significant ft in each of quintiles 2-4. Highest (5th) quintile (729- 1894 ppb) returned to baseline MN level, perhaps because of cytostasis or cytotoxicity. Patients treated with Fowler's solution had mean of 14.0 SCE/mitotic cell, while 44 normal controls had mean of 5.8 SCEs/mitotic cell. They saw no difference in chromosome breakage between the groups. ft in frequency of chromosomal breaks (i.e., chromatid and chromosome aberrations together) in psoriasis patients with inorganic arsenic treatment and an even bigger ft in comparison to healthy untreated males. Inorganic arsenic treatment had NSE on SCE frequency. Reference Banerjee et al., 2007 Moore et al., 1997b Burgdorf et al., 1977 Nordenson etal., 1979 C-22 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Induction of CAs (mostly airborne inorganic arsenic, lymphocytes) Induction of CAs and SCEs (lymphocytes) Induction of CAs andMN (lymphocytes for CAs) Induction of MN Induction of MN (chromosome breakage and/or aneuploidy) Population Sampled 9 workers exposed to inorganic arsenic at smelter in northern Sweden People in Fallen (exposed) and Reno (control), Nevada, USA People in Santa Ana (high dose) and Nazareno (control), Mexico People in Nevada, USA, with either very high or low exposure to inorganic arsenic in DW People in Nevada, USA, with either very high or low exposure to inorganic arsenic in DW Information on Exposure Levels and Durations and on Biomarkers Little information was presented except to say that there was no obvious relationship between exposure and CA frequencies. The exposed sample of 104 used DW containing >50 ppb arsenic (mostly >100 ppb As) for at least 5 years and the control sample of 86 used DW containing <50 ppb arsenic (and often much less) for the same period. The high-dose group used DW containing a mean of 408 ppb As, and the control (i.e., low dose) group used DW containing a mean of 30 ppb As. They also considered arsenic concentrations in urine and blood and concentrations of arsenic metabolites. The high-dose group of 18 used DW containing a mean of 13 12 ppb As, and the individually matched control (i.e., low-dose) group used DW containing a mean of 16 ppb As. They also considered the concentration of inorganic arsenic and methylated metabolites in urine. The high dose group of 18 used DW containing a mean of 1,3 12 ppb As, and the individually matched control (i.e., low-dose) group used DW containing a mean of 16 ppb As. They also considered the concentration of inorganic arsenic and methylated metabolites in urine. Results 87 C As/8 19 mitotic cells among smelter workers and 13 CAs/1012 mitotic cells in controls. Person with highest CA frequency had also been exposed to lead and selenium. No hint of any effect of inorganic arsenic on CA or SCE frequencies was seen, even though there was an approximately 9-fold difference in the mean inorganic arsenic concentrations in DW between the 2 groups. inorganic arsenic caused ft in CA (chromatid and isochromatid deletions) frequency in lymphocytes and an ft in MN frequency in exfoliated epithelial cells obtained from the oral mucosa and from urine samples. MN frequencies were higher in people with skin lesions, by a factor of 2.3 in oral mucosa and 4.3 in urothelial cells. There was also much more induction of MN in males than in females for both cell types. inorganic arsenic caused ft in MN frequency in exfoliated bladder cells to 1.8x (90% CI, 1.06x to 2.99x). The MN frequency was positively correlated with the urinary concentration of inorganic arsenic plus methylated metabolites. In contrast, inorganic arsenic had no effect on the MN frequency in epithelial cells obtained from the buccal mucosa. The exfoliated cell MN assay using FISH with a centromeric probe was applied: frequencies of MN containing acentric fragments (MN-) and those containing whole chromosomes (MN+) both showed ft, to 1.65x (statistically significant) and 1.37x (p = 0.15), respectively, suggesting that arsenic has clastogenic and possibly even aneuploidogenic properties. Effect was stronger in males than in females. Thus, in males the increases were 2.06x (p = 0.07) and 1.86x (p = 0.08), respectively. The frequencies of MN- and MN+ were both positively correlated with urinary arsenic and its metabolites. Reference Beckman et al., 1977 Vig et al., 1984 Gonsebatt et al., 1997 Warner et al., 1994 Moore et al., 1996 C-23 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Induction of CAs and SCEs (lymphocytes for CAs) Population Sampled People in Santa Ana (high dose) and Nuevo Leon (low- exposure group), Mexico Information on Exposure Levels and Durations and on Biomarkers The high-exposure group of 1 1 used DW containing a mean of 390 ppb arsenic (98% as Asv), and the low- exposure group of 13 used DW that ranged from 19 to 60 ppb As. They also considered arsenic concentrations in urine. Results Examined the levels of CAs and SCEs in peripheral blood lymphocytes. There were no skin lesions in the control subjects, but 4 of the 11 exposed subjects had cutaneous signs of arsenicism The percentages of total CAs and SCEs were similar in the two groups; however, the finding of a higher point estimate of the frequency of complex CAs (i.e., dicentrics, rings, and translocations) in the high- exposure group was considered suggestive of a possible effect of inorganic arsenic. Average generation times (AGT) of lymphocytes were 19.02 hr in the laboratory control, 19.90 hr in the low- exposure group, and 28.70 hr in the high-exposure group, with this difference being statistically significant. It was suggested that this effect might suggest an impairment of the immune response. Reference Ostrosky- Wegman et al., 1991 DNA Damage DNA damage detected using SCGE (comet) assay (lymphocytes) New Hampshire, USA Low-exposure (control) group had < 0.7 ppb arsenic in DW and high-exposure group had >13 (nd up to 93) ppb arsenic in DW. Using the SCGE (comet) assay, baseline DNA damage as well as the capacity of the lymphocytes from these subjects to repair damage induced by an in vitro challenge with 2-AAAF were assessed. 2- AAAF was used because its adducts are primarily repaired through the NER pathway. High-exposure group had ft in baseline damage (i.e., damage resulting from inorganic arsenic exposure only) to ~1.8x. Two hours after identical in vitro 2-AAAF treatments to cells from both high- and low- inorganic arsenic -exposure groups, cells from both groups showed big ft in DNA damage, with inorganic arsenic-high-exposure group showing -15% more DNA damage than control (NSE). Aftei 4-hr repair period, significantly more DNA damage remained in lymphocytes from individuals in high- exposure group (~1.54x), and essentially all 2- AAAF-induced DNA damage had been repaired in the control cells. Andrew et al., 2006 C-24 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Oxidative DNA damage Correlation of urinary 8-OHdGwith urinary metal elements and many other substances Levels of urinary 8-OHdG following acute arsenic poisoning incident Population Sampled Residents of Bayingnorme n (Ba Men), Inner Mongolia, China, with exposures to a wide range of concentration s of inorganic arsenic in DW 6 regions of Japan Wakayama, Japan Information on Exposure Levels and Durations and on Biomarkers Concentrations of inorganic arsenic in DW were determined for individuals; -70% of subjects used DW containing nondetectable arsenic through 200 ppb As, with the rest using DW containing up to -830 ppb As, with all exposures lasting at least 5 years. They also determined arsenic levels in toenail clippings as a biomarker of exposure. 128 men and 120 women from Japan who did not live within several kilometers of large chemical factories or garbage incinerator facilities 63 people were poisoned by eating food contaminated with arsenic trioxide, with 4 dying about 12 hours after eating. Doses in individuals were poorly known. Results OGG1 expression was used as an indicator of oxidative stress. OGG1 was selected because it codes for the enzyme 8-oxoguanine DNA glycosylase, which is involved in base excision repair of 8-oxoguanine residues that result from oxidative damage to DNA. The study found that OGG1 expression was closely linked to the levels of arsenic in the drinking water and toenails of the individuals examined, indicating a link between ROS damage to DNA and arsenic exposure in humans. There were no significant differences in arsenic-induced expression due to gender, smoking, or age. OGG1 expression was also associated with skin hyperkeratosis in males, and there was a hint of the same in females. There was an inverse relationship between OGG1 expression and Se levels in toenails, indicating possible protective effects of Se against arsenic -induced oxidative stress. The maximal OGG1 response appeared to be at a water arsenic concentration of 149 ppb, after which its expression leveled off and was gradually down-regulated. The association was investigated between urinary concentrations of 8-OHdG and urinary concentrations of As, Al, Cr, Ni, Hg, Zn, Cu, Pb (in ng of element/mg creatinine) as well as with 5 antioxidants and several other substances. Statistically significant positive correlations were found with As, Cr, and Ni and not with any other substances. (The correlation coefficient for arsenic was 0.25.) It thus appears that exposure of healthy people to these 3 metals under normal conditions may increase oxidative DNA damage. Urinary arsenic levels ranged from -0 to -230 ng As/mg creatinine. Some interesting observations were made among the 52 poisoned individuals who were tested for 8- OHdG levels in urine following acute poisoning. After 30 days, urinary 8-OHdG levels were maximal, with a mean for all patients of ~1.5xthe normal level in Japanese people. By 180 days after the poisoning, levels returned to normal. About 37% of the patients never showed any increase in the concentration of 8-OHdG in urine. The same paper documented a significant increase in urinary 8-OHdG in people from Outer Mongolia, China, who drank water contaminated with about 130 ppb As. The increase in urinary 8-OHdG disappeared after they drank "low-arsenic" water for 1 year. Reference Mo et al., 2006 Kimura et al., 2006 Yamauchi et al., 2004 C-25 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) DNA damage in peripheral blood lymphocytes detected by alkaline comet assay Population Sampled West Bengal, India Information on Exposure Levels and Durations and on Biomarkers Low-exposure (control) group had 7.7+0.5 ppb arsenic in DW. High- exposure group had 247+19 ppb arsenic in DW. They also considered arsenic levels in nails, hair, and urine. Results Used SCGE (comet) assay with DNA denaturation at pH >13. High-exposure group had significantly more DNA damage in lymphocytes. Assay was also combined with FPG enzyme digestion to demonstrate that arsenic induced oxidative base damage. Reference Basuetal., 2005 DNA Repair Inhibition or Stimulation Decreased DNA repair (lymphocytes) Decreased DNA repair (lymphocytes) New Hampshire, USA, and the towns of Esperanza and Colonia Allende, Mexico New Hampshire, USA Subjects from New Hampshire were from an ongoing epidemiological study of bladder cancer. Low-exposure (control) group had 0.007-5.3 ppb (average of 0.7) arsenic in DW. High-exposure group had 10.4-74.7 ppb (average of 32) arsenic in DW. Subjects from Colonia Allende had 5.5 ± 0.20 ppb arsenic in DW, and those from Esperanza had 43.3 ± 8.4 ppb arsenic in DW. Comparisons between the low (i.e., control) and high exposure groups used either 5 (for protein analysis) or 6 ppb (for mRNA analysis) as the dividing line between low and high. They also considered arsenic levels in urine and toenails. Subjects from New Hampshire were from an ongoing epidemiological study of bladder cancer. They compared levels of expression of 5 NER genes in 6 cases and 10 controls with the inorganic arsenic levels in their DW and in their toenails. Earlier work suggested that inorganic arsenic exposure was correlated with decreased expression of the nucleotide excision repair genes ERCC1, XPB, and XPF. This study focused on ERCC1 and, besides considering gene expression, it looked at both the protein and DNA repair functional levels (for latter, see part of study described in DNA damage part of this table). Inorganic arsenic exposure was associated with U in expression of ERCC1 in isolated lymphocytes both at the mRNA and protein levels. In combined data, there was a U to -0.71 x, with a significant effect in New Hampshire alone and in the total data. Estimate of effect in Mexico was U to ~0.84x (NSE). U in ERCC1 protein level to ~0.28x was also demonstrated in high-exposure group in New Hampshire. Toenail and DW arsenic levels were inversely correlated with expression of ERCC1, XPB, and XPF. The arsenic levels in toenails were more strongly negatively correlated with the changes in gene expression that the arsenic concentrations in DW. In these comparisons, expression levels were compared between high and low levels of arsenic exposure. By definition a high level in DW was anything >2 ppb arsenic and a high level in toenails was anything >2 ppm As. Andrew et al., 2006 Andrew et al., 2003 Effects Related to Oxidative Stress (ROS) Evidence of oxidative damage to DNA caused by As, but not necessarily from inorganic arsenic inDW Taichung County, Taiwan School children ages 10-12, with attention being given to possibility of oxidative stress to DNA from exposure to environmental pollutants As, Cr, and Ni. No information given on concentrations of inorganic arsenic in DW. When oxidative damage occurs in DNA, the excised 8-OHdG adduct is excreted into urine and is a biomarker of oxidative stress. In this cross-sectional study, subjects with higher urinary arsenic tended to have more (19% more, p = 0.09) urinary 8-OHdG than those with lower urinary As. Cr was also on the borderline of showing a significant ft; when both arsenic and Cr were at a higher level in urine, there was a highly significant ft of 39% in urinary 8- OHdG. Wongetal., 2005 C-26 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Evidence of oxidative damage to DNA caused by inorganic arsenic in DW, and the relationship of that DNA damage to arsenic -related skin lesions Population Sampled 2 villages in Wuyuan prefecture in Hetao Plain, Inner Mongolia, China Information on Exposure Levels and Durations and on Biomarkers Adults from low-arsenic- exposure village (mean of 5.3 ppb arsenic inDW) and from high-arsenic-exposure village (mean of 158.3 ppb arsenic in DW). They also measured levels of MMA and DMA in the urine, and the levels of those metabolites in the urine in the high-arsenic-exposure village were at least 17x higher than they were in the low-arsenic-exposure village. Results When oxidative damage occurs in DNA, the excised 8-OHdG adduct is excreted into urine and is a biomarker of oxidative stress. For subjects without arsenic-related skin lesions in the high-arsenic- exposure village, there was no statistically significant correlation found between inorganic arsenic, MMA, or DMA and 8-OHdG adducts in the urine. However, for subjects with arsenic -related skin lesions in the high-arsenic-exposure village, there was a significant positive correlation in urine between levels of each those 3 types of arsenic and the level of 8-OHdG adducts. There was so much individual variability that overall there was no excess of 8-OHdG adducts in urine in the high- As village compared to the low-As village, even if restricted to only those with arsenic -related lesions. An overall comparison did, however, show an excess of 8-OHdG adducts in urine in the high- arsenic village among those who had been drinking well water for more than 12 years when compared to those who had been drinking it for less than 12 years, regardless of whether they had skin lesions. Reference Fujino et al., 2005 Gene Mutations Induction of HGPRT mutations (isolated mononuclear cells) People in Santa Ana (high dose) and Nuevo Leon (low- exposure group), Mexico The high-exposure group of 1 1 used DW containing a mean of 390 ppb arsenic (98% as Asv), and the low- exposure group of 13 used DW that ranged from 19 to 60 ppb As. They also considered arsenic concentrations in urine. The frequency of monocytes resistant to thioguanine (i.e., mutants) was twice as high in the high-exposure group, but this suggestion of an ft was not statistically significant. Ostrosky- Wegman et al., 1991 Hypermethylation of DNA Extent of methylation of the promoters of tumor suppressor genes p53 and p!6 (relationship to arsenicosis) West Bengal, India Criteria for diagnosis of arsenicosis included a history of using DW containing > 50 ppb arsenic for more than 6 months and presence of skin lesions characteristic of chronic arsenic toxicity. Comparisons were made to individuals without skin lesions or those who live in non-arsenic affected areas. Methylation of the p53 promoter region of DNA obtained from blood samples was studied using methyl-sensitive restriction endonuclease HP AIL Methylation of p!6 was studied using bisulfite modification of the DNA followed by methyl sensitive PCR. Hypermethylation of the promoter region of both genes was observed in people with arsenicosis, and there was a positive dose-response for this hypermethylation. There was a strong suggestion that the promoter region of p53 is hypermethylated in individuals with arsenic -induced skin cancer in comparison to those with skin cancer unrelated to inorganic arsenic exposure, but this comparison did not reach statistical significance (p < 0.2) Chanda et al., 2006 C-27 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Relationship between epigenetic silencing of 3 tumor suppressor genes and exposure to arsenic in patients with bladder cancer Population Sampled New Hampshire patients with bladder cancer Information on Exposure Levels and Durations and on Biomarkers Estimated internal dose of arsenic exposure from toenail measurements. 18 patients with bladder cancer had >0.26 ppm arsenic in their toenails, and 318 had O.26 ppm arsenic in their toenails. 0.26 ppm was the 95th percentile of arsenic exposure in this population. Results They applied methylation-specific PCR. A significant relationship was identified between arsenic exposure and promoter methylation of RASSFlAandPRSSSbutnotpie™5^. The promoter hypermethylation was associated with advanced tumor state. Thus the data provide a potential link between arsenic exposure and epigenetic alterations in patients with bladder cancer. Reference Marsit et al., 2006 Hypomethylation of DNA Extent of methylation of the promoters of tumor suppressor genes p53 and p!6 (relationship to arsenicosis) West Bengal, India Criteria for diagnosis of arsenicosis included a history of using DW containing >50 ppb arsenic for more than 6 months and presence of skin lesions characteristic of chronic arsenic toxicity. Comparisons were made to individuals without skin lesions or those who live in non-arsenic-affected areas. Methylation of the p53 promoter region of DNA obtained from blood samples was studied using methyl-sensitive restriction endonuclease HP AIL Methylation of p!6 was studied using bisulfite modification of the DNA followed by methyl sensitive PCR. In the study described in the row above, a small number of people with high arsenic exposure showed hypomethylation. Hypomethylation occurs only after prolonged arsenic exposure at higher doses. The authors noted that cases of both hyper- and hypomethylation leading to silencing of tumor suppressor genes and activation of oncogenes have been documented in different types of cancers. Chanda et al., 2006 Immune System Response Association between biomarkers of lung inflammation and level of inorganic arsenic exposure from DW Ajo and Tucson, Arizona, USA 40 subjects were from the high-arsenic-exposure town of Ajo (20.3 +3. 7 ppb arsenic in DW), and 33 were from the low-arsenic- exposure town of Tucson (4.0 + 2.3 ppb arsenic in DW). They also measured inorganic arsenic levels in urine, with the mean in Ajo being 2.6 times higher than that in Tucson. Proteolytic enzymes including MMP-2 and MMP-9 are continually secreted in the airways, and their activities are regulated mainly by TIMP-1. The log- normalized concentrations of these 3 substances in induced sputum were not significantly different between these towns. However, after adjusting for town, asthma, diabetes, urinary MMA/inorganic arsenic, and smoking history, total urinary arsenic was negatively associated with MMP-2 and TIMP-1 levels and positively associated with the ratio of MMP-2/TIMP-landMMP-9/TIMP-l. This suggests an association between changes in sensitive markers of lung inflammation and levels of inorganic arsenic of only -20 ppb in DW. It appears that inorganic arsenic levels in DW and the extent of arsenic methylation may be important predictors of lung metalloproteinase concentrations. Josyula et al., 2006 Signal Transduction C-28 DRAFT—DO NOT CITE OR QUOTE ------- Topic(s) Association between TGF-a and/or EGFR and cumulative inorganic arsenic exposure from DW Population Sampled Taiwan Information on Exposure Levels and Durations and on Biomarkers 150 persons were selected from the arseniasis-endemic area in Ilan county in northeast Taiwan, with 30 each coming from those having residential well water in the following ranges (all inppb of As): 0- 50, >50-100, >100-300, >300-600, and >600. Of them, the 66 who agreed to participate in medical surveillance were compared to 35 healthy individuals with no known arsenic exposure. Those with arsenic exposure were further divided on the basis of cumulative arsenic dose (i.e., total DW inorganic arsenic levels x years of exposure) into the following 2 groups: 32 with <6 ppm- years and 34 with >6 ppm- years. Results Blood plasma was collected and tested for TGF-a and EGFR levels using immunoassays. No relationship between arsenic exposure and EGFR protein levels was found. However, both levels of plasma TGF-a and the proportion of individuals with TGF-a overexpression were significantly higher in the high cumulative arsenic exposure group than in the control group. After adjusting for age and sex, there was also a significant linear trend between cumulative arsenic exposure and the prevalence of plasma TGF-a overexpression. Reference Hsu et al., 2006 C-29 DRAFT—DO NOT CITE OR QUOTE ------- Table C-2. In vivo experiments on laboratory animals related to possible modes of action of arsenic in the development of cancer—only oral exposures Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference Aberrant Gene or Protein Expression Lung/mouse (C57BL/6) Lung/mouse (C57B16, male, 21 days of age at start of exposure) Urothelial cells/rat (F344, female) AsmSA AsmSA DMAV sodium cacodyl ate- trihydrat e * 5.8, 28.8 ppm (DW) 10, 50 ppb (DW) Note in ppb! *0.35, 1.4, 14,35 ppm (DW) 8wk 4wk 28 days 28.8 ppm 50 ppb 0.35 ppm mRNA levels were determined in a microchip analysis and validated using real-time PCR: 29 genes were up- regulated and 42 down-regulated. 15% of affected genes were associated with inflammation, including HSP27 and HSP90 (both up-regulated). Numerous extracellular matrix genes were affected, as reflected in phenotypic lung changes related to the organization of elastin and collagen. Protein levels were determined by a Western blot assay: ft for 4 genes, U for 14. No correlation was found between altered genes and altered proteins. Protein levels in BALF determined by proteomic analysis: it is unclear if samples from 10 ppb were examined, ft after dose of 50 ppb: peroxiredoxin-6 and enolase 1. U after dose of 50: GST- omega- 1, RAGE, contraspin, and apolipoproteins A-I and A-IV. Microarray analysis using chip for 4395 genes: gene trees generated by hierarchical clustering of the 510 responsive genes showed marked changes at every dose in comparison to the dose (or dose of 0) below it. Of the 510 genes, 38% were up-regulated and 9% down-regulated by >3-fold. Most affected genes related to the functional categories of apoptosis, cell cycle regulation, adhesion, signal transduction, stress response, or growth factor and hormone receptors. There was a change in the types of genes affected at the different doses, particularly when comparing the higher 2 doses (both cytotoxic) with the 2 non-cytotoxic doses. The dose with most genes affected was 14 ppm. At the lowest dose, 503 genes (11%) were significantly affected, of which 41% were up-regulated and 6% down- regulated by >3-fold. Lantz and Hays, 2006 Lantz et al., 2007 Senetal., 2005 C-30 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Liver cells/mouse (129/SvJ) Brain, liver, placenta/mouse (only pregnant ICR females drank the water) Fetal brain, fetal liver/mouse (only pregnant ICR females drank the water) Liver/mouse (BALB/c, male) Liver/mouse (BALB/c, male) Arsenic Species AsmSA AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 45 ppm (DW) * 4.35 mg/kg (gavage) * 4.35 mg/kg (gavage) 50, 100, 150 ug/mouse/day for 6 days/week (gavage) 50, 100, 150 ug/mouse/day for 6 days/week (gavage) Duration of Treatment 48 wk 1 time only on each of 9 days, gestation days 7 to 16 1 time only on each of 9 days, gestation days 7 to 16 3, 6, 9, 12 months 3, 6, 9, 12 months LOELb 45 ppm 4.35 mg/kg 4.35 mg/kg 50 at 9 and 12 months only 50 at 9 and 12 months Results Microarray analysis, RT-PCR, and immunochemistry : big ft in ER-a and cyclin Dl mRNA and protein levels. Of 588 genes tested in microarray analysis, 30 showed aberrant expression, including steroid-related genes, cytokines, apoptosis-related genes, cell cycle-related genes, and genes encoding for growth factors and hormone receptors. Activities of selenoenzymes GPx, TrxR, DI-I, DI-II, and DI-m in maternal tissues when examined on gestation day 17 of their litter: liver: U of DI-I to ~0.61x when Se- adequate diet; liver: U of DI-I to ~0.30x when Se- deficient diet; all other comparisons were either slight or NSE. Activities of selenoenzymes GPx, TrxR, DI-I, DI-II, and DI-m in fetal tissues when examined on gestation day 17: brain: ft of DI-II to ~4.1x when Se- deficient diet; liver: U of TrxR to ~0.78x when Se- deficient diet; all other comparisons were either slight or NSE. Levels of TNF-a and IL-6: NSE on either one at any dose in first 6 months. At 9 months: TNF-a: 50, ~1.2x; 100, ~1.2x; 150, IL-6: 50, ~2.0x; 100, ~2.5x; 150, ~2.7x. At 12 months: TNF-a: 50, ~1.9x; 100, ~2.3x; 150, ~3.0x; IL-6: 50, ~2.8x; 100, ~5.7x; 150, ~9.5x. Concentration of total collagen: At 3 months: NSE at all doses, but hint of ft at 100 (~1.2x) and 150 (~1.3x). At 6 months: NSE at all doses, but hint of ft at 100 (~1.3x) and 150 (~1.4x). At 9 months: 50, ~1.3x; 100, ~1.4x; 150, At 12 months: 50, ~1.5x; 100, ~1.9x; 150, ~2.1x. Reference Chen et al., 2004b Miyazaki et al., 2005 Miyazaki et al., 2005 Das etal., 2005 Das etal., 2005 C-31 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Liver, kidney, and lung/mouse (B6C3F1, female) Liver and kidney/mouse (B6C3F1, female) HCC cells/mouse (only pregnant C3H females drank the water, male offspring) HCC cells/mouse (only pregnant C3H females drank the water, male offspring) HCC cells/mouse (only pregnant C3H females drank the water, male offspring) Arsenic Species AsmSA Asv sodium arsenate DMAV AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) In all cases, dissolved in water and administered once by gavage: * 9.58 mg/kg for all * 9.58 mg/kg for all *391 mg/kg for all In all cases, dissolved in water and administered once by gavage: * 0.0749, 0.749, 2.25, 7.49 mg/kg for both 42.5, 85 ppm (DW); report did not state if molecular analysis was done on one or both of these doses combined 42.5, 85 ppm (DW); report did not state if molecular analysis was done on one or both of these doses combined 42.5, 85 ppm (DW); report did not state if molecular analysis was done on one or both of these doses combined Duration of Treatment One dose for all One dose 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 LOELb 9.58 mg/kg 9.58 mg/kg None 2.25 mg/kg in liver 7.49 mg/kg in kidney 42.5 or 85 ppm 42.5 or 85 ppm 42.5 or 85 ppm Results HMOX-1 activity 6 hr after the single oral dose was administered by gavage: Liver: Asm, ~7.5x; Asv, ~5.1x, DMAV, ~0.96x (NSE). Kidney: Asm, ~7.6x; Asv, ~3.2x, DMAV, ~1.03x(NSE). Lung: none of the arsenicals induced HMOX-1 activity. HMOX-1 activity in liver 6 hr after the single oral dose was administered by gavage: at 2 lower doses, NSE; 2.25, ~2.5x; 7.49, ~7.5x. HMOX-1 activity in kidney 4 hr after the single oral dose was administered by gavage: at 3 lower doses, NSE; 7.49, ~3.5x. Comparisons of gene expression based on microarray analysis, with comparisons being made between HCC tumors from offspring of exposed dams and normal liver tissue from offspring of unexposed dams: ft of AFP to -18. 5x; U ofIGF-lto0.78x; ft of IGFBP-1 to ~8.8x; ft of CK8 to ~2.4x; ft of CK18 to ~8.8x; U of BHMT to ~0.33x. Comparisons of gene expression based on microarray analysis, with comparisons being made between HCC tumors of offspring of exposed dams and spontaneous liver tumors of offspring of unexposed dams: ft of AFP to ~6.2x; NSE for IGF-1; ft of IGFBP-1 to ~1.7x; ft of CK8 to ft of CK18 to ~5.8x; U of BHMT to ~0.36x. Comparisons of gene expression based on microarray analysis, with comparisons being made between HCC tumors and normal-appearing liver cells of offspring of exposed dams: ft of AFP to ~7.4x; U of IGF-1 to ~0.68x; ft of IGFBP-1 to ~3.7x; ft of CK8 to ~1.3x; ft of CK18 to -7.0 x; UofBHMTto~0.32x. Reference Kenyon et al., 2005b Kenyon et al., 2005b Waalkes et al., 2004b Waalkes et al., 2004b Waalkes et al., 2004b C-32 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Liver cells/mouse (only pregnant C3H females drank the water, male offspring) HCC cells/mouse (only pregnant C3H females drank the water, male offspring) Uterus/mouse (only pregnant CD1 females drank the water, female offspring only) HCC cells/mouse (only pregnant C3H females drank the water, male offspring) Arsenic Species AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 42.5, 85 ppm (DW); report did not state if molecular analysis was done on one or both of these doses combined 42.5, 85 ppm (DW); report did not state if molecular analysis was done on one or both of these doses combined 85 ppm (DW) 42.5, 85 ppm (DW) Duration of Treatment 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 LOELb 42.5 or 85 ppm 42.5 or 85 ppm 85 ppm if also treated with DBS orTAM 42.5 ppm Results Comparisons of gene expression based on microarray analysis, with comparisons being made between normal-appearing liver cells in both offspring of exposed dams and unexposed dams: ft of AFP to ~2.5x; ft of IGF- 1 to -l.lx; ft of IGFBP-1 to ~2.4x; ft of CK8 to NSEforCKlSorBHMT. In general, the results in the 4 previous rows were confirmed by real-time RT- PCR analysis. Aberrant gene expression was also noted in the microarray analysis for numerous other genes including those related to cell proliferation, oncogenes, stress, and metabolism. Expression (by real-time RT-PCR) of various estrogen-related genes in uteri at 1 1 days of age: ft in ER-a to 1.56x. Some female offspring were also exposed by subcutaneous injection to DBS on the first 5 days after birth. DBS alone or (inorganic arsenic + DBS) did not significantly increase ER-a expression. Inorganic arsenic alone did not ft expression of pS2, CYP2A4, or lactoferrin. However, DBS alone caused large ft in expression of all 3 of these genes, and (inorganic arsenic + DBS) caused a further ft to 3.0 times, 7.8 times, and 1.47 times that of DBS alone, respectively. These and other results showed that inorganic arsenic acts with estrogens to enhance production of urogenital cancers in female mice. Comparisons of gene expression based on microarray analysis of RNA, with comparisons being made between HCC tumors from offspring of exposed dams and normal (i.e., non-tumorous) liver tissue from offspring of unexposed dams: 13.7% of 600 genes were significantly up-regulated or down- regulated. Only 7.7% of those 600 genes were similarly affected in spontaneous tumors in liver tissue from offspring of unexposed dams. The 600 genes studied included oncogenes and genes associated with cell proliferation, differentiation, or otherwise related to cancer outcome. Reference Waalkes et al., 2004b Waalkes et al., 2004b Waalkes et al., 2006a Liu et al., 2004 C-33 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species HCC cells/mouse (only pregnant C3H females drank the water, male offspring) Liver cells/mouse (only pregnant C3H females drank the water, male offspring) HCC cells/mouse (only pregnant C3H females drank the water, male offspring) Liver cells/mouse (only pregnant C3H females drank the water, male offspring) Arsenic Species AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 42.5, 85 ppm (DW) 42.5, 85 ppm (DW) 85 ppm (DW) 85 ppm (DW) Duration of Treatment 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 LOELb 42.5 ppm 42.5 ppm 85 ppm 85 ppm Results Comparisons of gene expression based on microarray analysis of RNA (see row above): up-regulated genes included oncogene/tumor suppressor genes and genes related to cell proliferation, hormone receptors, metabolism, stress, apoptosis, growth arrest, and DNA damage. A wide array of different types of genes was also down-regulated. Real-time RT-PCR analysis largely confirmed the findings of microarray analysis. The higher dose tended to yield more significant differences, but a positive dose-response was not always evident. Comparisons of gene expression based on microarray analysis of RNA, with comparisons being made between non- tumorous liver cells in both offspring of exposed dams and unexposed dams: -10% of 600 genes were significantly up-regulated or down-regulated. The 600 genes studied included oncogenes and genes associated with cell proliferation, differentiation, or otherwise related to cancer outcome. Comparisons of gene expression based on microarray analysis of RNA, with comparisons being made between HCC tumors from offspring of exposed dams and normal liver tissue from offspring of unexposed dams: statistically significant alterations in expression were seen for 2,540 genes. Real-time RT-PCR and Western blot analyses of selected genes or proteins showed >90% concordance. Affected gene expression included oncogenes, HCC biomarkers, cell proliferation-related genes, stress proteins, insulin-like growth factors, estrogen-linked genes, and genes involved in cell-cell communication. Comparisons of gene expression based on microarray analysis of RNA, with comparisons being made between non- tumorous liver cells in both offspring of exposed dams and unexposed dams: statistically significant alterations in expression were seen for 2010 genes. See row above for results in HCC cells. Reference Liu et al., 2004 Liu et al., 2004 Liu et al., 2006c Liu et al., 2006c C-34 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Fetal livers/mouse (only pregnant C3H females drank the water, male offspring) Livers of newborn males/mouse (only pregnant C3H females drank the water) Liver and liver tumors/mouse (only pregnant C3H females drank the water) Arsenic Species AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 85 ppm (DW) 85 ppm (DW) 85 ppm (DW) Duration of Treatment 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 LOELb 85 ppm 85 ppm Results Comparisons of gene expression based on microarray analysis of RNA from fetal livers just after treatment ended, with confirmation by real-time RT-PCR: alteration of expression of 187 genes (of 22,000 in array) was demonstrated, with -25% of them being related to either estrogen signaling or steroid metabolism — some with dramatic (here meaning »100x) up-regulation. Expression of some genes important in methionine metabolism was suppressed. Comparisons of gene expression based on microarray analysis of RNA from livers of newborn males, with confirmation by real-time RT-PCR: among 600 genes examined, marked alteration of expression of 40 genes was demonstrated. Affected genes included genes related to stress (several in the glutathione system), metabolism (several cytochrome P450 genes), growth factors (several insulin-like growth factor genes), and hormone metabolism. Samples from adults of both sexes were tested. Some had had a post-weaning 21-wk dermal treatment with TPA. Comparisons with the TPA- treatment-only control were made regarding gene expression based on microarray analysis of RNA, with confirmation by real-time RT-PCR. Alteration of expression of ~70 genes (of 588 in array) was demonstrated. There were generally similar gene alteration patterns in both sexes both in inorganic arsenic/TPA exposed non-tumorous livers and in inorganic arsenic/TPA-induced tumors. The tumors themselves generally had more pronounced alterations in gene expression than the normal tissue around them. In general, the inorganic arsenic/TPA- induced gene expression alterations were similar to those seen in liver samples from male mice exposed only to inorganic arsenic in utero. It should be noted that while in utero inorganic arsenic -exposed males developed hepatocellular carcinoma without the TPA treatment, in utero inorganic arsenic-exposed females only developed those tumors after TPA treatment. Reference Liuetal., 2007a Xie etal., 2007 Liu et al., 2006b C-3 5 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Bladder and liver/rat (Fisher 344, male) Lung/mice (C57BL/6J Oggl+/+wtmice and Oggl"'" knockout mice, both sexes, 14 weeks old at start of treatment) Liver cells/rat (Sprague Dawley) Liver cells/rat (Sprague Dawley) Skin/mouse (homozygous, strain Tg. AC, female) Arsenic Species MMAV DMAV TMAVO DMAV Asvas Na2HAs Cv 7H2O Asvas Na2HAs Cv 7H2O AsmSA Dose in Elemental Asa (in Units Stated) * 121ppm(DW)c * 109 ppm (DW)C * 110ppm(DW)c * 115.3 ppm (DW) * 0.24, 2.4, 24 ppm (DW) * 0.24, 2.4, 24 ppm (DW) 200 ppm (DW) Duration of Treatment 20 days for all 4 weeks 1 month 4 months 4, 10 wk LOELb Results Changes in gene expression observed in cDNA microarray analysis: MMAV caused ft for 20 genes and U for 1 gene in liver and ft for 5 genes and U for 5 genes in bladder. DMAV caused ft for 15 genes and U for 2 genes in liver and ft for 13 genes and U for 4 genes in bladder. TMAVO caused ft for 23 genes and U for 2 genes in liver and ft for 6 genes and U for 7 genes in bladder. Groups of genes affected by all arsenicals in both tissues included genes related to xenobiotic metabolism, growth factor receptors, and energy metabolism. In the liver, phase I and II metabolizing enzymes were induced to a lesser extent by MMAV and DMAV than by TMAVO, and in the bladder they were induced only by DMAV. CYP1A1 was only overexpressed by TMAVO and in liver. Results of an Affymetrix oligonucleotide microarray analysis: a change in expression was found for 165 and 182 genes in male and female knockout Oggl"'" mice, respectively. In DMAv-treated knockout Oggl"7" mice, there was marked induction of Polal, CYP7B1, NdfuaS, MMP-13 and other genes specific to cell proliferation, cell signaling, and xenobiotic metabolism. Various Various 200 ppm Determination of mRNA levels of cancer-related genes using real-time quantitative RT-PCR: ft cyclinDl at 2.4 only; ft p27Klpl at 2.4 only; ft ILK at 0.24 only; U PTEN at 0.24 only; U 3-catenin at 24 only. Determination of mRNA levels of cancer-related genes using real-time quantitative RT-PCR: ft cyclin Dl at 24 only; ft ILK at 0.24 and 2.4; ft p27Klpl at 0.24 only; U PTEN at all doses; U 3-catenin at all doses. Results were confirmed by protein levels. (Histograms were assumed to be correct for ILK and p27; the descriptions for them appear to have become reversed in the text.) Kinetics of mRNA expression based on RT-PCR: EGFR and TNF-a: ft by week 10; GM-CSF and TGF-a: ft by week 4; big ft by week 10; c-myc: NSE. Reference Kinoshita et al., 2007a Kinoshita et al., 2007b Cuietal., 2004b Cuietal., 2004b Germolec et al., 1998 C-36 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Heart/mouse (C57BL/6NCr, male) Blood plasma/mouse (C57BL/6NCr, male) Tumors that developed from B16-F10(GFP) melanoma tumor cells/ mice (NCr nu/nu, male) Arsenic Species AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 0.05, 0.25, 0.5 ppm (DW) 0.5 ppm (DW) 10, 50, 200 ppb (DW) Duration of Treatment 5, 10, 20 wk 20 wk 9wk LOELb Various 0.5 ppm 10 ppb Results mRNA levels determined by RT-PCR: VEGF165: ft at 0.25 and 0.5 at wk 5; ft at all doses at wk 10; NSE at wk 20; VEGFR1 : NSE at wk 5 and 10; big ft at 0.25 and big U at 0.5 at wk 20; VEGFR2: ft at 0.5 at wk 5; NSE at wk 10; ft at 0.05 and 0.25 and U at 0.5 at wk 20; PAI-1: NSE at wk 5; ft at 0.5 at wk 10; ft at 0.25 and 0.5 at wk 20; Endothelin-1: NSE at wk 5 and 10; ft at 0.05 and big ft at 0.25 at wk 20; MMP-9: NSE at wk 5; ft at 0.5 at wk 10; ft at all doses at wk 20. PAI-1 protein levels determined by ELISA assay: ftto~1.33x. Protein levels in primary melanoma tumors determined by immunohistochemical staining: ft HIF-la at 10 and 50 only; ft VEGF at 10 and 200 only, ft for both proteins was just locally around tumor blood vessels. Western blot assay of whole tumor lysates showed no more than barely detectable ft of HIF-la at any dose. Reference Soucy et al., 2005 Soucy et al., 2005 Kamat etal., 2005 Apoptosis Bladder and liver/rat (Fisher 344, male) Liver/rat (Wistar, male) MMAV DMAV TMAVO AsmSA * 121ppm(DW)c * 109 ppm (DW)C * 110ppm(DW)c *0.03, 1.4,2.9 ppm (DW) 5, 10, 15, and 20 days for all 60 days None Various Various 1.4 Apoptosis labeling index based on an immunochemistry method of staining single-stranded DNA: Bladder: ft on day 20 to ~1.5x for DMAV only; Liver: ft on day 20 to ~3.3x for TMAVO only. Induced apoptosis (experimental - control) based on TUNEL assay with PI staining and analysis using fluorescence microscopy: 0.03, 5.0; (NSE); 1.4, 14.9; 2.9, 22.3; these results were consistent with DNA ladder formation found by agarose gel electrophoresis for which there was an ft at 1.4; bigger ft at 2.9. There was also microscopic evidence of cell death by necrosis. Kinoshita et al., 2007a Bashir et al., 2006a C-37 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Kidney, leukocytes and liver/rat (albino Wistar, male) Kidney, leukocytes and liver/rat (albino Wistar, male) Splenocytes and thymocytes/mous e (C57BL/6, female) Brain and liver/rat (Wistar, male) Arsenic Species AsmSA AsmSA Asvas Na2HAs Cv 7H2O AsmSA Dose in Elemental Asa (in Units Stated) * 57.7 ppm (DW) * 57.7 ppm (DW) 0.5, 5, 50 ppm (DW) * 3.6, 6.1,7.3 mg/kg (gavage, with animals being killed 24 hr later for sample collection) Duration of Treatment 30 days 30 days 8, 12 wk One dose LOELb 57.7 ppm 57.7 ppm 50 at 8 wk for both cell types Various Results TNF-a levels: kidney, ft ~1.6x; leuko., ft ~2.2x; liver, ft ~1.9x; caspase-3 levels: kidney, ft ~3.2x; leuko., ft ~2.8x; liver, ft ~3.5x; effects on both endpoints in all 3 tissues were markedly reduced by co-treatment with AA and/or a-Toc. Induced percentage of DNA that was fragmented (experimental - control): kidney, ft -17.6%; leuko., ft -17.4%; liver, ft -2 1.8%. Induced percentage of TUNEL positive cells (experimental - control): kidney, ft -6.7%; leuko., ft -5.1%; liver, ft -8. 1%; effects on both endpoints in all 3 tissues were markedly reduced by co- treatment with AA and/or a-Toc. Confirmation of induced apoptosis in leukocytes shown by finding typical DNA ladders after agarose gel electrophoresis; co-treatment with AA and/or a-Toc abolished that effect. Induced apoptosis (experimental - control) determined by TUNEL method: Splenocytes: 8 wk: 0.14% of cells at dose of 50 (or 6.6x); 12 wk: 0.22% of cells at dose of 50 (or 5.4x). Thymocytes: 8 wk: 0.40% of cells at dose of 50 (or 4.0x); 12 wk: 0.28% (NSE) of cells at dose of 50 (or 2.5x). For both cell types, the data suggested a positive dose-response across all doses; however, the other results showed much variability. Brain: caspase-3 activity: ft to ~1.4x (NSE) at 3. 6, to ~2.0x at 6.1, and to ~2.6x at 7.3; Liver: caspase-3 activity: ft to ~1.8x at 3.6, to ~2.5x at 6.1, and to ~3.0x at 7.3. Both brain and liver: agarose gel electrophoresis showed DNA "nucleosomal ladder," suggesting induction of apoptosis; results were not quantified. Histopathological examination also showed evidence of cellular necrosis. Reference Ramanathan et al., 2005 Ramanathan et al., 2005 Stepnik et al., 2005 Bashir et al., 2006b C-3 8 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference Cancer Promotion Skin/mouse (homozygous, strain Tg. AC, female) Skin/mouse (hairless swiss- bald strain, male) Lung/mouse (ddY, male) AsmSA Asv sodium arsenate DMAV assumed to be dimethy lar-sinic acid 200 ppm (DW) * 11.4 ppm (DW) *217ppm(DW) 14 wk 25 wk 25 wk 200 ppm None, but 11.4 ppm if also treated with DMBA 217 ppm, but only following 4NQO treatment After low-dose application of TPA on 4 occasions over 2 weeks starting after 3 1 days of inorganic arsenic exposure, there was a marked ft in the number of skin papillomas compared to single treatments, whereas no papillomas developed in inorganic arsenic -treated Tg.AC mice without TPA treatment or in FVB/N mice with the combined treatment. Injection of neutralizing antibodies to GM-CSF after TPA application reduced the number of papillomas in Tg.AC mice. Inorganic arsenic acted like a co-promoter. PCNA protein levels determined by Western blotting: no PCNA was present following the inorganic arsenic treatment alone, compared to the baseline of 22 units of PCNA in the control (set equal to x). When mice were given 4 DMBA treatments (as an initiating carcinogen) during the first 2 weeks of the inorganic arsenic treatment, there was PCNA ft to ~5.3x. DMBA treatment alone caused ft to only 2.9x. Mice that were untreated or treated with inorganic arsenic alone developed no papillomas or skin tumors. DMBA treatment alone induced development of squamous cell papillomas. Combined inorganic arsenic and DMBA treatment caused development of well- differentiated squamous cell carcinomas. Inorganic arsenic acted as a skin tumor promoter by promoting abnormal cell proliferation. Findings suggest that inorganic arsenic is toxic to normal skin cells and that preneoplastic cells are more resistant to inorganic arsenic. Some of the mice were subcutaneously injected with 10 mg/kg of 4NQO just before the 25-wk DMA treatment began. Some of the mice ate only feed containing 0.05% of the antioxidant EGCG. Number out of 10 mice in each group bearing tumors: control, 0; DMA alone, 0; 4NQO alone, 7; EGCG alone, 0; (4NQO + DMA), 10; (4NQO + DMA + EGCG), 7. That last group had only 0.89 tumor/mouse compared to 3. 10 tumors/mouse in 4NQO group and 4.00 tumors/mouse in the (4NQO + DMA) group. Germolec et al., 1998 Motiwale et al., 2005 Mizoi et al., 2005 C-39 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference Cell Cycle Arrest or Reduced Proliferation Heart/mouse (C57BL/6NCr, male) AsmSA 0.5 ppm (DW) 5, 10, 20 wk 0.5 ppm at 20 wk Density of microvessels of <12 um diameter using histopathology and a digital-imaging subroutine: U to ~0.82x at 20 wk; hint of a U at 10 wk. Soucy et al., 2005 Cell Proliferation Stimulation Bladder/rat (F344, female) Bladder/rat (F344, female) Liver/rat (Fischer 344, male) (they used normal-appearing tissue) Bladder and liver/rat (Fisher 344, male) DMAV DMAV TMAVO MMAV DMAV TMAVO * 54.3 ppm (food) (assumes MW of chemical used was 138.0) * 54.3 ppm (food) (assumes MW of chemical used was 138.0) *27.5, 110.2 ppm (DW) Estimated total intakes: 351 and 1363 mg As/rat * 121ppm(DW)c * 109 ppm (DW)C * 110ppm(DW)c 2wk 26 wk 104 wk 5, 10, 15, and 20 days for all 54.3 ppm 54.3 ppm 110.2 ppm None Various Various Stimulation of proliferation determined by BrdU labeling assay: ft to 3.9x; co-treatment with DMPS (a chelator of trivalent arsenicals) completely eliminated the effect. Stimulation of proliferation determined by BrdU labeling assay: ft to 1.6x; co-treatment with DMPS (a chelator of trivalent arsenicals) completely eliminated the effect. Histological examination showed simple hyperplasia in 4 of 9 rats, compared to 0 of 10 rats in control and 1 of 10 rats with co-treatment with DMPS. Livers were stained for the analysis of PCNA by an immunohistochemical method, with the PCNA index being the number of positive cells/100 cells: ft in PCNA index to 2.0x, thereby suggesting that cell proliferation in the normal- appearing parenchyma was elevated. The point estimate of the index was also ft at lower dose, but the SE for it was large. PCNA labeling index based on an immunochemistry method: Bladder: ft on day 20 to ~1.8x for DMAV only; Liver: ft on day 20 to ~1.8x for TMAVO only. Cohen et al., 2002 Cohen et al., 2002 Shen et al., 2003 Kinoshita et al., 2007a C-40 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Bladder/mouse (only pregnant CD1 females drank the water, male offspring only) Kidney/mouse (only pregnant CD1 females drank the water, male offspring only) Bladder/mouse (only pregnant CD1 females drank the water, female offspring only) Arsenic Species AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 85 ppm (DW) 85 ppm (DW) 85 ppm (DW) Duration of Treatment 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 LOELb 85 ppm if also treated with DBS orTAM 85 ppm 85 ppm if also treated with DBS orTAM Results Some male offspring were also exposed by subcutaneous injection to DBS or TAM on the first 5 days after birth; all male offspring were held for 90 wk before examination. Induced (i.e., experimental - control) % of mice with bladder hyperplasia: inorganic arsenic alone, 9% (NSE); DBS alone, 12% (NSE); TAM alone, 10% (NSE); (inorganic arsenic + DBS), 45%; (inorganic arsenic + TAM), 30%. All induced percentages were the same for total proliferative lesions, except for (inorganic arsenic + TAM), which was 40%. The lesions induced by inorganic arsenic with either DES or TAM overexpressed ER-a. Some male offspring were also exposed by subcutaneous injection to DES or TAM on the first 5 days after birth; all male offspring were held for 90 weeks before examination. Induced (i.e., experimental - control) % of mice with cystic tubular hyperplasia: inorganic arsenic alone, 23%; DES alone, 0%; TAM alone, 0%; (inorganic arsenic + DES), 24%; (inorganic arsenic + TAM), 7%. Some female offspring were also exposed by subcutaneous injection to DES or TAM on the first 5 days after birth; all female offspring were held for 90 wk before examination. Induced (i.e., experimental - control) % of mice with bladder hyperplasia: inorganic arsenic alone, 12% (NSE); DES alone, 0% (NSE); TAM alone, -3% (NSE); (inorganic arsenic + DES), 26%; (inorganic arsenic + TAM), 23%. All induced percentages were the same for total proliferative lesions, except for (inorganic arsenic + DES), which was 35%, and (inorganic arsenic + TAM), which was 26%. Unlike in the male offspring, inorganic arsenic did not induce hyperplasia in kidneys. Reference Waalkes et al., 2006b Waalkes et al., 2006b Waalkes et al., 2006a C-41 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Lung/mice (C57BL/6J Oggl+/+wtmice, both sexes, 14 weeks old at start of treatment) Lung/mice (C57BL/6J Oggl '" knockout mice, both sexes, 14 weeks old at start of treatment) Bladder/mouse (C57BL/6, female) Bladder/mouse (C57BL/6, female) Bladder/mouse (C57BL/6, female) Blood vessels/chicken (Leghorn, chorioallantoic membranes of 10-day-old chicken embryos) Matrigel implants/mouse (C57BL/6NCr, male) Arsenic Species DMAV AsmSA AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) * 115.3ppm(DW) * 57.7 ppm (DW) * 57.7 ppm (DW) * 11.5, 57.7 ppm (DW) 0.00033, 0.001, 0.0033,0.01, 0.033,0.1,0.33, 1.0,3.3, 10 uM 0.001,0.005,0.01, 0.05 ppm (DW) Duration of Treatment 72 weeks 4wk 16 wk 16 wk 24 hr 5wk LOELb None 115.3 57.7 ppm 57.7 ppm 11.5 ppm 0.033 uM 0.001 ppm Results PCNA labeling index based on an immunochemistry method, x = wt control level: wt with inorganic arsenic treatment: ft to ~3x (NSE). Knockout Oggl"'" without inorganic arsenic: ft to ~6x. Knockout Oggl"7" with inorganic arsenic treatment: ft to ~17x. Results were confirmed in a study with only a 4 week exposure. All experimental mice developed mild hyperplasia of the urinary bladder epithelium, that being a 3- to 4-fold ft in the thickness of the transitional cell layer. ft in PCNA-stained nuclei in the bladder epithelium from 2% in control to 3 1% in experimental group, an indication of big ft in cell proliferation. Similar ft also seen at 4 weeks. Also consistent with ft in proliferation: ft in DNA binding of the AP- 1 transcription factor to ~1.9x and ~4.7x at the 2 doses, respectively. At one or both doses (not specified): 38% and 76% of the bladder cells stained positive for the c-jun and c-fos immunoreactive proteins, respectively, compared to only 2% in control mice. CAM assay to determine vascularity (i.e., bloodvessel density): ft to ~2.2x and remained at about that level to dose of 1; U to ~0.28x at dose of 3.3 and remained at about that level to dose of 10. Blood vessel no. determined in Matrigel implants surgically inserted during last 2 wk of inorganic arsenic treatment: probable ft to ~1.8x at dose of 0.001; statistically significant ft to ~2.4x at the higher 3 doses. Implants were supplemented with recombinant FGF-2; inorganic arsenic-enhanced neovascular- ization did not occur without FGF-2. Data suggest that inorganic arsenic potentiates, but does not directly cause, neovascularization in Matrigel implants. Reference Kinoshita et al., 2007b Simeonova et al., 2000 Simeonova et al., 2000 Simeonova et al., 2000 Soucy et al., 2003 Soucy et al., 2005 C-42 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Matrigel implants/mouse (C57BL/6NCr, male) Tumors that developed from B16-F10(GFP) melanoma tumor cells/mice (NCr nu/nu, male) Tumors that developed from B16-F10(GFP) melanoma tumor cells/mice (NCr nu/nu, male) Blood vessels/chicken (Leghorn, chorioallantoic membranes of 10-day-old chicken embryos) Skin/mouse (homozygous, strain Tg. AC, female) Arsenic Species AsmSA AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 0.05, 0.25, 0.5 ppm (DW) 10, 50, 200 ppb (DW) Note in ppb! 10, 50, 200 ppb (DW) Note in ppb! 0.33, 10 uM 200 ppm (DW) Duration of Treatment 5, 10, 20 wk 8wk 8wk 48 hr 10 wk LOELb 0.05 ppm for each duration 10 ppb 10 ppb 0.33 uM 200 ppm Results Blood vessel number determined in Matrigel implants surgically inserted during last 2 wk of inorganic arsenic treatment: at 5 wk: ft to ~2.6x, ~4.4x, and ~5.5x at the 3 doses in ascending order. For each longer duration treatment, there was still a strong ft at dose of 0.05 but a somewhat diminished ft at 2 higher doses. Tumor volume and tumor growth rate, after implantation of tumor cells (into external surface at the base of right ear) 5 wk after inorganic arsenic treatment began: Volume: 10, ~1.4x (NSE); 50, ~2.2x; 200, ~3.0x. Rate: 10, ~1.9x (NSE); 50, ~2.2x; 200, ~3.2x. Mean no. of lung metastases/lobe, after implantation of tumor cells (into external surface at the base of right ear) 5 wk after inorganic arsenic treatment began: 10, ~1.6x; 50, ~2.0x; 200, ~2.0x (statistically significant at 10 and 200); the metastases were significantly larger at the 2 lower doses. CAM assay to determine vascularity (i.e., blood vessel density): ft to ~1.8x at 0.33 but big U at dose of 10. At dose of 0.33, co-treatment with YC-1 or SU5416 (inhibitors of HIF and VEGF receptor-2 kinase) eliminated inorganic arsenic effect. 10 uM inorganic arsenic + YC-1 caused no change from control, but inorganic arsenic alone, or in addition to SU5416, resulted in U to ~0.28x. By 10 weeks the skin showed hyperkeratosis as well as ft in numbers of proliferating cells. A kinetic study with samples at weekly intervals demonstrated ft in number of BrdU- positive nuclei in skin after 4 weeks and number remained elevated through 10 weeks. Reference Soucy et al., 2005 Kamatetal., 2005 Kamatetal., 2005 Kamatetal., 2005 Germolec et al., 1998 Chromosomal Aberrations and/or Genetic Instability Bone marrow/rat (Rattus norvegicus, Charles foster strain) Asvas disodiu m hydroge n arsenate * 4.0 mg As/kg bw (unspecified route of administration) 15, 21 days 4.0 mg/kg Chromosomal analysis of Giemsa- stained cells, with few details provided: induction of gross CAs for both periods of treatment; induction of hyperploidy detected as aneuploids for longer treatment. Dattaetal., 1986 C-43 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Bone marrow/mouse (albino Swiss, male) Bone marrow/mouse (C57BL/6J/Han, female) Arsenic Species AsmSA Asvas Na2HAs Cv 7H2O Dose in Elemental Asa (in Units Stated) * 1.44 mg/kg x 4, 5, and 6 times at weekly intervals, (gavage) 50, 200, 500 ppb (DW) Note in ppb! Duration of Treatment Single dose each week 3,6, 12 months LOELb 1.44 x 4 None Results Significant ft in CA and probably also in polyploidy after 4, 5, and 6 gavage treatments. CA frequencies were significantly higher than control in all 3 comparisons at 2.5x, 2.7x, and 4.4x, respectively. Similar experiments with 7 and 8 exposures killed the mice. Daily treatments by gavage with a black tea infusion for one week before every inorganic arsenic treatment caused a significant reduction in the frequency of CAs after 4 and 6 inorganic arsenic treatments. Half of the mice were maintained on a low-Se diet. Mouse erythrocyte MN test: inorganic arsenic caused no induction of MN in PCEs and no change in the PCE:NCE ratio at any dose at any interval, with or without the low-Se diet. Reference Patra et al., 2005 Palus etal., 2006 Co-carcinogenesis Skin/mouse (Hairless mice, strain Skhl) Skin/mouse (Hairless CrL:SKl-hrBD, female, weanling) Starting 3 wk after inorganic arsenic treatment began; mice were irradiated thrice weekly with UV at a dose of 1.0 kJ/m2 (i.e., -30% of MED) AsmSA AsmSA *0.7, 1.4,2.9,5.8 ppm (DW) * 2.9 ppm (DW) 161 days beginning at 21 days of age 29 wk 0.7 ppm 2.9 ppm Starting 21 days after the As111 treatments began, mice had their dorsal skin exposed to 1.0 kJ/m2 of solar spectrum UV (a low nonerythemic dose) 3 times weekly. Untreated control mice and inorganic arsenic-treated mice unexposed to UV developed no skin tumors. Of mice exposed to UV, skin tumor yields per mouse at the different doses of inorganic arsenic were as follows: 0, 2.40; 0.7, 5.40; 1.4, 7.21; 2.9, 11. 10; 5.8, 6.80. More than 95% of tumors were squamous cell carcinomas. Mice in all dose groups exposed to UV and inorganic arsenic showed a 2.5-3x ft in epidermal hyperplasia above that caused by UV alone, with the highest point estimate at 0.7. Immunohistological determination of oxidative DNA damage shown by staining of 8-oxo-dG: Control: no effect. UV alone: very slight ft. inorganic arsenic alone at 5.8 ppm (earlier experiment): ft. inorganic arsenic + UV (this experiment): huge ft. Co-treatment with vitamin E or p-XSC: ft (i.e., a significant reduction in inorganic arsenic + UV effect). Above effects roughly paralleled those for SCC induction, except that no tumors were caused by arsenic alone. Burns et al., 2004 Uddin et al., 2005 C-44 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference Co-mutagenesis Skin/mouse (Fi offspring from cross of FVB/N carrying Gil FLAP transgene x C57BL/6J, both sexes) AsmSA *5.8ppm(DW) 10 wk None, but 5.8 ppm if co- treatment withB[a]P Frequencies of induction of PLAP+ cells (result from frameshift mutations) in (A) untreated control, (B) group with inorganic arsenic treatment alone, (C) group with skin painting with B [a]P 5 days/week during weeks 3-10 after start of experiment, and (D) group with both BandC: A = x; B, ~1.9x, was a NSE; C, ~3.2x, was a NSE; D, ~10.7x. Also, significantly more of the individual mutations arose as clusters in group D, which suggests that more mutations arose in stem cells. This assay in bladder, spleen, lung, kidney, and liver yielded no obvious effect. Oxidation of guanosines in poly G tracts of G:C base pairs is thought to be one cause of these frameshift mutations. Fischer etal., 2005 Cytotoxicity Bladder/rat (F344, female) Urothelium/rat (F344, female) DMAV DMAV as sodium cacodyl ate- trihydrat e * 54.3 ppm (food) (assumes MW of chemical used was 138.0) *0.35, 1.4, 14,35 ppm (DW) 2wk 28 days 54.3 ppm 14 ppm Evidence of cytotoxicity by SEM as frequency of class-5 bladders, which showed necrosis and piling up of rounded urothelial cells: 6 of 10 rats, compared to 0 of 10 in control. In group with co-treatment with DMPS (a chelator of trivalent arsenicals), only 1 in 10 rats had a class-1 bladder. In another experiment with the same dose for 26 weeks, none of the rats had class-5 bladders. By light and transmission electron microscopy, no alterations were detected at lower 2 doses. At higher 2 doses, urothelial cells showed signs of swelling, appearance of cytoplasmic vacuoles and a decreased number of mitochondria (all being signs of cytotoxicity), with a positive dose- response. Cohen et al., 2002 Sen etal., 2005 DNA Damage Liver/rat (Fischer 344, male) (they used normal-appearing tissue) TMAVO *27.5, 110.2 ppm (DW) Estimated total intakes: 351 and 1363 mg As/rat 104 wk 110.2 ppm 8-OHdG formation assessed by HPLC: ft to ~1.22x; point estimate was also ft at lower dose, but the SE for it was large. Shen et al., 2003 C-45 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Lung/mice (C57BL/6J Oggl+/+wtmice, both sexes, 14 weeks old at start of treatment) Lung/mice (C57BL/6J Oggl '" knockout mice, both sexes, 14 weeks old at start of treatment) Peripheral blood leukocytes/mous e (C57BL/6J/Han, female) Lung/mouse (ddY, male) Liver/rat (Fisher 344, male) Bladder/rat (Fisher 344, male) Arsenic Species DMAV Asvas Na2HAs 04- 7H2O DMAV assumed to be dimethy lar-sinic acid MMAV DMAV TMAVO MMAV DMAV TMAVO Dose in Elemental Asa (in Units Stated) * 115.3ppm(DW) 50, 200, 500 ppb (DW) Note in ppb! *217ppm(DW) * 121ppm(DW)c * 109 ppm (DW)C * 110ppm(DW)c * 121ppm(DW)c * 109 ppm (DW)C * 110ppm(DW)c Duration of Treatment 72 weeks 3,6, 12 months 4wk 5, 10, 15, and 20 days for all 20 days for all LOELb None 115.3 50 ppb 2 17 ppm None None 110 ppm None 109 ppm None Results 8-OHdG formation assessed by HPLC, x = level of wt control: wt with inorganic arsenic treatment: ft to ~1.6x(NSE); knockout Oggl"'" without inorganic arsenic: ft to ~7.8x; knockout Oggl"7" with inorganic arsenic treatment: ft to -13. Ix. Half of the mice were maintained on a low-Se diet. Alkaline SCGE (comet assay) was used to detect DNA fragmentation (SSBs) and alkaline labile sites as well as oxidative DNA base damage identified by using FPG and Enm enzymes. The only significant inorganic arsenic effects were seen at 3 months, perhaps because water consumption (and thus inorganic arsenic consumption) was lower at the last 2 times sampled. An ft in DNA fragmentation was observed only in the mice with the low-Se diet, but there was no positive dose-response. An ft in oxidative DNA damage was observed only in the mice with the normal-Se diet, and again there was no positive dose- response. 8-oxo-dG levels: ft to 1.42x; subcutaneous injection of 10 mg/kg 4NQO just before 4-wk DMA treatment had no significant effect on this level; it was 1.3 8x. Use of feed containing 0.05% of the antioxidant EGCG was tested. 8-oxo-dG level in the (4NQO + DMA + EGCG) group was only 1.09x. 8-OHdG formation assessed by HPLC: TMAVO: ft on day 15 to ~1.5x and on day20to~1.82x. 8-OHdG formation assessed by HPLC: DMAv:ftto~1.62x. Reference Kinoshita et al., 2007b Palus etal., 2006 Mizoi et al., 2005 Kinoshita et al., 2007a Kinoshita et al., 2007a C-46 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference Effects Related to Oxidative Stress (ROS) Brain, liver, RBCs/rat (Wistar, male) Liver/rat (Fisher 344, male) Kidney and liver/rat (Wistar, female) Kidney and liver/rat (Wistar, male) As111 as SA MMAV DMAV TMAVO Asm ATO Asm ATO * 57.7 ppm (DW) * 121ppm(DW)c * 109 ppm (DW)C * 110ppm(DW)c * 30.3 mg/kg, 15 times (gavage) * 30.3 mg/kg, 15 times (gavage) 12 wk 5, 10, 15, and 20 days for all Every other day for 30 days Every other day for 30 days 57.7 ppm None 109 ppm 110 ppm 30.3 mg/kg x!5 30.3 mg/kg x 15 In liver and brain: U GSH levels; ft GSSG levels; ft MDA levels. InRBCs: U GSH levels; U ALAD levels; ft MDA levels. Some, but not all, of these effects were mitigated by oral post-treatment with NAC and/or DMSA. Oxidative stress in microsomes shown by elevation of total cytochrome P450 content and/or by ft in hydroxyl radical levels: DMAV for P450: ft on day 10 only to DMAV for OH radicals: ft on day 15 only to ~1.18x. TMAVO for P450: ft on days 10-20, maximum ft on day 15 to ~1.25x. TMAVO for OH radicals: ft on days 15 and 20, maximum ft on day 20 to ~1.33x. Kidney: MDA level ft to 3.8x; GSH level U to 0.78x; GSSG level ft to 7.5x; GST activity U to 0.44x. Liver: MDA level ft to 2.0x; GSSG level ft to 5.3x; GST activity U to 0.52x. Co-treatment with L-ascorbate reduced the size of the inorganic arsenic-induced effects (either ft or U) on all 4 endpoints in kidneys and on all but GSH in livers. Kidney: MDA level ft to 3.4x; GSH level U to 0.62x; GSSG level ft to 8.5x; GST activity U to 0.49x. Liver: MDA level ft to 2.7x; GSH level U to 0.82x; GSSG level ft to 5.9x; GST activity U to 0.49x. Co-treatment with L-ascorbate reduced the size of the inorganic arsenic -induced effects (either ft or U) on all 4 endpoints in kidneys and on all but GSH in livers. Flora, 1999 Kinoshita et al., 2007a Sohini and Rana, 2007 Sohini and Rana, 2007 C-47 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Blood, kidney, liver/mouse (albino Swiss, male) Liver/rat (Wistar, male) Liver/mouse (BALB/c, male) Liver/mouse (BALB/c, male) Arsenic Species AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) * 57.7 ppm (DW) *0.03, 1.4,2.9 ppm (DW) 50, 100, 150 ug/mouse/day for 6 days/week (gavage) 50, 100, 150 ug/mouse/day for 6 days/week (gavage) Duration of Treatment 8wk 60 days 3 months 6 months LOELb 57.7 Various 50 for ft None None None 50 for ft 50 for ft None 100 for ft 100 for U 100 for ft 100 for U 100 for U Results Blood: ALAD activity U to 0.32x; GSH level U to 0.78x; ROS level ft to 2.82x. Kidney: SOD activity U to 0.38x; CAT activity U to 0.34x; TEARS level ft to 1.17x; GSH level U to ~0.39x; GSSG level ft to ~2.5x; GPx activity U 0.94x (NSE). Liver: SOD activity U to 0.33 x; CAT activity U to 0.54x; TEARS level ft to 1.25x; GSH level U to ~0.44x; GSSG level ft to -3.1 x; GPx activity U 0.76x (NSE); G-6-P activity U to ~0.73x. Cytochrome P450 activity: ft to 1.41x and l.Slx at 1.4 and 2.9, respectively. MDA level: ft to 1.39x and 1.55x at 1.4 and 2.9, respectively. GSH level: U to 0.59x, 0.47x, and 0.42x at 3 doses in ascending order. SOD activity: U to 0.76x, 0.60x, and 0.55x at 3 doses in ascending order. U in activities of CAT, GPx, GR, G-6-P, and GST, respectively, to 0.90x, 0.75x, 0.50x, 0.76x, and 0.6 Ix at 1.4 ppm and to 0.54x, 0.66x, 0.42x, 0.64x, and 0.45x at 2. 9 ppm. Changes in various components of antioxidant defense system: GSH level: 50, 1.14x; 100, 1.17x; 150, 1.25x. MDA level: NSE at any dose. PSH level: NSE at any dose. PC level: NSE at any dose. GPx activity: 50, 1.12x; 100, 1.15x; 150, 1.24x. CAT activity: 50, 1.06x; 100, l.OSx; 150, l.lOx. Changes in various components of antioxidant defense system: GSH level: NSE at any dose. MDA level: 50, NSE; 100, 1.39x; 150, 1.44x. PSH level: 50, NSE; 100, O.Slx; 150, 0.75x. PC level: 50, NSE; 100, 1.16x; 150, 1.30x. GPx activity: 50, NSE; 100, 0.91x; 150, 0.90x. CAT activity: 50, NSE; 100, 0.94x; 150, 0.92x. Reference Mittal and Flora, 2006 Bashiretal., 2006a Das etal., 2005 Das etal., 2005 C-48 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Liver/mouse (BALB/c, male) Liver/mouse (BALB/c, male) Blood/rat (Wistar, male) Liver/rat (Wistar, male) Blood, kidney, liver/rat (Wistar, male) Arsenic Species AsmSA As111 SA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) 50, 100, 150 ug/mouse/day for 6 days/week (gavage) 50, 100, 150 ug/mouse/day for 6 days/week (gavage) * 57.7 ppm (DW) * 57.7 ppm (DW) *1.15 mg/kg/day (gavage) Duration of Treatment 9 months 12 months 6 weeks 6 weeks 3 weeks LOELb 50 for U 50 for ft 50 for U 50 for ft 50 for U 50 for U 50 for U 50 for ft 50 for U 50 for ft 50 for U 50 for U 57.7 ppm 57.7 ppm Various Results Changes in various components of antioxidant defense system: GSH level: 50, O.SOx; 100, 0.77x; 150, 0.66x. MDA level: 50, 1.97x; 100, 2.06x; 150, 2.16x. PSH level' 50 0 80x' 100 0 75x' 150 0.71x. PC level: 50, 1.64x; 100, 1.78x; 150, 1.94x. GPx activity: 50, 0.95x; 100, 0.91x; 150, 0.87x. CAT activity: 50, 0.95x; 100, 0.93x; 150, 0.92x. Changes in various components of antioxidant defense system: GSH level: 50, 0.76x; 100, 0.72x; 150, 0.63x. MDA level: 50, 2.20x; 100, 3.03x; 150, 3.97x. PSH level: 50, 0.73x; 100, 0.63x; 150, 0.56x. PC level: 50, 2.09x; 100, 2.91x; 150, 3.46x. GPx activity: 50, 0.87x; 100, 0.84x; 150, 0.75x. CAT activity: 50, 0.93x; 100, 0.92x; 150, 0.88x. Effects on levels of biochemical variables indicative of disturbances in the heme synthesis pathway and oxidative stress: ALAD U to 0.12x; GSH U to 0.73x; RBC ROS ft to 1.35x; GPx showed NSE. Effects on levels of biochemical variables indicative of oxidative stress: GSH U to 0.69x; GSSG ft to 1.41x; TEARS ft to 1.16x; catalase showed NSE. There was NSE for any of these parameters in the kidney. ALAD activity: blood, 0.45x. CAT activity: kidney, 1.12x (NSE); liver, 1.1 6x. GSH level: blood and kidney, NSE; liver, 0.79x. TEARS level: kidney, NSE; liver, 1.28x;. Co-treatment with NAC (i.p. injection) and/or zinc sulfate (oral) reduced some effects, especially when used together. Reference Das etal., 2005 Das etal., 2005 Kaliaetal., 2007 Kaliaetal., 2007 Modi etal., 2006 C-49 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Brain/rat (albino Wistar, male) Brain/rat (albino Wistar, male) Brain/rat (albino Wistar, male) Kidney, liver, RBCs/rat (albino Wistar, male) Arsenic Species AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) * 57.7 ppm (DW) * 57.7 ppm (DW) * 57.7 ppm (DW) *5.8ppm(DW) Duration of Treatment 60 days 60 days 60 days 12 weeks LOELb 57.7 ppm 57.7 ppm 57.7 ppm 5.8 ppm, but for only some effects Results Effects on levels of chemicals indicative of oxidative stress in 5 regions of the brain (hippocampus, cortex, striatum, hypothalamus, and cerebellum): MDA ft to from 1.64x to 2.2 Ix; GSH U to from 0.43x to 0.58x; GPx U to from 0.77x to O.Slx; GR U to from 0.73x to 0.78x; G6PDH U to from 0.70x to 0.84x. Simultaneous treatment with DL-a- lipoic acid markedly reduced all of these effects. Effects on levels of chemicals indicative of oxidative stress in 5 regions of the brain (hippocampus, cortex, striatum, hypothalamus, and cerebellum): ROS based on DCF assay ft to from 1.62x to 2.18x; total SOD U to from 0.56x to 0.77x; Mn SOD U to from 0.36x to 0.55x; Cu/Zn SOD U to from 0.53x to 0.62x; CAT U to from 0.67x to O..80x. Simultaneous treatment with DL-a- lipoic acid markedly reduced all of these effects. (This is the same experiment as in the previous row; findings not already listed in that row are listed here.) Measures of protein oxidation: ft in protein carbonyl level: cerebellum, 1.23x; cortex, 1.32x; hippocampus, 1.48x; hypothalamus, 1.25x; striatum, 1.49x; U in membrane protein sulfhydryl content: cerebellum, 0.71x; cortex, 0.55x; hippocampus, 0.50x; hypothalamus, 0.79x; striatum, 0.6 Ix; essentially the same regional pattern of inorganic arsenic -induced loss occurred with total protein-bound sulfhydryls. Co-treatment with DL-a-lipoic acid mostly or completely abolished all of the above effects. MDA level: ft in kidney to -2. Ix, in liver to ~1.7x, and in RBCs to ~1.4x. CAT activity: U in kidney to -0.73 x, in liver to -0.9 Ix (NSE), and in RBCs to -0.78. SOD activities were measured but with NSE. Co-treatment with cysteine, methionine, AA, or thiamine usually decreased tissue arsenic concentrations (especially in kidney and liver) and blocked oxidative damage to variable degrees. Reference Shila et al., 2005a Shila et al., 2005b Samuel etal., 2005 Nandi et al., 2005 C-50 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Kidney/rat (albino Wistar, male) Liver/rat (albino Wistar, male) RBCs/rat (albino Wistar, male) Liver and kidney/rat (albino Wistar, male) Liver and kidney/rat (albino Wistar, male) Arsenic Species AsmSA AsmSA AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) *5.8ppm(DW) *5.8ppm(DW) *5.8ppm(DW) * 57.7 ppm (DW) * 57.7 ppm (DW) Duration of Treatment 4, 8, 12 weeks 4, 8, 12 weeks 4, 8, 12 weeks 30 days 30 days LOELb Various for ft and U Various for ft and U Various for ft and U 57.7 ppm 57.7 ppm Results MDA level: ft at 4 wk to ~1.27x (NSE), at 8 wk to ~1.54x, and at 12 wk to -2.1 Ix. CAT activity: ft at 4 wk to ~1.72x, at 8 wk to ~1.18x (NSE) but U at 12 wk to ~0.75x. SOD activity: ft at 4 wk to ~1.84x, at 8 wk to ~1.23x, but U at 12 wk to 0.91x (NSE). MDA level: ft at 4 wk to ~1.07x (NSE), at 8 wk to ~1.46x, and at 12 wk to ~1.49x. CAT activity: ft at 4 wk to ~1. 19x (NSE), at 8 wk to ~1.52x but U at 12 wk to -0.9 Ix (NSE). SOD activity: ft at 4 wk to ~1.52x, at 8 wk to ~1.16x, but NSE at 12 wk. MDA level: ft at 4 wk to ~1. 13x (NSE), at 8 wk to ~1.28x, and at 12 wk to ~1.41x. CAT activity: ft at 4 wk to ~1.36x, NSE at 8 wk, and U at 12 wk to -0.7 Ix. SOD activity: ft at 4 wk to ~1.81x, at 8 wk to ~1.59x, but NSE at 12 wk. Level of ROS determined by DCFH assay: ft in liver to ~3.6x and in kidney to ~3.5x. Level of MDA released per mg protein: ft in liver to ~1.5x and in kidney to Co-treatment with both DL-a-lipoic acid and DMSA markedly reduced all of these effects. Activities of antioxidant enzymes: U of SOD in liver to ~0.51x and in kidney to ~0.55x. U of CAT in liver to ~0.59x and in kidney to ~0.58x. U of GPx in liver to ~0.53x and in kidney to ~0.56x. Levels of non-enzymatic antioxidants: U of GSH in liver to ~0.56x and in kidney to ~0.67x. U of AA in liver to ~0.48x and in kidney to ~0.50x. U of a-Toc in liver to ~0.49x and in kidney to ~0.58x. U of total sulfhydryls in liver to ~0.53x and in kidney to ~0.59x. Co-treatment with both DL-a-lipoic acid and DMSA markedly reduced all of these effects. Reference Nandi et al., 2006 Nandi et al., 2006 Nandi et al., 2006 Kokilavani et al., 2005 Kokilavani et al., 2005 C-51 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Blood (whole), brain, kidney, liver/mice (Swiss albino, male) Blood (whole), brain, kidney, liver/rat (Wistar, male) Blood (whole), brain/rat (Wistar, male) Arsenic Species AsmSA AsmSA AsmSA Dose in Elemental Asa (in Units Stated) * 14.4 ppm (DW) * 11.5ppm(DW) * 57.7 ppm (DW) Duration of Treatment 3 months 4wk 10 wk LOELb 14.4 ppm 11.5 ppm 57.7 ppm Results Whole blood: U of ALAD activity to 0.37x; U of GSH level to 0.93x. Brain: ft in TEARS level to ~2.2x; U in GSH/GSSG ratio to ~0.96x. Kidney: ft in TEARS level to 1.65x. Liver: ft in TEARS level to 1.21x; U in SOD activity to 0.76x; U in CAT activity to 0.89x; U in GSH/GSSG ratio to 0.89x. Post-treatments with 3 different extracts ofHippophae rhamnoides L. (thought to have antioxidant properties) showed various levels of effectiveness in reducing some of the above effects in all but the kidney. Whole blood: U of ALAD activity to 0.24x; U of GSH level to 0.86x; ft of ZPP level to 1.30x. Brain: ft in TEARS level to 1.89x; U in GSH level to 0.85x; NSE on GSSG level; U in SOD activity to 0.75x; U in CAT activity to 0.75x. Kidney: ft in TEARS level to 1.39x; U in GSH level to 0.55x; ft in GSSG level to 1.59x. Liver: ft in TEARS level to 1.96x; U in GSH level to 0.6 Ix; ft in GSSG level to 2.00x; oral co-treatment with Centella asiatica (thought to have antioxidant properties) showed various levels of effectiveness in reducing some of the above effects. Whole blood: ft of ROS level to 2.63x; U of ALAD activity to 0.46x; U of GSH level to 0.85x; U of Hb as grams/dL to 0.79x. Brain: ft of ROS level to 4.03x; ft in TEARS level to 1.50x; U in GSH level to 0.82x; U in SOD activity to 0.92x (NSE); U of ALAD activity to 0.58x; ft of ALAS activity to 1.21 x; U of GPx activity to 0.84x (NSE); ft of GST activity to l.OSx (NSE); "considerable" but unqualified ft in DNA fragmentation (single-strand breaks) was detected by polyacrylamide gel electrophoresis. Postreatment with the thiol chelating agents DMSA, DMPS, and MiADMSA showed various levels of effectiveness in reducing some of the above effects. Reference Gupta and Flora, 2005 Gupta and Flora, 2006 Flora et al., 2005 C-52 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Liver/mouse (BALB/c, male) Lung/mouse (ddY, male) Liver/rat (Wistar, male) Arsenic Species Unspeci fied arsenica l,but from discuss! on assumed to be AsmSA DMAV AsmSA Dose in Elemental Asa (in Units Stated) * 1.8ppm(DW) *217.2ppm(DW) * 3.6, 6.1,7.3 mg/kg (gavage, with animals being killed 24 hr later for sample collection) Duration of Treatment 3,6,9,12, 15 months 2, 4, 8, 15, 25 wk One dose LOELb 1.8at>9 months for MDA 1.8at>6 months for GSH 217.2ppm at 8 wk or longer Various Results MDA cone: ft to ~1.7x at 9, ~1.9x at 12, and~2.2xat 15. GSH content: U to ~0.84x at 6, ~0.78x at 9, ~0.67x at 12, and ~0.58x at 15. U in activities were also noted for G6PDH, GPx, and plasma membrane Na+/K+ ATPase at 6 months, for CAT at 9 months, and for GST and GR at 12 and 15 months. It seems likely that the activities remained lower at later times than when each U was noted, but that was not stated. Immunohistochemical analysis of 4HNE adducts showed that lipid peroxidation occurred in 48.8%, 72.9%, and 77.6% of terminal bronchiolar Clara cells by 8, 15, and 25 weeks, respectively. (None before that.) The modified proteins were specifically in the secretory granules of those cells. 8-OHdG adducts (showing oxidative DNA) damage were also demonstrated in the same cells. Clara cells are the major target cell for DMA- induced oxidative stress, and the authors suggested that lipid peroxidation via the formation of ROS is involved in promotion of lung tumor (malignant adenocarcinoma) formation following initiation by 4NQO. Significant dose-related ft in total arsenic cone at all doses; cone in liver at highest dose was ~22 times that in brain. MDA cone: ft to 1.43x at 6.1 and 1.52x at 7.3. GSH level: U to 0.57x at 3.6, to 0.41x at 6.1,andto0.39xat7.3. Total cytochrome P450 activity: ft to 1.46x at 6.1 and 1.54x at 7.3. SOD level: U to 0.67x at both 6. 1 and 7.3. CAT activity: U to 0.54x at 6. 1 and 0.49xat7.3. GPx activity ft to 1. 15x at 3.6, 1.21x at 6.1,andl.27xat7.3. GST activity: U to 0.72x at 6. 1 and 0.62x at 7.3. NSE on either GR or G6PD activity. Reference Mazumder, 2005 Anetal., 2005 Bashir et al., 2006b C-53 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference Brain/rat (Wistar, male) AsmSA * 3.6, 6.1,7.3 mg/kg (gavage, with animals being killed 24 hr later for sample collection) One dose Various Significant ft in total arsenic cone at both higher doses. MDA cone: ft to 1.48x at 6.1 and 1.56x at 7.3. GSH level: U to 0.79x at 3.6, to 0.60x at 6.1,andto0.51xat7.3. SOD level: U to 0.73x at 6.1 and 0.70x at 7.3. CAT activity: U to 0.58x at 6.1 and 0.51xat7.3. GPx activity ft to 1.17x at 6.1, and 1.26x at 7.3. GST activity: U to 0.7Ix at 6.1 and 0.69x at 7.3. NSE on either GR or G6PD activity. Bashir et al.. 2006b Kidney, rat (Wistar, male) Asm ATO * 30.3 mg/kg, 15 times (gavage) Every other day for 30 days 30.3 x 15 GSH content U to ~0.59x. GST activity: NSE. Ranaand Allen, 2006 Gene Mutations Skin/mouse (Aprt+/~ hybrid mice of complex genotype needed for assay: see paper) AsmSA *5.7ppm(DW) 10 wk None Starting 2 wk after consumption of inorganic arsenic-contaminated water began, half of the mice were also exposed to B[a]P for 8 wk by skin painting. Skin was assayed for DAP- resistant (DAP1) colonies indicative of cells lacking Aprt activity as the result of loss of heterozygosity (LOH) at Aprt because of malsegregation or mitotic recombination in vivo. No significant differences were found because of inorganic arsenic and/or B[a]P exposure. and thus there was no evidence that inorganic arsenic alone, or by enhancement of a known mutagen (but not one + in this assay), caused such genetic changes. Curiously, the point estimate for most LOH was in the control (45%); it was 38% for B[a]P alone, 8% for inorganic arsenic alone, and 30% for them together. Because there was much variability, these seemingly large differences were not statistically significant. Fischer etal. 2006 C-54 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Wing/Drosophila melanogaster Arsenic Species DMAV Dose in Elemental Asa (in Units Stated) 0.05,0.1,0.25,0.5 mM (in medium) Duration of Treatment 72 hr LOELb 0.25 mM, regarding total spots Results SMART (somatic mutation and recombination test) wing spot assay: positive dose-response was found, but nature of induced mutations was uncertain. Was earlier shown that inorganic arsenic is inactive in this assay. They showed no biomethylation occurs in larvae or in growth medium. Results suggest importance of biomethylation as a determinant of genotoxicity of arsenic compounds, at least in Drosophila. Reference Rizkietal., 2006 Hypermethylation of DNA Lung/mice (A/J, male) Asvas Na2HAs Cv 7H2O * 0.24, 2.4, 24 ppm (DW) 18 months The LOEL was 0.24 ppm. Extent of hypermethylation of promoter regions of tumor suppressor genes p\6WK4a and RASSF1A in lung adenocarcinomas from inorganic arsenic exposed mice compared to the control, based on methylation- specific PCR: percentages of methylated promoters of pl6INK43 in jung tumors of Q Q 24^ 2 4 and 24 ppm dose groups were 11%, 30%, 36%, and 42%, respectively. Percentages of methylated promoters of RASSF1A in lung tumors of the same dose groups were 33%, 70%, 82%, and 89%, respectively. Reduced expression, or lack of expression, of these 2 genes was correlated with the extent of hypermethylation. There was constant expression of these genes in lungs without tumors in both control and inorganic arsenic-treated mice. They concluded that epigenetic changes of tumor suppressor genes are involved in inorganic arsenic-induced lung carcinogenesis. Cuietal., 2006 Hypomethylation of DNA Liver cells/mouse (129/SvJ) AsmSA 45 ppm (DW) 48 wk 45 ppm There was global DNA hypomethylation, as shown by 5- methylcytosine content of DNA and by using the methyl acceptance assay. In particular, there was a marked U in methylation within the ER-a gene promoter region, which was statistically significant in 8 of 13 CpG sites. Control had 28.3% of ER-a sites methylated, but experimental group had 2.9%. Chen et al., 2004b C-55 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Livers of newborn males/mouse (only pregnant C3H females drank the water) Arsenic Species AsmSA Dose in Elemental Asa (in Units Stated) 85 ppm (DW) Duration of Treatment 10 days, gestation days 8 to 18 LOELb 85 ppm Results Global DNA methylation status was not significantly altered based on methyl acceptance assay, which measures methylation in both quiescent and active areas of DNA. However, another assay showed that GC-rich regions globally were less methylated if they were from livers of newborn males exposed in utero to inorganic arsenic. Band intensity showing the extent of methylation was 0.20x after Rsal + Mspl digestion and 0.40x after Rsal + Hpall digestion. Mspl and Hpall are methylation sensitive enzymes. Reference Xieetal., 2007 Interference With Hormone Function Kidney, rat (Wistar, male) Asm ATO * 30.3 mg/kg, 15 times (gavage) Every other day for 30 days 30.3 x 15 T3 and T4 levels in serum: triodothyronine (T3) ft to ~4.8x; thyroxine (T4) ft to ~1.7x. Ranaand Allen, 2006 Signal Transduction Fetal lungs/mouse (only pregnant C3H females drank the water, female offspring only) Adenomas and adeno- carcinomas from lungs of adults exposed in wtero/mouse (only pregnant C3H females drank the water, female offspring only) AsmSA AsmSA 85 ppm (DW) 85 ppm (DW) 10 days, gestation days 8 to 18 10 days, gestation days 8 to 18 85 ppm 85 ppm ft in ER-a transcript (5.3x) and protein levels; ft in expression of the following estrogen-related genes: trefoil factor-3 (9.66x), anterior gradient-2 (3.21x); ft in expression of the following steroid metabolism genes: 17-p-hydroxysteroid dehydrogenase type 5 (3.55x) and aromatase (2.53x). (Expression of ER-a and the ER-linked genes was unchanged in male fetal lung as compared to control.) The insulin growth factor system was also activated, with transcripts for IGF-1, IGF-2, IGF-R1, IGF-R2, IGF- BP1, and IGF-BP5 all being increased to 1.6-2.5x. Also, there was overexpression of the following genes that have been associated with lung cancer: AFP (6.9x), EGFR (3.2x), L- myc (1.9x), and metallothionein-1 Based on immunohistochemical analysis: intense and widespread ft in nuclear ER- a expression; in contrast, normal adult lung and DENA-induced lung adenocarcinoma showed little evidence of ER-a expression. Shen et al., 2007 Shen et al., 2007 C-56 DRAFT—DO NOT CITE OR QUOTE ------- Tissue or Cell Type/Species Arsenic Species Dose in Elemental Asa (in Units Stated) Duration of Treatment LOELb Results Reference a When doses were reported in mg arsenic/L or in ppm As, it was assumed that the doses included adjustment for the amount of arsenic in solution. Because it was sometimes unclear from the papers whether a correction was needed, a "*" was put front of the doses listed in the table if those doses were corrected to the amount of arsenic in the dose. b Lowest observed effect level. 0 Estimates were based on the reported concentrations of MMAV, DMAV, and TMAVO in DW of 1.62, 1.45, and 1.47 mM, respectively, and on their molecular weights (MWs) of 139.969, 137.997, and 136.025 g and on the atomic weight of arsenic of 74.926 g. The paper stated that the concentrations of all arsenicals were 0.02% (or 200 ppm). For the arsenicals themselves, the concentrations were actually 226, 200, and 200 ppm, respectively, if based on the MWs just listed. Table C-3. In vitro studies related to possible MOA of arsenic in the development of cancer Type of Cell/Tissue Arsenic Species Concentration(s) Tested (\iM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in jiM Unless Noted) Reference Aberrant Gene or Protein Expression HaCaT cells TRL 1215 cells (normal rat liver) Hepa-1 cells (mouse hepatoma) As111 SA As111 SA As111 SA 0.5, 1.0 0.125,0.250, 0.500 1,3,10,30 20 passages 24 wk 30 min before 4hr co-treatment with 1 nM TCDD 0.5 0.500 for effects noted here 1 ft intracellular GSH quantities. U keratins 5, 6, 7, 8, 10, and 17. Using Atlas Rat cDNA expression microarrays, -80 of the 588 genes assayed were aberrantly expressed — including genes related to stress and DNA damage, signal transduction modulators and effectors, apoptosis- related proteins, cytokines and cytokine- related components, and growth factors and hormone receptors. Results of Northern blot analysis of mRNA: ft TCDD-inducible levels ofNqolmRNA; response was much higher at 3 and 10, but decreased markedly at 30 to slightly more than was present at 1. Chien et al., 2004 Chenetal., 2001 Maier et al., 2000 C-57 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Huh? cells Huh? cells, transfected for use in the DRE-CALUX bioassay PARP-1+/+ MEF cells PARP-I-'- MEF cells Arsenic Species As111 SA As111 SA As111 SA for both Concentration(s) Tested (nM) 0.5,1,3,5, 10,20 0.5,1,3,5, 10,20 11.5 for both Duration of Treatment 24 hr 24 hr 24 hr for both LOECa (HM) 3 3 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Following co-treatment withlOnMTCDD:li TCDD-inducible level of CYP1A1 activation to -45% of level without inorganic arsenic, then reached plateau of -18% at doses of 5-15 (based onEROD assay); inorganic arsenic did not affect CYP1A1 activation by itself . Following co-treatment withlOnMTCDD: U TCDD-inducible luciferase activity in the DRE-CALUX bioassay to -80% of level without inorganic arsenic, followed by a dose- related U to 42% at dose of 20. In a microarray gene chip analysis that analyzed the expression pattern of more than 34,000 genes, —311 genes were found to be differentially expressed among the different groups (i.e., control versus inorganic arsenic treatment or in comparisons between the 2 genotypes). Many of those genes belonged to the following groups: responders to stress and external stimuli, genes related to cell growth and maintenance, cell death, or DNA metabolism. While some genes were markedly up-regulated in both genotypes (sometimes to widely different amounts), other genes were up-regulated for one genotype and down-regulated for the other, and vice versa. Reference Chao etal., 2006b Chao etal., 2006b Poonepalli etal., 2005 C-58 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells PRCCs HEK293 cells Arsenic Species As111 ATO As111 ATO for both Concentration(s) Tested (nM) 0.5 0.1 1 Duration of Treatment 6, 12, 24, 48, and 72 hr for transcriptome analysis; 12 and 48 hr for proteomic analysis 10 min, 1, 6, 24 hr for both LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) In a microarray and 2-dimensional gel electrophoresis (with mass spectrometry) study aimed at understanding effects of therapies with ATO alone, retinoic acid alone, and their combined therapy, the main findings for ATO were as follows. At the transcriptome level, ATO affected regulation of 487 genes, many of which were probably related to essential aspects of cell-activity control such as induction of differentiation antigens, modulation of apoptosis regulators, and regulation of genes involved in cell-cycle and growth control. Other groups of affected genes included those involved with protein degradation, cell defense, stress response, protein modification and synthesis, and a group of 5 down- regulated HLA-class I genes. At the proteome level, ATO affected 982 protein spots, and there was often a time-dependent pattern of regulation, with much hr than at lower protein levels at 48 12 hr after treatment. A group of enzymes involved in biochemical metabolism was found to be significantly down-regulated, and there was a strong reduction of cytoskeleton proteins, implying a considerable reorganization of the cell nucleus and cytoplasmic structures. By comparison with relatively minor changes at many of the corresponding genes at the transcriptome level, the significant changes found at the proteomic level suggest that ATO particularly enhances mechanisms of post-transcriptional/translational 0. 1 at 6 hr Iat6hr modification. HMOX1 gene expression (mRNA levels measured by quantitative PCR): In PRCCs: NSE at 10 min or 1 hr; ~2.3x at 6 hr, ~2.8xat24hr. HEK293:NSEatlOmin or 1 hr; ~40x at 6 hr, ~54x at 24 hr. Reference Zheng et al., 2005 Sasaki et al., 2007 C-59 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PRCCs HEK293 cells PRCCs HEK293 cells HCT15 cells HeLa cells PLC/PR/5 cells Chang cells Arsenic Species As111 ATO for both As111 ATO for both As111 SA for all Concentration(s) Tested (nM) 0.1,0.5,2 for both 0.1 for both 278.33, the LC50 200.33, the LC50 376.66, the LC50 328.33, the LC50 Duration of Treatment 24 hr for both 10 min, 1, 6, 24 hr for both 24 hr for all LOECa (HM) 0.1 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) HMOX1 gene expression (mRNA levels measured by quantitative PCR): In PRCCs: 0.1, 2.2x, 0.5, 11.7x;2, 33 5X InHEK293:0.1, 1.2x, 0.5, 8.3x; 2, 224.9x. Western blot analysis for heme oxygenase 1 protein for dose of 1 for 24 hr: Huge ft in PRCCs andbigflinHEK293. Microarray analysis identified 73 genes whose expression changed in both types of cells, and for many expression increased in a time- dependent manner. These included HMOX1, Bax (involved in induction of apoptosis), and genes involved in many other biological processes including intracellular protein transport, signal transduction, differentiation, GSH metabolism, and protein complex assembly among others. Data were presented that suggest that heme oxygenase 1 protein confers a cytoprotective effect against inorganic arsenic treatment. 278.33 200.33 376.66 328.33 Western blot assay to determine eIF4E protein levels: for all cell lines, there was a reduction in the protein level to roughly 50%-60%ofthe corresponding control level. There was also a statistically significant, but smaller, U after 16 hr for all lines. Reference Sasaki et al., 2007 Sasaki et al., 2007 Othumpan gat et al., 2005 C-60 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HCT15 cells HeLa cells PLC/PR/5 cells Chang cells HeLa cells HeLa cells, HCT15 cells, CHO-K1 cells TR9-7 cells that were released from being mostly synchronized in G2 (using Hoechst 33342) shortly before inorganic arsenic treatment began Arsenic Species As111 SA for all As111 SA As111 SA for all As111 SA Concentration(s) Tested (nM) 278.33, the LC50 200.33, the LC50 376.66, the LC50 328.33, the LC50 200 Various 5 Duration of Treatment 24 hr for all 24 hr LOECa (HM) 278.33 200.33 376.66 None 200 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Quantitative real-time PCR to determine eIF4E mRNA levels: there was a statistically significant U only in lines HCT 15 and HeLa. Actual data on gene expression, in arbitrary units: HCT15: no inorganic arsenic, 0.099, with inorganic arsenic, 0.049. HeLa: no inorganic arsenic, 0.041, with inorganic arsenic, 0.029. PLC/PR/5: no inorganic arsenic, 0.051, with inorganic arsenic, 0.028. Chang: no inorganic arsenic, 0.018, with inorganic arsenic, 0.019. (Judging from their SEs, the result for PLC/PR/5 must have been of borderline significance.) Western blot assay to determine protein levels: Big U in cyclin Dl. ft in cellular levels of ubiquitin and in the process of ubiquitination. Additional experiments involving a genetically modified cell line, an siRNA that targeted expression of eIF4E, and proteasome inhibitors suggested (1) that the changes seen in eIF4E protein levels played a role in the observed cytotoxicity, (2) and that the inhibition of cyclin D 1 is mediated through the inhibition of eIF4E, and (3) that the inorganic arsenic stimulated ubiquitination and the resulting proteolysis play an important role in reducing eIF4E protein levels. 3-24 hr Conclusions based on determining protein levels using Western blot assays until 24 hr of inorganic arsenic treatment in cells made p53(+) or p53(-) by controlling tetracycline levels: big ft in ID 1, but it occurred only when there was p53 protein present, arsenic p53 protein level decreased, ID1 protein level decreased. The general finding was confirmed by microarray analysis. Work by others showed that ID1 protects against apoptosis through activation of the NF-KB signaling pathway. Reference Othumpan gat et al., 2005 Othumpan gat et al., 2005 Othumpan gat et al., 2005 McNeely et al., 2006 C-61 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TR9-7 cells that were released from being mostly synchronized in G2 (using Hoechst 33342) shortly before inorganic arsenic treatment began HeLa cells TRL1215 cells Arsenic Species As111 SA As111 ATO As111 SA Concentration(s) Tested (nM) 5 2 0.125,0.250, 0.500 Duration of Treatment 3hr 6 and 24 hr 24 weeks LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Conclusions based on microarray analysis (done by hybridizing fragmented cRNAs to U95Av2 GeneChips) in cells made p53(+) or p53(-) by controlling tetracycline levels: several genes were induced by inorganic arsenic independently of p53 status, of which some of the biggest effects were as follows (at both p53 conditions): HMOX1: huge ft by >25x; MT2A: ft by >3x; SLC30A1: ft by >3x. MKP-1 was induced only in p53(+) cells, and ubiquitin-conjugating enzyme E2N was induced only in p53(-) cells. In a cDNA microarray -based global transcription profiling experiment that compared the inorganic arsenic treatment with a co-treatment of the same inorganic arsenic dose with 30 uM emodin, the numbers of genes with an expression level that differed between the two treatments by more than a factor of 2 at the 2 time points were 793 and 480, respectively. The affected genes included genes involved in such things as cell signaling, organelle functions, cell- cycle control, redox regulation, and apoptosis. The manner of data presentation did not permit identification of genes affected exclusively by inorganic arsenic. Various mRNA levels determined by real time RT-PCR: effects on oncogenes AFP: ft at 0.250, big ft at 0.500; WT-1: ft at 0.125, big ft at 0.250 and 0.500. c-jun: ft at 0.250, big ft at0.500;H-ras:ftat 0.125, big ft at 0.250 and 0.500. (By 24 weeks of exposure at the higher 2 doses, these cells had undergone malignant transformation and were called CAsE cells.) Reference McNeely etal.,2006 Wanget al., 2005 Liu et al., 2006d C-62 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL 1215 cells TRL 1215 cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 0.125,0.250, 0.500 0.125,0.250, 0.500 Duration of Treatment 24 weeks 24 weeks LOECa (HM) Various Various Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) mRNA levels determined by real time RT-PCR: effects on stress-related genes HMOX-1: flat 0.125 and 0.250, big ft at 0.500; SOD: flat 0.250, big flat 0.500. MT-1: big ft at 0.250, ft at 0.500; GSTjc: ft at 0.125, big ft at 0.250 and 0.500. (By 24 weeks of exposure at the higher 2 doses, these cells had undergone malignant transformation and were called CAsE cells.) mRNA levels determined by real time RT-PCR: effects on cell cycle regulators CyclinDl:ftat0.125, then ft with dose to 0.500. PCNA: ft at 0.250, big ft at 0.500. p21 : big U at 0.125, then U with dose to 0.500. p!6: Hat 0.125, big U to -0% at 0.250 and 0.500. (By 24 weeks of exposure at the higher 2 doses, these cells had undergone malignant transformation and were called CAsE cells.) Reference Liu et al., 2006d Liu et al., 2006d C-63 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL 1215 cells TRL 1215 cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 0.125,0.250, 0.500 0.125,0.250, 0.500 Duration of Treatment 24 weeks 24 weeks LOECa (HM) Various Various Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) mRNA levels determined by real time RT-PCR: effects on growth factor genes c-met: big ft at 0.125, then ft with dose to 0.500. HGF: ft at 0.125, big ft at 0.250 and 0.500. FGFR1: huge U at 0.250, then U to -0% at 0.500. IGF-II: huge U to -0% at all doses. (By 24 weeks of exposure at the higher 2 doses, these cells had undergone malignant transformation and were called CAsE cells.) Protein levels determined using Western blots: AFP: slight ft at 0.125 through 0.500; WT-1: huge ft at 0.125 through 0.500. CyclinDl:ftat0.125 through 0.500; p!6: huge U at all doses. p21 :li at 0.125, then U with dose to 0.500. (By 24 weeks of exposure at the higher 2 doses, these cells had undergone malignant transformation and were called CAsE cells.) Reference Liu et al., 2006d Liu et al., 2006d C-64 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL 1215 cells CL3 cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 0.500 2 Duration of Treatment 24 weeks 24 hr LOECa (HM) Various 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) mRNA levels determined by real time RT-PCR: effects of 72 -hr post- treatment with 5 |iM 5- aza-dC (results were compared to cells with inorganic arsenic treatment alone) Mr-lift 19xover already elevated level. p21: ft 15x over what was a greatly reduced level, and level then far above that with no inorganic arsenic exposure pl6andIGF-II:NSE. (By 24 weeks of exposure at the higher 2 doses, these cells had undergone malignant transformation and were called CAsE cells.) ft Nqol mRNA level to 1.7x control; ft Nqol protein level to 6.4x control. Cells given this inorganic arsenic pretreatment became more sensitive to MMC- induced cytotoxicity and less sensitive to ADM- induced cytotoxicity. Co-treatment with MMC and the Nqol inhibitor DIG resulted in big ft in cell survival (even higher than after MMC treatment without an inorganic arsenic pretreatment). CL3R15 cells, which have much higher levels of Nqo 1 activity than CL3 cells, are also much more sensitive to MMC- induced cytotoxicity than CL3 cells. Reference Liu et al., 2006d Lin et al., 2006 C-65 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue H460 cells CL3 cells CL3R15 cells CL3R15 cells co-treated with 200 uM DIC for 6 hr to inhibit >95%ofthe high endogenous level of Nqo 1 activity Arsenic Species As111 SA for both As111 SA for both Concentration(s) Tested (nM) 2.5,5, 10,20 1,2.5,5, 10 50, 100, 200 for both Duration of Treatment 72 hr for both 6hr for both LOECa (HM) 2.5 1 100 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival determined by SRB assay: LC50s: H460, 9.0; CL3, 3.7; H460 cell have ~30x higher endogenous Nqol activity than CL3 cells, and unlike CL3 cells they showed no statistically significant induction of Nqol after 24-hr treatments with inorganic arsenic at doses of 2, 5, or 10. (Even at the highest level of induction in CL3 cells, the endogenous level of Nqol activity in H460 cells was still ~15x higher.) These findings raised question whether Nqol plays a role in inorganic arsenic resistance. Cell survival determined by colony-forming assay: LC50s: with DIC, -35; without DIC, 120. Reference Lin et al., 2006 Linetal., 2006 C-66 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SIK cells Arsenic Species As111 SA Concentration(s) Tested (nM) 2 Duration of Treatment 1,3,5,7,9 days LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Changes in protein levels detected at each of the 5 times using 2- dimensional gel electrophoresis of soluble proteins, with proteins identified by peptide mass mapping and other methods: -300 distinct protein spots were monitored with -40% showing >2-fold ft or U in silver staining intensity at every time point, about as many ft as U, with at least as many changes on day 1 as on other days. There were some changes as to the proteins affected over time. Of 10 proteins identified as showing prominent changes within first few days of inorganic arsenic treatment, enzymes of the glycolytic pathway were seen to be substantially elevated. This dose of inorganic arsenic suppressed differentiation but did not cause cell loss. Reference Lee et al., 2005 C-67 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL 1215 cells Arsenic Species MMAV DMAV TMAVO Concentration(s) Tested (nM) 1300 700 10000 Duration of Treatment 20 weeks for all LOECa (HM) 1300 700 10000 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft GST activity to 2.6x control, ft cellular GSH protein level to 2.2x control. ft GST activity to 1.7x control, U cellular GSH protein level to 43% of control. ft GST activity to 1.8x control, ft cellular GSH protein level to 2.4x control. All 3 treatments increased GST, MRP and MDR at the mRNA level, and all 3 treatments increased GST, Mrps, and P-gp at the protein level. GST and MRP have several forms. While not all forms responded in the same way, the overall responses were as noted. Experiments with inhibitors of GSH, Mrps, and P-gp led to the conclusion that increased arsenic excretion caused the resistance to arsenic- induced cytotoxicity that resulted from these treatments. Reference Kojima et al., 2006 C-68 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Ahr+/+MEFs Ahr+/+MEFs AG06 cells AG06 cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 1,2,5 Duration of Treatment 6hr LOECa (UM) 2 for Nqol only None for CYP1B1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) mRNA levels measured by real-time RT-PCR: ft NqolmRNAto4x control; 5 uMB[a]P increased Nqo 1 mRNA to 8x control; there was a synergistic interaction between them such that the dose of 2 of inorganic arsenic plus thedoseof5ofB[a]P increased Nqo 1 mRNA to 27x control. A synergistic interaction to 20x control also occurred with a dose of 1 of inorganic arsenic. At a dose of 5 of inorganic arsenic, the interaction became only additive. The interaction between inorganic arsenic and B[a]P regarding CYPlBlmRNAwas never more than additive. In Ahr A MEFs, there was no interaction of inorganic arsenic and B(a)P regarding Nqol mRNA; the combined treatment did not ft Nqol mRNA levels. Thus the synergistic interaction requires the wt Ahr gene. Following treatment with 2 uM inorganic arsenic, 5 uM B[a]P, or both, for an unspecified time, oligonucleotide microarray analysis of 13,332 sequences from annotated mouse genes: they identified 64 genes that were up-regulated or down-regulated by inorganic arsenic, B[a]P, or both; of these, 13 showed at least a 2x up-regulation and 12 caused at least a 2-fold down-regulation in gene expression because of the inorganic arsenic treatment alone. Many different types of genes were affected. One of the major consequences of exposure to these mixtures was the up-regulation of oxidative stress and protein chaperone responses and the down-regulation of the TGF-(3 pathway. Exposure to inorganic arsenic/B[a]P mixtures caused regulatory changes in the expression of detoxification genes that ultimately affect the metabolic activation and disposition of toxicants. 0.2, 1, 3, 10 3 24 hr 48 hr 48 hr 1 0.2 o J ft GSH concentration. Specific activities: GSTrc ftto~1.6xandyGCSto ~2.2xatdoseof3. Reference Kannet al., 2005a Kannet al., 2005a Snow et al., 1999 Snow et al., 1999 C-69 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue AG06 cells GM847 cells GM847 cells AG06 cells WI38 cells HaCaT cells HaCaT cells JB6 C141 cells K562 cells Arsenic Species As111 SA As111 SA As111 SA As111 SA As111 SA As111 SA, Asv MMAV, DMAV As111 SA As111 ATO Concentration(s) Tested (nM) 0.1,0.25,0.5, 1.0, 5, 10, 25 0.5, 1.0, 10, 25 0.2, 4, 20 0.3, 1.4,5.7,29 0.001,0.01,0.05, 0.1,0.5, 1.0 1.0 1.0 0.05, 0.2, 0.8, 3.125, 12.5,50, 200 2.5 Duration of Treatment 24hr 24hr 3hr 24 hr Not reported 2 days 14 days 2 days for all 15 min 6hr LOECa (HM) 1.0 0.25 0.5 for ft 0.5 for U 0.2 0.2 0.3 0.1 0.01 1.0 None 0.8 2.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft GR protein level to 2.9xatl. ft GR mRNA level to 1.3x and enzyme activity to 2.0x at 0.25. ft Trx, TrxR, GR mRNA levels; for TrxR and GR: ftto~2.7xby 10 and thenlito~1.5xby25). U GPx mRNA level to ~0.5xby land~0.2xby 25. APE/Ref-1 mRNA levels: at 3 hr: ft to ~2.7x at 0.2 and then only slight ft to ~3. Ox at 20. At 24 hr: ft to ~3. Ox at 0.2butUto~0.9xat20. (APE/Ref-1 is required forBER.) ft DNA Poly p level (both cytoplasmic and nuclear) to ~2xby 1.4 butUto~0.8xby29. (DNA Poly p is required forBER.) U p53 protein; ft mdm2 protein. U p53 protein; ft mdm2 protein. U p53 protein; ft mdm2 protein; (much bigger effect for As111). No significant change. ft Erk activation resulting from Erk phosphorylation; another experiment showed that overexpression of dominant negative Erk2 blocks arsenite-induced activation of Erk. ft GlycoA, HLA-DR, CD33, andCD34onthe cell surface, indicating maturation of myeloid cells. Reference Snow et al., 2001 Snow et al., 2001 Snow et al., 2001 Snow et al., 2001 Hamadeh etal., 1999 Hamadeh etal., 1999 Huang et al., 1999a Li and Broome, 1999 C-70 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MCF-7 cells H460 cells Primary cultures of rat cerebellar neurons MC/CAR (human multiple myeloma cell line) Arsenic Species As111 ATO As111 ATO As111 SA As111 ATO Concentration(s) Tested (nM) 3 10 10 2 Duration of Treatment 12hr 24hr 24 hr 72 hr LOECa (HM) 3 10 10 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Microtubule polymerization, with a major effect on the organization of the cellular microtubule network, resulting in the formation of long polymerized microtubule bundles; ft p34cdc2/cyclin B complex (both activation and accumulation); ft Bcl-2 phosphorylation. The following changes occurred only in mitotic cells (definitely not in interphase cells): ft caspase-3 activation, ft caspase-7 activation, cleavage of PARP and (3- catenin. These findings suggest that arsenic- induced mitotic arrest may be a requirement for the activation of apoptotic pathways. ft caspase activity (apoptosis is blocked in these cells if caspase is inhibited; there was a much bigger effect with a 48-hr treatment). ft caspase-3 activity, p21,andCDKl;up- regulation of cdc2 phosphorylation; U in CDK6, cdc2, cyclin A, and Bcl-2 levels; ft binding of p21 with CDK6, cdc2, and cyclins A and E; U activity of CDK6-associated kinase and cdc2 -associated kinase; loss of mitochondria! transmembrane potential (Av|/m); no change in p27, CDK2, CDK4, or cyclins B1,D1, or E levels. Reference Ling et al., 2002 Ling et al., 2002 Namgung and Xia, 2001 Parketal., 2000 C-71 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PCI-1 (human head and neck squamous cell carcinoma cell line) (Human myeloma-like cell lines) RPMI 8226 Karpas 707 U266 LAK effector cells WRL-68 (human hepatic cell line) Human aorta VSMCs (vascular smooth muscle cells) Human aorta VSMCs (vascular smooth muscle cells) WI38 cells WI38 cells Arsenic Species As111 ATO As111 ATO As111 ATO As111 SA As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 2 0.5 0.5 0.001,0.01,0.1, 10 2.5, 5, 10 2.5, 5, 10, 20 0.1 10, 20, 50 0.1 50 Duration of Treatment 3 days 72 hr 72 hr 16hr 4hr 4hr 14 days 18 hr 14 days 18 hr LOECa (HM) 2 0.5 0.5 0.1 0.001 ~5 ~5 0.1 50 0.1 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft p21 and its binding with cdc2; U protein levels of cdc2 and cyclin Bl; U activity of cdc2 kinase; no change in CDK2, CDK4, CDK6 and cyclins A, Dl, E. ft CD38 and CD54 (molecules involved in cell-cell interactions). ftCDllaandCD31 (molecules involved in cell-cell interactions, and the ligands [i.e., counter- receptors] of CD54 and CD38, respectively). ftGSH. ft CK18. ft p22phox mRNA expression (p22phox is 1 of at least 7 subunits of NADH oxidase.) U oc- actin mRNA expression. ft NADH oxidase activity. The effect was even stronger, with a LOEC of 1, in nonproliferating VSMCs. ft p53 (3 -fold increase). ft p53 (large increase). ft cyclin Dl; also treatment blocks ft in p21 that occurs follow exposure to 6 Gy of ionizing radiation. U cyclin D 1 ; also treatment mostly blocks ft inp21 that occurs follow exposure to 6 Gy of ionizing radiation. Reference Seoletal., 1999 Deaglio et al., 2001 Deaglio et al., 2001 Ramirez et al., 2000 Lynn et al., 2000 Lynn et al., 2000 Vogtand Rossman, 2001 Vogtand Rossman, 2001 C-72 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Untransforme dand immortalized RWPE-1 cells (human prostate epithelial cell line) PAEC from freshly harvested vessels Arsenic Species As111 SA As111 probably ATO, but called arsenite Concentration(s) Tested (nM) 5 5 Duration of Treatment Up to 30 wk 15 minto 3 hr depending on endpoint LOECa (HM) 5 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft MMP-9 activity (likely biomarker of when malignant transformation occurred); U in DNA methyltransferase activity but no change in DNA methyltransferase mRNA levels; ft K-ras mRNA and protein levels. Time course study suggested over- expression of K-ras preceded malignant transformation. There was no indication of mutations being induced in K-ras gene and no indication that hypomethylation of K- ras promoter region caused K-ras changes. The cells became tumorigenic after 29 weeks of treatment and were then called the CAsE-PE cell line. ft NF-KB dependent transcription, ft H2O2- dependent tyrosine phosphorylation (which was blocked by CAT), ftcSrc activation. MAP kinases, extracellular signal-regulated kinase, and p3 8 were only activated at a dose of 100, which causes cell death. Reference Benbrahim -Tallaa et al., 2005 Barchowsk y et al., 1999a C-73 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HeLa S3 cells TM3 cells E7 cells Arsenic Species As111 SA As111 SA As111 ATO Concentration(s) Tested (nM) 5 0.008, 0.77, 7.7 0.025,0.05,0.1, 0.25,0.51 Duration of Treatment 24 hr 70 days 4 weeks LOECa (HM) 5 Various 0.005 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Changes in cells that were arrested in mitosis by As111: c-Mos was hyperphosphorylated, cyclin A was degraded, cyclin B accumulated; ftftp34cdc2/cyclinB kinase activity. These and numerous other changes in mitotic proteins were similar to changes seen in cells arrested in mitosis by nocodazole, which is a known microtubule disassembly agent. Changes in expression of cell-cycle related genes: U at 7.7 for Cyclin Dl; for PCNA: flat 0.008, U at 0.77 and 7.7. Changes in expression of DNA repair genes: U at 0.77 and higher for ERCC6andOGGl; Uat7.7forXPC,MYH, and DNA polymerase-p. Changes in expression of other genes: U at 7.7 for, MnSoD, and Bax; for DNMT1: flat 0.008, NSE at 0.77, U at 7.7. ft Aurora-A protein expression level, with a positive dose-response, reaching 4.2x control at dose of 0.1; unreported data showed ft Aurora-A mRNA. Reference Huang and Lee, 1998 DuMond and Singh, 2007 Tseng et al., 2006 C-74 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue BEAS-2B cells Arsenic Species As111 AC Concentration(s) Tested (nM) 1.25, 2.5, 5, 10, 20 Duration of Treatment 12 hr LOECa (HM) 1.25 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft GADD45a protein expression level, with a positive dose-response; however, only a marginal ft in GADD45a transcription; pretreatment with NAC completely blocked the ft ofGADD45a. After inorganic arsenic dose of 20 for 4-20 hr: transitory activation of Akt and transitory ft phosphorylation of FoxOSa. Inorganic arsenic induced accumulation of GADD45a mRNA and did not affect the degradation of GADD45a protein. Inorganic arsenic stabilized GADD45a mRNA through nucleolin; it induced the binding of mRNA stabilizing proteins, nucleolin and less potently, HuR, to GADD45a mRNA. Inorganic arsenic did not affect the expression of nucleolin; inorganic arsenic treatment resulted in redistribution of nucleolin from nucleoli to nucleoplasm. Silencing of nucleolin reversed inorganic arsenic-induced stabilization of the GADD45a mRNA. Reference Zhang et al., 2006 C-75 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference Gclm+/+ MEF cells and Gclrn7- MEF cells, from GCLM knockout mice As111 SA for all See rows under Apoptosis and Cytotoxicity for this citation for experimental conditions. Analysis of global gene expression profiles revealed up-regulation or down-regulation of vast numbers of genes by inorganic arsenic. Significant changes were largely consistent with changes in the expression of DNA damage and repair genes, the suppression of TGF-(3 signals, inhibition of integrin-mediated cell adhesion, induction of multiple transcription factors, repression of co-repressors, and the derailment of cell cycle regulatory functions. Inorganic arsenic exposure also caused profound changes in protein levels in what appear to be conflicting regulatory changes. These changes go hand in hand with massive up-regulation of HSPs, metalloproteinases, and proteasome components, and the authors suggested that inorganic arsenic induces critical changes in protein folding and structure and that the cells mount a major effort to properly refold misfolded proteins or to eliminate them altogether. Global gene expression profiles also indicated that tBHQ is significantly effective in reversing inorganic arsenic-induced gene deregulation in Gclm+/+ but not in Gclm"7" MEFs. These results suggested that regulation of GSH levels by GCLM determines the sensitivity to inorganic arsenic-induced apoptosis and cytotoxicity by setting the overall ability of the cells to mount an effective antioxidant response. Kannet al., 2005b NB4 cells NB4-M-AsR2 cells As111 ATO for both 0.5, 1 2,4 16hrs for both 0.5 JNK activation leading to phosphorylation of c- jun, after treatment with ATO alone and co- treatment with 100 uM Trolox: At 0.5: slight ft alone, ft with Trolox. At 1: big ft alone, huge ft with Trolox. At 2: slight ft alone, ft with Trolox. At 4: big ft alone, huge ft with Trolox. Diazetal., 2005 JB6C141 PG13 cells JB6C141 PG13 cells exposed to 4 kJ/m2ofUVB at end of inorganic arsenic treatment As111 SA for both 1, 5, 10, 20 for both 24hrs for both in p53 activity with dose, reaching -30% of control at dose of 20. U in p53 activity with dose, reaching -5% of that with the UVB treatment alone at dose of 20. The UVB exposure strongly stimulated p53 activation (to ~9x the control level), and the inorganic arsenic treatment inhibited that increase, reducing it to a point estimate less than that of the untreated control at the dose of 20. Tang et al.. 2006 C-76 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue JB6C141 P+l-l cells JB6C141 P+l-l cells exposed to 4 kJ/m2ofUVB at end of inorganic arsenic treatment JB6C141 cells exposed to 4 kJ/m2 of UVB at end of inorganic arsenic treatment Arsenic Species As111 SA for both As111 SA for both Concentration(s) Tested (nM) 1, 5, 10, 20 0.1, 1,5, 10 5, 10 1,5,10 Duration of Treatment 24hrs for both 24hrs for both LOECa (HM) 5 5 5 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft in AP-1 activity to 2x control at 5 and to 5x control at 10, back to control level at 20. ft in AP-1 activity to 1.5x and 1.7x that with the UVB treatment alone at doses of 5 and 10, respectively. It should be noted that the UVB exposure strongly stimulated AP-1 activation (to ~6x the control level). llUVB-inducedp53 phosphorylation (at serines 15 and 392); bigger U at 10. llUVB-inducedp53 DNA binding activity; bigger U at 10. Other experiments not involving UVB showed that inorganic arsenic inhibited casein kinase 2a activity and decreased p5 3 -regulated p21 protein expression. Reference Tang et al., 2006 Tang et al., 2006 C-77 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SV-HUC-1 cells SVEC4-10 cells Arsenic Species As111 SA MMAm DMA111 As111 SA Concentration(s) Tested (nM) 0.5 0.05,0.1,0.2 0.2, 0.5 5, 10, 20 Duration of Treatment Subcultured twice weekly for 25 passages 24 hr LOECa (HM) 0.5 0.05 0.2 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Results of cDNA microarray analysis of -2000 genes: 114 genes were differentially expressed among the 6 groups; DMA111 had a substantially different gene profile from other 2. Gene coding for IL-1 receptor, type II, was the only gene with ft expression by all arsenicals. 11 genes had U expression by all arsenicals. For 2 of those 11, transcription was partially restored by treatment with 5-aza-dC, which suggests that the suppression resulted from epigenetic DNA hypermethylation. The treatments also caused differential morphological changes affecting cell size, extent of aggregation, and adhesion ability. Protein levels: a7-nAChR: slight U at 5, huge U at 10 and 20, with only a trace present at 20. eNOS: slight U at 5, huge U at 10 and 20, with none present at 20. ChAT: NSE. Reference Suetal., 2006 Hsu et al., 2005 C-78 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue BEAS-2B cells BEAS-2B cells BEAS-2B cells HT1 197 cells Arsenic Species As111 ATO As111 ATO As111 ATO As111 SA Concentration(s) Tested (nM) 10, 20, 50 10, 20, 50 10, 20, 50 10 Duration of Treatment 12hr 6hr 6hr 8hr LOECa (HM) 10 10 for all 10 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) HSP70 protein ft: fold increases over control by Western blotting after 12-hr recovery period: 2.6x, 2.5x, and 7.9x at doses of 10, 20, and 50, respectively; alternative ELISA analysis gave similar response but with much higher-fold increases over the control. Co-treatments with large doses of antioxidants CAT, SOD, NAC, or SF considerably reduced the arsenic effect, with the NAC treatment completely eliminating it. mRNA levels determined by RT-PCR, with no recovery time after exposure, fold ft over control: At 10: HSP70A, 4.4x; HSP70B, 4.3x; HSP70C, 3.6x. After 4-, 8-, and 12-hr recovery periods, mRNA levels usually U to levels closer to control and often NSE; however, all increases remained significantly higher than control at dose of 50. Intracellular GSH levels: U to 80% of control at 10, followed by dose- related decrease to 70% of control at dose of 50; co-treatment with NAC blocked this effect of inorganic arsenic. p53 protein levels: slight ft; at 24 hr at this dose: big ft to 4x control. p21 protein levels: ft to 7.5x control; also at this dose: at 12-20 hr, much smaller increases; at 24 hr, big U; at 4 hr, 2.4x control. Reference Han et al., 2005 Han et al., 2005 Hanetal., 2005 Hernandez -Zavala etal., 2005 C-79 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SVEC4-10 cells RAW264.7 cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 4, 8, 12, 16; separase was tested only at the highest dose 2.5,5 Duration of Treatment 24 hr 24 hr LOECa (HM) Various 2.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Effects on protein levels: Securin: U at 12 to 23%, U at 16 to 5%. Separase: ft to 1.2x control (of ?-able significance). Phospho-CDC2 (threonine-161): U at 16 to 34%. CDC2: U at 12 to 73%, U at!6to38%;cyclinBl: U at 16 to 11%. p53(DO-l):ftat4to2x control with positive dose-response reaching 8x control at dose of 16. TRAP histochemistry was done 3 days after the end of the inorganic arsenic treatment: huge ft in TRAP activity at both doses; this increased activity accompanied multinucleated cell formation and the beginning of osteoclast differentiation; the level of effect at both doses was comparable to (and, at the dose of 2.5, probably higher than) that caused by a RANKL treatment; co-treatment with CAT blocked most of the inorganic arsenic- induced effect. Reference Chao etal., 2006a Szymczyk etal., 2006 C-80 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HCT1 16 cells (securin +/+) HCT1 16 cells (securin -/-) RKO cells (p53 wt) SW480 cells (p53 mutant) FGC4 cells HepG2 cells Arsenic Species As111 SA for both As111 SA for both As111 SA As111 SA Concentration(s) Tested (nM) 4, 8, 12, 16 for both 8, 16 for both 50,65 Equivalent to <5% and 20-25% cytotoxicity 15,55 Equivalent to <5% and 20-25% cytotoxicity Duration of Treatment 24 hr for both 24 hr for both 24 hr 24 hr LOECa (HM) Various Various 16 16 Various Various Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Effects on protein levels: Securin: U at 4, then U with dose to -30% at 16. Phospho-p53 (serine 15): ft to 2x control at 4 and then ft with dose to 6x control at 16. p53 (DO-1): ft to 2x control at 12 and ft to 3. 4x control at 16. No securin present at any dose in -/- mutant. Phospho-p53 (serine 15): ft to 3.5x control at 4 and then ft with dose to 7x control at 16. p53 (DO-1): ft to 1.8x control at 4 and then ft with dose to 3.2x control at 16. Effects on protein levels of securin: rather similar U in both, reaching 27% and 13% of control in RKO and SW480, respectively. Effects on protein levels of SPs: MT, HSP60andHSP90: NSE at either dose. HSP25: big ft at 50, big ft at 65. HSP40: big ft at 50, big ft at 65. HSP70: big ft at 50, huge ft at 65. Effects on protein levels of SPs: MT: NSE at 15, very slight ft at 55. HSP60andHSP90:NSE at either dose. HSP27: slight ft at 15, ft at 55. HSP40: slight ft at 15, big ft at 55. HSP70: ft at 15, big ft at 55. Reference Chao etal., 2006a Chao etal., 2006a Gottschalg etal., 2006 Gottschalg et al., 2006 C-81 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Rat hepatocytes HELP cells MDAH 2774 cells UROtsa cells UROtsa cells Arsenic Species As111 SA As111 SA As111 ATO As111 SA MMAm Concentration(s) Tested (nM) 10,20 Equivalent to <5% and 20-25% cytotoxicity 0.1,0.5, 1,5, 10 1, 2, 5, 8 0.5, 5, 10, 25 0.05 Duration of Treatment 24 hr 3, 6, 12, 24, or48hr Probably 72hror 96 hr 24 hr 12 weeks LOECa (HM) Various Various lor 2 5 0.05 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Effects on protein levels of SPs: MT, HSP60andHSP90: NSE at either dose. HSP25: ft at 10, ft at 20. HSP40: NSE at 10, ft at 20. HSP70: NSE at 10, big ft at 20. HSP27 protein: ft at 0.5 and 1 after 12-hr treatment, but U at 5 and 10 after 48-hr treatment; HSP27 was said to be a chaperone whose expression protects against oxidative stress and is anti-apoptotic. HSP70 protein: U at 1 and 5 after 12-hr treatment, but ft at 5 and 10 after 24-hr treatment; an inducible form of HSP70 was said to be expressed at a high level in various malignant human tumors. U topoisomerase Ila to about half of control value at dose of 5 (paralleling degree of cytotoxicity) — there is some question about this result because band densities were not normalized to another protein; decrease possibly resulted from U in cell number. ft accumulation of high- molecular-weight Ub- conjugated proteins. Co-treatment with BSO: ftft in the same effect, which was then seen even at dose of 0.5. Huge ft COX-2 protein, with an even higher level after 24 weeks and still high level after 52 weeks. Reference Gottschalg et al., 2006 Yanget al., 2007 Askar et al., 2006 Bredfeldt et al., 2004 Eblin et al., 2007 C-82 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells UROtsa cells Arsenic Species As111 SA MMA111 As111 SA MMA111 Concentration(s) Tested (nM) 1, 10 0.01,0.05,0.1 1, 10 0.05,0.5,5 Duration of Treatment 4hr for both 30 min for both LOECa (HM) 1 0.01 1 0.05 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Big ft COX-2 protein level at both doses. Regarding COX-2 protein level: huge ft at 0.01, big ft over control at 0.05, ft over control at 0.1. Various experiments, including some with pharmacological inhibitors of various signal transduction pathways, led to the conclusion that MMA111 appears to stimulate ligand-independent activation of EGFR, subsequent ERK-1 and - 2 phosphorylation via MEK-1 and -2, as well as activation of PIS K, which leads to elevated levels of COX-2 protein. ft HSP70 protein (similar response at both doses; with lower dose, the level decreases from 60 to 240 min); ft MT protein (much bigger ft at higher dose). ft HSP70 protein (strong response at all doses). ft MT protein (much bigger ft at higher doses). Reference Eblin et al., 2007 Eblin et al., 2006 C-83 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells MEF cells MEF cells Arsenic Species As111 SA Asv MMAm DMA111 As111 SA As111 SA Concentration(s) Tested (nM) 2, 6, 10 1,5, 10 1,2,3 1,4,7 0.01,0.1,5, 10, 20,40 20 in most assays Duration of Treatment 24 hr for all 5hr LOECa (HM) 6,2 None 20 ,3 4 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Extent of selenium incorporation into selenoproteins determined using 75Se-selenite: LOECsof6and2forli TrxRl and U cGpx, respectively; big U at higher dose(s). NSE. LOECsof2and3forft TrxRl and U cGpx, respectively. ft of TrxRl and cGpx at dose of 4 and decrease for both proteins to near control levels at higher dose. ft eIF2a phosphorylation; ft ATF4 protein; ft ATF3 protein. At doses >10: ft GADD45a protein and ft CHOP protein. All effects showed substantial dose-related increases. Effects were mostly blocked by NAC pretreatment. (ATF3 was not tested.) GADD45a is a small protein implicated in the regulation of the cell cycle, DNA repair, genome stability, innate immunity, and apoptosis. Additional tests with modulators and genetic variants of MEF cells showed the following: ATF4 is required for an increase in GADD45a mRNA following inorganic arsenic exposure, and its induction is independent of p53. ATF4 binds to a GADD45a promoter element in response to inorganic arsenic stress. Exposure to inorganic arsenic reduces proteasome activity, which permits the increase in transcription of GADD45a to actually result in an increase in the protein level of GADD45a, which is labile. Reference Ganyc et al., 2007 Jiang et al., 2007 Jiang etal., 2007 C-84 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Protein extracts (membrane fraction) derived from BAEC cells N-18 cells N-18 cells Arsenic Species MMAm As111 SA, Asv, MMAvor DMAV As111 SA Asv potas- sium arsenate Concentration(s) Tested (nM) 1,2.5,5,7.5, 10, 15 10 5, 10, 20, 50 20 Duration of Treatment 5 min for all 6hr 6hr LOECa (HM) 1 None 5 for first effect noted 20 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) For MMAm only: U eNOS activity, IC50 = 2.1 and a 5-min treatment at dose of 10 caused -90% U; co-treatment with DTT substantially blocked the MMAm effect, resulting in only -50% U. ft synthesis of HSP proteins of 50, 73, 78, 89, 98, and 104 kDa. Other experiments demonstrated: ft activation of HSF1 DNA-binding (detected by EMSA) by dose of 20 (lowest dose tested) in 2 hr; ft induction of HSP70-luciferase reporter gene expression by dose of 20 (lowest dose tested) in 6 hr; an ft induction of HSP70 rnRNA by dose of 50 (lowest dose tested) in 1 hr. ft induction of HSP70- luciferase reporter gene expression (point estimates suggests weaker response from Asv than from same dose ofAsmSA). Reference Sumi et al., 2005 Khalil et al., 2006 Khalil et al., 2006 C-85 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue N-18 cells hsf^ immortalized MEF cells hsf7' immortalized MEF cells hsf7' immortalized MEF cells transfected withHSFl expression vector Arsenic Species As111 As111 SA for all Concentration(s) Tested (nM) 2, 5, 10, 20, 50, 100, 200, 500 5, 10, 20, 50, 100, 200, 500 for all Duration of Treatment 0.5, 1,2,3,6, or!2hr Ihr for all LOECa (HM) Various 50 None 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft induction of HSP70- luciferase reporter gene expression, with bell- shaped dose-response curves for each duration of treatment; e.g. for 1-hr treatment, the peak occurred at dose of 200 (highest peak seen); for 6-hr treatment, the peak occurred at dose of 20; the bell-shaped curves shifted to the left as the duration increased. Results on HSP70-firefly luciferase activity were normalized against that of Renilla luciferase to correct for differences in transfection efficiency and/or toxic and non- specific effects of the experimental treatment conditions. ft induction of HSP70- luciferase reporter gene expression: ft with dose up to peak at 200; still big ft at 500. No effect; clearly inorganic arsenic requires a functional HSF1 gene to induce HSP70-luciferase reporter gene expression. ft with dose up to peak at 200; still big ft at 500. Generally similar results were also found with treatment durations of 0.5 and 2 hr. Reference Khalil et al., 2006 Khalil et al., 2006 C-86 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue H1355 cells H1355 cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 5, 25, 50, 100, 200 100 Duration of Treatment 24 hr 24 hr LOECa (HM) Various Various Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Phosphorylation of ERK 1/2: ft at 50, huge ft at 100 and 200. Phosphorylation of INK: slight ft at 50, huge ft at 100 and 200. Phosphorylation of p38: slight ft at 100, big ft at 200. PARP cleavage: ft at 100 and 200. Survivin protein level: U at 100 and 200. Ubiquitination in total cell lysate: big ft at 100 (the only dose tested for it). Effects of pretreatments with specific inhibitors ofp38, JNK, MEK1/2 (upstream of ERK 1/2) or ubiquitin-proteasome showed that blockage of either p3 8 or JNK phosphorylation attenuated the ATO- induced down-regulation of survivin and increase of PARP cleavage; however, blockage of ERK 1/2 or ubiquitin- proteasome did not attenuate those same effects. Also, only inhibitors of p38 and JNK affected ATO- induced cytotoxicity, which was just slightly reduced (i.e., there was ~5%-8% more cell survival). The specific inhibitors of p38, JNK, and MEK 1/2 did block the phosphorylations of p3 8, JNK, and ERK 1/2, respectively. Reference Cheng et al., 2006 Cheng et al., 2006 C-87 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue A549 cells A549 cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 2 2 Duration of Treatment 48 hr 48 hr Results (Compared With Controls, With All Concentrations LOECa Being (|oM) in |iM Unless Noted) Protein levels and mRNA levels: 2 uM inorganic arsenic: NSE on survivin. 200 uM sulindac: NSE on survivin. (Sulindac is a NSAID that inhibits COX-2.) (2 uM inorganic arsenic + 200 uM sulindac): big U in survivin (by 72 hr almost no survivin was protein present). Protein levels only for combined treatment: big ft for p53 but NSE for XIAP, cIAP-1, cIAP-2, andBcl-2. Inhibition of p53 ft by siRNA blocked the down- regulation of survivin by the (2 uM inorganic arsenic + 200 uM sulindac) treatment. (It is known that p53 binds to the survivin promoter and suppresses its transcription.) Transfected cells with a survivin- luciferase reporter also showed the big U in survivin for the combined treatment and NSE for single treatments. Pretreatment with NAC mostly (or entirely) blocked the synergistic effect of a U of survivin protein (was shown both by Western blot and luciferase reporter assays). More about the synergistic effect between 2 uM inorganic arsenic and 200 uM sulindac: evidence that changes in survivin levels are related to synergistic big ft in cytotoxicity: (1) if marked overexpression of survivin by transfection, then U in cytotoxicity by 1/3, (2) if inhibition of survivin level by siRNA, then ft in cytotoxicity. (Sulindac is a NSAID that inhibits COX-2.) Reference Jinetal., 2006b Jinetal., 2006b C-88 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue N-18 cells NHEK cells NHEK cells NHEK cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 2, 5, 10, 20, 50, 100, 200, 500 0.1, 1,5, 10 0.1, 1,5, 10 1 Duration of Treatment 6hr 72 hr 7 days 24, 48, 72 hr LOECa (HM) 10 0.1 for ft Iforli 0.1 for ft Iforli 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of HSP70- luciferase reporter gene expression: big ft at 10, huge ft (peak) at 20, big ft at 50, then NSE. Effects of pretreatment + co-treatment with modulators: DTT: almost entirely blocked inorganic arsenic effect; slight ft at 20 and 50, questionable ft at 10 and 100. NAC and GSH (individually): ft at 10, big ft at 20, huge ft (peak) at 50, ft at 100, then NSE. Level of Pi-integrin protein: after a possible slight ft at 0.1, there was all to 61-63% of control level at other 3 doses. Level of Pi-integrin mRNA: after a possible slight ft at 0. 1, a dose- related 11 at other 3 doses reaching 47% of control at dose of 10. Level of FAK protein based on immunofluorescence: ft at 24 hr followed by U below control level at later times, with almost none present at 72 hr. Reference Khalil et al., 2006 Lee et al., 2006b Lee et al., 2006b Lee et al., 2006b C-89 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Normal human mammary epidermal keratinocytes Swiss 3T3 mouse cells UROtsa cells Arsenic Species As111 SA for all As111 SA As111 SA MMAV DMAV Concentration(s) Tested (nM) 0.005,0.5, 1,2.5 1,2.5,5 1, 2.5, 5, 10, 20, 40 5, 50 for all Duration of Treatment 4hr 8hr 16 hr 2 hr for all LOECa (HM) 0.005 2.5 1 5 for all Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft COX-2 mRNA (also at 8 and 24 hr). ft COX- protein (also at 12 hr), also under the same or similar conditions: ft PGE2 secretion, phosphorylation of p42/44 MAPK, and DNA synthesis. Tests with various modulators showed that inorganic arsenic111 elevates COX-2 at the transcriptional and translational levels. ft GSH synthesis; starting at 2. 5: cell retraction and loss of thick cables of actin filaments, U cytoskeletal protein synthesis; starting at 20: ft in protein sulfhydryl content of both cytoskeletal and cytosolic protein fractions, with the time course showing a slight decrease before the increase. There was also severe loss of microtubules. Increased DNA binding of the AP-1 transcription factor, which is often associated with the regulation of genes involved in cell proliferation. For all 3 chemicals the response was higher at dose of 50; the highest amount of binding was with SA. Reference Trouba and Germolec, 2004 Li and Chou, 1992 Simeonova etal.,2000 C-90 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells C-33A cells HeLa cells Jurkat cells LCL-EBV cells HeLa cells Arsenic Species As111 SA As111 SA for all As111 SA Concentration(s) Tested (nM) 10,50 1, 10, 25, 50 for all 100, 200, 400 Duration of Treatment 2hr 24 hr for all 30 min LOECa (HM) 10 None 10 1 1 100 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Use of a cDNA array consisting of 588 human genes, and other methods: At 10: ft activity of 7 genes; U activity in 6 genes. At 50: ft activity of 15 genes; U activity in 6 genes. Specifics: Genes affecting cell growth: ft for c-fos, c- jun, Pig 7, EGR-1, and Rho8. Genes affecting cell growth arrest: ft for GADD45 and GADD153. p53 protein expression: No ft, slight U at high doses, very high basal level. ft, peak at 25, low basal level. ft, peak at 10, moderate basal level. ft, peak at 10, very low basal level. Decreases above peak may result from cell death. ft GADD153 mRNA expression (harvested for RNA isolation after 4 hours of incubation following the arsenite treatment). This effect was increased by pretreatment with BSO, PHEN (slight increase), BCS, or mannitol (an HO" scavenger). Effect was completely blocked by pretreatment with NAC. Reference Simeonova et al., 2000 Salazar et al., 1997 Guyton et al., 1996 C-91 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue WI38 cells Simian vims 40 (SV40)- transformed subline of the above parental W138 line with twice the GPx specific activity of parental cells JB6 C141 cells JB6 C141 cells HFW cells (diploid human fibroblasts) HFW cells (diploid human fibroblasts) HFW cells (diploid human fibroblasts) Both HL-60 cells and HaCaT cells Arsenic Species As111 SA for both As111 SA Asv As111 SA Asv As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (jiM) 100, 200, 400 for both 3.125, 12.5,50, 200 3.125, 12.5,50, 200 200 200 5, 10, 20 1,2.5,5, 10,20 0.5,2, 10 0.5,20 Duration of Treatment 30 min 3hr 0 min 60 min 24hr 24 hr 24 hr 3 days LOECa (HM) 100 for both 50 50 200 200 5 1 See next column 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft GADD153 mRNA expression (harvested for RNA isolation after 4 hours of incubation following the arsenite treatment). The increase was cut approximately in half (i.e., half the slope) in the transformed cell line. Other parts of this study showed that AP-1 is critical to oxidative regulation of GADD153. ft activity of JNKs: stronger response at 50 for Asv (sodium arsenate); both forms shown some response by Ihr at dose of 200; arsenic did not induce p5 3 -dependent transactivation. ft phosphorylation of JNKs: stronger response for Asv (sodium arsenate). ft heme oxygenase activity (arsenic -induced synthesis of this enzyme was blocked by co- treatment with antioxidants sodium azideorDMSO);ft ferritin. ft GSH (by 20 level drops to control level). ft SOD activities, U catalase and GPx activities, with LOECs being 0.5, 2, and 10, respectively. ft hTERT protein expression; however U hTERT protein expression at 20 (i.e., significantly inhibited at higher concentration). Reference Guyton et al., 1996 Huang et al., 1999b Huang et al., 1999b Lee and Ho, 1995 Lee and Ho, 1995 Lee and Ho, 1995 Zhang et al., 2003 C-92 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells HL-60 cells NB4 cells NB4 cells NB4 cells HeLa cells LoVo cells MCF7 cells Arsenic Species As111 SA As111 SA As111 ATO As111 ATO As111 ATO As111 ATO for all Concentration(s) Tested (nM) 0.5, 10, 20 0.1,0.5, 1, 10,20 0.75 0.75 0.1,0.25 2 for all Duration of Treatment 3 days 3 days 8 days 2 days 12 days 14 days for all LOECa (HM) 0.5 0.1 0.75 0.75 0.1 2 for all Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft telomerase activity; however, telomerase activity was below control level at 10 and even lower at 20. ft telomerase activity; however, telomerase activity was below control level at 10 and even lower at 20. U telomerase activity; U hTERT mRNA and protein levels; U c-myc mRNA and protein levels; ft hTER mRNA level; no change in p53 mRNA or protein level; no change in Spl mRNA or protein levels. Further experiments showed that arsenic inhibits transcription of hTERT and inhibits the function of Spl in hTERT transcription. U hTERT mRNA. U hTERT mRNA. U hTERT mRNA and U c-myc mRNA for all. Reference Zhang et al., 2003 Zhang et al., 2003 Chou et al., 2001 Chou et al., 2001 Chou et al., 2001 Chou et al., 2001 C-93 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Normal human keratinocytes treated with 50 mJ/cm2 UVB before or after inorganic arsenic treatment NB4 cells SHE cells Arsenic Species As111 SA for both As111 ATO for both As111 SA Asv Concentration(s) Tested (nM) 1, as pretreatment 1, as post- treatment begun 24 hr after irradiation 1 0.5, 1.0, 1.5,2.0 6,8 50, 100, 150 Duration of Treatment 24 hr for both 2 days 3 days 48 hr for both LOECa (HM) 1 None 1 1.0 — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) No change from control in procaspase-8 and procaspase-9 protein levels or in caspase-3, caspase-8, and caspase-9 enzyme activities; this is considered an LOEC because the inorganic arsenic-pretreatment blocked the effects of UVB described below. U procaspase-8 protein level, slight U procaspase-9 protein level; ft caspase-8 enzyme activity; ft caspase-9 enzyme activity; ft caspase-9 enzyme activity; effects similar to with UVB alone. As a result of permeability changes in the outer mitochondria! membrane: slight release of cytochrome c into cytoplasm; complete release by 3 days of treatment. ft Cpp32 (was activated) as shown by U of its precursor. From among these treatment groups, 5 neoplastic transformed cell lines were produced that were shown to be tumorigenic. Of these: all had ft c-Ha-ras (oncogene) mRNA expression; 4 had ft c-myc (oncogene) mRNA expression; a few other arsenic- treated cell lines also showed the same effects. Reference Chenetal., 2005b Jing et al., 1999 Takahashi etal.,2002 C-94 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Peritoneal macrophages (PMs) from CDFi mice U118MG cells HaCaT cells (immortalized non- tumorigenic human keratinocyte cell line) arsenic-TL cells (arsenic- tolerant cells, which are HaCaT cells that were cultured for 28 weeks in 100 nM As111 SA) Arsenic Species As111 SA Asv MMAV DMAV TMAV As111 ATO As111 SA for both Concentration(s) Tested (nM) 1.25,2.5,5, 10 125, 250, 500, 1000 1.25,2.5,5, 10 mM 1.25,2.5,5, 10 mM 1.25,2.5,5, 10 mM 1,5,10,25 20 for both Duration of Treatment 48 hr for all 24 hr 6 hr for both LOECa (HM) 1.25 500 None 2.5 mM 5 mM 1 or 5 20 for both Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Changes in release of TNF-a from macrophages in the presence of both lipopolysaccharide and recombinant murine interferon y, which are two compounds known to increase secretory functions of PMs: U at 1.25, no change from control at 5; big ft at 10. big ft at 500 and much bigger ft at 1000. no effect. U at 2.5, 5 and 10 mM. U at 5 and 10 mM. Changes in protein expression: p53:ftatl, U at 5 or higher; Bcl-2: ft at 1 or higher. Bax: U at 1 or higher; HSP70: ft at 5 or higher. Co-treatment with lipoic acid blocked all of these effects at an inorganic arsenic111 dose of 5. ft caspase-3 activation. Much smaller ft in caspase-3 activation than in HaCaT cells. Reference Sakurai et al., 1998 Cheng et al., 2007 Pi et al., 2005 C-95 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference HUVEC cells As111 ATO 20 2hr 20 ft expression of 1C AM-1; effect was similarly strong after 24-hr treatment but weaker after 4- or 8-hr treatment (yet still ft above control level). Effect was completely blocked by a 1-hr pretreatment with 15 mMNAC followed by a co-treatment of NAC with the Asin-treatment. Griffin et al., 2003 Apoptosis K562 cells Asm ATO 2.5 12 hr 2.5 ft annexin V, an apoptotic marker. Li and Broome, 1999 NCI (human myeloma cell line) As111 ATO 24 hr Apoptosis was demonstrated by 4,6- diamidino-2- phenylindole staining, by the demonstration of typical DNA ladders corresponding to internucleosomal cleavage, and by annexin-V and PI staining. Various indications of induction of apoptosis were also presented (with less detail) for at least 1 other myeloma cell line and for fresh myeloma cells. In the NCI cells, [3H]thymidine incorporation was also used to assess proliferation: the 50% growth-inhibitory concentration (IC50) in NCI cells was found to be 0.3 uM, based on concentrations tested of 0.05,0.1,0.5, 1,5, 10 over 72 hr. Similar testing of 3 other human myeloma cell lines yielded IC50s of 0.1 for 1 line and ~1 for 2 other lines, with much less detail presented. Rousselot etal., 1999 C-96 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MGC-803 cells Primary cultures of rat cerebellar neurons MC/CAR (human multiple myeloma cell line) V79-C13 Chinese hamster cell line HL-60 cells HaCaT cells Arsenic Species As111 ATO As111 SA DMAV As111 ATO As111 SA As111 SA for both Concentration(s) Tested (nM) 0.01-1 5, 10 5 mM 1,2,5,10 10 0.1,0.5, 1, 10,20, 40 for both Duration of Treatment 24 hr 12 hr 48 hr 72 hr 24 hr 5 days for both LOECa (HM) 0.01 5 5 mM 2 10 lor possibly 0.5 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis detected by flow cytometry and by agarose gel electrophoresis of genomic DNA showing typical DNA ladder; at various doses apoptosis was also induced in 5 other human malignant cell lines. Demonstrated by "DNA ladders" with agarose gel electrophoresis and microscopic examination (nuclear fragmentation and/or condensation). Apoptosis was demonstrated by an analysis using a FACStar flow cytometer and by detection of cell membrane changes by labeling with annexin V- FITC and annexin PI. Apoptotic cells appeared by 6 hr after treatment began and included 40% of cells by 24 hr; frequency gradually decreased during 48 hr of observation after treatment ended. ByuseofHoechst/PI staining assay: ft in apoptosis for both; for both cell lines, there was the same general response, but to a lesser extent, when same treatments were given over 1 or 3 days. Reference Zhang et al., 1999 Namgung and Xia, 2001 Parketal., 2000 Sciandrello etal.,2002 Zhang et al., 2003 C-97 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HL-60 cells HaCaT cells SW13 cells SW480 cells HT1080 cells TRL 1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with SOuMBSO TRL 1215 cells Arsenic Species As111 SA for all MMAV for both DMAV Concentration(s) Tested (nM) 1, 10, 20, 40 for all 5 mM for both 5mM Duration of Treatment 5 days for all 24 hr for both 24 hr LOECa (HM) 1 10 -20 -20 1 None 5 mM 5mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ByuseofHoechst/PI staining assay: ft in apoptosis in all. SW13 and SW480 are telomerase negative cell lines, and they showed much less apoptosis at all concentrations than the other 3 cell lines. HT1080 is a telomerase positive cell line, and it was intermediate in the amount of apoptosis at all concentrations to HL- 60 (which was higher) andHaCaT. Thus there is a strong positive correlation between telomerase activity and susceptibility to arsenic- induced apoptosis. Apoptosis demonstrated by TUNEL staining: there was little evidence of induction of apoptosis by MMAV alone; however, the cells also treated with BSO showed considerable apoptosis. Apoptosis demonstrated by TUNEL staining: huge ft, much more extensive that that of the considerable level of apoptosis reported in row above for MMAV + BSO. Reference Zhang et al., 2003 Sakurai et al., 2005a Sakurai et al., 2005a C-98 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with 50uMBSO TRL 1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with SOuMBSO Arsenic Species MMAV for both MMAV for both Concentration(s) Tested (nM) 5mM for both 5mM for both Duration of Treatment 12, 24, 36, or 48 hr for both 24 hr for both LOECa (HM) 5mM 5 mM None 5 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis demonstrated by FACS analysis after annexin-V and PI staining: 5 mM MMAV alone caused some apoptosis after 48 hr; however, that response was slight compared to the response oftheMMAv + BSO group after only 24 hr, andtheMMAv + BSO group showed huge ft at 36 hr and even bigger ft at48hr. After 48 hr, the percentages of annexin- positive cells were as follows: control, 1.9%, BSO alone, 6.7%; MMAV alone, 10.6%; MMAV + BSO, 64%. The PI staining showed that by 48 hr there were also numerous induced necrotic cells in the MMAV + BSO group. Apoptosis demonstrated by agarose gel electrophoresis showing induced internucleosomal DNA fragmentation: substantial DNA fragmentation in MMAV + BSO group; no effect with MMAV alone. Reference Sakurai et al., 2005a Sakurai et al., 2005a C-99 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with 50uMBSO TRL 1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with 50uMBSO Primary keratinocytes (in third passage) obtained from foreskins of adults Arsenic Species DMAV for both MMAV for both As111 SA Concentration(s) Tested (nM) 5mM for both 5mM for both 1,5, 10 Duration of Treatment 24 hr for both 12 hr for both 48 hr LOECa (HM) 5mM 5mM None 5mM 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis demonstrated by agarose gel electrophoresis showing induced internucleosomal DNA fragmentation: massive ft with DMAV alone (many times more than with MMAv + BSOin previous row); slight ft in DMAV + BSD group (about the same as with MMAv + BSOin previous row). Cellular caspase-3 activation: ft to ~1.6x in MMAV + BSO group; no effect without BSO; other experiments showed that co-treatment with 150 uM Z-DEVD-FMK (a caspase 3 inhibitor) during preincubation period and during a 24-hr MMAV treatment blocked almost all or all of the cytotoxicity detected by AB assay (i.e., -35% survival without inhibitor, -92% survival with inhibitor); with a 48-hr MMAV + BSO treatment, Z-DEVD- FMK caused cytotoxicity to be markedly reduced (i.e., -7% survival without inhibitor, -42% survival with inhibitor). Apoptosis detected by the presence of DNA ladders after agarose gel electrophoresis: much bigger ft at two higher doses, which showed a similar effect. Reference Sakurai et al., 2005a Sakurai et al., 2005a Liao etal., 2004 C-100 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Primary keratinocytes (in third passage) obtained from foreskins of adults HeLa cells Arsenic Species As111 SA As111 ATO Concentration(s) Tested (nM) 1,5,10 2 Duration of Treatment 48 hr 3 days LOECa (HM) Various 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Protein levels detected by Western blotting: F ADD : ft at 1, bigger ft at 5 and 10. Caspase-8 (p!8, active): ft at 1, huge ft at 5 and 10. Caspase-3 (p20, active): huge ft at 5 and 10. Cleaved PARP (85 kD): ft at 5 and 10; additional experiments with and without modulators confirmed the involvement of the Fas- associated pathway in inorganic arsenic- induced apoptosis. Induced apoptosis (experimental - control) detected by Annexin V/PI flow cytometry: -13% for inorganic arsenic alone; -3% for 10 uM emodin alone; -41% for inorganic arsenic plus 10 uM emodin; -14% for inorganic arsenic with both 10 uM emodin and l.SmMNAC. Other experiments showed that the effect of emodin in enhancing inorganic arsenic-induced apoptosis involved a decrease of mitochondria! membrane potential. Emodin was used because it has a semiquinone structure that is likely to increase the generation of intracellular ROS. Reference Liao etal., 2004 Yi et al., 2004 C-101 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HeLa cells AR230-S cells, AR230-r cells, KCL22-S cells, KCL22- r cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 2 1 Duration of Treatment 3 days 24 hr LOECa (HM) 2 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (experimental - control) detected by Annexin V- FITC/PI flow cytometry: 27.0% for inorganic arsenic alone; 6.9% for 30 uM emodin alone; 44. 1% for inorganic arsenic plus 30 uM emodin; 20.4% for inorganic arsenic with both 30 uM emodin and l.SmMNAC. Emodin and inorganic arsenic synergistically interacted to greatly ft the ROS level and to cause cytotoxicity. Pretreatment or co- treatment with NAC blocked the synergism for both effects. A 2uM inorganic arsenic treatment of 90 min caused an ft in ROS to ~2.0x (with wide confidence limits) and, in a treatment lasting 48 hr, about 20% cytotoxicity. Apoptosis detected by Annexin V-FLUOS staining kit and flow cytometry: NSE in any of the 4 cell lines with ATOorlOOuMBSO treatments alone. For the combined treatment, induced rates (experimental - control) were: AR230-S, -35%; AR230-r, -35%; KCL22-S, -10%; KCL22-r, -13%. Reference Wang et al., 2005 Konig et al., 2007 C-102 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue AR230-r cells, KCL22-r cells U-937 cells NB4 cells HL-60 cells Arsenic Species As111 ATO As111 ATO for all Concentration(s) Tested (nM) 1 1, 2, 4, 8 0.5, 1,2,4 1,2,4 Duration of Treatment 24 hr 24 hr for all LOECa (HM) None 4 1 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Western blot analyses: inorganic arsenic alone caused NSE on protein levels of tyrosine phosphorylated Bcr-Abl or total cellular Bcr-Abl in either cell line. In both cell lines, combined treatment of inorganic arsenic with 100 uM B SO yielded huge U in both proteins. In non- imatinib resistant CML cells, unlike in these 2 imatinib-resistant cell lines, inorganic arsenic alone had been shown to suppress Bcr-Abl activity. Induced apoptosis (experimental - control) based on chromatin fragmentation: U-937 cells: 1, NSE; 2, -2%; 4, -14%; 8, -85%. NB4 cells: 0.5, NSE; 1, -5%; 2, -33%; 4, -63%. HL-60 cells: 1, NSE; 2, -5%; 4, -22%. Induction of apoptosis was potentiated by co- treatment with PI3K inhibitors LY294002 and wortmannin, and by the Akt inhibitor Akt,5. Reference Konig et al., 2007 Ramos et al., 2005 C-103 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U-937 cells HK-2 cells Arsenic Species As111 ATO As111 SA Asv Concentration(s) Tested (nM) 4 0.1, 1, 10 for both Duration of Treatment Various 6, 24 hr LOECa (HM) 4 0.1 at 24 hr Prob- ably 1 at 24 hr Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) U in Akt phosphorylation after 24 hr (not by 14 hr); ft in caspase 3 activity to ~3x after 24 hr; ft in cytochrome c protein (released from mitochondria) after 14 hr; big ft in activated Bax after 14 hr; big ft in HSP 27 after 14 and 24 hr; big ft in HSP 70 after 14 and24hr. The potentiation of apoptosis by inhibitors mentioned in prior row involved more extreme changes in the same direction for p- Akt, caspase 3, cytochrome c, and Bax activation as well as attenuation of HSP27 expression. It also involved increased disruption of the mitochondria! transmembrane potential. To assess mitochondria! function, depolarization of mitochondria! membrane was detected using MitoTracker Red, a mitochondrion selective dye. Effect of dose of lof As111 appeared equivalent to that of dose of 10 of Asv Effect increased with dose and time. Reference Ramos et a!., 2005 Peraza et a!., 2006 C-104 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HK-2 cells APL primary cells K562 cells NB4 cells Thymocytes from adult male BALB/cByJ mice Arsenic Species As111 SA As111 ATO As111 ATO Asv Concentration(s) Tested (nM) 0.1, 1, 10,25 3 5 for both Duration of Treatment 24 hr 24 hr 3, 10, 22 hr for both LOECa (HM) 0.1 3 for all None None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (experimental - control) detected by Annexin V- FITC/PI flow cytometry: 0.1, -36%; 1, -23%; 10, -15%; 25, -15%. Induced necrotic cells (experimental - control) detected by same method: 0.1, -2.5%; 1, -3%; 10, -6%; 25, -24%. Apoptotic cells detected in this way were said to be in early apoptosis. Examination by transmission electron microscopy showed that most such cells failed to complete apoptosis and ultimately underwent necrosis instead. They suggested that inorganic arsenic was so toxic to mitochondria that they lost "their ability to keep the cell on course for apoptotic cell death." Apoptosis rates (control rates were not provided), detected by FITC- annexin V and PI double-staining: 52.2% 27.6% 56.6% NSE at any time point for induction of apoptosis by any of the following types of analysis: (1) "Annexin V-FITC positive" without loss of membrane impermeance (i.e., "7-AAD negative") to identify early apoptotic cells, (2) DNA loss, and (3) both "Annexin V-FITC positive" and "7-AAD positive" for cells in the final stages of cell death. Reference Peraza et al., 2006 Sahu and Jena, 2005 Mondal et al., 2005 C-105 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Jurkat cells Namalwa cells NB4 cells U937 cells Jurkat cells Namalwa cells NB4 cells U937 cells NB4 cells U937 cells Arsenic Species As111 ATO for all As111 ATO for all As111 ATO for both Concentration(s) Tested (nM) 1,2 for all 2 for all 1, 2, 4, 6 for both Duration of Treatment 24 hr for all 24 hr for all 24 hr for both LOECa (HM) None 2 1 None None 2 2 None 1 4 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (experimental - control) detected by fluorescence microscope analysis after staining with AO and EB: Namalwa cells: 1, -1%; 2, -16%. NB4 cells: 1, -12%; 2, -26%. NSE at dose of 2 in Jurkat and U937 cells. Pretreatment with NAC or Z-VAD-FMK blocked induction of apoptosis in Namalwa and NB4 cells. Western blot analysis: ft in PARP-cleavage and U in procaspase-3 level in both Namalwa and NB4 cells but not in the other two cell lines; inorganic arsenic did not induce JNK phosphorylation. Induced apoptosis (experimental - control) detected by fluorescence microscope analysis after staining with AO and EB: NB4 cells: 1, -6%; 2, -30%; 4, -70%; 6, 85%. U937 cells: 1, -0%; 2, -4%; 4, -15%; 6, 12%. NB4 cells showed more severe cell growth inhibition at doses of >2. Also, Western blot analysis showed that inorganic arsenic induced PARP cleavage in a dose-dependent pattern in NB4 cells. In U937 cells there was only very slight PARP cleavage at the highest dose. JNK phosphorylation did not occur in either cell line. Reference Chenetal., 2006 Chenetal., 2006 Chenetal., 2006 C-106 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MEFs that are wt MEFs that are DKOs for Bax and Bak MEFs that are wt or DKOs for Bax and Bak Namalwa cells NB4 cells Arsenic Species As111 ATO both As111 ATO As111 ATO for both Concentration(s) Tested (nM) 10 for both 10, 125, 500, 1000 1 for both Duration of Treatment 8hr for both LOECa (HM) 10 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Various indicators of apoptosis: Induced (experimental - control) DNA fragmentation: wt, -7%; DKO, NSE. Cytochrome c release: ft in wt, NSE in DKO. Induced caspase-3 activity: wt, -140 units; DKO, none. Caspase-3 activity was only detected in DKO cells when they were permeabilized and incubated for 1 hr in the presence of 4 uM exogenous cytochrome c. These and other experiments showed that mitochondria! events associated with apoptotic cell death induced at concentrations such as 10 or less required Bax and/or Bak. Results from several experiments suggested that extramitochondrial thiol oxidation leading to changes in intracellular Ca2+ compartmentalization plays a critical role in inorganic arsenic-induced cytochrome c release. At concentrations of 125 and higher, Bax and Bak became irrelevant to the mechanism of cytotoxicity and cell death resulted from oxidative stress that led to necrosis. ROS seem to be implicated in a concentration-dependent mechanistic switch between apoptosis and necrosis. 24 hr for both 1 for both without BSO Induced apoptosis (experimental - control) detected by fluorescence microscope analysis after staining with AO and EB: Namalwa cells: inorganic arsenic, -6%; inorganic arsenic + 10 uM BSO, -29%. NB4 cells: inorganic arsenic, -8%; inorganic arsenic + 10 uM BSO, -47%. BSO treatments markedly reduced GSH levels. Reference Nuttetal., 2005 Nuttetal., 2005 Chenetal., 2006 C-107 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Jurkat cells U937 cells Jurkat cells Namalwa cells NB4 cells U937 cells Jurkat cells U937 cells Arsenic Species As111 ATO for both As111 ATO for all As111 ATO for both Concentration(s) Tested (nM) 1 for both 1 for all 1 for both Duration of Treatment 48 hr for both 24 hr for Namalwa and NB4 cells, 48 hr for other 2 lines Various, for 6-72 hr LOECa (HM) None for both without BSO 1 for all with BSO Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (experimental - control) detected by fluorescence microscope analysis after staining with AO and EB: Jurkat cells: inorganic arsenic, NSE; inorganic arsenic + 10 uM BSO, -25%. U937 cells: inorganic arsenic, NSE; inorganic arsenic + 10 uM BSO, -67%. BSO treatments markedly reduced GSH levels. Results of Western blot analysis in all 4 cell lines following co-treatment of inorganic arsenic with lOuMBSO: Big ft in PARP-cleavage; big U in procaspase-3 level. Big ft in JNK phosphorylation (the latter effect was not seen in absence of BSO co- treatment). Time course experiments for co- treatment with 10 uM BSO showed ft in PARP-cleavage; U in procaspase-3 level; strong ft in JNK phosphorylation. Induced apoptosis increased to -85% and -50% by 72 hr in U937 and Jurkat cells, respectively. Reference Chenetal., 2006 Chenetal., 2006 Chenetal., 2006 C-108 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U937 cells U937 cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 1 1 Duration of Treatment 48 hr 48 hr LOECa (HM) 1, but only with B SO co-treat- ment 1, but only with B SO co-treat- ment Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (experimental - control) detected by fluorescence microscope analysis after staining with AO and EB: -55% following the co-treatment with B SO; this ft was not significantly decreased by 4-hr treatments with either 10 mM NAC or 200 units of catalase even though those treatments substantially decreased H2O2 levels. Moreover, NAC and catalase did not block the JNK activation caused by the inorganic arsenic + B SO treatment. Results of Western blot analyses: huge ft inDR5, huge U in Bid, and U in IicBa following co- treatment with 10 uM BSO;NSEonthese3 proteins after inorganic arsenic or BSO alone. Experiments with inhibitors suggested that (1) both caspase- and JNK-mediated pathways (due to activation of NF- KB) participate in the induction of apoptosis that occurs following co- treatment with inorganic arsenic and BSO and (2) that JNK increases DR5 protein levels that in turn mediate that apoptosis. Reference Chenetal., 2006 Chenetal., 2006 C-109 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference NB4 cells, NB4-AsR, and APL primary cells As111 ATO A series of experiments was conducted involving 24-72 hr treatments with concentrations of inorganic arsenic of 0.125-10 uM. Tests of MEK1 mRNA knockdown using inorganic arsenic treatments and MEK1 inhibitors (namely, PD98059 at 40 uM and PD184352 at 1 uM) showed that MEK1 inhibitors and inorganic arsenic synergize to induce apoptosis. Although inorganic arsenic induces apoptosis, it also causes ERK1/2 activation, which tends to decrease the extent of apoptosis by causing phosphorylation at Serll2 of the proapoptotic Bad protein. Phosphorylated Bad protein does not heterodimerize with the Bel proteins. The only known function of the Bad protein is to bind (i.e., heterodimerize) with the death antagonist Bcl-2 family proteins, Bcl-2 and Bcl-xL, thereby blocking their antiapoptotic action by preventing them from binding to Bax/Bak. Because MEK1 inhibitors block this ERK1/2 activation and the phosphorylation of BAD, there is more nonphosphorylated Bad protein to heterodimerize with the Bcl-2 proteins and keep them from functioning to block apoptosis. In this way, exposure to inorganic arsenic in the presence of MEK1 inhibitors greatly increases the extent of apoptosis. Lunghi et al., 2005 Primary AML blasts from 25 patients with non-APL AML As111 ATO In experiments involving 48-hr treatments that used concentrations of inorganic arsenic of 0.125-10 uM in the presence or absence of concentrations of the MEK1 inhibitor PD184352 of 0.1-10 uM, synergistic, additive, or antagonistic interactions in the induction of apoptosis were found in primary cells from 13, 8, and 4 patients, respectively. The p53-related gene p73 was shown to be the molecular target of importance in this interaction, and the synergism had the following basis. Inorganic arsenic induced both the proapoptotic and antiproliferative TAp73 and the antiapoptotic and proproliferative ANp73 isoforms, with no net effect on apoptosis because the TAp73/ANp73 ratio did not change. The MEK1 inhibitor reduced the level of ANp73 and blunted the inorganic arsenic-induced up-regulation of ANp73, with the result that the TAp73/ANp73 ratio increased, leading to more apoptosis. At 1 uM, inorganic arsenic induced only p73, but at doses >2 uM) it also promoted accumulation of p53 protein levels, which also caused apoptosis. Lunghi et al., 2006 C-110 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO Kl cells Normal human keratinocytes treated with 50 mJ/cm2 UVB before or after inorganic arsenic treatment Arsenic Species As111 SA As111 SA for both Concentration(s) Tested (nM) 20, 40, 80 1, as pretreatment 1, as post- treatment begun 24 hr after irradiation Duration of Treatment 4hr 24 hr for both LOECa (HM) 20 1 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis detected by flow cytometry and by the presence of DNA ladders from internucleosomal DNA degradation — ladder effect did not appear until 24 hr after treatment. At dose of 40, it took 8 hours after treatment before apoptosis could be detected by flow cytometry. Reduced levels of apoptosis resulted from treatment with various modulators (antioxidants, a copper ion chelator, a protein kinase inhibitor, and a protein synthesis inhibitor) either simultaneously or, in some instances, immediately following the arsenic treatment. Apoptosis as detected by PI staining and TUNEL assay: the inorganic arsenic treatment alone did not induce a significant increase in apoptosis or cytotoxicity; U in the level of UV- induced apoptosis to control levels, with a corresponding U in cytotoxicity to control levels. A similar amount of apoptosis was seen as with UVB alone, or possibly apoptosis increased slightly; cytotoxicity was similar to that with UVB treatment alone or possibly slightly more extreme. Reference Wanget al., 1996 Chenetal., 2005b C-111 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue A mouse fibroblast cell line as well as various stable transfectants ofJB6C141 cells NB4 cells U937 cells HL-60 cells Mouse 291.03C keratinocytes Arsenic Species As111 SA Asv As111 ATO for all As111 SA for both Concentration(s) Tested (nM) Various 2 for all 5 5 Duration of Treatment — 2 days for all 48 hr 60 hr LOECa (HM) — 2 2 2 5 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Various tests indicated that p53 is not involved in arsenic -induced apoptosis. The pathway of JNKs was shown to play an essential role in arsenic-induced apoptosis. For example, such apoptosis was blocked by expression of the dominant-negative mutant of JNK1. Percentages of apoptosis determined by fluorescent microscopy, and units of basal activity of GSTjc, GPx, and CAT, respectively: 67.5%, 94.0, 28.3, 25.8. 5.6%, 212.1,67.6, 170.5. 5.8%, 138.6, 55.5, 198.3. These data and others showed that the higher the basal levels of these 3 enzymes, the less the inorganic arsenic- induced apoptosis. Higher activities of these enzymes decrease the amount of H2O2 in cells. Modulators that increase activities of these enzymes were shown to decrease apoptosis and vice versa. Apoptosis measured by flow cytometry: ft by 4.20% over control, which was 0.74%. ft by 7.3 1% over control. Reference Huang et al., 1999b Jing et al., 1999 Wuetal., 2005 C-l 12 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Mouse 291.03C keratinocytes irradiated immediately after the arsenic treatment with a single dose of 0.30k J/m2 UV HaCaT cells (immortalized , non- tumorigenic human keratinocyte cell line) arsenic-TL cells (arsenic- tolerant cells, which are HaCaT cells that were cultured for 28 weeks in 100 nM As111 SA) Arsenic Species As111 SA for all As111 SA for both Concentration(s) Tested (nM) None (i.e., UV only) 2.5 5.0 20, 40, 60, 80 for both Duration of Treatment — 24 hr 24 hr 24 hr for both LOECa (HM) — 2.5 5.0 20 40 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis measured by flow cytometry 24 hr after the dose of UV: ft by 26.87% over control, which was 0.74%. ft by 20.62% over control. ft by 9.78% over control. Thus, both pretreatments with As111 SA markedly reduced the amount of UV-induced apoptosis. In parallel with the above, UV-induced caspase 3/7 activity was also decreased by both treatments. Apoptosis detected using flow cytometry following staining with Annexin V and PI: ft in apoptosis. Much smaller ft in apoptosis. There was a significant decrease in apoptosis compared to HaCaT cells at all 4 dose levels. A similar resistance by arsenic-TL cells was seen to apoptosis induction by 25J/cm2ofUVA, as well as by cisplatin, etoposide, and doxorubicin. Arsenic- TL cells showed greatly increased stability of nuclear P-PKB, and pretreatment with chemicals that inhibit PKB phosphorylation blocked inorganic arsenic-induced acquired apoptotic resistance. Reference Wuetal., 2005 Pi et al., 2005 C-113 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MCF-7 cells U-2OS cells Arsenic Species As111 ATO As111 SA Concentration(s) Tested (nM) 3 0.1,1, 10 Duration of Treatment 36 hr 24 hr LOECa (HM) 3 0.1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis detected based on electrophoretic analysis of DNA fragmentation: -18% of the cells were apoptotic. TUNEL staining assay was used to detect apoptotic cells after 0, 24, or 48 hr of post- treatment culturing in arsenic -free medium. At dose of 0.1, apoptotic cells were -0%, -0.3%, and -3.6%, respectively. At dose of 1, apoptotic cells were -0%, -0.2%, and -3.4%, respectively. At dose of 10, apoptotic cells were -0%, -0%, and -0%, respectively. Note that a 24-hr treatment with SA affected apoptosis only if there was an additional 24-hr or longer period of culturing in SA-free medium between the end of the SA treatment and when the assay was done. Reference Ling et al., 2002 Komissaro va et al., 2005 C-l 14 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U-2OS cells Undifferentiat edPC12 cells PARP-1+/+ MEF cells PARP-I-'- MEF cells Arsenic Species As111 SA As111 ATO As111 SA for both Concentration(s) Tested (nM) 0.1, 1, 10 8 11.5,23 for both Duration of Treatment 24 hr 24 hr 24 hr for both LOECa (HM) 0.1 8 11.5 11.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Assay utilizing activation of cellular caspase-3 was used to detect apoptotic cells after 0, 24, or 48 hr of post-treatment culturing in arsenic-free medium: At dose of 0.1, apoptotic cells were -0%, -1.3%, and -6.2%, respectively. At dose of 1, apoptotic cells were -0%, -0.3%, and -5.4%, respectively. At dose of 10, apoptotic cells were -0%, -0%, and -0%, respectively. Note that a 24-hr treatment with SA affected apoptosis only if there was an additional 24-hr or longer period of culturing in SA-free medium between the end of the SA treatment and when the assay was done. Induction of apoptosis detected by annexin V binding and caspase activity: -55% of cells with apoptotic death, rest with necrotic death; at 6 hrs, -60% of dead cells were apoptotic. Induction of apoptosis detected by PI and RNase staining and flow cytometry, visualized as sub-Gl population and reported as % of apoptosis (controls were always -6% at 11. 5, -9% at 23 -11% at 11.5, -21% at 23. Reference Komissaro va et al., 2005 Pigaetal., 2007 Poonepalli etal.,2005 C-115 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PARP-1+/+ MEF cells PARP-r7' MEF cells JB6C141 cells, transfected with IKKP- KM to greatly reduce COX- 2 induction JB6C141 cells transfected with vector only JB6C141 cells, after knockdown of endogenous COX-2 expression to low levels by its specific siRNA JB6C141 cells transfected with mock vector for the siRNA, with normal COX- 2 expression Arsenic Species As111 SA for both As111 SA for both As111 SA for both Concentration(s) Tested (nM) 11.5,23 for both 20,40 for both 10,20 for both Duration of Treatment 48 hr for both 24 hr for both 36 hr for both LOECa (HM) 11.5 11.5 20 20 10 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of apoptosis detected by PI and RNase staining and flow cytometry, visualized as sub-Gl population and reported as % of apoptosis (controls were always 0 /o): -23% at 11. 5, -32% at 23. -40% at 11. 5, -62% at 23. Induction of apoptosis detected by PI staining and flow cytometry: ftft in apoptosis: medium alone, 0.83%; 20, 12.60%, 40, 41.33%;. Slight ft in apoptosis: medium alone, 1.03%; 20, 4.58%, 40, 7.23%. Similar conclusion was reached using TUNEL assay and flow cytometry. Induction of apoptosis detected by PI staining and flow cytometry: ftft in apopthosis: medium alone, 4.14%; 10, 28.45%, 20, 49.22%. Much smaller ft in apoptosis: medium alone, 1.86%; 10, 10.52%, 20, 26.60%. Another experiment showed that pretreatment of normal JB6 C141 cells with NS398, an inhibitor of COX-2, markedly ft amount of apoptosis. Reference Poonepalli etal.,2005 Ouyang et al., 2007 Ouyang et al., 2007 C-116 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MEF cells that were made IKKp-A so that they markedly overexpressed COX-2 MEF cells that had the vector only, with normal (low) level of COX-2 SY-5Y cells HEK 293 cells Arsenic Species As111 SA for both As111 ATO for both Concentration(s) Tested (nM) 20 for both 1 for both Duration of Treatment 36 hr for both 24 hr 48 hr 72 hr LOECa (HM) 20 20 1 for all Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of apoptosis detected by PI staining and flow cytometry: Slight ft in apoptosis: medium alone, 0.68%; 20, 6.35%. Big ft in apoptosis: medium alone, 0.87%; 20, 49.62%. Thus, COX-2 protects cells from apoptosis. Induction of apoptosis detected by Hoechst staining: Response as % of control in SY-5Y and HEK 293 cells, respectively, for each duration of treatment: 266%, 156%. 152%, 192%. 214%, 200%. There was NSE on the mitotic index at any time. Reference Ouyang et al., 2007 Florea et al., 2007 C-117 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PMs from CDFj mice TK6 cells Arsenic Species As111 SA Asv MMAV DMAV TMAV As111 SA As111 ATO Concentration(s) Tested (nM) 10 ImM 10 mM 10 mM 10 mM 0.1, 1 for both Duration of Treatment 48 hr for all 24 hr for both LOECa (HM) 10 ImM 10 mM 10 mM None 1 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis detected based on electrophoretic analysis of DNA fragmentation and by TUNEL staining. The particular assay shown in this row used cellular morphological changes to assess apoptosis and the AlamarBlue assay to measure cell death. Approximate resulting percentages of cell death (listed first) and apoptotic cells (listed second) for the 5 compounds follow: For As111 SA: 82% and 23%. For Asv: 65% and 17%. For MMAV: 10% and 7%. For DMAV: 100% and 100%. ForTMAv: 12% and none. Thus DMAV was unusual in causing almost entirely apoptotic cell death, while the inorganic arsenicals caused mainly necrotic cell death. Apoptosis identified using APO2.7 antibody: ft to 5.0% from 3.6% in control. ft to 5.5% from 3.6% in control. Reference Sakurai et al., 1998 Hornhardt etal.,2006 C-118 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TK6 cells irradiated with 1,2, or 4 Gyof69 cGy/min gamma radiation at beginning of inorganic arsenic treatment HCT1 16 cells (securin +/+) HCT1 16 cells (securin -/-) Arsenic Species As111 SA As111 ATO As111 SA for both Concentration(s) Tested (nM) 0.1, 1 for both 16 for both Duration of Treatment 24 hr for both 24 hr for both LOECa (HM) None 1 16 16 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Apoptosis identified using APO2.7 antibody: At dose of 1: 1 Gy, 9.1%;2Gy, 10.4%, 4 Gy, 22.6%; SA had no significant effect on any of them. At dose of 1: 1 Gy, 12.5%; 2 Gy, 21.75%, 4 Gy, 38.6%; ATO caused a significant increase over the control (no inorganic arsenic + radiation) at all 3 radiation doses. This was a synergistic interaction. Induced apoptosis (i.e., experimental - control) detected using fluorescent microscopy after Hoechst staining: securin +/+: -6%; securin -/-: -10%; with the amount of apoptosis in the null mutant being significantly higher. Reference Hornhardt etal.,2006 Chao etal., 2006a C-119 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells NB4-M-AsR2 cells DVI9 cells Gclm"7" MEF cells, from GCLM knockout mice Arsenic Species As111 ATO for all As111 SA for both durations Concentration(s) Tested (nM) 0.5, 1 for all 25 for both durations Duration of Treatment 48 hr for all 8hr 24 hr LOECa (HM) 0.5 1 0.5 25 for both dura-tions Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (i.e., experimental - control) for ATO alone and for ATO with 100 uM Trolox, detected using PI staining in binding buffer: At 0.5: -6% alone, -20% with Trolox; at 1: -16% alone, -55% with Trolox. At 0.5:0% alone, -11% with Trolox; at 1: -14% alone, -45% with Trolox. At 0.5: -1.5% alone, -4% with Trolox; at 1: -6% alone, -20% with Trolox. Additional support for the conclusion that Trolox enhanced ATO- mediated apoptosis was provided by an annexin V-FITC staining assay and by the observation that Trolox significantly enhanced the percentage of cells with activated caspase-3 and cleaved PARP. Induced apoptosis (i.e., experimental - control) detected by staining with FITC-labeled annexin-V and PI: At 8 hours: -5% early apoptotic, -38% late apoptotic, -8% necrotic. At 24 hours: -3% early apoptotic, -79% late apoptotic, -5% necrotic. Experiments in Gclm+/+ cells showed that co- treatment or pretreatment with tBHQ partially or completely blocked inorganic arsenic- induced apoptosis. Reference Diaz et al., 2005 Kannet al., 2005b C-120 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MEFs MEFs that were wt MEFs that were Bax" " and Bak"7" double knockout (DKO) cells Arsenic Species As111 ATO As111 ATO for both Concentration(s) Tested (nM) 2,3,5 10 for both Duration of Treatment 3 days 12 hr for both LOECa (HM) 2 10 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (i.e., experimental - control) for ATO alone and for ATO co-treatment with Trolox, detected by PI staining using flow cytometry: ATO alone: 2, -9%; 3, -22%; 5, -62%. ATO and Trolox: 2, -3%; 3, -3%; 5, -20%. Thus, in contrast to what happened in malignant cells, Trolox blocked the effects of ATO. Induced apoptosis (i.e., experimental - control) detected by PI staining and F ACS: -23% in wt and -7% in DKO; the results at dose of 500 are ignored here. wt: large ft in release of cytochrome c, which was mostly blocked by pretreatment with BAPTA-AM; DKO: trace ft in release of cytochrome c. Results showed that cytochrome c release and apoptosis occurred largely via a Bax/Bak- dependent mechanism. Reference Diaz et al., 2005 Bustamant e et al., 2005 C-121 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Isolated rat liver mitochondria loaded with Ca2+ SVEC4-10 cells RAW264.7 cells RAW264.7 cells Arsenic Species As111 ATO As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 10, 50, 100 20 5,25 5,25 Duration of Treatment 2 min 24 hr 24 hr 24 hr LOECa (HM) 10 20 5 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) There was a dose- dependent, cyclosporin A-sensitive release of cytochrome c via induction of mitochondrial permeability transition and subsequent swelling of mitochondria. Mitochondrial GSH did not seem to be a target for inorganic arsenic which, however, appeared to cause oxidative modification of thiol groups of pore- forming proteins, notably adenine nucleotide translocase. Induced apoptosis (i.e., experimental - control), apoptotic cells were counted by hemocytometer in a fluorescence microscope: -68%. Apoptosis detected by TUNEL assay; results were presented as mean densities of TUNEL staining: there was a positive dose-response. Apoptosis detected by fluorescence staining of caspase-3 activation: there was a positive dose-response. A 30- min pretreatment with DPIC (which inhibits H2O2 production) completely blocked caspase-3 activation at both inorganic arsenic doses, thus showing that it prevented induction of apoptosis by inorganic arsenic. Reference Bustamant e et al., 2005 Hsu et al., 2005 Szymczyk etal.,2006 Szymczyk etal.,2006 C-122 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NIH 3T3 cells HL-60 cells Arsenic Species As111 SA As111 ATO Concentration(s) Tested (nM) 5, 10, 20, 50, 100, 200 3 Duration of Treatment 6hr 48 hr LOECa (HM) 10 3 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of caspase 3/7 activity assayed using Caspase-Glo™ assay (an indicator of apoptosis): units of activity at 0, 10, 50, 100, and 200 were about 2.5, 4, 12, 17, and 36, respectively. Pre- induction of HSP by conditioning heat shock (2 hr at 42°C on prior day) or by constitutive expression of HSP70 markedly reduced the induction, as follows: With heat: NSE at any dose. With constitutive expression: at most a hint of induction at highest 3 doses. Induced apoptosis (i.e., experimental - control), based on TUNEL assay: 15%. Effect of intracellular AA (icAA): (cells were loaded with 4 mM icAA by incubating them with DHA prior to inorganic arsenic treatments, thus avoiding generation of extracellular ROS in tissue culture media caused by direct addition to it of AA) Induced apoptosis for inorganic arsenic + icAA = 1% (NSE). Results using annexin V/FITC assay gave a consistent but milder effect. Reference Khalil et al., 2006 Karasawa s et al., 2005 C-123 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue H22 cells BAEC cells NB4 cells NB4 cells Arsenic Species As111 ATO for both As111 ATO As111 ATO Concentration(s) Tested (nM) 0.5, 1,2,4 for both 3 1 Duration of Treatment 24 hr, 48 hr 24 hr, 48 hr 48 hr 48 hr LOECa (HM) 1,0.5 1,1 3 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis index (i.e., experimental - control), based on TUNEL assay: H22, 24 hr: 0.5, NSE; 1, -8%; 2, -22%; 4, -35%. H22, 48 hr: 0.5, -8%; 1, -20%; 2, -36%; 4, -45%. BAEC, 24 hr: 0.5, NSE; 1, -6%; 2, -22%; 4, -26%. BAEC, 48 hr: 0.5, NSE; 1, -8%; 2, -28%; 4, -40%. % of cells with nuclear fragmentation (NuFr): -80%. Effects of modulators at high doses: Co-treatments with 1000^000 uM DTT: dose-related marked U in NuFr reaching -20%. Co-treatments with 100- 400 uM DMSA: dose- related marked U in NuFr reaching -20%. Co-treatments with 50- 200 uMDMPS: dose- related marked U in NuFr reaching -27%. % of cells with NuFr: -20% for experiments with DTT and DMSA; about 12% in experiment withDMPS. Effects of modulators at low doses: Co-treatments with 12.5- 50 uM DTT: dose- related marked ft in NuFr reaching -90%. Co-treatments with 10- 40 uM DMSA: dose- related marked ft in NuFr reaching -75%. Co-treatments with 5-20 uM DMPS: dose-related marked ft in NuFr reaching -80%. Reference Liuetal., 2006e Jan et al., 2006 Jan et al., 2006 C-124 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue 293 cells SV-HUC-1 cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 2 2 Duration of Treatment 48 hr 48 hr LOECa (HM) 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) % of cells with sub-Gl DNA content: untreated = -5%; dose of 2: big ft to -53%. Effects of co-treatment (CoTr) with modulators at high doses: CoTr 200 uM DMSA: U from inorganic arsenic alone to -26%. CoTr 100 uMDMPS:ll from inorganic arsenic alone to -37%. Effects of CoTr with modulators at low doses: CoTr 20 uM DMSA: ft from inorganic arsenic alone to -83%. CoTrlOuMDMPS:ft from inorganic arsenic alone to -88%. % of cells with sub-Gl DNA content: untreated = -6%; dose of 2: big ft to -46%. Effects of CoTr with modulators at high doses: CoTr 200 uM DMSA: U from inorganic arsenic alone to -22%. CoTr 100 uMDMPS:U from inorganic arsenic alone to -28%. Effects of CoTr with modulators at low doses: CoTr 20 uM DMSA: ft from inorganic arsenic alone to -70%. CoTrlOuMDMPS:ft from inorganic arsenic alone to -72%. Reference Jan et al., 2006 Jan et al., 2006 C-125 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue A549 cells A549 cells WM9 cells OM431 cells LU1205 cells Arsenic Species As111 ATO As111 ATO As111 SA for all Concentration(s) Tested (nM) 1, 2, 5, 10, 20, 50 2 4 Duration of Treatment 48hr 48 hr 48 hr LOECa (HM) 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival determined by MTT assay: LC50 = -27. Cell survival determined by flow cytometry after annexin V and PI staining: inorganic arsenic at dose of 2: NSE. 200 uM sulindac: NSE. (2 uM inorganic arsenic + 200 uM sulindac): -40% cytotoxicity; pretreatment with NAC almost completely blocked this synergistic interaction. Regarding caspase 3/7 protein levels: 2 uM inorganic arsenic: NSE. 200 uM sulindac: NSE. (2 uM inorganic arsenic + 200 uM sulindac): ft to ~1.4x. Regarding caspase 9 protein levels: 2 uM inorganic arsenic: ft to 1.05x. 200 uM sulindac: NSE. (2 uM inorganic arsenic + 200 uM sulindac): ft to ~1.5x. There was also a clear synergistic interaction between these treatments in causing big U of both procaspase-3 and procaspase-9 protein levels. Pretreatment with NAC almost entirely blocked the caspase 3/7 and caspase 9 effects. 4 Induced apoptosis (i.e., experimental - control), based on PI staining and FACS analysis of hypo- diploid content of DNA in the pre-GO/Gl region: WM9, ~32%;OM431, -17%; LU1205, -18%. Treatment with soluble recombinant TRAIL was effective in inducing apoptosis; combined treatment with inorganic arsenic yielded no more than an additive effect. Reference Jin et al., 2006b Jinetal., 2006b Ivanov and Hei, 2006 C-126 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested (nM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Reference Cancer Promotion BALB/c 3T3 A3 1-1-1 cells (derived from mice) BALB/c 3T3 A3 1-1-1 cells (derived from mice) V79 cells As111 SA AsvDA MMAV DMAV As111 SA AsvDA MMAV DMAV As111 SA AsvDA MMAV DMAV 0.2, 0.5, 1, 2, 5 0.5, 1,2,5, 10 50, 100, 200, 500, 1000 10, 20, 50, 100, 200 1 5 500 50 0.15,0.3,0.7, 1.5, 2.5 0.5, 1.5,2.5,5, 10, 20 0.5, 1.5,2.5,5, 10, 20 mM 0.15,0.3,0.6, 1.3, 2.7, 5 mM 18 days for all 18 days for all 72 hrs for all 0.5 1 200 None 1 5 500 None 0.7 5 5 mM None Caused promotion in a two-stage transformation assay; based on a significant increase in the number of transformed cells after an initiating treatment of 0.2 ug/mL MCA for 3 days followed by post- treatment with an arsenic compound for 18 days. At doses above the LOEC, the responses increased no more than slightly with dose. For As111 SA there was a humped dose-response with a peak at the dose ofl. Caused promotion in a two-stage transformation assay; based on a significant increase in the number of transformed cells after an initiating treatment of 10 uM As111 SA for 3 days followed by post- treatment with an arsenic compound for 18 days. Inhibited gap-junctional intercellular communication, which is a mechanism linked to many tumor promoters; it is based on the metabolic cooperation assay, which detects chemicals that inhibit the transfer of the lethal metabolite of 6- thioguanine from HPRT- proficient to HPRT- deficient cells, thereby allowing recovery of the 6-thioguanine-resistant (HPRT-deficient) cells. Tsuchiya etal.,2005 Tsuchiya etal.,2005 Tsuchiya etal.,2005 C-127 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested (nM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Reference Cell Cycle Arrest or Reduced Proliferation MGC-803 (human gastric cancer) MC/CAR (human multiple myeloma cell line) UROtsa cells V79 cells As111 ATO As111 ATO As111 SA Asv MMAmO MMAV DMAmI DMAV DMAV 0.01-1 1,2,3,4,5 0.1,0.5, 1,5 1,200 0.1,0.5, 1,5 1,200 0.1,0.5, 1,5 1,200 1,2, 5mM 24 hr 72 hr 24 hr for all 12 hr 0.01 1 1 None 1 None 5 None ImM Growth inhibition (growth measured by MTT assay): at various doses, growth inhibition was also induced in 5 other human malignant cell lines. Growth inhibition (growth measured by MTT assay): About 60% inhibition at 2; cells were arrested in both Gl and G2-M phases. Growth inhibition was also induced in 7 other human multiple myeloma cell lines to various degrees. Extent of reduction of cell proliferation based on [3H]thymidine incorporation: Cell proliferation reductions at dose of 5 were approximately as follows: DMAmI, 15%; As111, 30%; MMAmO, 85%. Induction of mitotic delay and formation of aberrant mitotic spindles, including tripolar and quadripolar spindles: -18% aberrant spindles at 1 mM. y-tubulin was co-localized with the aberrant spindles. The following things were noted to occur after exposure of V79 cells to DMAV: multiple MTOCs, multipolar spindles, amoeboid cells, multinucleated cells, and cell death. Zhang et al., 1999 Parketal., 2000 Drobna et al., 2002 Ochi et al., 1999a C-128 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HeLa S3 cells U937 cells Arsenic Species As111 SA As111 SA for both Concentration(s) Tested (nM) 1, 3, 5, 10, 20 2.5,5, 10 for both durations Duration of Treatment 24hr 24 hr 48 hr LOECa (HM) o 5 2.5 for both dura-tions Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cells arrested at mitotic stage: At dose of 5, 35% of cells were arrested in that stage. Of 7 cell lines tested in this way, two others were almost as sensitive to this effect. Examination of cells arrested in mitosis showed abnormal mitotic figures and spindles, as well as deranged chromosomal congression. Cell numbers counted with a Coulter counter: After 24 hr at the doses of 2.5, 5, and 10, there were approximately 71%, 56%, and 43% as many cells as in the control group, respectively. After 48 hr at the doses of 2.5, 5, and 10, there were approximately 54%, 38%, and 23% as many cells as in the control group, respectively. There was little if any cytotoxicity even at 48 hr at the dose of 5. Reference Huang and Lee, 1998 McCollum et al., 2005 C-129 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U937 cells Arsenic Species As111 SA Concentration(s) Tested (nM) 5 Duration of Treatment 8hr Results (Compared With Controls, With All Concentrations LOECa Being (|oM) in |iM Unless Noted) The LOEC was 5. Centrifugal elutriation was used to enrich cells in different phases of the cell cycle so that the effect of inorganic arsenic could be tested on them. Progression of inorganic arsenic-treated cells from each cell cycle stage to the next was studied, and it was found that inorganic arsenic slowed cell growth in every phase of the cycle. For example, in asynchronous populations of untreated cells, DNA synthesis lasted 10 to 12 hr. However, in cells treated with 5 uM inorganic arsenic, it lasted 16 hr. In the presence of inorganic arsenic, cells in Gl entered the S phase more slowly, etc. Cell passage from any cell cycle phase to the next was inhibited by 5 uM inorganic arsenic arsenite. Clearly there was not induction of cell-cycle arrest at one specific checkpoint. The biggest inorganic arsenic -induced slowdown occurred between M and Gl, and the next biggest was between G2 and M. By looking at caspase activity, they showed that inorganic arsenic induced apoptosis specifically in cell populations delayed in the G2/M phase. Reference McCollum et al., 2005 C-130 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PARP-1+/+ MEF cells PARP-r7' MEF cells CGL-2 cells Arsenic Species As111 SA for both As111 SA Concentration(s) Tested (nM) 11.5,23 for both 1, 2, 3, 4, 5, 7, 10 Duration of Treatment 24 and 48 hr for both 24 hr LOECa (HM) 11.5 for both at both times 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) inorganic arsenic caused much disruption of cell cycle as shown by PI and RNase staining and flow cytometry when visualized as proportions of cells that were in G2/M, S, Gl,orsub-Gl (i.e., apoptotic) under the different conditions. Disruption was more extreme in PARP-l"'" MEF cells. Results for apoptosis, which are easier to quantify, are detailed in separate rows. Especially at the highest inorganic arsenic dose in PARP-1"7" cells, the proportion of G2/M cells became especially small, at least when the comparison was made to all cells and not just to non-apoptotic ones. Cell survival was determined using a colony-forming assay: LC50= 1.7. arsenic mitotic cells round-up, they can be separated from the attached interface cells by using the shake-off technique. When that technique was applied to a sample at the dose of 2, 96% of the attached cells were found to be alive, and 96% of the floating (i.e., mitotic) cells were found to be dead, thus indicating that inorganic arsenic induced mitosis- mediated cell death. Reference Poonepalli etal.,2005 Yihetal., 2005 C-131 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CGL-2 cells CGL-2 cells HeLa S3 cells Arsenic Species As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 1, 2, 3, 4, 5, 10 Duration of Treatment 24 hr LOECa (HM) 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Treatments caused a shift in percentages of cells in Gl, S, andG2/M, with a dose-dependent ft in G2/M cells over the range of doses of 0 (-25%) to 4 (-85%), followed by a U above a dose of 5 that reached -50% at dose of 10. G2/M cells were predominantly mitotic cells. Mitotic arrest was associated with inorganic arsenic -induced cell death (see row immediately above). When synchronized cells were treated with dose of 2, all cells, whether treated in the Gl, S, or G2 stage, progressed into and arrested at mitosis, where they were demonstrated to contain damaged DNA, as demonstrated by the appearance of the DNA double-strand-break marker phosphorylated histone H2A.X (y- H2AX). Following on from row above, other experiments showed that inorganic arsenic appears to inhibit activation of the G2 DNA damage checkpoint and thereby allows cells with damaged DNA to proceed from G2 into mitosis. The subsequent arresting of cells with damaged DNA in mitosis is thought to enhance the induction of apoptosis. 5, 10, 20, 50 Ihr 10 Inhibition of mitotic exit after cells were arrested in mitosis by treatment with nocodazole and the nocodazole was removed before arsenic treatment. This shows that such a dose interferes with mitosis. Reference Yihetal., 2005 Yih et al., 2005 Huang and Lee, 1998 C-132 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HLFC cells HLFK cells (Ku70 deficient) Human primary peripheral blood lymphocytes Human primary peripheral blood lymphocytes Arsenic Species As111 SA for both As111 SA Asv MMAm MMAV DMA111 DMAV As111 SA Asv MMAm MMAV DMA111 DMAV Concentration(s) Tested (nM) 1,2.5,5, 10 for both 1.25-160 1.25-500 0.1-2.7 10-10000 0.11-12.26 10-10000 1.25-160 1.25-500 0.1-2.7 10-10000 0.11-12.26 10-10000 Duration of Treatment 4hr for both 24 hr for all 24 hr for all LOECa (HM) 2.5 2.5 2.5 50 1.5 10000 1.02 3000 20 150 1.8 None 1.02 300 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Following the 4-hr As111 treatment, cells were incubated in arsenic-free medium for 24 hr before determining the proliferation index and the proportions of cells in different parts of cell cycle. Both cell types had U proliferation index and an ft in G0/Gi cells at dose of 2.5. Both effects were more extreme in HLFK than in HLFC cells at the 3 highest doses. Replicative index (RI): All 6 compounds induced significant slowing of the cell cycle. Methylated trivalent arsenicals were 3 orders of magnitude more potent than the methylated pentavalent arsenicals. Inorganic arsenic compounds were substantially more toxic than methylated pentavalent arsenicals. Mitotic index (MI): u. u. u. NSE. ft to peak of 3x at 3. 07. ft to peak of 6x at 1000. Both decreased abruptly near concentration at which RI showed ft proportion of first division metaphases. This is consistent with disruption of spindle integrity. Reference Liuetal., 2007b Kligerman etal.,2003 Kligerman etal.,2003 C-133 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human peripheral lymphocytes TR9-7 cells that were released from being mostly synchronized in G2 (using Hoechst 33342) shortly before inorganic arsenic treatment began PCI-1 cells CHO cells treated with MMS before or after inorganic arsenic treatment Arsenic Species As111 SA As111 SA As111 ATO As111 SA Concentration(s) Tested (nM) 5 5 1,2,3,4 10, as pretreatment 10, as post-treatment Duration of Treatment 24 hr 1-24 hr 3 days 24 hr 24 hr LOECa (HM) 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) There was delayed cell cycle progression. In treated cells, 73% and 32% were still in the first mitotic division at fixation times of 72 and 96 hr, respectively, whereas in untreated cells up to 90% were in second or subsequent divisions at these times. Conclusions based on mitotic indices determined over the 24-hr period in cells made p53(+) or p53(") by controlling tetracycline levels: inorganic arsenic delayed entry into mitosis in both p53(+) and p53(-) cells, with peak being delayed by ~3 hr from that of cells unexposed to inorganic arsenic. Mitotic exit occurred at a normal rate in inorganic arsenic-treated p53(+) cells but was markedly delayed in p53(-) cells and only reached the baseline level after 24 hr, by which time the inorganic arsenic-treated p53(+) cells had already reached that level and had begun to cycle again. 2 10 10 Growth inhibition (growth measured by MTT assay): About 50% inhibition at 2; cells were arrested in the G2-M phases. Growth inhibition was also induced in 3 other human head and neck squamous cell carcinoma cell lines. Inhibition of mitosis and cell proliferation: U in inhibition of both endpoints compared to MMS alone. ft in inhibition of both endpoints compared to MMS alone, synergistic interaction. Reference Jha et al., 1992 McNeely etal.,2006 Seoletal., 1999 Lee etal., 1986 C-134 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MCF-7 cells Human lymphocytes Chinese hamster V79 cells Arsenic Species As111 ATO As111 SA As111 SA DMAV Concentration(s) Tested (nM) 3 10'10, ID'8, ID'6, 10"4, 0.01, 1 5 2mM Duration of Treatment 24 hr 2hr 24 hr for both LOECa (HM) 3 10-io 5 2mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Treatment blocked the cell cycle in mitosis, resulting in a time- dependent accumulation of cells in G2/M, with about 50% in G2/M at this time. Induction of mitotic arrest: 4 of 5 donors showed statistically significant increase at lowest dose. All showed significant increase from dose of 10" 8 through 0.01. There was much inter- individual variation, but there was a positive dose-response within data for each donor. There was a almost no response at dose of 1 because of cytotoxicity. Accumulation of mitotic cells and other abnormal cells as follows (approximate percentages of cells of each type present after 24-hr treatment): Control (assumed same as distribution at starting time): 97% mononucleated, 3% metaphase. As111: 75% mononucleated, 11% metaphase, 10% binucleated, 4% multinucleated. DMAV: 24% mononucleated, 52% metaphase, 1% binucleated, 23% multinucleated. DMAV caused disappearance of microtubule network and abnormalities of mitotic microtubules (i.e., spindles) — there was a big ft increase in frequency of multipolar and aster-like spindles. Reference Ling et al., 2002 Vegaetal., 1995 Ochi et al., 1999b C-13 5 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SVEC4-10 cells HCT1 16 cells (securin +/+) HCT1 16 cells (securin -/-) SVEC4-10 cells HT1 197 cells Arsenic Species As111 SA As111 SA for both As111 SA As111 SA Concentration(s) Tested (nM) 2, 4, 8, 16 4, 8, 12, 16 for both 20 1,5, 10 Duration of Treatment 24 hr 24 hr for both 24 hr 24 hr LOECa (HM) 4 12 4 20 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Fraction of cells in G2/M phases of cell cycle: slight ft at 4, big ft at 8 and 16. Also an effect on rate of cell growth (tested at 4, 8, 12, 16): U at all doses, with a strong dose-response. Fraction of cells in G2/M phases of cell cycle: Similar ft at 12 and 16 to -39%. ft at 4 to -3 8% with a positive dose-response, reaching -49% at highest dose. Consistent with the conclusion, based on the above data, that securin protects against arsenic-induced cell cycle arrest, the -/- cells also showed a much bigger ft in the mitotic index and in the fraction of cells in "anaphase/mitosis." They also showed sister- chromatid separation. Cell numbers were counted using a hemocytometer: after 6 days of culturing after the inorganic arsenic treatment, there were -25% as many cells as in the control. Complete inhibition of cell proliferation occurred eventually at the dose of 10, with an accumulation of cells in S-phase. At the dose of 10, after 12 and 24 hr, 1.5x and 2. Ix more cells were in S-phase than in control, respectively, with a large deficit of cells inGl. Reference Chao etal., 2006a Chao etal., 2006a Hsu et al., 2005 Hernandez -Zavala etal., 2005 C-136 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CL3 cells, synchronous atGl Human lymphoblastoi d cells Lyophilized bovine tubulin Arsenic Species As111 SA As111 SA Asv MMAm MMAV DMA111 DMAV As111 SA Asv MMAm MMAV DMA111 DMAV Concentration(s) Tested (uM) 50 0.2, 0.4, 0.6, 1, 2.5, 5, 10 uM 0.5, 1,2.5,5,7.5, 10 mM 0.05,0.1,0.2,0.3, 0.4,0.5,1 uM 0.5, 1,2.5,5,7.5, 10 mM 0.05,0.1,0.2,0.3, 0.4, 0.5 uM 0.5, 1,2.5,5,7.5, 10 mM 0.1, 1, 10 mM 0.1, 1, 10 mM 1, 10, 100 uM 0.1, 1, 10 mM 1, 10, 100 uM 0.1, 1, 10 mM Duration of Treatment 3hr 6hr for all Time course over 1 hr LOECa (UM) 50 None 7.5 mM 0.4 uM None None 5mM ImM None 1 uM 0.1 mM 10 uM 0.1 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell proliferation, based on cell number: U to -35% of control. Survival was cut to 20%-25% by co- treatment with PD98059 orU0126. Effect on the mitotic index: NSE, but results were confounded by high toxicity. Slight statistically significant ft in slope. Statistically significant ft in slope. Slight statistically significant ft in slope. Equivocal, highly variable, effects probably because of toxicity. Statistically significant ft in slope. Effect on GTP-induced polymerization of lyophilized bovine tubulin: U at 1 mM, M at 10 mM. NSE. Slight ft at 1 uM, U at 10 uM, M at 100 uM. Slight ft at 0.1 and 1 mM, ft at 10 mM. U at 10 uM, M at 100 uM. U at 0. ImM, NSE at 1 mM, ft at 10 mM. Reference Li et al., 2006a Kligerman et al., 2005 Kligerman et al., 2005 Cell Proliferation Stimulation C-137 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue K562 cells (human erythroleukem ia cells) NHEK cells NHEK cells Arsenic Species As111 ATO As111 SA Asv, MMAV, DMAV As111 SA Concentration(s) Tested (nM) 2.5 0.001,0.005,0.01, 0.05,0.1,0.5, 1,5, 10 for all 0.2, 0.4, 0.8 Duration of Treatment 12hr 48 hr 24 hr for all Iday 2 days 3 days LOECa (HM) 2.5 2.5 0.005 None 0.2 0.4 0.4 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) -27% of cells are mitotic. (In control, only 4% of cells are mitotic.) -55% of cells are mitotic. Stimulation of cell proliferation, but with inhibition of cell proliferation at > 0.05. Stimulation was measured as incorporation of 3[H]thymidine into cellular DNA. No stimulation of cell proliferation; inhibition of cell proliferation at 0.05 or higher. Increase in proliferation based on cell counts: ft of 32%, 58%, and 50%, respectively. ft of 20% and 21% at doses of 0.4 and 0.8, respectively. ft of 27%, only at dose of 0.4. PI staining and FACS analysis after 2 days showed a significant shift from cells in Gl to cells in G2/S at both doses that showed an ft in proliferation. Reference Li and Broome, 1999 Vegaetal., 2001 Hwang et al., 2006 C-13 8 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells JB6C141 cells transfected as described for this assay JB6C141 cyclinDl-Luc massl cells HaCaT cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 1.25,2.5,5 1.25 5 0.5, 1.0 Duration of Treatment 48hr 72 hr 24 hr 20 passages LOECa (HM) 1.25 1.25 5 — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft in fraction of cells in S phase: at doses of 0, 1.25, 2.5, and 5, the percentages of cells in S were 24.9%, 29.8%, 33.8%, and 38.7%, respectively. Since there was a corresponding ft in fraction of cells in G2/M phase, it was concluded that inorganic arsenic promoted the transition fromGltoS. The 24-hr treatment caused a similar effect at the 2 higher doses. Proliferation was measured by using the CellTiter-Glo® Luminescent Cell Viability Assay: ft in proliferation index to ~1.62x. Fraction of cells in S phase and cell apoptosis (i.e., cell sub-Gl phase) were measured using PI staining with flow cytometry: ft in fraction of cells in S from -11.8% to -14.5%; there was no induction of apoptosis and no evidence of cytotoxicity. Not a significantly increased growth rate, but the trend was in that direction with accumulated population doublings of 58 to 67 in the control and 1.0 groups, respectively, with the value being -6 1 in the 0.5 dose group. Reference Ouyang et al., 2005 Ouyang et al., 2006 Ouyang et al., 2006 Chien et al., 2004 C-139 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue C3H 10T1/2 cell line (mouse cells with fibroblast morphology during routine culture but capable of differentiation into adipocytes) C3H 10T1/2 cell line (mouse cells with fibroblast morphology during routine culture but capable of differentiation into adipocytes) BothHL-60 cells and HaCaT cells UROtsa cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 6 6 0.1,0.5, 1, 10,20, 40 2,4 Duration of Treatment 8wk 8wk 5 days 72 hr LOECa (HM) 6 6 0.5 but possibly 0.1 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Marked increase in FBS- stimulated DNA synthesis (detected using [3H]thymidine incorporation) following dexamethasone/insulin treatment (to induce differentiation), but only after the arsenite exposure has been stopped — the increased mitogenic response is masked while the arsenite treatment continues. Marked increase in cell number compared to control cells following dexamethasone/insulin treatment (to induce differentiation), but increase only occurs after the arsenite exposure has been stopped. By use of MTT assay: ft in cell number, with peak at 0.5; U in cell number to below control level at 1, with a continuing decrease at higher concentrations. (Same general response, but to a lesser extent, with same treatments over 1 day or 3 days.) Increase in cell proliferation based on statistically significant increase in [3H]thymidine incorporation; also there was a significantly higher fraction of cells in S-phase of cell cycle. Reference Trouba et al., 2000 Trouba et al., 2000 Zhang et al., 2003 Simeonova etal.,2000 C-140 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NHEK cells Arsenic Species As111 SA MMAm DMA111 Concentration(s) Tested (nM) 0.2, 0.4, 0.6, 0.8, 1, 2, 4, 6, 12 0.1,0.2,0.4,0.5, 0.8, 1,2 0.1,0.2,0.4,0.5, 0.6, 0.7, 0.8, 1, 2, 3 Duration of Treatment 24 hr for all; index was then determined immediately LOECa (HM) 2 0.5 0.6 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Proliferation index based on MTT assay; the statistical comparison was with the untreated control: ft at 3 doses from LOEC through 6. ft at 2 doses from LOEC through 0.8. ft at 2 doses from LOEC through 0.7. Significant cytotoxicity occurred at 12 uM and higher for inorganic arsenic and at 1 uM and higher for the other arsenicals. Cell cycle distributions were changed in many different ways. Reference Mudipalli etal.,2005 C-141 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NHEK cells irradiated with 100 mJ/cm2 of UVB to arrest 94.5% of cells in GO/G! stages of cell cycle while only killing 2-3% of the cells. Postconfluent PAEC cells in a monolayer PAEC cells in mid- exponential growth in a monolayer Arsenic Species As111 SA MMAm DMA111 As111 SA for both Concentration(s) Tested (nM) 0.2, 0.4, 0.6, 0.8, 1, 2, 4, 6, 12 0.1,0.2,0.4,0.5, 0.8, 1,2 0.1,0.2,0.4,0.5, 0.6, 0.7, 0.8, 1, 2, 3 1, 2.5, 5, 10, 20 1, 2.5, 5, 10, 20 Duration of Treatment 24 hr for all; index was then determined Immediately 4 hr for both LOECa (UM) 0.6 0.4 0.4 1 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Proliferation index based on MTT assay; the statistical comparison was with the untreated control: ft at 6 doses from LOEC through 6. ft at 4 doses from LOEC through 1.0. ft at 5 doses from LOEC through 0.8. Significant cytotoxicity occurred at 12 uM for inorganic arsenic and at 1 uM and higher for the other arsenicals. At all doses showing a significant effect on the proliferation index after arsenical exposure, the point estimate was always higher in the cells with prior UVB exposure. Cell cycle distributions were changed in many different ways. Incorporation of [3H]thymidine into genomic DNA: ft at 1, 2.5, and 5, indicating a mitogenic response. Only the response at 5 is significantly higher, but the 2 lower doses are probably also higher; there was no effect at higher doses. U in rate of DNA synthesis. (In the absence of any treatment, such cells have a higher rate of DNA synthesis than the postconfluent cells in a monolayer.) Reference Mudipalli etal.,2005 Barchowsk y et al., 1996 C-142 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PAEC from freshly harvested vessels U-2OS cells SHE cells TM3 cells Arsenic Species As111 probably ATO, but called arsenite As111 SA DMAmI As111 SA Concentration(s) Tested (nM) 1,5,10 0.01,0.05,0.1, 0.25,0.5, 1,2.5 0.1,0.25,0.5, 1.0 0.000008, 0.00008, 0.0008, 0.008, 0.08, 0.77 Duration of Treatment 24 hr 24 hr Iday 72 hr LOECa (HM) 1 0.01 0.1 0.000008 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Extent of cell proliferation was estimated using fluorescent Cyquant assay: ft at 1 and 5, but U at 10. Cell survival was determined using the clonal survival treat-and- plate method: At doses of 0.01 and 0.05, clonal-forming ability was stimulated to 120%-124%ofthe control, p < 0.006. There was no increase at a 72-hr exposure or at higher doses with a 24-hr exposure. Similar results were found with the neutral red and MTT assays, and sometimes with those assays the point estimates still showed an increase at the dose of 0.01 after the 72-hr exposure. Cell growth (no. of viable cells): ft at both 0.1 and 0.25, and also big ft for them after 2 and 3 days. Increase by 1 day at dose of 0. 1 was ~8-fold. At dose of 1.0, -40% cytotoxicity. No clear effect at 0.5 until after 3 days, then -40% cytotoxicity. Increase in cell proliferation: a statistically significant increase at all doses except 0.77; the peak of -152% of control was at 0.00008. Reference Barchowsk y et al., 1999a Komissaro va et al., 2005 Ochi et al., 2004 DuMond and Singh, 2007 C-143 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells NHEK cells, both with and without irradiation with 100 mJ/cm2 of UVB to arrest 94.5% of cells in GO/G! stages of cell cycle while only killing 2%-3%ofthe cells HELP cells Arsenic Species MMAm for all As111 SA MMAm DMA111 As111 SA Concentration(s) Tested (nM) 0.05 for all 6 0.8 0.8 0.1,0.5, 1,5, 10 Duration of Treatment 12 weeks 24 weeks 52 weeks 24 hr for all 24 hr LOECa (HM) 0.05 for all 6 0.8 0.8 0.1 for ft 5forli Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Shortened cell population doubling times (hr) based on counting cells in trypan blue exclusion assay: (control doubling time = 42 hr) 27 hr. 25 hr. 21 hr. Examination of expression profiles of more than 10 cell cycle and cell signaling proteins that seem likely to influence cell proliferation showed that many large changes occurred following the UVB and arsenic treatments, arsenic examples, all 3 arsenicals caused a big ft in nuclear cyclin D 1 in UVB irradiated cells, and, for nuclear PCNA in UVB-irradiated cells, MMA and DMA caused a big ft while inorganic arsenic had no effect. Activation of JNK phosphorylation and increased EOF expression and phosphorylation of the EOF receptor occurred. Cell proliferation efficiency based on MTT assay: ft to 150% and 175% of control at 0.1 and 0.5, respectively; U to 60% of control at 5; significant stimulation of proliferation was also seen at dose of 0.5 after treatments of 12 and 48 hr. Reference Bredfeldt etal.,2006 Mudipalli etal.,2005 Yanget al., 2007 Chromosomal Aberrations and/or Genetic Instability C-144 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells HaCaT cells Primary Syrian hamster embryo cells (HEC) Primary Syrian hamster embryo cells (HEC) Human peripheral lymphocytes Human peripheral lymphocytes Human primary peripheral blood lymphocytes Arsenic Species As111 SA As111 SA As111 SA Asv As111 SA Asv As111 SA Asv As111 SA Asv As111 SA Asv MMAm MMAV DMA111 DMAV Concentration(s) Tested (nM) 0.5, 1.0 0.5, 1.0 0.38,3.8,7.7 3.2, 8, 16, 32 7.7 32 0.77, 1.9 16,32 7.7 32 1.25-160 1.25-500 0.1-2.7 10-10000 0.11-12.26 10-10000 Duration of Treatment 20 passages 20 passages 24 hr for both 24 hr for both 48 hr for both 48 hr for both 24 hr for all LOECa (HM) 0.5 0.5 0.38 16 7.7 32 0.77 16 7.7 32 None None None 1000 0.34 1000 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Comparative genomic hybridization showed that all 1 1 cell lines derived from tumors (see malignant transformation) showed significant loss of chromosome 9q, and 7 lines showed significant gain of chromosome 4q. ft MN; detected using the cytokinesis-block micronucleus assay, and scored only in binucleated cells. There was a positive dose- response. SCEs were induced; slight upward trend with dose. CAs were induced: mostly chromatid gaps and breaks, but some chromatid and chromosome exchanges. SCEs were induced; dose-independent response. CAs were induced: mostly chromatid and chromosome gaps and breaks, very few exchanges. SCE/metaphase Top 3 in list were negative. Potency of others: DMA111 > DMAV > MMAV. All were weak inducers of SCE, with the most potent inducing ~1 SCE/metaphase/|aM. Reference Chien et al., 2004 Chien et al., 2004 Larramend y et al., 1981 Larramend y et al., 1981 Larramend y et al., 1981 Larramend y et al., 1981 Kligerman etal.,2003 C-145 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human primary peripheral blood lymphocytes Syrian hamster embryo cells CHO Kl cells in late Gl of mitotic cycle Human peripheral lymphocytes Human peripheral lymphocytes Arsenic Species As111 SA Asv MMAm MMAV DMA111 DMAV As111 SA Asv As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 1.25-160 1.25-500 0.1-2.7 10-10000 0.11-12.26 10-10000 0.8,3.0,6.2,10 10, 20, 64, 96 40 1, 5, 10 0.5, 1.0, 1.5,2.0 Duration of Treatment 24 hr for all 24 hr for both 4hr 48 hr 48 hr LOECa (HM) 2.5 50 0.6 3000 1.35 3000 6.2 64 40 1 2.0 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Chromosomal aberrations: ft to 42.5% aberrant cells at 10.0. ft to 11.0% aberrant cells at 80.0. ft to 11.0% aberrant cells at 1.2. ft to 6.5% aberrant cells at 3000. ft to 22.0% aberrant cells at 2.70. ft to 57.0% aberrant cells at 10000. All 6 showed a positive dose-response. Chromatid and isochromatid deletions were most prevalent; exchanges were infrequent. CAs and endoreduplication (also, with 48 hr treatment, polyploidy). Mainly chromatid gaps, breaks, and exchanges, but a few chromosome- type aberrations (fragments and dicentrics). High frequency of CAs was induced; effect was markedly reduced by prior or simultaneous (but not by subsequent) treatment with 5 mM GSH. Induction of chromatid aberrations; there was a positive dose-response. Induction of chromosomal aberrations. Reference Kligerman et al., 2003 Barrett et al., 1989 Huang et al., 1993 Jha et al., 1992 Wiencke and Yager, 1992 C-146 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue AS52 cells G12 cells CHO cells Arsenic Species As111 SA MMAmO DMAmI As111 SA Asv Concentration(s) Tested (nM) 50, 100 0.2, 0.4, 0.6, 0.8, 1.0 0.1,0.2,0.3,0.4 0.01,0.1, 1, 10 0.01,0.1, 1, 10, 100 Duration of Treatment 4hr 3 days for both 12 hr for both LOECa (HM) 100 0.6 0.3 1 100 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of gpt mutations at the ypt locus: The mutation frequency was twice that of the spontaneous mutation frequency at a high level of cytotoxicity (15% of the relative survival of the control). Taken as very weak evidence that As111 is a gene mutagen; results are grouped here with CAs because most or all of the induced mutations were total deletions of the gene, perhaps caused by the cytotoxicity. Induction of mutations at the gpt locus: DMAmI: reached 5x control mutant frequency at 7% cell survival; MMAmO: reached 5x control mutant frequency at 11% cell survival. Taken as weak evidence that the arsenicals are gene mutagens with sub- linear dose-responses; results are grouped here with chromosomal aberrations because -80% of the induced mutations were deletions of the gene, perhaps caused by the cytotoxicity. -11% of non-deletion mutants exhibited altered DNA methylation. Induction of chromosomal aberrations: A positive dose- response; 36.7% of cells with aberrations at dose of 10. 8.0% of cells with aberrations at dose of 100. Reference Meng and Hsie, 1996 Klein et al., 2007 Kochhar et al., 1996 C-147 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO cells CHO cells MRC-5 cells MRC-5 cells Arsenic Species As111 SA Asv As111 SA Asv As111 SA DMAV Concentration(s) Tested (nM) 0.01,0.1, 1, 10 0.01,0.1, 1, 10, 100 0.01,0.1, 1, 10 0.01,0.1, 1, 10 2.5, 5, 10 125, 250, 500 Duration of Treatment 12 hr for both 12 hr for both 26 hr 26 hr LOECa (HM) 1 None 0.01 0.01 2.5 125 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of endoreduplication: A positive dose- response; 22.0% of cells with endoreduplication at dose of 10. Induction of SCEs: 10.94%/cell at lowest dose; 14.08%/cell at highest dose; slight upward trend with dose. 11.38%/cell at lowest dose; 12.84%/cell at highest dose; no dose- response. Induction of SCEs (frequencies): 0,3.24; 2.5, 5.23; 5, 6.2; 10, no surviving cells could be found to evaluate. There was also much cytotoxicity at dose of 5. High level of cytotoxicity was also reflected in the proliferation index. Induction of SCEs (frequencies): 0, 4.25; 125, 5.89; 250, 5.95; 500, 5.91; thus no dose-response for SCEs. There was a significant U in the proliferation index at the highest dose. Reference Kochhar et al., 1996 Kochhar et al., 1996 Mouron et al., 2006 Mouron et al., 2005 C-148 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human lymphocytes Arsenic Species As111 SA Concentration(s) Tested (nM) 10"10, 10'8, 10'6, 10'4, 0.01 Duration of Treatment 24 hr LOECa (HM) 10-io Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of hypoploid and hyperploid cells: There was a statistically significant increase in hyperploidy at all dose levels in both 1st and 2nd division cells. There was a positive (but shallow) dose-response. For example, in 2nd division cells, the frequency went from 2.3% at dose of lO'10 to 11.7% at dose of 0.01. The 4 donors showed variation, with 2 showing no effect at lowest dose. It is unclear at what dose level induction of hypoploidy became significant, but there was a slight positive dose-response for it also. DataonCAs, which were reported only briefly, showed that roughly 40% of cells had CAs at the dose of 0.01. A concentration of 1 only uM was highly cytotoxic in these cells with an exposure lasting only 2 hr. Reference Vegaetal., 1995 C-149 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human lymphocytes Primary cultured human umbilical cord fibroblasts Arsenic Species As111 SA As111 SA Asv MMAV DMAV TMAV Concentration(s) Tested (nM) 0.001,0.01,0.1 0.8,2.3,3.8,7.7 16, 32, 64, 160, 321 1.4, 3.6, 7.1 mM 0.7, 1.4, 3.6 mM 3.7,7.6, 14.7 mM Duration of Treatment 24 hr 24 hr for all LOECa (HM) 0.001 0.8 uM 16 uM 1.4 mM 0.7 mM 3.7 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Increase in hyperdiploid frequency (based on FISH analysis, there was a statistically significant dose-related increase for each of the 2 chromosomes tested from both donors). There was also an increase in hypodiploid frequency, but it was only seen (again at all doses) in 1 of the 2 chromosomes tested and in only 1 donor. A related experiment showed that As111 can disrupt the microtubule organization of lymphocytes at a dose as low as 0.001. Induction of CAs: The percentages of abnormal cells at the LOECs for the 5 chemicals in descending order, as listed to the left, were: 10%, 16%, 19%, 28%, and 26%. Depletion of GSH by pretreatment of cells with B SO increased induction of CAs by As111 SA, Asv, andMMAvbut decreased it for DMAV. In cells pretreated with BSO before treatment with DMAV, the presence of 5 mM or higher GSH in the medium markedly increased induction of CAs. Since GSH does not enter the cells itself, this suggests that some clastogenic chemical is generated in the medium by interaction of DMAV with GSH. Reference Ramirez et al., 1997 Oya-Ohta etal., 1996 C-150 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO-9 cells CHO-9 cells Human primary peripheral blood lymphocytes CHO Kl cells Human peripheral lymphocytes Arsenic Species As111 SA Asv MMAm MMAV DMA111 DMAV TMAV As111 SA Asv MMA111 MMAV DMA111 DMAV TMAV As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 10 to 10000 for all 10 to 10000 for all 0.8 20 1, 5, 10 Duration of Treatment 30 minfor all 30 min 48 hr 6hr 48 hr LOECa (HM) 1000 1000 10 None 50 None None 1000 1000 10 None 50 None None 0.8 20 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of chromosomal aberrations: Aberrations consisted mainly of chromatid exchanges and breaks; dicentrics and rings occurred rarely. Frequencies of aberrations per 100 cells at the most effective concentration for the 4 positive chemicals ranged from 44 to 74x that of the control. Induction of SCEs: For even the most potent inducers of SCE, the number of SCEs/cell was less than double that of the untreated control; thus they were weak inducers. SCEs were induced; simultaneous treatment with SOD (an oxygen radical scavenger) blocked induction of SCEs. SCEs were induced; simultaneous treatment with squalene at from 40 to 160 uM significantly and dose-dependently inhibited induction of SCEs. SCEs were induced; there was a positive dose-response. Reference Dopp et al., 2004 Dopp et al., 2004 Nordenson and Beckman, 1991 Fanetal., 1996 Jha et al., 1992 C-151 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human peripheral lymphocytes CHO Kl cells Mouse lymphoma cells (L5178Y/Tk+A -3.7.2Ccells) Arsenic Species As111 ATO As111 SA MMAm DMA111 Concentration(s) Tested (nM) 0.00036, 0.00072, 0.0014 5, 10, 20, 40 0.19,0.28,0.38, 0.47, 0.52, 0.57 0.65,0.83, 1.29, 1.51 Duration of Treatment 24 hr 6hr 4 hr for both LOECa (HM) 0.00036 5 0.28 1.51 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) SCEs were induced; there was a positive dose-response; co- treatment with retinyl palmitate at the highest dose of As111 caused a significant U to a SCE frequency like that seen at the middle dose; the same thing also occurred for PDT and ACT, showing that retinyl palmitate also reversed some of the arsenic- induced decrease in the rate of cell proliferation. Induction of MN in binucleated cells, using cytochalasin B after arsenic treatment to block cytokinesis: simultaneous treatment with 80 uM squalene significantly reduced the effect. Mutations at Tk+/" locus in mouse lymphoma agar assay without exogenous metabolic activation: ft to 2.0x at 0.28, with a positive dose-response, reaching 7.2x at 0.57. ft 2.4x control at maximum concentration tested. Both compounds showed large excess of small colonies, which is indicative of chromosomal aberrations; generally similar results were found in a mouse lymphoma microwell assay, which was complicated by higher toxicity. Reference Avaniand Rao, 2007 Fan et al., 1996 Kligerman et al., 2003 C-152 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human peripheral lymphocytes SY-5Y cells HEK 293 cells Mouse lymphoma cells (L5178Y/Tk+/- -3.7.2Ccells) Arsenic Species As111 SA Asv MMAm MMAV DMAV TMAVO As111 ATO for all As111 SA Asv MMAV DMAV Concentration(s) Tested (nM) 0.5, 1,2,4 4, 8, 16, 32 0.01,0.05,0.1, 0.5, 1,2 50, 100, 250, 500 50, 100, 250 400, 800, 1000 1 for all 2.3,5.4,7.7,8.5, 10.8, 14.6, 16.2 3.0, 15.2, 30.3, 45.5, 60.6, 75.8, 84.9 6.2, 12.3, 15.4, 18.5,24.7,30.9 mM 12.5,25.0,37.5, 50.0, 56.3, 62.5 mM Duration of Treatment 72 hr for all 24hr 48 hr 72 hr 4hr for all LOECa (HM) 2 8 1 100 250 None 1 for all 8.5 45.5 18.5 mM 56.3 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of MN in binucleated lymphocytes detected by the cytokinesis-block assay (using cytochalasin B): ft in 2 donors at 2 and in all 3 donors at 4. ft in 1 donor at 8 and in all 3 donors at 2 higher doses. ft in 1 donor at 1 and in all 3 donors at 2. ft in 2 donors at 100 and 250 and in all 3 donors at 500. ft in 1 donor at 250. NSE. Further analysis of MMAm showed ftft in centromere-positive micronuclei (-80% of total), which is an indicator of induced aneuploidy. Induction of MN detected by Hoechst staining; response in comparison to control in SY-5Y and HEK 293 cells, respectively, for each duration of treatment: At24hr:3.70x, 3.35x. At48hr:5.14x, 4.81x. At 72 hr: 4.00x, 3.16x. Mutations at Tk+/" locus in mouse lymphoma agar assay without exogenous metabolic activation. Very few, if any, large colony mutants were induced by all compounds. Induction of small colony mutants is indicative of induction ofCAs. Reference Colognato et al., 2007 Florea et al., 2007 Moore et al., 1997a C-153 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Mouse lymphoma cells (L5178Y/Tk+/- -3.7.2Ccells) SHE cells V79 cells Arsenic Species As111 SA Asv MMAV DMAV As111 SA As111 SA Asv Concentration(s) Tested (nM) 11.5, 13.1, 15.4 60.6, 69.7, 84.9 21.6, 24.7, 27.8 mM 50.0, 56.3, 62.5 mM 4,6,8 50, 100, 250, 500 for both Duration of Treatment 4hr for all 24 hr Ihr for both LOECa (HM) 11.5 60.6 24.7 mM None 4 for first two effects 50 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of CAs: Aberrations consisted mainly of chromatid exchanges and breaks; all concentrations reported showed induction of CAs except for DMAV, which gave results of borderline significance that were considered negative by the authors. Lower frequencies of induction were seen for MMAV than for the inorganic arsenics in spite of the much higher doses. Induction of CAs: 0, 1%; 4, 9%; 6, 15%; 8, 32%. Induction of polyploidy and endoreduplication: 0, 0%; 4, 6%; 6, 19%; 8, 27%. Colony -forming efficiency relative to control after 7 days of culturing post- As treatment: 6, 77%; 8, 49%. MI: 0, 9.2; 4, 10.9; 6, 8.7; 8, 1.3. Induction of CAs (no. of aberrations in 100 metaphase cells): 0, ~7; 50, -49; 100, -99; 250, -120; 500, -160. 0, -6; 50, -32; 100, -44; 250, -62; 500, -73. Aberrations were mainly chromatid breaks. Co- treatment or pretreatment with tea extracts reduced aberration frequencies by half or more, while post- treatments also reduced the level of effects, which was suggestive of enhanced repair. Tea extracts induced CAT and SOD activity. Reference Moore et al., 1997a Hagiwara etal.,2006 Sinha et al., 2005a C-154 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue V79 cells Mouse lymphoma cells (L5178Y/Tk+A -3.7.2Ccells) Arsenic Species As111 SA Asv DMAV As111 SA Asv MMAV DMAV Concentration(s) Tested (nM) 50, 100, 250, 500 for all 11.5, 13.1, 15.4 60.6, 69.7, 84.9 21.6, 24.7, 27.8 mM 50.0, 56.3, 62.5 mM Duration of Treatment Ihr for all 4hr for all LOECa (HM) 100 or possibly 50 for all None 60.6 24.7 mM None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of micronuclei (MN) in cytochalasin B assay: (No. of MN per 1000 binucleated cells): 50, -105; 100, -110; 250, -170; 500, -300. 50, -80; 100, -105; 250, -125; 500, -150. 50, 52; 100, 70; 250, 99; 500, 111. Co-treatments with tea extracts reduced MN frequencies by two-thirds or more for As111 and by half or more for Asv and DMAV. Pretreatments with tea extracts also caused a large U in MN frequencies for all 3 arsenicals. Post- treatments also reduced MN frequencies, which was suggestive of enhanced repair. The polyphenols EGCG and TF extracted from tea had similar effects in reducing MN frequencies. The LOECs are uncertain because no data were reported for the untreated controls. Induction of MN in binucleated cells, using cytochalasin B after arsenic treatment to block cytokinesis: As111 SA gave results of borderline significance that were considered negative by the authors. Reference Sinha et al., 2005b Moore et al., 1997a C-155 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Don Chinese hamster cells CHO cells Human lymphocytes CHO cells P388DJ macrophage cell line Human peripheral lymphocytes Human peripheral lymphocytes Arsenic Species As111 SA Asv arsenic pent- oxide Asv disodium arsenate As111 SA Asv As111 SA As111 SA As111 SA Asv DMAV As111 SA DMAV As111 SA Concentration(s) Tested (nM) 7.7 13.9 32.1 1,5, 10 50, 80, 100 0.5, 1.0, 5.0 1, 10 0.01,0.1, 1 0.1, 1, 10 1,10 1 1 0.5, 1.0, 1.5,2.0 Duration of Treatment 28 hr for all 24 hr for all 48 hr 24 hr 48 hr for all 48 hr 48 hr 48 hr LOECa (HM) 7.7 13.9 32.1 1 50 0.5 1 None None None 1 1 1.0 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) SCEs were induced for all 3 chemicals at 1.56, 1.61, and 1.46 times the control level, respectively. The concentrations tested for SCEs for all 3 chemicals were the "50% inhibition doses" following culturing for 72 hours and using a Giemsa test for viability. Chromosome aberrations (breaks and exchanges) were induced by both compounds with a dose- response relationship; As111 was 5-10 times more effective than Asv per unit dose; 80 (oM was -50% growth inhibition dose over 4 days for As111. Chromosome aberrations (breaks and exchanges) were induced. SCEs were induced with a dose-response relationship. No more than slight hints of induction of SCEs under any of these experimental conditions. Induction of SCEs. Induction of SCEs. Induction of SCEs: in 2 of the 3 donors, the LOECwasl.5. Cells from one donor were more sensitive. Reference Ohno et al., 1982 Wan et al., 1982 Wan et al., 1982 Wan et al., 1982 Andersen, 1983 Andersen, 1983 Wiencke and Yager, 1992 C-156 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue BrdU- substituted replicating human lymphocytes GO human lymphocytes 2BS cells V79-C13 Chinese hamster cell line NB4 cells Arsenic Species As111 SA Asv As111 SA Asv As111 SA As111 SA As111 ATO Concentration(s) Tested (nM) 0.77, 1.54 13.5,26.9 1.54 26.9 1.0, 3.0, 5.0, 10 10 0.75 Duration of Treatment 24 hr 24hr 24hr 24hr 5hr 24 hr 3wk LOECa (HM) 0.77 13.5 None None 1.0 10 0.75 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of SCEs, but only in 2 of 4 subjects. Induction of SCEs, but only in 1 of 4 subjects; in 2 subjects (1 at lower dose and 1 at higher dose) there was a slight but significant decrease in SCEs. No induction of SCEs with either treatment. (4 subjects in each group.) DNA-protein crosslinks detected by alkaline elution; peak effect at 3.0; no effect at 10.0; further testing of DNA showed the crosslinks to be protein-associated DNA-strand breaks. Cells examined after 6, 12, 18, and 24 hr and after 6, 24, and 48 hr of recovery: by 6 hrof treatment there were multinucleated cells and round cells, by 12 hr there were giant cells. Multinucleated cells persisted at high levels to 48 hr after treatment. Also saw abnormal spindles and persistent (i.e., up to 5 days observed) aneuploidy and hyperdiploidy, but no statistically significant changes in CAs or MI. Enlarged cells were found that contained chromosomal end-to-end fusions. In 80 karyotypes, there were an average of 2.4 fusion events per cell, and 32 cells had polyploidy. FISH analysis showed that fusions are associated with attrition of telomeres. Reference Crossen, 1983 Crossen, 1983 Dong and Luo, 1993 Sciandrello et al., 2002 Chou et al., 2001 C-157 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells HeLa cells PARP-1+/+ MEF cells PARP-I-'- MEF cells PARP-1+/+ MEF cells PARP-I-'- MEF cells PARP-1+/+ MEF cells PARP-1"'" MEF cells Arsenic Species As111 ATO for both As111 SA for both As111 SA for both As111 SA for both Concentration(s) Tested (nM) 0.25 1 11.5,23 for both 11.5,23 for both 11.5,23 for both Duration of Treatment 4, 5, 6 wk 3,4wk 24 hr for both 48 hr for both 24 hr for both LOECa (HM) 0.25 1 None 11.5 23 11.5 11.5 11.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Southern blot of digested genomic DNA: U telomere length at all 3 time points. U telomere length at both time points. Telomere length measured by flow FISH assay (point estimate comparisons were made to unexposed cells of the same genotype): -98% of control at 11. 5, -91% of control at 23; both are NSEs. -76% of control at 11. 5, -71% of control at 23. Telomere length measured by flow FISH assay (point estimate comparisons were made to unexposed cells of the same genotype): -99% of control at 11. 5, -79% of control at 23; the one at 11.5 was NSE. -79% of control at 11. 5, -41% of control at 23. inorganic arsenic- induced telomere attrition was thus much greater in PARP-l"7" MEFs. Induced (experimental - control) % of MN in binucleated cells (with cytochalasin B post- treatment to block cytokinesis): -4% at 11. 5, -5% at 23. -18% at 11.5, -13% at 23. Reference Chou et al., 2001 Poonepalli et al., 2005 Poonepalli etal.,2005 Poonepalli et al., 2005 C-158 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PARP-1+/+ MEF cells PARP-r7' MEF cells PARP-1+/+ MEF cells PARP-r7' MEF cells PARP-1+/+ MEF cells PARP-I-'- MEF cells Arsenic Species As111 SA for both As111 SA for both As111 SA for both Concentration(s) Tested (nM) 11.5,23 for both 11.5,23 for both 11.5,23 for both Duration of Treatment 48 hr for both 24 hr for both 48 hr for both LOECa (HM) 11.5 11.5 None 11.5 None 11.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced (experimental - control) % of MN in binucleated cells (with cytochalasin B post- treatment to block cytokinesis): -6% at 11. 5, -6% at 23. -27% at 11. 5, -15% at 23. Induced (experimental - control) frequency of CAs per cell, using FISH with a telomeric PNA probe: -0.04 at 11.5, -0.04 at 23;bothareNSEs. -0.09 at 11. 5, -0.05 at 23; only the one at 11.5 was statistically significant. CAs included end-to-end fusions, chromosome breaks, and fragments. Induced (experimental - control) frequency of CAs per cell, using FISH with a telomeric PNA probe: -0.04 at 11. 5, -0.04 at 23;bothareNSEs. -0.11 at 11. 5, -0.03 at 23; only the one at 11.5 was statistically significant. CAs included end-to-end fusions, chromosome breaks, and fragments. Reference Poonepalli etal.,2005 Poonepalli et al., 2005 Poonepalli etal.,2005 C-159 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue 293 cells 293 cells Arsenic Species As111 ATO MMAm Concentration(s) Tested (nM) 2 2 Duration of Treatment 24 hr 24 hr LOECa (HM) 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) No. MN/1000 binucleated cells (with cytochalasin B during treatments to block cytokinesis): untreated = -35; dose of 2: big ft to -260. Effects of co-treatment (CoTr) with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -155. CoTr 100 uMDMPS:U from inorganic arsenic alone to -170. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -605. CoTrlOuMDMPS:ft from inorganic arsenic alone to -670. No. MN/1000 binucleated cells (with cytochalasin B during treatments to block cytokinesis): untreated = -35; dose of 2: big ft to -230. Effects of CoTr with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -130. CoTr 100 uMDMPS:U from inorganic arsenic alone to -155. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -465. CoTrlOuMDMPS:ft from inorganic arsenic alone to -470. Reference Jan et al., 2006 Jan et al., 2006 C-160 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue 293 cells SV-HUC-1 cells Arsenic Species DMA111 As111 ATO Concentration(s) Tested (nM) 2 2 Duration of Treatment 24 hr 24 hr LOECa (HM) 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) No. MN/1000 binucleated cells (with cytochalasin B during treatments to block cytokinesis): untreated = -35; dose of 2: big ft to -315. Effects of CoTr with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -170. CoTr 100 uMDMPS:U from inorganic arsenic alone to -175. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -630. CoTrlOuMDMPS:ft from inorganic arsenic alone to -635. No. MN/1000 binucleated cells (with cytochalasin B during treatments to block cytokinesis): untreated = -35; dose of 2: big ft to -330. Effects of CoTr with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -150. CoTr 100 uMDMPS:U from inorganic arsenic alone to -150. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -680. CoTrlOuMDMPS:ft from inorganic arsenic alone to -645. Reference Jan et al., 2006 Jan et al., 2006 C-161 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SV-HUC-1 cells SV-HUC-1 cells Arsenic Species MMAm DMA111 Concentration(s) Tested (nM) 2 2 Duration of Treatment 24 hr 24 hr LOECa (HM) 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) No. MN/1000 binucleated cells (with cytochalasin B during treatments to block cytokinesis): untreated = -35; dose of 2: big ft to -270. Effects of CoTr with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -145. CoTr 100 uMDMPS:U from inorganic arsenic alone to -150. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -570. CoTrlOuMDMPS:ft from inorganic arsenic alone to -470. No. MN/1000 binucleated cells (with cytochalasin B during treatments to block cytokinesis): untreated = -35; dose of 2: big ft to -400. Effects of CoTr with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -160. CoTr 100 uMDMPS:U from inorganic arsenic alone to -145. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -620. CoTrlOuMDMPS:ft from inorganic arsenic alone to -650. Reference Jan et al., 2006 Jan et al., 2006 C-162 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SHE cells Primary rat hepatocytes CL3 cells, synchronous atGl CL3 cells, asynchronous (asyn) CL3 cells, synchronous atG2/M Arsenic Species As111 SA DMAmI As111 SA As111 SA for all Concentration(s) Tested (nM) 3, 10 0.5, 1.0 0.25,0.5, 1,2.5,5, 7.5, 10 50 for all Duration of Treatment 48 hr for both 27 hr 3hr for all LOECa (HM) 10 0.5 1 50 for all Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Aneuploidy detected by flow cytometry: Slight It Slight ft; iWl at 1.0. Other experiments showed that DMAmI caused abnormalities of mitotic spindles, centrosomes, and microtubule elongation. Induction of MN (mean no./lOOO cells): 17.4 at dose of 1, increasing with dose to 24. 4 at dose of 7.5; control = 13.7; too many cells were dead at dose of 10 to evaluate this endpoint. Co-treatment withlOor25uMSbmCl: 0 in micronucleus frequency below expectation of an additive interaction; that chemical also induced MN. Induction of MN; inorganic arsenic treatment was followed by culturing with cytochalasin B for 24 hr to block cytokinesis): induced no. of MN (experimental - controiyiOOO binucleated cells: Gl, -181; asyn, -141; G2/M, -125; when Gl cells were co-treated with inorganic arsenic and either PD98059 or UO 126, this number U from -181 to -75-80. Percentages of binucleated cells: Gl, 14%; asyn, 47%; G2/M, 39%. Reference Ochi et al., 2004 Hasgekar etal.,2006 Li et al., 2006a C-163 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CL3 cells, synchronous atGl V79-C13 Chinese hamster cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 50 10 Duration of Treatment 3hr 24 hr LOECa (HM) 50 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of MN; inorganic arsenic treatment was followed by culturing with cytochalasin B for 24 hr to block cytokinesis): induced frequency = -18 1/1000 binucleated cells (as in row above); percentage of binucleated cells: 14% (as in row above). Culturing of Gl cells with cytochalasin B for 36-48 hr (instead of 24) caused marked ft in percentages of binucleated cells and marked U in induced numbers of MN (1000 binucleated cells) from 181 to -40-70. Also, when cultured with cytochalasin B for 40 hr (instead of 24 hr) after the co-treatment of inorganic arsenic with PD98059orU0126, these 2 structurally dissimilar inhibitors of MEK1/2 caused no further U from inorganic arsenic alone. After being expanded through 120 generations in the absence of arsenic and then being cloned, acquired genetic instability persisted and often came to include dicentric chromosomes and telomeric associations. These same clones, which were often aneuploid, micronucleated and/or multinucleated, were affected by the DNA hypomethylation that was seen globally in the cells immediately after the 24-hr treatment. Reference Li et al., 2006a Sciandrello etal.,2004 C-164 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO cell lines: Kl (parental to the following line) XRS-5 (X-ray andH2O2 sensitive) CHO cell lines: Kl (parental to the following lines) XRS-6 (X-ray sensitive) XRS-5 (X-ray andH2O2 sensitive) Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 10, 20, 40, 80 for both 20, 40, 80 20, 40, 60 10, 20, 30, 40, 60 Duration of Treatment 4 hr for both 4 hr for all LOECa (HM) 80 10 40 20 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of MN in binucleated cells, using cytochalasin B after arsenic treatment to block cytokinesis: the much less responsive Kl cells have 6 times as much catalase activity as XRS-5 cells; both lines are similar in arsenic uptake and release, in GSH levels, and in GSH S-transferase activity. Frequencies of MN per thousand binucleated cells per uM of arsenic for Kl, XRS-6, and XRS-5 cells were 2.1, 4.5, and 10.8, respectively. (Cytochalasin B was used after arsenic treatment to block cytokinesis.) Kl cells have 5.8 times as much catalase activity and 5.4 times as much GPx activity as XRS-5 cells. Kl cells have 3.7 times as much catalase activity and 2. 1 times as much GPx activity as XRS-5 cells. The cells with intermediate amounts have an intermediate response. Co-treatment of XRS-5 cells with catalase or GPx eliminates induction of MNbyAsmSA. Treatment of Kl cells with inhibitors of catalase and GPx makes them much more sensitive to induction of MNbyAsmSA;when co-treated together, there is a synergistic effect. Reference Wang and Huang, 1994 Wanget al., 1997 C-165 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO-9 cells HFW cells HLFC cells HLFK cells (Ku70 deficient) Arsenic Species As111 SA Asv MMAm MMAV DMAV DMA111 TMAV As111 SA for both durations As111 SA for both Concentration(s) Tested (nM) 1, 5, 10, 50, 100, 500 for both 1, 5, 10, 30 1, 5, 10, 30, 100, 500, 5000 for both 1,5,10 1, 5, 10, 5000 1.25,2.5,5, 10 5, 10, 20, 40, 80 1,2.5,5, 10 for both Duration of Treatment Ihr for all 24hr 4hr 24 hr for both LOECa (HM) None None 10 5000 5000 1 5000 1.25 10 2.5 2.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of MN in binucleated cells: DMA111 was by far the most potent. Induction of MN, with about 70% being kinetochore-positive at maximum induction found at dose of 5. Induction of MN, with about 70% being kinetochore-negative at maximum induction found at dose of 40. Induction of micronuclei (% of cells with MN): Control, 5%; 1, 4%; 2.5, 8%; 5, 10%, 10, 15%. Control, 4%; 1, 6%; 2.5, 10%; 5, 21%, 10,27%. At the 2 higher doses the % is significantly higher in the HLFK cells. Ku70 is 1 of 3 subunits of DNA-dependent protein kinase, and the Ku70 protein plays an important role in repair of DNA double-strand breaks. Reference Dopp et al., 2004 Yihand Lee, 1999 Liuetal., 2007b C-166 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HLFC cells HLFK cells (Ku70 deficient) HFF cells HL-60 cells HaCaT cells Arsenic Species As111 SA for both As111 SA As111 SA Concentration(s) Tested (nM) 1,2.5,5, 10 for both 5 0.5, 10, 20 Duration of Treatment 24 hr for both 24 hr 3 days LOECa (HM) 5 2.5 5 10 for U 0.5 for ft, 10 U Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Formation of abnormal nuclei (% of cells with abnormal nuclei): Control, 7%; 1, 9%; 2.5, 10%; 5, 19%, 10, 23%. Control, 10%; 1, 12%; 2.5, 21%; 5, 37%, 10, 42%. At the 3 higher doses the % is significantly higher in the HLFK cells. Ku70 is 1 of 3 subunits of DNA- dependent protein kinase, and the Ku70 protein plays an important role in repair of DNA double-strand breaks. cen+ and cen- MN induced per 1000 cells: cen- MN: control, -10/1000; inorganic arsenic, -17/1000. cen+ MN: control, -2/1000; inorganic arsenic, -18/1000. Co-treatment with 170 nM SAM essentially eliminated induction of cen+ MN without having any effect on induction of cen- MN. Analysis of telomere length by TRF analysis using Southern blot assay: Telomeres were shortened compared to controls at 10 and 20. Telomeres were shortened compared to controls at 10 and 20, but in these cells only, the telomeres were slightly elongated at dose of 0.5. Reference Liuetal., 2007b Ramirez et al., 2007 Zhang et al., 2003 C-167 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human- hamster hybrid AL cells Human- hamster hybrid AL cells Human- hamster hybrid AL cells Arsenic Species As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 3.8,7.7, 15.4 11.5, 15.4 3.8 Duration of Treatment 1 day or 5 days 24 hr 24 hr LOECa (HM) Depends on locus 11.5 3.8 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of mutations at both loci, with both showing higher response after 5 -day treatment than after 1-day treatment. After only 1 day of treatment, the LOECs were 3.8 at SI locus and 15.4 at the HPRT locus. This effect is not grouped with gene mutations because most mutations were large deletions; about 28 times as many mutations occurred at the SI locus, and co-treatment with DMSO eliminated most of the mutation induction. Induction of mutations at CD59 locus (formerly known as SI locus); this effect is not grouped with gene mutations because most mutations were large multilocus deletions; co-treatment with SOD or catalase considerably reduced mutation induction. Induction of mutations at CD59 locus; pretreatment with BSO (to reduce GSH levels) increased mutation rate about 3 -fold. Reference Heietal., 1998 Kessel et al., 2002 Kessel et al., 2002 C-168 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Enucleated AL hybrid cells treated with As111 were fused with untreated nuclei to form reconstituted AL hybrid cells AL hybrid cells made highly deficient in mitochondria! DNA by long- term ditercalinium treatment; then called p° cells Human- hamster hybrid AL cells Arsenic Species As111 SA As111 SA As111 SA for both Concentration(s) Tested (nM) 15.4 7.7, 11.5, 13.5, 15.4 1.9,3.8,7.7 for both Duration of Treatment 3hr 18 hr 16 days 30 days LOECa (HM) 15.4 None 1.9 1.9 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of mutations at CD59 locus: Mutant frequency >2x the frequency in control cells reconstituted from untreated enucleated cells and untreated nuclei. Induction of ROS was demonstrated in inorganic arsenic- treated enucleated cells by using a fluorescent probe. These results suggest that mitochondria may be essential for induction of CD59" mutations (in nuclear DNA). No increase in CD59" mutations; there was a dose-related increase in cytotoxicity. Analysis of DNA showed that mtDNA was >95% depleted in the p° cells. Suggests that mitochondria! function may be necessary for induction of CD59" mutations by inorganic arsenic. Induction of mutations at the CD59" locus: increase in mutation frequency at all doses, with a positive dose- response and at least a doubling of the control frequency at the higher dose. These cells showed a dose-related increase in cytotoxicity, with never less than a 60% surviving fraction. After a 60-day exposure, there was an almost 3- fold increase in the number of MN observed over the untreated control, but details were not provided. Reference Liu et al., 2005 Liu et al., 2005 Partridge etal.,2007 C-169 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human- hamster hybrid AL cells Arsenic Species As111 SA Concentration(s) Tested (nM) 0.8,3.8,7.7,15.4 Duration of Treatment 24 hr LOECa (HM) 3.8 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of CD59 mutants: (Addition of BSD, which suppresses GSH, increased mutant frequencies more than 5- fold.) Reference Liuetal., 2001 Co-carcinogenesis Rat lung epithelial cell line Rat lung epithelial cell line exposed to 100 nM B[a]Pfor24 hr As111 SA for both 1.5 for both 12 wk without theB[a]P treatment or immediately following that treatment 1.5 for both Transformation (i.e., anchorage-independent growth in soft agar) occurred with 12-wk inorganic arsenic treatment alone or with B[a]P treatment alone. There was a synergistic interaction when the B[a]P treatment was followed by the 12-wk inorganic arsenic treatment, with the transformation rate then exceeding 500 and 200 times that of the inorganic arsenic or B[a]P treatments alone, respectively. Lauand Chiu, 2006 C-170 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Rat lung epithelial cell line Rat lung epithelial cell line exposed to 100 nM B[a]Pfor24 hr Arsenic Species As111 SA for both Concentration(s) Tested (nM) 1.5 for both Duration of Treatment 12 wk without theB[a]P treatment or immediately following that treatment LOECa (HM) 1.5 for both Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Changes in the proteome of the transformed cells detected by MALDI- TOF-MS analysis and other methods: inorganic arsenic andB[a]P treatments alone caused changes in most of the following proteins alone. The combined treatment often caused a synergistic interaction on the protein levels in the same direction as one or both treatments changed them alone. Affected proteins were as follows: 3 proteins belonging to intermediate filaments were down-regulated; 6 proteins belonging to antioxidative stress-, chaperone-, and glycolytic proteins were up-regulated. Also phosph-ERKl/2 and a- actinin, which are associated with promotion of cell proliferation and de- differentiation, were up- regulated. Reference Lauand Chiu, 2006 C-171 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue GM04312C cells Arsenic Species As111 SA Concentration(s) Tested (nM) 10,50 Duration of Treatment 24 hr LOECa (HM) 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) BPDE-DNA adducts were measured after a 30-min treatment with O.SuMBPDEthat followed the inorganic arsenic pretreatment. Compared to no pretreatment, increases in these adducts at the doses of 10 and 50 were 1.4x and 1.6x, respectively. In these NER-deficient cells, which could be used to dissect induction of DNA damage from DNA repair, it was shown that inorganic arsenic markedly increased the cellular uptake of BPDE in a dose-dependent manner. It was concluded that this effect contributes to the co- carcinogenesis in addition to arsenic's "well demonstrated inhibitory effect on DNA repair." Reference Shen et al., 2006 Co-mutagenesis E. coli WP2 irradiated with 5. 6 J/m2 of U Von plates that contained: As111 SA Asv 100, 250, 500, 750 100, 300, 500 — 100 None Plating protocol for Trp+ revertants: synergistic interaction in inducing Trp+ revertants at lower 3 dose levels for SA only, with peak effect at 250; synergistic interaction was seen only in a strain of E. coli that can carry out excision repair of pyrimidine dimers. FourE1. coli strains that did not meet that criterion were tested, with no synergism being seen. Rossman, 1981 C-172 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO Kl cells in late Gl of mitotic cycle exposed to 7 J/m2ofUV CHO cells exposed to 1, 2, 4, or 8 J/m2 ofUV CHO cells exposed to 1, 2, 4, or 8 J/m2 ofUV Human peripheral lymphocytes simultaneousl y treated with 6uMDEB Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 40 5, 10 5, 10 0.5, 1.0, 1.5,2.0 Duration of Treatment 2hr 24 hr 24 hr 48 hr LOECa (HM) 40 5 None 1.0 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) High frequency of chromosome aberrations was induced; effect was markedly reduced by prior or simultaneous (but not by subsequent) treatment with GSH. Induction of chromosomal aberrations: synergistic interaction was demonstrated at all dose levels of UVand inorganic arsenic except for 1 J/m2 with the 10 uM inorganic arsenic treatment. At other UV dose levels, the responses at 10 uM arsenic only slightly exceeded those at 5 uM. UV or inorganic arsenic alone induced mainly chromatid-type aberrations, but in cells treated with both agents there was an apparent increase of chromatid breaks, chromatid exchanges, chromatid gaps, and chromosome breaks. Induction of SCEs: no statistically significant effect of the inorganic arsenic treatment was observed. Induction of chromosomal aberrations: there was synergistic interaction between DEB and inorganic arsenic. Reference Huang et al., 1993 Lee et al., 1985 Lee et al., 1985 Wiencke and Yager, 1992 C-173 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human peripheral lymphocytes simultaneousl y treated with 6uMDEB CHO cells exposed to 2 or 4 J/m2 of UV CHO cells exposed to 2 or 4 J/m2 of UV CHO Kl cells exposed to 1.5 or 2. 5 J/m2ofUV CHO cells treated with MMS before or after inorganic arsenic treatment CHO cells treated with MMS before or after inorganic arsenic treatment Arsenic Species As111 SA As111 SA As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 0.5, 1.0, 1.5,2.0 5, 10 5, 10 10 10, as pretreatment 10, as posttreatment 5, 10, as pretreatments 5, 10, as posttreatments Duration of Treatment 48 hr 24 hr 24 hr 24 hr 24hr 24hr 24 hr 24 hr LOECa (HM) -1.0 5 5 10 10 10 None 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of SCEs: Unlike with CAs, there was not a synergistic interaction. Although no statistical comparisons were presented, the trends suggested additivity between the two mutagens. Induction of gene mutations to 6- thioguanine resistance: synergistic interaction was demonstrated at both dose levels of UV and inorganic arsenic. Induction of gene mutations to ouabain resistance: inorganic arsenic had no effect. Induction of 6-TGr gene mutations at the HPRT locus: synergistic interaction was demonstrated at both dose levels of UV; inorganic arsenic at doses of 10 to 40 had no effect on the mutation frequency by itself. Induction of gene mutations at the HGPRT locus: U compared to MMS alone. ft compared to MMS alone, synergistic interaction. Induction of chromosomal aberrations: No change from MMS alone. ft frequency compared to MMS alone, synergistic interaction with even bigger effect at 10. Reference Wiencke and Yager, 1992 Lee et al., 1985 Lee et al., 1985 Yanget al., 1992 Lee et al., 1986 Lee etal., 1986 C-174 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO cells treated with MMS before or after inorganic arsenic treatment Human peripheral lymphocytes VH16 cell line (human primary fibroblasts) exposed to 7.5 J/m2ofUV V79 cells treated with MNU Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 5, 10, as pretreatments 5, 10, as posttreatments 5 5 10 5 Duration of Treatment 24 hr 24 hr 2 hr before X-rays, 30 min after X-rays 24 hr 3hr 24 hr LOECa (HM) None None 5 5 10 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of SCEs: No change from MMS alone. No change from MMS alone. Synergistic interaction in causing dicentrics and rings in both donors; synergistic interaction in causing deletions in one of the donors and approximately an additive response in the other; doses of X-rays were 1 Gy or 2 Gy with the dose rate unspecified. inorganic arsenic exposure increased the frequencies of MN in binucleated cells and of SCEs over what they would have been with UV alone, but there was not a synergistic effect forMN. Induction of gene mutations at the HPRT locus: While neither inorganic arsenic treatment induced mutations by itself, as a post-treatment these inorganic arsenic treatments both caused an ft in the mutation frequency compared to MNU alone; there was a synergistic interaction. Reference Lee et al., 1986 Jha et al., 1992 Jha et al., 1992 Li and Rossman, 1989a C-175 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue V79 cells exposed to 5- 15 J/m2 of UVC V79 cells exposed to 55-220 KJ/m2 ofUVA V79 cells exposed to: 400-800 J/m2 ofUVB 200 J/m2 of UVB Mouse 291.03C keratinocytes irradiated immediately after the arsenic treatment with a single dose of0.30kJ/m2 UV Arsenic Species As111 SA As111 SA As111 SA for both As111 SA Concentration(s) Tested (nM) 10 10 10 5, 10, 15 2.5, 5.0 Duration of Treatment 3hr 3hr 3hr 24 hr 24 hr LOECa (HM) 10 10 None 10 5.0 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of gene mutations at the HPRT locus: While the inorganic arsenic treatment induced no mutations by itself, as a post-treatment it caused an ft in the mutation frequency compared to UVC irradiation alone; there was a synergistic interaction. Induction of gene mutations at the HPRT locus: While the inorganic arsenic treatment induced no mutations by itself, as a post-treatment it caused an ft in the mutation frequency compared to UVA irradiation alone; there was a synergistic interaction. Induction of gene mutations at the HPRT locus: While the inorganic arsenic treatments induced no mutations by themselves, the 24-hr post-treatment caused an ft in the mutation frequency compared to UVB irradiation alone; there was a synergistic interaction. Effect on repair rate of UV-induced photodamage to genomic DNA measured at 2 and 6 hr after the UV exposure ended: U in repair rate of 6-4PPs by 48%, but no effect on the repair of CPDs. Reference Li and Rossman, 1991 Li and Rossman, 1991 Li and Rossman, 1991 Wuetal., 2005 C-176 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TK6 cells TK6 cells irradiated with 1 or c3 Gyof69 cGy/min gamma radiation at beginning of inorganic arsenic treatment Arsenic Species As111 SA As111 ATO As111 SA As111 ATO Concentration(s) Tested (nM) 0.1, 1, 10 for both 0.1,1, 10 for both Duration of Treatment 24 hr for both 24 hr for both LOECa (HM) 10 10 1 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of MN using flow cytometry assay: ft to 24.7% from 3. 4% in control. ft to 17.4% from 3. 4% in control; the text noted that it was sometimes difficult to distinguish between the MN and necrotic cell fragments due to toxicity at the dose of 10 for SA and ATO. Induction of MN using flow cytometry assay: At dose of 1: 1 Gy, 10.2%; 3 Gy, 12.2%; 12.2% was significantly higher than 9. 8% in control. There was a statistically significant (additive) effect. At dose of 1: 1 Gy, 10.0%; 3 Gy, 16.3%; 16.3% was significantly higher than 9. 8% in control. There was a statistically significant (possibly slightly synergistic) effect. Interpretation of results at dose of 10 was complicated by difficulty of distinguishing micronuclei and necrotic cell fragments. Responses were extremely different for the 2 arsenicals at dose of3Gy:30.2%forSA and only 15. 9% for ATO. Reference Hornhardt etal.,2006 Hornhardt etal.,2006 Cytotoxicity NHEK cells As111 SA Asv, MMAV, DMAV 0.001,0.005,0.01, 0.05,0.1,0.5, 1,5, lOforall 24 hr 24hr 0.005 0.5 Extent of viability determined by neutral red assay; viability was significantly reduced. Vegaetal., 2001 C-177 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HOS cells AG06 cells W138 cells WI38 cells HaCaT cells HepG2 cells 17 human cancer cell lines: 4 bladder cell lines, 2 lung cell lines, 2 liver cell lines, 1 leukemia cell line, and various others Arsenic Species As111 SA, Asv As111 SA As111 SA Dimethyl- arsinate, the usual form of DMAvin this table Thio- DMAV (i.e., Thio- dimethyl- arsinate) As111 ATO Concentration(s) Tested (nM) IC50 determinations 0.25,0.5, 1,2 0.5, 1.0 0.01,0.1,0.5, 1,5, 10, 50 mM for both IC50 determinations Duration of Treatment 100 hr for all 7 days 20 passages 48 hr for both 96 hr LOECa (HM) — 0.25 0.5 0.5 mM 0.1 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Extent of viability determined by neutral red assay: IC5os:3.5forAsin, 11 for Asv IC50s: 1.1 for As111, 16 for Asv IC50s: 8.8 for As111, 30 for Asv Clonal survival determined by crystal violet assay: LD50: -1.85. ft resistance to cytotoxicity caused by exposure to concentrations of As111 of !-16|aMfor72hr. Cell survival was determined by WST-8 assay: LC50s: regular DMA, -0.2 mM; Thio-DMA, -0.02 mM. At 0.1 mM, regular DMA showed no cytotoxicity, but thio- DMA resulted in only 22% cell survival. Viability determined by sulphorhodamine B method: Bladder: IC50s: 0.34, 0.47,0.93, 1.38. Lung: :IC50s: 3.27, 4.17. Liver:IC50s:5.17, 7.17. Leukemia: IC50s: 0.64. All 17 lines: LC50 range was 0.34-7. 17. There was a strong positive correlation between GSH content of cells and magnitude of IC50: Reference Hu et al., 1998 Vogtand Rossman, 2001 Chien et al., 2004 Ramlet al., 2007 Yanget al., 1999 C-178 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue 4 of above 17 human cancer cell lines with high levels of GSH Hepa-1 cells (mouse hepatoma) NHEK cells AG06 cells Human- hamster hybrid AL cells Primary cultures of rat cerebellar neurons Arsenic Species As111 ATO As111 SA As111 SA As111 SA As111 SA As111 SA DMAV Concentration(s) Tested (nM) IC50 determinations 2, 5, 10, 25, 50 IC50 determinations 0.1,0.3, 1,3 0.8,3.8,7.7, 15.4 5, 10, 15 1, 5, 30 mM Duration of Treatment 96 hr 12 hr 24 hr 72hr 48hr 24hr 12 hr 48 hr LOECa (HM) — None 10 — 3 3 0.3 0.1 3.8 5 5 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Viability determined by sulphorhodamine B method: 10 |oM BSD, which depletes cellular GSH, was incubated with cells for 4 days, causing them all to become very sensitive to arsenic, as follows: IC50s without BSD: 0.47, 2.59, 2.08, 9.89. IC50s with BSD: 0.19, 0.14,0.40,0.20, respectively. Viability determined by LDH release method Extent of viability determined by neutral red assay: IC50: 10.8 Extent of viability determined by neutral red assay: Values below at 3 : -90% of cells viable if no pretreatment (pt) to change GSH level. -85% of cells viable if NACpt to ft GSH level. -20% of cells viable if BSD pt to U GSH level. -20% of cells viable if CHE pt to U GSH level. No. of colonies counted to determine surviving fraction: LC50 = about 7.7. (Addition of BSD, which suppresses GSH markedly, increased cytotoxicity.) Viability determined using MTT metabolism assay. Reference Yanget al., 1999 Maier et al., 2000 Snow et al., 1999 Snow et al., 1999 Liuetal., 2001 Namgung and Xia, 2001 C-179 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Chang human hepatocytes Chang human hepatocytes Chang human hepatocytes Raji cells (human B- lymphocytes) Jurkat cells A549 cells Arsenic Species As111 SA, Asv, MMAm, MMAV, DMAV As111 SA, Asv, MMAm, MMAV, DMAV As111 SA, Asv, MMA111, MMAV, DMAV As111 SA MMA111 DMA111 As111 SA MMA111 DMA111 As111 SA Asv DMAV Concentration(s) Tested (nM) LCso determinations LCso determinations LCso determinations 0.2, 1, 10, 20, 40, 100 for all 0.2, 1, 10, 20, 40, 100 for all 0.016, 0.08, 0.4, 2.0, 10 30, 100, 300 2, 20, 200, 2000 Duration of Treatment 24hr 24 hr 24 hr 4hr 4hr 2hr 4hr 4hr 2hr 7 days for all LOECa (HM) — — — 10 40 10 40 0.2 10 0.016 30 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) LC50s using LDH leakage assay in phosphate media: As111: 68.0. Asv: 1,628. MMA111: 6.0. MMAV: 8,235. DMAV: 9,125. LC50s using K+ leakage assay in phosphate media: As111: 19.8. Asv: 1,006. MMA111: 6.3. MMAV: 9,283. DMAV:4,109. LC50s using the XTT assay in phosphate media: As111: 164. Asv: 3,050. MMA111: 13.6. MMAV: 42,000. DMAV: 91,440. Extent of viability determined by trypan blue assay: Viabilities at maximum dose for each: As111: -85%. MMA111: -85%. DMA111: 60%. Extent of viability determined by trypan blue assay: Viabilities at maximum dose for each: As111: -95%. MMA111: -52%. DMA111: -58%. Colony-forming efficiency assay with Giemsa staining: LC50s: As111, -0.08; Asv, -100. Reference Petrick et al., 2000 Petrick et al., 2000 Petrick et al., 2000 Gomez et al., 2005 Gomez et al., 2005 Mass and Wang, 1997 C-180 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO Kl cells CHO cells: Wild-type V850 R120 CHO Kl cells CHO cell lines: Kl (parental to the following line) XRS-5 (X-ray and H2O2 sensitive) Arsenic Species As111 SA As111 SA for all As111 SA As111 SA Concentration(s) Tested (nM) 10 5, 10, 15, 20, 30, 50, 75, 100 for most 20, 40, 80 10, 20, 40 for both Duration of Treatment 4hr 48 hr for all 4hr 4hr LOECa (HM) None 5 20 10 (lowest for it) 20 40 20 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Clonogenic survival assay for cytotoxicity: 12-hr pretreatment with BSO depletes GSH; with BSO at 50 and 400 uM, survival was 9% and 1%, respectively; other experiments showed that an increase in GSH markedly reduced the cytotoxicity of an As111 treatment following UV irradiation. Comparative inhibition of cell growth was based on numbers of cells present compared to control: V 850 cells were adapted to 850 uM H2O2 over about 4 months of exposures to increasing concentrations; R 120 cells had then been cultured 4 months without exposure to H2O2. IC50 values: Wild- type, 17.2; V 850, 62.45; R 120, 26.6. Results after pretreatment with BSO suggest that intracellular thiol levels (GSH mainly) may account for the arsenic resistance seen in V 850 cells. Colony formation assay Clonogenic survival with crystal violet staining: ID50s: line Kl, 37.8; line XRS-5, 17.0; the much less responsive Kl cells have 6 times as much catalase activity as XRS-5 cells; both lines are similar in arsenic uptake and release, in GSH levels, and in GST activity. Reference Huang et al., 1993 Cantoni et al., 1994 Wanget al., 1996 Wang and Huang, 1994 C-181 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO cell lines: Kl (parental to the following lines) XRS-6 (X-ray sensitive) XRS-5 (X-ray andH2O2 sensitive) CHO-9 cells BFTC905 cells and NTUB1 cells CHO Kl cells Arsenic Species As111 SA As111 SA Asv MMAm MMAV DMA111 DMAV TMAVO As111 SA Asv MMA111 MMAV DMA111 DMAV As111 SA Concentration(s) Tested (nM) 20, 40, 80, 160 20, 40, 80, 160 20, 40, 80, 160 0.1, 1, 10, 100, 500 for all IC50 determinations 20 Duration of Treatment 4hr Ihr for all 7 days 6hr LOECa (HM) 160 80 20 1 1 500 100 0.1 500 None — 20 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Inhibition of cell growth: ID50 values were: Kl, 235; XRS-6, 108; XRS- 5,33. Kl cells have 5.8 times as much catalase activity and 5.4 times as much GPx activity as XRS-5 cells. Kl cells have 3.7 times as much catalase activity and 2.1 times as much GPx activity as XRS-5 cells. The cells with intermediate amounts have an intermediate response. Extent of viability determined by trypan blue assay: DMA111 was by far the most cytotoxic at all concentrations tested, with the percentages of living cells at 1, 10, and 100 being approximately 45, 41, and 0%, respectively. Clonogenic survival in a colony -forming assay, IC50 values in BFTC905 and NTUB1 cells, respectively: 0.13,0.16. 9.25, 9.00. 0.13,0.15. 3.04, 2.64. 0.52,0.58. 0.38,0.63. Colony -forming assay: this concentration caused ~32 % survival; squalene at up to 160 uM had no effect on cytotoxicity. Reference Wanget al., 1997 Dopp et al., 2004 Wang et al., 2007 Fan et al., 1996 C-182 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CHO cells treated with MMS before or after inorganic arsenic treatment CHO Kl cells exposed to 1.5 or 2.5 J/m2ofUV C-33A cells HeLa cells Jurkat cells LCL-EBV cells Jurkat cells and human lymphocytes Mouse 291.03C keratinocytes Arsenic Species As111 SA As111 SA As111 SA for all As111 SA As111 SA Concentration(s) Tested (nM) 5, 10, as pretreatments 5, 10, as post- treatments 10 0.1, 1, 10,25,50 for all 01 1 10 25 50 for both 0.05,0.1,0.5, 1,5 Duration of Treatment 24 hr 24 hr 24 hr 24 hr 24 hr for both 7 days LOECa (HM) None 5 10 10 50 0.1 10 0.1 for both 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Colony -forming assay: No change from MMS alone. ft in cytotoxicity compared to MMS alone, synergistic interaction with even less survival at 10. Colony-forming assay: Synergistic ft in cytotoxicity because of the inorganic arsenic post-treatment. Cell viability determined by Trypan blue exclusion: -35% viability at 50. -75% viability at 50. -55% viability at 50. -60% viability at 50. Cell viability determined by Trypan blue exclusion: When both of these cell types were transfected with mutant p53 genes (by electroporation) there was substantially increased cytotoxicity. This ft was already apparent at a dose of 0. 1 (i.e., theLOEQinthel p53 mutation tested in Jurkat cells and in 1 of 2 p53 mutations tested in PHA-stimulated lymphocytes. Cytotoxicity based on colony survival, using crystal violet staining: LC50 = 0.9; almost all dead at dose of 5. Reference Lee et al., 1986 Yanget al., 1992 Salazar et al., 1997 Salazar et al., 1997 Wuetal., 2005 C-183 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Chinese hamster V79 cells A2780 cells H460 cells MCF-7 cells BALB/c 3T3 cells (derived from mice) G12 cells U-2OS cells Arsenic Species As111 SA DMAV As111 ATO for all As111 SA Asv MMAV DMAV TMAV As111 SA MMAmO DMAmI As111 SA Concentration(s) Tested (nM) 1,2,5, 10 -0.8, 1, 2, 5, 10 mM IC50 determinations IC50 determinations 0.05,0.1,0.5, 1, 2.5,5, 10 0.2, 0.4, 0.6, 0.8, 1 0.1,0.2,0.3,0.4 0.01,0.05,0.1, 0.25,0.5, 1,2.5 Duration of Treatment 24hrs for both 72hr 18 hr 72 hr 10 days LOECa (HM) 1 2mM — — 1 0.2 0.1 0.05 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cytotoxicity based on number of viable cells compared to control: LC50s: As111, -5.5; DMAV: -3.5 mM. Cell survival was determined using the MTT assay: IC50 values: A2780, 2.80; H460, 14.60; MCF-7, 3.00. Cell survival was determined using the MTT assay: IC50 values: As111 SA, 16.9; Asv, 64; MMAV, 14.7 mM; DMAV, 4.35 mM; TMAV, >74 mM. Depletion of GSH in cells by co-treatment with 0.2 mM BSD markedly reduced the cytotoxicity of DMAV even though it markedly increased the cytotoxicity of the other 4 compounds. Cell survival was determined using the clonal survival assay: LC50 values: As111 SA, ~8;MMAmO, 0.51; DMAmI, 0.15. The 2 methylated forms were also tested at 4 and 24 hr and showed cytotoxicity at both; for MMAmO, cytotoxicity was >50% at both times at highest dose. Cell survival was determined using the clonal survival assay: LC50 = 0.68; 100% cell killing at 2.5. Reference Ochi et al., 1999b Ling et al., 2002 Ochi et al., 1994 Klein et al., 2007 Komissaro va et al., 2005 C-184 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U-2OS cells U118MG cells Undifferentiat edPC12 cells FGC4 cells HepG2 cells Rat hepatocytes SVEC4-10 cells Arsenic Species As111 SA for all As111 ATO As111 ATO As111 SA for all As111 SA Concentration(s) Tested (nM) LC50 determinations 1, 5, 10, 25, 50 1, 10, 100, 1000 50, 75, 100, 125 25, 50, 75, 100, 125 2, 10,25,35,45, 55 2, 4, 8, 12, 16 Duration of Treatment 24hr 48hr 72hr 24hr 24 hr 24 hr for all 24 hr LOECa (HM) — — — 5 1 75 50 25 4 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the clonal survival treat-and- plate (CSTP), neutral red (NR), and MTT assays, for the different durations: LC50:CSTP, 1.1; MTT, 3.8;NR, 4.8. LC50: CSTP, 0.9; MTT, 0.99; NR, 1.05. LC50: CSTP, 0.8; MTT, 0.8; NR, 0.84. Cell survival was determined using the MTT assay: slightly > 50% survival at dose of 5; co-treatment with lipoic acid blocked cytotoxicity. Other tests showed no ft in either apoptotic cell death or intracellular peroxide levels; cell death was shown to be autophagic. Cell survival was determined using the MTT assay: LC50 = 8. (At dose of 8, about 75% cell survival at 12 hr.) Effects of pretreatment or co-treatment with antioxidants on cytotoxicity: NAC: big but a-Toc, GSH, 17(3- estradiol, orBO653: NSE. Cell survival was determined by the NR uptake assay: LC50s: FGC4, -90; HepG2, -70; hepatocytes, -30. Cytotoxicity determined by the MTT assay: LC50 = -13. Reference Komissaro va et al., 2005 Cheng et al., 2007 Pigaetal., 2007 Gottschalg etal.,2006 Chao etal., 2006a C-185 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HCT1 16 cells (securin +/+) HCT1 16 cells (securin -/-) RKO cells (p53 wt) SW480 cells (p53 mutant) U-2OS cells Arsenic Species As111 SA for both As111 SA for both As111 SA for all Concentration(s) Tested (nM) 4, 8, 12, 16 for both 8, 16, 24, 32 for both 0.1, 1, 10 Duration of Treatment 24 hr for both 24 hr for both 24 hr LOECa (HM) 4 4 8 8 1 or 10; see explana- tion Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cytotoxicity determined by the MTT assay: LC50s = securin +/+, -17; securin-/-, -11. There was significantly more cytotoxicity in null mutant at doses of 8, 12 and 16. Cytotoxicity determined by the MTT assay: LC50s = RKO, -20; SW480, -27. There was significantly more cytotoxicity in wt p53 cell line at doses of 16, 24 and 32. Trypan blue exclusion assay to identify necrotic cells (which take up stain) after additional periods of post-treatment culturing of 0, 24, or 48 hr in arsenic -free medium: At dose of 0.1, no increase in necrotic cells at any time. At dose of 1, necrotic cells were -0%, -20%, and -40% of total cells, respectively. At dose of 10, necrotic cells were -70%, -95%, and -95% of total cells, respectively. Note that a 24-hr treatment with S A affected the amount of necrosis at a dose of 1 only if there was an additional 24-hr or longer period of culturing in SA-free medium between the end of the SA treatment and when the assay was done. Reference Chao etal., 2006a Chao etal., 2006a Komissaro vaetal., 2005 C-186 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells CRL1675 cells THP-1 + A23 187 cells HaCaT cells CRL1675 cells THP-1 + A23 187 cells Alveolar macrophages (AMs) from CDFi mice Peritoneal macrophages (PMs) from CDFj mice Arsenic Species As111 ATO for all As111 ATO for all As111 SA Asv MMAV DMAV TMAV Concentration(s) Tested (nM) LD10andLD25 determinations for each cell line LD10andLD25 determinations for each cell line IC50 determinations Duration of Treatment 72 hr 72 hr under chronic exposure conditions 48 hr LOECa (HM) — — — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cytotoxicity assessed using fluorescein diacetate assay: LD10= 1.9;LD25= 15.2. LD10=1.0;LD25=1.9. LD10=1.9;LD25 = 3.8. Testing for cytotoxicity was preceded by exposure to 1.0 uM As111 ATO for at least 8 passages to establish chronic-exposure conditions. Then, following exposures to various doses for 72 hr, cytotoxicity was assessed using fluorescein diacetate assay: LD10 = 2.0 ; LD25 = 4.0. LD10 = 0.5;LD25=1.3. LD10 = 0.5;LD25 = 5.1. Cell survival was determined using the AlamarBlue assay (said to be similar to the MTT assay): IC50 values of AM cells: As111 SA, 4; Asv, 400; MMAV, >10 mM; DMAV, 5 mM; TMAV, »10 mM. IC50 values of PM cells: As111 SA, 5; Asv, 650; MMAV, >10 mM; DMAV, 5 mM; TMAV, »10 mM. DMAV caused almost entirely apoptotic cell death, while the inorganic arsenicals caused mainly necrotic cell death. SOD, CAT. and a peptide inhibitor ICE inhibited the cytotoxicity of As111 but notofDMAv Reference Graham- Evans et al., 2004 Graham- Evans et al., 2004 Sakurai et al., 1998 C-187 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue RHMVE cells HLFC cells HLFK cells (Ku70 deficient) Arsenic Species MMAV DMAV TMAVO As111 SA for both Concentration(s) Tested (nM) 0.25,0.5, 1,2.5,5, 10, 25, 50, 100 mM for all 5, 10, 20, 40, 80 for both Duration of Treatment 24 hr 24 hr for both LOECa (HM) 25 mM ImM None 10 for both Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using a modified MTT assay: LC50s: MMAV, 33.6 mM; DMAV, 2.54 mM; TMAVO, cell number increased by dose of 1 mM, reaching 135% of control at dose of 25 mM. Another study showed LC50s: As111, 36; Asv, 220 (both uM). Co-treatment with NAC caused U in cellular arsenic content and cytotoxicity by DMAV but not by MMAV Co- treatment with B SO caused big ft in cytotoxicity of MMAV but slight U in cytotoxicity of DMAV. Viability was determined by trypsin blue exclusion assay: LC5os: HLFC, 27.38; HLFK, 21.80; cytotoxicity was significantly greater for HLFK than HLFC at doses of 20, 40 and 80. Reference Hirano et al., 2004 Liu et al., 2007b C-188 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells NB4-M-AsR2 cells IM9 cells MCF-7 cells, T47D cells, and MDA-MB- 231 cells Arsenic Species As111 ATO for all As111 ATO for all Concentration(s) Tested (nM) 0.5, 1 2,4 0.5, 1 IC50 determinations Duration of Treatment 6 days for all 3 days LOECa (HM) 0.5 4 0.5 — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell viability (% of control) for ATO alone and for ATO with 100 uM Trolox, determined using trypan blue exclusion: At 0.5: 75% alone, 43% with Trolox; at 1: 30% alone, 3% with Trolox. At 2: 100% alone, -80% with Trolox; at 4: -63% alone, -30% with Trolox. At 0.5: -80% alone, -70% with Trolox; at 1: -50% alone, -25% with Trolox. Thus, Trolox enhanced ATO-induced cytotoxicity (or growth inhibition) in all 3 cell lines. Cell viability for ATO without and with 100 uM Trolox co-treatment, respectively, determined using trypan blue exclusion assay: MCF-7: 2.07 and 1.02; T47D: 3.22 and 1.56; MDA-MB-23 1:2.27 and 0.98. Thus, co-treatment with Trolox enhanced ATO growth inhibition similarly to what was seen in the row above. Reference Diaz et al., 2005 Diaz et al., 2005 C-189 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human PBMCs cultured in various ways MDAH 2774 cells SVEC4-10 cells 1RB3AN27 cells BEAS-2B cells HT1 197 cells Arsenic Species As111 ATO As111 ATO As111 SA As111 SA As111 ATO As111 SA Concentration(s) Tested (nM) 1 1, 2, 5, 8 5, 10, 20, 40 0.1,0.5, 1,5, 10 10, 20, 50 1, 5, 10, 25, 50 Duration of Treatment 15 days 72hr 24hr 72hr 24 hr 24 hr LOECa (HM) 1 lor 2 10 1 10 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Colony-forming ability was assessed for ATO alone and for co- treatment with Trolox by counting CFU- erythrocytes, CFU- granulocytes- macrophages, and BFU- erythrocytes. Biggest effect of ATO alone: 62% U for CFU- erythrocytes. In all 3 cases, co-treatment with Trolox had little or no effect. Cytotoxicity assessed using trypan blue exclusion assay: uncertainty about LOEC exists because control value was not reported: LC50 = 5. Cell survival was determined using the MTT assay: LC50 = -13. Cell survival was determined using the MTT assay: there probably was cytotoxicity at dose of 1; statistically significant cytotoxicity at dose of 5; LC50 = ~8; all experiments on ROS or induction of transcription factors were at doses of <10 for <4 hr, and under those conditions, there was no cytotoxicity. Cell survival was determined using the MTT assay: LC50 = -15. Cell survival was determined using the trypan blue exclusion assay: LC50 = -35. Reference Diaz et al., 2005 Terek et al., 2006 Hsu et al., 2005 Felix et al., 2005 Hanetal., 2005 Hernandez -Zavala etal., 2005 C-190 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HL-60 cells U266 cells Arsenic Species As111 ATO for both Concentration(s) Tested (nM) 1,2,3,5,10 for both Duration of Treatment 24 hr for both LOECa (HM) 2 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the trypan blue exclusion assay: LC50s = HL-60, ~7; U266, ~2. Effects of modulators in both cell lines: (Cells were loaded with high concentrations of intracellular AA [icAA] by incubating them with DHA prior to inorganic arsenic treatments, thus avoiding generation of extracellular ROS in tissue culture media caused by direct addition to it of AA.) icAA caused big U in cytotoxicity of inorganic arsenic. GSH depletion by BSO treatment caused big ft in inorganic arsenic-induced cytotoxicity. icAA caused big U in cytotoxicity caused by inorganic arsenic in GSH-depleted cells. Extracellular AA caused big ft in inorganic arsenic-induced cytotoxicity, including after GSH depletion. Relatively limited data from a methylcellulose colony-forming assay in both cell lines (with 48- hr inorganic arsenic treatment and 10-14 days to form colonies) and from cytotoxicity testing of RPMI-8226 cells supported some of the above conclusions. Effect of NAC was tested in HL60 cells; it caused big U in inorganic arsenic-induced cytotoxicity. Reference Karasawa s etal., 2005 C-191 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Embryonic mesenchymal cells prepared fromCF-1 mouse conceptuses at gestation day 11 HCT15 cells HeLa cells PLC/PR/5 cells Chang cells K562 cells, AR230-S cells, AR230- r cells, KCL22-S cells, KCL22- r cells, NB4 cells Arsenic Species As111 SA As111 SA for all As111 ATO Concentration(s) Tested (nM) 5.8, 11.6, 15.4, 30.8 LCso determinations IC50 determinations for all Duration of Treatment 2hr 24 hr 3 days LOECa (HM) 5.8 — — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the MTT assay: LC50 = -27; 15-min pretreatment with 0.5% (v/v) DMSO completely blocked the inorganic arsenic effect at dose of 15.4, whereas 15-min pretreatment with 0.1% or 0.2% (v/v) DMSO partially blocked it. Cell survival determined by MTT cell proliferation assay: LC50s: HCT15, 278.33; HeLa, 200.33; PLC/PR/5, 376.66; Chang, 328.33. Antiproliferative activity as determined by MTS assay — some would interpret such results as cytotoxicity and present results as LC50s: IC50s: K562, 0.9; AR230-S, 2.6; AR230-r, 6.9; KCL22-S, 2.6; KCL22-r, 2.8; NB4, 0.4. A dose of 2 represents the upper margin of the clinically useful range for ATO. There was a positive correlation between GSH content of cells and resistance to the antiproliferative (i.e., cytotoxic) effect. Reference Perez- Pasten et al., 2006 Othumpan gat et al., 2005 Konig et al., 2007 C-192 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue AR230-S cells, AR230-r cells, KCL22-S cells, KCL22- r cells H1355 cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 1 3.125,6.25,12.5, 25, 50, 100, 200 Duration of Treatment 2 days 24 hr LOECa (HM) None 6.25 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined by trypan blue assay: 100 uM BSO treatment was shown to greatly U GSH levels in all 4 cells types both with and without inorganic arsenic exposure. In all 4 cell types, the inorganic arsenic + BSO treatment caused big to huge U in number of viable cells, whereas untreated cells or cells treated with inorganic arsenic or BSO showed ~2-fold H A similar assay in primary cultures of mononuclear cells from 4 patients in blast crisis with imatinib- resistant CML also showed maximum cytotoxicity for the combined inorganic arsenic + BSO treatment. Cell survival was determined using the MTT assay: Cytotoxicity increased with dose, with -57% cell survival at dose of 200. Reference Konig et al., 2007 Cheng et al., 2006 C-193 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL1215 cells = X in this row TRL1215 cells that had been treated with 1.3 mM MMAV for 20 weeks prior to acute arsenic treatments = Y in this row Arsenic Species As111 SA, Asv, DMAV for both Concentration(s) Tested (nM) LCso determinations for both Duration of Treatment 48 hr for both LOECa (HM) — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival based on AB assay: LC50sforAsni:X, 16.3; Y, 74.1. LC50s for Asv:X, 157.1; Y, 2743.8. LC50s for DMAV: X, 2090; Y, 6950. Thus the MMAV pretreatment caused marked resistance to cytotoxicity for all 3 arsenicals. Much of this resistance was lost if Y cells were cultured for 8 more weeks with no arsenic in media. The 20-week pretreatment caused no cytotoxicity, gave the Y cells an arsenic content of 135.4 + 12.0 ng/mg cellular protein, and did not induce malignant transformation. Arsenicals were not methylated or demethylated in these cells. Reference Kojima et al., 2006 C-194 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL1215 cells = X in this row TRL1215 cells that had been treated with 0.7 mM DMAV for 20 weeks prior to acute arsenic treatments = Y in this row Arsenic Species As111 SA, Asv, DMAV for both Concentration(s) Tested (nM) LC50 determinations for both Duration of Treatment 48 hr for both LOECa (HM) — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival based on AB assay: LC50sforAsni:X, 16.3; Y, 19.2. LC50s for Asv:X, 157.1; Y, 182.2. LC50s for DMAV: X, 2090; Y, 4730. Thus the DMAV pretreatment caused marked resistance to cytotoxicity for only the DMAV treatment, and the slight differences for the other 2 arsenicals were not statistically significant. WhenY cells were cultured for 8 more weeks with no arsenic in media, there was no change regarding the lack of resistance to As111, but the resistance to the other 2 arsenicals increased substantially. The 20-week pretreatment caused no cytotoxicity, gave the Y cells an arsenic content of41.8 + 2.5ng/mg cellular protein, and did not induce malignant transformation. Arsenicals were not methylated or demethylated in these cells. Reference Kojima et al., 2006 C-195 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL1215 cells = X in this row TRL1215 cells that had been treated with 10.0 mM TMAVO for 20 weeks prior to acute arsenic treatments = Y in this row Arsenic Species As111 SA, Asv, DMAV for both Concentration(s) Tested (nM) LCso determinations for both Duration of Treatment 48 hr for both LOECa (HM) — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival based on AB assay: LC50sforAsni:X, 16.3; Y, 54.8. LC50s for Asv:X, 157.1; Y, 684.1. LC50s for DMAV: X, 2090; Y, 4500. Thus the TMAVO pretreatment caused marked resistance to cytotoxicity for all 3 arsenicals. Much of this resistance was lost regarding DMAV, and all of it was lost regarding the other 2 arsenicals, if Y cells were cultured for 8 more weeks with no arsenic in media. The 20-week pretreatment caused no cytotoxicity, gave the Y cells an arsenic content of 543.8 + 12.0 ng/mg cellular protein, and did not induce malignant transformation. Arsenicals were not methylated or demethylated in these cells. Reference Kojima et al., 2006 C-196 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells 293 cells Arsenic Species As111 SA As111 ATO MMAm DMA111 As111 ATO Concentration(s) Tested (nM) 1,2,3,4 1,2,3,4 0.25,0.5, 1,2 2, 4, 6, 8 0.5,1,2,3,4 Duration of Treatment 72 hr 12 days LOECa (HM) 1 1 0.25 4 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the MTT assay: LC50s: As111 SA, -3.4; As111 ATO, -2.2; MMA111, -1.2; DMA111, -5.8. Co-treatment (CoTr) with 3000 uM DTT markedly decreased cytotoxicity of all arsenicals: Maximum cytotoxicities with 3000 uM DTT CoTr: As111 SA, -17%; As111 ATO, -12%; MMA111, -25%; DMA111, -12%. CoTr with 100 uM DTT markedly increased cytotoxicity of all arsenicals: LC50s with 100 uM DTT CoTr: As111 SA, -2.2; As111 ATO, -1.0; MMA111, -0.28; DMA111, -4.0. Cell survival was determined by colony- forming assay (% of cells forming colonies): -73% at dose of 4; LC2s = -3.6. Co-treatment with 200 uM DMSA increased survival: -87% at dose of 4. Co-treatment with 20 uM DMSA decreased survival: -61% at dose of 4; LC25 = ~1.6. Co- treatment with 100 uM DMPS increased survival: -86% at dose of 4. Co-treatment with 10 uM DMPS decreased survival: -50% at dose of4;LC25 = ~1.2. Reference Jan et al., 2006 Jan et al., 2006 C-197 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SV-HUC-1 cells HeLa cells Primary rat hepatocytes A431 cells RAW264.7 cells Arsenic Species As111 ATO As111 SA As111 SA As111 ATO As111 SA Concentration(s) Tested (nM) 0.5,1,2,3,4 10, 100 2.5,5,7.5, 10, 15, 20, 25, 30, 40, 50 1.25,2.5,5, 10,20 for both 2.5, 5, 10, 25 Duration of Treatment 12 days 24hr 24hr 24hr 48 hr 24 hr LOECa (HM) 0.5 10 7.5 2.5 1.25 2.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined by colony- forming assay (% of cells forming colonies): -62% at dose of 4; LC25 = -2.2. Co-treatment with 200uMDMSA increased survival: -73% at dose of 4; LC25 = -3.5. Co-treatment with 20 uM DMSA decreased survival: -43% at dose of4;LC25 = ~1.4. Co- treatment with 100 uM DMPS increased survival: -79% at dose of 4. Co-treatment with 10 uM DMPS decreased survival: -47% at dose of4;LC25 = ~1.2 Cell survival determined using a LIVE/DEAD viability/cytotoxicity kit: LC50: ~95. Cell survival was determined using the MTT assay: LC50 = -18. Cell survival was determined using the MTT assay: At 24 hr: LC50 = -20. At48hr:LC50 = ~3. Cell survival based on neutral red uptake assay: LC50 = ~13. Reference Jan et al., 2006 Hansen et al., 2006 Hasgekar etal.,2006 Huang et al., 2006 Szymczyk et al., 2006 C-198 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NIH 3T3 cells NHEK cells BAEC cells H22 cells BAEC cells Arsenic Species As111 SA As111 SA As111 SA As111 ATO for both Concentration(s) Tested (nM) 5, 10, 20, 50, 100, 200 0.2, 0.4, 0.8 1, 5, 10 0.5, 1,2,4 for both Duration of Treatment 6hr 1, 2, 3, 4 days 24 hr 48 hr 24 hr, 48 hr 24 hr, 48 hr LOECa (HM) 20 for U 0.2 for ft on all days 5 1 1,0.5 2,1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell viability assayed using CellTiter-Glo assay: possibly slight ft at 5 and 10; U at 20, LC50 = ~90. Pre-induction of HSP by conditioning heat shock (2 hr at 42°C on prior day) or by constitutive expression of HSP70 markedly reduced the cytotoxicity, as follows: with heat: LOEC = 100 and -80% viability at dose of 200, with constitutive expression: LOEC = 50 and -70% viability at dose of 200. Cell survival was determined using the NR uptake assay: 1ho~l.l-1.4x at doses of 0.2 and 0.4 on all days; point estimates at dose of 0.8 were always higher than control, but the ft was always a NSE. Cell survival was determined using a variation of the MTT assay:LC50s:~7.5at24 hr, ~5.0at48hr. Unlike co-treatment with Zn11, Fen,orCun, only co-treatment with Mn11 increased inorganic arsenic toxicity at concentrations at which it (the metal) did not cause cytotoxicity alone. Cell survival (also called the proliferation index) was determined using the MTT assay: LC50sforH22:~2.0at 24 hr, -1.2 at 48 hr. LC50sforBAEC:~4.5at 24 hr, ~2 at 48 hr Reference Khalil et al., 2006 Hwang et al., 2006 Bunderson etal.,2006 Liu et al., 2006e C-199 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HEK 293 cells HEK 293 cells transfected withOATP-C U937 cells TRL 1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with 50uMBSO Arsenic Species As111 SA Asv MMAV DMAV As111 SA for all MMAV for both Concentration(s) Tested (nM) LCso determinations 0.5, 1,2.5,5, 10, 20 for all 1.25,2.5,5, 10 mM for both Duration of Treatment 72 hr 24 hr 48 hr 72 hr 48 hr for both LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the MTT assay: LC50s without and with OATP-C, respectively: As111: 10.9, 5.6; Asv: 98.1,53; MMAV: 43 19.3, 421 1.6 (this comparison: NSE); DMAV: 994.1, 899.3 (this comparison: NSE). The OATP-C transfected cells accumulated 43% more As111 and 34% more Asv than the non-transfected cells while they did not accumulate more of the methylated arsenicals. Co-treatment of the As111- or Asv- treated cells with rifampin or taurocholic acid eliminated the difference between the two cell types. OATP-C can transport inorganic arsenic in a (GSH)-dependent manner but this may not be the major pathway for arsenic transport. 20 10 10 10 mM 2.5 mM Cell survival was determined using the Pi- exclusion assay: At 24 hr, -74% survival at dose of 20. At 48 hr, -62% survival at dose of 20. At 72 hr, -40% survival at dose of 20, LC50: -17.5. Cell survival was determined using the AB assay: Without BSO: -80% cell survival at dose of 10 mM; at 5 mM, survival may have been higher than that of control LC50 with BSO: 3.2 mM. Similar results were obtained using CV assay. Reference Luetal., 2006 McCollum etal.,2005 Sakurai et al., 2005a C-200 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Gclm+/+ MEF cells Gclm+/" MEF cells Gclm'7- MEF cells, from GCLM knockout mice HeLa cells U937 cells Primary human skin fibroblasts Arsenic Species As111 SA for all As111 ATO for all Concentration(s) Tested (nM) 4, 8, 16, 32, 64 for all 2 for all Duration of Treatment 8hr for all 3 days LOECa (HM) 16 or 64 16 or 64 4 2 2 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the MTT assay: It was unclear which of the first two genotypes had the LOEC of 16; one had an LOEC of 64, and the LOEC for the other one was 16; LC50s: +/+, 86; +/-, 86; - /-, ii; pretreatment with tBHQ protected Gclrn 7~ and Gclm+/" MEF cells from inorganic arsenic- induced cytotoxicity in a dose- and time- dependent manner. Cell survival was determined using the MTS assay: -77% survival in HeLa and -85% survival in U937; no hint of cytotoxicity in fibroblasts. Co- treatment with 10 uM emodin apparently sensitized HeLa and U937 cells (but not fibroblasts) to cytotoxicity. The addition of 1.5 mM NAC to the co-treatment of HeLa cells with 10 uM emodin and 2 uM inorganic arsenic eliminated all cytotoxicity; effect of NAC was not tested in U937 cells. Emodin was used because it has a semiquinone structure that is likely to increase the generation of intracellularROS. Reference Kannet al., 2005b Yi et al., 2004 C-201 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MCF-7 cells MYP3 cells Arsenic Species As111 ATO As111 SA Asv MMAm DMA111 DMAV TMAVO Concentration(s) Tested (nM) 0.5, 1, 2, 4, 8, 16 2,3 35,40 1,1.5 0.6, 1 0.6 mM, 1 mM 15 mM, 20 mM Duration of Treatment 24 hr, 48 hr, or 96 hr 7 days for all LOECa (HM) 2 at 24 hr; Iat48 and 96 hr 2 35 1 0.6 0.6 mM 15 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the MTTassay:LC50sat24, 48, and 96 hr were 8.6, 3. 3, and 1.86, respectively. Apoptosis was shown to be the mechanism of cell death after treatment with a dose of 5 for 3 days. Cell survival was determined using the MTT assay: -33% at 2, -9% at 3. -37% at 35, -28% at 40. -60% at 1, -7% at 1.5. -28% at 0.6, -10% at 1. -45% at 0.6 mM, -28% at 1 mM. ~28%atl5mM,~18% at 20 mM. Co-treatments with antioxidants that work by different mechanisms yielded the following results: melatonin slightly inhibited cytotoxicity of As111. NAC inhibited cytotoxicity of MMAm, DMA111, DMAV and TMAVO. Vitamin C inhibited cytotoxicity of As111, Asv, MMAm and DMA111. Tironand Trolox did not affect cytotoxicity of any arsenical. Reference Ye et al., 2005 Wei et al., 2005 C-202 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Primary keratinocytes (in third passage) obtained from foreskins of adults Huh? cells CL3 cells, synchronous atGl CL3 cells, asynchronous (asyn) CL3 cells, synchronous atS CL3 cells, synchronous atG2/M Arsenic Species As111 SA As111 SA As111 SA for all Concentration(s) Tested (nM) 0.001,0.01,0.1, 1, 5, 10, 100, 1000 0.5,1,3,5, 10,20 50, 100 for all Duration of Treatment 24hr 48hr 72hr 24 hr 3hr for all LOECa (HM) SforU lft;5U 0.1 ft; 5U 1 for ft 20 for U — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined using the XTT assay: ft in viability (proliferation) to 1.1 8x and 1.32x at dose of 1 at 48 and 72 hr, respectively. LC50s at 24, 48, and 72 hr were -160, -10, and -4.2, respectively. Cell survival was determined using the MTT assay: ftto-l.lxat doses of 1 and 3; U to 58% at dose of 20. In co-treatments with lOnM TCDD, inorganic arsenic doses of 5 and 10 caused 0% and 10% cytotoxicity, respectively. Cell survival was determined using a colony-forming assay: % survival at dose of 50: Gl, 45%; asyn, 35%; S, 29%; G2/M, 17%. Survival at dose of 50 in Gl cells was cut from 45% to 25% to 30% by co-treatment with PD98059orU0126, which are these 2 structurally dissimilar inhibitors of MEK1/2. Reference Liao etal., 2004 Chao etal., 2006b Li et al., 2006a C-203 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human neuroblastom a cell lines: IMR-32 SK-N-DZ SK- N-BE(2) SK-N-AS SH-SY5Y All 4 lines ± co-treatment with 25 uM DCHA HaCaT cells Arsenic Species As111 ATO As111 SA Asv MMAm DMA111 Concentration(s) Tested (uM) 1 0.5,1,1.5,2.5,4, 6, 7, 8, 10, 12, 13, 14, 16, 18, 20, 22 10, 20, 30, 40, 50, 60, 80, 100, 120, 160, 200, 240, 280, 320, 360 0.1,0.5, 1,2,2.5, 3,3.5,4,4.5,5, 5.5,6,6.5,8, 10 0.1,0.5, 1,2,3,4, 5,6,7,8,9, 10, 11 Duration of Treatment 48 hr 24 hr for all LOECa (UM) -or + DCHA None, 1 None, 1 1,1 1,1 1,1 0.5 for ft 12 for U 10 for ft 100 for U 0.1 for ft 2. 5 for U None for ft 3 for U Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival (% of control) was determined using the MTT assay: inorganic arsenic alone. DCHA alone, (inorganic arsenic + DCHA) NSE, NSE, 35%. NSE, NSE, 45%. 73%, NSE, 41%. 56%, NSE, 39%. 61%, NSE, 40%. co-treatment of (inorganic arsenic +DCHA) with vitamin E blocked much of the cytotoxicity in line IMR- 32. Cell survival was determined using the MTT assay, with a proliferative effect being seen at lower doses: Peak of 141% at dose of 1; first point estimate below 100% at dose of 8; about 50% cytotoxicity at 22. Peak of 145% at dose of 10; first point estimate below 100% at dose of 80; about 50% cytotoxicity at 320. Peak of 160% at dose of 1; first point estimate below 100% at dose of 2.5; about 50% cytotoxicity at 4. 5. About 60% cytotoxicity at 11. Reference Lindskog etal.,2006 Ganyc et al., 2007 C-204 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Mouse lymphoma cells (L5178Y/Tk+/- -3.7.2Ccells) V79 cells treated with MNU V79 cells exposed to UVA, UVB, orUVCover a wide range of doses Human- hamster hybrid AL cells Arsenic Species As111 SA Asv MMAV DMAV As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 2.3,3.1,4.6,6.2, 7.7, 9.2, 10.8, 11.5, 13.1, 13.9, 14.6, 15.4 6.1, 15.2,30.3, 45.5, 48.5, 54.6, 60.6, 66.7, 69.7, 72.8, 78.8, 84.9 12.3, 15.4, 18.5, 21.6, 24.7, 27.8 mM 12.5, 18.8,25.0, 31.3,37.5,43.8, 50.0, 56.3, 62.5 mM 10 5 10 3.8,7.7, 15.4 Duration of Treatment 4hr for all 3hr 24 hr 3hr 1 day or 5 days LOECa (HM) 4.6 15.2 12.3 mM 18.8 mM 10 5 10 3.8 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival, as percent of relative total growth compared to the vehicle control: Estimates of LC5os: As111 SA, -7.3 uM; Asv, -50.3 uM; MMAV, ~16.1mM; DMAV, -38.8 mM. Cell survival, percent of control: both inorganic arsenic treatments caused 4% or less cytotoxicity; however, as post- treatments they both considerably increased the cytotoxicity caused by the MNU treatments. Cell survival, percent of control: The inorganic arsenic treatments caused 8% or less cytotoxicity; however, as post- treatments they increased the cytotoxicity caused by the UV treatments. Colony-forming assay; -55% survival with 1- day treatment at 7.7. Reference Moore et al., 1997a Li and Rossman, 1989a Li and Rossman, 1991 Heietal., 1998 C-205 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue 1T1 cells MYP3 cells AG06 cells Human cells: AG06 (keratinocytes ) AG06 (keratinocytes ) HaCaT (keratinocytes NHEK (keratinocytes GM847 (fibroblasts) WI38 (fibroblasts) AG06 cells K562 cells Arsenic Species As111 SA Asv MMAmI2 MMAV DMAmI DMAV TMAVO As111 SA As111 SA MMAm As111 SA As111 SA As111 SA As111 SA As111 SA MMAm As111 ATO Concentration(s) Tested (nM) LC50 determinations for all 0.2, 4, 20 IC50 determinations 1, 5, 10, 20, 30 2.5 Duration of Treatment 7 days 24 hr pretreatment 48 hr 5hr 12 hr LOECa (HM) — 0.2 for ft 4forU — — 2.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) LC50s based on trypan blue assay for viability: AsmSA:4.8inlTl;0.4 inMYP3. Asv:31.3inlTl;5.3in MYP3. MMAmI2: 1.0 in IT 1; 0.8 inMYP3. MMAV: 1.7mMinboth lines. DMAmI: 0.8 in 1T1; 0.5 inMYP3. DMAv:0.50mMinlTl; LlmMinMYPS. TMAVO: l.VmMin lTl;4.5mMinMYP3. Extent of viability determined by NR assay: ft in viability over that seen for MNNG alone at -1-15 (oM MNNG. U in viability below that seen for MNNG alone at -1 5-40 (oM MNNG (synergistic interaction). Extent of viability determined by NR assay: IC50: 7.2. IC50: -7.5. IC50: 11.6. IC50: 12.3. IC50: 10.7. IC50: 11.2. Extent of viability determined by NR assay: -20 kills 20% of cells. -20 kills 50% of cells. -50% of cells die. Reference Cohen et al., 2002 Snow et al., 1999 Snow et al., 2001 Snow et al., 2001 Li and Broome, 1999 C-206 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human kidney carcinoma cell lines: UOK123 UOK109 UOK121 Human lung carcinoma cell line: A549 HFW cells (diploid human fibroblasts) HFW cells (diploid human fibroblasts) V79-C13 Chinese hamster cell line Syrian hamster embryo cells Arsenic Species As111 SA As111 SA As111 SA As111 SA As111 SA Asv Concentration(s) Tested (nM) IC50 determinations 2.5, 5, 10, 20 1.25,2.5,5, 10 5, 10, 20, 40, 80 5, 10, 20, 30, 40, 50,60 -0.7, 1.4, 2, 3, 4, 5,6 ~5, 10, 20, 50, 75, 100, 130, 160, 200 Duration of Treatment 7 days 6hr 24 hr 4hr 24 hr 7 days for all LOECa (HM) — 2.5 1.25 -10 10 0.7ft, 5U 10ft, lOOli Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Extent of viability determined by colony- formation efficiency assay: 0.020. 0.021. 0.020. 0.4. Cytotoxicity determined by a colony-forming assay; co-treatment with catalase (but not heat- inactivated catalase) at 100 ug/mL markedly reduced cytotoxicity; increasing GSH levels with B-mercaptoethanol reduced cytotoxicity; decreasing GSH levels with B SO increased cytotoxicity. Cytotoxicity determined by a colony-forming assay. Cytotoxicity determined by a colony-forming assay: survival at 10 was 76.3 ±2.61% of control; IC50: ~20. Cytotoxicity determined by measuring CFE: Small but reproducible ft from 0.7 to about 1.5 followed by a logarithmic decrease in CFE with a linear increase in dose. Small but reproducible ft from 10 to 50 followed by a logarithmic decrease in CFE with a linear increase in dose. Reference Zhong and Mass, 2001 Lee and Ho, 1995 Yihand Lee, 1999 Sciandrello etal.,2002 Barrett et al., 1989 C-207 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells UROtsa cells UROtsa cells BALB/c 3T3 A3 1-1-1 cells (derived from mice) TK6 cells HL-60 cells Arsenic Species As111 SA MMAm for all As111 SA Asv MMAmO MMAV DMAmI DMAV As111 SA AsvDA MMAV DMAV As111 SA As111 ATO As111 ATO Concentration(s) Tested (nM) 0.1, 10,25,50, 100, 200 0.5, 1,2,5, 10 for all 0.1,0.5, 1,5 1,200 0.1,0.5, 1,5 1,200 0.1,0.5, 1,5 1,200 2, 5, 10, 15, 20 10, 15, 20, 25, 30 1,2,5, 10 mM 0.5, 1, 2, 5mM 0.1,0.5, 1, 10, 100, 1000 0.1, 1, 10, 100 0.2, 0.4, 0.8, 1.6, 3.1,6.3, 13,25, 50, 100 Duration of Treatment 24 hr 24 hr 48 or 72 hr 24 hr for all 72 hr for all 24 hr for both 24 hr LOECa (HM) 50 5 5 None None 5 None None None 5 10 5 mM ImM 10 10 0.8 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Viability determined using MTT assay: IC50 = -100; doses <10 were said to stimulate mitochondria! activity (i.e., the curve went up; the assay tests mitochondria! function), but the stimulation was not statistically significant. Co-treatment with BSO: big ft in cytotoxicity, withIC50 = ~15. Viability determined using MTT assay: IC50 = ~5. All cells (or almost all cells) were dead at LOEC. Viability determined using MTT assay: WithMMAmO: 50% cytotoxicity was estimated to result from dose of about 2.5, with about 90% cytotoxicity at dose of 5. Cytotoxicity based on percent cell growth compared to treatment with distilled water: IC50 values: As111 SA, 4.8; AsvDA, 17; MMAV, 9.8 mM; DMAV, 3.2 mM. Cytotoxicity based on trypan blue exclusion assay: For both: LC50 between 3 and 4. Viability determined using MTT assay: LC50 = 32. Reference Bredfeldt etal.,2004 Bredfeldt etal.,2006 Drobna et al., 2002 Tsuchiya etal.,2005 Hornhardt etal.,2006 Yedjou and Tchounwo u, 2007 C-208 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue IEC cells (primary culture) IEC-6 cells MDAH 2774 cells HPBMs exposed to M-CSF for 7 days and considered M- macrophages HPBMs exposed to GM-CSF for 7 days and considered GM- macrophages Arsenic Species As111 SA for both As111 ATO As111 SA Asv MMAV DMAV As111 SA Asv MMAV DMAV Concentration(s) Tested (nM) 7.7, 15, 38, 77, 116, 154 for both 1, 2, 5, 8 LCso determinations LCso determinations Duration of Treatment 24 hr for both 72 hr 48 hr 48 hr LOECa (HM) 15 15 Ior2 (uncer- tain since control not shown) — — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Viability determined using MTT assay: At dose of 77: IEC, -45% dead; IEC-6, -55% dead; the cytotoxicity of the 2 cells types was almost identical at most doses; based on this and their rather similar concentration-dependent declines in membrane enzymes and constituents (e.g., alkaline phosphatase, hexose, sialic acid, cholesterol, and phospholipid), the primary and established cultures gave approximately similar toxic responses. Cytotoxicity estimated by XXT proliferation assay and alternatively by trypan blue dye- exclusion assay (for which treatment time was either 72 or 96 hr — it was unclear from methods): IC5o by both methods: 5. Viability based on AB assay: LC50 values: As111, 7.0; Asv, 1900; MMAV, 2500; DMAV, 800. Viability based on AB assay: LC50 values: As111, 5.8; Asv, 2800; MMAV, 2000; DMAV, 2000. Reference Upreti et al., 2007 Askar et al., 2006 Sakurai et al., 2006 Sakurai et al., 2006 C-209 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue The following cell lines: HL- 60, U-937, TIG-112, CRL-1609, RAW264.7, mouse normal embryo cells, mouse embryo cells that were MT +/+and MT -/-, and the following 3 types of human immune cells: peripheral T- lymphocytes, immature dendritic cells and multi- nucleated giant cells GM04312C cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) LC50 determinations 2.5, 10, 50 Duration of Treatment 48 hr 24 hr LOECa (HM) — 2.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Viability based on AB assay: LC50 values: HL-60, 13; U-937, 12; TIG-112, 25; CRL-1609, 17; RAW264.7, 25; MT +/+ cells, 4.8; MT - /- cells , 5.8; T-lymphocytes, 3.3; dendritic, 8.2; giant, 2.3. Viability based on neutral red assay: LC50 = —20. However, when viability was based on colony-forming assay: LC5o = ~6withLOECof 2.5 Reference Sakurai et al., 2006 Shen et al., 2006 C-210 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Primary mouse hepatocytes SV-HUC-1 cells JB6C141 cells JB6C141 cells exposed to 0.1, 0.2, 0.5, 1,2,3,4, 5, 6,7, or 8 kJ/m2ofUVB at end of pretreatment with inorganic arsenic Arsenic Species As111 SA As111 SA MMAm DMA111 As111 SA for both Concentration(s) Tested (nM) 60, 100, 200 0.5, 1,2,5, 10 0.1,0.25,0.5, 1 0.25,0.5, 1,2,5 0.1, 1,5, 10,20, 50, 100, 500, 1000 10 Duration of Treatment 24 hr 3 days for all 24 hr for both LOECa (HM) 60 1 0.25 0.5 5 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Viability determined using MTT assay: LC50 = -200 (LC50 = 30 for 48-hr treatment). Pretreatment with SFN caused big U in cytotoxicity. SFN activates transcription factor Nrf2 and causes significant ft of protein expressions responsible for excretion of arsenic into extracellular space. SFN caused big ft in intracellular GSH levels and big U in intracellular arsenic levels. Also, pretreatments with BSO, EA, orMK-571,whichft arsenic accumulation in hepatocytes, caused big ft in cytotoxicity. Viability determined by SRB assay: LC50 values: As111, 2.91; MMAm, 0.46; DMA111, 1.59. Viability determined by MTS assay: LC50 = -15, decreased with dose until reached -12% of control at top 3 doses. Probably some cytotoxicity at UVB dose of 5, and there was significant cytotoxicity at UVB dose of 6. Viability was -70% of control at highest UVB dose. Reference Shinkai et al., 2006 Suetal., 2006 Tang et al., 2006 DNA Damage WRL-68 (human hepatic cell line) As111 SA 0.001,0.01,0.1, 10 16 hr 0.001 Induction of DNA- protein crosslinks (methylated forms of arsenic could not be detected in the cells). Ramirez et al., 2000 C-211 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human aorta VSMCs (vascular smooth muscle cells) PHA- stimulated and unstimulated human lymphocytes L-132 cells (human diploid alveolar epithelial type II cells) L-132 cells (human diploid alveolar epithelial type II cells) Arsenic Species As111 SA AsmATO As111 SA MMAV DMAV As111 SA MMAV DMAV Concentration(s) Tested (nM) -1.2,2.5,5, 10 10 100 100 5, 10, 100 100 for all Duration of Treatment 4hr 2hr 6hr for all 3hr for all LOECa (HM) -1.2 10 None None 5 None None 100 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) DNA strand breaks (double and single strand breaks and alkali-labile sites) detected by comet assay; the effect was similar in nonproliferating VSMCs. Oxidative damage to DNA measured by the comet assay, including SSBs — after digestion with FPG, arsenic- induced base damage was converted to a large increase in SSBs and some FPG-created DSBs. (FPG cleaves purines including 7,8- dihydro-8-oxoguanine (8-oxoG), formamidopyrimidines, and AP sites.) Like the damage induced by H2O2, arsenic-induced DNA damage was repaired more slowly in unstimulated lymphocytes. Induction of DNA S SB resulting from inhibition of repair polymerization by polymerization inhibitors aphidicolin and hydroxyurea. DMAV induced them in a dose- dependent manner (measured by alkaline elution). Induction of DNA repair synthesis using the BrdU photolysis assay (single- strand DNA breaks induced by UV- irradiation were measured by alkaline elution). Follow-up experiment with same DMAV treatment for 1, 3, or 6 hr showed increases with longer durations of treatment. Reference Lynn et al., 2000 Li et al., 2001 Yamanaka etal., 1997 Yamanaka etal., 1997 C-212 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue L-132 cells (human diploid alveolar epithelial type II cells) 4>X174 RF I DNA Naked double- stranded circular DNA Human primary peripheral blood lymphocytes Arsenic Species MMAV As111 SA MMAm DMA111 DMAV As111 SA Asv MMAm MMAV DMA111 DMAV Concentration(s) Tested (nM) 100 with 10 mM SAM present 0.1, 1, 10, 100, 300 mM 10, 15, 20, 25, 30, 60 mM 40, 80, 150, 250 (oM 0.1, 1, 10, 100, 300 mM 1-1000 1-1000 1.25-80 Not reported-875 1.4-91 Not reported- 1000 Duration of Treatment 6hr 2hr for all 2hr for all LOECa (HM) 100 None 30 mM 150 (oM None Not reported for any of them Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of DNA repair synthesis using the BrdU photolysis assay (single- strand DNA breaks induced by UV- irradiation are measured by alkaline elution). This and other evidence strongly suggests that the DNA damage was not directly induced by MMAV but by dimethylated arsenic that was produced metabolically by reaction ofMMAvwithSAM. Nicked DNA in DNA nicking assay. Breaks and/or alkali- labile lesions in DNA detected in the single- cell gel comet assay — the relative potencies based on slopes are shown below (the larger the number, the bigger the effect): As111 1 Asv 1. MMAm 77 MMAV <1 DMA111 386 DMAV <1 As111 and Asv caused a significant effect, and they were not significantly different from each other. MMAm and DMA111 were thus 77 and 386 times more potent in causing DNA damage than SA. Reference Yamanaka etal., 1997 Mass et al., 2001 Mass et al., 2001 C-213 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue E. coli WP2s(X) (lonn, sulAi, trpE65, uvrA155, lamB+) Raji cells (human B- lymphocytes) Jurkat cells Arsenic Species MMAm DMA111 As111 SA MMA111 DMA111 As111 SA MMA111 DMA111 Concentration(s) Tested (nM) 0.01,0.10, 1.0, 10 for all 0.2, 1, 10, 20, 40, 100 for all 0.2, 1, 10, 20, 40, 100 for all Duration of Treatment Overnight for all 4hr 4hr 2hr 4hr 4hr 2hr LOECa (HM) None None 10 0.2 10 10 0.2 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Assay to test for induction of prophage with and without exogenous metabolic activation: No statistically significant induction of prophage by either compound. Extent of DNA damage detected by single-cell gel electrophoresis (comet) assay: At 0.2 and 1: MMA111 » DMA111 = As111. At 100: all 3 chemicals had roughly the same level of DNA damage as MMA111 had at 0.2, but MMA111 still has significantly more DNA damage than the other two chemicals. Extent of DNA damage detected by single-cell gel electrophoresis (comet) assay: At 0.2 and 1: MMA111 » DMA111 = As111. At 40 and 100: DMA111 > MMA111 > As111. Reference Kligerman et al., 2003 Gomez et al., 2005 Gomez et al., 2005 C-214 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells HL-60 cells CHO-K1 cells Human peripheral blood lymphocytes from 2 donors, with results reported separately Arsenic Species As111 SA for all As111 SA MMAm DMA111 Concentration(s) Tested (nM) 0.25,0.5, 1 0.25,0.5, 1 0.25,0.5, 1,2 5, 10 2.5, 5, 10, 20, 40, 80, 100 2.5, 5, 10, 20, 40, 80 Duration of Treatment 4 hr for all 4 hr for all LOECa (HM) 0.25 0.25 None None 2.5 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) The LOECS shown are for DNA strand breaks (termed ADSB by the authors) detected by the comet assay without any additional treatments of DNA to digest and reveal ODA or DPC. They also treated the damaged DNA with FPG or PK to yield estimates of ODA or DPC, respectively. TheLOEC was 0.25 for all 3 cell types for ODA, DPC, or ODA+DPC. Clearly much more DNA damage is revealed by treatments with FPG, PK, or both. DNA damage was induced at levels causing no cytotoxicity. The LOECs apply to the extent of DNA damage detected by SCGE (comet) assay at pH > 13. There was no cytotoxicity at doses up to 20. Much lower responses for all arsenicals were seen in comet assay at pH of 12.1, with the difference between this and pH 13 being defined as alkaline labile sites. DNA damage by both methylated arsenicals was markedly reduced by co-exposures to the antioxidants Se-Met or vitamin C. DNA-double strand breaks were not induced. Reference Wanget al., 2001 Soto- Reyes et al., 2005 C-215 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MRC-5 cells MRC-5 cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 2.5, 5, 10 2.5, 5, 10 Duration of Treatment 2hr 2hr LOECa (HM) 2.5 2.5 for SSBs 10 for protein- DNA adducts Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) DNA SSBs detected by the standard alkaline (pH > 13) comet assay: ft with dose of both tail length and tail moment at doses of 2.5 and 5, but a U for both effects at dose of 10 to less than effect seen at dose of 2.5. NSE on cytotoxicity at any of the tested doses. Protein-DNA adducts and DNA SSBs detected by alkaline (pH > 13) comet assay done with and without posttreatment with proteinase K, respectively: Experiment without proteinase K: ft of both tail length and tail moment at doses of 2.5 and 5, but a U of both effects at dose of 10 to less than effect seen at other doses. Experiment with proteinase K: ft of both tail length and tail moment at doses of 2.5 and 5, and a further large ft in both parameters at dose of 10. NSE on cytotoxicity at any of the tested doses in either experiment. Evidence for protein-DNA adducts (or crosslinks) came from ft observed at dose of 10, which is thus the LOEC for that effect. Proteinase K breaks the crosslinks that hinder the DNA fragmentation caused by the DNA SSBs. Reference Mouron et al., 2006 Mouron et al., 2006 C-216 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MRC-5 cells MRC-5 cells Arsenic Species DMAV DMAV Concentration(s) Tested (nM) 125, 250, 500 125, 250, 500 Duration of Treatment 2hr 2hr LOECa (HM) 500 for U in SSBs (see row below) 125 for both protein- DNA adducts and SSBs Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) DNA SSBs detected by the standard alkaline (pH > 13) comet assay: slight ft in tail moment (TM) at dose of 125 (a NSE); point estimates of TM were below control at 2 higher doses, with that at 500 being significantly below it; actual data: TMs: 0, 13.4; 125, 14.6; 250, 13.1; 500, 9.7. NSE on cytotoxicity at any of the tested doses. Protein-DNA adducts and DNA SSBs detected by alkaline (pH > 13) comet assay done with and without posttreatment with proteinase K, respectively: Experiment without proteinase K (buffer only): progressive U in tail moment (TM) with increasing dose; actual data: TMs: 0, 7.7; 125, 6.7; 250, 5.3; 500, 4.9. Experiment with proteinase K: ft in TM, with a positive dose- response; actual data: TMs: 0,8.3; 125, 11.9; 250, 22.2; 500, 23.3. NSE on cytotoxicity at any of the tested doses in either experiment. Proteinase K breaks the crosslinks that hinder the DNA fragmentation caused by the DNA SSBs. Reference Mouron et al., 2005 Mouron et al., 2005 C-217 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue L-132 cells (human alveolar type II cells) HepG2 cells NB4 cells HL-60 cells Arsenic Species DMAV As111 SA As111 SA As111 ATO MMAm DMA111 As111 ATOc Concentration(s) Tested (nM) 5,7.5, 10 mM 7.5 0.5 12.5, 25, 50 Duration of Treatment 12 hr 24 hr 30 min 24 hr LOECa (HM) 5mM 7.5 0.5 12.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) DNASSB detected by alkaline elusion: there was a dose-response. Early in the exposure period, there was marked suppression of replicative DNA synthesis, and the chain length of the nascent DNA was shorter than that of the control, which suggests that the template DNA was modified by more than just strand breaks. Induction of DNA DSBs by immunodetection of yH2A.Xfoci: ft to ~6x control level; co-treatment with 170 nM SAM did not change the induced DSB frequency. Experiments with EN111, FPG and NE (from NB4 cells) as well as experiments using immunodepletion of NE with antibodies directed against proteins known to be involved in excision repair suggest that these trivalent arsenicals induce only oxidative DNA adducts and that OGG1,MYH and APE are involved in the excision of the oxidative DNA adducts. DNA damage detected by alkaline SCGE (comet) assay: while the response was barely statistically significant at the lowest dose, it was strong at the other 2 doses, with a positive dose-response. Reference Tezuka et al., 1993 Ramirez et al., 2007 Puetal., 2007 Yedjou and Tchounwo u, 2007 C-218 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PARP-1+/+ MEF cells PARP-r7' MEF cells PARP-1+/+ MEF cells PARP-1"'" MEF cells HaCaT cells CRL1675 cells THP-1 + A23 187 cells Arsenic Species As111 SA for both As111 SA for both As111 ATO for all Concentration(s) Tested (nM) 11.5,23 for both 11.5,23 for both 72-hr LD10 and LD25 for each cell line: 1.9, 15.2 1.0, 1.9 1.9,3.8 Duration of Treatment 30 min for both 24 hr for both 72 hr for all LOECa (HM) 11.5 11.5 11.5 11.5 15.2 None 1.9 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Extent of DNA damage detected by SCGE (comet) assay at pH >13, reported as induced damage (experimental - control) in units of TM length: -0.4 at 11. 5, -0.7 at 23. -2.9 at 11. 5, -3. 4 at 23. All 4 estimates were statistically significant. Extent of DNA damage detected by single-cell gel electrophoresis (comet) assay at pH >13, reported as induced damage (experimental - control) in units of tail moment length: -2.0 at 11. 5, -3. 6 at 23. -4.8 at 11.5, -5. 5 at 23. All 4 estimates were statistically significant. DNA single-strand breaks detected by SCGE (comet assay) following alkaline treatment: NSE at LD10; U at LD25 (perhaps stimulates repair). NSEatLD10;NSEat LD25. ft at LD10; ft at LD25. Reference Poonepalli et al., 2005 Poonepalli et al., 2005 Graham- Evans et al., 2004 C-219 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells CRL1675 cells THP-1 + A23 187 cells 293 cells Arsenic Species As111 ATO for all As111 ATO Concentration(s) Tested (nM) 72-hr LD10 and LD25 for each cell line under chronic-exposure conditions, as follows: 2.0, 4.0 0.5, 1.3 0.5,5.1 1 Duration of Treatment Under chronic exposure conditions: 72 hr for all 6hr LOECa (HM) 2.0 1.3 0.5 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Testing for DNA single- strand breaks was preceded by exposure to 1.0uMAsmATOforat least 8 passages to establish chronic- exposure conditions. Then, following exposures to various doses for 72 hr, DNA single-strand breaks were detected by single- cell gel electrophoresis (comet assay) following alkaline treatment: ftatLD10;ftftatLD25. NSEatLD10;ftatLD25. ftft at LD10; ftft at LD25. DNA damage reported in units of tail moment in a comet assay that used nuclear extraction incubation: untreated = -11 units; dose of 1 : big ft to -58 units. Effects of co-treatment (CoTr) with modulators at high doses: CoTr 200 uMDMSA:li from inorganic arsenic alone to -38 units. CoTr 100 uMDMPS:U from inorganic arsenic alone to -39 units. Effects of CoTr with modulators at low doses: CoTr 20 uM DMSA: ft from inorganic arsenic alone to -104 units. CoTrlOuMDMPS:ft from inorganic arsenic alone to -84 units. Reference Graham- Evans et al., 2004 Jan et al., 2006 C-220 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SV-HUC-1 cells UROtsa cells UROtsa cells E. coli strain WP2S(X) Arsenic Species As111 ATO As111 SA MMAm As111 SA MMAm As111 SA Concentration(s) Tested (nM) 1 1, 10 0.05,0.5,5 1, 10 0.05,0.5,5 Up to 3.2mM Duration of Treatment 6hr 30 min for both 60 min for both 20 hr LOECa (HM) 1 1 0.05 10 0.05 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) DNA damage reported in units of tail moment in a comet assay that used nuclear extraction incubation: untreated = -10 units; dose of 1: big ft to -49 units. Effects of CoTr with modulators at high doses: CoTr 200 uMDMSA:ll from inorganic arsenic alone to -34 units. CoTr 100 uMDMPS:U from inorganic arsenic alone to -35 units. Effects of CoTr with modulators at low doses: CoTr20uMDMSA:ft from inorganic arsenic alone to -99 units. CoTrlOuMDMPS:ft from inorganic arsenic alone to -89 units. Detection of 8-oxo-dG (measure of oxidative DNA damage): ft to 3 x control at 1, ft to 2x control at 10. ft to 5x control at 0.05, ft to 4x control at 0.5, NSE at 5. Detection of 8-OHdG formation (measure of oxidative DNA damage): NSE at 1, big U from control at 10. ft to 3x control at 0.05, ft to 3. 3x control at 0.5, ft to 4.3 x control at 5. Thus MMAm showed a time delay just as it did for ROS production. No induction of I phage (part of "SOS" system) using 8 serial 2-fold dilutions from a concentration that inhibits growth. Reference Jan et al., 2006 Eblin et al., 2006 Eblin et al., 2006 Rossman etal., 1984 C-221 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human- hamster hybrid AL cells HaCaT cells TK6 cells 4>X174 RF I DNA Naked double- stranded circular DNA Arsenic Species As111 SA As111 SA Asv As111 SA As111 ATO As111 un- specified MMAm DMA111 Concentration(s) Tested (nM) 30.8 5, 10, 20, 30 10, 20, 30, 50, 100 0.1, 1, 10 for both 10|aM-30 mM in log increments 10, 20, 30, 40, 50 37.5, 75, 150, 300, 1000 Duration of Treatment 24 hr 24 hr 22 hr for both 24 hr for all LOECa (HM) 30.8 10 20 1 10 None 10 37.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induction of 8-OHdG; co-treatment with SOD or catalase considerably reduced induction of this oxidative DNA damage. Induction of 8-OHdG; pre-incubation with SOD, CATorDMSO almost completely blocked this. Oxidative DNA damage by 20 uM As111 SA: pre- incubation with MnTMPyP, Z-NAME or FeTMPyP substantially blocked such damage. Induction of DPCs detected by a decrease in DNA damage detected in the comet assay when an arsenic treatment was followed by exposure to lor2Gyof69cGy/min gamma radiation. The DPCs kept the damaged DNA from moving during electrophoresis. While both SA and ATO caused a significant effect, the effect was more pronounced for SA. Nicked DNA in DNA nicking assay. Reference Kessel et al., 2002 Ding et al., 2005 Hornhardt etal.,2006 Nesnow et al., 2002 C-222 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Supercoiled DNA (plasmid pBR 322); similar results were found for plasmid cpX174, but details were not reported Arsenic Species As111 SA Asv MMAm MMAV Mono- methyl- arsine DMA111 DMAV Dimethyl- arsine Tri- methyl- arsine Concentration(s) Tested (nM) >5mM >5mM >5mM >5mM > 5 mM <5 mM >5mM <0.5mM <0.5mM Duration of Treatment 2hr for all LOECa (HM) None None >5mM None > 5 mM <5 mM None <0.5mM <0.5mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Damage to DNA detected by agarose gel electrophoresis: The arsines were produced in aqueous reaction mixtures of sodium borohydride and the appropriate arsenical. Trimethylarsine and dimethylarsine were about 100 times more potent than DMA111. WhenNADHor NADPH, which are biological hydride donors, were incubated with DMA111 for 2 hr, DNA damage was increased by at least 10- fold, possibly because of the generation of dimethylarsine. Reference Andrewes et al., 2003 DNA Repair Inhibition or Stimulation CHO Kl cells As111 SA 5, 10, 20, 40, 80 6 hrs 5 DNA single-strand breaks detected by alkaline elution: those induced by MMS were repaired after incubation in a drug-free medium for 6 hr; however, posttreatment with sodium arsenite accumulated MMS- induced breaks with a dose-response for the arsenite exposure. Both alkali-labile sites and frank breaks were enhanced, with the latter occurring at higher doses of MMS and arsenite. Lee-Chen etal., 1993 C-223 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue V79 cells, strain 743 -3 -6 HeLa S3 cells Arsenic Species As111 SA for both MMAm MMAV DMA111 DMAV Concentration(s) Tested (nM) 10 uM 0.0001,0.001, 0.01,0.1, 1 0.01,0.1, 1, 10, 100, 500 0.0001,0.001, 0.01,0.1 0.01,0.1, 1, 10, 100, 250 Duration of Treatment Stu- Por all: 18hr + 5 min more while also being treated with 100 uM H202 LOECa (UM) 10 uM 0.001 None 0.001 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Similar decreases in inducible total nuclear DNA ligase activity and in inducible nuclear DNA ligase II activity were demonstrated after arsenic treatments given before or after MNU treatments, thereby demonstrating that most of the inhibited activity was DNA ligase II. Effect on H2O2-induced poly(ADP-ribosyl)ation: U with dose, 59% of control at dose of 1. NSE. U with dose, 49% of control at dose of 0. 1 . NSE. Other experiments showed that the above effects were real decreases (not merely delayed responses). All above measurements were at dose levels with little to no cytotoxicity. After 18 hr incubation, these arsenicals had NSE on the extent of gene expression of PARP-1 at doses up to 0.1 and 100 for methylated and pentavalent arsenicals, respectively. Reference Li and Rossman, 1989b Walter et al., 2007 C-224 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Isolated recombinant PARP-1 Jurkat cells Jurkat cells HLFC cells HLFK cells (Ku70 deficient) Arsenic Species As111 SA MMAm DMA111 As111 SA As111 SA As111 SA for both Concentration(s) Tested (nM) 10, 50, 100, 200, 500 for all 0.01,0.1, 1,5, 10 1 1,2.5,5, 10 for both Duration of Treatment For all: 10 min preincubation before PARP-1 reaction with a nicked plasmid as substrate 24hr 24 hr 2hr for both LOECa (HM) 10 10 10 0.01 1 2.5 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Effect on activity of PARP-1: U with dose, 58% of control at dose of 500. U with dose, 24% of control at dose of 500. U with dose, 15% of control at dose of 500. These data suggest that trivalent arsenicals inhibit cellular poly(ADP-ribosyl)ation by reducing PARP-1 activity. U ERCC1 mRNA level; not said to be statistically significant until dose of 1, but means + SDs suggest 45% U at 0.01 and 60% U at 0.1. Decreases of 60%, 95%, and 85% at doses of 1, 5, and 10, respectively U in repair following a 2- hr in vitro treatment with 4 uM 2-AAAF immediately after the inorganic arsenic treatment. DNA damage measured by SCGE (comet) assay: inorganic arsenic group had ft DNA damage after 2-hr 2-AAAF treatment and following a 4-hr repair period. No difference in DNA damage before 2-AAAF. DNA DSB damage as measured with neutral SCGE assay: This type of damage was significantly greater for HLFK than HLFC at all 4 doses. Reference Walter et al., 2007 Andrew et al., 2006 Andrew et al., 2006 Liu et al., 2007b C-225 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HLFC cells HLFK cells (Ku70 deficient) CHO-K1 cells GM847 cells HaCaT cells W138 cells for both Arsenic Species As111 SA for both As111 SA As111 SA for both As111 SA for both Concentration(s) Tested (nM) 5 for both 0.1,0.5, 1,5, 10 0.1,0.5, 1,5, 10 for both 0.1,0.5, 1,5, 10 for both Duration of Treatment 2hr for both 24 hr 24 hr for both 24 hr 48 hr LOECa (HM) 5 5 0.1 0.1 0.1 0.1 0.1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) The LOECs are for induction of DNA DSBs. After the 2-hr As111 treatment, cells were incubated in arsenic-free medium to measure repair of DNA DSBs using the neutral SCGE assay at 0.5, 1, 1.5, and 2 hr. At all time points there was significantly and substantially less repair in HLFK, showing that the Ku70 deficiency decreases the efficacy of repair of arsenic-induced DSBs. DNA polymerase (3 promoter activity: big ft at 0.1; slight ft at 0.5; no effect at 1; big U at 5 and 10. DNA polymerase (3 protein levels: Big ft at 0.1 and 0.5, slight ft at 1, no effect at SandbigUatlO. Big ft at 0.1 and 0.5, no effect at 1, big U at 5 and 10. DNA ligase activity: ft at 0.1, big ft at 0.5, huge ft at 1, U at 5, big U at 10. No effect at 0.1, big ft at 0.5 and 1, no effect at 5, big U at 10. Two other experiments of 72 and 96 hr duration showed generally even more subdued increases and decreases than the 48-hr experiment. Reference Liu et al., 2007b Snow et al., 2005 Snow et al., 2005 Snow et al., 2005 C-226 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested (nM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Reference Effects Related to Oxidative Stress (ROS) Hepa-1 cells (mouse hepatoma) stably transformed with pEpREpgeo WI38 (human fibroblasts) Purified thioredoxin enzyme from mouse liver; to test the NADPH- dependent reduction of DTNB Primary culture of rat hepatocytes Human- hamster hybrid AL cells As111 SA As111 SA As111 SA MMAm DMA111 Asv MMAV DMAV AsmSA MMAm As111 SA 0.1,1,5,25,50 0.05,0.5,5 (24 hr pretreatment) followed by 60 min exposure to H2O2 at 1, 10 or 50 mM for 1 hr and then 24 hr to recover -0.2-800 -0.2-800 -0.2-800 -10-6000 -10-6000 -10-6000 1-50 0.1-10 30.8 6hr 24 hr pretreatment — 30 min for both Within 5 min 5 0.05 -100 -0.2 ~3 -300 — — 30.8 Activated a |3- galactosidase gene reporter system: suggests there was induced oxidative stress — 5.6- fold response; progressively and markedly decreasing responses at 2 higher doses. Extent of viability determined by NR assay: Compared to control cells exposed to H2O2, with no pretreatment: ft viability at 1 mM H2O2 only. At dose of 5, there was an ft in viability at 10 mM H2O2 but a U in viability at 50 mm H2O2. Approximate IC50s (inhibition of enzyme activity): -200. -0.4. -30. -3000. Never more than —80% inactivation. Never more than —80% inactivation. Decreased thioredoxin enzyme activity (the NADPH-dependent reduction of DTNB) IC50: » 100. IC50: ~3. Production of ROS, measured by ESR and with about a 3 -fold increase in amplitude of signals; concurrent treatment with the radical scavenger DMSO eliminates the effect. Maier et al., 2000 Snow et al., 2001 Linetal., 1999 Lin et al., 2001 Liu et al., 2001 C-227 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human aorta VSMCs (vascular smooth muscle cells) HFW cells (diploid human fibroblasts) Jurkat cells Arsenic Species As111 SA As111 SA As111 SA MMAm DMA111 Concentration(s) Tested (nM) -1.2,2.5,5, 10 1.25,2.5,5, 10 5, 10, 20, 40, 80 0.2, 10, 20, 100 for all Duration of Treatment 4hr 24 hr 4hr 2hr 2hr 2hr LOECa (HM) -1.2 1.25 20 None 10 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Numerous experiments in this study led to the conclusion that arsenite activates NADH oxidase to produce superoxide, which then causes oxidative DNA damage. Micronuclei were induced in both protocols; the yield of micronuclei was greatly reduced by the presence of the antioxidants catalase or NAC (the precursor of GSH), which suggests that oxidative stress was involved in the induction of micronuclei. Level of intracellular peroxides determined by flow cytometry using cell permeable fluorogenic marker DHR123: At 10 and 20: DMA111 » MMAin»Asm. AtlOO:MMAin>DMAm about equal to As111. (Cell lysis may explain the reduction of DMA111 at dose of 100 to 1/3 level seen at 20.) Control value was not reported. If control value was actually 0 (and thus the baseline in the figure), then the LOEC for all 3 arsenicals would have been 0.2, with a rather similar slight response for all of them. Reference Lynn et al., 2000 Yihand Lee, 1999 Gomez et al., 2005 C-228 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Whole blood lymphocytes from 2 human donors, with results reported separately HaCaT cells L-132 cells Arsenic Species MMAm DMA111 As111 SA Asv DMAV Concentration(s) Tested (nM) 2.5, 5, 10, 20 for both 5, 10, 15, 20 for both 10 mM alone 10 mM + 0.5 mM PQ Duration of Treatment 4 hr for all 24hr 2hr Ihr LOECa (HM) 2.5 10 5 10 None 10 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Levels of MDA as lipid peroxidation marker in human plasma: For MMA111 both donors showed significant increase over control at all doses except 10, for which only 1 was significant. For DMA111 both donors showed significant increase over control at 20, but only 1 did at 10. There was no cytotoxicity at the dose levels tested. Induction of 3 -NT, which is a diagnostic marker for RNS in vivo; pre-incubation with SOD, MnTMPyP, L- NAMEorFeTMPyP almost completely blocked this protein damage by 20 uM As111 SA; pre-incubation with CAT or DMSO had no effect, in sharp contrast to what happened for ROS-damage to DNA. DNA single-strand breaks detected by alkaline elusion: co- exposure with PQ or sequential exposures of 1 hr (with either one first) yielded a strong response. Reference Soto- Reyes et al., 2005 Ding et al., 2005 Kawaguchi etal., 1996 C-229 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PAEC cells harvested from freshly isolated vessels NB4 cells PAEC from freshly harvested vessels HFW cells Cell free buffer Arsenic Species As111 SA As111 ATO As111 probably ATO, but called arsenite As111 SA DMAmI Concentration(s) Tested (nM) 5, 10 1 5 5, 10, 20 — Duration of Treatment Ihr 4hr 5-15 min 24 hr — LOECa (HM) 5 1 5 5 — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Various experiments showed that inorganic arsenic activates a NADPH-dependent oxidase located in the plasma membrane that results in superoxide accumulation. Both the subunits of the oxidase were shown to be essential for the response, and the oxidase is dependent on exogenous NAD(P)H for activity. The peak effect occurred within 1 hr and was higher at a dose of 5 than 10. Generation of ROS led to decrease (and eventual loss, with continued treatment) of mitochondria! membrane potential, with subsequent outer mitochondria! membrane permeability changes. ff in superoxide and H2O2 accumulation. DCF fluorescence to indicate formation of cellular oxidants; co- treatment with BHT (a radical scavenger) completely blocked this effect. Oxidative damage was induced in thymine to form cis-thymine glycol. SOD and CAT did not alter this reaction. Other tests suggest that the reaction requires the formation of a reactive arsenic peroxide, probably dimethylated arsenic peroxide. Reference Smith et al., 2001 Jing et al., 1999 Barchowsk y et al., 1999b Lee and Ho, 1995 Yamanaka etal.,2003 C-230 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Postconfluent PAEC cells in a monolayer Human- hamster hybrid AL cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (nM) 1, 2.5, 5, 10, 20 11.5, 15.4 Duration of Treatment 30 min 24 hr LOECa (HM) 1 11.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) DCF fluorescence as a direct measure of intracellular oxidant concentrations (i.e., accumulation of ROS): likely ft at all doses, with a peak at 5 that is -45% higher than control, a difference that is statistically significant. Induction of CD59" mutations: dose-related increase in mutation frequency; pretreatment + co-treatment with L- NMMA (a nitric oxide synthase inhibitor) substantially reduced the mutation frequencies at both doses. Similar treatment with D- NMMA (the inactive enantiomer) had no effect. These findings were taken as evidence that peroxynitrites have an important role in inorganic arsenic- induced genotoxicity. That conclusion was supported by a Western blot analysis of nitrotyrosine-modified proteins induced by inorganic arsenic treatments and mostly blocked by L-NMMA. Reference Barchowsk y et al., 1996 Liu et al., 2005 C-231 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HepG2 cells NB4 cells NB4-M-AsR2 cells IM9 cells Arsenic Species As111 ATO As111 ATO for all Concentration(s) Tested (nM) 20 0.5, 1 2,4 0.5, 1 Duration of Treatment 6hr 24hrs for all LOECa (HM) 20 0.5 2 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Analysis of 481 selected genes in a DNA microarray experiment: hierarchical clustering analysis showed that inorganic arsenic exposure closely resembled DMNQ exposure (and was extremely different from DMN or phenol exposure) regarding patterns of genes that were up-regulated and down-regulated. In phase 1 of this experiment, DMNQ was selected as a model chemical that generates ROS and is known to induce genes associated with cell proliferative responses. Exposure to inorganic arsenic caused significant up-regulation of 38 genes and down- regulation of 20 genes; dose used had >80% cell viability. HMOX-1 protein (a stress-responsive protein) levels after treatment with ATO alone and co-treatment with 100 uM Trolox: At 0.5: slight ft alone, big ft with Trolox; at 1: ft alone, huge ft with Trolox. At 2: slight ft alone, big ft with Trolox; at 4: ft alone, huge ft with Trolox. At 0.5: slight ft alone, big ft with Trolox; at 1: ft alone, huge ft with Trolox. Reference Kawata et al., 2007 Diazetal., 2005 C-232 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells BFTC905 cells and NTUB1 cells BFTC905 cells and NTUB1 cells Arsenic Species As111 ATO for all As111 SA Asv MMAm MMAV DMA111 DMAV As111 SA Asv MMAm MMAV DMA111 DMAV Concentration(s) Tested (nM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Regarding row above, other indications that Trolox potentiates ATO-mediated oxidative stress: bigger ft in protein carbonyls (indicator of oxidative damage to proteins) and 8-iso-PGF2a (indicator of lipid peroxidation) by combined ATO and Trolox treatment(s) than by ATO treatment(s) alone. Other experiments showed that the synergistic effect of Trolox on ATO-mediated apoptosis was not related to extracellular H2O2 production. ATO was shown to induce the formation of Trolox phenoxyl radicals by electronic spin resonance spectroscopy. 0.2 for all 0.2 for all 24 hr for all 24 hr for all 0.2 0.2 0.2 None 0.2 None 0.2 0.2 0.2 0.2 0.2 None Relative extent of oxidative damage (peroxidation) in lipids, measured as malonaldehyde formation; ranking of those with statistically significant ft over control (i.e., unranked arsenicals hadNSE): In BFTC905 cells: Asin>DMAin>MMAm» Asv in NTUB1 cells: DMAIII»MMAIII>Asm. Relative extent of oxidative damage (carbonylation) in proteins; ranking of those with statistically significant ft over control (i.e., unranked arsenicals hadNSE): In BFTC905 cells: MMAni>Asin>DMAm» Asv In NTUB1 cells: Asin>MMAin>DMAm» Asv>MMAv. Consistent with these effects, increased levels of nitric oxide, superoxide ions, hydrogen peroxide, and the cellular free iron pool were consistently detected in both cell lines after treatments by the 3 trivalent arsenicals. Reference Diaz et al., 2005 Wanget al., 2007 Wanget al., 2007 C-233 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue BFTC905 cells and NTUB1 cells Gclm"'" MEF cells, from GCLM knockout mice NB4 cells 1RB3AN27 cells Arsenic Species As111 SA Asv MMAm MMAV DMA111 DMAV As111 SA for all AsmATO As111 SA Concentration(s) Tested (nM) 0.2 for all Duration of Treatment 24 hr for all LOECa (HM) 0.2 0.2 0.2 0.2 0.2 0.2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Relative extent of oxidative damage (comet assay) in DNA; ranking of those with statistically significant ft over control (i.e., unranked arsenicals hadNSE): Without enzyme digestion: In BFTC905 cells: Asm = MMAm>MMAv> DMAV. In NTUB1 cells: Asm = MMAln>DMAm = MMAm = DMAV. WithEnm + FPG digestion: In BFTC905 cells: Asni>MMAin>DMAm> MMAV. In NTUB1 cells: Asln>MMAln>DMAm>M MAV> DMAV = Asv. See rows under Apoptosis and Cytotoxicity for this citation for experimental conditions. The high level of arsenic-induced oxidative stress from some treatments was not significantly decreased by tBHQ. Yet, tBHQ pretreatment or co-treatment greatly decreased inorganic arsenic induced apoptosis and cytotoxicity. 0.75 0.1,0.5, 1,5, 10 Results were obtained from various experiments, including Affymetrix oligonucleotide microarray analysis using a chip that contained 22,000 open reading frames from the human genome. Treatment for 10 days increased the expression of a set of genes responsible for ROS production. Genes were identified that responded to inorganic arsenic and H2O2 but whose response to inorganic arsenic was reversed by NAC. It was found that 26% of the genes significantly responsive to inorganic arsenic might have been directly altered by ROS. Inorganic arsenic treatment induced ROS, which in turn oxidized the Spl transcription factor, with a corresponding decrease in its in situ binding to the promoters of the 3 genes hTERT, C17, and c-Myc, with the result that their expressions were significantly suppressed (e.g., hTERT: U expression to < 1% normal). 2hr 0.5 ROS production using DCFH-DA assay: ft with a positive dose-response; the increase at dose of 1 was blocked by co- treatment with either NAC or a-Toc. Reference Wanget al., 2007 Kannet al., 2005b Chou et al., 2005 Felix etal., 2005 C-234 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue BEAS-2B cells Embryonic mesenchymal cells prepared fromCF-1 mouse conceptuses at gestation day 11 RAW264.7 cells HELP cells Arsenic Species As111 ATO As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 5, 10, 20 5.8, 11.6, 15.4 2.5,5, 10,25 0.1,0.5, 1,5, 10 Duration of Treatment 24hr 2hr 3hr 4hr LOECa (HM) 5 5.8 5 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Production of 8- isoprostane, a by-product of lipid peroxidation: ft with a positive dose- response; 2x control at 5, 6x control at 20. In addition, electron spin resonance studies (involving co-treatments with CAT, SF, NAC, or NADPH) and confocal microscope studies showed that inorganic arsenic can produce ROS, such as H2O2 and 'OH, as a result of reduction reactions within cells. Production of ROS detected by a variant of the DCF assay using CM-H2DCFDA: Induced RFUs (i.e., experimental - control): 5.8, -950; 11.6, -2050; 15.4, -2700. 15-min pretreatment with 0.2or0.5%(v/v)DMSO blocked all or almost all inorganic arsenic- induced production of ROS at dose of 15.4, whereas 15-min pretreatment with 0.1% (v/v) DMSO partially blocked it. Extracellular H2O2 production quantified using the Amplex Red Hydrogen Peroxide Assay: there was a positive dose-response, reaching ~1.4x control. Production of ROS detected by the DCFH- DA assay in the 15 min after inorganic arsenic treatment: ft with dose to >2x control at dose of 10. Reference Hanetal., 2005 Perez- Pasten et al., 2006 Szymczyk etal.,2006 Yanget al., 2007 C-23 5 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HELP cells NB4 cells Arsenic Species As111 SA As111 ATO Concentration(s) Tested (nM) 0.1,0.5, 1,5, 10 1,3 Duration of Treatment 3, 6, 12, 24, or48hr 16 hr LOECa (UM) Various 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) SOD activity after 24 hr: ft at 0.5, U at 5 and 10; hint of similar change of direction in response also in treatments of some other durations. GPx activity after 24 or 48 hr: U at 5 and 10; but hints of ft at lower doses and U at higher doses in treatments of some durations. MDA content (measure of LPO) after 24 or 48 hr: ft at 5 and 10; tended to increase with time and dose in treatments of all durations. Effect on cellular total antioxidant capacity determined using the ABTS assay (Troiloc -equivalent antioxidant capacity in units of nmol/mg protein): Control = -420; inorganic arsenic at dose of 3: -150; inorganic arsenic at dose of 1 : -240. Effects of co-treatment (CoTr): inorganic arsenic at 3 + CoTr with 1000 uM DTT: -275. inorganic arsenic at 3 + CoTr with 2000 uM DTT: -340. inorganic arsenic at 1 + CoTr with 25 uMDTT: -150. inorganic arsenic at 1 + CoTr with 50 uM DTT: -125. Reference Yanget al., 2007 Jan et al., 2006 C-236 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells BFTC905 cells and NTUB1 cells A549 cells Arsenic Species As111 ATO DMAV As111 ATO Concentration(s) Tested (nM) 0.5 1,2 2 Duration of Treatment 2hr 24 hr for all 48 hr LOECa (HM) 0.5 1 in at least one cell line for all 3 effects Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Intracellular H2O2 level (units of Amplex red assay): control = -20; inorganic arsenic -45. Effects of co-treatment (CoTr): CoTr with 80 uMDTT: -72. CoTr with 100 uM DMSA: -67. CoTr with 20 uM DMPS: -72. ft oxidative damage (peroxidation) in lipids, measured as malonaldehyde formation: at both doses in BFTC905 cells, at doseof2inNTUBl cells. ft oxidative damage (carbonylation) in proteins: at higher dose inBFTC905cells,at lower dose in NTUB1 cells. ft oxidative damage (comet assay) in DNA, without enzyme digestion: at both doses in both cell lines. Loss of MMP determined by flow cytometry using JC-1: 2 uM inorganic arsenic: ft to ~1.25x; 200 uM sulindac: ft to ~1.15x; (2 uM inorganic arsenic + 200 uM sulindac): ftto~1.9x. There was also a synergistic interaction between these treatments in causing big ft in cytochrome C protein level in the cytosol, which is thought to result from damage to mitochondria! membranes that permits cytochrome C release to the cytosol. Pretreatment with NAC almost entirely blocked the MMP and cytochrome C effects. (Sulindac is a NSAID that inhibits COX-2.) Reference Jan et al., 2006 Wang et al., 2007 Jinetal., 2006b C-237 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue A549 cells HeLa cells BAEC cells BAEC cells HEK 293 cells and SV-HUC-1 cells Arsenic Species As111 ATO As111 SA As111 SA As111 SA As111 SA MMAm DMA111 Concentration(s) Tested (nM) 2 10, 100 5, 10 10 0.2 for all Duration of Treatment 48hr 4hr Ihr 24 hr 24 hr for all LOECa (HM) 2 10 for Trxl and Trx2; none for GSH/ GSSG 5 10 0.2 0.2 0.2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Production of ROS using carboxy-H2DCFDA assay: control, -0.8 unit; 2 uM inorganic arsenic, -4.2 units; 200 uM sulindac: ~4.5x; (2 uM inorganic arsenic + 200 uM sulindac): ~7.5x. Thus there was only additivity. Pretreatment with NAC before combined treatment: U to -3.9 units. Effects on Trxl and Trx2 redox states determined using Redox Western blot methods: Trxl : ft in oxidation at 10, slightly bigger ft at 100. Trx2: huge ft in oxidation at 10, slightly bigger ft at 100. In contrast, inorganic arsenic had little effect on the GSH/GSSG redox state, as determined by HPLC. ft in peroxynitrite to ~1.4xand~1.6xat 5 and 10, respectively. ft in nitrotyrosine formation to ~1 . 15x. Relative extent of oxidative damage (peroxidation) in lipids, measured as malonaldehyde formation; ranking of those with statistically significant ft over control (all were significant): In HEK 293 cells: Asin»MMAin>DMAm. In SV-HUC-1 cells: Asin>DMAm>MMAni. Reference Jinetal., 2006b Hansen et al., 2006 Bunderson etal.,2006 Bunderson etal.,2006 Wanget al., 2007 C-23 8 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HEK 293 cells and SV-HUC-1 cells TRL1215 cells TRL 1215 cells pretreated with 50 uM BSOfor24hr to deplete GSH levels and then co- treated with 50uMBSO TRL 1215 cells TRL 1215 cells pretreated with 50 uM BSOfor24hr to deplete GSH levels and then co- treated with SOuMBSO Arsenic Species As111 SA MMAm DMA111 MMAV for both MMAV for both Concentration(s) Tested (nM) 0.2 for all 5mM 5 mM Duration of Treatment 24 hr for all 24 hr 48 hr LOECa (HM) 0.2 0.2 0.2 None 5mM None 5 mM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Relative extent of oxidative damage (comet assay) in DNA; ranking of those with statistically significant ft over control (all were significant): Without enzyme digestion: In HEK 293 cells: Asni>MMAm = DMA111. In SV-HUC-1 cells: MMAm = As111 = DMA111. WithEnm + FPG digestion: In HEK 293 cells: As111 = MMAm = DMA111. In SV-HUC-1: Asni>DMAm = MMAm. Cellular ROS levels based on DCFH-DA assay: MMAV: NSE. MMAV + BSD: ft to ~2.22x . Cell survival determined by AB assay: MMAV: 100% survival. MMAV + BSO:~3% survival. Co-treatment with 10 mM DMPO during the MMAV + BSO treatment blocked most of the cytotoxicity, resulting in -72% survival. DMPO effectively scavenged cellular radical molecules. Reference Wanget al., 2007 Sakurai et al., 2005a Sakurai et al., 2005a C-239 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue TRL1215 cells TRL 1215 cells pretreated with 50 uM BSO for 24 hr to deplete GSH levels and then co- treated with 50uMBSO HeLa cells HeLa cells Jurkat cells Namalwa cells NB4 cells U937 cells Arsenic Species MMAV for both As111 ATO As111 ATO As111 ATO for all Concentration(s) Tested (nM) 5mM 2 2 2 Duration of Treatment 48 hr Various up to24hr Ihr 24 hr for all LOECa (HM) None None with DMPO 2 2 None 2 2 None Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Caspase 3 activity (related to apoptosis): MMAV: NSE. MMAV + BSO: ft to ~1.66x. Co-treatment with 10 mM DMPO during the MMAV + BSO treatment completely blocked the ft of caspase 3 activity. DMPO effectively scavenged cellular radical molecules. ROS levels were shown by DCFH-DA assay to be significantly elevated by inorganic arsenic and to ft roughly 3x higher than for inorganic arsenic alone following a combined inorganic arsenic plus 10 uM emodin treatment; the addition of 1.5 mMNAC as a co-treatment attenuated (but did not completely block) that ft in ROS levels. Analysis of GSH/GSSG ratios showed that co- treatment of inorganic arsenic with emodin had a major oxidative impact on the cellular redox state, as shown by following ratios: control, -62; inorganic arsenic, -52; 10 uM emodin, -34; inorganic arsenic plus 10 uM emodin, -13; pretreatment with 1.5 mM NAC attenuated (but did not completely block) this effect. ft in H2O2 levels as detected by FACS after staining with DCFH-DA: large effect seen in Namalwa and NB4 cells only; NSE in other cell lines. Reference Sakurai et al., 2005a Yi et al., 2004 Yietal., 2004 Chenetal., 2006 C-240 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U937 cells HEK293 cells Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (nM) 1 2 Duration of Treatment 24 hr 48 hr LOECa (HM) None 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft in H2O2 levels as detected by FACS after staining with DCFH-DA: large effect was seen only following a co- treatment with B SO for 24 hr; the ft was substantially decreased by a 4-hr treatment with either 10 mM NAC or 200 units of catalase. Cell survival was determined by the WST- 1 cell proliferation assay: inorganic arsenic treatment resulted in -22% cell survival; co- treatment with 1 mM Tironor400U/mLCAT significantly ft cell survival although more than 60% of the cells still died; co-treatment with 200 U/mL SOD markedly U cell survival. These and other data suggested that inorganic arsenic induced both superoxide anion and H2O2 through the activation of NAD(P)H oxidase. Presence of superoxide anion in cells that resulted from inorganic arsenic treatment was confirmed. Reference Chenetal., 2006 Sasaki et al., 2007 C-241 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PRCCs HEK293 cells NB4 cells HL60 cells KMS12BM cells U266 cells Arsenic Species As111 ATO for both As111 ATO for all Concentration(s) Tested (nM) 0.01,0.05,0.1, 0.5,1,5, 10,20 for both -0.01,0.05,0.1, 0.5, 1, 2, 5, 10, 50 for first three -0.05,0.1,0.6, 1.2,6 Duration of Treatment 48 hr for both 48 hr for all LOECa (HM) 0.5 0.1 -0.1 ~5 -0.2 -0.6 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival was determined by the WST- 1 cell proliferation assay: LC50sinPRCC: inorganic arsenic, -10; co-treatment of inorganic arsenic with 10 uM a- lipoic acid, -25. LC50sinHEK293: inorganic arsenic, -1; co- treatment of inorganic arsenic with 10 uM a- lipoic acid, -7. In both cell types, this antioxidant markedly attenuated inorganic arsenic's cytotoxicity, andinHEK293cellsit was shown to suppress superoxide anion generation. Cell survival was determined by the WST- 1 cell proliferation assay: LC50s: NB4, -0.2; HL60, ~8;KMS12BM, -0.3; U266, -0.3. In all 4 cell lines, co-treatment of inorganic arsenic with 10 uM a-lipoic acid resulted in a remarkably similar dose-related pattern of cell survival to that seen with inorganic arsenic alone, this being in sharp contrast to the attenuation of cytotoxicity caused by it that was seen in PRCCs and HEK293 cells. Note that the LOEC is higher than the estimated LC50 of 0.3 for U266 cells because the next lower dose of 0. 1 had no effect, and the LC50 was estimated from the dose- response curve that was presented. Reference Sasaki et al., 2007 Sasaki et al., 2007 C-242 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue JAR cells JAR cells BEAS-2B cells Undifferentiat edPC12 cells Arsenic Species As111 ATO As111 ATO As111 SA As111 ATO Concentration(s) Tested (nM) 5 5 1, 2.5, 5 for mRNA 2.5, 5 for protein 8 Duration of Treatment 2, 4, 6, 16, 24hr 6hr 8hr for both Various up to 6hr LOECa (HM) 5 5 Ifor mRNA 2.5 for protein 8 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft in HMOX-1 protein level in cytoplasm by 2 hr, with time-related response becoming huge by 16 hr. Intracellular H2O2 level detected by DCFH-DA and flow cytometry assay: ft to 2x. BigftinHMOX-1 mRNA level at 1, bigger ft of the same at 2. 5, huge ft of the same at 5. BigftinHMOX-1 protein level at 2.5, huge ft of the same at 5. Detection of ROS shown by increase of DCF- fluorescence in DCFH- DA assay: ft to ~2x control for several time points during first hr; no hint of effect at 3-6 hr; fluorescence was observed before the onset of cell death. Reference Massrieh etal.,2006 Massrieh et al., 2006 O'Hara et al., 2006 Pigaetal., 2007 C-243 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells Human- hamster hybrid AL cells Arsenic Species As111 SA MMAm As111 SA For both Concentration(s) Tested (nM) 1, 10, 100 0.05,0.5,5 7.7 Duration of Treatment 10 min 50 min 60 days LOECa (HM) 1 0.05 7.7 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Detection of ROS using CM-H2DCFDA assay: Slight ft at 1, big ft at 10, huge ft at 100. When quantified at dose of 10 over 10 min: 20 RFU by 4.5 min, 110 RFU by 10 min. Pretreatment with PEG-SOD or PEG-CAT blocked most ROS production. ft at 0.05, huge ft at 0.5, slightly weaker response at dose of 5 than at dose of 0.05. When quantified at dose of 0.5 over 10 min: 0 RFU. When quantified at dose of 0.5 over 50 min: 10 RFU by 42 min, 65 RFU by 50 min. Pretreatment with PEG-CAT blocked most ROS production, and co-treatment with PEG-SOD blocked some ROS production; less effect for both than for inorganic arsenic111, suggesting a difference in the ROS they produce. Effects related to mitochondria: fluorescence microscopy showed that arsenic treatment led to considerable variation in the distribution of mitochondria between cells and caused the fraction of them with elongated morphology to increase from 6% to 66%; -50% U in COX activity; -40% U in oxygen consumption; -40% ft in citrate synthase activity. Reference Eblin et al., 2006 Partridge et al., 2007 C-244 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human- hamster hybrid AL cells Splenic lymphocytes from SodltmlLeb knockout mice Lyophilized bovine tubulin Arsenic Species As111 SA for both As111 SA DMA111 Concentration(s) Tested (nM) 1.9,3.8,7.7 1.9,3.8,7.7 50, 100, 200 50 Duration of Treatment 60 days Iday 2hr Time course over 1 hr LOECa (HM) 3.8 for copy # 1.9 for deletions 50 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) mtDNA copy number: U to~0.84xat3.8;Uto ~0.65x at 7.7. SA induced large heteroplasmic deletions in mitochondrial DNA, and the frequencies of induction increased with dose and time of exposure. Breaks and/or alkali- labile lesions in DNA detected in the single- cell gel (comet) assay: big ft in effect in the SOD -/- mice, which were also shown to have big U in levels of SOD in spleens (and also in livers and kidneys). SOD +/- mice were intermediate in SOD levels and DNA damage. Results suggest ROS may be involved in Asm- induced DNA damage. Big U in GTP-induced polymerization of lyophilized bovine tubulin. Effects of modulators: NAC blocked the inhibition by DMA111, while AA, CAT, DMSO, Tiron, or Trolox® had NSE on it, which suggests that ROS is not involved in the inhibition. Premixingof . V inorganic arsenic , MMAV, or DMAV for 2 hr with a 5 -fold molar excess of GSH greatly decreased the polymerization of tubulin (i.e., increased the inhibition). Reference Partridge etal., 2007 Kligerman and Tennant, 2007 Kligerman and Tennant, 2007 C-245 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue W138 cells and HaCaT cells Arsenic Species As111 SA Concentration(s) Tested (|oM) 0.5 Duration of Treatment 24 hr LOECa (HM) 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ROS (peroxide) levels based on DCF assay :U in both cell lines compared to control, and less in W138 than in HaCaT. The average activities of 3 important intracellular redox agents, GSH, GR, and GST are ~3X higher in WI38 cells than in HaCaT cells. After the inorganic arsenic treatment, there was a 60-min menadione treatment followed by a 60-min recovery period. During this 120 min, ROS levels in W138 cells never reached control levels, while the control level was substantially exceeded in HaCaT cells after 60 min of the menadione treatment and later. Reference Snow et al., 2005 C-246 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NB4 cells HL-60 cells PAEC cells isolated from freshly harvested vessels CHO Kl cells Arsenic Species As111 SA for all As111 SA As111 SA Concentration(s) Tested (|oM) 2 for all assays, which tested effects of various co-treatments described in Results column 5 20, 40, 80, 160 Duration of Treatment 4hr for all Up to 20 min 4hr LOECa (HM) — — 5 40 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) This row relates only to the effects seen after co- treatments in an attempt to learn how SA causes DNA damage. They assayed DNA strand breaks (AD SB) detected using the comet assay. In the absence of a co- treatment, a significant increase would be expected with a dose of only 0.25. Conclusions always were supported by data on ODA and DPC individually. Chemicals used individually in co- treatments were: catalase, calcium chelators, and inhibitors of nitric oxide synthase, SOD, and myeloperoxidase. On the basis of the large reduction in DNA strand breaks seen following the co-treatments, they concluded that arsenite induces DNA adducts through calcium- mediated production of peroxy nitrite, hypochlorous acid, and hydroxyl radicals. ft oxygen consumption associated with ft superoxide (O2~) formation; ft extracellular accumulation of H2O2, with same time and dose dependence as superoxide formation. Pretreatment of the cells with DPI, apocynin, or SOD abolished arsenite- stimulated superoxide (O2~) formation. ft intracellular peroxide level (strong hint of same effect at dose of 20) Reference Wang et al., 2001 Barchowsk y et al., 1999b Wang et al., 1996 C-247 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue 4>X174 RF I DNA Naked double- stranded circular DNA in presence of ROS inhibitors Both HL-60 cells and HaCaT cells HL-60 cells Arsenic Species MMAm DMA111 As111 SA As111 SA Concentration(s) Tested (nM) 10, 20, 30, 40, 50 37.5, 75, 150, 300, 1000 0.1,0.5, 1, 10,20, 40 10 Duration of Treatment 24 hr for all 5 days 3 days LOECa (HM) — — 0.5 but possibly 0.1 — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) This row relates only to the effects seen after co- treatments in an attempt to learn how SA causes DNA damage. Significant (and usually complete) reduction in nicked DNA (in DNA nicking assay) was found when ROS inhibitors Trolox, melatonin, or Tiron were present individually during the arsenic treatment. Spin trap agent DMPO was also effective in preventing DNA nicking by these compounds. Thus, production of ROS by these chemicals is associated with their DNA-cutting activity. Genotoxicity is an indirect effect via the generation of ROS. By use of MTT assay, in presence of 2.5 mM DMPO: ft in cell number, with peak at 0.5 (DMPO has no effect); U in cell number to below control level at 1 for HL- 60 and at 10 for HaCaT, and DMPO significantly lessens reduction in cell number at >10 (possibly 1) for HL-60 and at >20 (possibly 10) for HaCaT. Analysis of TRF using Southern blot assay in presence of 2.5 mM DMPO: With DMPO present, telomere length was longer than it was with arsenic alone; interpreted to mean that DMPO provided some protection against arsenic-induced telomere shortening. Reference Nesnow et al., 2002 Zhang et al., 2003 Zhang et al., 2003 C-248 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HL-60 cells HaCaT cells Arsenic Species As111 SA for both Concentration(s) Tested (|oM) 0.1,0.5, 1, 10,20, 40 for both Duration of Treatment 5 days for both LOECa (HM) Ibut possibly 0.5 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ByuseofHoechst/PI staining assay, in presence of 2.5 mM DMPO: ft in apoptosis for both; however, DMPO significantly reduced the amount of apoptosis at >1 for HL-60 and at >10 forHaCaT. Reference Zhang et al., 2003 Enzyme Activity Inhibition AG06 cells were pretreated for 24 hr with unspecified low dose of As, and then extracts of the cells were tested for activity of: GSH peroxidase and ligase Cell-free system using purified human enzymes As111 SA As111 SA Asv IC50 determinations IC50 determinations Rate over 6 min Rate of reaction over 6 min — — IC50s: 2.0 (was 0. 13 mM for purified enzyme with no arsenic pretreatment) 14.5 (was 6.5 mM with no arsenic pretreatment). The same paper presented the IC50s for a similar treatment with Asv for GSH peroxidase, and it was 173 uM. The paper also presented IC50s for numerous purified enzymes with both SA and Asv, but they were almost all far above a physiologically interesting range and are thus not presented here. Most were in the range of 6.3 to 381 mM for SA and usually even higher for Asv. Inhibition of PDH: IC50s: 5.6 (oM for inorganic arsenic111, 206 mM for Asv; 7 other enzymes involved in aspects of DNA repair and/or cellular stress response hadIC50sforAsniof6.3- 381mM. Only PDH, with its lipoic acid cofactor, was inhibited by physiologically relevant, micromolar concentrations of As111. Snow et al., 1999 Hu et al., 1998 C-249 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Cell-free system using purified human enzymes Cell-free system using purified porcine heart PDH Cell-free system using hamster kidney PDH Arsenic Species As111 SA Asv As111 SA MMA111 As111 SA MMA111 Concentration(s) Tested (nM) -0.0007, 0.001, 0.007, 0.01, 0.07, 0.1 -0.01,0.07,0.1, 1, 10, 25, 75, 100, 125 25, 75, 100, 200 (all approximate) 8, 16, 30, 50, 100 -20 to -400 -20 to -400 Duration of Treatment Rate of reaction over 1 min 30 min for both 30 min for both LOECa (HM) -0.001 -25 -25 o — o — Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Inhibition of PDH. Inhibition of PDH (IC50s): 106.1. 17.6. Inhibition of PDH (IC50s): 115.7. 61.0. Reference Hu et al., 1998 Petrick et al., 2001 Petrick et al., 2001 Gene Amplification Mouse 3T6 cells AG06 cells AG06 cells Human osteosarcoma TE85 (HOS) cells As111 SA Asv As111 SA As111 SA As111 SA 0.2, 0.4, 0.8, 1.6, 3.2,6.4 1, 2, 4, 8, 16 7, 10, 17, 20 6 0.0125, 0.025, 0.05, 0.1 for both durations Not reported 3.5 hr Assay's maximal response time 6 wk 8wk 0.4 2 None 6 0.025 0.0125 Gene amplification of dhfr gene detected by MTX-selection assay: Both compounds showed positive dose-response extending to highest concentrations tested. Amplification of SV40: none observed at concentrations causing from 40% to 98% cytotoxicity. Amplification of endogenous dhfr genes (determined by MTX- selection assay): highly effective at this concentration, which caused 50% survival. "Amplification factor" was -3 even though it was 1 (i.e., no induction) for same concentration for amplification of SV40. Amplification of endogenous dhfr genes (determined by MTX- selection assay): dose- response was the same for both durations beginning with 0.025; it increased with dose to 0.05 and then plateaued. Barrett et al., 1989 Rossman and Wolosin, 1992 Rossman and Wolosin, 1992 Mure et al., 2003 C-250 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SHE cells Arsenic Species As111 SA Asv Concentration(s) Tested (nM) 6,8 50, 100, 150 Duration of Treatment 48 hr for both LOECa (HM) 6 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) From among these treatment groups, 5 neoplastic transformed cell lines were produced that were shown to be tumorigenic. Of these: 3 had c-Ha-ras (oncogene) gene amplification; 2 had c-myc (oncogene) gene amplification; a few other arsenic- treated cell lines also showed this same gene amplification. Reference Takahashi etal.,2002 Gene Mutations E. coli (several strains) V79 cells G12 cells As111 SA As111 SA As111 SA Up to 25 mM 0.5 5, 20, 100 5, 10, 15 10, 25, 50 Various 2 days Uptol.Shr 24 hr 3hr None None None None None Several assays (spot tests, treat and plate protocols, and fluctuation tests) for Trp+ revertants yielded no evidence of induction of gene mutations. Also, there was no induction of I prophage. In several assays, ouabain resistance and thioguanine resistance were used as genetic markers. No evidence was found of induction of gene mutations. Only the dose of 100 caused cy totoxicity (3 3 . 1 % the survival of the control). No statistically significant induction of mutations at the gpt locus in an assay that can detect multilocus deletions, point mutations, and small deletions (tested up to cy totoxicity of 61.9% of cells killed). Rossman etal., 1980 Rossman etal., 1980 Li and Rossman, 1989a C-251 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Salmonella typhimurium strains TA98, TA100, TA104 Syrian hamster embryo cells Human osteosarcoma TE85 (HOS) cells TM3 cells Arsenic Species As111 SA Asv MMAm MMAV DMA111 DMAV As111 SA Asv As111 SA MMA111 As111 SA for both Concentration(s) Tested (|oM) Tested up to concentrations limited by cytotoxicity or to the limit concentration for the assay -0.8, 1.6, 3, 3.5, 5 ~8, 16, 32, 64, 128 0.0125, 0.025, 0.05,0.1 0.00625, 0.0125, 0.025, 0.05 0.008, 0.77, 7.7 for both Duration of Treatment 3 days for all Not reported 8 wk for all -25 days -75 days LOECa (HM) None None None 0.0125 None 0.008 0.008 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Salmonella mutagenicity plate incorporation assay with and without exogenous metabolic activation: There was no indication of any induction of gene mutations over background levels by any of the compounds. Gene mutation assays for the Na+/K+ ATPase and HPRT loci. Mutations in the HPRT gene: positive dose- response to highest concentration for As111; no increase until almost 15 generations of continuous exposure. Detection of DNA changes by RAPD-PCR: gain or loss of loci and changes in the intensity of loci were detected at the DNA sequence level; although the nature of the "mutations" and whether they were actual gene mutations is unknown. Reference Kligerman et al., 2003 Barrett et al., 1989 Mure et al., 2003 Singh and DuMond, 2007 C-252 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested (|oM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Reference Hypermethylation of DNA Human kidney carcinoma cell lines: UOK123 UOK109 Human lung carcinoma cell line: A549 A549 cells (human adenocarcino ma) As111 SA for all As111 SA Asv DMAV 0.010, 0.020, 0.050 0.007, 0.021, 0.093 0.08, 0.4, 2.0 0.08, 0.4, 2.0 3, 10, 30, 100, 300 2, 20, 200, 2000 4wk 4wk 2wk 7 days for all <0.050 <0.093 <2.0 0.08 30 None The number of specific DNA sequences shown to undergo hypermethylation changes by methylation sensitive AP-PCR following exposure to SA: lfromlineUOK123,4 from line UOK 109, and 1 from line A549. The concentrations used to treat these lines were known to be the IC30, IC50, and IC80 concentrations for UOK cells and the IC2o, IC50, and IC8o concentrations for A549 cells. It was not reported which concentrations yielded the hypermethylation changes, but the LOECs could not be higher than the maximum concentration used for each cell line. Hypermethylation within a 341 -base-pair fragment of the promoter of p53 . For the two inorganic forms, there was a positive dose-response throughout the range of concentrations tested. Zhong and Mass, 2001 Mass and Wang, 1997 Hypomethylation of DNA TRL 1215 cells (normal rat liver) As111 SA 0.125,0.250, 0.500 19 wk 0.125 Global DNA hypomethylation, thought to be caused by continuous methyl depletion. Zhao etal., 1997 C-253 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human kidney carcinoma cell line: UOK121 Human lung carcinoma cell line: A549 Untransforme dand immortalized RWPE-1 cells (human prostate epithelial cell line) Arsenic Species As111 SA for all As111 SA Concentration(s) Tested (|oM) 0.009, 0.020, 0.074 0.08, 0.4, 2.0 5 Duration of Treatment 4wk 2wk 30 wk LOECa (HM) <0.074 <2.0 5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) The number of specific DNA sequences shown to undergo hypomethylation changes by methylation sensitive AP-PCR following exposure to SA: 1 from line UOK121 and 1 from line A549. The concentrations used to treat these lines were known to be the IC30, IC50, and IC80 concentrations for UOK121 cells and the IC20, ICso, and IC80 concentrations for A549 cells. It was not reported which concentrations yielded the hypermethylation changes, but the LOECs could not be higher than the maximum concentration used for each cell line. Global hypomethylation of DNA (up to 131% increase in unmethylated DNA compared to the control); hypomethylation still present 6 weeks after end of exposure. The cells became tumorigenic after 29 weeks of treatment and were then called the CAsE-PE cell line. Reference Zhong and Mass, 2001 Benbrahim -Tallaa et al., 2005 C-254 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SHE cells TM3 cells HaCaT cells Arsenic Species As111 SA Asv As111 SA for both As111 SA Concentration(s) Tested (|oM) 6,8 50, 100, 150 0.008, 0.77, 7.7 for both 0.2 Duration of Treatment 48 hr for both -25 days -75 days For 10 serial passages in folic-acid- depleted media LOECa (HM) — 0.008 0.008 0.2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) From among these treatment groups, 5 neoplastic transformed cell lines were produced that were shown to be tumorigenic. Testing of them using the methylation-sensitive restriction endonuclease isoschizomers Hpall and Mspl revealed hypomethylation of c- myc and c-Ha-ras in the 5'-CCGG sequence. Both of these oncogenes were often shown to exhibit gene amplification and ft mRNA expression. Detection of methylation changes in DNA by RAPD-PCR using methylation-sensitive restriction endonuclease isoschizomers Hpall and Mspl: methylation changes were detected at 18 loci, with some showing hypomethylation and others hypermethylation. Some loci were only affected by the shorter- term exposure, and vice- versa. Genomic hypomethylation as demonstrated by a 27% U in the level of 5-methyl-dCMP compared with cells cultured for the same number of passages in medium without As111. This dose was too low to have much, if any, effect on the proliferation rate. Reference Takahashi etal.,2002 Singh and DuMond, 2007 Reichard etal.,2007 C-255 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HaCaT cells Arsenic Species As111 SA Concentration(s) Tested (|oM) 0.5, 1.5,5 Duration of Treatment 72 hr LOECa (HM) Various Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) J]DNMTlmRNAat0.5, and progressively larger decreases at 2 higher doses; U DNMT3A mRNA at 1.5, and larger U at dose of 5. These cells did not show any detectable quantities of the other 2 mammalian DNA methyltransferases. Big ft HMOX-1 RNA at 1.5 with very big ft at 5. Reference Reichard et al., 2007 Immune System Response (Human myeloma-like cell lines) RPMI 8226 Karpas 707 U266 HPBMs co- exposed to M-CSF HPBMs co- exposed to GM-CSF As111 ATO As111 SA Asv MMAV DMAV As111 SA Asv MMAV DMAV 0.5, 1,2 IC50 determinations IC50 determinations 72 hr 7 days 7 days 0.5 — — Induction of cell lysis by LAK effector cells was apparent by 36 hours and maximal at 72 hours. The extent of lysis was determined by the 51Cr release assay. At these concentrations, arsenic trioxide had no effect on viability (using trypan- blue assay) or apoptosis. Viability of M-type macrophages based on AB assay was used to estimate the arsenic concentration at which maturation into M-type macrophages was inhibited by 50%: IC50 values: As111, 0.06; Asv, 200; MMAV, 750; DMAV, 300. Viability of GM-type macrophages based on AB assay was used to estimate the arsenic concentration at which maturation into GM-type macrophages was inhibited by 50%: IC50 values: As111, 0.38; Asv, 300; MMAV, 700; DMAV, 550. Deaglio et al., 2001 Sakurai et al., 2006 Sakurai et al., 2006 C-256 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference HPBMs co- exposed to GM-CSF and IL-4 Asm SA determination 7 days Viability of immature dendritic cells based on AB assay was used to estimate the arsenic concentration at which maturation into immature dendritic cells was inhibited by 50%: IC50 value: 0.70. Sakurai et al., 2006 HPBMs co- exposed to GM-CSF and IL-4 Asm SA IC50 determination 14 days Viability of multinucleated giant cells based on AB assay was used to estimate the arsenic concentration at which maturation into multinucleated giant cells was inhibited by 50%: IC50 value: 0.33. Sakurai et al., 2006 HPBMs co- exposed to GM-CSF As111 SA With regard to 4 rows immediately above this one, SA at doses of 0.05 to 0.5 induced abnormal morphological changes in the HPBMs to form small nonadhesive and CD14-positive cells called arsenite-induced cells that displayed a dendritic morphology with delicate membrane projections. This response was not produced by treatments with many other metallic compounds (e.g., chromium, mercury, and zinc) including inorganic arsenicv, MMAV and DMAV. This effect was not seen at doses exceeding 1. Sakurai et al., 2006 HPBMs co- exposed to GM-CSF As111 SA 0.5 7 days 0.5 In comparison to the cells not treated with inorganic arsenic, there was 43.3% less metabolic activity, 0.6% as much adherent ability, a 76% higher cellular GSH concentration, 256% as much NO2 production, 185% as much IL-la production in the supernatant, 412% as much IL-la production in the lysate, and 576 ng/g cellular protein of IL-12 in the lysate even though none was detected in arsenic- untreated cells. Sakurai et al., 2006 HUVECs As111 SA 0.5 3hr 0.5 Both HUVECs and PMNs were pretreated for24hrwithGLN (glutamine) at 0, 300, 600, or 1000 uM. Those HUVECs were then exposed to the same concentration of GLN with or without the Hou et al., 2005 C-257 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested (|oM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) inorganic arsenic treatment for 3 hr. The pretreated PMNs were added to wells and allowed to migrate across the pretreated HUVECs for 2 hr, after which surface expressions on HUVECs of 1C AM- 1 and VCAM- 1 were measured, with the following results: ICAM-1: ft in inorganic arsenic only group and huge ft at all 3 dose levels of GLN; VCAM- 1 : NSE in inorganic arsenic only group and ft at all 3 dose levels of GLN, with largest ft at 300 uM. Clearly HUVECs were activated by inorganic arsenic. Also at this time, PMN expressions of CD 1 Ib and IL-8 receptor were measured, with the folio wing results: CD 11- b: ft in inorganic arsenic only group and bigger ft at all 3 dose levels of GLN; IL-8 receptor: ft in inorganic arsenic only group and at all 3 dose levels of GLN. Clearly PMNs were activated by the inorganic arsenic treatment of the HUVECs. Effects on PMN migration: In absence of GLN pretreatment, inorganic arsenic caused slight U from 36% to 30% migrated. In the inorganic arsenic + 300 uM GLN group: ft from -40% (for GLN alone) to -50% migrated (for inorganic arsenic + GLN), which was the most migration observed. Reference C-258 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PBMCs co- treated with GM-CSF PBMCs co- treated withM-CSF Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (|oM) 0.125,0.25,0.5, 1, 2 1 Duration of Treatment 6 days 6 days LOECa (HM) 0.125 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (i.e., experimental - control) detected using A5/SG assays: approximate induced frequencies at the 5 doses: 6%, 20%, 29%, 48%, and 62%, respectively, with all being statistically significant except first one. Induced frequency of necrotic cells was -20% at the highest dose, and there were smaller numbers of necrotic cells induced at the lower doses. After dose of 1 for 3 days: ft caspase-3 activity, ft caspase-8 activity, big ft in active caspase-3 subunitpl?. ATO was shown to reduce DNA binding of the transcriptionally active p65 NF-KB subunit to the KB consensus sites in GM- CSF treated PBMCs, which was thought to be important in development of apoptosis. Other experiments showed that ATO inhibited macrophagic differentiation of PBMCs. Induced apoptosis (i.e., experimental - control) detected using A5/SG assays: -44%. The induced frequency of necrotic cells was -23%. Reference Lemarie et al., 2006a Lemarie et al., 2006a C-259 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue U937 cells co-treated with PMA U937 cells co-treated with PMA PBMCs co- treated with GM-CSF Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Arsenic Species As111 ATO As111 ATO As111 ATO As111 ATO As111 ATO Concentration(s) Tested (|oM) 1,4 4 1 1,4 1,4 1 Duration of Treatment 4 days 4 days 3 days 3 days 6 days 6 days LOECa (HM) 4 4 1 4 4 None for u Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Induced apoptosis (i.e., experimental - control) detected using A5/SG assays: approximate induced frequencies at the 2 doses: 3% and 35%, respectively, with the higher one being statistically significant. Induced frequency of necrotic cells was ~9% at the highest dose. Other experiments showed (1) that ATO induced apoptosis through inhibition of NF-KB signals and (2) that ATO inhibited macrophagic differentiation of U937 cells. U FLIPL protein level, U XI AP protein level. U FLIPL protein level and U FLIPL mRNA level; U XIAP protein level and U XIAP mRNA level. Induced apoptosis (i.e., experimental - control) detected using A5/SG assays: No induced apoptosis at dose of 1 at either time. At dose of 4: -22% and -50% after 3 and 6 days, respectively; thus these cells are resistant to induction of apoptosis by ATO at low doses. NSE regarding FLIPL protein level; big ft XIAP protein level. Reference Lemarie et al., 2006a Lemarie et al., 2006a Lemarie et al., 2006a Lemarie et al., 2006a Lemarie et al., 2006a C-260 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Arsenic Species As111 ATO As111 ATO As111 ATO Concentration(s) Tested (|oM) 4 0.25,0.5, 1 0.5, 1,2,4 Duration of Treatment 3 6 days 6 days LOECa (HM) 4 0.25 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Big U FLIPL protein level; big U XIAP protein level. Major alterations in the morphology, adhesion, and actin organization with the impression that inorganic arsenic "de- differentiated" macrophages back into monocytic cells. The effect was time- dependent with rounded and contracted morphology first observed at dose of 1 after only 8 hr. By 6 days at dose of 1 only 3 1% as many cells were adherent as in control. Inorganic arsenic induced a reorganization of theF-actin cytoskeleton. The series of experiments suggested that the effects occurred because of the activation ofaRhoA/ROCK pathway. Induced apoptosis (i.e., experimental - control) detected using A5/SG assays: approximate induced frequencies at the 4 doses: 0%, 0%, 20%, and 50%, respectively. Induced frequency of necrotic cells was ~4% at the highest dose. 18 days of treatment at dose of 1 caused no cytotoxicity. Reference Lemarie et al., 2006a Lemarie et al., 2006b Lemarie et al., 2006b C-261 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (|oM) 1 1 Duration of Treatment 6 days 6 days LOECa (HM) 1 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Changes in surface markers: CD 14: ft 5. Ix; CD71: U to 45% of control; CD29: U to 49% of control; CD lib: Uto 42% of control. Changes in major functions: marked U in endocytosis and phagocytosis. Changes in surface markers and morphology were shown to be reversible when inorganic arsenic was removed and cells were cultured with GM-CSF for 6 days. Ability to secrete inflammatory cytokines in response to co- treatment of inorganic arsenic (dose of 1) and 200 ng/mL LPS for 8 or 24 hr (control = macrophages treated with LPS only): TNF-a secretion: ft ~3. Ox and ~3. Ox at 8 and 24 hr, respectively. IL-8 secretion: ft ~3x and ~4.5x at 8 and 24 hr, respectively. Much more extreme potentiation was demonstrated for both cytokines at the mRNA level at 8 hr. The text implies that the potentiation of both secretion and mRNA production does not occur without the 6-day inorganic arsenic treatment. Reference Lemarie et al., 2006b Lemarie et al., 2006b C-262 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days before inorganic arsenic treatment Arsenic Species As111 ATO As111 ATO Concentration(s) Tested (|oM) 1 1 Duration of Treatment 6 days 8hr LOECa (HM) 1 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) The inorganic arsenic- treated macrophages differentiated into dendritic-like cells when exposed to GM-CSF and IL-4 in the absence of inorganic arsenic for 6 days. This conclusion was based on the ~9x increase in the expression of the typical dendritic marker CD la. The increase was similar to that seen in PBMCs treated with GM-CSF and IL-4 for 6 days, and in both cases the dendritic-like cells were nonadherent. In contrast, fully differentiated macrophages (i.e., PBMCs treated with GM-CSF for 6 days without inorganic arsenic) did not show this response. ft GTP-binding fraction ofRhoA; ft phospho-Moesin protein level. (Phosphorylated-Moesin is a major cytoskeleton adaptor protein involved in RhoA regulation. RhoA is a small GTPase protein known to regulate the actin cytoskeleton in the formation of stress fibers.) Reference Lemarie et al., 2006b Lemarie et al., 2006b C-263 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Differentiated macrophages developed from PBMCs treated with GM-CSF for 6 days and then pretreated with ROCK inhibitor Y- 27632 for 2 hr before inorganic arsenic treatment HepG2 cells Arsenic Species As111 ATO As111 SA Concentration(s) Tested (|oM) 1 0.04, 0.4, 4, 40 Duration of Treatment 72 hr 48-hr pretreatment LOECa (HM) 1 4 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Pretreatment with the ROCK inhibitor prevented both the F- actin reorganization and cellular rounding of macrophages treated with inorganic arsenic. It also considerably blunted damage to the phagocytosis function caused by the inorganic arsenic treatment. After the inorganic arsenic pretreatment, there was a 30-min treatment with IL-6, which induced STATS activity unless inhibited by the pretreatment. Level of STATS activity: huge U at 4; no activity at 40. Reference Lemarie et al., 2006b Cheng et al., 2004 C-264 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue HepG2 cells Arsenic Species As111 SA Concentration(s) Tested (|oM) 0.04, 0.4, 4, 40, 400 Duration of Treatment 48-hr pretreatment LOECa (HM) 40 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) After the inorganic arsenic pretreatment, there was a 30-min treatment with IL-6, which induced both STATS tyrosine phosphorylation and STATS serine phosphorylation. Only the tyrosine phosphorylation was inhibited by the inorganic arsenic pretreatment, with slight U at 40 and huge U at 400. Inorganic arsenic is thought to inactivate the JAK-STAT signaling pathway by means of inhibition of STATS tyrosine phosphorylation. Other inflammatory stimulants, stress agents, and cadmium failed to induce similar effects on the tyrosine phosphorylation of STATS. Reference Cheng et al., 2004 C-265 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference HepG2 cells As111 SA 4, 40, 400 30-min pretreatment and 1-hr co-treatment with IL-6 Huge U in Cis mRNA and in SOCS mRNAs for 5 of 6 forms tested (U for the other form); the U at higher doses was usually the same or more severe; UinSTATmRNAsfor4 of 6 forms tested, the U at higher doses was usually the same or more severe. The decreases for STAT mRNAs were very slight compared to those for SOCS. The inhibition of induction of SOCS mRNA confirmed that JAK-STAT signaling had been turned off. Other experiments showed that the effect of inorganic arsenic on JAK-STAT inactivation is independent of ligand- receptor action and is a result of the direct action of arsenic on the JAK1 protein. Cheng et al., 2004 HepG2 cells As111 SA 0.04, 0.4, 4, 400 40, inorganic arsenic (in co-treatment with IL-6 for unknown duration) activated all 3 subfamilies of MAP kinases (i.e., there was phosphorylation of ERK 1/2, p38, and JNKs) with LOECs of 40, 0.04, and 0.04 respectively. Such activation was independent of IL-6 stimulation at least at higher doses. Experiments with specific inhibitors of the 3 MAP kinases showed that inorganic arsenic selectively targeted JAK tyrosine kinase and that the inhibition of JAK-STAT activity by inorganic arsenic did not require the participation of any MAP kinases. Cheng et al., 2004 PBMCs treated with 1000 U/mL of M-CSFatthe same time as with inorganic arsenic As111 SA 0.005, 0.010, 0.050,0.10,0.50 7 days 0.050 Cell survival demonstrated by trypan blue exclusion assay: LC50: 0.22; about 25% survival at dose of 0.5. The cells differentiated into adhesive M-type macrophages that were elongated and had a spindle-like morphology. Sakurai et al., 2005b C-266 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PBMCs treated with 5000 U/mL of GM-CSF at the same time as with inorganic arsenic PBMCs treated with 5000 U/mL of GM-CSF at the same time as with inorganic arsenic Arsenic Species As111 SA As111 SA Concentration(s) Tested (|oM) 0.005, 0.010, 0.050,0.10,0.50 0.50 Duration of Treatment 7 days 7 days LOECa (HM) 0.10 0.50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cell survival demonstrated by trypan blue exclusion assay: -85% survival at 2 highest doses; up to dose of 0.050, all cells differentiated into GM- Mp, which had a round saucer-like appearance; at dose of 0.10, -80% of living cells were GM-Mp and the rest were abnormal "arsenite- induced cells"; at dose of 0.50, -10% of living cells were GM-Mp and the rest were "arsenite- induced cells." In comparison to controls (i.e., PBMCs treated with 5000 U/mL of GM-CSF and no inorganic arsenic), the resulting morphologically, phenotypically, and functionally altered "arsenite-induced cells" had:ftHLA-DRto5.0x; U CD lib to 0.7 Ix; ftCD14tol.4x;UCD54 to 42% of control; big U in phagocytic ability; ft in effectiveness in inducing allogeneic or autologous T-cell responses; and huge ft in response to bacterial LPS by inflammatory cytokine release. Reference Sakurai et al., 2005b Sakurai et al., 2005b C-267 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PBMCs treated with 5000 U/mL of GM-CSF at the same time as with inorganic arsenic PBMCs treated with 1000 U/mL of M-CSF or 5000 U/mL of GM-CSF at the same time as with inorganic arsenic Arsenic Species As111 SA Asv Concentration(s) Tested (|oM) 0.50 LCso determinations Duration of Treatment 7 days 7 days LOECa (HM) 0.50 >1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) The resulting high numbers of "arsenite- induced cells" were markedly reduced by co- treatment with DMPO, DMSO, orBHT, all of which are membrane- permeable radical trapping reagents. Further indication that ROS might impact development of the "arsenite-induced cells" was that by using DCFH-DA it was shown that ROS levels were much higher throughout the 7 days of culturing and >2x higher on days 1-4 of that period. Cell survival demonstrated by trypan blue exclusion assay: LC50: 300 for simple cytotoxicity for both treatments and with no toxic effect on differentiation into macrophages up to dose of 1. Reference Sakurai et al., 2005b Sakurai et al., 2005b C-268 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PBMCs stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment PBMCs stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment Arsenic Species As111 SA As111 SA Concentration(s) Tested (|oM) 1,2,3,4,5 1,2,3,4,5 Duration of Treatment 120 hr 120 hr LOECa (HM) 1 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Number of rounds of cell division estimated using CFSE dilution assay with FACS (control had 6 rounds): 5, 4, 3, 2, and 1 rounds of cell division were observed after doses of 1, 2, 3, 4, and 5, respectively; there was a marked dose-related U in both proliferation and the percentage of divided cells. Additional staining with 7-AAD revealed that, at even the higher doses, most cells were viable but unable to divide. The reduced proliferation resulted from an ft in the fraction of non-dividing cells and a delay in the cell cycle, with only a comparative small ft in the number of dead cells. At the highest dose, 63% of the cells were non-dividing, and 2/3 of them were alive. Expression of CD4 and CDS molecules was determined using CFSE staining during the inorganic arsenic treatment and then, after the 96 hr incubation, by indirect immunofluorescence using OKT4 or OKT8 hybridoma supernatants and goat anti-mouse IgG-PE, followed by 7- AAD staining and FACS analysis. The dose of 1 slightly modified the expression of both CD4 and CDS. At doses >3: a marked U in number of cells expressing CD4; at doses >4: a marked U in number of cells expressing CDS. Reference Tenorio and Saavedra, 2005 Tenorio and Saavedra, 2005 C-269 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue PBMCs stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment Human CD4+ cells stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment Human CD8+ cells stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment Arsenic Species As111 SA As111 SA for both Concentration(s) Tested (|oM) 1,2,3,4,5 1,2,3,5 for both Duration of Treatment 120 hr 120 hr for both LOECa (HM) 1 1 for both: a slight but signifi- cant effect Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Evaluation of blast transformation of both CD4+ and CD8+ T cells suggested that they have different sensitivities to inorganic arsenic. There was an accumulation of resting CD8+ cells with a positive dose-response; that accumulation was not seen for CD4+ cells. Number of rounds of cell division estimated using CFSE dilution assay with FACS (control CD4+and CD8+ cells had 6 and 5 rounds, respectively): At a dose of 1: only 5 rounds in CD4+ but no U in rounds in CD8+; however, CD8+ cells had U in cell number in the last 3 rounds, arsenic doses increased in both cell types: decreasing numbers of cell divisions and of numbers of cells in each round. Effects were generally more extreme in CD8+ cells. Reference Tenorio and Saavedra, 2005 Tenorio and Saavedra, 2005 C-270 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Human CD4+ cells stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment Human CD8+ cells stimulated with PHA for 96 hr starting 24 hr after the beginning of the inorganic arsenic treatment PBMCs stimulated with PHA during the inorganic arsenic treatment, CD4+ and CD8+T cells were analyzed separately Arsenic Species As111 SA for both As111 SA Concentration(s) Tested (|oM) 1,2,3,4,5 for both 1,2,3,4,5 Duration of Treatment 120 hr for both 24 hr 48 hr 72 hr LOECa (HM) 1 for both: a slight but signifi- cant effect 2 1 1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) CFSE dilution assay with 7-AAD staining and FACS: In both cell types there were apparent differences from the control at the dose of 1, and there was a progressive U in viable proliferating cells with increasing dose, arsenic doses increased from 0 to 3, there was a much faster ft in the fraction of resting cells that was alive among CD8+ cells than among CD4+ cells, and that fraction remained higher. LOECs were based on FACS patterns that seemed substantially different as to kinetics of expression of CD25 and CD69 in CD4+ T cells. Inorganic arsenic delayed both the expression of CD25 and the down-regulation of CD69, suggesting that inorganic arsenic delays the activation kinetics of CD4+T cells. CD4+T cells exposed to the highest dose for 72 hr showed a very similar pattern to that seen in non-inorganic arsenic- exposed cells stimulated for only 24 hr. A similar analysis of CD8+T cells showed similar results; however, with them there were somewhat more CD25"CD69~ cells (i.e., cells unable to activate) as dose increased. Reference Tenorio and Saavedra, 2005 Tenorio and Saavedra, 2005 C-271 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SV-HUC-1 cells SV-HUC-1 cells Arsenic Species As111 SA As111 SA Concentration(s) Tested (|oM) 2, 4, 8, 10, 40 2, 4, 8, 10, 40 Duration of Treatment 48 hr 48 hr LOECa (HM) 2 for all effects 2 for all effects Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Labeling indices (LIs) for immunochemistry of cells: Bcl-6: ft at 2, increases with dose as follows: LIs ofO, 1.04,3.05,6.01, 8.24, and 23. 94 for control and doses listed, respectively. JAK2: U at 2, decreases with dose as follows: LIs of 100, 58.1,48.9, 13.0, 5.1, and 0.8 for control and doses listed, respectively. p-STAT3 (Tyrosine 705): ft at 2 with peak at dose of 4 before decreasing, as follows: LIs of 100, 111.7, 151.0, 125.2, 119.0, and 50.8 for control and doses listed, respectively. All experimental LIs above differed from control, p<0.05. Effects on protein levels determined by Western blotting: Bcl-6: ft at 2 and increases with dose. JAK2:Uat2and decreases with dose. P-STAT3 (Tyrosine 705): ft at 2, peak at 4, less than control at 40. Results at different doses were highly consistent with results obtained using immunochemistry, as shown in row above. Reference Huang et al., 2007b Huang et al., 2007b C-272 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SV-HUC-1 cells BAEC cells Thymocytes (freshly isolated) Splenocytes (freshly isolated) Arsenic Species As111 SA As111 SA Asv Asv Concentration(s) Tested (|oM) 2, 4, 8, 10, 40 10 67, 150,315,680, 1000, 2000 67, 150,315,680, 1000, 2000 Duration of Treatment 48hr 48hr 24 hr 24 hr LOECa (HM) 2 for all effects 10 315 150 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Microscopy and immunochemistry showed Bcl-6 and p- STAT3 (Tyrosine 705) to be localized in the nucleus and JAK-2 to be localized in the cytoplasm. Morphological changes began to appear at dose of 2. At dose of 4, cells became round and exhibited nuclear condensation. At highest two doses, there was cellular shrinkage and cytoplasmic vacuolization. ftinLTE4to~5x. Co- treatment with 50 uM Mn11, which caused ~9x ft by itself, caused an approximately additive ft to~12x. Addition of L- NAME to the inorganic arsenic/Mn co-treatment boosted LTE4 level to ~24x. Addition of ETU to inorganic arsenic/Mn co-treatment boosted LTE4 level to slightly above that of inorganic arsenic/Mn combination. Addition of AA-861 to inorganic arsenic/Mn co- treatment reduced LTE4 level by -80%. Cell survival determined using XTT assay: LC50: 442. Cell survival determined using XTT assay: LC50: 427. Reference Huang et al., 2007b Bunderson etal.,2006 Stepnik et al., 2005 Stepnik et al., 2005 C-273 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Thymocytes (freshly isolated) Splenocytes (freshly isolated) Arsenic Species Asv Asv Concentration(s) Tested (jiM) 67,315,680 67,315,680 Duration of Treatment 24 hr 24 hr LOECa (HM) 315 315 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Point estimates of induced apoptosis (experimental minus control) determined by TUNEL staining: 5% at 67(NSE); 16% at 3 15 (NSE); and 24% at 680. 27% of control cells were apoptotic. Agarose gel electrophoresis of DNA showed high (and indistinguishable) levels of apoptosis in control group and at the 3 experimental dose levels. Point estimates of induced apoptosis (experimental minus control) determined by TUNEL staining: 1% at 67 (NSE); 16% at 3 15; and 33% at 680. 29% of control cells were apoptotic. Agarose gel electrophoresis of DNA showed high (and indistinguishable) levels of apoptosis in control group and at the 3 experimental dose levels. Reference Stepnik et al., 2005 Stepnik et al., 2005 Inhibition of Differentiation C3H 10T1/2 cell line (mouse cells with fibroblast morphology during routine culture but capable of differentiation into adipocytes) As111 SA 6 8wk 6 Complete inhibition of differentiation into adipocytes induced by dexamethasone/insulin (dexl) treatment. The effect is the same if arsenic is removed just before the dexl treatment. Trouba et al., 2000 C-274 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue C3H 10T1/2 cell line (mouse cells with fibroblast morphology during routine culture but capable of differentiation into adipocytes) SIK cells treated in surface cultures beginning when they reached confluence, which is when their rate of division decreases as differentiation increases hEp cells treated in surface cultures beginning when they reached confluence, which is when their rate of division decreases as differentiation increases Arsenic Species As111 SA As111 SA As111 SA Concentration(s) Tested (|oM) 0.1, 1,3,6, 10 2 2 Duration of Treatment 48 hr Various Various LOECa (HM) 3 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Dose-related inhibition of differentiation into adiopocytes induced by dexamethasone/insulin (dexl) treatment. These concentrations do not cause cytotoxicity. CFE based on assay using Rhodanile blue staining: on 1 day post- confluence both experimental and control groups had CFEs of -11%, by 4 days their CFEs were -9.2% and -5.2%, and by 14 days they were -4.7% and -0.6%, respectively. Thus, inorganic arsenic decreased the exit of cells from the germinative compartment under conditions that promote differentiation. CFE based on assay using Rhodanile blue staining: at 4 days post- confluence experimental and control groups had CFEs of -1.1% and 0.25%, by 11 days their CFEs were -1.0% and -0.05%, and by 14 days they were -1.0% and -0%, respectively. Thus, inorganic arsenic decreased the exit of cells from the germinative compartment under conditions that promote differentiation. Reference Trouba et al., 2000 Patterson etal.,2005 Patterson etal.,2005 C-275 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SIK cells, with inorganic arsenic treatment beginning 1 day before suspension and continuing while cells were in suspension, which drives such cells prematurely into the differentiation pathway hEp cells, with inorganic arsenic treatment beginning 1 day before suspension and continuing while cells were in suspension, which drives such cells prematurely into the differentiation pathway SIK cells, with inorganic arsenic treatment beginning when they were put into suspension, which drives such cells prematurely into the differentiation pathway Arsenic Species As111 SA As111 SA As111 SA Concentration(s) Tested (|oM) 2 2 0.1,0.2,0.5,2 Duration of Treatment 1,2, 3, 4 or 5 days, when including the 1 day of treatment before being put into suspension 1 or 2 days, when including the 1 day of treatment before being put into suspension 4 days LOECa (HM) 2 2 0.1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) CFE based on assay using Rhodanile blue staining comparison with control (C): At 1 day : C, -11.0%, inorganic arsenic, -10.8%. At 2 days: C, -0.5%; inorganic arsenic, -2.3%. At 3 days: C, -0.1%; inorganic arsenic, -2.0%. At 4 days: C, -0%; inorganic arsenic, -1.3%. At 5 days: C, -0%; inorganic arsenic, -0.8%. CFE based on assay using Rhodanile blue staining comparison with control (C): At 1 day: C, -1.15%, inorganic arsenic, -1.37%. At 2 days: C, -0.08%; inorganic arsenic, -0.68%. CFE based on assay using Rhodanile blue staining (control CFE = -0.03%): Experimental CFEs: 0.1, -0.10%; 0.2, -0.23%; 0.5, - 0.40%; 2, - 0.80%. Reference Patterson etal.,2005 Patterson etal.,2005 Patterson etal.,2005 C-276 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue RACs from either the SIK or hEp cell line (did not specify which) treated in surface culture SACs from either the SIK or hEp cell line (did not specify which) treated in surface culture SIK cells, with inorganic arsenic treatment beginning when cultures reached 90% confluence Arsenic Species As111 SA As111 SA As111 SA Concentration(s) Tested (nM) 2 2 2 Duration of Treatment 4,7, 11, or 14 days 4,7, 11, or 14 days 3 days LOECa (HM) 2 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Mean no. of colonies present, comparison with control (C): At 4 days: C, -3.0; inorganic arsenic, -18.2. At 7 days: C, -0.8; inorganic arsenic, -10.5. At 11 days: C, -0.7; inorganic arsenic, -7.8. At 14 days: C, -0.4; inorganic arsenic, -5.0. Mean no. of colonies present, comparison with control (C): At 4 days: C, -1.3; inorganic arsenic, -5.5. At 7 days: C, -0.5; inorganic arsenic, -1.8. At 11 days: C, -0.6; inorganic arsenic, -1.5. At 14 days: C, -0.3; inorganic arsenic, -1.3. Relative CFEs based on Rhodanile blue assay, with values relative to the CFE of untreated cells in medium normally contained insulin (was set at 1): ft to -2.6; if inorganic arsenic + EGF in medium: ft to -4.1. If EGF alone: -1.9; if no insulin in medium (± EGF): -3.5. Thus, inorganic arsenic delays differentiation and preserves the proliferative potential of keratinocytes. Reference Patterson etal.,2005 Patterson etal.,2005 Patterson and Rice, 2007 C-277 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue hEp cells, with inorganic arsenic treatment beginning when cultures reached 90% confluence SIK cells, with inorganic arsenic treatment beginning when cultures reached 90% confluence Arsenic Species As111 SA As111 SA Concentration(s) Tested (|oM) 2 2 Duration of Treatment 3 days 9 days LOECa (HM) 2 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Relative CFEs based on Rhodanile blue assay, with values relative to the CFE of untreated cells in medium normally contained insulin (was set at 1): ft to -2.6; if inorganic arsenic + EOF in medium: ft to -4.1; if EOF alone: -2.1; if neither EOF nor insulin: -2.1; if EOF but no insulin: -5.3. Thus, inorganic arsenic delays differentiation and preserves the proliferative potential of keratinocytes. Relative CFEs based on Rhodanile blue assay, with values relative to the CFE of untreated cells in medium normally contained insulin (was set at 1): ft to -3.8; if inorganic arsenic + EGF in medium: ft to -5.1; if EGF alone: -1.3; if no insulin in medium: -5.5. In the absence of insulin, EGF substantially augmented CFE while inorganic arsenic had no effect. Thus, inorganic arsenic delays differentiation and preserves the proliferative potential of keratinocytes. Reference Patterson and Rice, 2007 Patterson and Rice, 2007 C-278 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SIK and hEPcells, with inorganic arsenic treatment beginning when cultures reached 90% confluence Arsenic Species As111 SA Concentration(s) Tested (|oM) 2 Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) With regard to effects in the previous 3 rows, EGFR inhibitors AG1478 and PD 158780 kept inorganic arsenic from increasing the CFE regardless of the addition of EOF. They did not affect CFE in the presence of insulin but largely prevented the ft in CFE resulting from insulin removal. Inorganic arsenic treatment caused big ft in active Ras protein, a downstream effector of EGFR; co-treatment with AG1478 blocked that effect. Other experiments showed that the inorganic arsenic treatment blocked the U in active EGFR protein and the U in active (3-catenin that normally occur after confluence as cells exit the proliferative pool and differentiate. Also, expression of a dominant negative (3-catenin suppressed the ft in colony -forming ability and yield of putative stem cells induced by inorganic arsenic and EOF. Reference Patterson and Rice, 2007 Interference With Hormone Function EDR3 cells transfected as described in paper COS-7 cells transfected as described in paper As111 SA As111 SA 0.045,0.09,0.18, 0.27, 0.36, 0.45, 0.54, 0.675, 0.9, 1.8,2.7 0.1,0.5, 1.0,2.0, 3.0 -18 hr -18 hr -0.09 None ft of hormone-activated OCR-mediated gene transcription of reporter genes containing TAT glucocorticoid response elements in the presence of activated OCR; peak response was at -0.5; however, inorganic arsenic was inhibitory at doses of 1.8 and 2.7. Other experiments showed a similar effect on the endogenous TAT gene and also that the central DNA binding domain of the OCR is the minimal region required for the arsenic effect. inorganic arsenic had no effect on transcriptional repression by OCR. That is, arsenic had no effect on the ability of hormone-activated OCR to inhibit API expression or NF-KB-mediated gene expression. Bodwell et al., 2004 Bodwell et al., 2004 C-279 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue EDR3 cells transfected as described in paper EDR3 cells transfected as described in paper EDR3 cells transfected as described in paper NHEK cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Asv, MMAV, DMAV Concentration(s) Tested (|oM) 0.045,0.09,0.18, 0.27, 0.36, 0.45, 0.54, 0.675, 0.9, 1.8,2.7 0.045,0.09,0.18, 0.27, 0.36, 0.45, 0.54, 0.675, 0.9, 1.8,2.7 Duration of Treatment ~18hr ~18hr LOECa (HM) -0.09 -0.09 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft of hormone-activated OCR-mediated gene transcription of reporter genes containing TAT response elements in the presence of activated PR; peak response was at -0.5; however, inorganic arsenic was inhibitory at doses of 0.9, 1.8 and 2.7. ft of hormone-activated OCR-mediated gene transcription of reporter genes containing TAT response elements in the presence of activated MCR; peak response was at -0.5; however, inorganic arsenic was inhibitory at doses of 1.8 and 2.7. For all 3 steroid receptors tested (OCR, PR and MCR — see 3 rows immediately above this one), the degree of stimulation at lower inorganic arsenic concentrations or repression at higher inorganic arsenic concentrations was highly dependent on, and inversely related to, the amount of activated steroid receptor within cells. The relative increases in transcription noted above, which were up to ~2x or more above control levels, were at the lowest levels of activated steroid receptor within cells that were tested. Other studies showed that iA (1) had no significant effect on cellular steroid levels or on binding of steroid to the receptor, (2) did not activate or act as an agonist for OCR, (3) did not act as a competitive antagonist, (4) did not appear to affect the ability of the hormone to bind to or activate OCR, (5) did not appear to affect hormone-stimulated nuclear translocation of OCR, and (6) did not significantly alter the level of OCR for either cells expressing endogenous OCR or those expressing stably integrated OCR. Dimerization is not critical for the response to inorganic arsenic. In summary, it is clear that inorganic arsenic can simultaneously disrupt multiple hormone receptor systems. 0.001,0.005,0.01, 0.05,0.1,0.5, 1,5 for all 24 hr 24 hr 0.001 0.001-0.5 For cytokines GM-CSF, TNF-a, andIL-6: substantial ft at 0.001- 0.01, but no change or U (sometimes markedly) at 0.05-5. No change or U (sometimes markedly). Reference Bodwell et al., 2006 Bodwell et al., 2006 Bodwell et al., 2006 Vegaetal., 2001 C-280 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Hormone- responsive H4IIE (rat hepatoma cell line) Arsenic Species As111 SA Concentration(s) Tested (nM) 0.3, 1.0,2.0,3.3 Duration of Treatment 2hr LOECa (HM) 0.3 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) U in hormone-inducible expression of GRE2-Luc with a 2-hr As111 pretreatment before an 18-hr Dex treatment. The pretreatment did not block the normal Dex- induced nuclear translocation of glucocorticoid receptor. As111 selectively inhibited glucocorticoid-receptor- mediated transcription. Reference Kaltreider etal.,2001 Malignant Transformation or Morphological Transformation HaCaT cells TRL 1215 cells (normal rat liver) JB6 C141 cells simultaneousl y treated with 10 ng/mL EOF JB6 C141 cells Primary Syrian hamster embryo cells (HEC) As111 SA As111 SA As111 SA Asv As111 SA Asv 0.5, 1.0 0.125,0.250, 0.500 25, 50, 200 12.5, 50, 200 25, 50, 100 13,27 20 passages 18 wk 14 days for both 4 wk followed by 4 wk at lower concentration 7-8 days 0.5 0.250 50 12.5 25 13 Cells became tumorigenic; tumors were produced by 2 months after injection of cells into Balb/c nude mice; cells from tumors were much more malignant. Transformed cells produced aggressive tumors capable of metastasis after inoculation into nude mice. Inhibition of EGF- induced cell transformation: The effect was much stronger for Asv (sodium arsenate) with complete blockage of transformation at 50 and 200. Transformed cells, as shown by growth of colonies in soft agar; transformation did not occur at the 2 higher doses; SA-induced transformation was blocked by introduction of dominant negative Erk2. Morphologically transformed cells. Chien et al., 2004 Zhao etal., 1997 Huang et al., 1999b Huang et al., 1999a DiPaolo and Casto, 1979 C-281 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Syrian hamster embryo cells Human osteosarcoma TE85 (HOS) cells Untransforme dand immortalized RWPE-1 cells (human prostate epithelial cell line) SHE cells SHE cells Arsenic Species As111 SA Asv As111 SA MMAm As111 SA As111 SA DMAmI As111 SA Asv Concentration(s) Tested (|oM) -0.8, 1.6, 3, 3.5, 5 ~8, 16, 32, 64, 128 0.0125, 0.025, 0.05,0.1 0.00625, 0.0125, 0.025, 0.05 5 1,3, 10 0.1,0.2,0.4, 1.0 3, 6, 8, 10 50, 100, 150 Duration of Treatment 7 days for all 6 and 8 wk for both 29 wk 48 hr for both 48 hr for both LOECa (HM) 0.8 8 0.025 at 8 wk; 0.05 at 6 wk None 5 1 0.1 6 50 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Morphological transformation: For both chemicals: a positive dose-response throughout the dose range tested. Transformation to anchorage-independence in soft agar As111: positive dose-response to highest concentration; 8 weeks was -40 generations; MMAm was more toxic than inorganic arsenic111. Aggressive tumors were produced after cells showing ft secretion of MMP-9 were inoculated into nude mice. Morphological transformation (% of surviving colonies transformed at each concentration): 1,0. 11%; 3, 0.23%; 10, 0.48%. 0.1, 0.28%; 0.2, 0.51%; 0.4, 3.41%, 1.0, 3.35%. Neoplastic transformation based on anchorage-independent growth and/or tumorigenicity in newborn hamsters. All 5 anchorage-independent cultures tested were tumorigenic. Reference Barrett et al., 1989 Mure et al., 2003 Achanzar etal.,2002 Ochi et al., 2004 Takahashi etal.,2002 C-282 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells NIH 3T3 cells Arsenic Species MMAm As111 SA Concentration(s) Tested (|oM) 0.05 2, 5, 10, 20, 50, 100, 200 Duration of Treatment 52 weeks 7 days LOECa (HM) 0.05 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Anchorage-independent growth as detected by colony formation in soft agar; cells from those colonies showed enhanced tumorigenicity in SCID mouse xenografts. After only 24 weeks there was also much anchorage- independent growth, but those cells did not show the enhanced tumorigenicity. Anchorage-independent growth in soft agar assayed using AlamarBlue dye assay and microscopic examination: ft to ~1.4x control at 2 and 5 ; NSE at 10, marked dose- related U at higher doses. A daily 2-hr 42°C heat shock (which would induce HSPs) boosted induction of anchorage- independent growth for up to 3 repetitions, but 5 heat-shock repetitions markedly reduced such growth. When the same experiment was repeated in R-3T3 (transformed) cells, there was NSE by inorganic arsenic or heat shock on the already high level of anchorage- independent growth; inorganic arsenic caused U at dose of 20, and at higher doses the U became marked, as it did also at all doses following 5 daily repetitions of the heat- shock treatment. Reference Bredfeldt etal.,2006 Khalil et al., 2006 C-283 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue BALB/c 3T3 A3 1-1-1 cells (derived from mice) Arsenic Species As111 SA AsvDA MMAV DMAV Concentration(s) Tested (nM) 2, 5, 10, 15, 20 10, 15, 20, 25, 30 1,2,5, 10 mM 0.5, 1,2, 5 mM Duration of Treatment 72 hr for all LOECa (HM) 5 15 10 mM ImM Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Caused initiation in a two-stage transformation assay; based on a significant increase in the number of transformed cells after an initiating treatment with an arsenic compound for 72 hr followed by post- treatment with 0.1 ug/mL TPA for 18 days. Except for As111 SA, responses were stronger at higher doses; with it, the peak response was at 10, with a steep decline by 20. Slight but significant transformation occurred even without TPA at the 2 highest doses of As111 SA and for 2 mM DMAV. The ranges of positive effects in foci/dish in the two-stage transformation assay (from the LOEC to the peak) for each arsenical were as follows: As111 SA, 1.80- 3.90; AsvDA, 1.20- 2.90; MMAV, 1.10 (only 1 positive value); DMAV, 1.0-3.10. The control value was 0.30. Reference Tsuchiya et al., 2005 Signal Transduction MGC-803 (human gastric cancer) Primary cultures of rat cerebellar neurons As111 ATO As111 SA DMAV 0.01-1 10 5mM 48hr 4hr 8hr 0.01 10 5mM Increase in intracellular Ca2+ as measured by a Ca2+ sensitive fluorescent probe Indo- 1/AM in flow cytometric assays, which parallels the effect on apoptosis. For both: ft in activated p38MAPkinase. Zhang et al., 1999 Namgung and Xia, 2001 C-284 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Primary cultures of rat cerebellar neurons SY-5Y cells HEK 293 cells Arsenic Species As111 SA As111 ATO for both Concentration(s) Tested GoM) 10 0.1, 1 for both Duration of Treatment Ihr ~lhr for both LOECa (HM) 10 0.1 0.1 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft JNK3 MAP kinase. No change in JNK1 and JNK2 MAP kinases. (Blocking the p38 and JNK signaling pathways inhibited arsenite- induced apoptosis.) The Ca2+ concentration in cells was substantially increased (and by rather similar amounts) by both doses; inorganic arsenic triggered 3 different kinds of Ca2+ signals: slow (sustained), transient elevations, and calcium spikes. The irreversible increases were independent of extracellular Ca2+ and dependent on internal Ca stores, which could become depleted. Little or no cytotoxicity resulted from these doses during the time of measuring Ca2+ concentrations. Reference Namgung and Xia, 2001 Florea et al., 2007 C-285 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue UROtsa cells Postconfluent PAEC cells in a monolayer Arsenic Species As111 SA Asv MMAmO MMAV DMAmI DMAV As111 SA Concentration(s) Tested (|oM) 0.1,0.5, 1,5 1, 10, 100 0.1,0.5, 1,5 1, 10, 100 0.1,0.5, 1,5 1, 10, 100 0.5,2,5 Duration of Treatment Up to 2hr for all Ihr LOECa (HM) Various None Various None Various None 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Phosphorylation of ERK2: ft with potencies: MMAmO » DMAmI » As111 AP-1 binding activity: for As111: U at 0.1 and 0.5; forMMAmO,bigftat 0.1,0.5, 1.0 but no increase at 5; for DMAmI: UatO.l, 0.5, and 1, and big ft at 5. Phosphorylation of c- Jun: for As111: Hat 0.1 and 0.5 and ft at 1 and 5; forMMAmO,bigftatl and5;forDMAmI:ftat 0.1, big ft at 5. Also trivalent arsenicals caused changes in Fra-1 and induced AP-1 dependent gene transcription. There was no effect on c-Jun N- terminal kinases or p38 kinases. EMS A analysis: ft nuclear retention of NF- KB binding proteins; ft nuclear translocation of NF-KB binding proteins. Supershift analysis showed that p65/p50 heterodimers accounted for the majority of proteins binding consensus KB sequences in cells treated with As111 or oxidants. These and other experiments suggest that As111 initiates vascular dysfunction by activating oxidant- sensitive endothelial cell signaling. Increased binding of proteins to genomic KB sites could induce a mitogenic or inflammatory response. Reference Drobna et al., 2002 Barchowsk y et al., 1996 C-286 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Gclm+/+ MEF cells Hepa-lclc7 cells 1RB3AN27 cells 1RB3AN27 cells Arsenic Species As111 SA for all As111 SA As111 SA As111 SA Concentration(s) Tested (|oM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) See rows under Apoptosis and Cytotoxicity for this citation for experimental conditions. Inorganic arsenic inhibits NFicB activation and nuclear translocation. Co-treatment or pretreatment with tBHQ appears to reverse the inorganic arsenic -mediated inhibition of NF-KB translocation, and it triggers the nuclear accumulation of the transcription factor Nrf2. tBHQ may cause its cytoprotective effects by inducing gene expression changes though activation of at least NF-KB and Nrf2. 6, 12, 25, 50 0.1,0.5, 1,5, 10 0.1,0.5, 1 Ihr 240 min 120 min 6 Various 0.1 ft AhR nuclear translocation, with a positive dose-response; other experiments showed that the translocation occurs by different mechanisms from those followed by ligands and that AhR- dependent gene expression is only weakly up-regulated by inorganic arsenic. Activation of nuclear transcription factors detected by EMSA: NF- KB: slight ft at 0.5, flat 1, huge ft at 5 and 10; effect at dose of 1 was considerably suppressed by co-treatment with NAC. Inorganic arsenic- induced degradation of IicBawas demonstrated in the cytosolic fraction. AP-1: ft at 0.1, slight ft at 0.5, huge ft at 1, 5, and 10; effect at dose of 1 was completely blocked by co-treatment with NAC. Phosphorylation (activation) of ERK detected by EMSA: huge ft at all 3 doses. Reference Kannet al., 2005b Kannet al., 2005a Felix etal., 2005 Felix et al., 2005 C-287 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue JB6C141 cells Hepa-lclc7 cells Arsenic Species As111 SA for both As111 SA Concentration(s) Tested (|oM) 20,40 40 2, 5, 10 0.1, 1,2,5, 10 2, 5, 10 2, 5, 10 Duration of Treatment 12 hr for both 5hr 48hr 5hr 5hr LOECa (HM) 20 40 2 0.1 None 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) ft COX-2 protein level. ft COX-2 transcription. However, deletion of NF-KB binding sites from the COX-2 promoter blocked this effect. Other experiments, including some in MEF cells, confirmed the requirement of the IKKJ3/NF-KB pathway for the induction of COX-2 by As111 (shown at protein and transcription levels). ftNqolmRNA expression, with a positive dose-response. ft Nqol enzyme activity, with a slightly higher and rather similar response at doses 1-10. NSEonNrf2mRNA levels. ft Nrf2 protein level, with a positive dose- response. These and other experiments showed that Nqol induction occurred through the Nrf2/ARE pathway with the following important steps: (1) inorganic arsenic markedly stabilizes Nrf2; (2) inorganic arsenic disrupts the Nrf2-Keapl- Cul3 complex in the nucleus, and (3) inorganic arsenic recruits NrfZandMaftothe ARE of Nqol. Inorganic arsenic does not recruit Keapl, Cul3, ubiquitin, c-Jun, or c-Fos to the ARE of Nqol. Reference Ouyang et al., 2007 He et al., 2006 C-288 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference WM9 cells OM431 cells K1735-SW1 cells and other melanoma cell lines As111 SA 2,4,6 A series of experiments (usually with durations of treatment between 30 min and 16 hr) demonstrated that inorganic arsenic up-regulated TRAIL-mediated apoptosis. Inorganic arsenic up-regulated surface levels of death receptors, TRAIL-R1 and TRAIL-R2, through increased translocation of these proteins from cytoplasm to cell surface. Furthermore, activation of cJun and suppression of NF-KB by inorganic arsenic caused up- regulation of the endogenous TRAIL and down- regulation of cFLIP gene expression, which was followed by cFLIP protein degradation and, finally, by acceleration of TRAIL-induced apoptosis. cFLIP is one of the main anti-apoptotic proteins in melanomas. Ivanov and Hei, 2006 HeLa cells As111 SA 100 4hr 100 Big ft in autophosphorylation (activation) of ASK 1 determined by autoradiography. Hansen et al., 2006 A431 cells As111 ATO 20 Many experiments, usually at dose of 20, and over various durations, and often involving inhibitors or other modulators, yielded the following information and conclusions: inorganic arsenic had following effects: ft p21 promoter activity, ft p21 mRNA level, ft p21 protein level. Transfection with a p21 siRNA reduced inorganic arsenic-induced p21 expression and reduced the inorganic arsenic-induced cytotoxicity after 24 hr by half. Conclusions: inorganic arsenic induced p21 activation via the EGFR-Ras-Raf-ERKl/2 pathway. ERK1/2 and JNK may differentially contribute to inorganic arsenic-induced p21 expression via the EGFR- Ras-Raf-ERKl/2 pathway. The ERK 1/2 and JNK pathways play opposing roles in inorganic arsenic- induced cytotoxicity. Huang et al., 2006 NHEK cells As111 SA 0.4 1,3,5,7 0.4 on days 3 and 5 only CyclinDmRNA level: ftto~3.2xonday3;ftto ~1.5x on day 5; NSE on other days. Hwang et al., 2006 NHEK cells As111 SA 0.4 1,3,5,7 0.4 on day 3 only Binding of transcription factors to their respective binding motifs within the cyclin D1 promoter by demonstrated by EMSA: ft for API to 1.9x;NSE on other days; ft for CREEP to 1.6x; NSE on other days. Note the correspondence with ft in mRNA level in row above; there was a hint of an ft for both on day 7. Hwang et al., 2006 C-289 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue NHEK cells DU145 cells HaCaT cells, trans-fected for use in a luciferase reporter assay HaCaT cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (|oM) 0.1,0.2,0.4 50, 100, 200 1.25,2.5,5 0.31, 1.25,5 Duration of Treatment 1,2,3,4,7 Ihr 12 hr 12 hr LOECa (HM) 0.4 on days 2-7 50 1.25 0.31 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Cyclin D protein level: ft to~1.35xondays2-4;ft to 2x on day 7. AMPK activation: ft at 50, big ft at 100, ft at 200; activation was blocked by preincubation with CAT, GSH, orNAC; tests with a dominant negative form of AMPK showed that AMPK activity is necessary for inorganic arsenic- induced VEGF expression. Other experiments showed that the arsenic -induced AMPK signaling pathway is independent ofthep38MAPkinase and PI-3 kinase pathways and that the blocking of AMPK activation markedly increased cytotoxicity from inorganic arsenic exposures of 50 or 100. ft cyclinDl transcription to 1.9x and then ft with dose to 2.4x at highest dose. Protein levels determined by Western blot assay: ft cyclin D 1 and then ft with dose to highest dose; other experiments showed that induction of cyclinDl required activation of NF-KB and also required IKK|3. It was suggested that the inorganic arsenic- induced stimulation of the transition from Gl to S phase that was reported in this paper occurred through a IKK|3/NF-KB/cyclin D 1 - dependent pathway. Reference Hwang et al., 2006 Lee et al., 2006c Ouyang et al., 2005 Ouyang et al., 2005 C-290 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference Primary keratinocytes (in third passage) obtained from foreskins of adults As111 SA 0.1, 1,5, 10 48 hr Various Enzyme activities detected by luciferase assays: NF-KB: ft to -l.SxatO.l, ft to ~5x at 1, ft to ~3.8x at5,NSEatlO. AP-1: NSEatO.l, ftto~1.7xat 1, ft to ~2.7x at 5, ft to ~4.5x at 10. All results were confirmed at the protein level. Liao etal., 2004 H9c2 cells As111 SA 1,2.5,5 for 1 or 2 days There was a large dose-related U in cell migration rates at the 2 higher doses at both durations. NSE on viability (based on MTT assay) at these 3 doses, but at the dose of 10, which was not tested for other effects, there was a U in viability. There was a dose-related U in focal adhesions per cell at all doses and a U in F-actin content of cells at the dose of 5. At doses of 2.5 and 5, there was a U in tyrosine phosphorylation of FAK, a U in phosphorylation of FAK at phosphotyrosine 397, and a U in tyrosine phosphorylation of FAK's substrate paxillin. Inorganic arsenic affected focal adhesion structure or formation and not the turnover or amounts of focal adhesion proteins. Focal adhesions are involved in integrin signaling, and the inorganic arsenic- induced changes may disrupt cell contraction and signaling. It was concluded that inorganic arsenic decreases cell migration through an effect on focal adhesions and by disruption of cell interactions with the extracellular matrix. Yancy et al., 2005 MEFs from wild type or Ikk(3 gene knockout As111 (AsCl3) Various between 1.25 and 50 mouse embryos In a series of experiments lasting for 2-32 hr, the main findings were as follows. In knockout MEFs, which exhibit NF-KB inhibition due to IKKp deficiency, (1) there was a big ft in basal levels of mRNAs of the following genes: gadd45a, gadd45p, gadd45y and gadd!53; (2) there was a big ft in inorganic arsenic- induced (usually at 20 uM for 4 hr) levels of mRNAs for gadd45a and gadd!53; and (3) there was no induction by arsenic (same conditions) of mRNAs for gadd45p and gadd45y. It appears that NF-KB activation is an inhibitory signal for the expression of gadd45a and gadd!53. C-myc expression was reduced in knockout cells, and IKKp deficiency did not affect p53 or Akt signaling and the expression of FoxO3a. Zhang et al., 2005 JAR cells As111 ATO 0.5, 1,2.5,5, 10 6hr 0.5 Big ft in nuclear Nrf2 protein level, with dose- related ft becoming huge by dose of 10; also similar ft in cytoplasmic Nrf2 protein level. Massrieh et al., 2006 C-291 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue JAR cells BEAS-2B cells SVEC4-10 cells SVEC4-10 cells Arsenic Species As111 ATO As111 SA As111 SA As111 SA Concentration(s) Tested (|oM) 5 5 10 10 Duration of Treatment 2, 4, 6, 16, 24 hr 4hr 3 min to 4 hr 3 min to 4 hr LOECa (HM) Various 5 10 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Big ft in nuclear Nrf2 protein level at first 4 time points, but small ft at24hr. Slight ft in MafF protein level at many time points, but NSE on 2 other dimerization partners of Nrf2, namely MafG and MafK. Experiments done in part in HEK293T cells suggested that in JAR cells there is an ft in binding of endogenous Nrf2/smallMafDNA- binding complexes to a StRE site. Huge ft in nuclear Nrf2 protein level. Other experiments showed inorganic arsenic caused ft in Nrf2 transcriptional complex binding to the HMOX-1 ARE cis element. inorganic arsenic induced actin filament reorganization to form lamellipodia and filopodia structures at the leading edge of the cells and rosette-like structures in the cell bodies. Effects were noted after only 3 min; longer treatments did more damage. Reorganization of actin filament occurred through the activation of Cdc42. Huge ft in activation of Cdc42 already after 3 min and level of activation stayed almost as high for at least 1 hr; by 4 hr the level of activation was similar to that of control. See comment in row above. Reference Massrieh et al., 2006 O'Hara et al., 2006 Qian et al., 2005 Qian et al., 2005 C-292 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SVEC4-10 cells JB6C141 cyclinDl-Luc massl cells JB6C141 cyclin Dl-Luc massl cells JB6C141 cells Arsenic Species As111 SA As111 SA As111 SA As111 SA Concentration(s) Tested (|oM) 10 5 2 0.1,0.5, 1,5, 10 Duration of Treatment Various 24hr 12 hr Ihr LOECa (HM) Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) As more information about series of experiments described in 2 rows immediately above, it was shown that inorganic arsenic-stimulated Cdc42- induced actin filament reorganization regulated the activation of NADPH oxidase. Authors suggested that the formation of superoxide anion radicals observed after inorganic arsenic treatment occurred through the activation of NADPH oxidase. Rac activities were required for Cdc42- mediated superoxide anion radical production, and NADPH oxidase activity was involved in inorganic arsenic-stimulated cell migration via Cdc42-mediated actin filament reorganization. 5 2 Various Protein level determined by Western blot assay: huge ft in cyclin Dl; separate treatments with vanadate, cadmium, or NiCl2: NSE. mRNA level determined by luciferase reporter assay: ft in cyclin Dl to~3.2x. Protein levels determined by Western blot assay: Phosphorylation of Akt Ser473:ftat0.1-5,bigft at 10. Phosphorylation of Akt Thr308: slight U at 0.1, U at 0.5, ft at 1 and 5, big ft at 10. Phosphorylation of p70S6KThr389:bigftat 0.1-10. Phosphorylation of p70S6KThr421/Ser424:ft at 0.1-10. Reference Qian et al., 2005 Ouyang et al., 2006 Ouyang et al., 2006 Ouyang et al., 2006 C-293 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue JB6C141 cells HEL cells A549 cells Arsenic Species As111 SA As111 ATO As111 SA Concentration(s) Tested (|oM) 5 0.25,0.5, 1,2,5, 8, 10forP-STAT3 0.5, 1, 2, 4, 6, 8, 10 forHSPVO 1, 5, 10, 20 Duration of Treatment 20min 6hr for both 24 hr LOECa (HM) 5 0.5 1 10 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Protein level determined by Western blot assay: big ft in PI-3K activation as shown by big ft in PIPS; inhibition experiments showed that inorganic arsenic triggered a PI-3K/Akt/IKKp/NF-KB signal cascade that played essential roles in inducing cyclin D 1 expression. Western blot analysis: U P-STAT3 protein level (IC50s=1.334);3HSP90 inhibitors all markedly potentiated the effect with iC50s of 0.0468, 0.395, and 0.745. ft HSP70 protein level. Dose of -2.9 doubled the control level. 3 HSP90 inhibitors all markedly potentiated the effect. HSP70 inhibits apoptosis. Much more inorganic arsenic was needed to kill half the cells in 6 hr (LC50, e t = 80) than to down-regulate P-STAT3 by 50% in 6 hr (1.334, as above). The trypan blue assay was used to determine cell survival. inorganic arsenic activated the binding of IRP-1 to IRE: ft to 1.35x, with smaller ft to 1.25x at dose of 20; 10 and 20 uM inorganic arsenic caused a slight ft in HIF- 1 a protein level (only -2% above control). Inorganic arsenic at dose of 20 had NSE on ferritin protein level. Reference Ouyang et al., 2006 Wetzler et al., 2007 Li et al., 2006b C-294 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue CL3 cells, synchronous atGl CL3 cells, synchronous atGl Arsenic Species As111 SA As111 SA Concentration(s) Tested (|oM) 50, 100 5, 10, 25, 50 Duration of Treatment 3hr 3hr LOECa (HM) 50 Varied Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) At 50, ft in phosphorylation (activation) of ERK1/2 to 1.7x right after treatment and to 2.0x after a 24-hr recovery period. At 100: ft to 2.3x immediately; no activation of ERK1/2 occurred following co- treatment with PD98059 orU0126. Dose-related ft in phosphorylation (activation) of ERK1/2 to~1.45xat25and ~1.8xat50,LOEC=10. Dose-related U in efficiency of synthesis of NER to -50% and -41% of control at doses of 25 and 50, respectively. LOEC = 25;forboth ERKl/2andNER,the changes at 5 and 10 were NSE. NER synthesis efficiency was determined based on whole cell extracts of treated cells in an assay with UV-irradiated pUC19 plasmids. Co- treatment of inorganic arsenic with either PD98059orU0126 blocked much of the phosphorylation of ERK1/2 and restored 50%-80%oftheNER synthesis efficiency. In summary, co-treatments of inorganic arsenic with inhibitors that blocked activation of ERK1/2 did the following: (1) ft NER synthesis efficiency, (2) U induction of micro nuclei, (3) U survival, and (4) U rate of proliferation. Reference Li et al., 2006a Lietal., 2006a C-295 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue Arsenic Species Concentration(s) Tested QaM) Duration of Treatment LOECa (HM) Results (Compared With Controls, With All Concentrations Being in nM Unless Noted) Reference Effects detected by Western blot assay following 30-min treatment: increases in p-EGFR, p-ERK, p-JNK, A431 cells As111 ATO pp-38, andp21 WAFl/CIPI (an immunoblot assay was used 20 for p21). NSE for INK. A series of experiments involving modulators and reporter genes showed that: (1) EGFR activation can mediate inorganic arsenic- induced p21 expression, (2) activation of EGFR by inorganic arsenic occurred later than by EOF, (3) c-Src was involved in inorganic arsenic-induced ERK activation and p21 expression, (4) the EGFR/Ras/Raf/ERK pathway is involved in inorganic arsenic-induced p21 gene expression, (5) Spl binding sites in the promoter are essential for inorganic arsenic- induced p21 activation, and (6) a post-transcriptional or post-translational stabilization mechanism is essential for inorganic arsenic-induced p21 expression. Liu and Huang, 2006 MDA-MB- 435 cells As111 SA 1, a non-cytotoxic dose 0.5 hr, Ihr 2hr 4hr 6hr Effects on the nuclear binding of the following 4 transcription factors, relative to control and in sequential order of the 5 time periods: AP-LNSE, NSE, ft 2.5x, ft 2.5x, NSE. NF-KB: NSE, U 0.5x, ft 3.5x, ft 3.5x, ft 1.5x. Spl:li0.5x, Uo.5x, ft 3x, NSE, NSE. YB-1:NSE,NSE, ft 9x, ft 3x, ft 3x. Another experiment using a highly cytotoxic dose of 100 resulted in markedly different patterns over time within approximately the same ranges of effect. Kaltreider etal., 1999 C-296 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue H41 IE cells SIK cells, with inorganic arsenic treatment beginning 1 day before suspension and continuing while cells were in suspension Arsenic Species As111 SA As111 SA Concentration(s) Tested (|oM) 0.33, a non-cytotoxic dose 2 Duration of Treatment 0.5 hr, Ihr 2hr 4hr 6hr 2 or 5 days, when including the 1 day of treatment before being put into suspension LOECa (HM) — 2 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Effects on the nuclear binding of the following 4 transcription factors, relative to control and in sequential order of the 5 time periods: AP-LNSE, ft5.5x, ft 8.5x, NSE, U 0.5x. NF-KB:ft 1.3x,NSE, ft 1.5x,NSE,NSE. Spl: NSE at any time. YB-1: ft 3x, ft 3x, ft 3.2x,NSE, Uo.5x. Another experiment using a highly cytotoxic dose of 333 resulted in markedly different patterns over time within approximately the same ranges of effect. Protein levels determined by immunoblotting assay: (3-catenin: ft to 3.2x on day 2 and to 3.6x on day 5; (31-integrin: ft to 2.7x on day 2 and to 4.0xonday5;p-GSK3p (the inactive form): ft to 2.5x on day 2 and to 2.2x on day 5. On day 1, in cells harvested before suspension, there was ft inp-GSK3(3to l.Sxand NSE for other two proteins. Consistent with inorganic arsenic maintaining the cell's proliferative potential, levels of these 3 proteins decreased much less rapidly during the 4 days in suspension if treated with inorganic arsenic. Reference Kaltreider etal., 1999 Patterson et al., 2005 C-297 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue SIK cells treated while being maintained in surface cultures B16-F10 cells Arsenic Species As111 SA As111 SA for all Concentration(s) Tested (|oM) 2 0.01,0.1, 1, 10 0.01,0.1, 1, 10 0.01,0.033,0.1 Duration of Treatment Various 4hr 72 hr 7 days LOECa (HM) 2 None 0.01 0.01 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Protein levels determined by immunoblotting assay: nuclear (3-catenin: ft to 3. 4x on day 9; cytoplasmic (3-catenin: ft to 2.5x on day 11; p-(3- catenin: U to 0.45x on day 1; pl-integrin: ft to 1.6x on day 7 and to 4.5x on day 11; P-GSK3P (the inactive form): ft to 1.8x on day 7 and to 3. Ix on day 11. Consistent with inorganic arsenic maintaining the cell's proliferative potential, inorganic arsenic decreased the rate of post-confluent loss of all of these proteins except P-p-catenin. HIF-la protein levels: No ft seen; no mention if there were decreases. Small ft, but decreased to no change from control at 0.1, and at higher doses a U from control. Big ft, but decreased to almost 2 times control at 0.033, and there was no change from control at 0.1. Reference Patterson et al., 2005 Kamat et al., 2005 C-298 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue B16-F10 cells J82 cells HMEC-1 cells SMC cells J82 cells HMEC-1 cells SMC cells H22 cells Arsenic Species As111 SA for all As111 SA for all As111 SA for all As111 ATO Concentration(s) Tested (|oM) 0.01,0.1, 1, 10 0.01,0.1, 1, 10 0.01,0.1 0.01,0.1 0.01,0.1 0.01,0.1 0.01,0.1 0.01,0.1 0.01,0.1 0.5, 1,2,4 Duration of Treatment 4hr 72 hr 7 days 7 days for all 7 days for all 36 hr LOECa (HM) 0.1 0.01 0.01 0.01 0.01 0.01 0.01 0.01 None 0.5 Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) VEGF secretion: Small ft seen at two middle doses. Big ft; much smaller increase over control at 0.1; no change from control at 1.0, and a U from control at 10. Big ft, but U from control at 0.1. Also, both 7-day treatments showed ft HRE transactivation that was mostly or completely blocked by co-treatment with YC-1 . HIF- la protein levels: ft, but U from control at 0.1. ft, but U from control at 0.1. Big ft at both dose levels. VEGF secretion: ft, but no change from control at 0.1. ft, but no change from control at 0.1. Huge ft in VEGF protein level in cell lysates, with similar response at all doses (Westen blot assay). Reference Kamat et al., 2005 Kamat et al., 2005 Kamat et al., 2005 Liu et al., 2006e C-299 DRAFT—DO NOT CITE OR QUOTE ------- Type of Cell/Tissue MCF-7 cells MCF-7 cells HeLa cells Arsenic Species As111 ATO As111 ATO As111 ATO Concentration(s) Tested (nM) 0.5, 2, 5, 10 2,5 2 Duration of Treatment 60 min LOECa (UM) Various Results (Compared With Controls, With All Concentrations Being in |iM Unless Noted) Results of Western blot analysis: ft phosphorylation (i.e., activation) of ERK1/2 at 2 with a dose-related increase to 10; ft phosphorylation (i.e., activation) of p38 at 2 with a dose-related increase to 10; NSE on JNK1/2. Thus, inorganic arsenic activates the prosurvival mitogen- activated MEK/ERK pathway. The findings in the row above prompted investigation whether a combination treatment with either of the MEK inhibitors U0126 (at 10 uM) and PD98059 (at 20 uM) could lead to enhanced growth inhibition and apoptosis. They did, augmenting apoptosis approximately 2x compared to the effects of inorganic arsenic and either inhibitor alone, based by Hoechst 33258 or annexin V/PI staining and flow cytometry. Treatment with a p38 inhibitor did not prevent inorganic arsenic-induced apoptosis; there was a slight but nonsignificant ft in apoptosis. Various None Experiments showed that NF-KB and AP-1 activation served as prosurvival or antiapoptotic forces and that their activation by PMA was suppressed by co-treatment with 10 uM emodin plus inorganic arsenic, whereas emodin or inorganic arsenic alone had rather little or no suppressive effect. The synergistic suppression was thought to be caused, at least in part, by cellular ROS because pretreatment or co-treatment with NAC reduced the effect. Reference Yeetal., 2005 Yeetal., 2005 Yietal., 2004 a Lowest observed effect concentration. C-300 DRAFT—DO NOT CITE OR QUOTE ------- APPENDIX D. IMMUNOTOXICITY 1 Arsenic has been observed to affect the immune system. While changes to the immune 2 system are not directly related to any specific disease or cancer endpoint, disruptions to the 3 immune function can impact the individual and likely increase their risk for developing a disease 4 or cancer outcome. This may, in part, be why there are so many cancers and diseases associated 5 with arsenic exposure. In addition, arsenic's effects on the immune response may play some role 6 in acting as a co-carcinogen with other compounds. Therefore, immunotoxicity is listed as a key 7 event in this review, and many studies are detailed in the MOA section (4.4.1). The effects, 8 however, on the immune system are important to note in and of themselves, and a few are 9 detailed here. 10 Gonsebatt et al. (1994) selected two populations from Comarca Lagunera (Mexico) to 11 study the labeling, mitotic, and replication indexes (LI, MI, and RI, respectively) of peripheral 12 blood lymphocytes. The exposed population consisted of 33 individuals from Santa Ana, 13 Coahuila, who had arsenic levels in the drinking water ranging from 375 to 392 ppb (92% in the 14 form of Asv and 8% in the form of As111) for several years. Approximately 50% of the exposed 15 individuals had cutaneous signs of arsenic poisoning. The control population consisted of 30 16 individuals (selected based on similar proportions of age and sex as the exposed population) 17 from Nazareno, Durango, who had arsenic levels from the preceding 2 years ranging from 19 to 18 26 ppb (>95% as Asv). Urine and blood samples were obtained from all subjects. Average total 19 arsenic in the urine and blood of the control group was 36.7 and 37.2 ug/L, respectively, and 20 758.4 and 412.0 ug/L, respectively, in exposed subjects. Peripheral blood lymphocyte counts 21 were significantly greater in the exposed population (3.1 x 106 cells) compared to the control 22 population (2.6 x 106 cells), with a greater increase noted in females. In females, the average 23 36-hour LI was lower in the exposed population than the control population, regardless of the 24 presence or absence of skin lesions (Table D-l below). Only exposed males with skin lesions, 25 however, exhibited a lower average 36-hour LI; males without skin lesions had an increase in LI. 26 MI were significantly increased in both sexes after a 72-hour incubation period (Table D-l), but 27 were not after 48 or 60 hours of incubation. Exposed females had a significantly lower RI after 28 48, 60, and 72 hours of incubation, while males were unaffected. D-1 DRAFT—DO NOT CITE OR QUOTE ------- Table D-l. Lymphocyte counts and labeling, mitotic, and replication indexes (mean ± se) in the peripheral blood lymphocytes in populations exposed to low (control) and high (exposed) levels of arsenic (Gonsebatt et al., 1994) Lymphocyte Count (xlO6 cells/ml) Labeling Index (36 hours) with skin lesions without skin lesions Mitotic Index 48 hours 60 hours 72 hours Replication Index 48 hours 60 hours 72 hours Control Males 2.7±1.2 3.32±1.06 1.15±0.23 2.53±0.30 3.52±0.37 1.07±0.01 1.43±0.03 1.93±0.09 Females 2.4±1.1 4.77±1.06 2.52±0.48 4.90±0.79 3.96±0.52 1.16±0.02 1.54±0.04 2.08±0.04 Total 2.6±1.1 3.37±0.61 1.89±0.30 3.85±0.50 3.78±0.34 1.12±0.01 1.49±0.03 2.01±0.04 Exposed Males 3.0±1.2 2.14±0.86 4.05±1.31 1.59±0.29 3.35±0.39 6.00±0.55b 1.05±0.01 1.39±0.04 1.89±0.09 Females 3.1±0.8a 2.74±0.32 3.90±0.49 1.59±0.26 3.65±0.48 6.60±0.69b 1.10±0.02a 1.37±0.04a 1.84±0.05a Total 3.1±1.0a 2.42±0.49a 3.95±0.56 1.59±0.20 3.50±0.32 6.34±0.45b 1.08±0.01a 1.38±0.02a 1.86±0.05a a Statistically different (p < 0.05) from the control (two-tailed Mann-Whitney U-test). b Statistically different (p < 0.001) from the control (two-tailed Mann-Whitney U-test). 1 A previous study by Ostrosky-Wegman et al. (1991), in which cell-cycle kinetics of 2 peripheral blood lymphocytes showed a significantly longer average generation time (AGT) for 3 the highly exposed group as compared to the control group. The AGT was 19.90 hours in the 4 low-exposure group compared to 28.70 hours in the high-exposure group. The AGT in the 5 control group was 19.02 hours. The exposed group consisted of 11 individuals (9 females and 2 6 males) from Santa Ana, Coahuila, with drinking water containing an average of 390 ppb (98% as 7 AsV). The control group consisted of 13 individuals (11 females and 2 males) from Nuevo 8 Leon, Coahuila (drinking water concentrations not reported). Average urine arsenic/creatinine 9 levels were 0.121 ug/mL in the control group and 1.565 ug/ml in the exposed group. There were 10 no incidence of skin lesions in the control subjects, but 4 of the 11 exposed subjects had skin 11 lesions (i.e., hyperkeratosis, hypo- and hyperpigmentation, and skin horns). 12 IgG and IgE levels were significantly elevated in arsenic-exposed individuals who 13 presented clinical symptoms of arsenic exposure (i.e., skin lesions) (Islam et al., 2007). As 14 exposure duration increased, so did the severity of the skin lesions. The level of IgE also was D-2 DRAFT—DO NOT CITE OR QUOTE ------- 1 greater with longer durations of arsenic exposure. IgG levels were highest during the initial 2 stages of skin lesions. There was a smaller, but significant, increase in IgA in individuals with 3 arsenicosis compared to the control group, but no change was observed in IgM levels. Arsenic- 4 exposed individuals also had a greater incidence (63% of subjects) of respiratory complications, 5 such as chest sounds, shortness of breath and breathing complications, irritation of the upper and 6 lower respiratory tract, cough, bronchitis, and asthma than the control group (7%). The IgE level 7 in individuals with respiratory complications was greater than in arsenic-exposed individuals 8 without complications. Because the difference in IgE levels could not be explained by 9 differences in eosinophil levels, it was suggested that the reason may be inflammatory reactions 10 due to arsenic exposure. 11 Yu et al. (1998) found that patients with Bowen's disease (skin carcinoma in situ) from 12 an arsenic-endemic area in the southwest coast of Taiwan had significantly decreased T-cells, 13 increased B-cells, decreased T-helper cells, decreased IFN-y release, decreased TNF-a release, 14 increased IL-2 release, decreased soluble IL-2 receptor release, and changes in soluble CD4 and 15 soluble CDS release (increases in spontaneous release, but decreases in phytohaemagglutinin- 16 induced release) in comparison to normal controls, as well as non-Bowen's patients in the 17 endemic region. Results indicate a depressed cell-mediated immunity in patients with Bowen's 18 disease. The deficient immune response appears to be related to an impairment of the membrane 19 IL-2R expression in lymphocytes after stimulation. This study, however, could not associate 20 arsenic with these changes because individuals without Bowen's disease who lived in the 21 endemic region did not demonstrate the same effects. In addition, a cause and effect relationship 22 could not be determined. Since arsenic has been demonstrated to affect the immune response in 23 other studies, it is possible that individuals developing Bowen's disease were more susceptible to 24 the effects of arsenic on the immune system. 25 The development of skin lesions is a typical symptom of arsenic-exposed individuals. 26 However, not all individuals exposed, even those within the same family, develop skin lesions. 27 Mahata et al. (2004) examined some effects on peripheral lymphocytes in arsenic-exposed 28 individuals with or without skin lesions and compared those results to a group of subjects that 29 were unexposed. Six individuals (3 males and 3 females) were selected from each group: 30 symptomatic (with lesions and exposure), asymptomatic (without lesions and exposure), and 31 unexposed. Where possible, symptomatic and asymptomatic were selected from the same 32 family. The 6 controls (3 males and 3 females) were selected for similar socio-economic status, 33 age, and gender. Levels of arsenic in urine, nail, and hair demonstrated that the control group 34 had little exposure to arsenic. Individuals with skin lesions were noted to have less arsenic in 35 their urine and more in their hair and nails. This indicated individual differences in distribution 36 and excretion (for more information on this see Section 4.7.3.1 on genetic polymorphism) that 37 may be related to the individual's susceptibility to developing skin lesions. When the blood of D-3 DRAFT—DO NOT CITE OR QUOTE ------- 1 the individuals from all three groups was exposed to further arsenite in vitro, all groups 2 demonstrated a dose-dependent increase in chromosomal aberrations in the lymphocytes, with a 3 significant increase observed even at the lowest concentration (1 uM). Untreated lymphocytes, 4 however, had a greater level of chromosome aberrations in arsenic-exposed individuals. In 5 addition, individuals with skin lesions had a 1.7-fold increase in "background" chromosomal 6 aberrations compared to asymptomatic individuals. Although the arsenic-exposed individuals 7 had more chromosomal aberrations in the absence of additional arsenic exposure in vitro, the in 8 vitro exposure to arsenite caused a smaller increase in chromosome aberrations in lymphocytes 9 of exposed individuals compared to unexposed individuals, indicating a greater sensitivity in the 10 control lymphocytes to the in vitro effects of As111. 11 The JAK-STAT pathway is essential in mediating the normal functions of different 12 cytokines in the hematopoietic and immune systems. In vitro studies by Cheng et al. (2004) 13 suggest that arsenic exposure in the range of 0.4 to 400 uM caused inactivation of the JAK- 14 STAT signaling pathway in HepG2 cells (a human hepatocarcinoma cell line). This inactivation 15 was caused by the direct interaction of arsenic with JAK tyrosine kinase and was independent of 16 arsenic activation of mitogen-activated protein (MAP) kinase. Exposure to sodium arsenite 17 abolished the STAT activity-dependent expression of cytokine signaling suppressors by 18 inhibiting IL-6-inducible STAT3 tyrosine phosphorylation. This effect on the STAT3 tyrosine 19 phosphorylation induced by arsenic was not observed with other inflammatory stimulants, stress 20 agents, or cadmium (metal). 21 Harrison and McCoy (2001) performed an in vitro study to examine the role of apoptosis 22 and enzyme inhibition in arsenic's suppression of the immune response. Cysteine cathepsins are 23 lysosomal enzymes that are critical in antigen processing. Because of As111 interactions with 24 sulfhydryl groups, cathepsin L (a member of the papain superfamily of cysteine proteases and a 25 major lysosomal protease involved in cleaving exogenous protein antigens into peptide 26 fragments) was examined as a potential target for arsenic inhibition. As111 caused a dose- 27 dependent inhibition of cathepsin L, both as a purified enzyme and in active murine B cells. 28 Inhibition occurred in TA3 cells even at concentrations that did not affect cell viability; greater 29 inhibition was obtained with the purified enzyme. Addition of DTT caused a complete reversal 30 of the inhibition. AsV was not able to inhibit cathepsin L, therefore, indicating the involvement 31 of the sulfhydryl groups. Although cathepsin L was inhibited by 4 hours, exposure for 18 hours 32 led to increases in apoptosis even at the lowest concentration (0.8 uM). Apoptosis was observed 33 at concentrations about 100-fold lower than those inhibiting cathepsin L, suggesting that 34 apoptosis likely plays a more important role in immunosuppression than inhibition of lysosomal 35 cathepsins. 36 De La Fuente et al. (2002) found a significant increase in apoptosis in PMBCs from 37 healthy donors at concentrations of 15 uM As111 after 48 hours of exposure; an increase also was D-4 DRAFT—DO NOT CITE OR QUOTE ------- 1 noted at 5 uM, but did not reach statistical significance. Results did not show a dose-response; 2 instead apoptosis levels were similar between 15 and 75 uM of arsenite. Lower concentrations 3 of As111 (i.e., 1 uM and 2.5 uM) also were able to increase apoptosis levels, but required at least 4 96 hours of exposure compared to only 16 hours of exposure needed with 15 uM of As111. 5 Measuring the levels of apoptosis in the PMBCs of children chronically exposed to arsenic 6 (urinary levels of arsenic between 94 and 240 ug/g of creatine) also demonstrated an increase in 7 apoptosis when compared to the control group. However, exposing the cells of chronically 8 exposed children to arsenic in vitro demonstrated a decrease in apoptosis when compared to 9 controls. Therefore, these data support the data of Mahata et al. (2004), which suggested that 10 control PMBCs are more sensitive to in vitro arsenic exposure. 11 In contrast, Gonzalez-Rangel et al. (2005) found the opposite response. Although their 12 data also show an increase in basal apoptosis in PMBCs lymphocytes and monocytes (but not 13 natural killer [NK] cells) in an exposed individual compared to six non-exposed individuals, the 14 data also show an increased sensitivity to in vitro arsenite-mediated apoptosis in lymphocytes 15 and NK cells in the exposed individual. This study, however, used a higher concentration of 16 arsenite (30 uM) compared to the other studies (which used at most 15 uM) and only used one 17 exposed individual compared to 6 unexposed individuals. Therefore, results could be different 18 due to dose or because of inter-individual variation. 19 Although Harrison and McCoy (2001) and De La Fuente et al. (2002) observed an 20 increase in the apoptosis in PMBCs, Chen et al. (2005b) did not observe any effect on the 21 apoptosis of human keratinocytes (obtained from the adult foreskin through routine 22 circumcision) with 1 uM of arsenite. When cells were exposed to As111 for 24 hours prior to 23 exposure to UVB, however, the As111 protected against UVB-induced apoptosis, indicating a 24 possible mechanism for arsenic's observed co-carcinogenic nature. Exposing the cells to As111 25 after UVB exposure did not cause a reduction in apoptosis and possibly increased the level of 26 apoptosis. 27 Galicia et al. (2003) isolated PBMC from healthy, non-smoking, males (22-40 years old) 28 who were not exposed to arsenic to examine the effects of As111 on cell proliferation. Although 29 there was individual variability, a dose-dependent decreased in cell proliferation in PHA-induced 30 cells was observed. In all cases, the highest concentration used (1 uM) decreased the percent of 31 dividing cells, with a reduction of 12% to 54%. Although cell proliferation was affected, there 32 was relatively little affect on cell viability. After further examination, it was suggested that 33 proliferation of T lymphocytes was affected and there was a reduction in CD3+ cells producing 34 IL-2. Although arsenic prevents cells from entering the cell cycle and slows down the 35 progression through the cell cycle, no alteration in the expression of CD69 or CD25 activation 36 molecules was observed. Thus, it was concluded that the reduction in T cell proliferation was 37 caused by a decrease in the production and secretion of IL-2, thereby blocking the T cells in Gl. D-5 DRAFT—DO NOT CITE OR QUOTE ------- 1 Di Gioacchino et al. (2007) examined the effects of arsenic compounds (i.e., As111, AsV, 2 MMAV, and DMAV) on PBMC proliferation and cytokine release. As111 had the greatest effect 3 on the cells. At 10-4 M, As111 caused the greatest decrease in PHA-induced cell proliferation and 4 the greatest reduction in IFN-y and TNF-a release. At 10-4 M, the effect on cell proliferation by 5 compound was Asin>AsV>DMAv>MMAv. At 10-7 M, however, As111 caused a significant 6 increase in cell proliferation. DMAV also caused a significant increase in cell proliferation at 10- 7 7 M, but had no effect on cell proliferation at 10-4 M. DMAV and AsV caused a significant 8 decrease in IFN-y release at 10-4 M, but did not affect TNF-a release. Although the text 9 indicates that dose-response analyses were performed, the article provides no data. It was 10 concluded that the immunotoxicity of arsenic was dependent on the chemical speciation of 11 arsenic. 12 AsV (0.5, 5, or 50 mg As/L) administered for 12 weeks via drinking water to female 13 C57BL/6J/Han mice (8-12 weeks old) was determined to decrease NO and superoxide 14 production (Arkusz et al., 2005). While there was a concentration-dependent decrease in ROS 15 production, the decrease observed in NO was similar across the three doses. The AsV did not 16 appear to affect TNF-a production. It should be noted, that in testing for the NO and superoxide 17 production, 2 x 105 cells/well were plated. Therefore, a cell to cell comparison was made 18 between the isolated macrophages from the control group and arsenate-treated mice. The AsV 19 treatment, however, was noted to cause a concentration-dependent increase in the number of 20 peritoneal macrophages isolated. The percent increase compared to control (55%, 77%, and 21 101%, respectively) was such that it may have compensated for the changes noted in NO and 22 superoxide production. This, however, was not tested. 23 Nayak et al. (2007), however, did test the immune response in zebra fish to virus or 24 bacterial infection. Zebra fish embryos (one-cell stage) were exposed to 2 or 10 ppb As111 in 25 water until 4 days post-fertilization. Seven days later fish were infected or left uninfected. Viral 26 and bacterial loads were then examined. Viral load was significantly increased in both As111 27 treatment groups compared to the infected control group, with a concentration-dependent 28 increase in the viral load. There also was a significant increase in the bacterial load in treated 29 fish; however, the increase was similar across both treatments. As111 was also found to decrease 30 ROS burst, interferon, MX mRNA expression, IL-lp, and TNF-a mRNA levels. The maximum 31 response for these cytokines was also found to be delayed compared to the controls. D-6 DRAFT—DO NOT CITE OR QUOTE ------- APPENDIX E. QUANTITATIVE ISSUES IN THE CANCER RISK ASSESSMENT FOR INORGANIC ARSENIC 1 As discussed in Section 5.3, the arsenic cancer risk assessment involved two distinct 2 phases. The first phase involved the derivation and fitting of dose-response models using the 3 Taiwanese epidemiological data of Chen et al. (1988a, 1992) and Wu et al. (1989). The outputs 4 of this phase of the analysis were arsenic dose-response coefficients that described the 5 relationship between estimated arsenic intake in the Taiwanese population and proportional 6 increases in age-specific lung and bladder cancer mortality risk. A key assumption underlying 7 this relative risk model is that the risk of arsenic-related cancer is a constant multiplicative 8 function of arsenic dose and the "background" age profile of risks. 9 The second phase of the risk assessment involved the estimation of arsenic-related cancer 10 risks in a (hypothetical) U.S. population exposed to arsenic at varying levels in drinking water. 11 This phase of the analysis involved the application of the dose-response coefficients for arsenic 12 derived from the Taiwanese data to the age-specific background population risks for the U.S. 13 population. In addition, the risk estimates were converted from mortality-based values to 14 incidence-based estimates. The following sections describe each of these phases. E.I. CANCER RISK ASSESSMENT FOR THE TAIWANESE POPULATION 15 The calculation of cancer risks from the Taiwanese epidemiological data was performed 16 using Excel workbook files. The files contained the input data for the dose-response models and 17 spreadsheets to accept user-specified inputs, perform calculations, and summarize outputs from 18 the assessment. Input data included male and female lung and bladder cancer mortality and 19 person-years at risk (PYR) data for arsenic-exposed populations from 42 villages obtained from 20 Morales et al. (2000), village water arsenic concentrations (minimum, median, and maximum 21 data sets), and southwest Taiwan and all Taiwan reference population mortality and PYR data. 22 The user first specifies drinking water consumption and body weights for the Taiwanese 23 population in the "Poisson Model" page of the risk calculation files. Solver® is then invoked to 24 estimate the age coefficients (al, a2, and a3) and the arsenic dose-response coefficient (b) in 25 equation E-l by maximizing the likelihood function that is coded into the spreadsheets. Solver is 26 then reconfigured to calculate the upper confidence limit (UCL) on "b" using the profile 27 likelihood method (see below). The resulting UCL value is then input to the "BEIR Model" 28 sheet and the LEDM for cancer incidence is calculated, again using Solver®. The LED0i value is 29 transferred to the "Summary" sheet, where other risk metrics (unit risk, cancer risks at different 30 drinking water concentrations, and the drinking water concentration corresponding to 10-4 31 lifetime risk) are calculated. Risk metrics are calculated based on user-specified drinking water 32 intake and body weight for the U.S. population. Likelihood calculations for most of the E-1 DRAFT—DO NOT CITE OR QUOTE ------- 1 endpoints were replicated using a different optimizer in the Non-Linear Estimation module of the 2 Statistica® software package. E.2. MLE ESTIMATION OF DOSE-RESPONSE PARAMETERS 3 The Taiwan risk model spreadsheets calculate the dose-response parameters for the 4 Poisson model, fitting separate models for each endpoint: 5 6 h(x,t) = exp(ai + a2 x age + a3 x age2) x (l + b x dose) (Equation E-l) 7 8 In this model, the midpoints of the age group strata are normalized (placed on a "z- 9 scale") before risk is estimated; age is thus treated as a "nuisance parameter" in the model. 10 Dose, as noted above, is calculated from dietary arsenic and village water concentrations and is 11 expressed in terms of mg/kg-day. Each age-dose group is represented by a row on the 12 spreadsheet. There are 42 villages with arsenic well water data and the reference population, 13 each divided into 13 age strata, for a total of 559 population groups. The model begins with 14 randomly selected values for the four parameters and then calculates the Poisson log likelihood 15 values for each group: 16 17 log likelihood = observed x ln[hcuRRENi(x,t)] - predicted (Equation E-2) 18 19 where: 20 observed = the number of cancer deaths in groups age t, exposed at dose x 21 hcuRRENx(x,t) = the estimated total cancer risk in age group t at dose x, based on 22 the current parameter estimates 23 predicted = the predicted number of cancer deaths in age group t at dose x, = 24 hCuRRENx(x,t) x PYR, where PYR = person-years at risk 25 26 The sum of the log likelihood across all the age groups is then maximized using 27 standard optimization methods (Excel Solver®) to provide the MLE estimates of the age and 28 dose parameters. 29 E.2.1. Estimation of Upper Confidence Limits on the Arsenic Dose-Response Parameters 30 ED01 values are derived based on the MLE dose-response parameter estimates. The 31 LEDoi estimates are derived from the 95% upper confidence limits (UCLs) on the dose-response 32 parameters. The UCLs on the dose-response "b" parameters were estimated using the "profile 33 likelihood" method (Venson and Moolgavkar, 1988). In this approach, the value of the dose 34 parameter, b, was varied from its estimated mean value, and the changes in log-likelihood were 35 calculated. The ratio of the log likelihood for the best-fit model to the log likelihood for other 36 values of "b" is known to follow an approximate chi-squared distribution with one degree of 37 freedom. Thus, the 5th and 95th confidence limits on the dose coefficient "b" correspond to the 38 values where the likelihood ratio is equal to 1.92. Upper and lower confidence limits were E-2 DRAFT—DO NOT CITE OR QUOTE ------- 1 calculated using Solver®. The fact that the profile likelihood method ignores the likelihood 2 impact of the age "nuisance parameters" implies that the calculated confidence limits are only 3 approximate. Confidence limit calculations using other methods (empirical Bayesian simulation 4 and "bootstrap-t") gave similar values (within a few percent). E.3. ESTIMATION OF RISK FOR U.S. POPULATIONS EXPOSED TO ARSENIC IN DRINKING WATER 5 LEDoi values were calculated using a life-table method that is a variation on the "BEIR 6 IV" model recommended by NRC (2001). Specifically, the approach includes a modification 7 suggested by Gail et al. (1999) for obtaining more accurate estimates of incidence within multi- 8 year age strata. The BEIR IV relative risk model takes as its inputs the arsenic dose-response 9 "b" coefficient from the Poisson model, background cancer incidence data, along with age- 10 specific mortality data to directly estimate lifetime bladder and lung cancer incidence for the 1 1 target (U. S. adult) population. Lung and bladder cancer incidence reference data for the years 12 2000-2003 were obtained from the National Cancer Institute's SEER program (NCI, 2006). 13 U.S. gender and age-specific population data and all-causes mortality data came from the 14 National Center for Health Statistics (NCHS, 2000). 15 Formulas for calculating LEDOT values were implemented on separate Excel spreadsheets 16 for each endpoint. The following calculations were implemented in separate lines on each 17 spreadsheet. In all of the equations, the subscript "i" refers to age group: 18 L(x) = lifetime risk of cancer incidence at dose x 19 i (Equation E-3) 20 21 Numerator Terms: 22 23 q(x) = cancer incidence hazard at dose (x), age interval (i) 24 25 c;(x) = Ci(0) x (1 + beta x dose) (beta comes from the linear Poisson model) 26 27 c;(0) = background cancer incidence; / cancer free population; 28 29 Background cancer incidence c ; comes from SEER, cancer-free population;, see (7) 30 31 b; = background cancer incidence; / alive population; (SEER data) 32 33 Y; = exp (- 5b;) 34 35 F; = initial estimated background probability of survival without cancer to the 36 end of interval (i - 1) E-3 DRAFT—DO NOT CITE OR QUOTE ------- (Equation E-4) 2 3 G; = initial estimated background probability of survival without cancer to 4 the middle of interval (i) 5 f^- rr-(-2.5*bi) 6 CzZ = PI (Equation E-5) 1 8 Cancer-free population; = alive population; x G; 9 10 Denominator Terms: 11 12 s;(x) = total noncancer mortality and cancer incidence hazard, at dose (x) 13 in age interval (i) 14 15 s;(x) = background noncancer mortality (x, i) + cancer incidence hazard (x, i) 16 17 s;(x) = (d; - h;) + Ci(0) x (1 + beta x dose) 18 19 d; = total mortality (background) in age interval (i) 20 21 d; = total deaths; / population; (Census, Vital Stat. U.S.) 22 23 h; = cancer deaths; / population; (Census, Vital Stat. U.S.) 24 25 Survival (Ti and n) Estimation: 26 27 T; = probability of survival without cancer to end of interval (i - 1) 28 ;'=!-! ;=2-i ;=2-i Ti = Y\ri= [] Wr; * [] Wib 29 >'=1 >=1 >=1 (Equation E-6) 30 31 r; = probability of survival cancer free through interval (i), given survival to beginning of 32 interval (i) 33 34 n = Wi x Wib 35 36 W; = exp (-5d; + 5h; - 5c;) 37 38 Wib = exp (-5c; x Beta x x) 39 40 To calculate ED0i values, the value of the daily arsenic dose used to calculate h;(x), and 41 hence L(x), was varied until L(x) = 0.01 (1%). For the MLE estimation, Solver was used to 42 estimate LED0i values in the model spreadsheets. E-4 DRAFT—DO NOT CITE OR QUOTE ------- APPENDIX F. RISK ASSESSMENT FOR TOWNSHIPS AND LOW-EXPOSURE TAIWANESE POPULATIONS F.I. RECENT STUDIES OF THE TAIWANESE POPULATIONS THAT DO NOT FIND CONSISTENT EXPOSURE-RESPONSE RELATIONSHIPS 1 As discussed in Section 5.3.8.5, several recently published studies have called into 2 question the strength and significance of the exposure-response relationship for arsenic in the 3 Taiwanese population studies (Chen et al., 1988a, 1992; Wu et al., 1989). This appendix 4 provides a brief analysis of some of these concerns. 5 Based on "graphical and regression analysis," Lamm et al. (2003) found no significant 6 dose-response relationship for arsenic-related bladder cancer in the subset of the Taiwanese 7 population with median drinking water well concentrations less than 400 ppb (ug/L). 8 Significant, positive dose-response slopes were found for villages with median well 9 concentrations above 400 ppb. They also observed that all of the villages "solely dependent" on 10 artesian wells had median arsenic concentrations above 350 ppb, and that the median well 11 concentrations in villages not solely dependent on artesian wells were generally below this 12 value. Based on these observations, Lamm et al. (2003) suggested that the nature of the villages' 13 water sources (artesian vs. non-artesian), rather than arsenic concentration, explained the 14 observed variations in bladder cancer risk in the Taiwanese population. 15 Kayajanian (2003) also argued that EPA is misinterpreting the data from the Taiwanese 16 population. Kayajanian stratified median well arsenic concentration into 10 ranges from 10 to 17 934 ppb. The author then calculated combined mortality rates for lung, bladder, and liver cancer 18 for each stratum of the population. They calculated that crude (age-unadjusted) cancer mortality 19 for both males and females was significantly elevated in the lowest exposure groups, decreased 20 to minimums for villages with water arsenic concentrations between 42 and 60 ppb, and then 21 again increased with increasing arsenic exposure. They argued on this basis (and based on the 22 analysis of cancer mortality data from another epidemiological study) that health standards for 23 arsenic should be set in the vicinity of 50 ug/L (ppb) in order to minimize the risk of arsenic- 24 associated cancer, and that lower exposures would actually result in increased risk in the U.S. 25 population. 26 In a more recent study, Lamm et al. (2006) reported additional analyses of the 27 relationship between cancer risks and drinking water arsenic in the same Taiwanese population. 28 In this analysis, they divided the epidemiological data according to six "township" designations 29 provided by the original Chinese investigators (townships 0, 2, 3, 4, 5, and 6).1 They stratified 30 the data into two groups: townships that (by their characterization) "showed a significant dose- 31 response relationship" with arsenic (2, 4, 6) and townships "that did not" (0, 3, and 5). They 1 Each township included subsets of the 42 "villages" used as the basic units of analysis in the current assessment. F-1 DRAFT—DO NOT CITE OR QUOTE ------- 1 then applied linear regression to characterize the relationship between combined bladder and 2 lung cancer standardized mortality ratios (SMRs) and arsenic exposures in the Taiwanese 3 villages. They found that (1) dummy variables related to township were significant (along with 4 arsenic well concentration) when all the townships were included in the analysis, and (2) the 5 dose-response parameter for arsenic exposure became insignificant for arsenic well 6 concentrations less than 151 ppb when only townships 2, 4, and 6 were included in the 7 regression. Based on these results, they concluded that location (township) was an important 8 explanatory variable for cancer risks and that 151 ppb represented a "threshold" well arsenic 9 concentration below which no exposure-response relationship for arsenic could be detected. F.2. LIMITATIONS OF THE RECENT STUDIES 10 The studies discussed above all have significant limitations, relating both to the methods 11 used to select or stratify data for the risk assessment and to the methods used in analyzing 12 exposure-response data. In the first place, it is important to recognize the complexity and 13 limitations of the data. Cancer mortality and person-years at risk observations are provided for a 14 large number (559) of relatively small age- and village-stratified populations (median person- 15 years at risk ~ 340 for both males and females). Most population groups have zero cancer 16 deaths, and the data are very "noisy." Cancer mortality is strongly age-dependent, and 17 simultaneously evaluating the age-and dose-dependence of cancer mortality based on a data set 18 in which cancer deaths are "rare events" requires appropriately structured models. All of these 19 features of the data drove the selection of the Poisson regression methods described in Section 5, 20 and the use of simpler models (linear regression, for example) can (and did) lead to misleading 21 results. 22 With regard to the Lamm et al. (2003) paper, it is likely that the use of linear regression 23 and the failure to account for the age-dependency of bladder cancer risks combined to make it 24 impossible to detect a significant exposure-response relationship in villages with water arsenic 25 levels less than 400 ppb. In addition, it should be noted that Lamm et al. (2003) did not have 26 data regarding the actual sources of drinking water in the various villages; instead they relied on 27 the arsenic concentration to assess the degree of dependency of specific villages on artesian 28 (generally high-arsenic) versus shallow (low-arsenic) wells. When defined in this circular 29 fashion, it is inevitable that including the degree of "artesian well dependence" in a multiple 30 regression along with arsenic concentration would deprive the latter variable of much of its 31 explanatory power and statistical significance. Finally, the rationale for excluding valid data on 32 southwestern or all-Taiwan reference populations from the analysis is highly questionable, and 33 again lowers the likelihood of detecting significant exposure-response relationships. 34 The major limitation of Kayajanian's (2003) analysis of the Taiwanese data is that it 35 breaks the data into strata that are too small to be used to calculate reliable mortality risks, and 36 that it is very sensitive to the specific way that the data are stratified. The relatively high cancer F-2 DRAFT—DO NOT CITE OR QUOTE ------- 1 mortality risks seen in the low-dose strata are associated with a small number of villages that 2 happen to have a (relatively) large number of deaths. The observed trend in cancer mortality 3 versus arsenic dose would be very different if only few cancer deaths were misclassified, or if 4 the pattern of cancer deaths had been slightly different by chance. Again, failure to use a model 5 that adequately addresses the distribution of cancer deaths as rare events (or that incorporates 6 age dependence) resulted in results that are misleading. 7 Lamm et al.'s (2006) failure to find a significant exposure-response relationship in 8 villages with arsenic water concentrations below 151 ppb can also be explained by (1) the use of 9 linear regression without age-adjustment and (2) the omission of data from three of the six 10 townships from some of the regressions. Lamm et al. (2006) did not explain the specific criteria 11 for determining if a township "showed a dose-response relationship," but based on the 12 description of their methods provided in the article, it may be assumed that they used linear 13 regression to characterize the relationship between SMRs and arsenic exposure in each village in 14 the various townships. Given the small number of villages in each township, this approach and 15 the rationale for leaving townships 0, 3, and 5 out of the analysis appear arbitrary and 16 unjustified. In the following sections, we present alternative analyses that further investigate 17 the nature of arsenic exposure-response relationships in the various townships and in villages 18 with low arsenic drinking water concentrations. F.3. CALCULATIONS OF RISKS FOR TOWNSHIP GROUPS 19 To address the issues raised by Lamm et al. (2003, 2006), EPA compared the patterns of 20 cancer risks for subjects in the two groups of townships (0, 3, and 5 vs. 2, 4, and 6) to see 21 whether there were any differences. As noted above, it is not believed that Lamm et al.'s 22 approach to omitting townships because they lack an internal dose-response relationship is valid, 23 so EPA did not do so. 24 First, to get a rough idea of the patterns in cancer incidence versus exposure, the crude 25 cancer risks (population-weighted deaths per person-year for all age groups) and population- 26 weighted average arsenic exposure concentrations were calculated for each of the six villages. 27 The results are shown in Figure F-l. This figure simply illustrates that, even without age- 28 adjustments, arsenic dose-response relationships across the villages are quite robust. F-3 DRAFT—DO NOT CITE OR QUOTE ------- 2.E-03 "re £ 2.E-03 "re -o r-2 1.E-03 o> c o = 5.E-04 re O -*- Villages 0,3,5 -•-Villages 2,4,6 A Reference Population O.E+00 t 0 100 200 300 400 500 600 Population-Weighted Drinking Water As, micrograms/L 1 2 3 4 5 6 7 8 9 10 1 1 12 13 14 15 16 17 18 19 20 21 22 Figure F-l. Lifetime crude total cancer risk (male + female) for the low- and high-exposure villages For both sets of villages (low- and high-exposure), crude cancer risks increase with average arsenic drinking water concentration. Age distributions were very similar in all cohorts, so the lack of age-adjustment did not seriously bias the results. While total cancer risks are dominated by male lung cancer, the other endpoints showed generally the same pattern. This finding suggests that the positive exposure-response relationship for arsenic is not being seriously confounded by a "village effect." Given the small populations, populations at risk, and numbers of cancer deaths in the individual villages, it is not clear that analyzing exposure- response relationships within these villages (as defined by Lamm et al.) is justified. Exposure-response relationships in the various townships were also investigated using a variant of the multiple regression method applied by Lamm et al. (2006). In this analysis, however, the non-linear relationship between cancer risk and age was explicitly recognized, and the analysis was conducted for township both "with" and "without" significant exposure- response relationships by Lamm et al.'s definition. First, male and female combined cancer mortality risks (bladder + lung) were regressed against the same non-linear age dependency incorporated into the Poisson model shown in Equation 5-2. That is, the following equation was fit to both the male and female cancer data from the various age groups in the low- and high- exposure villages: risk (age) = exp(ai + a2 x age + age2) Then, the residuals from these regressions (the cancer risks with the effect of age removed) were regressed against estimated arsenic dose levels. The dose levels were calculated F-4 DRAFT—DO NOT CITE OR QUOTE ------- 1 assuming a nonwater arsenic intake of 10 |J,g/day for exposed and reference populations, which 2 is consistent with the assumptions outlined in Section 5.3.5. The regressions were population- 3 (person-year) weighted, in effect giving a linear regression of age-adjusted cancer risks versus 4 arsenic dose. The results are shown in Table F-l. Table F-l. Coefficients from linear regressions of age-adjusted cancer risk versus arsenic doses for townships identified by Lamm et al. (2006) Township Numbers Reference Population" Male arsenic dose coefficient (p-value) Female arsenic dose (p-value) All Townships Included 0.035 (0.043) 0.12 (0.0002) Excluded 0.032 (0.068) 0.12 (0.0004) Townships 2, 4, and 6 Included 0.092 (0.0002) 0.11 (0.0001) Excluded 0.091 (0.001) 0.12 (0.0001) Townships 0, 3, and 5 Included -0.0093 (0.787) 0.14(0.015) Excluded -0.002 (0.487) 0.13 (0.026) aSouthwest Taiwan. 5 The estimated dose coefficients for age-adjusted women's cancer risk (the linear "slope" 6 of the relationship between cancer mortality, with the effect of age removed, and arsenic dose2) 7 are positive and statistically significant for all combinations of townships. Coefficients for male 8 age-adjusted cancer risk are positive and significant when all townships are included (although 9 marginally significant when the reference population is excluded). Similarly, age-adjusted male 10 cancer risk coefficients are positive and highly significant for townships 2, 4, and 6, with or 11 without the reference population. In contrast, the arsenic dose-response coefficients for male 12 age-adjusted cancer risks are negative, but very small and not significant, for townships 0, 3, and 13 5. 14 This analysis illustrates that, even using the less-desirable linear regression approach, 15 when the cancer risk for the genders separated, and with proper age adjustment, female arsenic 16 dose-response relationships are robust and significant for both village groups. For males, the 17 arsenic dose-response relationships are significant when a reference population is included, 18 except for townships 0, 3, and 5. As noted above, the rationale for analyzing groups of 19 townships separately is questionable, as is the omission of a reference population. The results 20 showing apparently insignificant associations between male cancer risks and arsenic exposure This approach is not particularly desirable from the standpoint of finding the best fit to the data because it restricts the effect of arsenic on cancer risk to being linear, and assumes that regression residuals are normally distributed, which is unlikely to be true. This approach has been used to illustrate that even using simple models, positive dose- response relationships can be detected in the data. Due to the different form of this model, the slope coefficients derived in this section are also not comparable to those shown in Tables 5-3 and 5-4. F-5 DRAFT—DO NOT CITE OR QUOTE ------- 1 more than anything reflect the limitations of this less-than-optimal approach to risk modeling for 2 these data. F.4. CALCULATION OF ARSENIC-RELATED CANCER RISKS FOR LOW- EXPOSURE VILLAGES 3 Rather than stratify the Taiwanese population by township, a better way to test the 4 significance of exposure-response relationships at low doses is to simply restrict the analysis to 5 the villages with low arsenic water concentrations, but use the appropriate Poisson regression 6 methodology. In the analysis summarized in Table F-2, the Poisson model shown in Equation 5- 7 2 was fit to data from the approximately one-half of subject groups with median arsenic drinking 8 water concentrations less than 150 ppb. Lamm et al. (2006) considered this concentration to be a 9 natural breakpoint because the median arsenic concentrations in the Wu et al. (1989) and Chen et 10 al. (1992) population cluster into two groups, one group with 10-126 ppb and the other with 11 256-934 ppb. Arsenic "b" coefficients (the dose coefficients in the Poisson model) were 12 estimated separately for lung and bladder cancer and for both endpoints combined, for men and 13 women. Table F-2. Arsenic dose coefficients for study populations with median well water arsenic concentrations less than 127 ppb Endpoint Male lung Male bladder Male combined Female lung Female bladder Female combined Arsenic "b" Coefficient (95% UCL, LCL) 85.7(13.1, 172.1) 586 (335, 877) 160 (83.4, 247) 615 (412, 836) 2639 (2021, 3307) 924(721, 1139) 14 15 For all of the endpoints, the arsenic dose coefficients are positive with lower confidence 16 limits that are also positive.3 This finding indicates that for population groups with water arsenic 17 concentrations less than or equal to 126 ppb, the dose-response relationships are positive and 18 statistically significant. 19 On the whole, the analyses presented in this section provide support for statistically 20 significant dose-response relationships for arsenic-related cancer, even in the population groups 21 with relatively low exposures. When the data are artificially stratified, when no reference 22 population is included, and when inappropriate statistical models are employed, it is possible to As in Section 5.3.8, the upper and lower confidence limits were calculated using profile likelihood; similar results are obtained using bootstrap methods. F-6 DRAFT—DO NOT CITE OR QUOTE ------- 1 find insignificant or negative dose-response relationships for arsenic for some portions of the 2 data. When appropriate models are used, however, the Taiwanese data show robust and 3 significant positive associations between arsenic exposures and cancer risks for all of the 4 endpoints analyzed, even in low-exposure groups. No evidence was found that either 400 ppb or 5 150 ppb represent "threshold" arsenic concentrations in drinking water below which cancer risks 6 are not increased. Likewise, the analyses do not support the existence of a "village effect" 7 related to the degree of dependence on artesian versus shallow wells. F-7 DRAFT—DO NOT CITE OR QUOTE ------- |