DRAFT - DO NOT CITE OR QUOTE EPA/635/R-09/013
WWW.EPA.GOV/IRIS
&EPA
TOXICOLOGICAL REVIEW
OF
METHANOL
(CAS No. 67-56-1)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
October 2009
NOTICE
This document is an Interagency Science Consultation 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.
1 U.S. Environmental Protection Agency
2 Washington, DC
1 -1 DRAFT-DO NOT CITE OR QUOTE
-------
DISCLAIMER
1 This document is a preliminary draft for review purposes only. This information is
2 distributed solely for the purpose of pre-dissemination peer review under applicable information
3 quality guidelines. It has not been formally disseminated by EPA. It does not represent and
4 should not be construed to represent any Agency determination or policy. Mention of trade
5 names or commercial products does not constitute endorsement or recommendation for use.
1 -2 DRAFT-DO NOT CITE OR QUOTE
-------
CONTENTS TOXICOLOGICAL REVIEW OF METHANOL
(CAS NO. 67-56-1)
1 1. INTRODUCTION 1-26
2 2. CHEMICAL AND PHYSICAL INFORMATION 2-1
3 3. TOXICOKINETICS 3-1
4 3.1. Overview 3-1
5 3.2. Key Studies 3-5
6 3.3. Human Variability In Methanol Metabolism 3-10
7 3.4. Physiologically Based Toxicokinetic Models 3-15
8 3.4.1. Model Requirements for EPA Purposes 3-15
9 3.4. l.l.MOA and Selection of a Dose Metric 3-15
10 3.4.1.2. Criteria for the Development of Methanol PBPK Models 3-17
11 3.4.2. Methanol PBPK Models 3-18
12 3.4.2.1. Ward etal. (1997) 3-19
13 3.4.2.2. Bouchard etal. (2001) 3-19
14 3.4.2.3. Gentry et al. (2003, 2002) and Clewell et al. (2001) 3-20
15 3.4.3. Selected Modeling Approach 3-21
16 3.4.3.1. Available PK Data 3-22
17 3.4.3.2. Model Structure 3-23
18 3.4.3.3. Model Parameters 3-25
19 3.4.4. Mouse Model Calibration and Sensitivity Analysis 3-26
20 3.4.5. Rat Model Calibration 3-31
21 3.4.6. Human Model Calibration 3-35
22 3.4.6.1. Inhalation Route 3-35
23 3.4.6.2. Oral Route 3-41
24 3.4.7. Monkey PK Data and Analysis 3-42
25 3.4.7.l.PK Model Analysis for Monkeys 3-44
26 3.4.8. Summary and Conclusions 3-47
27 4. HAZARD IDENTIFICATION 4-1
28 4.1. Studies in Humans - Case Reports, Occupational and Controlled Studies 4-1
29 4.1.1. Case Reports 4-1
30 4.1.2. Occupational Studies 4-9
31 4.1.3. Controlled Studies 4-11
32 4.2. Acute, Subchronic and Chronic Studies and Cancer Bioassays in Animals—Oral and
33 Inhalation 4-13
34 4.2.1. Oral Studies 4-13
35 4.2.1.1. Acute Toxicity 4-13
36 4.2.1.2. Subchronic Toxicity 4-13
37 4.2.1.3. Chronic Toxicity 4-14
38 4.2.2. Inhalation Studies 4-20
39 4.2.2.1. Acute Toxicity 4-20
40 4.2.2.2. Subchronic Toxicity 4-21
41 4.2.2.3. Chronic Toxicity 4-24
42 4.3. Reproductive AND Developmental Studies-Oral and Inhalation 4-32
43 4.3.1. Oral Studies 4-32
44 4.3.2. Inhalation Studies 4-34
1 -3 DRAFT-DO NOT CITE OR QUOTE
-------
1 4.3.3. Other Reproductive and Developmental Toxicity Studies 4-46
2 4.4. Neurotoxicity 4-50
3 4.4.1. Oral Studies 4-51
4 4.4.2. Inhalation Studies 4-53
5 4.4.3. Studies Employing In Vitro, S.C. and IP. Exposures 4-60
6 4.5. Immunotoxicity 4-64
7 4.6. Mechanistic Data and Other Studies in Support of the MO A 4-69
8 4.6.1. Role of Methanol and Metabolites in the Developmental Toxicity of Methanol ..4-69
9 4.6.2. Role of Folate Deficiency in the Developmental Toxicity of Methanol 4-74
10 4.6.3. Methanol-Induced Formation of Free Radicals, Lipid Peroxidation, and Protein
11 Modifications 4-75
12 4.6.4. Exogenous Formate Dehydrogenase as a Means of Detoxifying the Formic Acid that
13 Results from Methanol Exposure 4-78
14 4.6.5. Mechanistic Data Related to the Potential Carcinogenicity of Methanol 4-78
15 4.6.5.1. Genotoxicity 4-78
16 4.6.5.2. Lymphoma Responses Reported in ERF Life span Bioassays of Compounds
17 Related to Methanol, Including an Analogue (Ethanol), Precursors (Aspartame
18 and Methyl Tertiary Butyl Ether), and a Metabolite (Formaldehyde) 4-81
19 4.6.5.2.1. Ethanol 4-83
20 4.6.5.2.2. Aspartame 4-84
21 4.6.5.2.3. MTBE 4-89
22 4.6.5.2.4. Formaldehyde 4-90
23 4.7. Synthesis of Major Noncancer Effects 4-92
24 4.7.1. Summary of Key Studies in Methanol Toxicity 4-92
25 4.7.1.1. Oral 4-95
26 4.7.1.2. Inhalation 4-96
27 4.8. NONCANCER MOA Information 4-99
28 4.9. Evaluation of Carcinogenicity 4-102
29 4.9.1. Summary of Overall Weight-of-Evidence 4-102
30 4.9.2. Synthesis of Human, Animal, and Other Supporting Evidence 4-103
31 4.9.3. MOA Information 4-109
32 4.10. Susceptible Populations and Life Stages 4-110
33 4.10.1. Possible Childhood Susceptibility 4-110
34 4.10.2. Possible Gender Differences 4-111
35 4.10.3. Genetic Susceptibility 4-112
36 5. DOSE-RESPONSE ASSESSMENT AND CHARACTERIZATION 5-1
37 5.1. Inhalation RfC for Effects Other than Cancer 5-1
38 5.1.1. Choice of Principal Study and Critical Effect(s) 5-1
39 5.1.1.1. Key Inhalation Studies 5-1
40 5.1.1.2. Selection of Critical Effect(s) 5-2
41 5.1.2. Methods of Analysis for the POD—Application of PBPK and BMD Models 5-6
42 5.1.2.1. Applicati on of the BMD/BMDL Approach 5-7
43 5.1.2.2. BMD Approach Applied to Brain Weight Data in Rats 5-9
44 5.1.3. RfC Derivation - Including Application of Uncertainty Factors 5-13
45 5.1.3.1. Comparison Between Endpoints andBMDL Modeling Approaches 5-13
46 5.1.3.2. Applicati on of UFs 5-15
47 5.1.3.2.1. Interindividual Variation UFH 5-15
48 5.1.3.2.2. Animal-to-Human Extrapolation UFA 5-17
1 -4 DRAFT-DO NOT CITE OR QUOTE
-------
1 5.1.3.2.3. Database UFD 5-17
2 5.1.3.2.4. Extrapolation from Subchronic to Chronic and LOAEL-to NOAEL
3 Extrapolation UFs 5-18
4 5.1.4. Previous RfC Assessment 5-18
5 5.2. Oral RfD 5-18
6 5.2.1. Choice of Principal Study and Critical Effect-with Rationale and Justification... 5-19
7 5.2.1.1. Expansion of the Oral Database by Route-to-Route Extrapolation 5-20
8 5.2.2. RfD Derivation-Including Application of UFs 5-21
9 5.2.1.2. Consideration of Inhalation Data 5-21
10 5.2.1.3. Selection of Critical Effect(s) from Inhalation Data 5-21
11 5.2.1.4. Selection of the POD 5-22
12 5.2.2. RfD Derivation-Application of UFs 5-22
13 5.2.3. Previous RfD Assessment 5-22
14 5.3. Uncertanties in the Inhalation RfC and Oral RfD 5-23
15 5.3.1. Choice of Endpoint 5-24
16 5.3.2. Choice of Dose Metric 5-25
17 5.3.3. Choice of Model for BMDL Derivations 5-26
18 5.3.4. Choice of Animal-to-Human Extrapolation Method 5-27
19 5.3.5. Route-to-Route Extrapolation 5-27
20 5.3.6. Statistical Uncertainty at the POD 5-28
21 5.3.7. Choice of Bioassay 5-28
22 5.3.8. Choice of Species/Gender 5-28
23 5.3.9. Human Population Variability 5-28
24 5.4. 5.4. Cancer Assessment 5-29
25 5.4.1. Oral Exposure 5-29
26 5.4.1.1. Choice of Study/Data—with Rationale and Justification 5-29
27 5.4.1.2. Dose-Response Data 5-29
28 5.4.1.3. Dose Adjustments and Extrapolation Method 5-30
29 5.4.1.4. Oral Slope Factor 5-33
30 5.4.2. Inhalation Exposure 5-33
31 5.4.2.1. Choice of Study/Data-with Rationale and Justification 5-33
32 5.4.2.2. Dose-Response Data 5-34
33 5.4.2.3. Dose Adjustments and Extrapolation Method 5-35
34 5.4.2.4. IUR 5-36
35 5.4.3. Uncertainties in Cancer Risk Assessment 5-36
36 5.4.3.1. Quality of Studies that are the Basis for the PODs 5-37
37 5.4.3.2. Interpretation of Results of the Studies that are the Basis for the PODs 5-40
38 5.4.3.3. Consistency across Chronic Bioassays for Methanol 5-43
39 5.4.3.4. Choice of Endpoint for POD Derivation 5-45
40 5.4.3.5. Choice of Species/Gender 5-45
41 5.4.3.6. Choice of Model for POD Derivation 5-45
42 5.4.3.7. Choice of Dose Metric 5-46
43 5.4.3.8. Choice of Animal-to-Human Extrapolation Method 5-47
44 5.4.3.9. Human Relevance of Cancer Responses Observed in Rats and Mice 5-48
45 6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
46 RESPONSE 6-1
47 6.1. Human Hazard Potential 6-1
48 6.2. Dose Response 6-3
1 -5 DRAFT-DO NOT CITE OR QUOTE
-------
1 6.2.1. Noncancer/Inhalation 6-3
2 6.2.2. Noncancer/Oral 6-5
3 6.2.3. Cancer/Oral and Inhalation 6-6
4 7. REFERENCES 7-1
5 APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC COMMENTS
6 AND DISPOSITION A-l
7 APPENDIX B. DEVELOPMENT, CALIBRATION AND APPLICATION OF A METHANOL
8 PBPK MODEL B-l
9 APPENDIX C. RFC DERIVATION OPTIONS C-l
10 APPENDIX D. RFC DERIVATION - COMPARISON OF DOSE METRICS D-l
11 APPENDIX E. EVALUATION OF THE CANCER POTENCY OF METHANOL E-1
1 -6 DRAFT-DO NOT CITE OR QUOTE
-------
LIST OF TABLES
1 Table 2-1. Relevant physical and chemical properties of methanol 2-1
2 Table 3-1. Background blood methanol and formate levels in humans 3-11
3 Table 3-2. Human blood methanol and formate levels following methanol exposure 3-12
4 Table 3-3. Monkey blood methanol and formate levels following methanol exposure 3-13
5 Table 3-4. Mouse blood methanol and formate levels following methanol exposure 3-13
6 Table 3-5. Rat blood methanol and formate levels following methanol exposure 3-14
7 Table 3-6. Plasma formate concentrations in monkeys 3-15
8 Table 3-7. Serum folate concentrations in monkeys 3-15
9 Table 3-8. Routes of exposure optimized in models - optimized against blood concentration data3-20
10 Table 3-9. Key methanol kinetic studies for model validation 3-22
11 Table 3-10. Parameters used in the mouse, rat and human PBPK models 3-25
12 Table 3-11. Primate Kms reported in the literature 3-38
13 Table 3-12. Parameter estimate results obtained using acslXtreme to fit all human data using
14 either saturable or first-order metabolism 3-38
15 Table 3-13. Comparison of LLFs for Michaelis-Menten and first-order metabolism 3-40
16 Table 3-14. Monkey group exposure characteristics 3-47
17 Table 4-1. Mortality rate for subjects exposed to methanol-tainted whiskey in relation to their
18 level of acidosisaa 4-2
19 Table 4-2. Incidence of carcinogenic responses in Sprague-Dawley rats exposed to methanol in
20 drinking water for up to 2 years 4-18
21 Table 4-3. Incidence of malignant lymphoma responses in Swiss mice exposed to methanol in
22 drinking water for life 4-20
23 Table 4-4. Histopathological changes in tissues of B6C3F1 mice exposed to methanol via
24 inhalation for 18 months 4-28
25 Table 4-5. Histopathological changes in lung and adrenal tissues of F344 rats exposed to
26 methanol via inhalation for 24 months 4-31
27 Table 4-6. Reproductive and developmental toxicity in pregnant Sprague-Dawley rats exposed to
28 methanol via inhalation during gestation 4-36
29 Table 4-7. Reproductive parameters in Sprague-Dawley dams exposed to methanol during
30 pregnancy then allowed to deliver their pups 4-37
31 Table 4-8. Developmental effects in mice after methanol inhalation 4-39
32 Table 4-9. Benchmark doses at two added risk levels 4-40
33 Table 4-10. Reproductive parameters in monkeys exposed via inhalation to methanol during
34 prebreeding, breeding, and pregnancy 4-42
35 Table 4-11. Mean serum levels of testosterone, luteinizing hormone, and corticosterone (± S.D.)
36 in male Sprague-Dawley rats after inhalation of methanol, ethanol, n-propanol or n-
37 butanol at threshold limit values 4-43
38 Table 4-12. Maternal and litter parameters when pregnant female C57BL/6J mice were injected
39 i.p. with methanol 4-46
40 Table 4-13. Reported thresholds concentrations (and author-estimated ranges) for the onset of
41 embryotoxic effects when rat and mouse conceptuses were incubated in vitro with
42 methanol, formaldehyde, and formate 4-49
43 Table 4-14. Brain weights of rats exposed to methanol vapors during gestation and lactation 4-58
44 Table 4-15. Effect of methanol on Wistar rat acetylcholinesterase activities 4-61
45 Table 4-16. Effect of methanol on neutrophil functions in in vitro and in vivo studies in male
46 Wistar rats 4-65
1 -7 DRAFT-DO NOT CITE OR QUOTE
-------
1 Table 4-17. Effect of intraperitoneally injected methanol on total and differential leukocyte
2 counts and neutrophil function tests in male Wistar rats 4-66
3 Table 4-18. Effect of methanol exposure on animal weight/organ weight ratios and on cell counts
4 in primary and secondary lymphoid organs of male Wistar rats 4-68
5 Table 4-19. The effect of methanol on serum cytokine levels in male Wistar rats 4-69
6 Table 4-20. Developmental outcome on GD10 following a 6-hour 10,000 ppm (13,104 mg/m3)
7 methanol inhalation by CD-mice or formate gavage (750 mg/kg) on GD8 4-70
8 Table 4-21. Summary of ontogeny of relevant enzymes in CD-I mice and humans 4-71
9 Table 4-22. Dysmorphogenic effect of methanol and formate in neurulating CD-I mouse
10 embryos in culture (GD8) 4-72
11 Table 4-23. Time-dependent effects of methanol administration on serum liver and kidney
12 function, serum ALT, AST, BUN, and creatinine in control and experimental groups
13 of male Wistar rats 4-77
14 Table 4-24. Effect of methanol administration on male Wistar rats on malondialdehyde
15 concentration in the lymphoid organs of experimental and control groups and the
16 effect of methanol on antioxidants in spleen 4-77
17 Table 4-25. Summary of genotoxicity studies of methanol 4-80
18 Table 4-26. Incidence of carcinogenic responses in Sprague-Dawley rats exposed to ethanol in
19 drinking water for up to 2 years 4-84
20 Table 4-27. Incidence of lymphomas and leukemias in Sprague-Dawley rats exposed to
21 aspartame via the diet 4-85
22 Table 4-28. Incidence of combined dysplastic hyperplasias, papillomas and carcinomas of the
23 pelvis and ureter and of malignant schwannomas in peripheral nerve in Sprague-
24 Dawley rats exposed to aspartame via the diet 4-86
25 Table 4-29. Incidence of tumors in Sprague-Dawley rats exposed to aspartame from GD12 to
26 natural death 4-87
27 Table 4-30. Comparison of the incidence of combined lymphomas and leukemias in female
28 Sprague-Dawley rats exposed to aspartame in feed for a lifetime, either pre- and
29 postnatally or postnatally only 4-88
30 Table 4-31. Incidence of Ley dig cell testicular tumors and combined lymphomas and leukemias
31 in Sprague-Dawley rats exposed to MTBE via gavage for 104 weeks 4-90
32 Table 4-32. Incidence of hemolymphoreticular neoplasms on Sprague-Dawley rats exposed to
33 formaldehyde in drinking water for 104 weeks 4-91
34 Table 4-33. Incidence of leukemias in breeder and offspring Sprague-Dawley rats exposed to
35 formaldehyde in drinking water for 104 weeks (TestBT 7005) 4-92
36 Table 4-34. Summary of studies of methanol toxicity in experimental animals (oral) 4-92
37 Table 4-35. Summary of studies of methanol toxicity in experimental animals (inhalation
38 exposure) 4-93
39 Table 5-1. Summary of studies considered most appropriate for use in derivation of an RfC.... 5-5
40 Table 5-2. The EPA PBPK model estimates of methanol blood levels (AUC) in rats following
41 inhalation exposures 5-10
42 Table 5-3. Comparison of benchmark dose modeling results for decreased brain weight in male
43 rats at 6 weeks of age using modeled AUC of methanol as a dose metric 5-11
44 Table 5-4. Summary of PODs for critical endpoints, application of UFs and conversion to HEC
45 values using BMD and PBPK modeling 5-14
46 Table 5-5. Summary of uncertainties in methanol noncancer risk assessment 5-23
47 Table 5-6. Incidence data for lymphoma, lympho-immunoblastic, and all lymphomas in male and
48 female Sprague-Dawley rats 5-30
49 Table 5-7. BMD results and oral CSF using all lymphoma in male rats 5-33
1 -8 DRAFT-DO NOT CITE OR QUOTE
-------
1 Table 5-8. Incidence data for tumor responses in male and female F344 rats 5-34
2 Table 5-9. BMD results and IUR using pheochromocytoma in female rats 5-36
3 Table 5-10. Summary of uncertainty in the methanol cancer risk assessment 5-37
4 Table B-l. Parameters used in the mouse and human PBPK models B-5
5 Table B-2. Primate kms reported in the literature B-29
6 Table B-3. Parameter estimate results obtained using acslXtreme to fit all human data using
7 either saturable or first-order metabolism B-31
8 Table B-4. Comparison of LLF for Michaelis-Menten and first-order metabolism B-32
9 Table B-5. PBPK model predicted Cmax and 24-hour AUC for mice and humans exposed to
10 MeOH B-35
11 Table B-6. Mouse total MeOH metabolic clearance/metabolites produced following inhalation
12 exposures B-90
13 Table B-7. Human total MeOH metabolic clearance/metabolites produced from inhalation
14 exposures B-91
15 Table B-8. Human total MeOH metabolic clearance/metabolites produced following oral
16 exposures B-91
17 Table B-9. Repeated daily oral dosing of humans with MeOH B-93
18 Table C-l. EPA's PBPK model estimates of methanol blood levels (AUC) in rats following
19 inhalation exposures C-3
20 Table C-2. Comparison of BMDiso results for decreased brain weight in male rats at 6 weeks of
21 age using modeled AUC of methanol as a dose metric C-4
22 Table C-3. Comparison of BMD0s results for decreased brain weight in male rats at 6 weeks of
23 age using modeled AUC of methanol as a dose metric C-9
24 Table C-4. EPA's PBPK model estimates of methanol blood levels (Cmax) in rats following
25 inhalation exposures C-l3
26 Table C-5. Comparison of BMDiso results for decreased brain weight in male rats at 8 weeks of
27 age using modeled Cmax of methanol as a dose metric C-14
28 Table C-6. Comparison of BMDos modeling results for decreased brain weight in male rats at 8
29 weeks of age using modeled Cmax of methanol as a common dose metric C-20
30 Table C-7. EPA's PBPK model estimates of methanol blood levels (Cmax) in mice following
31 inhalation exposures C-24
32 Table C-8. Comparison of BMD modeling results for cervical rib incidence in mice using
33 modeled Cmax of methanol as a common dose metric C-25
34 Table C-9. Comparison of BMD modeling results for cervical rib incidence in mice using
35 modeled Cmax of methanol as a common dose metric C-33
36 Table C-10. Comparison of BMD modeling results for VDR in female monkeys using AUC
37 blood methanol as the dose metric C-42
38 Table E-l. Calculation of mg/kg-day doses E-12
39 Table E-2. Incidence for neoplasms considered for dose-response modeling E-12
40 Table E-3. Results from multistage (1°) quantal modeling rat data using mg/kg-day exposures
41 and default HED derivation method E-13
42 Table E-4. Results from time-to-tumor modeling data using mg/kg-day exposures and default
43 HED derivation method E-13
44 Table E-5. PBPK model estimated dose-metrics for doses E-14
45 Table E-6a. Results from time-to-tumor modeling of data using PBPK dose metrics E-14
46 Table E-6b. HEDs from time-to-tumor modeling of data using PBPK dose metrics E-14
47 Table E-7. Results of Multistage (1°) quantal modeling of data using PBPK dose metrics E-15
48 Table E-8. Application of human PBPK model to derive HEDs from results of multistage (1°)
49 quantal modeling of data using PBPK dose metrics E-15
1 -9 DRAFT-DO NOT CITE OR QUOTE
-------
1 Table E-9. Incidence for neoplasms considered for dose-response modeling E-16
2 Table E-10. PBPK dose metrics for doses E-16
3 Table E-l 1. Benchmark results from multistage quantal dose-response modeling data using
4 PBPK dose-metrics E-17
5 Table E-12. Application of human PBPK model to derive HECs from BMDLio estimates in
6 Table E-ll using multistage quantal modeling E-17
7 Table E-13. Incidence for malignant lymphoma (Apaja, 1980) E-17
8 Table E-14. PBPK dose metrics for doses in Apaja (1980) E-18
9 Table E-l5. Benchmark results from Multistage-cancer dose-response modeling data for
10 malignant lymphoma in Swiss Webster mice (Apaja, 1980) using PBPK dose-metricsE-18
11 Table E-16. Application of human PBPK model to derive HEDs from BMDLio estimates of
12 Table E-15, Multistage (1°) modeling of malignant lymphoma in Swiss mice (Apaja,
13 1980) using PBPK dose metrics E-19
14 Table E-17. Benchmark results for all tumor types using BMDS 2.1 multistage "background
15 dose" models and PBPK dose-metrics E-19
1-10 DRAFT-DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
1 Figure 3-1. Methanol metabolism and key metabolic enzymes in primates and rodents 3-3
2 Figure 3-2. Folate-dependent formate metabolism. Tetrahydrofolate (THF)-mediated one carbon
3 metabolism is required for the synthesis of purines, thymidylate, and methionine.... 3-4
4 Figure 3-3. Plot of fetal (amniotic) versus maternal methanol concentrations in GD20 rats 3-7
5 Figure 3-4. Schematic of the PBPK model used to describe the inhalation, oral, and i.v. route
6 pharmacokinetics of methanol 3-24
7 Figure 3-5. Model fits to data sets from GD6, GD7, and GD10 mice for 6- to 7-hour inhalation
8 exposures to 1,000-15,000 ppm methanol 3-27
9 Figure 3-6. Simulation of inhalation exposures to methanol in NP mice from Perkins et al.
10 (1995a) (8-hour exposures) and GD8 mice from Dorman et al. (1995)(6-hour
11 exposures) 3-28
12 Figure 3-7. Conceptus versus maternal blood AUC values for rats and mice plotted (A) on a log-
13 linear scale, as in Figure 8 of Ward et al. (1997), and (B) on a linear-linear scale... 3-30
14 Figure 3-8. NP rat i.v. route methanol blood kinetics 3-33
15 Figure 3-9. Model fits to data sets from inhalation exposures to 200 (triangles), 1,200
16 (diamonds), or 2,000 (squares) ppm methanol in male F-344 rats 3-33
17 Figure 3-10. Model fits to datasets from oral exposures to 100 and 2,500 mg/kg methanol in
18 female Sprague-Dawley rats 3-35
19 Figure 3-11. Urinary methanol elimination concentration (upper panel) and cumulative amount
20 (lower panel) following inhalation exposures to methanol in human volunteers 3-37
21 Figure 3-12. Data showing the visual quality of the fit using optimized first-order or Michaelis-
22 Menten kinetics to describe the metabolism of methanol in humans 3-39
23 Figure 3-13. Inhalation exposures to methanol in human volunteers 3-41
24 Figure 3-14. Blood methanol concentration data from NP and pregnant monkeys 3-44
25 Figure 3-15. Chamber concentration profiles for monkey methanol exposures 3-46
26 Figure 4-1. Hemolymphoreticular neoplasms in male and female Sprague-Dawley rats in
27 formaldehyde 4-110
28 Figure 5-1. Hill model BMD plot of decreased brain weight in male rats at 6 weeks age using
29 modeled AUC of methanol in blood as the dose metric, 1 control mean S.D 5-12
30 Figure 5-2. PODs (in mg/m3) for selected endpoints with corresponding applied UFs 5-15
31 Figure 5-3. All lymphomas versus methanol metabolized (mg/day) for female and male rats. 5-31
32 Figure 5-4. All lymphomas versus Cmax (mg/L) for female and male rats 5-31
33 Figure 5-5. All lymphomas versus AUC (hr x mg/L) for male and female rats 5-31
34 Figure 5-6. Lympho-immunoblastic Lymphoma minus "lung-only" response in rats of Soffritti et
35 al. (2002a) methanol study versus methanol metabolized (mg/day) 5-42
36 Figure 5-7. Total amount metabolized per day (after periodicity is reached) in a 420 g rat 5-47
37 Figure B-l. Schematic of the PBPK model used to describe the inhalation, oral, and i.v. route
38 pharmacokinetics of MeOH B-3
39 Figure B-2. Model fits to data sets from GD6, GD7, and GD10 mice for 7-hour inhalation
40 exposures to 1,000-15,000 ppm MeOH B-7
41 Figure B-3. Simulation of inhalation exposures to MeOH in NP mice from Perkins et al. (1995a)
42 (8-hour exposures) and Dorman et al. (1995), (6-hour exposures) B-8
43 Figure B-4. Oral exposures to MeOH in pregnant mice on GD8 (Dorman et al., 1995) or NP and
44 GDIS B-10
45 Figure B-5. Mouse intravenous route MeOH blood kinetics B-ll
46 Figure B-6. Mouse model inhalation route sensitivity coefficients for metabolic parameters.B-l5
1-11 DRAFT-DO NOT CITE OR QUOTE
-------
1 Figure B-7. Mouse model inhalation route sensitivity coefficients for flow rates (QCC: cardiac
2 output; QPC: alveolar ventilation), and partitioning to the body (PR) compartment are
3 reported for blood MeOH AUC B-15
4 Figure B-8. Mouse model sensitivity coefficients for oral exposures to MeOH B-16
5 Figure B-9. Mouse daily drinking water ingestion pattern (A) and resulting predicted blood
6 concentration for a 6 d/wk exposure (B) B-17
7 Figure B-10. NP rat i.v.-route methanol blood kinetics B-19
8 Figure B-l 1. Model fits to data sets from inhalation exposures to 200 (triangles), 1,200
9 (diamonds), or 2,000 (squares) ppm MeOH in male F-344 rats B-20
10 Figure B-12. Model fits to data sets from oral exposures to 100 (squares) or 2,500 (diamonds)
11 mg/kg MeOH in female Sprague-Dawley rat B-22
12 Figure B-13. Simulated Sprague-Dawley rat inhalation exposures to 500, 1,000, or 2,000 ppm
13 MeOH B-24
14 Figure B-14. Simulated rat oral exposures of Sprague-Dawley rats to 65.9, 624.2, or 2,177 mg
15 MeOH/kg/day B-25
16 Figure B-15. Rat daily drinking water ingestion pattern (A) and resulting predicted blood
17 concentration for a 2-day exposure (B) B-26
18 Figure B-16. Urinary MeOH elimination concentration (upper panel) and cumulative amount
19 (lower panel), following inhalation exposures to MeOH in human volunteers B-28
20 Figure B-17. Control (nonexposed) blood methanol concentrations B-30
21 Figure B-18. Data showing the visual quality of the fit using optimized first-order or Michaelis-
22 Menten kinetics to describe the metabolism of MeOH in humans B-31
23 Figure B-19. Inhalation exposures to MeOH in human volunteers B-33
24 Figure B-20. Predicted 24-hour AUC (upper left), Cmax (upper right), and amount metabolized
25 (lower) for MeOH inhalation exposures in the mouse (average over a 10-day exposure
26 at 7 hours/day) and humans (steady-state values for a continuous exposure) B-36
27 Figure B-21. Fit of the model to i.v. data using different clearance and uptake parameter
28 optimizations B-41
29 Figure B-22. Fit of the model to oral data using different clearance and uptake parameter
30 optimizations B-42
31 Figure B-23. Fit of the model to inhalation data using different clearance and uptake parameter
32 optimizations B-42
33 Figure B-24. Simulated human oral exposures to 0.1 mg MeOH/kg/-day comparing the first few
34 days for four exposure scenarios B-92
35 Figure C-l. Hill model, BMR of 1 Control Mean S.D. - Decreased Brain weight in male rats at 6
36 weeks age versus AUC, Fl Generation inhalational study C-7
37 Figure C-2. Hill model, BMR of 0.05 relative risk - Decreased Brain weight in male rats at 6
38 weeks age versus AUC, FI Generation inhalational study C-12
39 Figure C-3. Hill model, BMR of 1 Control Mean S.D. - Decreased Brain weight in male rats at 8
40 weeks age versus Cmax, Gestation only inhalational study C-18
41 Figure C-4. Exponential model, BMR of 0.05 relative risk - Decreased Brain weight in male
42 rats at 8 weeks age versus Cmax, Gestation only inhalational study C-23
43 Figure C-5. Nested Logistic Model, 0.1 Extra Risk - Incidence of Cervical Rib in Mice versus
44 Cmax Methanol, GD 6-15 inhalational study C-32
45 Figure C-6. Nested Logistic Model, 0.05 Extra Risk - Incidence of Cervical Rib in Mice versus
46 Cmax Methanol, GD 6-15 inhalational study C-39
47 Figure C-7. 3rd Degree Polynomial Model, BMR of 1 Control Mean S.D. - VDR in female
48 monkeys using AUC blood methanol as the dose metric C-44
1-12 DRAFT-DO NOT CITE OR QUOTE
-------
1 Figure D-l. Exposure-response data for methanol-induced CR plus SNR malformations in mice
2 at various concentration-time combinations D-2
3 Figure D-2. Internal dose-response relationships for methanol-induced CR plus SNR
4 malformations in mice at various concentration-time combinations for three dose
5 metrics D-3
6 Figure E-l. Female rat survival E-20
7 Figure E-2. Male rats survival E-20
8 Figure E-3. Female - Lymphomas lympho-immunoblastic - Multistage Weibull Model -
9 Approach 1 E-21
10 Figure E-4. Female - All lymphomas - Multistage Weibull Model - Approach 1 E-21
11 Figure E-5. Male - Hepatocellular carcinoma - Multistage Weibull Model - Approach 1 E-22
12 Figure E-6. Male - Lymphomas lympho-immunoblastic - Multistage Weibull Model -
13 Approach 1 E-22
14 Figure E-7. Male - All lymphomas - Multistage Weibull Model - Approach 1 E-23
15 Figure E-8. Female rats -All lymphomas time-to-tumor model fit and Kaplan Meier curves.E-23
16 Figure E-9. Male rats -All lymphomas time-to-tumor model fit and Kaplan Meier curves. ...E-24
17 Figure E-10. Male rats-All lymphomas; dose = amount metabolized (mg/day); 1° multistage
18 model E-24
19 Figure E-l 1. Female rat survival E-25
20 Figure E-12. Male rat survival E-25
21 Figure E-13. Female rats- pheochromocytomas; dose = amount metabolized (mg/d); 3°
22 multistage model E-26
23 Figure E-14. Plots for female (top; p= 0.55) and male (bottom; p= 0.21) mice - malignant
24 lymphoma; dose=amount metabolized (mg/d); 2° multistage model E-26
1-13 DRAFT-DO NOT CITE OR QUOTE
-------
LIST OF ACRONYMS
ACGIH American Conference of Governmental and Industrial Hygienists
ADH alcohol dehydrogenase
ADH1 alcohol dehydrogenase-1
ADH3 formaldehyde dehydrogenase-3
AIC Akaike Information Criterion
ALD aldehyde dehydrogenase
ALDH2 mitochondrial aldehyde dehydrogenase-2
ALT alanine aminotransferase
ANO VA analy si s of vari ance
AP alkaline phosphatase
AST aspartate aminotransferase
ATP adenosine triphosphate
ATSDR Agency for Toxic Substances and Disease Registry
AUC area under the curve, representing the cumulative product of time
and concentration for a substance in the blood
P-NAG N-acetyl-beta-D-glucosaminidase
BAA butoxyacetic acid
BAL butoxyacetaldehyde
BMC benchmark concentration
BMCL benchmark concentration, 95% lower bound
BMD benchmark dose(s)
BMDiso BMD for response one standard deviation from control mean
BMDL 95% lower bound confidence limit on BMD (benchmark dose)
BMDLiso BMDL for response one standard deviation from control mean
BMDS benchmark dose software
BMR benchmark response
BSO butathione sulfoximine
BUN blood urea nitrogen
BW, bw body weight
Ci pool one carbon pool
Cmax peak concentration of a substance in the blood during the
exposure period
C-section Cesarean section
CA chromosomal aberrations
CAR conditioned avoidance response
CASRN Chemical Abstracts Service Registry Number
CAT catalase
CERHR Center for the Evaluation of Risks to Human Reproduction
CH3OH methanol
CHL Chinese hamster lung (cells)
1-14
DRAFT-DO NOT CITE OR QUOTE
-------
CHR contact hypersensitivity response
CI confidence interval
Cls clearance rate
CNS central nervous system
CC>2 carbon dioxide
con-A concanavalin-A
CR crown-rump length
CSF Cancer slope factor
CT computed tomography
CYP450 cytochrome P450
d, 5, A delta, difference, change
D2 dopamine receptor
DA dopamine
DIPE diisopropylether
DMDC dimethyl dicarbonate
DNA deoxyribonucleic acid
DNT developmental neurotoxicity test(ing)
DOPAC dihydroxyphenyl acetic acid
DPC days past conception
DTH delayed-type hypersensitivity
EFSA European Food Safety Authority
EKG electrocardiogram
EO Executive Order
EPA U.S. Environmental Protection Agency
ERF European Ramazzini Foundation
EtOH ethanol
F fractional bioavailability
FQ parental generation
FI first generation
F2 second generation
F344 Fisher 344 rat strain
FAD folic acid deficient
FAS folic acid sufficient
FD formate dehydrogenase
FP folate paire
FR folate reduced
FRACIN fraction inhaled
FS folate sufficient
FSH follicular stimulating hormone
y-GT gamma glutamyl transferase
1-15
DRAFT-DO NOT CITE OR QUOTE
-------
g gravity
g, kg, mg, jig gram, kilogram, milligram, microgram
G6PD glucose-6-phosphate dehydrogenase
GAP43 growth-associated protein (neuronal growth cone)
GD gestation day
GFR glomerular filtration rate
GI gastrointestinal track
GLM generalized linear model
GLP good laboratory practice
GSH glutathione
HAP hazardous air pollutant
Hb hemoglobin
HCHO formaldehyde
HCOO formate
Hct hematocrit
HEC human equivalent concentration
HED human equivalent dose
HEI Health Effects Institute
HH hereditary hemochromatosis
5_HIAA 5-hydroxyindolacetic acid
HMGSH S-hydroxymethylglutathione
Hp haptoglobin
HPA hypothalamus-pituitary-adrenal (axis)
HPLC high-performance liquid chromatography
HSDB Hazardous Substances Databank
HSP70 biomarker of cellular stress
5-HT serotonin
IL interleukins
i.p. intraperitoneal
IPCS International Programme on Chemical Safety
IQ intelligence quotient
IRIS Integrated Risk Information System
IUR inhalation unit risk
i.v. intravenous
KI first order rate loss
K1C first order clearance of methanol from the blood to the bladder for
urinary elimination
KAI first order uptake from the intestine
KAS first order methanol oral absorption rate from stomach
KBL rate constant for urinary excretion from bladder
KIA first order uptake from intestine
1-16 DRAFT-DO NOT CITE OR QUOTE
-------
KLH keyhole limpet hemocyanin
KLL alternate first order rate constant
Km substrate concentration at half the enzyme maximum
velocity (Vmax)
Km2 Michaelis-Menten rate constant for low affinity metabolic
clearance of methanol
KSI first order transfer between stomach and intestine
L, dL, mL liter, deciliter, milliliter
LD50 median lethal dose
LDH lactate dehydrogenase
LH luteinizing hormone
LLF (maximum) log likelihood function
LMI leukocyte migration inhibition (assay)
LOAEL lowest-observed-adverse-effect level
M, mM, jiM molar, millimolar, micromolar
MAA 2-methoxyacetic acid
MCH mean corpuscular hemoglobin
MCHC mean corpuscular hemoglobin concentration
MCV mean cell volume
MeOH methanol
MLE maximum likelihood estimate
M-M Michaelis-Menten
MN micronuclei
MOA mode of action
4-MP 4-methylpyrazale messenger RNA
MRI magnetic resonance imaging
MTBE methyl tertiary butyl ether
MTX methotrexate
N2O/O2 nitrous oxide
NAD+ nicotinamide adenine dinucleotide
NADH reduced form of nicotinamide adenine dinucleotide
NET nitroblue tetrazolium (test)
NCEA National Center for Environmental Assessment
ND not determined
NEDO New Energy Development Organization (of Japan)
NIEHS National Institute of Environmental Health Sciences
NIOSH National Institute of Occupational Safety and Health
NK natural killer
nmol nanomole
NOAEL no-observed-adverse-effect level
NOEL no-observed-effect level
1-17
DRAFT-DO NOT CITE OR QUOTE
-------
NP nonpregnant
NR not reported
NRC National Research Council
NS not specified
NTB nitroblue tetra zolium
NTP National Toxicology Program
NZW New Zealand white (rabbit)
8-OHdG 8-hydroxydeoxyguanosine
OR osmotic resistance
OSF oral slope factor
OU ocular uterque (each eye)
OXA oxazolone
P, p probability
PBPK physiologically based pharmacokinetic
PEG polyethylene glycol
PFC plaque-forming cell
PK pharmacokinetic
PMN polymorphonuclear leucocytes
PND postnatal day
POD point of departure
ppb, ppb parts per billion, parts per million
PWG Pathology Working Group of the NTP of NIEHS
Q wave the initial deflection of the QRS complex when such deflection is
QCC cardiac output
QPC pulmonary (alveolar) ventilation scaling coefficient
QRS portion of electrocardiogram corresponding to the depolarization
of ventricular cardiac cells.
R2 square of the correlation coefficient, a measure of the reliability
of a linear relationship.
RBC red blood cell
RfC reference concentration
RfD reference dose
RNA ribonucleic acid
ROS reactive oxygen species
S9 microsomal fraction from liver
SAP serum alkaline phosphatase
s.c. subcutaneous
SCE sister chromatid exchange
S.D. standard deviation
S.E. standard error
SEM standard error of mean
1-18 DRAFT-DO NOT CITE OR QUOTE
-------
SGPT serum glutamate pyruvate transaminase
SHE Syrian hamster embryo
SOD superoxide dismutase
SOP standard operating procedure(s)
t time
Ty2, t/2 half-life
T wave the next deflection in the electrocardiogram after the QRS
complex; represents ventricular repolarization
TAME tertiary amyl methyl ether
TAS total antioxidant status
Tau taurine
THF tetrahydrofolate
TLV threshold limit value
TNFa tumor necrosis factor-alpha
TNP-LPS trinitrophenyl-lipopolysaccharide
TRI Toxic Release Inventory
U83836E vitamin E derivative
UF(s) uncertainty factor(s)
UF associated with interspecies (animal to human) extrapolation
UF associated with deficiencies in the toxicity database
UF associated with variation in sensitivity within the human
population
UFs UF associated with subchronic to chronic exposure
Vd volume of distribution
Vmax maximum enzyme velocity
VmaxC maximum velocity of the high-affinity/low-capacity pathway
v/v volume/volume
VDR visually directed reaching test
VitC vitamin C
VYS visceral yolk sac
WBC white blood cell
WOE weight of evidence
w/v weight/volume
%2 chi square
1-19
DRAFT-DO NOT CITE OR QUOTE
-------
FOREWORD
The purpose of this Toxicological Review is to provide scientific support and rationale
for the hazard and dose-response assessment in IRIS pertaining to chronic exposure to methanol.
It is not intended to be a comprehensive treatise on the chemical or toxicological nature of
methanol.
The intent of Section 6, Major Conclusions in the Characterization of Hazard and Dose
Response, is to present the major conclusions reached in the derivation of the reference dose,
reference concentration and cancer assessment, where applicable, and to characterize the overall
confidence in the quantitative and qualitative aspects of hazard and dose response by addressing
the quality of data and related uncertainties. The discussion is intended to convey the limitations
of the assessment and to aid and guide the risk assessor in the ensuing steps of the risk
assessment process.
For other general information about this assessment or other questions relating to IRIS,
the reader is referred to EPA's IRIS Hotline at (202) 566-1676 (phone), (202) 566-1749 (fax), or
hotline.iris@epa.gov.
1 -20 DRAFT-DO NOT CITE OR QUOTE
-------
AUTHORS, CONTRIBUTORS, AND REVIEWERS
CHEMICAL MANAGER
Jeffrey Gift, Ph.D.
National Center for Environmental Assessment
U.S Environmental Protection Agency
Research Triangle Park, NC
AUTHORS
Stanley Barone, Ph.D.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Allen Davis, MSPH
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Jeffrey Gift, Ph.D.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Annette lannucci, M.S.
Sciences International (first draft)
2200 Wilson Boulevard
Arlington, VA
Paul Schlosser, Ph.D.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
CONTRIBUTORS
Bruce Allen, Ph.D.
Bruce Allen Consulting
101 Corbin Hill Circle
Chapel Hill, SC
Hugh Barton, Ph.D.
National Center for Computational Toxicology
U.S. Environmental Protection Agency
Research Triangle Park, NC
1 -21 DRAFT-DO NOT CITE OR QUOTE
-------
J. Michael Davis, Ph.D.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Robinan Gentry, M.S.
ENVIRON International
650 Poydras Street
New Orleans, LA
Susan Goldhaber, M.S.
Alpha-Gamma Technologies, Inc.
3301 Benson Drive
Raleigh, NC
Mark Greenberg, Ph.D.
Senior Environmental Employee Program
U.S. Environmental Protection Agency
Research Triangle Park, NC
George Holdsworth, Ph.D.
Oak Ridge Institute for Science and Education (second draft)
Badger Road
Oak Ridge, TN
Angela Howard, Ph.D.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Connie Meacham, M.S.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Greg Miller
Office of Policy, Economics & Innovation
U.S. Environmental Protection Agency
Washington, DC
Sharon Oxendine, Ph.D.
Office of Policy, Economics & Innovation
U.S. Environmental Protection Agency
Washington, DC
Torka Poet, Ph.D.
Battelle, Pacific Northwest National Laboratories (Appendix B)
902 Battelle Boulevard
Richland, WA
1 -22 DRAFT-DO NOT CITE OR QUOTE
-------
John Rogers, Ph.D.
National Health & Environmental Effects Research laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC
Reeder Sams, II, Ph.D.
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Roy Smith, Ph.D.
Air Quality Planning & Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC
Frank Stack
Alpha-Gamma Technologies, Inc.
3301 Benson Drive
Raleigh, NC
Justin TeeGuarden, Ph.D.
Battelle, Pacific Northwest National Laboratories
902 Battelle Boulevard
Richland, WA
Chad Thompson, Ph.D., MBA
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Washington, DC
Cynthia VanLandingham, M.S.
ENVIRON International (Appendix E)
650 Poydras Street
New Orleans, LA
Deborah Wales
National Center for Environmental Assessment
U.S. Environmental Protection Agency
Research Triangle Park, NC
Lutz Weber, Ph.D., DABT
Oak Ridge Institute for Science and Education
Badger Road
Oak Ridge, TN
1 -23 DRAFT-DO NOT CITE OR QUOTE
-------
Errol Zeiger, Ph.D.
Alpha-Gamma Technologies, Inc.
3301 Benson Drive
Raleigh, NC
REVIEWERS
1 This document has been reviewed by EPA scientists, interagency reviewers from other
2 federal agencies and White House offices.
INTERNAL EPA REVIEWERS
Jane Caldwell, Ph.D.
National Center for Environmental Assessment
U.S Environmental Protection Agency
Washington, DC
Ila Cote, Ph.D., DABT
National Center for Environmental Assessment
U.S Environmental Protection Agency
Washington, DC
Robert Dewoskin Ph.D., DABT
National Center for Environmental Assessment
U.S Environmental Protection Agency
Research Triangle Park, NC
Joyce Donahue, Ph.D.
Office of Water
U.S. Environmental Protection Agency
Washington, DC
Marina Evans, Ph.D.
National Health and Environmental Effects Research Laboratory
U.S Environmental Protection Agency
Research Triangle Park, NC
Lynn Flowers, Ph.D., DABT
National Center for Environmental Assessment
U.S Environmental Protection Agency
Research Triangle Park, NC
BrendaFoos, M.S.
Office of Children's Health Protection and Environmental Education
U.S Environmental Protection Agency
Washington, DC
Jennifer Jinot, Ph.D.
National Center for Environmental Assessment
1 -24 DRAFT-DO NOT CITE OR QUOTE
-------
U.S Environmental Protection Agency
Washington, DC
Eva McLanahan, Ph.D.
National Center for Environmental Assessment
U.S Environmental Protection Agency
Research Triangle Park, NC
John Vandenberg, Ph.D.
National Center for Environmental Assessment
U.S Environmental Protection Agency
Research Triangle Park, NC
Debra Walsh, M.S.
National Center for Environmental Assessment
U.S Environmental Protection Agency
Research Triangle Park, NC
1 -25 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 methanol.
3 IRIS Summaries may include oral reference dose (RfD) and inhalation reference concentration
4 (RfC) values for chronic and other exposure durations, and a carcinogenicity assessment.
5 The RfD and RfC, if derived, provide quantitative information for use in risk assessments
6 for health effects known or assumed to be produced through a nonlinear (presumed threshold)
7 mode of action (MOA). The RfD (expressed in units of milligrams per kilogram per day
8 [mg/kg-day]) is defined as an estimate (with uncertainty spanning perhaps an order of
9 magnitude) of a daily exposure to the human population (including sensitive subgroups) that is
10 likely to be without an appreciable risk of deleterious effects during a lifetime. The inhalation
11 RfC (expressed in units of milligrams per cubic meter [mg/m3]) is analogous to the oral RfD but
12 provides a continuous inhalation exposure estimate. The inhalation RfC considers toxic effects
13 for both the respiratory system (portal-of-entry) and for effects peripheral to the respiratory
14 system (extrarespiratory or systemic effects). Reference values are generally derived for chronic
15 exposures (up to a lifetime), but may also be derived for acute (<24 hours), short-term (>24
16 hours up to 30 days), and subchronic (>30 days up to 10% of lifetime) exposure durations, all of
17 which are derived based on an assumption of continuous exposure throughout the duration
18 specified. Unless specified otherwise, the RfD and RfC are derived for chronic exposure
19 duration.
20 The carcinogenicity assessment provides information on the carcinogenic hazard
21 potential of the substance in question, and quantitative estimates of risk from oral and inhalation
22 exposure may be derived. The information includes a weight-of-evidence (WOE) judgment of
23 the likelihood that the agent is a human carcinogen and the conditions under which the
24 carcinogenic effects may be expressed. Quantitative risk estimates may be derived from the
25 application of a low-dose extrapolation procedure. If derived, the oral slope factor is a plausible
26 upper bound on the estimate of risk per mg/kg-day of oral exposure. Similarly, an inhalation unit
27 risk (IUR) is a plausible upper bound on the estimate of risk per microgram per cubic meter
28 (ug/m3) air breathed.
29 Development of these hazard identification and dose-response assessments for methanol
30 has followed the general guidelines for risk assessment as set forth by the National Research
31 Council (NRC) (1983). EPA Guidelines and Risk Assessment Forum Technical Panel Reports
32 that may have been used in the development of this assessment include the following: Guidelines
33 for the Health Risk Assessment of Chemical Mixtures (U.S. EPA, 1986a), Guidelines for
34 Mutagenicity Risk Assessment (U.S. EPA, 1986b), Recommendations for and Documentation of
35 Biological Values for Use in Risk Assessment (U.S. EPA, 1988), Guidelines for Developmental
3 6 Toxicity Risk Assessment (U.S. EPA, 1991), Interim Policy for Particle Size and Limit
1 -26 DRAFT-DO NOT CITE OR QUOTE
-------
1 Concentration Issues in Inhalation Toxicity Studies (U.S. EPA, 1994a), Methods for Derivation
2 of Inhalation Reference Concentrations and Application of Inhalation Dosimetry (U.S. EPA,
3 1994b), Use of the Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 1995),
4 Guidelines for Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for
5 Neurotoxicity Risk Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk
6 Characterization (U.S. EPA, 2000a), Benchmark Dose Technical Guidance Document
1 (U.S. EPA, 2000b), Supplementary Guidance for Conducting Health Risk Assessment of
8 Chemical Mixtures (U.S. EPA, 2000c), A Review of the Reference Dose and Reference
9 Concentration Processes (U.S. EPA, 2002a), Guidelines for Carcinogen Risk Assessment
10 (U.S. EPA, 2005a), Supplemental Guidance for Assessing Susceptibility from Early-Life
11 Exposure to Carcinogens (U.S. EPA, 2005b), Science Policy Council Handbook: Peer Review
12 (U.S. EPA, 2006a), and A Framework for Assessing Health Risks of Environmental Exposures to
13 Children (U. S. EPA, 2006b).
14 The literature search strategy employed for this compound was based on the Chemical
15 Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
16 scientific information submitted by the public to the IRIS Submission Desk was also considered
17 in the development of this document. The relevant literature was reviewed through January,
18 2009.
1 -27 DRAFT-DO NOT CITE OR QUOTE
-------
2. CHEMICAL AND PHYSICAL INFORMATION
1 Methanol is also known as methyl alcohol, wood alcohol; Carbinol; Methylol; colonial
2 spirit; Columbian spirit; methyl hydroxide; monohydroxymethane; pyroxylic spirit; wood
3 naphtha; and wood spirit (Chemfmder, 2002). Some relevant physical and chemical properties
4 are listed below (Hazardous Substances Data Bank [HSDB], 2002, International Programme on
5 Chemical Safety [IPCS], 1997).
Table 2-1. Relevant physical and chemical properties of methanol
CASRN: 67-56-1
Empirical formula: CHsOH
Molecular weight: 32.04
Vapor pressure: 160 mmHg at 30 °C
Vapor Density: 1.11
Specific gravity: 0.7866 g/mL (25 °C)
Boiling point: 64.7 °C
Melting point: -98 °C
Water solubility: Miscible
Log octanol-water partition -0.82 to -0.68
coefficient:
Conversion factor (in air): 1 ppm =1.31 mg/m3;
1 mg/m3 = 0.763 ppm
6 Methanol is a clear, colorless liquid that has an alcoholic odor (IPCS, 1997). Endogenous
7 levels of methanol are present in the human body as a result of both metabolism1 and dietary
8 sources such as fruit, fruit juices, vegetables and alcoholic beverages,2 and can be measured in
9 exhaled breath and body fluids (Turner et al., 2006; CERHR 2004; IPCS 1997). Dietary
10 exposure to methanol also occurs through the intake of some food additives. The artificial
11 sweetener aspartame and the beverage yeast inhibitor dimethyl dicarbonate (DMDC) release
12 methanol as they are metabolized (Stegink et al., 1989). In general, aspartame exposure does not
13 contribute significantly to the background body burden of methanol (Butchko, 2002). Oral,
14 dermal, or inhalation exposure to methanol in the environment, consumer products, or workplace
15 also occur.
16 Methanol is a high production volume chemical with many commercial uses and it is a
17 basic building block for numerous chemicals. Many of its derivatives are used in the
1 Methanol is generated metabolically through enzymatic pathways such as the methyltransferase system (Fisher
et al., 2000).
2 Fruits and vegetables contain methanol. Further, ripe fruits and vegetables contain natural pectin, which is
degraded to methanol in the body by bacteria present in the colon (Siragusa et al., 1988). Increased levels of
methanol in blood and exhaled breath have also been observed after the consumption of ethanol (Fisher et al., 2000).
2-1 DRAFT-DO NOT CITE OR QUOTE
-------
1 construction, housing or automotive industries, with its use as a transportation fuel accounting
2 for -20% of its demand (Methanol Institute, 2006). Consumer products that contain methanol
3 include varnishes, shellacs, paints, windshield washer fluid, antifreeze, adhesives, de-icers, and
4 Sterno heaters. Methanol is used to produce other chemicals and is among the highest
5 production volume chemicals in the EPA's Toxic Release Inventory (TRI), which reported that
6 183,000 pounds of methanol was released or disposed of in the United States in 2003, making
7 methanol among the top 10 chemicals on the list entitled "On-site and Off-site Reported
8 Disposed of or Otherwise Released (in pounds), for Facilities in All Industries, for All
9 Chemicals" (U.S. EPA, 2006c). Natural gas is the primary feedstock for methanol and as such,
10 the major portion of its production cost. Due to the high cost of natural gas in North America,
11 North America methanol production capacity is down from 9 million metric tons at the turn of
12 the century to less than 3 million metric tons at the end of 2005 and is expected to be nonexistent
13 by 2010 (Methanol Institute, 2006). While production has switched to other regions of the
14 world, demand for methanol is growing steadily in almost all end uses. A large reason for the
15 increase in demand is its use in the production of biodiesel, a low-sulfur, high-lubricity fuel
16 source that is gaining favor in nearly every major region of the world. Starting from less than
17 100,000 metric tons in 2005, global demand for methanol into biodiesel is expected to reach as
18 much as 1.5 million metric tons by the year 2010. Power generation and fuel cells could also be
19 large end users of methanol in the near future (Methanol Institute, 2006).
2-2 DRAFT-DO NOT CITE OR QUOTE
-------
3. TOXICOKINETICS
3.1. OVERVIEW
1 As has been noted, methanol occurs naturally in the human body as a product of
2 metabolism and through intake of fruits, vegetables, and alcoholic beverages (Turner et al., 2006;
3 CERHR 2004; IPCS 1997). Table 3-1 summarizes background blood methanol levels in healthy
4 humans which were found to range from 0.25-4.7 mg/L. One study reported a higher
5 background blood methanol level in females versus males (Batterman and Franzblau, 1997), but
6 most studies did not evaluate gender differences. Formate, a metabolite of methanol, also occurs
7 naturally in the human body (IPCS 1997). Table 3-1 outlines background levels of formate in
8 human blood. In most cases, methanol and formate blood levels were measured in healthy adults
9 following restriction of methanol-producing foods from the diet.3
10 The absorption, excretion, and metabolism of methanol are well known and have been
11 consistently summarized in reviews such as CERHR (2004), IPCS (1997), U.S. EPA (1996),
12 Kavet and Nauss (1990), HEI (1987), and Tephly and McMartin (1984). Therefore, the major
13 portion of this toxicokinetics overview is based upon those reviews.
14 Studies conducted in humans and animals demonstrate rapid absorption of methanol by
15 inhalation, oral, and dermal routes of exposure. Table 3-2 outlines increases in human blood
16 methanol levels following various exposure scenarios. Blood levels of methanol following
17 various exposure conditions have also been measured in monkeys, mice, and rats, and are
18 summarized in Tables 3-3, 3-4, and 3-5, respectively. Once absorbed, methanol pharmacokinetic
19 (PK) data and physiologically based pharmacokinetic (PBPK) model predictions indicate rapid
20 distribution to all organs and tissues according to water content, as an aqueous-soluble alcohol.
21 Tissue:blood concentration ratios for methanol are predicted to be similar through different
22 exposure routes, though the kinetics will vary depending on exposure route and timing (e.g.,
23 bolus oral exposure versus longer-term inhalation). Because smaller species generally have
24 faster respiration rates relative to body weight than larger species, they are predicted to have a
25 higher rate of increase of methanol concentrations in the body when exposed to the same
26 concentration in air.
27 At doses that do not saturate metabolic pathways, a small percentage of methanol is
28 excreted directly in urine. Because of the high blood:air partition coefficient for methanol and
29 rapid metabolism in all species studied, the bulk of clearance therefore occurs by metabolism,
30 though exhalation and urinary clearance will become more significant when doses or exposures
31 are sufficiently high to saturate metabolism. Metabolic saturation and the corresponding
32 clearance shift have not been observed in humans and nonhuman primates because doses used
3 In general, background levels among people who are on normal/non-restricted diets will be higher than those
reported.
3-1 DRAFT-DO NOT CITE OR QUOTE
-------
1 were limited to the linear range, but the enzymes involved in primate metabolism are also
2 saturable.
3 The primary route of methanol elimination in mammals is through a series of oxidation
4 reactions that form formaldehyde, formate, and carbon dioxide (Figure 3-1). As noted in
5 Figure 3-1, methanol is converted to formaldehyde by alcohol dehydrogenase-1 (ADH1) in
6 primates and by catalase (CAT) and ADH1 in rodents. Although the first step of metabolism
7 occurs through different pathways in rodents and nonhuman primates, Kavet and Nauss (1990)
8 report that the reaction proceeds at similar rates (Vmax =30 and 48 mg/h/kg in rats and nonhuman
9 primates, respectively). In addition to enzymatic clearance, methanol can react with hydroxyl
10 radicals to spontaneously yield formaldehyde (Harris et al., 2003). Mannering et al. (1969) also
11 reported a similar rate of methanol metabolism in rats and monkeys, with 10 and 14% of a 1 g/kg
12 dose oxidized in 4 hours, respectively; the rate of oxidation by mice was about twice as fast, 25%
13 in 4 hours. In an HEI study by Pollack and Brouwer (1996), the metabolism of methanol was
14 2 times as fast in mice versus rats, with a Vmax for elimination of 117 and 60.7 mg/h/kg,
15 respectively. Despite the faster elimination rate of methanol in mice versus rats, mice
16 consistently exhibited higher blood methanol levels than rats when inhaling equivalent methanol
17 concentrations (See Tables 3-4 and 3-5). Possible explanations for the higher methanol
18 accumulation in mice include faster respiration (inhalation rate/body weight) and increased
19 fraction of absorption by the mouse (Perkins et al., 1995a). Because smaller species generally
20 have faster breathing rates than larger species, humans would be expected to absorb methanol via
21 inhalation more slowly than rats or mice inhaling equivalent concentrations. If humans eliminate
22 methanol at a comparable rate to rats and mice, then humans would also be expected to
23 accumulate less methanol than those smaller species. However, if humans eliminate methanol
24 more slowly than rats and mice, such that the ratio of absorption to elimination stays the same,
25 then humans would be expected to accumulate methanol to the same internal concentration but to
26 take longer to reach that concentration.
27 In all species, formaldehyde is rapidly converted to formate, with the half-life for
28 formaldehyde being ~1 minute. Formaldehyde is oxidized to formate by two metabolic
29 pathways (Teng et al., 2001). The first pathway (not shown in Figure 3-1) involves conversion
30 of free formaldehyde to formate by the so-called low-affinity pathway (affinity = l/Km =
31 0.002/|jM) mitochondrial aldehyde dehydrogenase-2 (ALDH2). The second pathway
32 (Figure 3-1) involves a two-enzyme system that converts glutathione-conjugated formaldehyde
33 (S-hydroxymethylglutathione [HMGSH]) to the intermediate S-formylglutathione, which is
34 subsequently metabolized to formate and glutathione (GSH) by S-formylglutathione hydrolase.4
35 The first enzyme in this pathway, formaldehyde dehydrogenase-3 (ADH3), is rate limiting, and
36 the affinity of HMGSH for ADH3 (affinity = l/Km = 0.15/nM) is about a 100-fold higher than
4 Other enzymatic pathways for the oxidation of formaldehyde have been identified in other organisms, but this is
the pathway that is recognized as being present in humans (Caspi et al., 2006; http://metacvc. org)
3-2 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
that of free formaldehyde for ALDH2. In addition to the requirement of GSH for ADH3 activity,
oxidation by ADH3 is nicotinamide adenine dinucleotide- (NAD+-)dependent. Under normal
physiological conditions NAD+ levels are about two orders of magnitude higher than NADH,
and intracellular GSH levels (mM range) are often high enough to rapidly scavenge
formaldehyde (Svensson et al., 1999; Meister and Adnerson, 1983); thus, the oxidation of
HMGSH is favorable. In addition, genetic ablation of ADH3 results in increased formaldehyde
toxicity (Deltour et al., 1999). These data indicate that ADH3 is likely to be the predominant
enzyme responsible for formaldehyde oxidation at physiologically relevant concentrations,
whereas ALDHs likely contribute to formaldehyde clearance at higher concentrations (Dicker
and Cederbaum, 1986).
Primates
Alcohol dehydrogenase
(ADH1)
Formaldehyde dehydrogenase
(ADH3)
S-formylglutathione hydrolase
Folate dependent pathway
(see Figure 3-2)
CH3OH (Methanol)
I
HCHO
(Formaldehyde)
( + GSH)
HMGSH
(hydroxymethyl-GSH)
4
(^-formyl glutathione)
|(-GSH)
HCOO (Formate)
4
I
Rodents
CAT and ADH1
Formaldehyde dehydrogenase
(ADH3)
S-formylglutathione hydrolase
CAT-peroxide and
Folate-dependent pathway
(see Figure 3-2)
CO2 (Carbon dioxide)
Figure 3-1. Methanol metabolism and key metabolic enzymes in primates and rodents.
Source: IPCS (1997).
Rodents convert formate to carbon dioxide (CO2) through a folate-dependent enzyme
system and a CAT-peroxide system (Dikalova et al., 2001). Formate can undergo adenosine
triphosphate- (ATP-) dependent addition to tetrahydrofolate (THF), which can carry either one or
two one-carbon groups. Formate can conjugate with THF to form jV10-formyl-THF and its
3-3 DRAFT-DO NOT CITE OR QUOTE
-------
1 isomer 7V5-formyl-THF, both of which can be converted to TV5, jV70-methenyl-THF and
2 subsequently to other derivatives that are ultimately incorporated into DNA and proteins via
3 biosynthetic pathways (Figure 3-2). There is also evidence that formate generates CO2" radicals,
4 and can be metabolized to CC>2 via CAT and via the oxidation of jV10-formyl-THF (Dikalova
5 etal.,2001).
Cytoplasm
Mitochondria
10-formylTHF
ff \C~~^r forarcate^
methenyiTHF THF
\>£- — • senne "Sfi
^(mSHMT
methyleneTHF glycine ~
MTHFDl //
yyfFDH
~ fonrate,_/V (
// \,Q methenyiTHF
THF ' Z ii
..,i~..v. serine — -sS^c-MWT* H
=ss glycine •* methyleneTHF d
I
1 MTHf'R
5-methylTHF
horaocysteme
AdoHyc
y dTMP
- — ^
^x
Methionine
)
AdoMet ^
Figure 3-2. Folate-dependent formate metabolism. Tetrahydrofolate (THF)-mediated
one carbon metabolism is required for the synthesis of purines, thymidylate, and
methionine.
Source: Montserrat et al. (2006).
7 Unlike rodents, formate metabolism in primates occurs solely through a folate-dependent
8 pathway. Black et al. (1985) reported that hepatic THF levels in monkeys are 60% of that in rats,
9 and that primates are far less efficient in clearing formate than are rats and dogs. Studies
10 involving [14C]formate suggest that -80% is exhaled as 14CO2, 2-7% is excreted in the urine, and
11 -10% undergoes metabolic incorporation (Hanzlik et al., 2005, and references therein). Mice
12 deficient in formyl-TFIF dehydrogenase exhibit no change in LDso (via intraperitoneal [i.p.]) for
13 methanol or in oxidation of high doses of formate. Thus it has been suggested that rodents
14 efficiently clear formate via folate-dependent pathways, peroxidation by CAT, and by an
15 unknown third pathway; conversely, primates do not appear to exhibit such capacity and are
16 more sensitive to metabolic acidosis following methanol poisoning (Cook et al., 2001).
17 Blood methanol and formate levels measured in humans under various exposure
18 scenarios are reported in Table 3-2. As noted in Table 3-2, 75-minute to 6-hour exposures of
19 healthy humans to 200 parts per million (ppm) methanol vapors, the American Council of
3-4
DRAFT-DO NOT CITE OR QUOTE
-------
1 Governmental Industrial Hygienists (ACGIH) threshold limit value (TLV) for occupational
2 exposure (ACGIH, 2000), results in increased levels of blood methanol but not formate. A
3 limited number of monitoring studies indicate that levels of methanol in outdoor air are orders of
4 magnitude lower than the TLV (IPCS, 1997). Table 3-3 indicates that exposure of monkeys to
5 600 ppm methanol vapors for 2.5 hours increased blood methanol but not blood formate levels.
6 Normal dietary exposure to aspartame, which releases 10% methanol during metabolism, is
7 unlikely to significantly increase blood methanol or formate levels (Butchko et al., 2002). Data
8 in Table 3-2 suggest that exposure to high concentrations of aspartame is unlikely to increase
9 blood formate levels; no increase in blood formate levels were observed in adults ingesting
10 "abusive doses" (100-200 mg/kg) of aspartame (Stegink et al., 1981). Kerns et al. (2002)
11 studied the kinetics of formate in 11 methanol-poisoned patients (mean initial methanol level of
12 57.2 mmol/L or 1.83 g/L) and determined an elimination half-life of 3.4 hours for formate.
13 Kavet and Nauss (1990) estimated that a methanol dose of 11 mM or 210 mg/kg is needed to
14 saturate folate-dependent metabolic pathways in humans. There are no data on blood methanol
15 and formate levels following methanol exposure of humans with reduced ADH activity or
16 marginal folate tissue levels, a possible concern regarding sensitive populations. As discussed in
17 greater detail in Section 3.2, a limited study in folate-deficient monkeys demonstrated no
18 increase in blood formate levels following exposure to 900 ppm methanol vapors for 2 hours. In
19 conclusion, limited available data suggest that typical occupational, environmental, and dietary
20 exposures are likely to increase baseline blood methanol but not formate levels in most humans.
3.2. KEY STUDIES
21 Some recent toxicokinetic and metabolism studies (Burbacher et al., 2004a, 1999b;
22 Medinsky et al., 1997; Pollack and Brouwer, 1996; Dorman et al., 1994) provide key information
23 on interspecies differences, methanol metabolism during gestation, metabolism in the nonhuman
24 primate, and the impact of folate deficiency on the accumulation of formate.
25 As part of an effort to develop a physiologically based toxicokinetic model for methanol
26 distribution in pregnancy, Pollack and Brouwer (1996) conducted a large study that compared
27 toxicokinetic differences in pregnant and nonpregnant (NP) rats and mice. Methanol disposition5
28 was studied in Sprague-Dawley rats and CD-I mice that were exposed to 100-2,500 mg/kg of
29 body weight pesticide-grade methanol in saline by intravenous (i.v.) or oral routes. Exposures
30 were conducted in NP rats and mice, pregnant rats on gestation days (GD)7, GDI4, and GD20,
31 and pregnant mice on GD9 and GDIS. Disposition was also studied in pregnant rats and mice
32 exposed to 1,000-20,000 ppm methanol vapors for 8 hours. Three to five animals were
33 examined at each dose and exposure condition.
5 Methanol concentrations in whole blood and urine were determined by gas chromatography with flame ionization
detection (Pollack and Kawagoe, 1991)
3-5 DRAFT-DO NOT CITE OR QUOTE
-------
1 Based on the fit of various kinetic models to methanol measurements taken from all
2 routes of exposure, the authors concluded that high exposure conditions resulted in nonlinear
3 disposition of methanol in mice and rats.6 Both linear and nonlinear pathways were observed
4 with the relative contribution of each pathway dependent on concentration. At oral doses of
5 100-500 mg/kg of body weight, methanol was metabolized to formaldehyde and then formic
6 acid through the saturable nonlinear pathway. A parallel, linear route characteristic of passive-
7 diffusion accounted for an increased fraction of total elimination at higher concentrations.
8 Nearly 90% of methanol elimination occurred through the linear route at the highest oral dose of
9 2,500 mg/kg of body weight.
10 Oral exposure resulted in rapid and essentially complete absorption of methanol. No
11 significant change in blood area under the curve (AUC) methanol was seen between NP and
12 GD7, GDI4 and GD20 rats exposed to single oral gavage doses of 100 and 2,500 mg/kg, nor
13 between NP and GD9 and GDIS mice at 2,500 mg/kg. The data as a whole suggested that the
14 distribution of orally and i.v. administered methanol was similar in rats versus mice and in
15 pregnant rodents versus NP rodents with the following exceptions:
16 • There was a statistically significant increase in the ratio of apparent volume of
17 distribution (Vd) to fractional bioavailability (F) by -20% (while F decreased but not
18 significantly), between NP and GD20 rats exposed to 100 mg/kg orally. However, this
19 trend was not seen in rats or mice exposed to 2,500 mg/kg, and the result in rats at
20 100 mg/kg could well be a statistical artifact since both Vd and F were being estimated
21 from the same data, making the model effectively over-parameterized.
22 • There were statistically significant decreases in the fraction of methanol absorbed by the
23 fast process (resulting in a slower rise to peak blood concentrations, though the peak is
24 unchanged) and Vmax for clearance between NP and GDIS mice. No such differences
25 were observed between NP and GD9 mice.
26 • The authors estimated a twofold higher Vmax for methanol elimination in mice versus rats
27 following oral administration of 2,500 mg/kg methanol, suggesting that similar oral doses
28 would result in lower methanol concentrations in the mouse versus rat.
29 Methanol penetration from maternal blood to the fetal compartment was examined in
30 GD20 rats by microdialysis.7 A plot of the amniotic concentration versus maternal blood
31 concentration (calculated from digitization of Figure 17 of Pollack and Brouwer [1996] report) is
32 shown in Figure 3-3. The ratio is slightly less than 1:1 (dashed line in plot) and appears to be
33 reduced with increasing methanol concentrations, possibly due to decreased blood flow to the
6 A model incorporating parallel linear and nonlinear routes of methanol clearance was required to fit the data from
the highest exposure groups.
7 Microdialysis was conducted by exposing the uterus (midline incision), selecting a single fetus in the middle of the
uterine horn and inserting a microdialysis probe through a small puncture in the uterine wall proximal to the head of
the fetus.
3-6 DRAFT-DO NOT CITE OR QUOTE
-------
1 fetal compartment. Nevertheless, this is a very minor departure from linearity, consistent with a
2 substrate such as methanol that penetrates cellular membranes readily and distributes throughout
3 total body water.
50001
O)
4000
c
O
£ 3000
,' *
= -4E-05>? + 1.0782x
1000 2000 3000 4000 5000
Maternal blood concentration (mg/L)
Figure 3-3. Plot of fetal (amniotic) versus maternal methanol concentrations in GD20
rats. Data extracted from Figure 17 by digitization, and amniotic concentration
obtains as ("Fetal Amniotic Fluid/Maternal Blood Methanol")x("Maternal
Methanol").
Source: Pollack and Brouwer (1996).
4 Inhalation exposure resulted in less absorption in both rats and mice as concentrations of
5 methanol vapors increased, which was hypothesized to be due to decreased breathing rate and
6 decreased absorption efficiency from the upper respiratory tract.8 Based on blood methanol
7 concentrations measured following 8-hour inhalation exposures to concentrations ranging from
8 1,000-20,000 ppm, the study authors (Pollack and Brown, 1996) concluded that methanol
9 accumulation in the mouse occurred at a two- to threefold greater rate compared to the rat. They
10 speculated that faster respiration rate and more complete absorption in the nasal cavity of mice
11 may explain the higher methanol accumulation and greater sensitivity to developmental toxicity
12 (see Section 4.3.2).
8 Exposed mice spent some exposure time in an active state, characterized by a higher ventilation rate, and the
remaining time in an inactive state, with lower (~i/2 of active) ventilation. The inactive ventilation rate was
unchanged by methanol exposure, but the active ventilation showed a statistically significant methanol-
concentration-related decline. There was also some decline in the fraction of time spent in the active state, but this
too was not statistically significant.
3-7 DRAFT-DO NOT CITE OR QUOTE
-------
1 The Pollack and Brouwer (1996) study was useful for comparing effects in pregnant and
2 NP rodents exposed to high doses, but the implication of these results for humans exposed to
3 ambient levels of methanol is not clear (CERHR, 2004).
4 Burbacher et al. (2004b, 1999a) examined toxicokinetics mMacaca fascicularis
5 monkeys prior to and during pregnancy. The study objectives were to assess the effects of
6 repeated methanol exposure on disposition kinetics, determine whether repeated methanol
7 exposures result in formate accumulation, and examine the effects of pregnancy on methanol
8 disposition and metabolism. Reproductive and developmental toxicity associated with this study
9 were also examined and are discussed in Sections 4.3.2 and 4.4.2. In a 2-cohort design, 48 adult
10 females (6 animals/dose/group/cohort) were exposed to 0, 200, 600, or 1,800 ppm methanol
11 vapors (99.9% purity) for 2.5 hours/day, 7 days/week for 4 months prior to breeding and during
12 the entire breeding and gestation periods. Six-hour methanol clearance studies were conducted
13 prior to and during pregnancy. Burbacher et al. (2004b, 1999a) reported that:
14 • At no point during pregnancy was there a significant change in endogenous methanol
15 blood levels, which ranged from 2.2-2.4 mg/L throughout.
16 • PK studies were performed initially (Study 1), after 90 days of pre-exposure and prior to
17 mating (Study 2), between GD66 and GD72 (Study 3), and again between GD126 and
18 GDI32 (Study 4). These studies were analyzed using classical PK (one-compartment)
19 models.
20 • Disproportionate mean, dose-normalized, and net blood methanol dose-time profiles in
21 the 600 and 1,800 ppm groups suggested saturation of the metabolism-dependent
22 pathway. Data from the 600 ppm group fit a linear model, while data from the 1,800 ppm
23 group fit a Michaelis-Menten model.
24 • Methanol clearance rates modestly increased between Study 1 and Study 2 (90 days prior
25 to mating). This change was attributed to enzyme induction from the subchronic
26 exposure.
27 • Blood methanol levels were measured every 2 weeks throughout pregnancy, and while
28 there was measurement-to-measurement variation, there was no significant change or
29 trend over the course of pregnancy. There appears to be an upward trend in elimination
30 half-life and corresponding downward trend in blood methanol clearance between Studies
31 2, 3, and 4. However, the changes are not statistically significant and the time-courses
32 for blood methanol concentration (elimination phase) appear fairly similar.
33 • Significant differences between pre-breeding and gestational blood plasma formate levels
34 were observed but were not dose dependent (Table 3-6).
35 • Significant differences in serum folate levels in periods prior to and during pregnancy
36 were not dose dependent (Table 3-7).
3-8 DRAFT-DO NOT CITE OR QUOTE
-------
1 An HEI review committee (Burbacher et al., 1999a) noted that this was a quality study
2 using a relevant species. Although the study can be used to predict effects in adequately
3 nourished individuals, the study may not be relevant to persons who are folate deficient.
4 A series of studies by Medinsky et al. (1997) and Dorman et al. (1994) examined
5 metabolism and pharmacokinetics of [14C]methanol and [14C]formate in normal and folate-
6 deficient cynomolgus, M. fascicularis monkeys that were exposed to environmentally relevant
7 concentrations of [14C]methanol through an endotracheal tube while anesthetized. In the first
8 stage of the study, 4 normal 12-year-old cynomolgus monkeys were each exposed to 10, 45, 200,
9 and 900 ppm [14C]methanol vapors (>98% purity) for 2 hours. Each exposure was separated by
10 at least 2 months. After the first stage of the study was completed, monkeys were given a folate-
11 deficient diet supplemented with 1% succinylsulfathiozole (an antibacterial sulfonamide used to
12 inhibit folic acid biosynthesis from intestinal bacteria) for 6-8 weeks in order to obtain folate
13 concentrations of <3 ng/mL serum and <120 ng/mL erythrocytes. Folate deficiency did not alter
14 hematocrit level, red blood cell count, mean corpuscular volume, or mean corpuscular
15 hemoglobin level. The folate-deficient monkeys were exposed to 900 ppm [14C]methanol for
16 2 hours. The results of the Medinsky et al. (1997) and Dorman et al. (1994) studies showed:
17 • Dose-dependent changes in toxicokinetics and metabolism did not occur as indicated by a
18 linear relationship between inhaled [14C]methanol concentration and end-of-exposure
19 blood [14C]methanol level, [14C]methanol AUC and total amounts of exhaled
20 [14C]methanol and [14C]carbon dioxide.
21 • Methanol concentration had no effect on elimination half-life (<1 hour) and percent
22 urinary [14C]methanol excretion (<0.01%) at all doses.
23 • Following exposure to 900 ppm methanol, urinary excretion or exhalation of
24 [14C]methanol did not differ significantly between monkeys in the folate sufficient and
25 deficient state. There was no significant [14C] formate accumulation at any dose.
26 • Peak blood [14C]formate levels were significantly higher in folate-deficient monkeys, but
27 did not exceed endogenous blood levels reported by the authors to be between 0.1 and
28 0.2 mmol/L (4.6 to 9.2 mg/L).
29 An FIEI review committee (Medinsky et al., 1997) noted that absolute values in this study
30 cannot be extrapolated to humans because the use of an endotracheal tube in anesthetized
31 animals results in an exposure scenario that is not relevant to humans. However, the data in this
32 study suggest that a single exposure to an environmentally relevant concentration of methanol is
33 unlikely to result in a hazardous elevation in formate levels, even in individuals with moderate
34 folate deficiency.
3-9 DRAFT-DO NOT CITE OR QUOTE
-------
3.3. HUMAN VARIABILITY IN METHANOL METABOLISM
1 The ability to metabolize methanol may vary among individuals as a result of genetic,
2 age, and environmental factors. Reviews by Agarwal (2001), Burnell et al. (1989), Bosron and
3 Li (1986), and Pietruszko (1980), discuss genetic polymorphisms for ADH. Class IADH, the
4 primary ADH in human liver, is a hetero- or homodimer composed of randomly associated
5 polypeptide units encoded by three separate gene loci (ADH1 A, ADH1B, and ADH1C).
6 Polymorphisms have been found to occur at the ADH1B (ADH1B*2, ADH1B*3) and ADH1C
7 (ADH1C*2) gene loci; however, no human allelic polymorphism has been found in ADH1 A.
8 The ADH1B*2 phenotype is estimated to occur in -15% of Caucasians of European descent,
9 85% of Asians, and <5% of African Americans. Fifteen percent of African Americans have the
10 ADH1B*3 phenotype, while it is found in <5% of Caucasian Europeans and Asians. To date,
11 there are two reports of polymorphisms in ADH3 (Cichoz-Lach et al., 2007; Hedberg et al.,
12 2001), yet the functional consequence(s) for these polymorphisms remains unclear.
13 Although racial and ethnical differences in the frequency of the occurrence of ADH
14 alleles in different populations have been reported, ADH enzyme kinetics (Vmax and Km) have not
15 been reported for methanol. There is an abundance of information pertaining to the kinetic
16 characteristics of the ADH dimers to metabolize ethanol in vitro; however, the functional and
17 biological significance is not well understood due to the lack of data documenting metabolism
18 and disposition of methanol or ethanol in individuals of known genotype. While potentially
19 significant, the contribution of ethnic and genetic polymorphisms of ADH to the interindividual
20 variability in methanol disposition and metabolism can not be reliably quantified at this time.
21 Because children generally have higher baseline breathing rates and are more active, they
22 may receive higher methanol doses than adults exposed to equivalent concentrations of any air
23 pollutant (CERHR, 2004). There is evidence that children under 5 years of age have reduced
24 ADH activity. A study by Pikkarainen and Raiha (1967) measured liver ADH activity using
25 ethanol as a substrate and found that 2-month-old fetal livers have -3-4% of adult ADH liver
26 activity. ADH activity in 4-5 month old fetuses is -10% of adult activity, and an infant's activity
27 is -20% of adult activity. ADH continues to increase in children with age and reaches a level
28 that is within adult ranges at 5 years of age. Adults were found to have great variation in ADH
29 activity (1,625-6,530/g liver wet weight or 2,030-5,430 mU/100 mg soluble protein). Smith
30 et al. (1971) also compared liver ADH activity in 56 fetuses (9-22 weeks gestation), 37 infants
31 (premature to <1 year old), and 129 adults (>20 years old) using ethanol as a substrate. ADH
32 activity was 30% of adult activity in fetuses and 50% of adult activity in infants. There is
33 evidence that some human infants are able to efficiently eliminate methanol at high exposure
34 levels, however, possibly via CAT (Tran et al., 2007).
35 ADH3 exhibits little or no activity toward small alcohols, thus the previous studies
36 provide no information about the ontogeny of formaldehyde clearance. While such data on
3-10 DRAFT-DO NOT CITE OR QUOTE
-------
1 ADH3 activity does not exist, ADH3 mRNAis abundantly expressed in the mouse fetus (Ang
2 et al., 1996) and is detectible in human fetal tissues (third trimester), neonates and children
3 (Hines et al., 2002; Estonius et al., 1996).
4 As noted earlier in this section, folate-dependent reactions are important in the
5 metabolism of formate. Individuals who are commonly folate deficient include those who are
6 pregnant or lactating, have gastrointestinal (GI) disorders, have nutritionally inadequate diets, are
7 alcoholics, smoke, have psychiatric disorders, have pernicious anemia, or are taking folic acid
8 antagonist medications such as some antiepileptic drugs (CERHR, 2004; IPCS, 1997). Groups
9 which are known to have increased incidence of folate deficiencies include Hispanic and African
10 American women, low-income elderly, and mentally ill elderly (CERHR, 2004). A
11 polymorphism in methylene tetrahydrofolate reductase reduces folate activity and is found in
12 21% of Hispanics in California and 12% of Caucasians in the United States. Genetic variations
13 in folic acid metabolic enzymes and folate receptor activity are theoretical causes of folate
14 deficiencies.
Table 3-1. Background blood methanol and formate levels in humans
Description of human subjects
12 males on restricted diet (no
methanol-containing or methanol-
producing foods) for 12 hr
22 adults on restricted diet (no
methanol-containing or methanol-
producing foods) for 24 hr
3 males who ate a breakfast with no
aspartame-containing cereals and no
juice
5 males who ate a breakfast with no
aspartame-containing cereals and no
juice (second experiment)
Adults who drank no alcohol for 24 hr
12 adults who drank no alcohol for
24 hr
4 adult males who fasted for 8 hr, drank
no alcohol for 24 hr, and took in no
fruits, vegetables, or juices for 18 hr
30 fasted adults
24 fasted infants
Methanol (mg/L)
mean + S.D.
(Range)
0.570 + 0.305
(0.25-1.4)
1.8 + 2.6
(No range data)
1.82+1.21
(0.57-3.57)
1.93+0.93
(0.54-3.15)
1.8 + 0.7
(No range data)
1.7 + 0.9
(0.4-4.7)
No mean data
(1.4-2.6)
<4
(No range data)
<3.5
(No range data)
Formate (mg/L)
mean + S.D.
(Range)
3.8 + 1.1
(2.2-6.6)
11.2 + 9.1
(No range data)
9.08 + 1.26
(7.31-10.57)
8.78 + 1.82
(5.36-10.83)
No data
No data
No data
19.1
(No range data)
No data
Reference
Cook etal. (1991)
Osterloh et al. (1996);
Chuwers et al. (1995)
Lee et al. (1992)
Lee et al. (1992)
Batterman etal. (1998)
Batterman and
Franzblau (1997)
Davoli etal. (1986)
Stegink etal. (1981)
Stegink etal. (1983)
Source: CERHR (2004).
3-11
DRAFT-DO NOT CITE OR QUOTE
-------
Table 3-2. Human blood methanol and formate levels following methanol exposure
Human subjects;
type of sample collected b'c
Adult males and females
administered aspartame; peak
methanol level and range of
formate levels up to 24 hr after
dosing
Infants administered
aspartame; peak exposure
level
Adult males administered
aspartame; range of peak
serum methanol levels in all
subjects
Males; postexposure samples
Males and females;
postexposure serum levels
Males without exercise;
postexposure blood methanol
and plasma formate
Males with exercise;
postexposure blood methanol
and plasma formate
Females; postexposure
samples
Exposure
route
Oral
Oral
Oral
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Exposure
duration or
method
1 dose in
juice
1 dose in
beverage
1 dose in
water
75 min
4hr
6hr
6hr
8hr
Methanol
exposure
concentration
0
3.4mg/kgbwa
10 mg/kg bwa
15 mg/kg bwa
20 mg/kg bwa
0
3.4mg/kgbwa
5 mg/kg bwa
10 mg/kg bwa
0
0.6-0.87
mg/kg bwa
0
191 ppm
0
200 ppm
0
200 ppm
0
200 ppm
0
800 ppm
Blood
methanol
mean or
range
(mg/L)
<4
12.7
21.4
25.8
<3.5
3.0
10.2
1.4-2.6
2.4-3.6
0.570
1.881
1.8
6.5
1.82
6.97
1.93
8.13
1.8
30.7
Blood
formate
mean or
range
(mg/L)
19.1
No data
No data
No data
8.4-22.8
No data
No data
3.8
3.6
11.2
14.3
9.08
8.70
8.78
9.52
No data
Reference
Stegink et al.
(1981)
Stegink et al.
(1983)
Davoli et al.
(1986)
Cook et al.
(1991)
Osterloh
et al. (1996)
Lee et al.
(1992)
Batterman
etal.(1998)
aMethanol doses resulting from intake of aspartame.
bUnless otherwise specified, it is assumed that whole blood was used for measurements.
Information about dietary restrictions is included in Table 3-1.
Source: CERHR (2004).
3-12
DRAFT-DO NOT CITE OR QUOTE
-------
Table 3-3. Monkey blood methanol and formate levels following methanol exposure
Strain-sex
Monkey; Cynomolgus;
female; mean blood
methanol and range of
plasma formate at 30 min
post daily exposure
during premating, mating,
and pregnancy
Monkey; Rhesus male;
postexposure blood level
Exposure
route
Inhalation
Inhalation
Exposure
duration
2.5 hr/day,
7days/wk
during
premating,
mating, and
gestation
(348 days)
6hr
JMethanol
exposure
concentration
o
200 ppm
600 ppm
1 800 ppm
200 ppm
1,200 ppm
2,000 ppm
Blood
methanol
mean
in mg/L
2 4
5
11
35
3.9
37.6
64.4
Blood
formate
mean
in mg/L
8 7
8.7
8.7
10
5.4-13.2
at all doses
Reference
Burbacher et al.
(2004b, 1999a)
Horton et al.
(1992)
Source: CERHR (2004).
Table 3-4. Mouse blood methanol and formate levels following methanol exposure
Species/strain/sex
Mouse;CD-l;female;
postexposure plasma
methanol and peak
formate level
Mouse;CD-l;female;
postexposure blood
methanol level
Mouse;CD-l;female;
mean postexposure plasma
methanol level
Mouse;CD-l;female;
plasma level 1 hr
postdosing
Mouse;CD-l;female; peak
plasma level
Exposure
route
Inhalation
Inhalation
Inhalation
Gavage
Oral-
Gavage
Exposure
duration
6 hron
GD8
8hr
GD6-GD15
GD6-GD15
GD8
Methanol exposure
concentration
10,000 ppm
10,000 ppm + 4-MP
15,000 ppm
2,500 ppm
5,000 ppm
10,000 ppm
15,000 ppm
0
1,000 ppm
2,000 ppm
5,000 ppm
7,500 ppm
10,000 ppm
15,000 ppm
4,000 mg/kg bw
l,500mg/kgbw
1,500 mg/kg bw +
4-MP
Blood
methanol
mean
(mg/L)
2,080
2,400
7,140
1,883
3,580
6,028
11,165
1.6
97
537
1,650
3,178
4,204
7,330
3,856
1,610
1,450
Blood
formate
mean
(mg/L)
28.5
23
34.5
No data
No data
No data
35
43
Reference
Dorman et al.
(1995)
Pollack and
Brouwer (1996);
Perkins et al.
(1995a)
Rogers et al.
(1993a)
Dorman et al.
(1995)
4-MP=4-methylpyrazole
Source: CERHR (2004).
3-13
DRAFT-DO NOT CITE OR QUOTE
-------
Table 3-5. Rat blood methanol and formate levels following methanol exposure
Species;strain/sex:
type of sample
collected
Rat; Sprague-Dawley ;
female; postexposure
blood methanol level
on 3 days
Rat;Sprague-Dawley;
female; postexposure
blood methanol level
Rat;LongEvans;female;
postexposure plasma
levelonGD7-GD12
Rat;LongEvans;female;
1 hr postexposure
blood level
Rat;Long-Evans;male
and female; 1 hr
postexposure blood
level in pups
Rat/Fischer-3 44/male ;
postexposure blood
level
Rat;Long-Evans;male;
post- exposure serum
level
Rat;Long-Evans;male;
peak blood formate
Rat;Long-Evans;male;
peak blood methanol
and formate
Exposure
route
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Oral-
gavage
Exposure
duration
7 hr/day for
19 days
8hr
7 hr/day on
GD7- GD
19
6 hr/day on
GD6-
PND21
6 hr/day on
PND1-PND
71
6hr
6hr
6hr
Single dose
Methanol exposure
concentration
5,000 ppm
10,000 ppm
20,000 ppm
1,000 ppm
5,000 ppm
10,000 ppm
15,000 ppm
20,000 ppm
0
15,000 ppm
4,500 ppm
4,500 ppm
200 ppm
1,200 ppm
2,000 ppm
200 ppm
5,000 ppm
10,000 ppm
OFS
OFS
1,200 ppm-FS
1,200 ppm-FR
2,000 ppm-FS
2,000 ppm-FR
3,500mg/kgbw-FS
3,500mg/kgbw-FP
3,500 mg/kg bw-FR
3,000 mg/kg
bw/day-FS
3, 000 mg/kg bw/day
FR
2,000 mg/kg bw/day
FS
2,000 mg/kg bw/day
FR
Blood
methanol
level in mg/L
1,000-2,170
1,840-2,240
5,250-8,650
83
1,047
1,656
2,667
3,916
2.7-1.8
3,826-3,169
555
1,260
3.1
26.6
79.7
7.4
680-873
1,468
No data
4 800
4,800
4,800
Blood
formate
level in
mg/L
No data
No data
No data
No data
No data
5.4-13.2 at
all doses
No data
80
.3
10 1
8.3
46
8.3
83
Baseline
level
382
860
9.2
9 2
538
Reference
Nelson et al.
(1985)
Pollack and
Brouwer
(1996);
Perkins et al.
(1995a)
Stanton et al.
(1995)
(1996)
Horton et al.
(1992)
Cooper et al.
(19921
Lee et al.
(1994)
FS = Folate sufficient; FR = Folate reduced; FP = Folate paire
Source: CERHR (2004).
3-14
DRAFT-DO NOT CITE OR QUOTE
-------
Table 3-6. Plasma formate concentrations in monkeys
Exposure group
Control
200 ppm
600 ppm
1800 ppm
Mean plasma formate level (mg/L) during each exposure period
Baseline
8.3
7.4
6.9
6.4
Pre-breeding
7.8
8.3
7.8
8.7
Breeding
10
9.7
9.2
11
Pregnancy
8.3
7.8
8.7
10
Source: Burbacher et al. (1999a).
Table 3-7. Serum folate concentrations in monkeys
Exposure group
Control
200 ppm
600 ppm
1800 ppm
Mean serum folate level (jig/L) during each exposure period
Baseline
14.4
11.9
12.5
12.6
Day 70
Pre-pregnancya
14.0
13.2
15.4
14.8
Day 98
Pre-pregnancya
13.4
12.9
13.4
15.3
Day 55
Pregnancy"
16.0
15.5
14.8
15.9
Day 113
Pregnancy"
15.6
13.4
16.4
15.7
aNumber of days exposed to methanol
Source: Burbacher et al. (1999a).
3.4. PHYSIOLOGIC ALLY BASED TOXICOKINETIC MODELS
1 In accordance with the needs of this human health risk assessment, particularly the
2 derivation of human health effect benchmarks from studies of the developmental effects of
3 methanol inhalation exposure in mice (Rogers et al., 1993a) and rats (NEDO, 1987) and
4 carcinogenic effects of methanol in rats exposed via drinking water (Soffritti et al., 2002a) and
5 inhalation (NEDO, 1987, 1985/2008b), mouse and rat models were developed to allow for the
6 estimation of mouse and rat internal dose metrics. A human model was developed to extrapolate
7 those internal metrics to inhalation and oral exposure concentrations that would result in the
8 same internal dose in humans (human equivalent concentrations [HECs] and human equivalent
9 doses [HEDs]). The procedures used for the development, calibration and use of these models
10 are summarized in this section, with further details provided in Appendix B, "Development,
11 Calibration and Application of a Methanol PBPK Model."
3.4.1. Model Requirements for EPA Purposes
3.4.1.1. MO A and Selection of a Dose Metric
12 Dose metrics closely associated with one or more key events that lead to the selected
13 critical effect are preferred for dose-response analyses compared to metrics not clearly
14 correlated. For instance, internal (e.g., blood, target tissue) measures of dose are preferred over
15 external measures of dose (e.g., atmospheric or drinking water concentrations), especially when,
16 as with methanol, blood methanol concentrations increase disproportionally with dose (Rogers
3-15
DRAFT-DO NOT CITE OR QUOTE
-------
1 et al., 1993a). This is likely due to the saturable metabolism of methanol in rodents. In addition,
2 respiratory and GI absorption may vary between and within species. Mode of action (MOA)
3 considerations can also influence whether to model the parent compound with or without its
4 metabolites for selection of the most adequate dose metric.
5 As discussed in Section 4, developmental effects following methanol exposures have
6 been noted in both rats and mice, but are not as evident or clear in primate exposure studies
7 (Burbacher et al., 2004a; Clary, 2003; Rogers et al., 1993a, 1993b; Andrews et al., 1987; Nelson
8 et al., 1985), and carcinogenic effects have been observed in a drinking water study of Sprague-
9 Dawley rats (Soffritti et al., 2002a) and an inhalation study of F344 rats (NEDO, 1985/2008b).
10 The report of the New Energy Development Organization (NEDO, 1987) of Japan, which
11 investigated developmental effects of methanol in rats, indicated that there is a potential that
12 developing rat brain weight is reduced following maternal and neonatal exposures. These
13 exposures included both in utero and postnatal exposures. The methanol PBPK models
14 developed for this assessment do not explicitly describe these exposure routes. Mathematical
15 modeling efforts have focused on the estimation of human equivalent external exposures that
16 would lead to internal blood levels of methanol or its metabolites presumed to be associated with
17 developmental effects as reported in rats (NEDO, 1987) and mice (Rogers et al., 1993a), and
18 carcinogenic effects as reported in rats by Soffritti et al. (2002a).
19 In a recent review of the reproductive and developmental toxicity of methanol, a panel of
20 experts concluded that methanol, not formate, is likely to be the proximate teratogen and
21 determined that blood methanol level is a useful biomarker of exposure (CERHR, 2004; Dorman
22 et al., 1995). The CERHR Expert Panel based their assessment of potential methanol toxicity on
23 an assessment of circulating blood levels (CERHR, 2004). While recent in vitro evidence
24 indicates that formaldehyde is more embryotoxic than methanol and formate (Harris et al., 2004,
25 2003), the high reactivity of formaldehyde would limit its unbound and unaltered transport as
26 free formaldehyde from maternal to fetal blood (Thrasher and Kilburn, 2001), and the capacity
27 for the metabolism of methanol to formaldehyde is likely lower in the fetus and neonate versus
28 adults (see discussion in Section 3.3). Thus, even if formaldehyde is ultimately identified as the
29 proximate teratogen, methanol would likely play a prominent role, at least in terms of transport
30 to the target tissue. Further discussions of methanol metabolism, dose metric selection, and
31 MOAissues are covered in Sections 3.3, 4.6, 4.8 and 4.9.2.
32 It has been suggested that the lymphomas observed in Sprague-Dawley rats following
33 methanol exposure are associated with formaldehyde because formaldehyde and other
34 compounds that metabolize to formaldehyde have been reported to cause lymphomas in Sprague-
35 Dawley rats (Soffritti et al., 2005). Given the reactivity of formaldehyde, models that predict
36 levels of formaldehyde in the blood are difficult to validate. However, production of
37 formaldehyde or formate following exposure to methanol can be estimated by summing the total
3-16 DRAFT-DO NOT CITE OR QUOTE
-------
1 amount of methanol cleared by metabolic processes.9 This metric of formaldehyde or formate
2 dose has limited value since it ignores important processes that may differ between species, such
3 as clearance of these two metabolites, but it can be roughly be equated to the total amount of
4 metabolites produced and may be the more relevant dose metric if formaldehyde is found to be
5 the proximate toxic moiety. Thus, both blood methanol and total metabolism metrics are
6 considered to be important components of the PBPK models. Dose metric selection and MOA
7 issues are discussed further in Sections 3.3, 4.6, 4.8 and 4.9.2.
3.4.1.2. Criteria for th e Developm en t ofMethan ol PBPK Models
8 The development of methanol PBPK models that would meet the needs of this
9 assessment was organized around a set of criteria that reflect: 1) the MOA(s) being considered
10 for methanol; 2) absorption, distribution, metabolism, and elimination characteristics; 3) dose
11 routes necessary for interpreting toxicity studies or estimating HECs; and 4) general parameters
12 needed for the development of predictive PK models.
13 The criteria with a brief justification are provided below:
14 1) Must simulate blood methanol concentrations and total methanol metabolic clearance.
15 Blood methanol is the recommended dose metric for developmental effects, but total
16 metabolic clearance may be a useful metric, particularly for cancer endpoints.
17 2) Must be capable of simulating experimental blood methanol and total metabolic
18 clearance (mg/day) data for the inhalation route of exposure in mice and rats (a) and
19 humans (b), and the oral route in rats (c) and humans (d). These routes are important for
20 determining dose metrics in the most sensitive test species under the conditions of the
21 toxicity study and in the relevant exposure routes in humans.
22 3) The model code should easily allow designation of respiration rates during inhalation
23 exposures. A standard variable in inhalation route risk assessments is ventilation rate.
24 Blood methanol concentrations will depend strongly on ventilation rate, which varies
25 significantly between species.
26 4) Must address the potential for saturable metabolism of methanol. Saturable metabolism
27 has the potential to bring nonlinearities into the exposure:tissue dose relationship.
28 5) Model complexity should be consistent with modeling needs and limitations of the
29 available data. Model should adequately describe the biological mechanisms that
30 determine the internal dose metrics (blood methanol and metabolic clearance) to assure
31 that it can be reliably used to predict those metrics in exposure conditions and scenarios
32 where data are lacking. Compartments or processes should not be added that cannot be
33 adequately characterized by the available data.
9 This assumption is more likely to be appropriate for formaldehyde than formate as formaldehyde is a direct
metabolite of methanol.
3-17 DRAFT-DO NOT CITE OR QUOTE
-------
1 Although the rat and mouse models are useful for the evaluation of the dose metrics
2 associated with methanol's developmental effects and the relevant toxicity studies, including
3 gestational exposures, no pregnancy-specific PBPK model exists for methanol, and inadequate
4 data exists for the development and validation of a fetal/gestational/conceptus compartment.
5 However, EPA determined that nonpregnancy models for the appropriate species and routes of
6 exposure could prove to be valuable because levels of methanol in NP, pregnant and fetal blood
7 are expected to be similar following the same oral or inhalation exposure. Pollack and Brouwer
8 (1996) determined that methanol distribution in rats and mice following repeated oral and i.v.
9 exposures up to day 20 of gestation is "virtually unaffected by pregnancy, with the possible
10 exception of the immediate perinatal period." The critical window for methanol induction of
11 cervical rib malformations in CD-I mice has been identified as occurring between GD6 and GD7
12 (Rogers and Mole, 1997; Rogers et al., 1993b), a developmental period roughly equivalent to
13 week 3 of human development (Chernoff and Rogers, 2004). Methanol blood kinetics measured
14 during and after inhalation exposure in NP and pregnant mice on GD6-GD10 and GD6-GD15
15 (Perkins et al., 1996, 1995a; Dorman et al., 1995; Rogers et al., 1993a) are also similar. Further,
16 the available data indicate that the maternal blood:fetal partition coefficient is approximately 1 at
17 dose levels most relevant to this assessment (Ward et al., 1997; Horton et al., 1992). The same
18 has been found in rat (Zorzano et al., 1989; Guerri and Sanchis, 1985) and sheep (Brien et al.,
19 1985; Cumming et al., 1984) studies of ethanol, a structurally related chemical that also
20 penetrates cellular membranes readily and distributes throughout total body water.
21 Consequently, fetal methanol concentrations are expected to be roughly equivalent to that in the
22 mother's blood. Thus, pharmacokinetics and blood dose metrics for NP mice and humans are
23 expected to provide reasonable approximations of pregnancy levels and fetal exposure,
24 particularly during early gestation, that improve upon default estimations from external exposure
25 concentrations.
3.4.2. Methanol PBPK Models
26 As has been discussed, methanol is well absorbed by both inhalation and oral routes and
27 is readily metabolized to formaldehyde, which is rapidly converted to formate in both rodents
28 and humans. As was discussed in Section 3.1, the enzymes responsible for metabolizing
29 methanol are different in rodents and humans. Several rat, mouse and human PBPK models
30 which attempt to account for these species differences have been published (Fisher et al., 2000;
31 Ward et al., 1997; Perkins et al., 1995a; Horton et al., 1992). In addition, a gestational model for
32 a similar water soluble compound, isopropanol, with the potential to be adapted to methanol
33 pharmacokinetics, was of interest (Gentry et al., 2003, 2002; Clewell et al., 2001). Three PK
34 models (Gentry et al., 2002; Bouchard et al., 2001; Ward et al., 1997) were identified as
35 potentially appropriate for use in animal-to-human extrapolation of methanol metabolic rates and
36 blood concentrations. An additional methanol PBPK model by Fisher et al. (2000) was
3-18 DRAFT-DO NOT CITE OR QUOTE
-------
1 considered principally because it had an important feature - pulmonary compartmentalization
2 (see below for details) - worth adopting in the final model.
3.4.2.1. Wardetal. (1997)
3 The PBPK model of Ward et al. (1997) describes inhalation, oral and i.v. routes of
4 exposure and is parameterized for both NP and pregnant mice and rats (Table 3-8). The model
5 has not been parameterized for humans.
6 Respiratory uptake of methanol is described as a constant infusion into arterial blood at a
7 rate equal to the minute ventilation times the inhaled concentration and includes a parameter for
8 respiratory bioavailability, which for methanol is <100%. This simple approach is nonstandard
9 for volatile compounds but is expected to be appropriate for a compound like methanol, for
10 which there is little clearance from the blood via exhalation. Oral absorption is described as a
11 biphasic process, dependent on a rapid and a slow first-order rate constant. This is conceptually
12 similar to the isopropanol model discussed below (Gentry et al., 2002; Clewell et al., 2001),
13 which also employs slow and fast absorption processes but functionally separates them into
14 stomach and duodenal compartments.
15 Methanol elimination in the Ward et al. (1997) model is primarily via saturable hepatic
16 metabolism. The parameters describing this metabolism come from the literature, primarily
17 previous work by Ward and Pollack (1996) and Pollack et al. (1993). A first-order elimination of
18 methanol from the kidney compartment includes a lumped metabolic term that accounts for both
19 renal and pulmonary excretion.
20 The model adequately fits the experimental blood kinetics of methanol in rat and mice
21 and is therefore suitable for simulating blood dosimetry in the relevant test species and routes of
22 exposure (oral and i.v.). The Ward et al. (1997) model meets criteria 1, 2a, 2c, 3, 4, and 5. The
23 most significant limitation is the absence of parameters for the oral and inhalation routes in the
24 human. A modified version of this model that includes human parameters and a standard PBPK
25 lung compartment might be suitable for the purposes of this assessment.
3.4.2.2. Bouchard et al. (2001)
26 The Bouchard et al. (2001) model is not actually a PBPK model but is an elaborate
27 classical PK model, since the transfer rates are not determined from blood flows, ventilation,
28 partition coefficients, and the like. The Bouchard et al. (2001) model uses a single compartment
29 for methanol: a central compartment represented by a volume of distribution where the
30 concentration is assumed to equal that in blood. The model was developed for inhalation and i.v.
31 kinetics only. Methanol is primarily eliminated via saturable metabolism. The model adequately
32 simulates blood kinetics in NP rats and humans following inhalation exposure and in NP rats
33 following i.v. exposure; there is no description for oral absorption. Because methanol distributes
34 with total body water (Ward et al., 1997; Horton et al., 1992), this simple model structure is
35 sufficient for predicting blood concentrations of methanol following inhalation and i.v. dosing.
3-19 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
The Bouchard et al. (2001) model has the advantage of simplicity, reflecting the
minimum number of compartments necessary for representing blood methanol pharmacokinetics.
Because volume of distribution can be easily and directly estimated for water-soluble compounds
like methanol or fit directly to experimental kinetics data, concern over the scalability of this
parameter is absent. The model has been parameterized for a required human exposure route,
inhalation (Table 3-8). The model meets criteria 1, 2b, 3, 4, and 5 described in Section 3.4.1.2.
However, the Bouchard model has specific and significant limitations. The model has neither
been parameterized for the mouse, a test species of concern (Table 3-8), nor for the oral route in
humans. As such, the model cannot be used to conduct the necessary interspecies extrapolation.
Table 3-8. Routes of exposure optimized in models - optimized against blood
concentration data
Route
i.v.
Inhalation
Oral
Ward et al.
Mouse
P/NP
P/NP
P/NP
Rat
P/NP
-
NP
Human
-
-
-
Bouchard et al.
Mouse
-
-
-
Rat
NP
NP
-
Human
-
NP
-
10
11
12
13
14
15
16
17
18
19
20
21
22
23
P = Pregnant
NP = Nonpregnant
Source: Bouchard et al. (2001); Ward et al. (1997).
3.4.2.3. Gentry et al. (2003, 2002) and Clewell et al. (2001)
The model described in these three papers is for isopropanol, not methanol, and therefore
lacks any immediately useful parameterization for the purposes of a methanol risk assessment.
Although the overall model structure, the description of kinetics for both parent compound and
primary metabolite, gestational compartments, lactational transfer, oral and i.v. routes, etc., are
attractive for application to methanol, this model is not ideal. In particular, the model structure is
more elaborate than necessary; because methanol partition coefficients are near 1 for all tissues
except fat, there is no need to individually represent these tissues. Similarly, a fetal compartment
may not be necessary because methanol kinetics in the fetus (conceptus) is expected to parallel
maternal blood concentrations in the rodent. However, even if a fetal model was considered
necessary, other than the partition coefficient, there are insufficient data to identify conceptus
compartment parameters for methanol. This model would require the most modification and
parameterization to be useful for methanol risk assessment since parameters would have to be
estimated for all relevant species (at least mouse and humans) and for several routes of exposure.
Therefore the isopropanol model was not considered further.
3-20
DRAFT-DO NOT CITE OR QUOTE
-------
3.4.3. Selected Modeling Approach
1 As discussed earlier regarding model criteria, fetal methanol concentrations can
2 reasonably be assumed to equal maternal blood concentration. Thus, as discussed in Section
3 3.4.2.3, methanol pharmacokinetics and blood dose metrics for NP mice and humans are
4 expected to improve upon default extrapolations from external exposures as estimates of fetal
5 exposure during early gestation. The same level of confidence cannot be placed on the whole-
6 body rate of metabolism, in particular as a surrogate for formaldehyde dose. Because of
7 formaldehyde's reactivity and the limited fetal metabolic (ADH) activity (see Sections 3.3 and
8 4.10.1), fetal formaldehyde concentration increases (from methanol) will probably not equal
9 maternal increases in formaldehyde concentration. But since there is no model that explicitly
10 describes formaldehyde concentration in the adult, let alone the fetus, the metabolism metric is
11 the closest one can come to predicting fetal formaldehyde dose. This metric is expected to be a
12 better predictor of formaldehyde dose than applied methanol dose or even methanol blood levels,
13 which do not account for species differences in conversion of methanol to formaldehyde.
14 Most of the published rodent kinetic models for methanol describe the metabolism of
15 methanol to formaldehyde as a saturable process but differ in the description of metabolism to
16 and excretion of formate (Bouchard et al., 2001; Fisher et al., 2000; Ward et al., 1997). The
17 model of Ward et al. (1997) used one saturable and one first-order pathway to describe methanol
18 elimination in mice. The saturable pathway described in Ward et al. (1997) can specifically be
19 ascribed to metabolic formation of formaldehyde in the liver, while the renal first-order
20 elimination described in the model represents nonspecific clearance of methanol (e.g.,
21 metabolism, excretion, or exhalation). The model of Ward et al. (1997) does not describe
22 kinetics of formaldehyde subsequent to its formation and does not include any description of
23 formate.
24 Bouchard et al. (2001) employed a metabolic pathway for conversion of methanol to
25 formaldehyde and a second pathway described as urinary elimination of methanol in rats and
26 humans. They then explicitly describe two pathways of formaldehyde transformation, one to
27 formate and the other to "other, unobserved formaldehyde byproducts." Finally, formate
28 removal is described by two pathways, one to urinary elimination and one via metabolism to CO2
29 (which is exhaled). All of these metabolic and elimination steps are described as first-order
30 processes, but the explicit descriptions of formaldehyde and formate kinetics significantly
31 distinguish the model of Bouchard et al. (2001) from that of Ward et al. (1997), which only
32 describes methanol.
33 There are two other important distinctions between the Ward et al. (1997) and Bouchard
34 et al. (2001) models. The former is currently capable of simulating blood data for all exposure
35 routes in mice but not humans, while the latter is capable of simulating human inhalation route
36 blood pharmacokinetics but not those in mice. The Ward et al. (1997) model has more
3-21 DRAFT-DO NOT CITE OR QUOTE
-------
1 compartments than is necessary to adequately represent methanol disposition but has been fit to
2 PK data in pregnant and NP mice for all routes of exposure (i.v., oral, and inhalation). The Ward
3 et al. (1997) model has also been fit to i.v. and oral route PK data in rats. Based primarily on the
4 extensive amount of fitting that has already been demonstrated for this model, it was determined
5 that a modified Ward et al. (1997) model, with the addition of a lung compartment as described
6 by Fisher et al. (2000), should be used for the purposes of this assessment. See Appendix B for a
7 more complete discussion of the selected modeling approach and modeling considerations.
3.4.3.1. Available PK Data
8 Although limited human data are available, several studies exist that contain PK and
9 metabolic data in mice, rats, and nonhuman primates for model parameterization. Table 3-9
10 contains references that were used to verify the model fits as reported in Ward et al. (1997).
Table 3-9. Key methanol kinetic studies for model validation
Reference
Battermanetal., 1998
Batterman & Franzblau 1997
Burbacher, 2004a, 2004b
Osterlohetal.,1996;
drawers etal., 1995;
D'Alessandro etal., 1994
Medinsky et al., 1997;
Dormanetal., 1994
Gonzalez-Quevedo et al.,
2002
Hortonetal., 1992
Perkins et al., 1996, 1995a,
1995b
Pollack and Brouwer, 1996;
Pollack etal., 1993
Rogers and Mole, 1997;
Rogers etal., 1993a;
Sedivec et al., 1981
Ward et al., 1997; Ward and
Pollack, 1996
i.v. dose
(mg/kg)
100 (rats
only)
100-2,500
100, 500
(Rat), 2,500
(Mouse)
Inhalation
(ppm)
800 (8 hr)
0-1,800 (2.5 hr, 4
mo)
200 (4 hr)
10-900 (2 hr)
50-2,000 (6 hr)
1,000-20,000(8
hr)
1,000-20,000(8
hr)
1,000-15,000(7
hr, 10 days)
78-231 (8hr)
Oral/dermal/
IP
Dermal
IP: 2 mg/kg-
day, 2 wk
Oral:
100-2,500
mg/kg
Oral:
2,500 mg/kg
Species
Human Male/female
Monkeys
Cynomolgus
Pregnant, NP
Human Male/female
Monkeys
Cynomolgus Folate
deficient
Rat
Rat & Monkey
Rhesus
Mouse and Rat
Rat: Sprague-
Dawley, & Mouse;
CD-I Pregnant, NP
Mouse
Pregnant
Human
GDI 8 Mouse,
GDI 4 & GD20 Rats
Samples
Blood
Blood,
urine,
exhaled
Blood
Blood, urine
Blood,
urine,
exhaled
Blood
Blood,
urine,
exhaled
Blood, urine
Blood
Blood
Urine,
blood
Blood,
conceptus
Digitized
figures"
Figure 1
Figure 1,
Osterloh
etal., 1996
Figure 7
Figures 2, 3,
6,7,8
adata obtained from the reported figure
3-22
DRAFT-DO NOT CITE OR QUOTE
-------
3.4.3.2. Model Structure
1 A model was developed which includes compartments for alveolar air/blood methanol
2 exchange, liver, fat, bladder (human simulations) and the rest of the body (Figure 3-4). This
3 model is a revision of the model reported by Ward et al. (1997), reflecting significant
4 simplifications (removal of compartments for placenta, embryo/fetus, and extraembrionic fluid)
5 and three elaborations (addition of an intestine lumen compartment to the existing stomach
6 lumen compartment, use of a saturable rate of absorption from the stomach (but not intestine),
7 and addition of a bladder compartment which impacts simulations for human urinary excretion),
8 while maintaining the ability to describe methanol blood kinetics in mice, rats, and humans. A
9 fat compartment was included because it is the only tissue with a tissue:blood partitioning
10 coefficient appreciably different than 1, and the liver is included because it is the primary site of
11 metabolism. A bladder compartment was also added for use in simulating human urinary
12 excretion to capture the difference in kinetics between changes in blood methanol concentration
13 and urinary methanol concentration. The model code describes inhalation, oral, and i.v. dose
14 routes, and data exist (Table 3-9) that were used to fit parameters and evaluate model predictions
15 for all three of those routes in both mice and rats. In humans, inhalation exposure data were
16 available for model calibration and validation but not oral or i.v. data. However, oral exposures
17 were simulated in humans, assuming a continuous, zero-order ingestion rate, thereby obviating
18 the need for oral uptake parameters.
19 PK data from exposure routes other than inhalation and oral were used to test or further
20 refine the parameters for methanol clearance. Monkey data were evaluated for insight into
21 primate kinetics. Data from Osterloh et al. (1996) and Sedivec et al. (1981) were used to validate
22 the modified Ward et al. (1997) model parameterization for humans. The fact that optimized
23 human parameters were similar to those predicted in monkeys was important to the validation
24 process (Bourchard et al., 2001)(see section 3.4.7 and Appendix B). Blood levels of methanol
25 have been reported following i.v., oral, and inhalation exposure in rats and mice and inhalation
26 exposure in nonhuman primates and humans.
27 The metabolic clearance of methanol was represented in mice, rats, and humans by
28 specifying separate rate constants for the species-specific enzymes: two saturable processes for
29 mice and rats10 and one for humans. The requirement for two saturable processes in the mouse
30 and rat models may reflect saturation of CAT and ADH1. Simulated methanol clearance by these
31 metabolic processes is not linked to production of formaldehyde or formate; it is simply cleared
32 from the methanol mass balance leaving the system. Metabolism of formaldehyde is not
33 explicitly simulated by the model, and this model tracks neither formate nor formaldehyde.
34 Since the metabolic conversion of formaldehyde to formate is rapid (<1 minute) in all species
10 The need for two saturable metabolic pathways in the mouse model was confirmed through simulation and
optimization. High exposure (>2,000 ppm methanol) and low exposure (1,000 ppm methanol) blood data could not
be fit visually, or by more formal optimization, without the second saturable metabolic pathway.
3-23 DRAFT-DO NOT CITE OR QUOTE
-------
1 (Kavet and Nauss, 1990), the methanol clearance rate may approximate a formate production
2 rate, though this has not been verified.
3
Inhalal
Expos
RlnHHar ^ K1
Diauuel ^ —
KBL| (human only)
Urine IV-"*
KLLor
Vm^Y _-
tion
ure
^
Venous Blood
_^— —
f
4
^
rf"
^ — '
Alveolar Air
Lung Blood
Body
Fat
Liver
-»E;
*
xha
Arterial Blood
^
Vmax2
Km2
Metabolites
Kfec
/KAI
——~~*,
Intestine
^KASorVmAS.KMAS
i*1
Stomach
Oral
Exposure
Feces
(rat only)
Figure 3-4. Schematic of the PBPK model used to describe the inhalation, oral, and
i.v. route pharmacokinetics of methanol. KAS, first-order oral absorption rate from
stomach; VmAS and KMAS, Michaelis-Menten rate constants for saturable
absorption from stomach; KAI, first-order uptake from the intestine; KSI,
first-order transfer between stomach and intestine; Vmax and Km and Vmax2 and
Km2, Michaelis-Menten rate constants for high affinity/low capacity and low
affinity/high capacity metabolic clearance of methanol; KLL, alternate first-
order rate constant; KBL, rate constant for urinary excretion from bladder.
Both metabolic pathways were used to describe methanol clearance in the
mouse, while a single metabolic pathway describes clearance in the human.
4
5 The primary purpose of this assessment is for the determination of noncancer and cancer
6 risk associated with increases in the levels of methanol or its metabolites (e.g., formate,
7 formaldehyde). Thus, the focus of model development is on obtaining accurate predictions of
8 increased body burdens over background. The PBPK models do not describe or account for
9 background levels of methanol, formaldehyde or formate, and background levels were subtracted
10 from the reported data before use in model fitting or validation (if not already subtracted by
11 study authors), as described below. This approach is not expected to have a significant impact on
12 PBPK model parameter estimates as background levels of methanol and its metabolites are low
3-24
DRAFT-DO NOT CITE OR QUOTE
-------
1 relative to exposure levels used in methanol bioassays. Further, while it is possible that
2 background levels of methanol or its metabolites contribute to background responses for some
3 adverse effects, the results of dose-response modeling of cancer endpoints using "background
4 dose" models suggest that this contribution is relatively small (see discussion in Appendix E,
5 Section E.4).
3.4.3.3. Model Par am eters
6 The EPA methanol model uses a consistent set of physiological parameters obtained
7 predominantly from the open literature (Table 3-10); the Ward et al. (1997) model employed a
8 number of data-set specific parameters.11 Parameters for blood flow, ventilation, and metabolic
9 capacity were scaled as a function of body weight raised to the 0.75 power, according to the
10 methods of Ramsey and Andersen (1984).
Table 3-10. Parameters used in the mouse, rat and human PBPK models
Body weight (kg)
Mouse
0.03"
Rat
SD | F344
0.2756
Human
70
Source
Measured/estimated
Tissue volume (% body weight)
Liver
Blood arterial
venous
Fat
Lung
Rest of body
5.5
1.23
3.68
7.0
0.73
72.9
3.7
1.85
4.43
7.0
0.50
73.9
2.6
1.98
5.93
21.4
0.8
58.3
Brown etal. 1997
Calculated'
Flows (L/hr/kg0'75)
Alveolar ventillation''
Cardiac output
25.4
25.4
16.4
16.4
16.5
24.0
Perkins et al. 1995a; Brown et al. 1997;
U.S. EPA, 2004
Percentage of cardiac output
Liver
Fat
Rest of body
25.0
5.0
70.0
Biochemical constants8
VmaxC (mg/hr/kg075)
Km (mg/L)
Vmax2C (mg/hr/kg075)
Km2 (mg/L)
K1C (BW025/hr)
19
5.2
3.2
660
NA
25.0
7.0
68
5.0 0
6.3 NA
8.4 22.3
65 100
NA
22.7
5.2
72.1
1s*
, saturable
order
NA 33.1
NA 23.7
NA
NA
0.0373 0.0342
Brown etal. 1997
Calculated
Fitted
11 Some data sets provided in the Ward et al., (1997) model code were corrected to be consistent with figures in the
published literature describing the experimental data.
3-25
DRAFT-DO NOT CITE OR QUOTE
-------
KLLC (BW025/hr/
Mouse
NA
Rat
SD | F344
NA
Human
95.7 NA
Source
Oral absorption
VmASC (mg/hr/kg075)
KMASC (mg/kg)
KSI (hr"1)
KAI (hr1)
Kfec (hr'1)
1830
620
2.2
0.33
0
5570
620
7.4
0.051
0.029
377
620
3.17
3.28
0
Mouse and rat fitted (mouse and human
KMASC assumed = rat); other human
values are those for ethanol from Sultatos et
al. (2004), with VmASC set so that for a
70-kg person VmAS/KMAS = the first-
order constant of Sultatos et al.
Partition coefficients
LiverBlood
Fat:Blood
Blood:Air
Body:Blood
Lung:Blood
KBL (hr1), bladder time-
constant J
FRACIN (%), nhalation
fractional availability
1.06
0.083
1350'
0.66
1
1.06
0.083
1350
0.66
1
NA
0.665
0.20
0.583*
0.142
1626
0.805
1.07
0.564 0.612
0.866*
Ward et al., 1997; Fiserova-Bergerova and
Diaz, 1986
Horton et al., 1992; Fiserova-Bergerova and
Diaz, 1986
Rodent: estimated; human: Fiserova-
Bergerova and Diaz, 1986 (human "body"
assumed = muscle)
Fitted (human)
Rodent: fitted; human
Ernstgard et al., 2005
NA - Not applicable for that species
"Both sources of mouse data report body weights of approximately 30 g
*The midpoints of rat weights reported for each study was used and ranged from 0.22 to 0.33 kg
The volume of the other tissues was subtracted from 91% (whole body minus a bone volume of approximately 9%) to get
the volume of the remaining tissues
''Minute ventilation was measured and reported for much of the data from Perkins et al. (1996) and the average alveolar
ventilation (estimated as 2/3 minute ventilation) for each exposure concentration was used in the model. When ventilation
rates were not available, a mouse QPC (Alveolar Ventilation/BW°75) of 25.4 was used (average from Perkins et al., 1995).
The QPC used to fit the human data was obtained from U.S. EPA (2004). This QPC was somewhat higher than calculated
from Brown et al. (1997) (-13 L/hr/kg075)
eVmax, Km, and Vmax2, Km2 represent the two saturable metabolic clearance processes assumed to occur solely in the liver.
The Vmax used in the model = VmaxC (mg/kg°75-hr) XBW075. K1C is the first-order loss from the blood for human
simulations that represents urinary elimination. Allometric scaling for first-order clearance processes was done as previously
described (Teeguarden et al., 2005); The Kl used in the model= K1C / BW°25
fKLLC - alternate human first-order metabolism rate (used only when VmaxC = Vmax2C = 0)
^uman oral simulations used a zero order dose rate equal to the mg/kg-day dose
*Human liverblood estimated from correlation to (measured) fatblood, based on data from 28 other solvents
7Rat partition coefficient used for mice as done by Ward et al. (1997)
JKBL - a first-order rate constant for clearance from the bladder compartment, used to account for the difference between
blood kinetics and urinary excretion data as observed in humans
*For human exposures, the fractional availability was from Sedivec et al. (1981), corrected for the fact that alveolar
ventilation is 2/3 of total respiration rate
3.4.4. Mouse Model Calibration and Sensitivity Analysis
1 The process by which the mouse, rat, and human inhalation and oral models were
2 calibrated is discussed in more detail in Appendix B, "Development, Calibration and Application
3 of a Methanol PBPK Model." The calibrated mouse inhalation model predicted blood methanol
4 blood concentration time-course agreed well with measured values in adult mice in the critical
3-26 DRAFT-DO NOT CITE OR QUOTE
-------
1 inhalation studies of Rogers and Mole (1997) (Figure 3-5), Perkins et al. (1995a, 1995b), and
2 Rogers et al. (1993a), as well as in NP and early gestation (GD8) mice of Dorman et al. (1995)
3 (Figure 3-6). Parameter values used in the calibrated model are given in Table 3-10.
4 The mouse model was also calibrated for the oral route by fitting all but one of the rate
5 constants for oral uptake of methanol to the oral-route blood methanol kinetics of Ward et al.
6 (1997, 1995). The best model fit to the mouse oral route blood methanol PK data was obtained
7 using a two-compartment GI tract model, as depicted in Figure 3-4. Because the oral data in rats
8 led to the conclusion that a saturable rate of uptake from the stomach lumen was necessary (see
9 section 3.4.5), the same equation was used for uptake in the mouse. But attempts to identify the
10 uptake saturation constant, KMASC, from the mouse data were unsuccessful; therefore KMASC
11 for the mouse was set equal to the value obtained for rats. Adjusting the other mouse oral uptake
12 parameters gave an adequate fit to those data. This calibration allows inhalation to oral dose-
13 route extrapolations in the mouse, which can then be extrapolated to identify human oral route
14 exposures equivalent to mouse inhalation exposures (if equivalent human exposures exist).
10,000
1,000
I
o
0.1
o
o
m
100
10
— 1000 pprn sim
— 2000 ppm sim
— 5000 ppm sim
— 7500 ppm sim
— 10000 pprn sim
15000 ppm sim
i 1000 ppm data (GD6 & 10)
• 2000 ppm data (GD6 & 10)
• 5000 ppm data (GD6 & 10)
* 7500 ppm data (GD6 & 10)
A 10000 pprn data (GD6 & 10)
T 10000 ppm data (GD7)
p 15000 ppm data (GD6 & 10) J
12 15
Time (hr)
IB
21
24
27
Figure 3-5. Model fits to data sets from GD6, GD7, and GD10 mice for 6- to 7-hour
inhalation exposures to 1,000-15,000 ppm methanol. Maximum concentrations are
from Table 2 in Rogers et al. (1993a). The dataset for GD7 mice exposed to
10,000 ppm is from Rogers and Mole (1997) and personal communication.
Symbols are concentration + SEM of a minimum of N=4 mice/concentration.
Default ventilation rates (Table 3-10) were used to simulate these data.
Source: Rogers and Mole (1997); Rogers et al. (1993a)
3-27
DRAFT-DO NOT CITE OR QUOTE
-------
8,000-
7,000-
J 6,000-
o>
E, 5,000-
O 4,000-
v
* 3,000-
o
I 2-ooo^
1,000-
0-
+ 2,500 ppm NP mouse
A 5,000 ppm IMP mouse
*_t ^r fW V |J- |-r i i i • • iff v> ^^v
o 10,000 ppm NP mouse
n 6 hr, lbrOOO ppm GD 8 mouse (Dorman)
15 20
Time (hr)
35
Figure 3-6. Simulation of inhalation exposures to methanol in NP mice from Perkins
et al. (1995a) (8-hour exposures) and GD8 mice from Dorman et al. (1995)(6-hour
exposures). Data points are measured blood methanol levels and lines represent
PBPK model simulations. Digitizlt (Sharlt! Inc., Greensburg, PA) was used to
digitize data from Figure 2 of Perkins et al. (1995a) and Figure 2 from Dorman
et al. (1995). Default ventilation rates (Table 3-10) were used to simulate the
Dorman data. The alveolar ventilation rate for each data set from Perkins et al.
(1995a) was set equal to the measured value reported in that manuscript. For
the 2,500, 5,000, and 10,000 ppm exposure groups, the alveolar ventilation rates
were 29, 24, and 21 (L/hours/kg0'75), respectively. The cardiac output for these
simulations was set equal to the alveolar ventilation rate.
Source: Dorman et al. (1995); Perkins et al. (1995a).
1 The parameterization of methanol clearance (high-and low-affinity metabolic pathways)
2 was also verified by simulation of datasets describing the pharmacokinetics of methanol
3 following i.v. administration. The results of this calibration of the methanol PBPK model are
4 described in Appendix B and were generally consistent with both the available inhalation and
5 oral-route data. Up to 20 hours postexposure, blood methanol kinetics appears similar for NP
6 and pregnant mice. However, some data suggests that clearance in GDIS mice is slower than in
7 NP and earlier in gestation (GD10 and less), particularly beyond 20 hours postexposure (see the
8 i.v. and oral data of Ward et al. [1997] in Appendix B).
9 Intravenous-route blood methanol kinetic data in NP mice were only available for a
10 single i.v. dose of 2,500 mg/kg, but were available for GDIS mice following administration of a
11 broader range of doses: 100, 500, and 2,500 mg/kg. The i.v. maternal PK data in GDIS mice
12 appeared to show an unexpected dose-dependent nonlinearity in initial blood concentrations.
13 Before discussing the nonlinearity, it is first noted that data values used here were obtained from
14 a computational "command file" provided by Ward et al. (1997). These values appear to be
3-28 DRAFT-DO NOT CITE OR QUOTE
-------
1 consistent with the plots in their publication but are inconsistent with some of the values in their
2 Table 6 (Ward et al., 1997). In particular, the initial maternal blood concentration (i.e., the Cmax)
3 after the 2,500 mg/kg i.v. is listed as 4,250 mg/L in their command file but as 3,251 mg/L in their
4 published table. The corresponding data point in their Figure 5 A is distinctly centered above
5 4,000 mg/L (digitizing yields 4,213 mg/L), and so must be 4,250 rather than 3,251 mg/L.
6 Therefore the data values listed in the command file were used in the subsequent analysis, rather
7 than those in the published table.
8 After i.v. dosing the ratio of the administered doses to the first concentrations measured
9 by Ward et al. (1997) (5-minute time points) were 0.588 L/kg, 0.585 L/kg, and 0.397 L/kg at
10 doses of 2,500, 500, and 100 mg/kg, respectively. The discrepancy between the first two values
11 and the third value suggests either a dose dependence in the Vd or some source of experimental
12 variability.12 It may be that Vd, which is not impacted by any other PBPK parameters and is only
13 determined by the biochemical partitioning properties of methanol, is 1.5-fold lower at 100
14 mg/kg than at the higher concentrations, while the Vd at 500 and 2,500 mg/kg are exactly as
15 predicted by the PBPK model without adjustment. However, it was found that the PBPK model,
16 obtained with measured partition coefficients and otherwise calibrated to inhalation data, could
17 adequately fit the data at the nominal dose of 100 mg/kg without other parameter adjustment
18 simply by simulating a dose of 200 mg/kg, as shown in Figure B-5. The fact that the alternate
19 dose (200 mg/kg) differs by a factor of 2 from the nominal dose suggests that the data could also
20 be the result of a simple dilution error in dose preparation. If the first two of the
21 dose/concentration values were not virtually identical (0.588 and 0.585 L/kg), but instead the
22 500 mg/kg value was more intermediate between those for 2,500 and 100 mg/kg, then a regular
23 dose dependence in Vd would seem more likely. However, based on these values, the U.S. EPA
24 has concluded that the apparent dose dependency is probably the result of a dosing error and
25 therefore, that dose-dependent parameter changes (e.g., in the partition coefficients) should not
26 be introduced in an attempt to otherwise better fit these data.
27 Further, the nominal "nonlinearity" between the maternal blood and conceptus shown in
28 Figure 8 of Ward et al. (1997) is the result of those data being plotted on a log-y/linear-x scale.
29 Replotting the data from Tables 5 and 6 (using the value of 4,250 mg/L from the command file as
30 the GDIS maternal Cmax for the 2,500 mg/kg) shows the results to be linear, especially in the
31 low-dose region which is of the most concern (Figure 3-7). Therefore, the current model uses a
32 consistent set of parameters that are not varied by dose and fit the 2,500 and 500 mg/kg i.v. data
33 adequately, although they do not fit the 100 mg/kg i.v. data unless, as noted above, a presumed
34 i.v. dose of 200 mg/kg is employed. With that exception, both the single set of parameters used
35 herein and the assumption that maternal blood methanol is a good metric of fetal exposure are
36 well supported by the data.
12 It is possible that Ward et al., (1997) were unaware of that discrepancy because they plotted the results for each
dose in separate figures, and it only becomes obvious when all the data and simulations are plotted together.
3-29 DRAFT-DO NOT CITE OR QUOTE
-------
10000 -
re
1,1000-
Conceptus AUC (
o 8
A
•„•- ' D
{
k
>
) 1000 2000 3000
" \
+ rat
• mouse
-y = x
4000 5000
Maternal AUC (ug/mL-day)
_ 3500 -
•§ 3000
i 2500
3
y 2000
< 1500
(A
a. 1000
8
§ 500
0
(
B
^
/
n/
) 1000 2000 3000
D
> rat
• mouse
-y = x
4000 5000
Maternal AUC (ug/mL-day)
1
2
3
4
Figure 3-7. Conceptus versus maternal blood AUC values for rats and mice plotted
(A) on a log-linear scale, as in Figure 8 of Ward et al. (1997), and (B) on a linear-linear
scale. In both panels the line y = x is plotted (dashed line) for comparison. Thus the
apparent "nonlinear" relationship indicated by Ward et al. (1997) is seen to be
primarily a simple artifact of the choice of axes. However, as evident in panel
B, there appears to be some nonlinearity at the two highest doses in the mouse
(results of 2,500 mg/kg i.v. in GDIS mice and 15,000 ppm exposure to GD8
mice), where distribution from the dam to the conceptus is below 1:1.
Source: Ward et al. (1997).
To summarize the mouse model calibration: using the single set of parameters listed for
the mouse in Table 3-10, the PBPK model has been shown to adequately fit or reproduce
methanol PK data from a variety of laboratories and publications, including both NP mice and
pregnant mice up to GD10. Two saturable metabolic pathways are thus described by the model
3-30
DRAFT-DO NOT CITE OR QUOTE
-------
1 and supported by the data. Also, it is thereby demonstrated that a model based on NP mouse
2 physiology adequately describes (predicts) dosimetry in the pregnant mouse dam through GD10.
3 Finally, as illustrated in Figure 3-7b, methanol PK in the conceptus and dam of both mice
4 (including lower doses at GDIS) and rats (GDI4 and GD20) are virtually identical, except for
5 the very highest doses in mice. Thus the existing model appears to be adequate for predicting
6 internal methanol doses, including fetal exposures, at bioassay conditions.
7 An evaluation of the importance of selected parameters on mouse model estimates of
8 blood methanol AUC was performed by conducting a sensitivity analysis using the subroutines
9 within acslXtreme v2.3 (Aegis Technologies, Huntsville, Alabama). The analysis was conducted
10 by measuring the change in model output corresponding to a 1% change in a given model
11 parameter when all other parameters were held fixed. Sensitivity analyses were conducted for
12 the inhalation and oral routes. The inhalation route analysis was conducted under the exposure
13 conditions of Rogers and Mole (1997) and Rogers et al. (1993a): 7-hour inhalation exposures at
14 the no-observed-effect level (NOEL) concentration of 1,000 ppm. The oral route sensitivity
15 analysis was conducted for an oral dose of 1,000 mg/kg.
16 The parameters with the largest sensitivity coefficients for the inhalation route at
17 1,000 ppm (absolute values >1) were pulmonary ventilation scaling coefficient (QPC) and
18 maximum velocity of the high-affinity/low-capacity pathway (VmaxC). The sensitivity
19 coefficient for QPC increases during the exposure period as metabolism begins to saturate.
20 Following oral exposure, mouse blood methanol AUC was sensitive to the rate constants for oral
21 uptake. Blood AUC was most sensitive to the maximum and saturation rate constants for uptake
22 from the stomach (VmASC and KMASC). The sensitivity coefficient for VmASC decreased
23 during the first hours after exposure from 1 to less than 0.1 at the end of exposure. Blood
24 methanol AUC was also modestly sensitive to first-order uptake from the intestine (KAI), and
25 first-order transfer between stomach and intestine (KSI), the rate constants for uptake from the
26 intestine and transfer rates between compartments, respectively. For a more complete
27 description of this sensitivity analysis for the mouse methanol PBPK model see Appendix B.
3.4.5. Rat Model Calibration
28 The rat model was calibrated to fit data from i.v., inhalation, and oral exposures in rats,
29 using data provided in the command file of Ward et al. (1997) and obtained from figures in
30 Horton et al. (1992) using Digitizlt. Holding other parameters constant, the rat PBPK model was
31 initially calibrated against the entire set of i.v.-route blood PK data (Figure 3-7) by fitting
32 Michaelis-Menten constants for one high-affinity/low-capacity and one low-affinity/high-
33 capacity enzyme to both the Ward et al. (1997) data for Sprague-Dawley (SD) rats and the
34 Horton et al. (1992) data for Fischer 344 (F344) rats, assuming that any difference between the
35 two data sets (100 mg/kg data) were from experimental variability and that a single set of
36 parameters could be fit to data for both strains of rat. However when the resulting parameters
3-31 DRAFT-DO NOT CITE OR QUOTE
-------
1 were then used to simulate the F344 inhalation uptake data of Horton et al. (with the fractional
2 absorption for inhalation, FRACIN, adjusted to fit those data), it was found that the clearance
3 rate predicted after the end of inhalation exposure was much more rapid than shown by the data.
4 More careful examination of the i.v. data then revealed that there too the clearance for F344 rats
5 was slower than for SD rats, and that the metabolic parameters obtained from fitting the
6 combined i.v. data best represented the SD rat data. It was concluded that the combined data set
7 indicated a true strain difference in metabolic parameters. The metabolic parameters for SD rats
8 were then obtained by fitting only the Ward et al. (1997) i.v. data (both doses).
9 The 100 mg/kg i.v. data of Horton et al. (1992) were combined with their inhalation data
10 and a simultaneous optimization of the metabolic parameters and FRACIN for F344 rats was
11 attempted over that data set. For this data set, however, the optimization either converged with
12 the metabolic Vmax for the high affinity (low Km) pathway at zero, or with that Km value
13 increasing to be statistically indistinguishable from the high Km value. Therefore the Vmax for
14 the high affinity pathway was allowed to be zero, the Km for that pathway was not estimated,
15 and only a single Vmax and low affinity (high Km) were fit to those data, with a simultaneous
16 identification of FRACIN. Since there are no inhalation data for SD rats, this value of FRACIN
17 was assumed to apply for both strains. The optimized parameters for both strains of rats are
18 given in Table 3-10.
19 When the model was calibrated using the available inhalation and i.v. data for F344 rats
20 (Horton et al., 1992), a low fractional absorption of 20% was optimized to best fit the data, vs.
21 66.5% for the mouse. This lower fractional absorption is consistent with values presented by
22 Perkins et al. (1995), who also found that the fractional absorption of methanol from inhalation
23 studies was lower in rats than in mice.
24
25
26
27
3-32 DRAFT-DO NOT CITE OR QUOTE
-------
5000
2500 mg/kg IV in rats
16 24 32 40 48
Time (hr)
100 mg/kg IV in rats
cu
O
o>
o
_o
CD
Ward '97 data
SD rat model fit
Morton '92 data
F344 rat model fit
246
Time (hr)
Figure 3-8. NP rat i.v. route methanol blood kinetics. Methanol (MeOH) was infused
into: female Sprague-Dawley rats (275 g; solid diamonds and lines) at target
doses of 100 or 2,500 mg/kg (Ward et al., 1997); or male F-344 rats (220 g; open
triangles and dashed line) at target doses of 100 mg/kg (Horton et al., 1992).
Data points represent measured blood concentrations and lines represent PBPK
model simulations.
Source: Ward et al. (1997); Horton et al. (1992).
100
10
o
o
0,1
2000 ppm data
* 1200 ppm data
A 200 ppm data
^— 2000 ppm, 7-hr simulation
X
— 2000 ppm simulation
— 1200 ppm simulation
— 200 ppm simulation
10
15
Time (hr)
Figure 3-9. Model fits to data sets from inhalation exposures to 200 (triangles), 1,200
(diamonds), or 2,000 (squares) ppm methanol in male F-344 rats. The model was
calibrated against all three sets of concentration data, though it converged to
parameter values that only fit the lower two data sets well. Symbols are
concentrations obtained from Horton et al. (1992) using Digitizlt! Lines
represent PBPK model fits. Since the 2000 ppm data peak occurred at 7 hr, a 7-
hr simulated exposure is also shown for comparison.
Source: Horton et al. (1992).
3-33
DRAFT-DO NOT CITE OR QUOTE
-------
1 Finally, oral absorption parameters were optimized to the oral absorption data reported by
2 Ward et al. (1997), also using the optimization routines in acslXtreme v2.5.0.6 (Aegis
3 Technologies, Huntsville, Alabama) (Table 3-10: Figure 3-9). While the two-compartment GI
4 model (Figure 3-4) allows for both slow and fast absorption modes, it was not possible to fit both
5 the 100 mg/kg data and the first several hours of the 2500 mg/kg data with that model structure
6 using linear absorption and inter-compartment transfer rates. In particular the shorter-time data
7 for 2500 mg/kg indicate a much slower rate of increase in blood levels than the linear-absorption
8 model (top, thick line in upper panel of Figure 3-10), but the 100 mg/kg data (lower panel of
9 Figure 3-10) are indeed consistent with a linear model, showing a rapid rise to a fairly narrow
10 peak, then dropping rapidly. As long as linear rate equations were used, the shape of the
11 absorption curve at 2500 mg/kg would mirror that at 100 mg/kg, but the data show a clear
12 difference. It was concluded that the rate of absorption must at least partly saturate at the higher
13 dose, and hence that Michaelis-Menten kinetics should be used.
14 Even with the addition of saturable absorprtion from the stomach, it was also found that
15 the 2500 mg/kg model simulations over-predicted all of those data (result not shown) and it was
16 hypothesized that fecal elimination might become significant at such a high exposure level, so a
17 term for fecal elimination from the intestine compartment was added. When that fecal rate
18 constant and the saturable absorption from the stomach compartment were both used, the
19 resulting fit to the data (thin, dashed line in upper panel) was considerably improved with an
20 almost identical (excellent) fit to the 100 mg/kg data (saturable curve can be distinguished from
21 the linear curve just after the peak is reached in the lower panel of Figure 3-4). For the purpose
22 of scaling across individuals, strains, and species, the Km for absorption from the stomach
23 (KMAS) was assumed to scale in proportion to the stomach (lumen) volume; i.e., with BW1.
24 The Vmax (VmAS) was assumed to scale as BW0'75, with the result that for low doses the
25 effective linear rate constant (VmAS/KMAS) scales as BW"0'25, which is a standard assumption
26 for linear rates. Since the quantity on which the rate depends is the total amount in the stomach
27 (mg methanol), the resulting scaling constant for the Km, KMASC, conveniently has units of
28 mg/kg BW; i.e., the standard units for oral dosing.
3-34 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
E
3E
2,700
2,400 -
2,100-
1,800
1,500-
1,200-
900-
600-
300-
0-
2500 mg/kg linear uptake simulation
2500 mg/kg saturable uptake simulation
2500 mg/kg data
100 mg/kg linear uptake simulation
100 mg/kg saturable uptake simulation
100 mg/kg data
0
12
18
Time (hr)
24
30
36
o
E
••_••
O
Qj
100 H
80
60
40
3 20
— 100 mg/kg linear uptake simulation
100 mg/kg saturable uptake simulation
n 100 mg/kg data
012345678
Time (hr)
Figure 3-10. Model fits to datasets from oral exposures to 100 and 2,500 mg/kg
methanol in female Sprague-Dawley rats. Symbols are concentration data obtained
from the command file. Lines represent PBPK model fits.
Source: Ward et al. (1997).
3.4.6. Human Model Calibration
3.4.6.1. Inhalation Route
The mouse model was scaled to humans by setting either a standard human body weight
(70 kg) or study-specific body weights and using human tissue compartment volumes and blood
flows, and then calibrated to fit the human inhalation exposure data available from the open
literature, which comprised data from four publications (Ernstgard et al., 2005; Batterman et al.,
1998; Osterloh et al., 1996; Sedivec et al., 1981).
Since the human data included time-course data for urinary elimination, a first-order rate
of loss of methanol from the blood (Kl) was used to provide an estimate of methanol elimination
to the bladder compartment in humans, and the rate of elimination from that compartment then
3-35
DRAFT-DO NOT CITE OR QUOTE
-------
1 characterized by a second constant (KBL). Note that the total amount eliminated by this route
2 depends only on Kl, while KBL affects the rate at which the material cleared from the blood
3 then appears in the urine. Inhalation-route urinary methanol kinetic data described by Sedivec
4 et al. (1981) (Figure 3-11) was used in the model calibration to inform this rate constant.
5 Without use of the bladder compartment and rate constant, the fit of the model predictions to the
6 data in Figure 3-11 is quite poor (results not shown), and a statistical test on the improvement of
7 fit obtained by introducing the additional parameter (KBL) is significant (p < 0.0001).
8 Conversion between the PBPK-model-predicted rate of urinary excretion (mg/hours) or
9 cumulative urinary excretion (mg) and the urine methanol concentration data reported by the
10 authors was achieved by assuming 0.5 mL/hours/kg body weight total urinary output (Horton
11 et al., 1992). The resulting values of K1C and KBL, shown in Table 3-10, differ somewhat
12 depending on whether first-order or saturable liver metabolism is used. These are only calibrated
13 against a small dataset and should be considered an estimate. Urinary elimination is a minor
14 route of methanol clearance with little impact on blood methanol kinetics.
15 Although the high doses used in the mouse studies clearly warrant the use of a second
16 metabolic pathway with a high Km, the human exposure data all represent lower concentrations
17 and may not require or allow for accurate calibration of a second metabolic pathway. Horton
18 et al. (1992) employed two sets of metabolic rate constants to describe human methanol
19 disposition, similar to the description used for rats and mice, but in vitro studies using monkey
20 tissues with nonmethanol substrates were used as justification for this approach. Although
21 Bouchard et al. (2001) described their metabolism using Michaelis-Menten metabolism, Starr
22 and Festa (2003) reduced that to an effective first-order equation and showed adequate fits.
23 Perkins et al. (1995a) estimated a Km of 320 + 1273 mg/L (mean + S.E.) by fitting a
24 one-compartment model to data from a single estimated oral dose. In addition to the extremely
25 high standard error, the large standard error for the associated Vmax (93 + 87 mg/kg/hours)
26 indicates that the set of Michaelis-Menten constants was not uniquely identifiable using this data.
27 Other Michaelis-Menten constants have been used to describe methanol metabolism in various
28 models for primates (Table 3-11).
29
3-36 DRAFT-DO NOT CITE OR QUOTE
-------
10
9 H
-231 ppm saturable sim
-157 ppm saturable sim
-78 ppm saturable sim
• 231 ppm linear sim
• 157 linear sim
• 78 ppm linear sim
231 ppm data
157 ppm data
O)
E,
o
is
-t
§
o
8
Sedivec et al., urine
concentration
O 78 ppm data
20
24
2.5 -
Sedivec et al.,
total urinary excretion
8 12 16
Time (h)
20
24
Figure 3-11. Urinary methanol elimination concentration (upper panel) and
cumulative amount (lower panel) following inhalation exposures to methanol in
human volunteers. Data points in lower panel represent estimated total urinary
methanol elimination from humans exposed to 78 (diamonds), 157 (triangles), and 231
(circles) ppm methanol for 8 hours, and lines represent PBPK model simulations.
Source: Sedivec et al. (1981).
3-37
DRAFT-DO NOT CITE OR QUOTE
-------
Table 3-11. Primate Kms reported in the literature
Km (mg/L)
320 +12733
716 + 4893
278
252+1163
33.9
0.66
23.7 + 8.7"'*
Reference
Jacobsenetal., 1988
Nokeretal., 1980
Makar et al., 1968
Eellsetal., 1983
Hortonetal., 1992
Fisher etal., 2000
(This analysis.)
Note
Human: oral poisoning, estimated dose
Cynomolgus Monkey: 2 g/kg dose
Rhesus Monkey: 0.05-1 mg/kg dose
Cynomolgus Monkey: 1 g/kg dose
PBPK model: adapted from rat Km
PBPK model, Cynomolgus Monkey: 10-900 ppm
PBPK model, human: 100-800 ppm
"The values reported are mean ± S.D.
bThis Km was optimized while varying Vmax, K1C, and KBL, from all of the at-rest human inhalation data as a part
of this project. The S.D. given for this analysis is based on the Optimize function of acslXtreme, which assumes all
data points are discrete and not from sets of data obtained over time; therefore a true S.D. would be higher. The
final value reported in Table B-l (21 mg/L) was obtained by sequentially rounding and fixing these parameters,
then re-optimizing the remaining ones.
Source: Perkins etal. (1995b).
Table 3-12. Parameter estimate results obtained using acslXtreme to fit all human
data using either saturable or first-order metabolism
Parameters
Michaelis-Menten (optimized)
Km
vmaxc
First order
KLLC
Optimized value
23.7
33.1
95.7
S.D.
8.9
10.1
5.4
Correlation coefficient
-0.994
NA
LLF
-24.1
-31.0
Note. The S.D.s are based on the Optimize function of acslXtreme v2.3, which assumes all data points are discrete
and not from sets of data obtained over time. Therefore a true S.D. would be a higher value.
4 To estimate both Michaelis-Menten and first-order rates, all human data under
5 nonworking conditions (Batterman et al., 1998; Osterloh et al., 1996; Sedivec et al., 1981) were
6 used (Table 3-12). The metabolic (first-order or saturable) and urinary elimination constants
7 were numerically fit to the human datasets, while holding the value for FRACIN at 0.8655
8 (estimated from the results of Sedivec et al. [1981]) and holding the ventilation rate constant at
9 16.5 L/hours/kga75 and QPC at 24 L/hours/kg0'75 (values used by EPA [2000d] for modeling the
10 inhalation-route kinetics of vinyl chloride). Other human-specific physiological parameters were
11 used,as reported in Table 3-10. Final fitted parameters that have been used in the saturable
3-38 DRAFT-DO NOT CITE OR QUOTE
-------
1 model are given in Table 3-10. The resulting fits of two different possible parameterizations
2 (first-order ["linear"] or optimized Km/Vmax) are shown in Figures 3-11 and 3-12.
3
Batterman et al., blood concentration
(800 ppm exposures)
-Saturable metabolsim
First-order metabolism
2 h data
1 h data
30 min data
Osterloh et al.
(200 ppm inhalation)
O Data
Saturable metabolsim
First-order metabolism
Time (hr)
4
5
6
7
Figure 3-12. Data showing the visual quality of the fit using optimized first-order or
Michaelis-Menten kinetics to describe the metabolism of methanol in humans. Rate
constants used for each simulation are given in Table 3-12.
Source: Batterman et al. (1998: top); Osterloh et al. (1996: bottom).
Use of a first-order rate has the advantage of resulting in a simpler (one fewer variable)
model, while providing an adequate fit to the data; however, the saturable model clearly fits
some of the data better. To discriminate the goodness-of-fit resulting of the inclusion of an
additional variable necessary to describe saturable metabolism versus using a single first-order
3-39 DRAFT-DO NOT CITE OR QUOTE
-------
1 rate, a likelihood ratio test was performed.13 The hypothesis that one metabolic description is
2 better than another is calculated using the likelihood functions evaluated at the maximum
3 likelihood estimates. Since the parameters are optimized in the model using the maximum log
4 likelihood function (LLF), the resultant LLF is used for the statistical comparison of the models.
5 The equation states that two times the log of the likelihood ratio follows a chi square (%2)
6 distribution with r degrees of freedom:
7 -2[log(/l(modell)//l(model2))] =-2[log/l(modell)-log/l(model2)] = %}
8 The likelihood ratio test states that if the two times the difference between the maximum LLFs of
9 the two different descriptions of metabolism is greater than the $ distribution then the model fit
10 has been improved (Devore, 1995; Steiner et al., 1990).
11 At greater than a 99.95% confidence level, using two metabolic rate constants (Km and
12 VmaxC) is preferred over using a single rate constant (Table 3-13). Forcing the model to use the
13 Km calculated by Perkins et al. (1995b) would result in model fits indistinguishable from the
14 first-order case (results not shown). While the correlation coefficients (Table 3-12) indicate that
15 VmaxC, and Km are highly correlated, that is not unexpected, and the S.D.s (Table B-3) indicate
16 that each is reasonably bounded. If the data were indistinguishable from a linear system, Km in
17 particular would not be so bounded from above since the Michaels-Menten model becomes
18 indistinguishable from a linear model as VmaxC and Km tend to infinity. Further, the internal dose
19 candidate points of departure (PODs), for example the BMDLio for the inhalation-induced brain-
20 weight changes from NEDO (1987) with methanol blood AUC as the metric, is 90.9 mg-hr/L,
21 which corresponds to an average blood concentration of 3.8 mg/L. Therefore, the Michaelis-
22 Menten metabolism rate equation appears to be sufficiently supported by the existing data with
23 values in a concentration range in which the nonlinearity has an impact.
Table 3-13. Comparison of LLFs for Michaelis-Menten and first-order metabolism
LLF (logX) for
M-M
-24.1
LLF (log!) for
1st order
-31.0
LLF
1st versus M-Ma
34.1
X2r (99%
confidence)1"
13.8
X2, (99.95%
confidence)1"
12.22
24
Note. Models were optimized for all human datasets under non working conditions. M-M: Michaelis-Menten
Obtained using this equation: - 2[log /l(model l) - log /l(model 2)]
bsignificance level at r=l degree of freedom.
25 While the use of Michaelis-Menten kinetics might allow predictions across a wide
26 exposure range (into the nonlinear region), extrapolation above 1,000 ppm is not suggested since
27 the highest human exposure data are for 800 ppm. Extrapolation to higher concentrations is
13 Models are considered to be nested when the model structures are identical except for the addition of complexity,
such as the added metabolic rate. Under these conditions, the likelihood ratio can be used to compare the relative
ability of the two models to describe the data, as described in "Reference Guide for Simusolv" (Steiner et al., 1990).
3-40 DRAFT-DO NOT CITE OR QUOTE
-------
1 potentially misleading since the nonlinearity in the exposure-internal-dose relationship for
2 humans is uncertain above this point. The use of a BMD or internally applied uncertainty factors
3 (UFs) should place the exposure concentrations well within the linear range of the model.
4 The data from Ernstgard et al. (2005) were used to assess the use of the first-order
5 metabolic rate constant to a dataset collected under conditions of light work. Historical measures
6 of QPC (52.6 L/hours/kga75) and QCC (26 L/hours/kg0'75) for individuals exposed under
7 conditions of 50 watts of work from that laboratory (52.6 L/hours/kg0'75) (Ernstgard, 2005;
8 Corley et al., 1994; Johanson et al., 1986) were used for the 2-hour exposure period (Figure. 3-
9 12). Otherwise, there were no changes in the model parameters (no fitting to these data). The
10 results are remarkably good, given the lack of parameter adjustment to data collected in a
11 different laboratory and using different human subjects than those to which the model was
12 calibrated.
13
14
15
Ernstgard et al. (2005)
Figure 3-13. Inhalation exposures to methanol in human volunteers. Data points
represent measured blood methanol concentrations from humans (4 males and
4 females) exposed to 100 ppm (open symbols) or 200 ppm (filled symbols) for 2
hours during light physical activity. Solid lines represent PBPK model
simulations with no fitting of model parameters. For the first 2 hours, a QPC
of 52.6 L/hours/kg075 (Johansen et al., 1986), and a QCC of 26 L/hours/kg075
(Corley et al., 1994) was used by the model.
Source: Ernstgard et al. (2005).
3.4.6.2. Oral Route
There were no methanol human data available for calibration or validation of the oral
route for the human model. In the absence of methanol data to estimate rate constants for oral
uptake, human oral absorption parameters reported values for ethanol (Sultatos et al., 2004) are
3-41
DRAFT-DO NOT CITE OR QUOTE
-------
1 set in the code, except that saturable absorption from the stomach was retained with the KMASC
2 equal to the mouse value. The maximum rate of absorption form the stomach, VMASC, was
3 then set such that for a 70-kg person, VMAS/KM (the effective first-order rate constant at low
4 doses) matched the first-order absorption rate from Sultatos et al. (0.21 hr"1). Also, while
5 Sultatos et al. included a rate of metabolism for ethanol in the stomach, the corresponding fecal
6 elimination rate was set to zero here, effectively assuming 100% absorption of methanol for
7 humans. However, human oral dosimetry was described as zero-order uptake, in which
8 continuous infusion at a constant rate into the stomach equal to the daily dose/24 hours was
9 assumed and human internal doses were computed at steady state. Since absorption is 100% for
10 the human model, at steady state the net rate of absorption must equal the rate of infusion to the
11 stomach, irrespective of the other parameter values. (Changes in the absorption constants simply
12 cause the amount of methanol in each GI compartment at steady state to change until the net rate
13 of absorption from the somtach and intestine equals the rate of infusion.) Thus the human
14 absorption constants were set to what is considered a reasonable estimate, given the lack of
15 human oral PK data, but the simulations are conducted in a way that makes the result insensitive
16 to their values; having human values set does allow for simulations of non-constant infusion,
17 should such be desired. Since the AUC was computed for a continuous oral exposure, its value
18 is just 24 hours times the steady-state blood concentration at a given oral uptake rate.
3.4.7. Monkey PK Data and Analysis
19 In order to estimate internal doses (blood AUCs) for the monkey health-effects study of
20 Burbacher et al. (1999b) and further elucidate the potential differences in methanol
21 pharmacokinetics between NP and pregnant individuals (2nd and 3rd trimester), a focused
22 reanalysis of the data of Burbacher et al. (1999a) was performed. Individual blood concentration
23 measurements prior to and following exposure are shown in scatter plots in Appendix B of
24 Burbacher et al. (1999a). More specifically, the monkeys in the study were exposed for 2.5
25 hours/day, with the methanol concentration raised to approximately the target concentration for
26 the first 2 hours of each exposure and the last 30 minutes providing a chamber "wash-out"
27 period, when the exposure chamber concentration was allowed to drop to 0. Blood samples
28 were taken and analyzed for methanol concentration at 30 minutes, 1, 2, 3, 4, and 6 hours after
29 removal from the chamber (or 1, 1.5, 2.5, 3.5, 4.5, and 6.5 hours after the end of active
30 exposure). These data were analyzed to compare the PK in NP versus pregnant animals, and
31 fitted with a simple PK model to estimate 24-hour blood AUC values for each exposure level.
32 Dr. Burbacher graciously provided the original data, which were used in this analysis.
33 Two cohorts of monkeys were examined, but the data (plots) did not indicate a systematic
34 difference between the two, so the data from the two cohorts were combined. The data from the
35 scatter plots of Burbacher et al. (1999a) for the NP (pre-pregnancy), first pregnancy (2nd
36 trimester), and second pregnancy (3rd trimester) studies are compared in Figure 3-13, along with
3-42 DRAFT-DO NOT CITE OR QUOTE
-------
1 model simulations (explained below). Since the pregnancy time points were from animals that
2 had been previously exposed for 87 days plus the duration of pregnancy to that time point, the
3 pre-exposed NP animals were used for comparison, rather than naive animals, with the
4 expectation that effects due to changes in enzyme expression (i.e., induction) from the
5 subchronic exposure would not be a distinguishing factor. Note that each exposure group
6 included a pre-exposure baseline or background measurement, also shown. To aid in
7 distinguishing the data visually, the NP data are plotted at times 5 minutes prior to the actual
8 blood draws and the 3rd trimester at 5 minutes after each blood draw.
9 Overall there appears to be no significant or systematic difference among the NP and
10 pregnant groups. The solid lines are model simulations calibrated to only the 2nd trimester data
11 (details below), but they just as adequately represent average concentrations for the NP and 3rd
12 trimester data. Likewise, a PK model calibrated to the NP PK data adequately predicted the
13 maternal methanol concentrations in the pregnant monkeys (results not shown). Since any
14 maternal:fetal methanol differences are expected to be similar in experimental animals and
15 humans (with the maternal:fetal ratio being close to one due to methanol's high aqueous
16 solubility and relatively limited metabolism by the fetus), the predicted levels for the 2nd
17 trimester maternal blood are used in place of measured or predicted fetal concentrations.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3-43 DRAFT-DO NOT CITE OR QUOTE
-------
1800 ppm
64.0 -i
-200 -100
0 100 200
Time (min)
300
400
600 ppm
24.0-i
18.0 -
±1/1
'
-200 -100
x P re-pregnancy
+ 2ndTrimester
- SrdTrimester
Simulation
0 100 200
Time (min)
300
400
200 ppm
10.0-1
_x&i/ §
*+ m
-200 -100
0 100 200
Time (min)
300
400
1
2
3
4
5
6
Figure 3-14. Blood methanol concentration data from NP and pregnant monkeys. NP
and 3rd trimester data are plotted, respectively, at 5 minutes before and after
actual collection times to facilitate comparison. Solid line is from simple PK
model, fit to 2nd trimester data only.
Source: Burbacher et al. (1999a; Figure B-4).
3.4.7.1. PK Model Analysis for Monkeys
To analyze and integrate the PK data of Burbacher et al. (1999a), the one-compartment
model for Michaelis-Menten kinetics used by Burbacher et al. (1999a, 1999b) was extended by
the addition of a chamber compartment to capture the kinetics of concentration change in the
exposure chamber, as shown in Figure 3-14. The data in Figure 3-14 (digitized from Figure 5 of
Burbacher et al., 1999a, 1999b) show an exponential rise to and fall from the approximate target
concentration during the exposure period. The use of a single-compartment model for the
3-44
DRAFT-DO NOT CITE OR QUOTE
-------
1 chamber allows this dynamic to be captured, so that the full concentration-time course is used in
2 simulating the monkey internal concentration rather than an approximate step function (i.e. rather
3 than assuming an instantaneous rise and fall). The pair of equations representing the time-course
4 in the chamber and monkey are as follows (bolded parameters are fit to data):
5 Chamber: dCch/dt = [(CCM- S - Cch> Fch - Rmh]/Vch
6 Monkey: dCmk/dt = [Rmh - Vmax- Cmk/(Km + (?„*)]/(Vmk- BW)
7 with Rmh = Cch-Rc-(1000-BW)0-74 -F and Cnet = Cmk + Cbg.
8 d: delta, change
9 Ccn: instantaneous chamber concentration (mg/L)
10 t: time (hour)
11 CCM: chamber in-flow methanol concentration (mg/L), which was set to the concentrations
12 corresponding to those reported in Table 2 of Burbacher et al. (1999a), using the
13 "Breeding" column for the NP (87 days pre-exposed; values in Table 3-14)
14 S: exposure switch, set to 1 when exposure is on (first 2 hours) and 0 when off
15 Fch:chamber air-flow, 25,200 L/hours, as specified by Burbacher et al. (1999a,1999b)
16 Rinh: net rate of methanol inhalation by the monkeys (mg/hr)
17 VCh (1,220 L): chamber volume, initially set to 1,380 L ("accessible volume" stated by
18 Burbacher et al. [1999a, 1999b]), but allowed to vary below that value to account for
19 volume taken by equipment, monkey, and to allow for imperfect mixing
20 C^: instantaneous inhalation-induced monkey blood methanol concentration (mg/L); this is
21 added to the measured background/endogenous concentration before comparison to data
22 Vmax (39.3 mg/hr): fitted (nonscaled) Michaelis-Menten maximum elimination rate
23 Km (14.6 mg/L): fitted (nonscaled) Michaelis-Menten saturation constant
24 Vmk (0.75 L/kg): fitted volume of distribution for monkey
25 BW: monkey body weight (kg); for NP monkeys set to group average values in data of
26 Burbacher et al. (1999a, 1999b; personal communication)
27 RC: allometric scaling factor for total monkey respiration (0.12 L/hours/g0'74 =
28 2 mL/minute/ga74), as used by Burbacher et al. (1999a, 1999b)(note that scaling is to BW
29 in g, not kg)
30 F: fractional absorption of inhaled methanol, set to 0.6 (60%), the (rounded) value measured
31 in humans by Sedivec etal. (1981); F and Vmk cannot be uniquely identified, given the
32 model structure, so F was set to the (approximate) human value to obtain a realistic
3 3 estimate of Vmk
34 Cnet: net blood concentration, equal to sum of the inhalation-induced concentration (Cmk) and
35 the background blood level (Cbg) (mg/L)
36 Cbg: background (endogenous) methanol concentration, set to the pre-exposure group-
37 specific mean from the data of Burbacher et al. (1999a, 1999b; personal communication)
3-45 DRAFT-DO NOT CITE OR QUOTE
-------
b
Q.
Q.
C
O
1
0
O
C
O
O
2000 -
1800
1600
1400
1200
1000
800
600
400
onn
c
A
f 1800 ppm
Chamber concentration profile
k
I
600 ppm
-•- -m--m--m---m — m--m--m--m---m — •--,
[,•
' 200 ppm
I I I
J 0.5 1 1.5 2
Time (hr)
A
'•vSv
™"
Figure 3-15. Chamber concentration profiles for monkey methanol exposures. Lines
are model simulations. Indicated concentrations are target concentrations;
measured concentrations differed slightly (see Table 3-14).
Source: Burbacher et al. (1999a).
1 The model was specifically fit to the 2nd trimester monkey data, assuming that the
2 parameters were the same for all the exposure groups and concentrations. While the discussion
3 above and data show little difference between the NP and two pregnancy groups, the 2nd
4 trimester group was presumed to be most representative of the average internal dosimetry over
5 the entire pregnancy. Further, the results of Mooney and Miller (2001) show that developmental
6 effects on the monkey brain stem following ethanol exposure are essentially identical for
7 monkeys exposed only during early pregnancy versus full-term, indicating that early pregnancy
8 is a primary window of vulnerability.
9 Model simulation results are the lines shown in Figures 3-13 and 3-14. The model
10 provides a good fit to the monkey blood and chamber air concentration data. While the chamber
11 volume was treated as a fitted parameter, which was not done by Burbacher et al. (1999a), the
12 chamber concentration data support this estimate. The model does an adequate job of fitting the
13 data for all exposure groups without group-specific parameters. In particular, the data for all
14 exposure levels can be adequately fit using a single value for the volume of distribution (Vmk) as
15 well as each of the metabolic parameters. While one may be able to show statistically distinct
16 parameters for different groups or exposure levels (by fitting the model separately to each), as
17 was done by Burbacher et al. (1999a), it is unlikely that such differences are biologically
18 significant, given the fairly large number of data points and the large variability evident in the
19 blood concentration data. Thus, the single set of parameters listed with the parameter
20 descriptions above will be used to estimate internal blood concentrations for the dose-response
3-46
DRAFT-DO NOT CITE OR QUOTE
-------
1 analysis. The chamber concentrations for "pregnancy" exposures recorded by Burbacher et al.
2 (1999a; Table 2) and average body weights for each exposure group at the 2nd trimester time
3 point were used along with the model to calculate 24-hour blood methanol AUCs (Table 3-14).
Table 3-14. Monkey group exposure characteristics
Exposure concentration (ppm)a
206
610
1,822
Group average BW (kg)b
3.46
4.08
3.83
24-hr blood methanol AUC
(mg-hr/L)c
6.73
28.28
138.11
aFrom Burbacher et al. (1999a,1999b), Table 2, "pregnancy" exposure.
bFrom Burbacher, original data (personal communication).
Calculated using the two-compartment PK model as described above.
3.4.8. Summary and Conclusions
4 Mouse, rat, and human versions of a methanol PBPK model have been developed and
5 calibrated to data available in the open literature. The model simplifies the structure used by
6 Ward et al. (1997), while adding specific refinements such as a standard lung compartment
7 employed by Fisher et al. (2000) and a two-compartment GI tract.
8 Although the endpoints of concern are developmental effects which occur during in utero
9 and (to a lesser extent) lactational exposure, no pregnancy-specific PBPK model exists for
10 methanol and inadequate data exists for the development and validation of a
11 fetal/gestational/conceptus compartment. The fact that the unique physiology of pregnancy and
12 the fetus/conceptus are not represented in a methanol model would be important if methanol
13 pharmacokinetics differed significantly during pregnancy or if the observed partitioning of
14 methanol into the fetus/conceptus versus the mother showed a concentration ratio significantly
15 greater than or less than 1. Methanol pharmacokinetics during GD6-GD10 in the mouse are not
16 different from NP mice (Pollack and Brouwer, 1996), and the maternal blood:fetus/conceptus
17 partition coefficient is reported to be near 1 (Ward et al., 1997; Horton et al., 1992). At GDIS in
18 the mouse, maternal blood levels are only modestly different from those in NP animals (see
19 Figures B-4 and B-5 [Appendix B] for examples), and in general the PBPK model simulations
20 for the NP animal match the pregnancy data as well as the nonpregnancy data. Likewise,
21 maternal blood kinetics in monkeys differs little from those in NP animals (see Section 3.4.7 for
22 details). Further, in both mice and monkeys, to the extent that late-pregnancy blood levels differ
23 from NP for a given exposure, they are higher; i.e., the difference between model predictions and
24 actual concentrations is in the same direction. These data support the assumption that the ratio of
25 actual target-tissue methanol concentration to (predicted) NP maternal blood concentrations will
26 be about the same across species, and hence, that using NP maternal blood levels in place of fetal
27 concentrations will not lead to a systematic error when extrapolating risks.
3-47 DRAFT-DO NOT CITE OR QUOTE
-------
1 The findings in the mouse (similar blood methanol kinetics between NP and pregnant
2 animals prior to GDIS and a maternal blood:fetal partition coefficient close to 1) are assumed to
3 be applicable to the rat. However, the critical gestational window for the reduced brain weight
4 effect observed in the NEDO (1987) rat study is broader than for the mouse cervical rib effect.
5 In addition, NEDO (1987) rats were exposed not only to methanol gestationally but also
6 lactationally and via inhalation after parturition. The additional routes of exposure presented to
7 the pups in this study present uncertainties and suggest that average blood levels in pups might
8 be greater than those of the dam.
9 Methanol is transported directly from the maternal circulation to fetal circulation via the
10 placenta, but transfer via lactation involves distribution to the breast tissue, then milk, then
11 uptake from the pup's GI tract. Therefore blood or target-tissue levels in the breast-feeding
12 infant or pup are likely to differ more from maternal levels than do fetal levels. In addition, the
13 health-effects data indicate that most of the effects of concern are due to fetal exposure, with a
14 relatively small influence due to postbirth exposures. Further, it would be extremely difficult to
15 distinguish the contribution of postbirth exposure from pre birth exposure to a given effect in a
16 way that would allow the risk to be estimated from estimates of both exposure levels, even if one
17 had a lactation/child PBPK model that allowed for prediction of blood (or target-tissue) levels in
18 the offspring. Finally, one would still expect the target-tissue concentrations in the offspring to
19 be closely related to maternal blood levels (which depend on ambient exposure and determine
20 the amount delivered through breast milk), with the relationship between maternal levels and
21 those in the offspring being similar across species. Further, as discussed to a greater extent in
22 Sections 5.1.2 and 5.3, it is likely that the difference in blood levels between rat pups and dams
23 would be similar to the difference between mothers and human offspring. Therefore, it is
24 assumed that the potential differences between pup and dam blood methanol levels do not have a
25 significant impact on this risk assessment and the estimation of HECs. The use of the full
26 intrahuman UF of 10 is also expected to account for PK differences between children and adults.
27 Therefore, the development of a lactation/child PBPK model appears not to be necessary,
28 given the minimal change that is likely to result in risk extrapolations, and use of (NP) maternal
29 blood levels as a measure of risk in the offspring is considered preferable over use of default
30 extrapolation methods. In particular, the existing human data allow for predictions of maternal
31 blood levels, which depend strongly on the rate of maternal methanol clearance. Since bottle-fed
32 infants do not receive methanol from their mothers, they are expected to have lower or, at most,
33 similar overall exposures for a given ambient concentration than the breast-fed infant, so that use
34 of maternal blood levels for risk estimation should also be adequately protective for that group.
35 The model fits to the mouse oral-route methanol kinetic data, using a consistent set of
36 parameters (Figure B-4 in Appendix B), are fairly good for doses of 1,500 mg/kg but
37 underpredict blood levels by 30% or more after a dose of 2,500 mg/kg. In particular, the oral
38 mouse model consistently underpredicts the amount of blood methanol reported in two studies
3-48 DRAFT-DO NOT CITE OR QUOTE
-------
1 (Ward et al., 1997, 1995). Ward et al. (1997) utilized a different Vmax for each oral absorption
2 dataset; the GDIS and the GD8 data from Dorman et al. (1995) were both fit using a Vmax of
3 -80 mg/kg/hours (body weights were not listed; the model assumed that GD8 and GDIS mice
4 were both 30 g; Ward et al. [1997] did not scale by body weight). Additionally, lower partition
5 coefficients for placenta (1.63 versus 3.28) and embryonic fluid (0.0037 versus 0.77) were used
6 for GD8 and GDIS. The current refined model adequately fits the oral PK data using a single set
7 of parameters that is not varied by dose or source of data.
8 The rat models were able to adequately predict the limited inhalation, oral and i.v.
9 datasets available. Low-dose exposures were emphasized in model optimization due to their
10 greater relevance to risk assessment. Based on a rat inhalation exposure to 500 ppm, the HEC
11 would be 281 ppm (by applying an AUC of 201.3 [Figure B-12] to Equation 1 of Appendix B).
12 The final mouse, rat, and human methanol PBPK models fit multiple datasets for
13 inhalation, oral, and i.v., from multiple research groups using consistent parameters that are
14 representative of each species but are not varied within species or by dose or source of data.
15 Also, a simple PK model calibrated to NP monkey data, which were shown to be essentially
16 indistinguishable from pregnant monkey PK data, was used to estimate blood methanol AUC
17 values (internal doses) in that species. In Section 5, the models and these results are used to
18 estimate chronic human exposure concentrations from internal dose metrics.
3-49 DRAFT-DO NOT CITE OR QUOTE
-------
4. HAZARD IDENTIFICATION
4.1. STUDIES IN HUMANS - CASE REPORTS, OCCUPATIONAL AND
CONTROLLED STUDIES
4.1.1. Case Reports
1 An extensive library of case reports has documented the consequences of acute
2 accidental/intentional methanol poisoning. Nearly all have involved ingestion, but a few have
3 involved percutaneous and/or inhalation exposure. As many of the case reports demonstrate, the
4 association of Parkinson-like symptoms with methanol poisoning is related to the observation
5 that lesions in the putamen are a common feature both in Parkinson's disease and methanol
6 overexposure. These lesions are commonly identified using computed tomography (CT) or by
7 Magnetic Resonance Imaging (MRI). Other areas of the brain (e.g., the cerebrum, cerebellum,
8 and corpus callosum) also have been shown to be adversely affected by methanol overexposure.
9 Various therapeutic procedures (e.g., ethanol infusion, sodium bicarbonate or folic acid
10 administration, and hemodialysis) have been used in many of these methanol overexposures, and
11 the reader is referred to the specific case reports for details in this regard. The reader also is
12 referred to Kraut and Kurtz (2008) and Barceloux et al. (2002) for a more in-depth discussion of
13 the treatments in relation to clinical features of methanol toxicity. A brief discussion of the
14 terms cited in case report literature follows.
15 Basal ganglia, a group of interconnected subcortical nuclei in each cerebral hemisphere,
16 refers to various structures in the grey matter of the brain that are intimately involved, for
17 example, in coordinating motor function, maintaining ocular and respiratory function, and
18 consciousness. The connectivity within the basal ganglia involves both excitatory and inhibitory
19 neurotransmitters such as dopamine (associated with Parkinson's disease when production is
20 deficient).
21 The structures comprising the basal ganglia include but are not limited to: the putamen
22 and the globus pallidus (together termed the lentiform nuclei), the pontine tegmentum, and the
23 caudate nuclei. Dystonia or involuntary muscle contraction can result from lesions in the
24 putamina; if there are concomitant lesions in the globus pallidus, Parkinsonism can result (Bhatia
25 and Marsden, 1994). Bhatia and Marsden (1994) have discussed the various behavioral and
26 motor consequences of focal lesions of the basal ganglia from 240 case-study reports. Lesions in
27 the subcortical white matter adjacent to the basal ganglia often occur as well (Airas et al., 2008;
28 Rubinstein et al., 1995; Bhatia and Marsden, 1994). In the case reports of Patankar et al. (1999),
29 it was noted that the severity and extent of necrosis in the lenticular nuclei do not necessarily
30 correlate with clinical outcome.
4-1 DRAFT-DO NOT CITE OR QUOTE
-------
1 In one of the earliest reviews of methanol overexposure, Bennett et al. (1953) described a
2 mass accidental poisoning when 323 persons, ranging in age from 10 to 78 years, in Atlanta,
3 Georgia, consumed "whisky" adulterated with as much as 35-40% methanol. In all, 41 people
4 died. Of the 323 individuals, 115 were determined to be acidotic with symptoms (visual
5 impairment, headache [affecting ~62%], dizziness [affecting ~30%], nausea, abdominal pain and
6 others) beginning around 24 hours postexposure. Visual impairment was mostly characterized
7 by blurred or indistinct vision; some who were not acidotic experienced transient visual
8 disturbances. The cardiovascular parameters were unremarkable. The importance of acidosis to
9 outcome is shown in Table 4-1. Among the key pathological features were cerebral edema, lung
10 congestion, gastritis, pancreatic necrosis, fatty liver, epicardial hemorrhages, and congestion of
11 abdominal viscera.
12 In another early investigation of methanol poisoning (involving 320 individuals), Benton
13 and Calhoun (1952) reported on methanol's visual disturbances.
Table 4-1. Mortality rate for subjects exposed to methanol-tainted whiskey in relation
to their level of acidosisaa
Subjects
All patients
Acidotic (CO2 <20 mEq)
Acidotic (CO2 <10 mEq)
Number
323
115
30
Percent deaths
6.2
19
50
aThese data do not include those who died outside the hospital or who were moribund on arrival.
Source: Bennett et al. (1953).
14 Riegel and Wolf (1966), in a case report involving a 60-year-old woman who ingested
15 methanol, noted that nausea and dizziness occurred within 30 minutes of ingestion. She
16 subsequently passed out and remained unconscious for 3 days. Upon awakening she had
17 paralysis of the vocal cords and was clinically blind in one eye after 4 months. Some aspects of
18 Parkinson-like symptoms were evident. There was a pronounced hypokinesia with a mask-like
19 face resembling a severe state of Parkinson's disease. The patient had difficulty walking and
20 could only make right turns with difficulty. There was no memory loss.
21 Treatment of a 13-year-old girl who ingested an unspecified amount of a windshield-
22 washer solution containing 60% methanol was described by Guggenheim et al. (1971). She
23 displayed profound acidosis; her vital signs, once she was treated for acidosis, were normal by
24 36 hours after hospital admission. During the ensuing 6 months after discharge from the
25 hospital, visual acuity (20/400, both eyes) worsened, and she experienced muscle tremors, arm
26 pain, and difficulty in walking. A regimen of levadopa treatment greatly improved her ability to
27 function normally.
4-2 DRAFT-DO NOT CITE OR QUOTE
-------
1 Ley and Gali (1983) also noted symptoms that are Parkinson like following methanol
2 intoxication. In this case report respiratory support was needed; the woman was in a coma.
3 Once stabilized, she exhibited symptoms similar to those noted in other case study reports, such
4 as blurred vision, movement difficulty, and tremors. Computerized Axial Tomography scan
5 findings highlighted the central nervous system (CNS) as an important site for methanol
6 poisoning.
7 Rubinstein et al. (1995) presented evidence that a methanol blood level of 36 mg/dL
8 (360 mg/L) is associated with a suite of CNS and ocular deficits that led to a 36-year-old man
9 (who subsequently died) becoming comatose. CT scans at 1-2 days following ingestion were
10 normal. However, MRI scans at day 4 revealed lesions in the putamen and peripheral white
11 matter of the cerebral and cerebellar hemispheres. Bilateral cerebellar cortical lesions had been
12 reported in an earlier case of methanol poisoning by Chen et al. (1991).
13 Finkelstein and Vardi (2002) reported that long-term inhalation exposure of a woman
14 scientist to methanol without acute intoxication resulted in a suite of delayed neurotoxic
15 symptoms (e.g., hand tremor, dystonia, bradykinesia, and other decrements in body movement).
16 Despite treatment with levadopa, an increase in the frequency and severity of effects occurred.
17 Exposure to bromine fumes was concomitant with exposure to methanol.
18 Hantson et al. (1997a) found, in four cases, that MRI and brain CT scans were important
19 tools in revealing specific brain lesions (e.g., in the putamina and white matter). The first subject
20 was a 57-year-old woman who complained of blurred vision, diplopia, and weakness 24 hours
21 after ingesting 250 mL of a methanolic antifreeze solution. Upon hospital admission she was
22 comatose and in severe metabolic acidosis. An MRI scan at 9 days indicated abnormal
23 hyperintense foci in the putamina (decreased in size by day 23) and subtle lesions (no change by
24 day 23) in the white matter. Upon her discharge, bilateral deficits in visual acuity and color
25 discrimination persisted.
26 Similar deficits (metabolic acidosis, visual acuity, and color discrimination) were seen in
27 a man who ingested 300 mL of 75% methanol solution. His blood methanol level was
28 163 mg/dL (1,630 mg/L). An MRI administered 24 hours after hospital admission revealed
29 abnormal hyperintense foci in the putamina, with less intense lesions in the white matter. Like
30 the first subject, a subsequent MRI indicated the foci decreased in size over time, but visual
31 impairments persisted.
32 The third individual, a male, ingested an unspecified amount of a methanolic solution.
33 His blood methanol level was 1,290 mg/dL (12,900 mg/L), and he was in a coma upon hospital
34 admission. An MRI revealed lesions in the putamina and occipital subcortical white matter. A
35 follow-up CT scan was performed after 1 year and showed regression of the putaminal lesions
4-3 DRAFT-DO NOT CITE OR QUOTE
-------
1 but no change in the occipital lesions. Upon his discharge, severe visual impairment remained
2 but no extrapyramidal signs were observed.
3 The last case was a man who became comatose 12 hours after ingesting 100 mL
4 methanol. His blood methanol level at that time was 60 mg/dL (600 mg/L). An MRI revealed
5 lesions in the putamina; at 3 weeks these lesions were observed to have decreased in size. Upon
6 his discharge, the neurological signs had improved but optic neuropathy (in visual evoked
7 potential) was observed.
8 In a separate publication, Hantson et al. (1997b) reported a case of a 26-year-old woman
9 who had ingested 250-500 mL methanol during the 38th week of pregnancy. Her initial blood
10 methanol level was 230 mg/dL (2,300 mg/L) (formate was 33.6 mg/dL or 336 mg/L), yet only a
11 mild metabolic acidosis was indicated. No distress to the fetus was observed upon gynecologic
12 examination. Six days after therapy was initiated (methanol was not present in blood), she gave
13 birth. No further complications with either the mother or newborn were noted.
14 There have been several case reports involving infant or toddler exposures to methanol
15 (De Brabander et al., 2005; Wu et al., 1995; Brent et al., 1991; Kahn and Blum, 1979). The
16 report by Wu et al. (1995) involved a 5-week-old infant with moderate metabolic acidosis and a
17 serum methanol level of 1,148 mg/dL (11,480 mg/L), a level that is ordinarily fatal. However,
18 this infant exhibited no toxic signs and survived without any apparent permanent problems. De
19 Brabander et al. (2005) reported the case of a 3-year-old boy who ingested an unknown amount
20 of pure methanol; at 3 hours after ingestion, the blood methanol level was almost 30 mg/dL (300
21 rng/L). Ethanol infusion as a therapeutic measure was not well tolerated; at 8 hours after
22 ingestion, fomepizole was administered, and blood methanol levels stabilized below 20 mg/dL
23 (200 mg/L), a level above which is considered to be toxic by the American Academy of Clinical
24 Toxicology (Barceloux et al., 2002). Neither metabolic acidosis nor visual impairment was
25 observed in this individual. Hantson et al. (1997a), in their review, touted the efficacy of
26 fomepizole over ethanol in the treatment of methanol poisoning
27 Bilateral putaminal lesions, suggestive of nonhemorrhagic necrosis in the brain of a man
28 who accidentally ingested methanol, were reported by Arora et al. (2005). Approximately
29 10 hours after MRI examination, he developed blurred vision and motor dysfunction. After
30 5 months, visual deficits persisted along with extrapyramidal symptoms. Persistent visual
31 dysfunction was also reported in another methanol poisoning case (Arora et al., 2007); the vision
32 problems developed -46 hours subsequent to the incident.
33 Vara-Castrodeza et al. (2007) applied diffusion-weighted MRI on a methanol-induced
34 comatose woman. Diffusion-weighted MRI provides an image contrast distinct from standard
35 imaging in that contrast is dependent on the molecular motion of water (Schaefer et al., 2000).
36 The neuroradiological findings were suggestive of bilateral putaminal hemorrhagic necrosis,
4-4 DRAFT-DO NOT CITE OR QUOTE
-------
1 cerebral and intraventricular hemorrhage, diffuse cerebral edema, and cerebellar necrosis.
2 Diffusion-weighted MRI allows for differentiation of restricted diffusion which is indicative of
3 nonviable tissue. In this case, treatment for acidosis (blood methanol levels had risen to
4 1,000 mg/L) was unsuccessful and the patient died.
5 Emergency treatment was unable to save the life of a 38-year-old man who presented
6 with abdominal pain and convulsions after methanol intoxication (Henderson and Brubacher,
7 2002). A review of a head CT scan performed before the individual went into respiratory arrest
8 revealed bilateral globus pallidus ischemia.
9 Discrete lesions of the putamen, cerebral white matter, and corpus callosum were
10 observed upon MRI (8 days post ingestion) in a man exposed to methanol (blood level 370
11 mg/L) complaining of vision loss (Keles et al., 2007). Standard treatments corrected the acidosis
12 (pH 6.8), and at 1-month follow-up, his cognitive function improved but blindness and bilateral
13 optic atrophy were described as permanent. The follow-up MRI showed persistent putaminal
14 lesions with cortical involvement.
15 Fontenot and Pelak (2002) described a case of a woman who presented with persistent
16 blurred vision and a worsening mental status 36 hours after ingestion of an unspecified amount
17 of methanol. The initial CT scan revealed mild cerebral edema. The blood methanol level at this
18 time was 86 mg/dL (860 mg/L). A repeat CT scan 48 hours after presentation showed
19 hypodensities in the putamen and peripheral white matter. One month after discharge, cognitive
20 function improved, and the patient experienced only a mild lower-extremity tremor.
21 Putaminal necrosis and edema of the deep white matter (the corpus callosum was not
22 affected) was found upon MRI examination of a 50-year-old woman who apparently ingested an
23 unknown amount of what was believed to be pure laboratory methanol (Kuteifan et al., 1998).
24 Her blood methanol level was 39.7 mM (127 mg/dL; 1,272 mg/L) upon hospital admission and
25 dropped to 102 mg/dL (1,020 mg/L) at 10 hours and to 71 mg/dL (710 mg/L) at 34 hours. The
26 woman, a chronic alcoholic, was in a vegetative state when found and did not improved over the
27 course of a year.
28 MRI and CT scans performed on a 51-year-old man with generalized seizures who had a
29 blood methanol level of 95 mM (304 mg/dL; 3,044 mg/L) revealed bilateral hemorrhagic
30 necrosis of the putamen and caudate nuclei (Gaul et al., 1995). In addition, there was extensive
31 subcortical necrosis and bilateral necrosis of the pontine tegmentum and optic nerve. The patient
32 died several hours after the scans were performed.
33 The relation of methanol overexposure to brain hemorrhage was a focus of the report by
34 Phang et al. (1988), which followed the treatment of 7 individuals, 5 of whom died within
35 72 hours after hospital admission. In two of the deceased individuals, CT scans and autopsy
36 revealed putaminal hemorrhagic necrosis. The investigators postulated that the association of
4-5 DRAFT-DO NOT CITE OR QUOTE
-------
1 methanol with hemorrhagic necrosis may be complicated by the use of heparin during
2 hemodialysis treatment for acidosis
3 Treatment of two men who had drunk a solution containing 58% methanol and presented
4 with impaired vision, coma, and seizures was discussed in a case report by Bessell-Browne and
5 Bynevelt (2007). A CT scan on one individual revealed bilateral putaminal and cerebral lesions.
6 Blood methanol levels were 21 mg/L. This individual, despite standard treatments, never
7 regained consciousness. The second individual, upon MRI, showed scattered hemorrhage at the
8 grey-white interface of the cerebral hemispheres.
9 There have been two case reports (Adanir et al., 2005; Downie et al., 1992) that involved
10 percutaneous and inhalation exposure. Use of a methanol-containing emollient by a woman with
11 chronic pain led to vision loss, hyperventilation and finally, coma (Adanir et al., 2005).
12 Subsequent to standard treatment followed by hospital discharge, some visual impairment and
13 CNS decrements remained. The methanol blood threshold for ocular damage and acidosis
14 appeared to be -20 mg/L. Dutkiewicz et al. (1980) have determined the skin absorption rate to
15 be 0.192 mg/cm2/minute. In the case report of Aufderheide et al. (1993), two firefighters were
16 transiently exposed to methanol by inhalation and the percutaneous route. Both only complained
17 of a mild headache and had blood methanol levels of 23 and 16 mg/dL (230 and 160 mg/L),
18 respectively.
19 Bebarta et al. (2006) conducted a prospective observational study of seven men who had
20 purposefully inhaled a methanol-containing product. Four had a blood methanol level upon
21 hospital presentation of >24 mg/dL (240 mg/L); the mean formic acid level was 71 |ig/dL. One
22 individual had a blood methanol level of 86 mg/dL (860 mg/L) and a blood formic acid level of
23 250 |ig/mL upon hospital admission. This latter individual was treated with fomepizole. No
24 patient had an abnormal ophthalmologic examination. All seven stabilized quickly and acidosis
25 was normalized in 4 hours.
26 Numerous other case reports documenting putaminal necrosis/hemorrhage and/or
27 blindness have been reported (Blanco et al., 2006; Feany et al., 2001; Hsu et al., 1997; Pelletier
28 etal., 1992; Chen etal., 1991).
29 Hovda et al. (2005) presented a combined prospective and retrospective case series study
30 of 51 individuals in Norway (39 males and 12 females, many of whom were alcoholics) who
31 were hospitalized after consuming tainted spirits containing 20% methanol and 80% ethanol. In
32 general, serum methanol concentrations were highest among those most severely affected. The
33 poor outcome was closely correlated with the degree of metabolic acidosis. It was noted by the
34 investigators that the concomitant consumption of ethanol prevented more serious sequelae in
35 2/5 individuals who presented with detectable ethanol levels and were not acidotic despite 2
36 having the highest blood methanol levels. However, others with detectable levels of ethanol
4-6 DRAFT-DO NOT CITE OR QUOTE
-------
1 along with severe metabolic acidosis (two of whom died) presumably had subtherapeutic levels
2 of ethanol in their system.
3 In a later report, Hovda et al. (2007) focused on formate kinetics in a 63-year-old male
4 who died 6 days after being admitted to the hospital with headache, vomiting, reduced vision,
5 and dizziness. The investigators speculated that the prolonged metabolic acidosis observed (T1/2
6 for formic acid was 77 hours before dialysis, compared to a typical normal range of 2.5-
7 12 hours) may have been related to retarded formate elimination.
8 Hovda and colleagues (Hunderi et al., 2006) found a strong correlation between blood
9 methanol concentration and the osmolal gap (R2 = 0.92) among 17 patients undergoing dialysis
10 after consuming methanol-contaminated spirits. They concluded that the osmolal gap could be
11 taken as a priori indication of methanol poisoning and be used to guide initiation and duration of
12 dialysis. As they indicated, many hours of dialysis could be safely dispensed with. The osmolal
13 gap pertains to the effect that methanol (and other alcohols) has on the depression of the freezing
14 point of blood in the presence of normal solutes. Braden et al. (1993) demonstrated in case
15 studies that the disappearance of the osmolal gap correlates with the correction of acidosis; they
16 cautioned that methanol and ethanol should not be assumed to be the main factors in causing
17 osmolal gap as glycerol and acetone and its metabolites can as well. A more detailed discussion
18 of the anion and osmolal gap has been provided by Henderson and Brubacher (2002).
19 Hassanian-Moghaddam et al. (2007) compiled data on the prognostic factor relating to
20 outcome in methanol-poisoning cases in Iran. They examined 25 patients, 12 of whom died; 3
21 of the survivors were rendered blind. There was a significant difference in mean pH of the first
22 arterial blood gas measurements of those who subsequently died compared with survivors. It
23 was concluded that poor prognosis was associated with pH <7, coma upon admission, and
24 >24-hours delay from intake to admission.
25 The use of blood methanol levels as predictors of outcome is generally not recommended
26 (Barceloux et al., 2002). These investigators cited differences in sampling time, ingestion of
27 ethanol, and levels of toxic (e.g., formic acid) metabolites among the complicating factors. As
28 an illustration, the case report by Prabhakaran et al. (1993) cites two women who ingested a
29 methanol solution (photocopying diluent) at about the same time, were admitted to the hospital
30 about the same time (25-26 hours after ingestion) and had identical plasma methanol
31 concentrations (83 mg/dL; 830 mg/L) upon admission, but different outcomes. Patient #1 was in
32 metabolic acidosis and had an unstable conscious state even after treatment. Upon discharge at
33 day 6, there were no apparent sequelae. Patient #2 had severe metabolic acidosis, fixed and
34 dilated pupils, and no brain stem reflexes. This patient died at day 3 even though therapeutic
35 measures had been administered.
4-7 DRAFT-DO NOT CITE OR QUOTE
-------
1 In a discussion of 3 fatal methanol-overexposure cases, Andresen et al. (2008) found
2 antemortem blood methanol levels of 540 and 740 mg/dL (5,400 and 7,400 mg/L) in 2
3 individuals. At autopsy brain stem blood levels were 738 and 1,008 mg/dL (7,380 and
4 10,080 mg/L), respectively. These brain levels were much higher than blood levels postmortem.
5 Autopsy revealed brain and pulmonary edema in all three individuals; in the two who had the
6 longer survival times, there was hemorrhagic necrosis of the putamen and hemorrhages of the
7 tissue surrounding the optic nerve. In their study of 26 chronic users of methylated spirits,
8 Meyer et al. (2000) found that the best predictor of death or a poor outcome in chronic abusers
9 was a pH <7.0; there was no correlation between blood methanol levels and outcome. Mahieu
10 et al. (1989) considered a latency period before treatment exceeding 10 hours and a blood
11 formate level >50 mg/dL (500 mg/L) as predictive of possible permanent sequelae. Liu et al.
12 (1998) in their examination of medical records of 50 patients treated for methanol poisoning over
13 a 10-year period found that 1) deceased patients had a higher mean blood methanol level than
14 survivors; and 2) initial arterial pH levels <7.0 (i.e., severe metabolic acidosis). Coma or seizure
15 was also associated with higher mortality upon hospital admission.
16 Numerous cases of methanol poisoning have been documented in a variety of countries.
17 In Tunisia, 16 cases of methanol poisoning were discussed by Brahmi et al. (2007). Irreversible
18 blindness occurred in two individuals, with others reporting CNS symptoms, GI effects, visual
19 disturbances, and acidosis. Putaminal necrosis was also described in case reports from Iran
20 (Sefidbakht et al., 2007). Of 634 forensic autopsies carried out in Turkey during 1992-2003, 18
21 appeared to be related to methanol poisoning (Azmak, 2006). Brain edema and focal necrosis of
22 the optic nerve were among various sequelae noted. Dethlefs and colleagues (Naraqi et al.,
23 1979; Dethlefs and Naraqi, 1978) described permanent ocular damage in 8/24 males who
24 ingested methanol in Papua New Guinea.
25 In summary, most cases of accidental/intentional methanol poisoning reveal a common
26 set of symptoms, many of which are likely to be presented upon hospital admission. These
27 include:
28 • blurred vision and bilateral or unilateral blindness
29 • convulsions, tremors, and coma
30 • nausea, headache, and dizziness
31 • abdominal pain
32 • diminished motor skills
33 • acidosis
34 • dyspnea
35 • behavioral and/or emotional deficits
36 • speech impediments
4-8 DRAFT-DO NOT CITE OR QUOTE
-------
1 Acute symptoms generally are nausea, dizziness, and headache. In the case reports cited
2 above, the onset of symptom sets as well as their severity varies depending upon how much
3 methanol was ingested, whether or not and when appropriate treatment was administered, and
4 individual variability. A longer time between exposure and treatment, with few exceptions,
5 results in more severe outcomes (e.g., convulsions, coma, blindness, and death). The diminution
6 of some acute and/or delayed symptoms may reflect concomitant ingestion of ethanol or how
7 quickly therapeutic measures (one of which includes ethanol infusion) were administered in the
8 hospital setting.
9 Those individuals who are in a metabolic acidotic state (e.g., pH <7.0) are typically the
10 individuals who manifest the more severe symptoms. Many case reports stress that, unlike blood
11 pH levels <7.0, blood levels of methanol are not particularly good predictors of health outcome.
12 According to a publication of the American Academy of Clinical Toxicology (Barceloux et al.,
13 2002), "the degree of acidosis at presentation most consistently correlates with severity and
14 outcome."
15 As the case reports demonstrate, those individuals who present with more severe
16 symptoms (e.g., coma, seizures, severe acidosis) generally exhibit higher mortality (even after
17 treatment) than those without such symptoms. In survivors of poisoning, persistence or
18 permanence of vision decrements and particularly blindness often have been observed
19 Correlation of symptomatology with blood levels of methanol has been shown to vary
20 appreciably between individuals. Blood methanol levels in the case reports involving ingestion
21 ranged from values of 30 to over 1,000 mg/dL (300 to over 10,000 mg/L). The lowest value
22 (20 mg/dL; 200 mg/L) reported (Adanir et al., 2005) involved a case of percutaneous absorption
23 (with perhaps associated inhalation exposure) that led to vision and CNS deficits after hospital
24 discharge. In one case report (Rubinstein et al., 1995) involving ingestion, coma and subsequent
25 death were associated with an initial blood methanol level of 36 mg/dL (360 mg/L).
26 Upon MRI and CT scans, the more seriously affected individuals typically have focal
27 necrosis in both brain white matter and more commonly, in the putamen. Bilateral hemorrhagic
28 and nonhemorrhagic necrosis of the putamen is considered by many radiologists as the most
29 well-known sequelae of methanol overexposure.
4.1.2. Occupational Studies
30 Occupational health studies have been carried out to investigate the potential effects of
31 chronic exposure to lower levels of methanol than those seen in acute poisoning cases such as
32 those described above. For example, Frederick et al. (1984) conducted a health hazard
33 evaluation on behalf of the National Institute for Occupational Safety and Health (NIOSH) to
4-9 DRAFT-DO NOT CITE OR QUOTE
-------
1 determine if vapor from duplicating fluid (which contains 99% methanol) used in mimeograph
2 duplicating machines caused adverse health effects in exposed persons. A group of 84 teacher's
3 aides were selected for study, 66 of whom responded with a completed medical questionnaire. A
4 group of 297 teachers (who were not exposed to methanol vapors to the same extent as the
5 teacher's aides) completed questionnaires as a control group. A 15-minute breathing zone
6 sample was taken from 21 duplicators, 15 of which were greater than the NIOSH-recommended
7 short term ceiling concentration of 800 ppm (1048 mg/m3). The highest breathing zone
8 concentrations were in the vicinity of duplicators for which no exhaust ventilation had been
9 provided (3,080 ppm [4,036 mg/m3] was the highest value recorded). Upon comparison of the
10 self-described symptoms of the 66 teacher's aides with those of 66 age-matched teachers chosen
11 from the 297 who responded, the number of symptoms potentially related to methanol were
12 significantly higher in the teacher's aides. These included blurred vision (22.7 versus 1.5%),
13 headache (34.8 versus 18.1%), dizziness (30.3 versus 1.5%), and nausea (18 versus 6%). By
14 contrast, symptoms that are not usually associated with methanol exposure (painful urination,
15 diarrhea, poor appetite, and jaundice) were similar in incidence among the groups.
16 To further investigate these disparities, NIOSH physicians (not involved in the study)
17 defined a hypothetical case of methanol toxicity by any of the following four symptom
18 aggregations: 1) visual changes; 2) one acute symptom (headache, dizziness, numbness,
19 giddiness, nausea or vomiting) combined with one chronic symptom (unusual fatigue, muscle
20 weakness, trouble sleeping, irritability, or poor memory); 3) two acute symptoms; or 4) three
21 chronic symptoms. By these criteria, 45% of the teacher's aides were classified as being
22 adversely affected by methanol exposure compared to 24% of teachers (p < 0.025). Those
23 teacher's aides and teachers who spent a greater amount of time using the duplicators were
24 affected at a higher rate than those who used the machines for a lower percentage of their work
25 day.
26 Tanner (1992) reviewed the occupational and environmental causes of Parkinsonism,
27 spotlighting the potential etiological significance of manganese, carbon monoxide, repeated head
28 trauma (such as suffered by boxers), and exposure to solvents. Among the latter, Tanner (1992)
29 discussed the effects of methanol and n-hexane on the nervous system. Acute methanol
30 intoxication resulted in inebriation, followed within hours by GI pain, delirium, and coma.
31 Tanner (1992) pinpointed the formation of formic acid, with consequent inhibition of
32 cytochrome oxidase, impaired mitochondrial function, and decreased ATP formation as relevant
33 biochemical and physiological changes for methanol exposure. Nervous system injury usually
34 includes blindness, Parkinson-like symptoms, dystonia, and cognitive impairment, with injury to
35 putaminal neurons most likely underlying the neurological responses.
4-10 DRAFT-DO NOT CITE OR QUOTE
-------
1 Kawai et al. (1991) carried out a biomarker study in which 33 occupationally exposed
2 workers in a factory making methanol fuel were exposed to concentrations of methanol of up to
3 3,577 ppm (4,687 mg/m3), as measured by personal samplers of breathing zone air. Breathing
4 zone exposure samples were correlated with the concentrations of methanol in urine at the end of
5 the shift in 38 exposed individuals and 30 controls (r = 0.82). Eleven of 22 individuals who
6 experienced high exposure to methanol (geometric mean of 459 ppm [601 mg/m3]) complained
7 of dimmed vision during work while 32% of this group of workers experienced nasal irritation.
8 These incidences were statistically significant compared to those of persons who worked in low-
9 exposure conditions (geometric mean of 31 ppm [41 mg/m3]). One 38-year-old female worker
10 who had worked at the factory for only 4 months reported that her visual acuity had undergone a
11 gradual impairment. She also displayed a delayed light reflex.
12 Lorente et al. (2000) carried out a case control study of 100 mothers whose babies had
13 been born with cleft palates. Since all of the mothers had worked during the first trimester,
14 Lorente et al. (2000) examined the occupational information for each subject in comparison to
15 751 mothers whose babies were healthy. Industrial hygienists analyzed the work histories of all
16 subjects to determine what, if any, chemicals the affected mothers may have been exposed to
17 during pregnancy. Multivariate analysis was used to calculate odds ratios, with adjustments
18 made for center of recruitment, maternal age, urbanization, socioeconomic status, and country of
19 origin. Occupations with positive outcomes for cleft palate in the progeny were hairdressing
20 (OR = 5.1, with a 95% confidence interval [CI] of 1.0-26) and housekeeping (OR = 2.8, with a
21 95% CI of 1.1-7.2). Odds ratios for cleft palate only and cleft lip with or without cleft palate
22 were calculated for 96 chemicals. There seemed to be no consistent pattern of association for
23 any chemical or group of chemicals with these impairments, and possible exposure to methanol
24 was negative for both outcomes.
4.1.3. Controlled Studies
25 Two controlled studies have evaluated humans for neurobehavioral function following
26 exposure to -200 ppm (262 mg/m3) methanol vapors in a controlled setting. The occupational
27 TLV established by the American Conference of Governmental Industrial Hygienists (ACGIH,
28 2000) is 200 ppm (262 mg/m3). In a pilot study by Cook et al. (1991), 12 healthy young men
29 (22-32 years of age) served as their own controls and were tested for neurobehavioral function
30 following a random acute exposure to air or 191 ppm (250 mg/m3) methanol vapors for
31 75 minutes. The majority of results in a battery of neurobehavioral endpoints were negative.
32 However, statistical significance was obtained for results in the P-200 and N1-P2 component of
33 event-related potentials (brain wave patterns following light flashes and sounds), the Sternberg
34 memory task, and subjective evaluations of concentration and fatigue. As noted by the Cook et
4-11 DRAFT-DO NOT CITE OR QUOTE
-------
1 al. (1991), effects were mild and within normal ranges. Cook et al. (1991) acknowledged
2 limitations in their study design, such as small sample size, exposure to only one concentration
3 for a single duration time, and difficulties in masking the methanol odor from experimental
4 personnel and study subjects.
5 In a randomized double-blind study, neurobehavioral testing was conducted on 15 men
6 and 11 women (healthy, aged 26-51 years) following exposure to 200 ppm (262 mg/m3)
7 methanol or water vapors for 4 hours (Chuwers et al., 1995); subjects served as their own
8 controls in this study. Exposure resulted in elevated blood and urine methanol levels (up to peak
9 levels of 6.5 mg/L and 0.9 mg/L, respectively) but not formate concentrations. The majority of
10 study results were negative. No significant findings were noted for visual, neurophysiological,
11 or neurobehavioral tests except for slight effects (p < 0.05) on P-300 amplitude (brain waves
12 following exposure to sensory stimuli) and Symbol Digit testing (ability to process information
13 and psychomotor skills). Neurobehavioral performance was minimally affected by methanol
14 exposure at this level. Limitations noted by Chuwers et al. (1995) are that studies of alcohol's
15 affect on P-300 amplitude suggest that this endpoint may be biased by unknown factors and
16 some experimenters and subjects correctly guessed if methanol was used.
17 Although the slight changes in P-200 and P-300 amplitude noted in both the Chuwers
18 et al. (1995) and Cook et al. (1991) studies may be an indication of moderate alterations in
19 cognitive function, the results of these studies are generally consistent and suggest that the
20 exposure concentrations employed were below the threshold for substantial neurological effects.
21 This is consistent with the data from acute poisoning events which have pointed to a serum
22 methanol threshold of 200 mg/L for the instigation of acidosis, visual impairment, and CNS
23 deficits.
24 Mann et al. (2002) studied the effects of methanol exposure on human respiratory
25 epithelium as manifested by local irritation, ciliary function, and immunological factors. Twelve
26 healthy men (average age 26.8 years) were exposed to 20 and 200 ppm (26.2 and 262 mg/m3,
27 respectively) methanol for 4 hours at each concentration; exposures were separated by 1-week
28 intervals. The 20 ppm (26.2 mg/m3) concentration was considered to be the control exposure
29 since previous studies had demonstrated that subjects can detect methanol concentrations of
30 20 ppm (26.2 mg/m3) and greater. Following each single exposure, subclinical inflammation was
31 assessed by measuring concentrations of interleukins (IL-8, IL-lp, and IL-6) and prostaglandin
32 E2 in nasal secretions. Mucociliary clearance was evaluated by conducting a saccharin transport
33 time test and measuring ciliary beat frequency. Interleukin and prostaglandin data were
34 evaluated by a 1-tailed Wilcoxon test, and ciliary function data were assessed by a 2-tailed
35 Wilcoxon test. Exposure to 200 (262 mg/m3) versus 20 ppm (26.2 mg/m3) methanol resulted in a
36 statistically-significant increase in IL-lp (median of 21.4 versus 8.3 pg/mL) and IL-8 (median of
4-12 DRAFT-DO NOT CITE OR QUOTE
-------
1 424 versus 356 pg/mL). There were no significant effects on IL-6 and prostaglandin E2
2 concentration, ciliary function, or on the self-reported incidence of subjective symptoms of
3 irritation. The authors concluded that exposure to 200 ppm (262 mg/m3) methanol resulted in a
4 subclinical inflammatory response.
5 In summary, adult human subjects acutely exposed to 200 ppm (262 mg/m3) methanol
6 have experienced slight neurological (Chuwers et al., 1995) and immunological effects
7 (increased subclinical biomarkers for inflammation) with no self-reported symptoms of irritation
8 (Mann et al., 2002). These exposure levels were associated with peak methanol blood levels of
9 6.5 mg/L (Chuwers et al., 1995), which is approximately threefold higher than background
10 methanol blood levels reported for adult human subjects on methanol-restrictive diets
11 (Table 3-1). Nasal irritation effects have been reported by adult workers exposed to 459 ppm
12 (601 mg/m3) methanol (Kawai et al., 1991). Frank effects such as blurred vision, bilateral or
13 unilateral blindness, coma, convulsions/tremors, nausea, headache, abdominal pain, diminished
14 motor skills, acidosis, and dyspnea begin to occur as blood levels approach 200 mg methanol/L,
15 while 800 mg/L appears to be the threshold for lethality. Data for subchronic, chronic or in utero
16 human exposures are very limited and inconclusive.
4.2. ACUTE, SUBCHRONIC AND CHRONIC STUDIES AND CANCER BIO ASSAYS IN
ANIMALS—ORAL AND INHALATION
17 A number of studies in animals have investigated the acute, subchronic, and chronic
18 toxicity of methanol. Most are via the inhalation route. Presented below are summaries of these
19 investigations.
4.2.1. Oral Studies
4.2.1.1. Acute Toxicity
20 Although there are few studies that have examined the short-term toxic effects of
21 methanol via the oral route, a number of median lethal dose (LD50) values have been published
22 for the compound. As listed in Lewis (1992), these include 5,628 mg/kg in rats, 7,300 mg/kg in
23 mice, and 7,000 mg/kg in monkeys.
4.2.1.2. Subchronic Toxicity
24 An oral repeat dose study was conducted by the EPA (1986c) in rats. Sprague-Dawley
25 rats (30/sex/dose) were gavaged with 0, 100, 500, or 2,500 mg/kg-day of methanol. Six weeks
26 after dosing, 10 rats/sex/dose group were subjected to interim sacrifice, while the remaining rats
27 continued on the dosing regimen until the final sacrifice (90 days). This study generated data on
28 weekly body weights and food consumption, clinical signs of toxi city, ophthalmologic
29 evaluations, mortality, blood and urine chemistry (from a comprehensive set of hematology,
4-13 DRAFT-DO NOT CITE OR QUOTE
-------
1 serum chemistry, and urinalysis tests), and gross and microscopic evaluations for all test animals.
2 Complete histopathologic examinations of over 30 organ tissues were done on the control and
3 high-dose rats. Histopathologic examinations of livers, hearts, and kidneys and all gross lesions
4 seen at necropsy were done on low-dose and mid-dose rats. There were no differences between
5 dosed animals and controls in body weight gain, food consumption, or upon gross or microscopic
6 evaluations. Elevated levels (p < 0.05 in males) of serum alanine transaminase (ALT)14 and
7 serum alkaline phosphatase (SAP), and increased (but not statistically significant) liver weights
8 in both male and female rats suggest possible treatment-related effects in rats bolus dosed with
9 2,500 mg methanol/kg-day despite the absence of supportive histopathologic lesions in the liver.
10 Brain weights of high-dose group (2,500 mg/kg-day) males and females were significantly less
11 than those of the control group at terminal sacrifice. Based on these findings, 500 mg/kg-day of
12 methanol is considered an NOEL from this rat study.
4.2.1.3. Chronic Toxicity
13 A report by Soffritti et al. (2002a) summarized a European Ramazzini Foundation (ERF)
14 chronic duration experimental study of methanol15 in which the compound was provided to
15 100 Sprague-Dawley rats/sex/group ad libitum in drinking water at concentrations of 0, 500,
16 5,000, and 20,000 ppm (v/v). The animals were 8 weeks old at the beginning of the study. In
17 general, ERF does not randomly assign animals to treatment groups, but assigns all animals from
18 a given litter to the same treatment group (Bucher, 2002). All rats were exposed for up to
19 104 weeks, then maintained until they died naturally. Rats were housed in groups of 5 in
20 Makrolon cages (41 x 25 x 15 cm) in a room that was maintained at 23 ± 2°C and 50-60%
21 relative humidity. The in-life portion of the experiment ended at 153 weeks with the death of the
22 last animal. Mean daily drinking water, food consumption, and body weights were monitored
23 weekly for the first 13 weeks, every 2 weeks thereafter for 104 weeks, then every 8 weeks until
24 the end of the experiment. Clinical signs were monitored 3 times/day, and the occurrence of
25 gross changes was evaluated every 2 weeks. All rats were necropsied at death then underwent
26 histopathologic examination of organs and tissues.16
14 Also known as serum glutamate pyruvate transaminase (SGPT)
15 Soffritti et al. (2002a) report that methanol was obtained from J.T. Baker, Deventer, Holland, purity grade 99.8%.
Histopathology was performed on the following organs and tissues: skin and subcutaneous tissue, brain, pituitary
gland, Zymbal glands, parotid glands, submaxillary glands, Harderian glands, cranium (with oral and nasal cavities
and external and internal ear ducts) (5 sections of head), tongue, thyroid and parathyroid, pharynx, larynx, thymus
and mediastinal lymph nodes, trachea, lung and mainstem bronchi, heart, diaphragm, liver, spleen, pancreas,
kidneys, adrenal glands, esophagus, stomach (fore and glandular), intestine (four levels), urinary bladder, prostate,
gonads, interscapular fat pad, subcutaneous and mesenteric lymph nodes, and any other organs or tissues with
pathologic lesions.
4-14 DRAFT-DO NOT CITE OR QUOTE
-------
1 Soffritti et al. (2002a) reported no substantial dose-related differences in survival, but no
2 data were provided. Using individual animal data available from the ERF website,17 Cruzan
3 (2009) reports that male rats treated with methanol generally survived better than controls, with
4 50% survival occurring at day 629, 686, 639 and 701 in the 0, 500, 5,000, and 20, 000 mg/L
5 groups, respectively. There were no significant differences in survival between female control
6 and treatment groups, with 50% survival occurring at day 717, 691, 678 and 708 in the 0, 500,
7 5,000, and 20, 000 mg/L groups, respectively. Body weight and water and food consumption
8 were monitored in the study, but the data were not documented in the published report.
9 However, based on data available from the ERF website, average doses of 0, 53.2, 524, and
10 1,780 mg/kg-day in males and 0, 66.0, 624.1, and 2,177 mg/kg-day in females could be
11 calculated (see Appendix E) from drinking water concentrations of 0, 500, 5,000, and
12 20,000 ppm.
13 Soffritti et al. (2002a) reported that water consumption in high-dose females was reduced
14 compared to controls between 8 and 56 weeks and that the mean body weight in high-dose males
15 tended to be higher than that of control males. Overall, there was no pattern of compound-
16 related clinical signs of toxicity, and the available data did not provide any indication that the
17 control group was not concurrent with the treated group (Cruzan, 2009). Soffritti et al. (2002a)
18 further reported that there were no compound-related signs of gross pathology or histopathologic
19 lesions indicative of noncancer toxicological effects in response to methanol.
20 Soffritti et al. (2002a) reported a number of oncogenic responses to methanol (Table 4-2),
21 principally hemolymphoreticular neoplasms, the majority of which were reported to be lympho-
22 immunoblastic lymphomas. In ERF bioassays, including this methanol study,
23 hemolymphoreticular neoplasms are generally divided into specific histological types
24 (lymphoblastic lymphoma, lymphoblastic leukemia, lymphocytic lymphoma, lympho-
25 immunoblastic lymphoma, myeloid leukemia, histocytic sarcoma, and monocytic leukemia) for
26 identification purposes. According to Soffritti et al. (2007), the overall incidence of
27 hemolymphoreticular tumors (lymphomas/leukemias) in ERF studies is 13.3% (range, 4.0-
28 25.0%) in female historical controls (2,274 rats) and 20.6% (range, 8.0-30.9%) in male historical
29 controls (2,265 rats). The high-dose responses, shown in Table 4-2, of 28% and 40% for females
30 and males, respectively, are above their corresponding historical ranges.18
31 The National Toxicology Program (NTP) does not routinely subdivide lymphomas into
32 specific histological types as was done by the ERF. In 2004, a Pathology Working Group
33 (PWG) of National Institute of Environmental Health Sciences (NIEHS) performed a limited
34 review of about 75 slides provided by ERF as representive of lesions in Sprague-Dawley rats
17 http://www.ramazzini.it/fondazione/foundation.asp.
18 While historical control data can be informative, for reasonably well-conducted studies, it should not take
precedence over concurrent controls or appropriate statistical dose-response trend tests.
4-15 DRAFT-DO NOT CITE OR QUOTE
-------
1 associated with aspartame exposure (EPS A, 2006; Hailey, 2004). The primary objective of this
2 review was to "provide a second opinion for this set of lesions by a group of pathologists
3 experienced in Toxicologic Pathology."19 Eleven of the slides reviewed by the PWG were
4 related to lymphomas, and three of these had been classified by ERF as lympho-immunoblastic.
5 The PWG concluded that "The diagnoses of lymphatic and histocytic neoplasms in the cases
6 reviewed were generally confirmed" (Hailey, 2004). In particular, the PWG accepted the more
7 specific diagnoses of ERF when the lesions were considered to be consistent with a neoplasm of
8 lymphocytic, histocytic, monocytic, and/or myeloid origin. The PWG noted, however, that while
9 lymphoblastic lymphomas, lymphocytic lymphomas, lympho-immunoblastic lymphomas, and
10 lymphoblastic leukemias as malignant lymphomas can be combined, myeloid leukemias,
11 histocytic sarcomas, and monocytic leukemia should be treated as separate malignancies and not
12 combined with the other lymphomas since they are of different cellular origin (Hailey, 2004).
13 McConnell et al. (1986) and Cruzan (2009) have also noted that myeloid leukemia, histocytic
14 sarcoma, and monocytic leukemia are of a different cell line and are not typically combined with
15 other lymphomas for statistical significance or dose-response modeling. Consistent with these
16 judgments, EPA has not included the myeloid leukemia, histocytic sarcoma, and monocytic
17 leukemia in combination with lymphoblastic lymphoma, lymphoblastic leukemia, lymphocytic
18 lymphoma, and lympho-immunoblastic lymphoma in its consideration of tumorgenic responses
19 reported by ERF (see Section 5.4.1; Table 5-6). Thus, EPA's analysis of this tumorogenic
20 response differs from the lymphoreticular tumor response shown in Table 4-2 and reported by
21 Soffriti etal. (2002a). As described in Section 5.4.1.1, EPA's reassessment indicates a
22 significant increase in tumor response at the two highest doses for males and across all doses for
23 females (Fisher's exact,p < 0.05), as well as a significant dose-response trend (Cochran
24 Armitage trend test; p < 0.05).
25 Schoeb et al. (2009) have suggested that the interpretation of lesions in ERF studies,
26 including the Soffritti et al. (2002a) methanol study, may have been confounded by a respiratory
27 infection referred to as Mycoplasma pulmonis (M. pulmonis) disease and that lesions of this
28 disease were interpreted as lymphoma. They noted that lympho-immunoblastic lymphoma is not
29 listed as a lymphoma type in rats in available reference sources and that the cellular morphology
30 of the lung lympho-immunoblastic lymphomas reported by ERF for aspartame (Soffritti et al.,
31 2005) and MTBE (Belpoggi et al., 1999) studies are more consistent withM pulmonis disease.
32 As noted above, an NIEHS PWG (Hailey, 2004) has confirmed the ERF diagnosis of the several
33 lymphomas, including three lymphomas from the lung, thymus and medullary lymph node and
34 mesenteric lymph node that were characterized by ERF as "lympho-immunoblastic." Hailey
19 This review was not considered a "peer review" of the pathology data from this study. As noted by Hailey (2004),
"a peer review would necessitate a review of the study data by a second party, and selection and examination of
lesions based upon that data review."
4-16 DRAFT-DO NOT CITE OR QUOTE
-------
1 (2004) reports that the PWG "accepted their [ERF's] more specific diagnosis if the lesion was
2 considered to be consistent with a neoplasm of lymphocytic, histocytic, monocytic, and/or
3 myeloid origin." The concerns of Schoeb et al. (2009) regarding the possibility of infection
4 confounding the interpretation of lung lesions in the ERF study are not unfounded. Chronic
5 inflammatory changes are apparently a common finding in ERF studies (Caldwell et al., 2008),
6 probably caused by the ERF bioassay design that does not employ specific pathogen-free (SPF)
7 rats (EFSA, 2006) and allows the rats to live out their "natural life span" in the absence of
8 disease barriers (e.g., fully enclosed cages). However, an infection of the ERF colony with
9 M. pulmonis has not been confirmed (Caldwell et al., 2008) and, without confirmation, cannot be
10 used to discount an existing dose-related trend (U.S. EPA, 2005). Further, even if the rats of the
11 ERF methanol study were suffering from a respiratory infection that confounded the
12 interpretation of lung lesions, 60% of reported lymphoma incidences involved other organ
13 systems, and the dose-response for lymphomas in other organ systems is not remarkably
14 different than for all lymphomas (see analysis in Section 5.4.3.2).
15 Another cancer response, reported by Soffritti et al. (2002a), that is considered to be
16 potentially related to methanol exposure was an increase in rare hepatocellular carcinomas in
17 male rats. Although the increase was not statistically increased compared to concurrent controls,
18 EPA has analyzed historical data for this tumor type in this species and determined that the
19 incidence in all dose groups was significantly elevated relative to historical controls (Fisher's
20 exact/? < 0.05 for all doses and/? < 0.01 for the high-dose group). The historical control group
21 (n = 407) used was the combined control groups from ERF studies for which individual animal
22 pathology data have been made available via the ERF website20 and include data for methanol,
23 formaldehyde, aspartame, MTBE, and TAME.
24 As noted in Table 4-2, increased incidences of carcinomas of the ear ducts and
25 osteosarcomas of the head were reported for both female and male rats, with a statistically
26 significant increase in only the high-dose male ear duct carcinomas. Ear duct carcinomas are a
27 rare finding in Charles River rats and NTP historical databases of Sprague-Dawley rats (Cruzan,
28 2009). In their limited review of pathology slides from the ERF aspartame bioassay (Soffritti
29 et al., 2006, 2005), NTP pathologists interpreted a majority of such head pathologies, including
30 in the ear duct, as being hyperplastic in nature, not carcinogenic (EFSA, 2006; Hailey, 2004).
31 Soffritti et al. (2002a) also noted an increased incidence of testicular hyperplasia in high-dose
32 males and uterine sarcomas in high-dose females compared to controls. However, these
33 increases were not statistically significant and were within historical control ranges for this
34 species and strain (NTP, 2007, 1999; Haseman et al., 1998). The group-specific total number of
35 malignant tumors was also shown to increase with dose in both sexes of rats.
20 http://www.ramazzini.it/fondazione/foundation. asp.
4-17 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-2. Incidence of carcinogenic
methanol in drinking water for up to
responses in Sprague-Dawley rats exposed to
2 years
Tissues/affected sites
Ear duct (carcinomas)
Head (osteosarcomas)
Hemolymphoreticular tumors
Liver (hepatocarcinomas)
Testis (interstitial cell adenomas)
Total malignant tumors
Dose (mg/kg-day)
Males
0
9/100
6/100
28/100
0/100
12/100
50/100
53.2
13/100
6/100
35/100
2/100
9/100
55/100
524
17/100
13/100
36/100
2/100
13/100
64/100
1780
24/100b
11/100
40/100
3/100
17/100
70/100b
Females
0
9/100
1/100
13/100
0/100
43/100
66.0
8/100
4/100
24/100
0/100
48/100
624.1
16/100
3/100
24/100
1/100
48/100
2177
19/100
6/100
28/1003
0/100
63/100b
<_;
"p < 0.05 using the ^ test.
V < 0.01 using the -£ test.
Source: Soffritti et al. (2002a).
1 Apaja (1980) performed dermal and drinking water chronic bioassays in which male and
2 female Eppley Swiss Webster mice (25/sex/dose group; 8 weeks old at study initiation) were
3 exposed 6 days per week until natural death to various concentrations of malonaldehyde and
4 methanol. The stated purpose of the study was to determine the carcinogen! city of
5 malonaldehyde, a product of oxidative lipid deterioration in rancid beef and other food products
6 in advanced stages of degradation. However, due to its instability, malonaldehyde was obtained
7 from the more stable malonaldehyde bis(dimethyacetal), which was hydrolyzed to
8 malonaldehyde and methanol in dilute aqueous solutions in the presence of a strong mineral acid.
9 In the drinking water portion of this study, mice were exposed to 3 different concentrations of
10 the malonaldehyde/methanol solution and three different control solutions of methanol alone,
1 1 0.222%, 0.444% and 0.889% methanol in drinking water (222, 444 and 889 ppm, assuming a
12 density of 1 g/ml), corresponding to the stoichiometric amount of methanol liberated by
13 hydrolysis of the acetal in the three test solutions. The methanol was described as Mallinckrodt
14 analytical grade. No unexposed control groups were included in these studies. However, the
15 author provided pathology data from historical records of untreated Swiss mice of the Eppley
16 colony used in two separate chronic studies, one involving 100 untreated males and 100
17 untreated females (Toth et al., 1977) and the other involving 100 untreated females
18 histopathological analyzed by Apaja (Apaja, 1980).
19 Mice in the Apaja (1980) study were housed five/plastic cage and fed Wayne Lab-Blox
20 pelleted diet. Water was available ad libitum throughout life. Liquid consumption per animal
21 was measured 3 times/week. The methanol dose in the dermal study (females only) was 21.3 mg
22 (532 mg/kg-day using an average weight of 0.04 kg as approximated from Figure 4 of the study),
23 three times/week. The methanol doses in the drinking water study were reported as 22.6, 40.8
4- 1 8 DRAFT-DO NOT CITE OR QUOTE
-------
1 and 84.5 mg/day (560, 1000 and 2100 mg/kg-day using an average weight of 0.04 kg as
2 approximated from Figures 14-16 of the study) for females, and 24.6, 43.5 and 82.7 mg/day
3 (550, 970, and 1800 mg/kg-day using an average weight of 0.045 kg as approximated from
4 Figures 14-16 of the study) for males, 6 days/week. The animals were checked daily and body
5 weights were monitored weekly. The in-life portion of the experiment ended at 120 weeks with
6 the death of the last animal. Like the Soffritti et al. (2002a) study, test animals were sacrificed
7 and necropsied when moribund.21
8 The authors reported that survival of the methanol exposed females of the drinking water
9 study was lower than untreated historical controls (p < 0.05), but no significant differences in
10 survival was noted for males. An increase in liver parenchymal cell necrosis was reported in the
11 male and female high-dose groups, with the incidence in females (8%) being significant
12 (p < 0.01) relative to untreated historical controls. Incidence of acute pancreatitis was higher in
13 high-dose males (p <0.001), but did not appear to be dose-related in females, increasing at the
14 mid- (p <0.0001) and low-doses (p <0.01) when compared to historical controls but not
15 appearing at all in the high-dose females. Significant increases relative to untreated historical
16 controls were noted in amyloidosis of the spleen, nephropathy and pneumonia, but the increases
17 did not appear to be dose related.
18 The author reported incidences of malignant lymphoma in females of 15%, 16%, 36%,
19 and 40% for 532 mg/kg-day (dermal), 560, 1,000, and 2,100 mg/kg-day (drinking water),
20 respectively. Males from the drinking water study had incidences of malignant lymphoma of 4,
21 24, and 16% for 550, 970, and 1,800 mg/kg-day. The lymphomas were classified according to
22 Rappaport's classification (Rappaport, 1966), but location of the lymphoma (organ system) was
23 not reported. The distributions of lymphomas according to subclasses reported by the author are
24 shown in Table 4-3 for historical untreated and methanol exposed mice in the drinking water
25 studies. The author indicates that the incidences in both males and females were "within the
26 normal range of occurrence of malignant lymphomas in Eppley Swiss mice," but provides no
27 references or supporting data for this statement and reports elsewhere that the response in high-
28 dose females and mid-dose males were significantly different from unexposed mice from
29 "historical data of untreated controls (Table 9)" of Toth et al. (1977) (p < 0.05). Though not
30 statistically significant (Fishers exact p = 0.06), the malignant lymphoma response in the mid-
31 dose females was also more than double that of untreated controls from another study (18/100)
32 for which the histopathology was also performed by Apaja (Apaja, 1980).
21
The following tisues were fixed in 10% formalin (pH 7.5), embedded in paraffin, sectioned, stained routinely with
hematoxylineosin (special stains used as needed) and histologically evaluated: skin, lungs, liver spleen, pancreas,
kidneys, adrenal glands, esophagus, stomach, small and large intestines, rectum, urinary bladder, uterus and ovaries
or testes, prostate glands and tumors or other gross pathological lesions.
4-19 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-3. Incidence of malignant lymphoma responses in Swiss mice exposed to
methanol in drinking water for life
Malignant Lymphoma
Lymphocytic, well diff.
Lymphocytic moderately diff.
Lymphocytic, poorly diff.
Mixed cell type
Histocytic type
Unclassified
Total
Dose (mg/kg-day)
Males (%)
Oa
n=100
8
550
n=25
4
4
970
n=25
4
4
4
4
8
24C
1800
n=24
8.3
4.2
4.2
17
Females (%)
Oa
n=100
20
Ob
n=100
8
3
7
18
560
n=25
4
4
12
16
1000
n=25
4
4
4
4
12
8
36d
2100
n=25
12
4
4
8
8
4
40C
aToth et al. (1977); Hinderer et al. (1979)
cp < 0.05 as reported by author compared with Toth et al. (1977); Female response is also significant (p < 0.05;
Fishers exact test) versus untreated controls from Hinderer et al. (1979) and combined controls from both studies.
p = 0.06 by Fishers exact test versus untreated controls from Hinderer et al. (1979) and combined controls from
both studies.
4.2.2. Inhalation Studies
4.2.2.1. Acute Toxicity
1 Lewis (1992) reported a 4-hour median lethal concentration (LCso) for methanol in rats of
2 64,000 ppm (83,867 mg/m3).
3 Japan's NEDO sponsored a series of toxicological tests on monkeys (M. fascicularis),
4 rats, and mice, using inhalation exposure.22 A short-term exposure study evaluated monkeys (sex
5 unspecified) exposed to 3,000 ppm (3,931 mg/m3), 21 hours/day for 20 days (1 animal),
6 5,000 ppm (6,552 mg/m3) for 5 days (1 animal), 5,000 ppm (6,552 mg/m3) for 14 days (2
7 animals), and 7,000 and 10,000 ppm (9,173 and 13,104 mg/m3, respectively) for up to 6 days (1
8 animal at each exposure level) (NEDO, 1987, unpublished report). Most of the experimental
9 findings were discussed descriptively in the report, without specifying the extent of change for
10 any of the effects in comparison to seven concurrent controls. However, the available data
11 indicate that clinical signs of toxicity were apparent in animals exposed to 5,000 ppm (all
12 exposure durations) or higher concentrations of methanol. These included reduced movement,
13 crouching, weak knees, involuntary movements of hands, dyspnea, and vomiting. In the
22 In their bioassays, NEDO (1987) used inbred rats of the F344 or Sprague-Dawley strain, inbred mice of the
B6C3F1 strain and wild-caught M. fascicularis monkeys imported from Indonesia. The possibility of disease among
wild-caught animals is a concern, but NEDO (1987) state that the monkeys were initially quarantined for 9 weeks
and measures were taken throughout the studies against the transmission of pathogens for infectious diseases. The
authors indicated that "no infectious disease was observed in monkeys" and that "subjects were healthy throughout
the experiment."
4-20 DRAFT-DO NOT CITE OR QUOTE
-------
1 discussion section of the summary report, the authors stated that there was a sharp increase in the
2 blood levels of methanol and formic acid in monkey exposed to >3,000 ppm (3,931 mg/m3)
3 methanol. They reported that methanol and formic acid concentrations in the blood of monkeys
4 exposed to 3,000 ppm or less were 80 mg/L and 30 mg/L, respectively.23 In contrast, monkeys
5 exposed to 5,000 ppm or higher concentrations of methanol had blood methanol and formic acid
6 concentrations of 5,250 mg/L and 1,210 mg/L, respectively. Monkeys exposed to 7,000 ppm and
7 10,000 ppm became critically ill and had to be sacrificed prematurely. Food intake was said to
8 be little affected at 3,000 ppm, but those exposed to 5,000 ppm or more showed a marked
9 reduction. Clinically, the monkeys exposed to 5,000 ppm or more exhibited reduced movement,
10 weak knees, and involuntary movement of upper extremities, eventually losing consciousness
11 and dying.
12 There were no significant changes in growth, with the exception of animals exposed to
13 the highest concentration, where body weight was reduced by 13%. There were few compound-
14 related changes in hematological or clinical chemistry effects, although animals exposed to 7,000
15 and 10,000 ppm showed an increase in white blood cells. A marked change in blood pH values
16 at the 7,000 ppm and 10,000 ppm levels (values not reported) was attributed to acidosis due to
17 accumulation of formic acid. A range of histopathologic changes to the CNS was apparently
18 related to treatment. Severity of the effects was increased with exposure concentration. Lesions
19 included characteristic degeneration of the bilateral putamen, caudate nucleus, and claustrum,
20 with associated edema in the cerebral white matter. Necrosis of the basal ganglia was noted
21 following exposure to 5,000 ppm for 5 days (1 animal) and 14 days(l animal). Exposure to
22 3,000 ppm was considered to be close to the threshold for these necrotic effects, as the monkeys
23 exposed at this level experienced little more than minimal fibrosis of responsive stellate cells of
24 the thalamus, hypothalamus and basal ganglion. The authors reported that no clinical or
25 histopathological effects of the visual system were apparent, but that exposure to 3,000 ppm
26 (3,931 mg/m3) or more caused dose-dependent fatty degeneration of the liver, and exposure to
27 5,000 ppm (6,552 mg/m3) or more caused vacuolar degeneration of the kidneys, centered on the
28 proximal uniferous tubules.
4.2.2.2. Subchronic Toxicity
29 A number of experimental studies have examined the effects of subchronic exposure to
30 methanol via inhalation. For example, Sayers et al. (1944) employed a protocol in which 2 male
31 dogs were repeatedly exposed (8 times daily for 3 minutes/exposure) to 10,000 ppm
32 (13,104 mg/m3) methanol for 100 days. One of the dogs was observed for a further 5 days
33 before sacrifice; the other dog was observed for 41 days postexposure. There were no clinical
23 Note that Burbacher et al. (2004b, 1999a) measured blood levels of methanol and formic acid in control monkeys
of 2.4 mg/L and 8.7 mg/L, respectively (see Table 3-3).
4-21 DRAFT-DO NOT CITE OR QUOTE
-------
1 signs of toxicity, and both gained weight during the study period. Blood samples were drawn on
2 a regular basis to monitor hematological parameters, but few if any compound-related changes
3 were observed. Ophthalmoscopic examination showed no incipient anomalies at any point
4 during the study period. Median blood concentrations of methanol were 65 mg/L (range 0-280
5 mg/L) for one dog, and 140 mg/L (70-320 mg/L) for the other.
6 White et al. (1983) exposed 4 male Sprague-Dawley rats/group, 6 hours/day, 5 days/week
7 to 0, 200, 2,000, or 10,000 ppm (0, 262, 2,621, and 13,104 mg/m3) methanol for periods of 1, 2,
8 4, and 6 weeks. Additional groups of 6-week-exposure animals were granted a 6-week
9 postexposure recovery period prior to sacrifice. The lungs were excised intact and lavaged
10 6 times with known volumes of physiological saline. The lavage supernatant was then assayed
11 for lactate dehydrogenase (LDH) and 7V-acetyl-/?-Z)-glucosamidase (/?-NAG) activities. Other
12 parameters monitored in relation to methanol exposure included absolute and relative lung
13 weights, lung DNA content, protein, acid RNase and acid protease, pulmonary surfactant,
14 number of free cells in lavage/unit lung weight, surface protein, LDH, and /?-NAG. As discussed
15 by the authors, none of the monitored parameters showed significant changes in response to
16 methanol exposure.
17 Andrews et al. (1987) carried out a study of methanol inhalation in 5 Sprague-Dawley
18 rats/sex/group and 3 M. fascicularis monkeys/sex/group, 6 hours/day, 5 days/week, to 0, 500,
19 2,000, or, 5,000 ppm (0, 660, 2,620, and 6,552 mg/m3) methanol for 4 weeks. Clinical signs
20 were monitored twice daily, and all animals were given a physical examination once a week.
21 Body weights were monitored weekly, and animals received an ophthalmoscopic examination
22 before the start of the experiment and at term. Animals were sacrificed at term by
23 exsanguination following i.v. barbiturate administration. A gross necropsy was performed,
24 weights of the major organs were recorded, and tissues and organs taken for histopathologic
25 examination. As described by the authors, all animals survived to term with no clinical signs of
26 toxicity among the monkeys and only a few signs of irritation to the eyes and nose among the
27 rats. In the latter case, instances of mucoid nasal discharges appeared to be dose related. There
28 were no differences in body weight gain among the groups of either rats or monkeys, and overall,
29 absolute and relative organ weights were similar to controls. The only exception to this was a
30 decrease in the absolute adrenal weight of female high-concentration monkeys and an increase in
31 the relative spleen weight of mid-concentration female rats. These changes were not considered
32 by the authors to have biological significance. For both rats and monkeys, there were no
33 compound-related changes in gross pathology, histopathology, or ophthalmoscopy. These data
34 suggest a NOAEL of 5,000 ppm (6,600 mg/m3) for Sprague-Dawley rats and monkeys under the
35 conditions of the experiment.
4-22 DRAFT-DO NOT CITE OR QUOTE
-------
1 Two studies by Poon et al. (1995, 1994) examined the effects of methanol on Sprague-
2 Dawley rats when inhaled for 4 weeks. The effects of methanol were evaluated in comparison to
3 those of toluene and toluene/methanol mixtures (Poon et al., 1994), and to gasoline and
4 gasoline/methanol mixtures (Poon et al., 1995). In the first case (Poon et al., 1994), 10 Sprague-
5 Dawley rats/sex/group were exposed via inhalation, 6 hours/day, 5 days/week to 0, 300, or
6 3,000 ppm (0, 393, 3,930 mg/m3) methanol for 4 weeks. Clinical signs were monitored daily,
7 and food consumption and body weight gain were monitored weekly. Blood was taken at term
8 for hematological and clinical chemistry determinations. Weights of the major organs were
9 recorded at necropsy, and histopathologic examinations were carried out. A 10,000 x g Hver
10 supernatant was prepared from each animal to measure aniline hydroxylase, aminoantipyrine N-
11 demethylase, and ethoxyresorufin-O-deethylase activities. For the most part, the responses to
12 methanol alone in this experiment were unremarkable. All animals survived to term, and there
13 were no clinical signs of toxicity among the groups. Body weight gain and food consumption
14 did not differ from controls, and there were no compound-related effects in hematological or
15 clinical chemistry parameters or in hepatic mixed function oxidase activities. However, the
16 authors described a reduction in the size of thyroid follicles that was more obvious in female than
17 male rats. The authors considered this effect to possibly have been compound related, although
18 the incidence of this feature for the 0, 300, and 3,000 ppm-receiving females was 0/6, 2/6, and
19 2/6, respectively.
20 The second experimental report by Poon et al. (1995) involved the exposure of
21 15 Sprague-Dawley rats/sex/group, 6 hours/day, 5 days/week for 4 weeks to 0 or 2,500 ppm (0
22 and 3,276 mg/m3) to methanol as part of a study on the toxicological interactions of methanol
23 and gasoline. Many of the toxicological parameters examined were the same as those described
24 in Poon et al. (1994) study. However, in this study urinalysis featured the determination of
25 ascorbic and hippuric acids. Additionally, at term, the lungs and tracheae were excised and
26 aspirated with buffer to yield bronchoalveolar lavage fluid that was analyzed for ascorbic acid,
27 protein, and the activities of gamma-glutamyl transferase (y-GT), AP and LDH. Few if any of
28 the monitored parameters showed any differences between controls and those animals exposed to
29 methanol alone. However, two male rats had collapsed right eyes, and there was a reduction in
30 relative spleen weight in females exposed to methanol. Histopathologic changes in methanol-
31 receiving animals included mild panlobular vacuolation of the liver in females and some mild
32 changes to the upper respiratory tract, including mucous cell metaplasia. The incidence of the
33 latter effect, though higher, was not significantly different than controls in rats exposed to
34 2,500 ppm (3,267 mg/m3) methanol. However, there were also signs of an increased severity of
35 the effect in the presence of the solvent. No histopathologic changes were seen in the lungs or
36 lower respiratory tract of rats exposed to methanol alone.
4-23 DRAFT-DO NOT CITE OR QUOTE
-------
4.2.2.3. Chronic Toxicity
1 Information on the chronic toxicity of methanol has come from NEDO (1987,
2 unpublished report) which includes the results of experiments on 1) monkeys exposed for up to 3
3 years, 2) rats and mice exposed for 12 months, 3) mice exposed for 18 months, and 4) rats
4 exposed for 2 years.
5 In the monkeys, 8 animals (sex unspecified) were exposed to 10, 100, or 1,000 ppm (13,
6 131, and 1,310 mg/m3) methanol, 21 hours/day, for 7 months (2 animals), 19 months,
7 (3 animals), or 29 months (3 animals). There was no indication in the NEDO (1987) report that
8 this study employed a concurrent control group. One of the 3 animals receiving 100 ppm
9 methanol and scheduled for sacrifice at 29 months was terminated at 26 months. Clinical signs
10 were monitored twice daily, body weight changes and food consumption were monitored weekly,
11 and all animals were given a general examination under anesthetic once a month. Blood was
12 collected for hematological and clinical chemistry tests at term, and all animals were subject to a
13 histopathologic examination of the major organs and tissues.
14 While there were no clinical signs of toxicity in the low-concentration animals, there was
15 some evidence of nasal exudate in monkeys in the mid-concentration group. High-concentration
16 (1,000 ppm) animals also displayed this response and were observed to scratch themselves over
17 their whole body and crouch for long periods. Food and water intake, body temperature, and
18 body weight changes were the same among the groups. NEDO (1987) reported that there was no
19 abnormality in the retina of any monkey. When animals were examined with an
20 electrocardiogram, there were no abnormalities in the control or 10 ppm groups. However, in the
21 100 ppm group, one monkey showed a negative change in the T wave. All 3 monkeys exposed
22 to 1,000 ppm (1,310 mg/m3) displayed this feature, as well as a positive change in the Q wave.
23 This effect was described as a slight myocardial disorder and suggests that 10 ppm (13.1 mg/m3)
24 is a NOAEL for chronic myocardial effects of methanol and mild respiratory irritation. There
25 were no compound-related effects on hematological parameters. However, 1 monkey in the
26 100 ppm (131 mg/m3) group had greater than normal amounts of total protein, neutral lipids,
27 total and free cholesterol, and glucose, and displayed greater activities of ALT and aspartate
28 transaminase (AST). The authors expressed doubts that these effects were related to methanol
29 exposure and speculated that the animal suffered from liver disease.24
30 Histopathologically, no degeneration of the optical nerve, cerebral cortex, muscles, lungs,
31 trachea, tongue, alimentary canal, stomach, small intestine, large intestine, thyroid gland,
32 pancreas, spleen, heart, aorta, urinary bladder, ovary or uterus were reported (neuropathological
33 findings are discussed below Section 4.4.2. Most of the internal organs showed no compound-
24 Ordinarily, the potential for liver disease in test animals would be remote, but may be a possibility in this case
given that these monkeys were captured in the wild.
4-24 DRAFT-DO NOT CITE OR QUOTE
-------
1 related histopathologic lesions. However, there were signs of incipient fibrosis and round cell
2 infiltration of the liver in monkeys exposed to 1,000 ppm (1,310 mg/m3) for 29 months. NEDO
3 (1987) indicated that this fibrosis occurred in 2/3 monkeys of the 1,000 ppm group to a "strictly
4 limited extent." They also qualitatively reported a dose-dependent increase in "fat granules" in
5 liver cells "centered mainly around the central veins" at all doses, but did not provide any
6 response data. The authors state that 1,000 ppm (1,310 mg/m3) represents a chronic lowest-
7 observed-adverse-effect level (LOAEL) for hepatic effects of inhaled methanol, suggesting that
8 the no effect level would be 100 ppm (131 mg/m3). However, this is a tenuous determination
9 given the lack of information on the pathological progression and significance of the appearance
10 of liver cell fat granules at exposures below 1,000 ppm and the lack detail (e.g., time of sacrifice)
11 for the control group.
12 Dose-dependent changes were observed in the kidney; NEDO (1987) described the
13 appearance of Sudan-positive granules in the renal tubular epithelium at 100 ppm (131 mg/m3)
14 and 1,000 (1,310 mg/m3) and hyalinization of the glomerulus and penetration of round cells into
15 the renal tubule stroma of monkeys exposed to methanol at 1,000 (1,310 mg/m3). The former
16 effect was more marked at the higher concentration and was thought by the authors to be
17 compound-related. This would indicate a no effect level at 10 ppm (13.1 mg/m3) for the chronic
18 renal effects of methanol. The authors observed atrophy of the tracheal epithelium in four
19 monkeys. However, the incidence of these effects was unrelated to dose and therefore, could not
20 be unequivocally ascribed to an effect of the solvent. No other histopathologic abnormalities
21 were related to the effects of methanol. Confidence in these determinations is considerably
22 weakened by concern over whether a concurrent control group was used in the chronic study.25
23 NEDO (1987) describes a 12-month inhalation study in which 20 F344 rats/sex/group
24 were exposed to 0, 10, 100, or 1,000 ppm (0, 13.1, 131, and 1,310 mg/m3) methanol,
25 approximately 20 hours/day, for a year. Clinical signs of toxicity were monitored daily; body
26 weights and food consumption were recorded weekly for the first 13 weeks, then monthly.
27 Blood samples were drawn at term to measure hematological and clinical chemistry parameters.
28 Weights of the major organs were monitored at term, and a histopathologic examination was
29 carried out on all major organs and tissues. Survival was high among the groups; one high-
30 concentration female died on day 337 and one low-concentration male died on day 340. As
31 described by the authors, a number of procedural anomalies arose during this study. For
32 example, male controls in two cages lost weight because of an interruption to the water supply.
33 Another problem was that the brand of feed was changed during the study. Fluctuations in some
34 clinical chemistry and hematological parameters were recorded. The authors considered the
35 fluctuations to be minor and within the normal range. Likewise, a number of histopathologic
25 All control group responses were reported in a single table in the section of the NEDO (1987) report that describes
the acute monkey study, with no indication as to when the control group was sacrificed.
4-25 DRAFT-DO NOT CITE OR QUOTE
-------
1 changes were observed, which, in every case, were considered to be unrelated to exposure level
2 or due to aging.
3 A companion experiment featured the exposure of 30 B6C3F1 mice/sex/group for 1 year
4 to the same concentrations as the F344 rats (NEDO, 1987). Broadly speaking, the same suite of
5 toxicological parameters was monitored as described above, with the addition of urinalysis.
6 10 mice/sex/group were sacrificed at 6 months to provide interim data on the parameters under
7 investigation. A slight atrophy in the external lacrimal gland was observed in both sexes and was
8 significant in the 1,000 ppm male group compared with controls. An apparently dose-related
9 increase in moderate fatty degeneration of hepatocytes was observed in males (1/20, 4/20, 6/20
10 and 8/20 in the 0, 10, 100, and 1,000 ppm dose groups, respectively) which was significantly
11 increased over controls at the 1,000 ppm dose. However the incidence of moderate to severe
12 fatty degeneration was observed in untreated animals maintained outside of the chamber. In
13 addition, there was a clear correlation between fatty degeneration and body weight (a change
14 which was not associated with treatment at 12 months); heavier animals tended to have more
15 severe cases of fatty degeneration. The possibility of renal deficits due to methanol exposure
16 was suggested by the appearance of protein in the urine. However, this effect was also seen in
17 controls and did not display a dose-response effect. Therefore, it is unlikely to be a consequence
18 of exposure to methanol. NEDO (1987) reported other histopathologic and biochemical (e.g.,
19 urinalysis and hematology) findings that do not appear to be related to treatment, including a
20 number of what were considered to be spontaneous tumors in both control and exposure groups.
21 NEDO (1987, 1985/2008a, unpublished reports)26 exposed 52 male and 53 female
22 B6C3F1 mice/group for 18 months at the same concentrations of methanol (0, 10, 100 and
23 1,000 ppm) and with a similar experimental protocol to that described in the 12-month studies.27
24 The fact that the duration of this study was only 18 months and not the more typical 2 years
25 limits its ability to detect carcinogenic responses with relatively long latency periods. Animals
26 were sacrificed at the end of the 18-month exposure period. NEDO (1985/2008a) reported that
27 "there was no microbiological contamination that may have influenced the result of the study"
28 and that the study included an assessment of general conditions, body weight change, food
29 consumption rate, laboratory tests (urinalysis, hematological, and plasma biochemistry) and
30 pathological tests (pathological autopsy,28 organ weight check and histopathology29). As stated
26 This study is described in a summary report (NEDO, 1987) and a more detailed, eight volume translation of the
original chronic mouse study report (NEDO, 1985/2008a).
27 The authors reported that "[t]he levels of methanol turned out to be ~4 ppm in low level exposure group (10 ppm)
for ~11 weeks from week 43 of exposure due to the analyzer malfunction" and that "the average duration of
methanol exposure was 19.1 hours/day for both male and female mice."
28 Autopsy was performed on all cases to look for gross lesions in each organ.
4-26 DRAFT-DO NOT CITE OR QUOTE
-------
1 in the summary report (NEDO, 1987), a few animals showed clinical signs of toxicity, but the
2 incidence of these responses was not related to dose. Likewise, there were no compound-related
3 changes in body weight increase, food consumption,30 urinalysis, hematology, or clinical
4 chemistry parameters. High-concentration males had lower testis weights compared to control
5 males. Significant differences were detected for both absolute and relative testis weights. One
6 animal in the high-dose group had severely atrophied testis weights, approximately 25% of that
7 of the others in the dose group. Exclusion of this animal in the analysis still resulted in a
8 significant difference in absolute testis weight compared to controls but resulted in no difference
9 in relative testis weight. High-concentration females had higher absolute kidney and spleen
10 weights compared to controls, but there was no significant difference in these organ weights
11 relative to body weight. At necropsy, there were signs of swelling in spleen, preputial glands,
12 and uterus in some animals. Some animals developed nodes in the liver and lung although,
13 according to the authors, none of these changes were treatment-related. NEDO (1985/2008a)
14 reported that all nonneoplastic changes were "nonspecific and naturally occurring changes that
15 are often experienced by 18-month old B6C3F1 mice" and that fatty degeneration of liver that
16 was suspected to occur dose-dependently in the 12-month NEDO (1987) study was not observed
17 in this study. Similarly, though the study found various neoplastic changes across dose groups,
18 there was no compound-related formation of tumors in any organ or tissue.
19 EPA reviewed the cancer findings documented in a recent translation of the original
20 NEDO report on this chronic mouse study (NEDO, 1985/2008a) to identify possible compound-
21 related effects. Hyperplastic and neoplastic histopathological findings have been tabulated and
22 are as shown in Table 4-4.
29 Complete histopathological examinations were performed for the control group and high-dose (1,000 ppm)
groups. Only histopathological examinations of the liver were performed on the low- and medium-level exposure
groups because no chemical-related changes were found in the high-level exposure group and because liver changes
were noted in the 12-month mouse study (NEDO, 1987).
30 NEDO (1985/2008a) reports sporadic reductions in food consumption of the 1,000 ppm group, but no associated
weight loss or abnormal test results.
4-27 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-4. Histopathological changes in tissues of B6C3F1 mice exposed to methanol
via inhalation for 18 months
Tissues/tumor type
Exposure concentration (ppm)
0
10
100
1,000
0
10
100
1,000
Number of animals affected/number examined
Males
Females
Lung
Adenomatosis
Pulmonary adenoma
0/52
4/52
0/3
0/3
0/3
0/3
0/52
7/52
0/53
3/53
0/0
0/0
0/5
0/0
1/53
2/53
Liver
Hepatocellular adenoma
Hepatocellular carcinoma
Neoplastic nodule
3/52
2/52
16/52
2/52
4/52
13/52
2/52
0/52
16/52
4/52
1/52
20/52
1/53
3/53
1/53
1/52
0/52
0/52
1/53
3/53
0/53
4/53
2/53
1/53
Source: NEDO (1985/2008a).
1 There is no clear evidence for treatment-related carcinogenic effects in the mice in this
2 study. However, the fact that the study duration was limited to 18 months rather than the
3 traditional 2-year bioassay makes it difficult to draw a definitive conclusion, particularly
4 regarding pulmonary adenomas, which were marginally increased in high-dose male mice of this
5 study and were also increased in male rats of the NEDO chronic rat study (NEDO, 1987,
6 1985/2008b, unpublished reports). In this study, the lack of adenomatosis in control or treated
7 male mice supports the conclusion of the authors that the observed tumors were probably
8 unrelated to methanol exposure. There was no apparent relationship to treatment in any
9 neoplastic findings in the liver. Of relevance to the findings of treatment-related lymphomas and
10 leukemias in Sprague-Dawley rats receiving methanol in drinking water in the Soffritti et al.
11 (2002a) study, few lymphomas and leukemias were identified in the NEDO (1987, 1985/2008b)
12 study reports, with no sign of a dose-related trend.
13 Another study reported in (NEDO (1987, 1985/2008b)31 was a 24-month carcinogenicity
14 bioassay in which 52 F344 rats/sex/group were kept in whole body inhalation chambers
15 containing 0, 10, 100, or 1,000 ppm (0, 13.1, 131, and 1,310 mg/m3) methanol vapor. Animals
16 were maintained in the exposure chambers for approximately 19.5 hours/day for a total of 733-
17 736 days (males) and 740-743 days (females). Animals were monitored once a day for clinical
18 signs of toxicity, body weights were recorded once a week, and food consumption was measured
19 weekly in a 24-animal subset from each group. Urinalysis was carried out on the day prior to
20 sacrifice for each animal, the samples being monitored for pH, protein, glucose, ketones,
21 bilirubin, occult blood, and urobilinogen. Routine clinical chemistry and hematological
31 This study is described in a summary report (NEDO, 1987) and a more detailed, eight-volume translation of the
original chronic rat study report (NEDO, 1985/2008b).
4-28 DRAFT-DO NOT CITE OR QUOTE
-------
1 measurements were carried out and all animals were subject to necropsy at term, with a
2 comprehensive histopathological examination of tissues and organs.32
3 There was some fluctuation in survival rates among the groups in the rat study, though
4 apparently unrelated to exposure concentration.33 In all groups, at least 60% of the animals
5 survived to term. A number of toxicological responses were described by the authors, including
6 atrophy of the testis, cataract formation, exophthalmia, small eye ball, alopecia, and paralysis of
7 the hind leg. However, according to the authors, the incidence of these effects were unrelated to
8 dose and more likely represented effects of aging. NEDO (1985/2000b) reported a mild,
9 nonsignificant (4%) body weight suppression among 1,000 ppm females between 51 and
10 72 weeks, but that body weight gain was largely similar among the groups for the duration of the
11 experiment. Food consumption was significantly lower than controls in high-concentration male
12 rats during the day 210-365 time interval, but no corresponding weight loss was observed.
13 Among hematological parameters, mid- and high-concentration females had a significantly
14 (p> 0.05) higher differential leukocyte count than controls, but dose dependency was not
15 observed. Serum total cholesterol, triglyceride, free fatty acid, and phospholipid concentrations
16 were significantly (p > 0.05) lower in high-concentration females compared to controls.
17 Likewise, serum sodium concentrations were significantly (p > 0.05) lower in mid- and high-
18 concentration males compared to controls. High-concentration females had significantly lower
19 (p > 0.05) serum concentrations of inorganic phosphorus but significantly (p > 0.05) higher
20 concentrations of potassium compared to controls. Glucose levels were elevated in the urine of
21 high-concentration male rats relative to controls, and female rats had lower pH values and higher
22 bilirubin levels in mid- and high-concentration groups relative to controls. In general, NEDO
23 (1987, 1985/2008b) reported that these variations in urinary, hematology, and clinical chemistry
24 parameters were not related to chemical exposure.
25 NEDO (1987) reported that there was little change in absolute or relative weights of the
26 major organs or tissues. When the animals were examined grossly at necropsy, there were some
27 signs of swelling in the pituitary and thyroid, but these effects were judged to be unrelated to
28 treatment. The most predominant effect was the dose-dependent formation of nodes in the lung
29 of males (2/52, 4/52, 5/52, and 10/52 \p < 0.01] for control, low-, mid-, and high-concentration
30 groups, respectively). Histopathologic examination pointed to a possible association of these
31 nodes with the appearance of pulmonary adenoma (1/52, 5/52, 2/52, and 6/52 for control, low-,
32 mid- and high-concentration groups, respectively) and a single pulmonary adenocarcinoma in the
32 Complete histopathological examinations were performed on the cases killed on schedule (week 104) among the
control and high-exposure groups, and the cases that were found dead/ killed in extremis of all the groups. Because
effects were observed in male and female kidneys, male lungs as well as female adrenal glands of the high-level
exposure group, these organs were histopathologically examined in the low- and mid-exposure groups.
33Survival at the time of exposure termination (24 months) was 69%, 65%, 81%, and 65% for males and 60%, 63%,
60% and 67% for females of the control, low-, mid- and high-exposure groups, respectively.
4-29 DRAFT-DO NOT CITE OR QUOTE
-------
1 high-dose group (1/52). Other examples of tumor formation that were increased in high-
2 concentration animals versus controls included an increased incidence of pituitary adenomas in
3 high-concentration males (17/52 compared to 12/52 controls), hyperplastic change in the testis in
4 high-concentration males (10/52 compared to 4/52 controls), and chromaffmoma
5 (pheochromocytomas)34 in the adrenals of high-concentration females (7/52 compared to 2/52
6 controls). Individually, these changes did not achieve statistical significance, and in general, the
7 authors concluded that few if any of the observed changes were effects of methanol.
8 EPA reviewed the cancer findings of this study that are documented in a recent translation
9 of the original NEDO (1985/2008b) report to identify possible compound-related effects. High-
10 dose incidences of pituitary adenomas (17/52; 33%) and hyperplastic change in testes (10/52;
11 19%) mentioned above were within historical incidences for this rat strain.35 However, the
12 observed incidence rate for pulmonary adenoma/adenocarcinoma in high-dose males of 13.5%
13 (7/52) was significantly elevated (Fisher's exact test/? < 0.05) over the concurrent control rate of
14 2% (1/52) and historical control rates of 2.5% ± 2.6% (n = 1054) and 3.84% ± 2.94% (n = 1199)
15 reported by NTP for the pre-1995 control F344 male rats fed NIH-07 diet (NTP, 1999) and post-
16 1994 control F344 male rats fed NTP-2000 diet (NTP, 2007), respectively. Also, the incidence of
17 pulmonary adenoma/adenocarcinoma in male rats exhibited a dose-response trend (Cochrane-
18 Armitage/? < 0.05). While the observed incidence rate for pheochromocytomas in high-dose
19 females of 13.7% (7/51) was not significantly elevated over the concurrent control rate of 4%
20 (2/50), it was significantly elevated (Fisher's exact test/? < 0.05) over NTP historical control
21 rates for total (benign, complex and malignant) pheochromocytomas of 2.5% ± 2.6% (n = 1054)
22 and 3.84% ± 2.94% (n = 1199) reported by NTP for pre-1995 control F344 female rats fed NIH-
23 07 diet (NTP, 1999) and post-1994 control F344 female rats fed NTP-2000 diet (NTP, 2007),
24 respectively.36 Also, the incidence of pheochromocytomas in female rats exhibited a dose-
25 response trend (Cochrane-Armitage/? < 0.05). The histopathological incidences for pulmonary
26 and adrenal effects reported by NEDO (1987, 1985/2008b) are shown in Table 4-5.
34 There were some differences in nomenclature used in the NEDO (1985/2008b) report translation versus those
used in the older summary report (NEDO, 1987). For example, it is probable that the adrenal chromaffinoma
referred to in NEDO (1987) are the same lesions as the pheochromocytoma referred to in NEDO (1985/2008b).
35 NTP reports high incidences in historical control male F344 rats of pituitary gland adenomas, ranging from 45.4%
± 20.19% (NTP, 2007) to 63.4% ± 18.3% (NTP, 1999). While control incidences for testicular hyperplasia are not
reported, historical incidences of testicular ademoma ranged from 70.1% ±11.2% (NTP, 1999) to 86.32% ±9.34%
(NTP, 2007) in this rat strain.
36 NEDO (1987, 1985/2008b) does not categorize reported chromoffinoma (pheochromocytomas) as benign,
complex or malignant. The historical rates for complex and malignant tumors are much lower, ranging from 0.1%
to 0.7 % for female F344 rats (NTP, 2007; NTP, 1999; Haseman et al., 1998).
4-30 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-5. Histopathological changes in lung and adrenal tissues of F344 rats exposed
to methanol via inhalation for 24 months
Tissues/
tumor type
Exposure concentration (ppm)
0
10
100
1000
0
10
100
1000
Number of animals affected/number examined
Males
Females
Lung
Pulmonary adenoma
Pulmonary adenocarcinoma
Combined pulmonary
adenoma/adenocarcinoma
Adenomatosis
Epithelial swelling
1/52
0/52
1/52
4/52
3/52
5/50
0/50
5/50
1/50
2/50
2/52
0/52
2/52
5/52
1/52
6/52
1/52
7/52 a'b
4/52
1/52
2/52
0/52
2/52
3/52
0/52
0/19
0/19
0/19
2/19
0/19
0/20
0/20
0/20
1/20
0/20
0/52
0/52
0/52
1/52
0/52
Adrenal glands
Pheochromocytoma
Medullary hyperplasia
7/52
0/52
2/16
0/16
2/10
0/10
4/51
2/51
2/50
2/50
3/51
3/51
2/49
7/49
7/51b'c
2/51
ap < 0.05 over concurrent controls using the Fisher's Exact test.
bp < 0.05 for Cochrane-Armitage test of overall dose-response trend.
°p < 0.05 over NTP historical controls for total (benign, complex and malignant) pheochromocytomas using the
Fisher's Exact test
Source: NEDO (1987, 1985/2008b).
1 In contrast to the conclusions of the NEDO (1987) summary report that there were no
2 compound-related changes in F344 rats exposed to methanol via inhalation, EPA identifies
3 potential treatment-related changes in the lungs of male rats and the adrenal medulla of female
4 rats in the more detailed translation of the original report (NEDO, 1985/2008b). The NEDO
5 (1987) summary report did not report the statistically significant combined pulmonary adenoma
6 and adenocarcinoma finding in the high-dose group of male rats. Table 6 (page 146) of the
7 NEDO (1987) summary reports only "Tumural changes occurring at a rate of over 5%." The
8 lung response of the male rats as shown in Table 4-5 suggests a proliferative change in cells of
9 the alveolar epithelium involving a progression towards adenoma and adenocarcinoma that
10 appears to be more pronounced with increasing methanol exposure and considerably elevated
11 over historical controls. Similarly, for female rats, the observed increase in medullary
12 hyperplasia in the 100 ppm dose group, in conjunction with a higher incidence of
13 pheochromocytoma in the adrenal gland is suggestive of a methanol-induced progressive change
14 leading to a carcinogenic response.
4-31
DRAFT-DO NOT CITE OR QUOTE
-------
4.3. REPRODUCTIVE AND DEVELOPMENTAL STUDIES-ORAL AND INHALATION
1 Many studies have been conducted to investigate the reproductive and developmental
2 toxicity of methanol. The purpose of these studies was principally to determine if methanol has
3 a similar toxicology profile to another widely studied teratogen, ethanol.
4.3.1. Oral Studies
4 Three studies were identified that investigated the reproductive and developmental effects
5 of methanol in rodents via the oral route (Fu et al., 1996; Sakanashi et al., 1996; Rogers et al.,
6 1993a). Two of these studies also investigated the influence of folic acid-deficient (FAD) diets on
7 the effects of methanol exposures (Fu et al., 1996; Sakanashi et al., 1996).
8 Rogers et al. (1993a) conducted a developmental toxicity study in which methanol in
9 water was administered to pregnant female CD-I mice via gavage on GD6-GD15. Eight test
10 animals received 4 g/kg-day methanol given in 2 daily doses of 2g/kg; 4 controls received
11 distilled water. By analogy to the protocol of an inhalation study of methanol that was described
12 in the same report, it is assumed that dams were sacrificed on GD17, at which point implantation
13 sites, live and dead fetuses, resorptions/litter, and the incidences of external and skeletal
14 anomalies and malformations were determined. In the brief summary of the findings provided
15 by the authors, it appears that cleft palate (43.5% per litter versus 0% in controls) and
16 exencephaly (29% per litter versus 0% in controls) were the prominent external defects
17 following maternal methanol exposure by gavage. Likewise, an increase in totally resorbed
18 litters and a decrease in the number of live fetuses per litter were evident. However, it is possible
19 that these effects may have been caused or exacerbated by the high bolus dosing regimen
20 employed. It is also possible that effects were not observed due to the limited study size. The
21 small number of animals in the control group relative to the test group limits the power of this
22 study to detect treatment-related responses.
23 Sakanashi et al. (1996) tested the influence of dietary folic acid intake on various
24 reproductive and developmental effects observed in CD-I mice exposed to methanol. Starting
25 5 weeks prior to breeding and continuing for the remainder of the study, female CD-I mice were
26 fed folic acid free diets supplemented with 400 (low), 600 (marginal) or 1,200 (sufficient) nmol
27 folic acid/kg. After 5 weeks on their respective diets, females were bred with CD-I male mice.
28 On GD6-GD15, pregnant mice in each of the diet groups were given twice-daily gavage doses of
29 2.0 or 2.5 g/kg-day methanol (total dosage of 4.0 or 5.0 g/kg-day). On GDIS, mice were
30 weighed and killed, and the liver, kidneys and gravid uteri removed and weighed. Maternal liver
31 and plasma folate levels were measured, and implantation sites, live and dead fetuses, and
32 resorptions were counted. Fetuses were weighed individually and examined for cleft palate and
33 exencephaly. One third of the fetuses in each litter were examined for skeletal morphology.
4-32 DRAFT-DO NOT CITE OR QUOTE
-------
1 They observed an approximate 50% reduction in liver and plasma folate levels in the mice fed
2 low versus sufficient folic acid diets in both the methanol exposed and unexposed groups.
3 Similar to Rogers et al. (1993a), Sakanashi et al. (1996) observed that an oral dose of 4-5 g/kg-
4 day methanol during GD6-GD15 resulted in an increase in cleft palate in mice fed sufficient
5 folic acid diets, as well as an increase in resorptions and a decrease in live fetuses per litter. They
6 did not observe an increase in exencephaly in the FAS group at these doses, and the authors
7 suggest that this may be due to diet and the source of CD-I mice differing between the two
8 studies.
9 In the case of the animals fed the folate deficient diet, there was a 50% reduction in
10 maternal liver folate concentration and a threefold increase in the percentage of litters affected by
11 cleft palate (86.2% versus 34.5% in mice fed sufficient folic acid) and a 10-fold increase in the
12 percentage of litters affected by exencephaly (34.5% versus 3.4% in mice fed sufficient folic
13 acid) at the 5 g/kg methanol dose. Sakanashi et al. (1996) speculate that the increased methanol
14 effect from the FAD diet could have been due to an increase in tissue formate levels (not
15 measured) or to a critical reduction in conceptus folate concentration following the methanol
16 exposure. Plasma and liver folate levels at GDIS within each dietary group were not
17 significantly different between exposed versus unexposed mice. However, these measurements
18 were taken 3 days after methanol exposure. Dorman et al. (1995) observed a transient decrease
19 in maternal red blood cells (RBCs) and conceptus folate levels within 2 hours following
20 inhalation exposure to 15,000 ppm methanol on GD8. Thus, it is possible that short-term
21 reductions in available folate during GD6-GD15 may have affected fetal development.
22 Fu et al. (1996) also tested the influence of dietary folic acid intake on reproductive and
23 developmental effects observed in CD-I mice exposed to methanol. This study was performed
24 by the same laboratory and used a similar study design and dosing regimen as Sakanashi et al.
25 (1996), but exposed the pregnant mice to only the higher 2.5 g/kg-day methanol (total dosage of
26 5.0 g/kg-day) on GD6-GD10. Like Sakanashi et al. (1996), Fu et al. (1996) measured maternal
27 liver and plasma folate levels on GDIS and observed similar, significant reductions in these
28 levels for the FAD versus FAS mice. However, Fu et al. (1996) also measured fetal liver folate
29 levels at GDIS. This measurement does not address the question of whether methanol exposure
30 caused short-term reductions in fetal liver folate because it was taken 8 days after the GD6-
31 GD10 exposure period. However, it did provide evidence regarding the extent to which a
32 maternal FAD diet can impact fetal liver folate levels in this species and strain. Significantly, the
33 maternal FAD diet had a greater impact on fetal liver folate than maternal liver folate levels.
34 Relative to the FAS groups, fetal liver folate levels in the FAD groups were reduced 2.7-fold for
35 mice not exposed to methanol (1.86 ± 0.15 nmol/g in the FAD group versus 5.04 ± 0.22 nmol/g
36 in the FAS group) and 3.5-fold for mice exposed to methanol (1.69 ± 0.12 nmol/g in the FAD
4-33 DRAFT-DO NOT CITE OR QUOTE
-------
1 group versus 5.89 ± 0.39 nmol/g in the FAS group). Maternal folate levels in the FAD groups
2 were only reduced twofold both for mice not exposed (4.65 ± 0.37 versus 9.54 ± 0.50 nmol/g)
3 and exposed (4.55 ± 0.19 versus 9.26 ± 0.42 nmol/g). Another key finding of the Fu et al. (1996)
4 study is that methanol exposure during GD6-GD10 appeared to have similar fetotoxic effects,
5 including cleft palate, exencephaly, resorptions, and decrease in live fetuses, as the same level of
6 methanol exposure administered during GD6-GD15 (Sakanashi et al., 1996; Rogers et al.,
7 1993a). This is consistent with the hypothesis made by Rogers et al. (1993b) that the critical
8 period for methanol-induced cleft palate and exencephaly in CD-I mice is within GD6-GD10.
9 As in the studies of Sakanashi et al. (1996) and Rogers et al. (1993a), Fu et al. (1996) reported a
10 higher incidence of cleft palate than exencephaly.
4.3.2. Inhalation Studies
11 Nelson et al. (1985) exposed 15 pregnant Sprague-Dawley rats/group to 0, 5,000, 10,000,
12 or 20,000 ppm (0, 6,552, 13,104, and 26,209 mg/m3) methanol (99.1% purity) for 7 hours/day.
13 Exposures were conducted on GDI-GDI9 in the two lower concentration groups and GD7-
14 GDIS in the highest concentration group, apparently on separate days. Two groups of 15 control
15 rats were exposed to air only. Day 1 blood methanol levels measured 5 minutes after the
16 termination of exposure in NP rats that had received the same concentrations of methanol as
17 those animals in the main part of the experiment were 1.00 ± 0.21, 2.24 ± 0.20, and 8.65 ± 0.40
18 mg/mL for those exposed to 5,000, 10,000 and 20,000 ppm methanol, respectively. Evidence of
19 maternal toxicity included a slightly unsteady gait in the 20,000 ppm group during the first few
20 days of exposure. Maternal body weight gain and food intake were unaffected by methanol.
21 Dams were sacrificed on GD20, and 13-30 litters/group were evaluated. No effect was observed
22 on the number of corpora lutea or implantations or the percentage of dead or resorbed fetuses.
23 Statistical evaluations included analysis of variance (ANOVA) for body weight effect, Kruskal-
24 Wallis test for endpoints such as litter size and viability and Fisher's exact test for
25 malformations. Fetal body weight was significantly reduced at concentrations of 10,000 and
26 20,000 ppm by 7% and 12-16%, respectively, compared to controls. An increased number of
27 litters with skeletal and visceral malformations were observed at > 10,000 ppm, with statistical
28 significance obtained at 20,000 ppm. Numbers of litters with visceral malformations were 0/15,
29 5/15, and 10/15 and with skeletal malformations were 0/15, 2/15, and 14/15 at 0, 10,000, and
30 20,000 ppm, respectively. Visceral malformations included exencephaly and encephaloceles.
31 The most frequently observed skeletal malformations were rudimentary and extra cervical ribs.
32 The developmental and maternal NOAELs for this study were identified as 5,000 ppm (6,552
33 mg/m3) and 10,000 ppm (13,104 mg/m3), respectively.
4-34 DRAFT-DO NOT CITE OR QUOTE
-------
1 NEDO (1987) sponsored a teratology study in Sprague-Dawley rats that included an
2 evaluation of postnatal effects in addition to standard prenatal endpoints. Thirty-six pregnant
3 females/group were exposed to 0, 200, 1,000, or 5,000 ppm (0, 262, 1,310, and 6,552 mg/m3)
4 methanol vapors (reagent grade) on GD7-GD17 for 22.7 hours/day. Statistical significance of
5 results was evaluated by t-test, Mann-Whitney U test, Fisher's exact test, and/or Armitage's $
6 test.
7 Contrary to the Nelson et al. (1985) report of a 10,000 ppm NOAEL for this rat strain, in
8 the prenatal portion of the NEDO (1987) study, reduced body weight gain and food and water
9 intake during the first 7 days of exposure were reported for dams in the 5,000 ppm group.
10 However, it was not specified if these results were statistically significant. One dam in the
11 5,000 ppm group died on GDI9, and one dam was sacrificed on GDIS in moribund condition.
12 On GD20, 19-24 dams/group were sacrificed to evaluate the incidence of reproductive deficits
13 and such developmental parameters as fetal viability, weight, sex, and the occurrence of
14 malformations. As summarized in Table 4-6, adverse reproductive and fetal effects were limited
15 to the 5,000 ppm group and included an increase in late-term resorptions, decreased live fetuses,
16 reduced fetal weight, and increased frequency of litters with fetal malformations, variations, and
17 delayed ossifications. Malformations or variations included defects in ventricular septum,
18 thymus, vertebrae, and ribs.
19 Postnatal effects of methanol inhalation were evaluated in the remaining 12 dams/group
20 that were permitted to deliver and nurse their litters. Effects were only observed in the
21 5,000 ppm group, and included a 1-day prolongation of the gestation period and reduced post-
22 implantation survival, number of live pups/litter, and survival on PND4 (Table 4-7). When the
23 delay in parturition was considered, methanol treatment had no effect on attainment of
24 developmental milestones such as eyelid opening, auricle development, incisor eruption, testes
25 descent, or vaginal opening. There were no adverse body weight effects in offspring from
26 methanol treated groups. The weights of some organs (brain, thyroid, thymus, and testes) were
27 reduced in 8-week-old offspring exposed to 5000 ppm methanol during prenatal development.
28
29
30
31
4-3 5 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
Table 4-6. Reproductive and developmental toxicity in pregnant Sprague-Dawley rats
exposed to methanol via inhalation during gestation
Effect
Exposure concentration (ppm)
0
200
1,000
5,000
Reproductive effects
Number of pregnant females examined
Number of corpora lutea
Number of implantations
Number of resorptions
Number of live fetuses
Sex ratio (M/F)
Fetal weight (male)
Fetal weight (female)
Total resorption rate (%)
19
17.0 ±2.6
15.7±1.6
0.79 ±0.85
14. 95 ±1.61
144/140
3.70 ±0.24
3. 51 ±0.19
11.2±9.0
24
17.2 ±2.7
15.0 ±3.0
0.71 ±1.23
14.25 ±3. 54
177/165
3.88 ±0.23
3.60 ±0.25
15.6±21.3
22
16.4 ±1.9
15.5 ±1.2
0.95 ±0.65
14. 55 ±1.1
164/156
3.82 ±0.29
3.60 ±0.30
10.6 ±8.4
20
16. 5 ±2.4
14.5 ±3.3
1.67 ±2.03
12.86±4.04a
134/136
3.02±0.27C
2.83±0.26C
23.3±22.7a
Soft tissue malformations
Number of fetuses examined
Abnormality at base of right subclavian
Excessive left subclavian
Ventricular septal defect
Residual thymus
Serpengious urinary tract
136
0.7 ±2.87(1)
0
0
2.9 ±5. 91 (4)
43.0 ±24.64 (18)
165
0
0
0.6 ±2. 96(1)
2.4 ±5.44 (4)
35.2 ±31.62 (19)
154
0
0
0
2.6 ±5.73 (4)
41. 8 ±38.45 (15)
131
0
3.5 ±9.08 (3)
47.6 ± 36.51 (16)b
53.3±28.6(20)b
22.1 ±22.91 (13)
Skeletal abnormalities
Number of fetuses examined
Atresia of foramen costotransversarium
Patency of foramen costotransversium
Cleft sternum
Split sternum
Bifurcated vertebral center
Cervical rib
Excessive sublingual neuropore
Curved scapula
Waved rib
Abnormal formation of lumbar
vertebrae
148
23. 5 ±5.47 (3)
0
0
0
0.8 ±3.28(1)
0
0
0
0
0
177
7.7 ±1.3 (8)
0
0
0
1.6 ±5.61 (2)
0
0
0
0
0
165
3. 5 ±8. 88 (4)
0.6 ±2.67(1)
0
0
3.0 ±8.16 (3)
0
0
0
0
0
138
45.2±25.18(20)b
13.7 ±20.58 (7)
5.6 ±14. 14 (3)
7.0 ±14.01 (5)
14.5 ± 16.69 (ll)b
65.2±25.95(19)b
49. 9 ±27.31 (19)
0.7±3.19(1)
6.1 ±11. 84 (5)
0.7±3.19(1)
ap < 0.05, bp < 0.01, °p < 0.001, as calculated by the authors.
Values are means ± S.D. Values in parentheses are the numbers of litters.
Source: NEDO (1987).
4-36
DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-7. Reproductive parameters in Sprague-Dawley dams exposed to methanol
during pregnancy then allowed to deliver their pups
Effect
Number of dams
Duration of gestation (days)
Number of implantations
Number of pups
Number of live pups
Number of live pups on PND4
Sex ratio (M/F)
Postimplantation embryo
survival rate
Exposure concentration (ppm)
0
12
21.9 ±0.3
15.8 ±1.6
15.2 ±1.6
15.2 ±1.6
15.0 ±1.7 (2)
88/94
96.3 ±4.2
200
12
21.9 ±0.3
14.8 ±1.2
14.4 ±1.3
14.1 ±1.4
13.8 ±1.5 (3)
87/85
94.9 ±5.1
1,000
12
21.9 ±0.3
15.3 ±1.3
14.5 ±1.4
14.3 ±1.4
14.2 ±1.6(1)
103/703
93.6 ±6.1
5,000
12
22.6±0.5C
14.6 ±l.la
13.1±2.2a
12.6±2.5b
10.3 ± 2.8 (9)c
75/81
86.2±16.2a
ap < 0.05, bp < 0.01, °p < 0.001, as calculated by the authors.
Values are means ± S.D. Values in parentheses are the numbers of litters.
Source: NEDO(1987).
1
2 NEDO (1987) contains an account of a two-generation reproductive study that evaluated
3 the effects of pre- and postnatal methanol (reagent grade) exposure (20 hours/day) on
4 reproductive and other organ systems of Sprague-Dawley rats. The F0 generation (30 males and
5 30 females per exposure group)37 was exposed to 0, 10, 100, and 1,000 ppm (0, 13.1, 131, and
6 1,310 mg/m3) from 8 weeks old to the end of mating (males) or to the end of lactation period
7 (females). The FI generation was exposed to the same concentrations from birth to the end of
8 mating (males) or to weaning of F2 pups 21 days after delivery (females). Males and females of
9 the F2 generation were exposed from birth to 21 days old (one animal/sex/litter was exposed to
10 8 weeks of age). NEDO (1987) noted reduced brain, pituitary, and thymus weights, and early
11 testicular descent in the offspring of F0 and FI rats exposed to 1,000 ppm methanol. The early
12 testicular descent is believed to be an indication of earlier fetal development as indicated by the
13 fact that it was correlated with increased pup body weight. However, no histopathologic effects
14 of methanol were observed. As discussed in the report, NEDO (1987) sought to confirm the
15 possible compound-related effect of methanol on the brain by carrying out an additional study in
16 which Sprague-Dawley rats were exposed to 0, 500, 1,000, and 2,000 ppm (0, 655, 1,310, and
17 2,620 mg/m3) methanol from the first day of gestation through the FI generation (see
18 Section 4.4.2).
A second control group of 30 animals/sex was maintained in a separate room to "confirm that environmental conditions inside
the chambers were not unacceptable to the animals."
4-37
DRAFT-DO NOT CITE OR QUOTE
-------
1 Rogers et al. (1993a) evaluated development toxicity in pregnant female CD-I mice
2 exposed to air or 1,000, 2,000, 5,000, 7,500, 10,000, or 15,000 ppm (0, 1,310, 2,620, 6,552,
3 9,894, 13,104, and 19,656 mg/m3) methanol vapors (> 99.9% purity) in a chamber for
4 7 hours/day on GD6-GD15 in a 3-block design experiment. The numbers of mice exposed at
5 each dose were 114, 40, 80, 79, 30, 30, and 44, respectively. During chamber exposures to air or
6 methanol, the mice had access to water but not food. In order to determine the effects of the
7 chamber exposure conditions, an additional 88 control mice were not handled and remained in
8 their cages; 30 control mice were not handled but were food deprived for 7 hours/day on GD6-
9 GDIS. Effects in dams and litters were statistically analyzed using the General Linear Models
10 procedure and multiple t-tesi of least squares means for continuous variables and the Fisher's
11 exact test for dichotomous variables. An analysis of plasma methanol levels in 3 pregnant
12 mice/block/treatment group on GD6, GD10, and GDIS revealed a dose-related increase in
13 plasma methanol concentration that did not seem to reach saturation levels, and methanol plasma
14 levels were not affected by gestation stage or number of previous exposure days. Across all
15 3 days, the mean plasma methanol concentrations in pregnant mice were approximately 97, 537,
16 1,650, 3,178, 4,204, and 7,330 |ig/mL in the 1,000, 2,000, 5,000, 7,500, 10,000, and 15,000 ppm
17 exposure groups, respectively.
18 The dams exposed to air or methanol in chambers gained significantly less weight than
19 control dams that remained in cages and were not handled. There were no methanol-related
20 reductions in maternal body weight gain or overt signs of toxicity. Dams were sacrificed on
21 GDI 7 for a comparison of developmental toxicity in methanol-treated groups versus the chamber
22 air-exposed control group. Fetuses in all exposure groups were weighed, assessed for viability,
23 and examined for external malformations. Fetuses in the control, 1,000, 2,000, 5,000, and
24 15,000 ppm groups were also examined for skeletal and visceral defects. Incidence of
25 developmental effects is listed in Table 4-8. A statistically significant increase in cervical
26 ribs/litter was observed at concentrations of 2,000, 5,000, and 15,000 ppm. At doses of
27 > 5,000 ppm the incidences of cleft palates/litter and exencephaly/litter were increased with
28 statistical significance achieved at all concentrations with the exception of exencephaly which
29 increased but not significantly at 7,500 ppm.38 A significant reduction in live pups/litter was
30 noted at >7,500 ppm, with a significant increase in fully resorbed litters occurring at
31 > 10,000 ppm. Fetal weight was significantly reduced at > 10,000 ppm. Rogers et al. (1993a)
32 identified a developmental NOAEL and LOAEL of 1,000 ppm and 2,000 ppm, respectively.
33 They also provide BMD maximum likelihood estimates (benchmark concentration [BMC];
34 referred to by the authors as MLE) and estimates of the lower 95% confidence limit on the BMC
38 Due to the serious nature of this response and the relative lack of a response in controls, all incidence of
exenceaphaly reported in this study at 5,000 ppm or higher are considered biologically significant.
4-3 8 DRAFT-DO NOT CITE OR QUOTE
-------
1 (benchmark concentration, 95% lower bound [BMCL]; referred to as benchmark dose [BMD] by
2 Rogers et al. (1993a) for 5% and 1% added risk, by applying a log-logistic dose-response model
3 to the mean percent/litter data for cleft palate, exencephaly and resorption. The BMCos and
4 BMCLos values for added risk estimated by Rogers et al. (1993a) are listed in Table 4-9. From
5 this analysis, the most sensitive indicator of developmental toxicity was an increase in the
6 proportion of fetuses per litter with cervical rib anomalies. The most sensitive BMCL and BMC
7 from this effect for 5% added risk were 305 ppm (400 mg/m3) and 824 ppm (1,080 mg/m3),
8 respectively.39
Table 4-8. Developmental effects in mice after methanol inhalation
Endpoint
No. live pups/litter
No. fully resorbed litters
Fetus weight (g)
Cleft palate/ litter (%)
Exencephaly /litter (%)
Exposure concentration (ppm)
0
9.9
0
1.20
0.21
0
1,000
9.5
0
1.19
0.65
0
2,000
12.0
0
1.15
0.17
0.88
5,000
9.2
0
1.15
8.8b
6.9a
7,500
8.6b
3
1.17
46.6C
6.8
10,000
7.3C
5a
1.04C
52.7C
27.4C
15,000
2.2C
14C
0.70C
48.3C
43.3C
Anomalies
Cervical ribs/litter (%)
Sternebral defects/litter (%)
Xiphoid defects/litter (%)
Vertebral arch defects/litter (%)
Extra lumbar ribs/litter (%)
28
6.4
6.4
0.3
8.7
33.6
7.9
3.8
ND
2.5
49.6b
3.5
4.1
ND
9.6
74.4C
20.2C
10.9
1.5
15.6
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
60.0a
100C
73.3C
33.3C
40.0C
Ossifications (values are means of litter means)
Sternal
Caudal
Metacarpal
Proximal phalanges
Metatarsals
Proximal phalanges
Distal phalanges
Supraoccipital score+
5.96
5.93
7.96
7.02
9.87
7.18
9.64
1.40
5.99
6.26
7.92
7.04
9.90
7.69
9.59
1.65
5.94
5.71a
7.96
7.04
9.87
6.91
9.57
1.57
5.81
5.42
7.93
6.12
9.82
5.47
8.46b
1.48
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
5.07C
3.20a
7.60b
3.33C
8.13C
Oc
4.27C
3.20C
ND = Not determined. = on a scale of 1-4, where 1 is fully ossified and 4 is unossified.
Statistical significance: ap < 0.05, bp < 0.01, cp < 0.001, as calculated by the authors.
Source: Rogers etal. (1993a).
39 The BMD analysis of the data described in Section 5 was performed similarly using, among others, a similar
nested logistic model. However, the Rogers et al. (1993a) analysis was performed using added risk and external
exposure concentrations, whereas the analyses in Section 5 used extra risk and internal dose metrics that were then
converted to human equivalent exposure concentrations.
4-39 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-9. Benchmark doses at two added risk levels
Endpoint
Cleft Palate (CP)
Exencephaly (EX)
CPandEX
Resorptions (RES)
CP, EX, and RES
Cervical ribs
BMCos (ppm)
4,314
5,169
3,713
5,650
3,667
824
BMCL05(ppm)
3,398
3,760
3,142
4,865
3,078
305
BMCoi (ppm)
2,717
2,122
2,381
3,749
2,484
302
BMCLoi (ppm)
1,798
784
1,816
2,949
1,915
58
Source: Rogers etal. (1993a).
1 Rogers and Mole (1997) investigated the critical period of sensitivity to the
2 developmental toxicity of inhaled methanol in the CD-I mouse by exposing 12-17 pregnant
3 females to 0 or 10,000 ppm (0 and 13,104 mg/m3), 7 hours/day on 2 consecutive days during
4 GD6-GD13, or to a single exposure to the same methanol concentration during GD5-GD9.
5 Another group of mice received a single 7-hour exposure to methanol at 10,000 ppm. The latter
6 animals were sacrificed at various time intervals up to 28 hours after exposure. Blood samples
7 were taken from these animals to measure the concentration of methanol in the serum. Serum
8 methanol concentrations peaked at ~4 mg/mL 8 hours after the onset of exposure. Methanol
9 concentrations in serum had declined to pre-exposure levels after 24 hours. All mice in the main
10 body of the experiment were sacrificed on GDI 7, and their uteri removed. The live, dead, and
11 resorbed fetuses were counted, and all live fetuses were weighed, examined externally for cleft
12 palate, and then preserved. Skeletal abnormalities were determined after the carcasses had been
13 cleaned and eviscerated. Cleft palate, exencephaly, and skeletal defects were observed in the
14 fetuses of exposed dams. For example, cleft palate was observed following 2-day exposures to
15 methanol on GD6-GD7 through GDI 1-GD12. These effects also were apparent in mice
16 receiving a single exposure to methanol on GD5-GD9. This effect peaked when the dams were
17 exposed on GD7. Exencephaly showed a similar pattern of development in response to methanol
18 exposure. However, the data indicated that cleft palate and exencephaly might be competing
19 malformations, since only one fetus displayed both features. Skeletal malformations included
20 exoccipital anomalies, atlas and axis defects, the appearance of an extra rudimentary rib on
21 cervical vertebra No.7, and supernumerary lumbar ribs. In each case, the maximum time point
22 for the induction of these defects appeared to be when the dams were exposed to methanol on or
23 near GD7. When dams were exposed to methanol on GD5, there was also an increased
24 incidence of fetuses with 25 presacral vertebrae (26 is normal). However, an increased incidence
25 of fetuses with 27 presacral vertebrae was evident when dams were exposed on GD7. These
26 results indicate that gastrulation and early organogenesis is a period of increased embryonic
27 sensitivity to methanol.
4-40
DRAFT-DO NOT CITE OR QUOTE
-------
1 Burbacher et al. (1999a, 1999b) carried out toxicokinetic and reproductive/developmental
2 studies of methanol in M. fascicularis monkeys that were published by the Health Effects
3 Institute (HEI) in a two-part monograph. Some of the data were subsequently published in the
4 open scientific literature (Burbacher et al., 2004a, 2004b). The experimental protocol featured
5 exposure to 2 cohorts of 12 monkeys/group to low exposure levels (relative to the previously
6 discussed rodent studies) of 0, 200, 600, or 1,800 ppm (0, 262, 786, and 2,359 mg/m3) methanol
7 vapors (99.9% purity), 2.5 hours/day, 7 days/week, during a premating period and mating period
8 (-180 days combined) and throughout the entire gestation period (-168 days). The monkeys
9 were 5.5-13 years old and were a mixture of feral-born and colony-bred animals. The study
10 included an evaluation of maternal reproductive performance and tests to assess infant postnatal
11 growth and newborn health, reflexes, behavior, and development of visual, sensorimotor,
12 cognitive, and social behavioral function (see Section 4.4.2 for a review of the developmental
13 neurotoxicity findings from this study). Blood methanol levels, clearance, and the appearance of
14 formate were also examined and are discussed in Section 3.2.
15 With regard to reproductive parameters, there was a statistically significant decrease
16 (p = 0.03) in length of pregnancy in all treatment groups, as shown in Table 4-10. Maternal
17 menstrual cycles, conception rate, and live birth index were all unaffected by exposure. There
18 were also no signs of an effect on maternal weight gain or clinical toxicity among the dams. The
19 decrease in pregnancy length was largely due to complications of pregnancy requiring Cesarean
20 section (C-section) deliveries in the methanol exposure groups. The C-section deliveries were
21 performed in response to signs of difficulty in the pregnancy and thus may serve as supporting
22 evidence of reproductive dysfunction in the methanol-exposed females.
23 While pregnancy duration was virtually the same in all exposure groups, there were some
24 indications of increased pregnancy duress only in methanol-exposed monkeys. C-sections were
25 done in 2 monkeys from the 200 ppm group and 2 from the 600 ppm group due to vaginal
26 bleeding, presumed, but not verified, to be from placental detachment.40 A monkey in the
27 1,800 ppm group also received a C-section after experiencing nonproductive labor for 3 nights.
28 In addition, signs of prematurity were observed in 1 infant from the 1,800 ppm group that was
29 born after a 150-day gestation period. The authors speculated that the shortened gestation length
30 could be due to a direct effect of methanol on the fetal hypothalamus-pituitary-adrenal (HPA)
31 axis or an indirect effect of methanol on the maternal uterine environment. Other fetal
32 parameters such as crown-rump length and head circumference were unchanged among the
33 groups. Infant growth and tooth eruption were unaffected by prenatal methanol exposure.
40 Burbacher et al. (2004a, 2004b) note, however, that in studies of pregnancy complication in alcohol- exposed
human subjects, an increased incidence of uterine bleeding and abrutio placenta has been reported.
4-41 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-10. Reproductive parameters in monkeys exposed via inhalation to methanol
during prebreeding, breeding, and pregnancy
Exposure (ppm)
0
200
600
1,800
Conception rate
9/11
9/12
9/11
10/12
Weight gain (kg)
1.67 ±0.07
1.27 ±0.14
1.78 ±0.25
1.54 ±0.20
Pregnancy duration (days)3
168 ±2
160 ± 2b
162 ± 2b
162 ± 2b
Live born delivery rate
8/9
9/9
8/9
9/10
Values are means ± SE.;
aLive-born offspring only; bp < 0.05, as calculated by the authors.
Source: Burbacher et al. (2004a).
1 In later life, 2 females out of the total of 9 offspring in the 1,800 ppm group experienced
2 a wasting syndrome at 12 and 17 months of age. Food intake was normal and no cause of the
3 syndrome could be determined in tests for viruses, hematology, blood chemistry, and liver,
4 kidney, thyroid, and pancreas function. Necropsies revealed gastroenteritis and severe
5 malnourishment. No infectious agent or other pathogenic factor could be identified. Thus, it
6 appears that a highly significant toxicological effect on postnatal growth can be attributed to
7 prenatal methanol exposure at 1,800 ppm (2,300 mg/m3).
8 In summary, the Burbacher et al. (2004a, 2004b, 1999b) studies suggest that methanol
9 exposure can cause reproductive effects, manifested as a shortened mean gestational period due
10 to pregnancy complications that precipitated delivery via a C-section, and developmental
11 neurobehavioral effects which may be related to the shortened gestational period (see
12 Section 4.4.2). The low exposure of 200 ppm may signify a LOAEL for reproductive effects.
13 However, the decrease in gestational length was marginally significant and largely the result of
14 human intervention (C-section) for reasons (presumably pregnancy complications) that were not
15 objectively confirmed with clinical procedures (e.g., placental ultrasound). Also, this effect did
16 not appear to be dose related, the greatest gestational period decrease having occurred at the
17 lowest (200 ppm) exposure level. Thus, a clear NOAEL or LOAEL cannot be determined from
18 this study.
19 In a study of the testicular effects of methanol, Cameron et al. (1984) exposed 5 male
20 Sprague-Dawley rats/group to methanol vapor, 8 hours/day, 5 days/week for 1, 2, 4, and 6 weeks
21 at 0, 200, 2,000, or 10,000 ppm (0, 262, 2,620, and 13,104 mg/m3). The authors examined the
22 possible effects of methanol on testicular function by measuring blood levels of testosterone,
23 luteinizing hormone (LH), and follicular stimulating hormone (FSH) using radioimmunoassay.
24 When the authors tabulated their results as a percentage of the control value for each duration
25 series, the most significant changes were in blood testosterone levels of animals exposed to
26 200 ppm methanol, the lowest concentration evaluated. At this exposure level, animals exposed
27 for 6 weeks had testosterone levels that were 32% of those seen in controls. However, higher
4-42
DRAFT-DO NOT CITE OR QUOTE
-------
1 concentrations of methanol were associated with testosterone levels that were closer to those of
2 controls. However, the lack of a clear dose-response is not necessarily an indication that the
3 effect is not related to methanol. The higher concentrations of methanol could be causing other
4 effects (e.g., liver toxicity) which can influence the results. Male rats exposed to 10,000 ppm
5 methanol for 6 weeks displayed blood levels of LH that were about 3 times higher (mean ± S.D.)
6 than those exposed to air (311 ± 107% versus 100 ± 23%). In discussing their results, the authors
7 placed the greater emphasis on the fact that an exposure level equal to the ACGIH TLV
8 (200 ppm) had caused a significant depression in testosterone formation in male rats.
9 A follow-up study report by the same research group (Cameron et al., 1985) described the
10 exposure of 5 male Sprague-Dawley rats/group, 6 hours/day for either 1 day or 1 week, to
11 methanol, ethanol, n-propanol, or n-butanol at their respective TLVs. Groups of animals were
12 sacrificed immediately after exposure or after an 18-hour recovery period, and the levels of
13 testosterone, LH, and corticosterone measured in serum. As shown in Table 4-11, the data were
14 consistent with the ability of these aliphatic alcohols to cause a transient reduction in the
15 formation of testosterone. Except in the case of n-butanol, rapid recovery from these deficits can
16 be inferred from the 18-hour postexposure data.
Table 4-11. Mean serum levels of testosterone, luteinizing hormone, and
corticosterone (± S.D.) in male Sprague-Dawley rats after inhalation of methanol,
ethanol, n-propanol or n-butanol at threshold limit values
Testosterone (as a percentage of control)
Condition
Control
Methanol
Ethanol
n-Propanol
n-Butanol
TLV
(ppm)
200
1,000
200
50
Single-day exposure
End of exposure
100 ±17
41±16a
64 ± 12a
58±15a
37±8a
18 hr
postexposure
100 ± 20
98 ±18
86 ±16
81 ±13
52 ± 22a
One-week exposure
End of exposure
100 ± 26
81 ±22
88 ±14
106 ± 28
73 ±34
18 hr
postexposure
100 ± 17
82 ±27
101 ±13
89 ±17
83 ±18
Luteinizing hormone
Control
Methanol
Ethanol
n-Propanol
n-Butanol
200
1,000
200
50
100 ±30
86 ±32
110 ±22
117±59
124 ±37
100 ±35
110 ±40
119±54
119±83
115±28
100 ± 28
78 ±13
62 ±26
68 ±22
78 ±26
100 ± 36
70 ±14
81 ±17
96 ±28
98 ±23
4-43
DRAFT-DO NOT CITE OR QUOTE
-------
Testosterone (as a percentage of control)
Condition
TLV
(ppm)
Single-day exposure
End of exposure
18 hr
postexposure
One-week exposure
End of exposure
18 hr
postexposure
Corticosterone
Control
Methanol
Ethanol
n-Propanol
n-Butanol
200
1,000
200
50
100 ± 20
115±18
111±32
112±21
143 ±Ha
ND
ND
ND
ND
ND
100 ±21
74 ±26
60 ±25
79 ±14
85 ±26
ND
ND
ND
ND
ND
ND = Nodata.;
ap < 0.05, as calculated by the authors.
Source: Cameron etal. (1985).
2 In a series of studies that are relevant to the reproductive toxicity of methanol in males,
3 Lee et al. (1991) exposed 8-week-old male Sprague-Dawley rats (9-10/group) to 0 or 200 ppm
4 (0 and 262 mg/m3) methanol, 8 hours/day, 5 days/week, for 1, 2, 4, or 6 weeks to measure the
5 possible treatment effects on testosterone production. Study results were evaluated by one factor
6 ANOVA followed by Student's ^-test. In the treated rats, there was no effect on serum
7 testosterone levels, gross structure of reproductive organs, or weight of testes and seminal
8 vesicles. Lee et al. (1991) also studied the in vitro effect of methanol on testosterone production
9 from isolated testes, but saw no effect on testosterone formation either with or without the
10 addition of human chorionic gonadotropin hormone.
11 In a third experiment from the same report, Lee et al. (1991) examined testicular
12 histopathology to determine if methanol exposure produced lesions indicative of changing
13 testosterone levels; the effects of age and folate status were also assessed. This is relevant to the
14 potential toxicity of methanol because folate is the coenzyme of tetrahydrofolate synthetase, an
15 enzyme that is rate limiting in the removal of formate. Folate deficiency would be expected to
16 cause potentially toxic levels of methanol, formaldehyde, and formate to be retained. The same
17 authors examined the relevance of folate levels, and by implication, the overall status of formate
18 formation and clearance in mediating the testicular functions of Long-Evans rats. Groups of
19 4-week-old male Long-Evans rats were given diets containing either adequate or reduced folate
20 levels plus 1% succinylsulfathiazole, an antibiotic that, among other activities,41 would tend to
21 reduce the folate body burden. At least 9 rats/dietary group/dose were exposed to 0, 50, 200, or
22 800 ppm (0, 66, 262, and 1,048 mg/m3) methanol vapors starting at 7 months of age while 8-
23 12 rats/dietary group/dose were exposed to 0 or 800 ppm methanol vapors at 15 months of age.
41 Succinylsulfathiazone antibiotic may have a direct impact on the effects being measured, the extent of which was
not addressed by the authors of this study.
4-44
DRAFT-DO NOT CITE OR QUOTE
-------
1 The methanol exposures were conducted continuously for 20 hours/day for 13 weeks. Without
2 providing details, the study authors reported that visual toxicity and acidosis developed in rats
3 fed the low folate diet and exposed to methanol. No methanol-related testicular lesions or
4 changes in testes or body weight occurred in rats that were fed either the folate sufficient or
5 deficient diets and were 10 months old at the end of treatment. Likewise, no methanol-lesions
6 were observed in 18-month-old rats that were fed diets with adequate folate. However, the
7 incidence but not severity of age-related testicular lesions was increased in the 18-month-old rats
8 fed folate-deficient diets. Subcapsular vacuoles in germinal epithelium were noted in 3/12
9 control rats and 8/13 rats in the 800 ppm group. One rat in the 800 ppm group had atrophied
10 seminiferous tubules and another had Ley dig cell hyperplasia. These effects, as well as the
11 transient decrease in testosterone levels observed by Cameron et al. (1985, 1984), could be the
12 result of chemically-related strain on the rat system as it attempts to maintain hormone
13 homeostasis.
14 Dorman et al. (1995) conducted a series of in vitro and in vivo studies of developmental
15 toxicity in ICR BR (CD-I) mice associated with methanol and formate exposure. The studies
16 used HPLC grade methanol and appropriate controls. PK and developmental toxicity parameters
17 were measured in mice exposed to a 6-hour methanol inhalation (10,000 or 15,000 ppm),
18 methanol gavage (1.5 g/kg) or sodium formate (750 mg/kg by gavage) on GD8. In the in vivo
19 inhalation study, 12-14 dams/group were exposed to 10,000 ppm methanol for 6 hours on
20 GD8,42 with and without the administration of fomepizole (4-methylpyrazole) to inhibit the
21 metabolism of methanol by ADH1. Dams were sacrificed on GD10, and folate levels in
22 maternal RBC and conceptus (decidual swelling) were measured, as well as fetal neural tube
23 patency (an early indicator of methanol-induced dysmorphogenic response). The effects
24 observed included a transient decrease in maternal RBC and conceptus folate levels within
25 2 hours following exposure and a significant (p < 0.05) increase in the incidence of fetuses with
26 open neural tubes (9.65% in treated versus 0 in control). These responses were not observed
27 following sodium formate administration, despite peak formate levels in plasma and decidual
28 swellings being similar to those observed following the 6-hour methanol inhalation of
29 15,000 ppm. This suggests that these methanol-induced effects are not related to the
30 accumulation of formate. As this study provides information relevant to the identification of the
31 proximate teratogen associated with developmental toxicity in rodents, it is discussed more
32 extensively in Section 4.6.1.
42 Dorman et al. (1985) state that GD8 was chosen because it encompasses the period of murine neurulation and the
time of greatest vulnerability to methanol-induced neural tube defects.
4-45 DRAFT-DO NOT CITE OR QUOTE
-------
4.3.3. Other Reproductive and Developmental Toxicity Studies
1
2
3
4
5
6
7
8
9
10
11
12
Additional information relevant to the possible effects of methanol on reproductive and
developmental parameters has been provided by experimental studies that have exposed
experimental animals to methanol during pregnancy via i.p. injections (Rogers et al., 2004).
Relevant to the developmental impacts of the chemical, a number of studies also have examined
the effects of methanol when included in whole-embryo culture (Hansen et al., 2005; Harris
et al., 2003; Andrews etal., 1998, 1995, 1993).
Pregnant female C57BL/6J mice received 2 i.p. injections of methanol on GD7 (Rogers
et al., 2004). The injections were given 4 hours apart to provide a total dosage of 0, 3.4, and 4.9
g/kg. Animals were sacrificed on GDI7 and the litters were examined for live, dead, and
resorbed fetuses. Rogers et al. (2004) monitored fetal weight and examined the fetuses for
external abnormalities and skeletal malformations. Methanol-related deficits in maternal and
litter parameters observed by Rogers et al. (2004) are summarized in Table 4-12.
Table 4-12. Maternal and litter parameters when pregnant female C57BL/6J mice
were injected i.p. with methanol
Parameter
No. pregnant at term
WtgainGD7-GD8(g)
WtgainGD7-GD10(g)
Live fetuses/litter
Resorbed fetuses/litter
Dead fetuses/litter
Fetal weight (g)
Methanol dose (g/kg)
0
43
0.33 ±0.10
1.63 ±0.18
7.5 ±0.30
0.4 ±0.1
0.1±0.1
0.83 ±0.02
3.4
13
0.37 ±0.15
2.20 ± 0.20
6.3±0.5a
1.3±0.4a
0
0.82 ±0.03
4.9
24
-0.24±0.14a
1.50 ±0.20
3.7±0.4a
4.4±0.4a
0.1±0.1
0.70 ± 0.02a
13
14
15
16
17
18
19
20
21
Values are means ± SEM.
ap < 0.05, as calculated by the authors.
Source: Rogers et al. (2004).
Rogers et al. (2004) used a number of sophisticated imaging techniques, such as confocal
laser scanning and fluorescence microscopy, to examine the morphology of fetuses excised at
GD7, GD8, and GD9. They identified a number of external craniofacial abnormalities, the
incidence of which was, in all cases, significantly increased in the high-dose group compared to
controls. For some responses, such as microanophthalmia and malformed maxilla, the incidence
was also significantly increased in animals receiving the lower dose. Fifteen compound-related
skeletal malformations were tabulated in the report. In most cases, a dose-response effect was
evident, resulting in statistically significant incidences in affected fetuses and litters, when
compared to controls. Apparent effects of methanol on the embryonic forebrain included a
4-46
DRAFT-DO NOT CITE OR QUOTE
-------
1 narrowing of the anterior neural plate, missing optical vesicles, and holoprosencephaly (failure of
2 the embryonic forebrain to divide). The authors noted that there was no sign of incipient cleft
3 palate or exencephaly, as had been observed in CD-I mice exposed to methanol via the oral and
4 inhalation routes (Rogers et al., 1993a).
5 In order to collect additional information on cell proliferation and histological changes in
6 methanol-treated fetuses, Degitz et al. (2004a) used an identical experimental protocol to that of
7 Rogers et al. (2004) by administering 0, 3.4, or 4.9 g methanol/kg in distilled water i.p. (split
8 doses, 4 hours apart) to C57BL/6J mice on GD7. Embryos were collected at various times on
9 GD8 and GD10. Embryos from dams exposed to 4.9 g/kg and examined on GD8 exhibited
10 reductions in the anterior mesenchyme, the mesenchyme subjacent to the mesencephalon and the
11 base of the prosencephalon (embryonic forebrain), and in the forebrain epithelium. The optic
12 pits were often lacking; where present their epithelium was thin and there were fewer neural
13 crest cells in the mid- and hindbrain regions.
14 At GD9, there was extensive cell death in areas populated by the neural crest, including
15 the forming cranial ganglia. Dose-related abnormalities in the development of the cranial nerves
16 and ganglia were seen on GD7. In accordance with an arbitrary dichotomous scale devised by
17 the authors, scores for ganglia V, VIII, and IX were significantly (not otherwise specified)
18 reduced at all dose levels, and ganglia VII and X were reduced only at the highest dose. At the
19 highest dose (4.9 g/kg), the brain and face were poorly developed and the brachial arches were
20 reduced in size or virtually absent. Flow cytometry of the head regions of the embryos from the
21 highest dose at GD8 did not show an effect on the proportion of cells in S-phase.
22 Cell growth and development were compared in C57BL/6J and CD-I mouse embryos
23 cultured in methanol (Degitz, et al., 2004b). GD8 embryos, with 5-7 somites, were cultured in
24 0, 1, 2, 3, 4, or 6 mg methanol/mL for 24 hours and evaluated for morphological development.
25 Cell death was increased in both strains in a developmental stage- and region-specific manner at
26 4 and 6 mg/mL after 8 hours of exposure. The proportions of cranial region cells in S-phase
27 were significantly (p < 0.05) decreased at 6 mg/mL following 8- and 18-hour exposures to
28 methanol. After 24 hours of exposure, C57BL/6J embryos had significantly (p < 0.05) decreased
29 total protein at 4 and 6 mg/kg. Significant (p < 0.05) developmental effects were seen at 3, 4,
30 and 6 mg/kg, with eye dysmorphology being the most sensitive endpoint. CD-I embryos had
31 significantly decreased total protein at 3, 4, and 6 mg/kg, but developmental effects were seen
32 only at 6 mg/kg. It was concluded that the C57BL/6J embryos were more severely affected by
33 methanol in culture than the CD-I embryos.
34 Andrews et al. (1993) carried out a comparative study of the developmental toxicity of
35 methanol in whole Sprague-Dawley rat or CD-I mouse embryos. Nine-day rat embryos were
36 explanted and cultured in rat serum containing 0, 2, 4, 8, 12, or 16 mg/mL methanol for 24 hours
4-47 DRAFT-DO NOT CITE OR QUOTE
-------
1 then transferred to rat serum alone for a further 24 hours. Eight-day mouse embryos were
2 cultured in 0, 2, 4, 6, or 8 mg/mL methanol in culture medium for 24 hours. At the end of the
3 culture period, embryos were examined for growth, development and dysmorphogenesis. For
4 the rats, doses of 8 mg/mL and above resulted in a concentration-related decrease in somite
5 number, head length, and developmental score. Some lethality was seen in embryos incubated at
6 12 mg/mL methanol. For the mouse embryos, incubation concentrations of 4 mg/mL methanol
7 and above resulted in a significant decrease in developmental score and crown-rump length. The
8 high concentration (8 mg/mL) was associated with embryo lethality. These data suggest that
9 mouse embryos are more sensitive than rat embryos to the developmental effects of methanol.
10 Using a similar experimental system to examine the developmental toxicity of formate and
11 formic acid in comparison to methanol, Andrews et al. (1995) showed that the formates are
12 embryotoxic at doses that are four times lower than equimolar doses of methanol. Andrews et al.
13 (1998) showed that exposure to combinations of methanol and formate was less embryotoxic
14 than would be expected based on simple toxicity additivity, suggesting that the embryotoxicity
15 observed following low-level exposure to methanol is mechanistically different from that
16 observed following exposure to formate.
17 A study by Hansen et al. (2005) determined the comparative toxicity of methanol and its
18 metabolites, formaldehyde and sodium formate, in GD8 mouse (CD-I) and GD10 rat (Sprague-
19 Dawley) conceptuses. Incubation of whole embryos was for 24 hours in chemical-containing
20 media (mouse: 4-12 mg/mL methanol, 1-6 |ig/mL formaldehyde, 0.5-4 mg/mL sodium formate;
21 rat: 8-20 mg/mL, 1-8 jig/mL, 0.5-8 mg/mL). Subsequently, the visceral yolk sac (VYS) was
22 removed and frozen for future protein and DNA determination. The embryos were examined
23 morphologically to determine growth and developmental parameters such as viability, flexure
24 and rotation, crown-rump length, and neuropore closure. In other experiments, the chemicals
25 were injected directly into the amniotic space. For each response, Table 4-13 provides a
26 comparison of the concentrations or amounts of methanol, formaldehyde, and formate that
27 resulted in statistically significant changes in developmental abnormalities compared to controls.
28 For a first approximation, these concentrations or amounts may be taken as threshold-
29 dose ranges for the specific responses under the operative experimental conditions. The data
30 show consistently lower threshold values for the effects of formaldehyde compared to those of
31 formate and methanol. The mouse embryos were more sensitive towards methanol toxicity than
32 rat embryos, consistent with in vivo findings, whereas the difference in sensitivity disappeared
33 when formaldehyde was administered. Hansen et al. (2005) hypothesized that, while the MOA
34 for the initiation of the organogenic defects is unknown, the relatively low threshold levels of
35 formaldehyde for most measured effects suggest formaldehyde involvement in the embryotoxic
36 effects of methanol. By contrast, formate, the putative toxicant for the acute effects of methanol
4-48 DRAFT-DO NOT CITE OR QUOTE
-------
1 poisoning (acidosis, neurological deficits), did not appear to reproduce the methanol-induced
2 teratogenicity in these whole embryo culture experiments.
Table 4-13. Reported thresholds concentrations (and author-estimated ranges) for the
onset of embryotoxic effects when rat and mouse conceptuses were incubated in vitro
with methanol, formaldehyde, and formate
Parameter
Viability (%)
Normal rotation (%)
CRa length
Neural tube closure (%)
Reduced embryo protein
Reduced VYSb protein
Reduced embryo DNA
Reduced VYS DNA
Mouse
Methanol
Formaldehyde
Formate
Rat
Methanol
Formaldehyde
Formate
In vitro incubation (mg/mL)
8.0
4.0
No change
8.0
8.0
10.0
8.0
4.0
0.004
0.003
No change
0.001
0.003
0.004
0.003
0.001
NS
0.5
No change
2.0
4.0
4.0
No change
0.5
16.0
8.0
No change
12.0
8.0
12.0
12.0
12.0
0.006
0.003
No change
No change
0.004
0.004
0.003
0.003
2.0
4.0
No change
No change
2.0
NR
NR
NR
Microinjection (author-estimated dose ranges in jig)
Viability (%)
Normal rotation (%)
CRa length
Neural tube closure (%)
Reduced embryo protein
Reduced VYSb protein
Reduced embryo DNA
Reduced VYSb DNA
46-89
1-45
No change
1-45
1-45
135-178
46-89
1-45
0.003-0.5
0.003-0.5
No change
0.003-0.5
0.501-1.0
1.01-1.5
0.501-1.0
0.003-0.5
1.01-1.5
0.03-0.5
No change
1.01-1.5
No change
No change
No change
0.03-0.5
46-89
46-89
No change
No change
No change
No change
No change
No change
1.01-1.5
1.01-1.5
No change
No change
1.51-2.0
No change
No change
No change
1.51-4.0
0.51-1.0
No change
1.01-1.5
0.51-1.0
1.01-1.5
0.51-1.0
0.51-1.0
4
5
6
7
8
9
10
11
12
aCR = crown-rump length,
bVYS = visceral yolk sac.
NR = not reported
Source: Hansen et al. (2005); Harris et al. (2004) (adapted).
Harris et al. (2003) provided biochemical evidence consistent with the concept that
formaldehyde might be the ultimate embryotoxicant of methanol by measuring the activities of
enzymes that are involved in methanol metabolism in mouse (CD-I) and rat (Sprague-Dawley)
whole embryos at different stages of development. Specific activities of the enzymes ADH1,
ADH3, and CAT, were determined in rat and mouse conceptuses during the organogenesis period
of 8-25 somites. Activities were measured in heads, hearts, trunks, and VYS from early- and
late-stage mouse and rat embryos. While CAT activities were similar between rat and mouse
embryos, mouse ADH1 activities in the VYS were significantly lower throughout organogenesis
when compared to the rat VYS or embryos of either species. ADH1 activities of heads, hearts,
and trunks from mouse embryos were significantly lower than those from rats at the 7-12 somite
4-49
DRAFT-DO NOT CITE OR QUOTE
-------
1 stage. However, these interspecies differences were not evident in embryos of 20-22 somites.
2 ADH3 activities were lower in mouse versus rat VYS, irrespective of the stage of development.
3 However, while ADH3 activities in mouse embryos were markedly lower than those of rats in
4 the early stages of development, the levels of activity were similar to at the 14-16 somite stage
5 and beyond. A lower capacity to transform formaldehyde to formate might explain the increased
6 susceptibility of mouse versus rat embryos to the toxic effects of methanol. The hypothesis that
7 formaldehyde is the ultimate embryotoxicant of methanol is supported by the demonstration of
8 diminished ADH3 activity in mouse versus rat embryos and by the demonstration by Hansen
9 et al. (2005) that formaldehyde has a far greater embryotoxicity than either formate or methanol
10 itself.
11 That formate can induce similar developmental lesions in whole rat and mouse
12 conceptuses was demonstrated by Andrews et al. (1995), who evaluated the developmental
13 effects of sodium formate and formic acid in rodent whole embryo cultures in vitro. Day 9 rat
14 (Sprague-Dawley) embryos were cultured for 24 or 48 hours and day 8 mouse (CD-I) cultures
15 were incubated for 24 hours. As tabulated by the authors, embryos of either species showed
16 trends towards increasing lethality and incidence of abnormalities with exposure concentration.
17 Among the anomalies observed were open anterior and posterior neuropores, plus rotational
18 defects, tail anomalies, enlarged pericardium, and delayed heart development.
4.4. NEUROTOXICITY
19 A substantial body of information exists on the toxicological consequences to humans
20 who consume or are exposed to large amounts of methanol. As discussed in Section 4.1,
21 neurological consequences of acute methanol intoxication in humans include Parkinson-like
22 responses, visual impairment, confusion, headache, and numerous subjective symptoms. The
23 occurrence of these symptoms has been shown to be associated with necrosis of the putamen
24 when neuroimaging techniques have been applied (Salzman, 2006). Such profound changes
25 have been linked to tissue acidosis that arises when methanol is metabolized to formaldehyde
26 and formic acid through the actions of ADH1 and ADH3. However, the well-documented impact
27 of the substantial amounts of formate that are formed when humans and animals are exposed to
28 large amounts of methanol may obscure the potentially harmful effects that may arise when
29 humans and animals exposed to smaller amounts. Human acute exposure studies (Chuwers et al.,
30 1995; Cooketal., 1991) (See Section 4.1.3) at TLV levels of 200 ppm would indicate that some
31 measures of neurological function (e.g., sensory evoked potentials, memory testing and
32 psychomotor testing) were impaired in the absence of measureable formate production.
4-50 DRAFT-DO NOT CITE OR QUOTE
-------
4.4.1. Oral Studies
1 Two rodent studies investigated the neurological effects of developmental methanol
2 exposure via the oral route (Aziz et al., 2002; Infurna and Weiss, 1986). One of these studies also
3 investigated the influence of FAD diets on the effects of methanol exposures (Aziz et al., 2002).
4 In the first, Infurna and Weiss (1986) exposed 10 pregnant female Long-Evans rats/dose to 2%
5 methanol (purity not specified) in drinking water on either GDIS-GDI? or GD17-GD19. Daily
6 methanol intake was calculated at 2,500 mg/kg-day by the study authors. Dams were allowed to
7 litter and nurse their pups. Data were analyzed by ANOVA with the litter as the statistical unit.
8 Results of the study were equivalent for both exposure periods. Treatment had no effect on
9 gestational length or maternal bodyweight. Methanol had no effect on maternal behavior as
10 assessed by the time it took dams to retrieve pups after they were returned to the cage following
11 weighing. Litter size, pup birth weight, pup postnatal weight gain, postnatal mortality, and day
12 of eye opening did not differ from controls in the methanol treated groups. Two neurobehavioral
13 tests were conducted in offspring. Suckling ability was tested in 3-5 pups/treatment group on
14 PND1. An increase in the mean latency for nipple attachment was observed in pups from the
15 methanol treatment group, but the percentage of pups that successfully attached to nipples did
16 not differ significantly between treatment groups. Homing behavior, the ability to detect home
17 nesting material within a cage containing one square of shavings from the pup's home cage and
18 four squares of clean shavings, was evaluated in 8 pups/group on PND10. Pups from both of the
19 methanol exposure groups took about twice as long to locate the home material and took less
20 direct paths than the control pups. Group-specific values differed significantly from controls.
21 This study suggests that developmental toxicity can occur at this drinking water dose without
22 readily apparent signs of maternal toxicity.
23 Aziz et al. (2002) investigated the role of developmental deficiency in folic acid and
24 methanol-induced developmental neurotoxicity in PND45 rat pups. Wistar albino female rats
25 (80/group) were fed FAD43 and FAS diets separately. Following 14-16 weeks on the diets, liver
26 folate levels were estimated and females exhibiting a significantly low folic acid level were
27 mated. Throughout their lactation period, dams of both the FAD and the FAS group were given
28 0,1, 2, or 4% v/v methanol via drinking water, equivalent to approximately 480, 960 and
29 1,920 mg/kg-day.44 Pups were exposed to methanol via lactation from PND1-PND21. Litter size
30 was culled to 8 with equal male/female ratios maintained as much as possible. Liver folate
31 levels were determined at PND21 and neurobehavioral parameters (motor performance using the
43
Along with the FAD diet, 1% succinylsulphathiazole was also given to inhibit folic acid biosynthesis from
intestinal bacteria.
44 Assuming that Wistar rat drinking water consumption is 60 mL/kg-day (Rogers et al., 2002), 1% methanol in
drinking water would be equivalent to 1% x 0.8 g/mL x 60 mL/kg-day = 0.48 g/kg-day = 480 mg/kg-day.
4-51 DRAFT-DO NOT CITE OR QUOTE
-------
1 spontaneous locomotor activity test and cognitive performance using the conditioned avoidance
2 response [CAR] test), and neurochemical parameters (dopaminergic and cholinergic receptor
3 binding and dopamine levels) were measured at PND45. The expression of growth-associated
4 protein (GAP 43), a neuro-specific protein in the hippocampus that is primarily localized in
5 growth cone membranes and is expressed during developmental regenerative neurite outgrowth,
6 was examined using immunohistochemistry and western blot analysis.
7 A loss in body weight gain was observed at PND7, PND14, and PND21 in animals
8 exposed to 2% (11, 15 and 19% weight gain reduction) and 4% (17, 24 and 29% weight gain
9 reduction) methanol in the FAD group and only at 4% (9, 14 and 17% weight gain reduction)
10 methanol in the FAS group. No significant differences in food and water intake were observed
11 among the different treatment groups. Liver folate levels in the FAD group were decreased by
12 63% in rats prior to mating and 67% in pups on PND21.
13 Based on reports of Parkinson-like symptoms in survivors of severe methanol poisoning
14 (see Section 4.1), Aziz et al. (2002) hypothesized that methanol may cause a depletion in
15 dopamine levels and degeneration of the dopaminergic nigrostriatal pathway.45 Consistent with
16 this hypothesis, they found dopamine levels were significantly decreased (32% and 51%) in the
17 striatum of rats in the FAD group treated with 2% and 4% methanol, respectively. In the FAS
18 group, a significant decrease (32%) was observed in the 4% methanol-exposed group.
19 Methanol treatment at 2% and 4% was associated with significant increases in activity, in
20 the form of distance traveled in a spontaneous locomotor activity test, in the FAS group (13%
21 and 39%, respectively) and more notably, in the FAD group (33% and 66%, respectively) when
22 compared to their respective controls. Aziz et al. (2002) suggest that these alterations in
23 locomotor activity may be caused by a significant alteration in dopamine receptors and
24 disruption in neurotransmitter availability. Dopamine receptor (D2) binding in the hippocampus
25 of the FAD group was significantly increased (34%) at 1% methanol, but was significantly
26 decreased at 2% and 4% methanol exposure by 20% and 42%, respectively. In the FAS group,
27 D2 binding was significantly increased by 22% and 54% in the 2% and 4% methanol-exposed
28 groups.
29 At PND45, the CAR in FAD rats exposed to 2% and 4% methanol was significantly
30 decreased by 48% and 52%, respectively, relative to nonexposed controls. In the FAS group, the
31 CAR was only significantly decreased in the 4% methanol-exposed animals and only by 22% as
32 compared to their respective controls. Aziz et al. (2002) suggest that the impairment in CAR of
33 the methanol-exposed FAD pups may be due to alterations in the number of cholinergic
34 (muscarinic) receptor proteins in the hippocampal region of the brain. Muscarinic receptor
45 The nigrostriatal pathway is one of four major dopamine pathways in the brain that are particularly involved in the
production of movement. Loss of dopamine neurons in the substantia nigra is one of the pathological features of
Parkinson's disease (Kim et al., 2003),
4-52 DRAFT-DO NOT CITE OR QUOTE
-------
1 binding was significantly increased in the 2% (20%) and 4% (42%) methanol-exposed group in
2 FAD animals, while FAS group animals had a significant increase in cholinergic binding only in
3 the 4% methanol exposed group (21%). High concentrations of methanol may saturate the
4 body's ability to remove toxic metabolites, including formaldehyde and formate, and this may be
5 exacerbated in FAD pups having a low store of folate.
6 Immunohistochemistry showed an increase in the expression of GAP-43 protein in the
7 dentate granular and pyramidal cells of the hippocampus in 2% and 4% methanol-exposed
8 animals in the FAD group. The FAS group showed increased expression only in the 4%
9 methanol-exposed group. The Western blot analysis also confirmed a higher expression of
10 GAP-43 in the 2% and 4% methanol-exposed FAD group rats. Aziz et al. (2002) suggested that
11 up-regulation of GAP-43 in the hippocampal region may be associated with axonal growth or
12 protection of the nervous system from methanol toxicity.
13 The Aziz et al. (2002) study provides evidence that hepatic tetrahydrofolate is an
14 important contributing factor in methanol-induced developmental neurotoxicity in rodents. The
15 immature blood-brain barrier and inefficient drug-metabolizing enzyme system make the
16 developing brain a particularly sensitive target organ to the effects of methanol exposure.
4.4.2. Inhalation Studies
17 A review by Carson et al. (1981) has summarized a number of older reports of studies on
18 the toxicological consequences of methanol exposure. In one example relevant to neurotoxicity,
19 the review cites a research report of Chen-Tsi (1959) who exposed 10 albino rats/group (sex and
20 strain unstated) to 1.77 and 50 mg/m3 (1.44 and 40.7 ppm) methanol vapor, 12 hours/day, for
21 3 months. Deformation of dendrites, especially the dendrites of pyramidal cells, in the cerebral
22 cortex was included in the description of histopathological changes observed in adult animals
23 following exposure to 50 mg/m3 (40.7 ppm) methanol vapor. One out often animals exposed to
24 the lower methanol concentration also displayed this feature.
25 Information on the neurotoxicity of methanol inhalation exposure in adult monkeys
26 (M fascicularis) has come from NEDO (1987) which describes the results of a number of
27 experiments. The study included an acute study, a chronic study monkeys, and a repeated
28 exposure experiment (of variable duration depending upon exposure level), followed by recovery
29 period (1-6 months), and an experiment looking at chronic formaldehyde exposure (1 or 5 ppm),
30 a combustion product of methanol. This last experiment was apparently only a pilot and included
31 only one monkey per exposure condition.
32 In the chronic experiment 8 monkeys were included per exposure level (control, 10, 100,
33 1,000 ppm or 13, 131, and 1,310 mg/m3, respectively, for 21 hours/day); however, animals were
34 serially sacrificed at 3 time points: 7 months, 19 months, or >26 months. This design reduced
4-53 DRAFT-DO NOT CITE OR QUOTE
-------
1 the number of monkeys at each exposure level to 2 subjects at 7 months and 3 subjects at the
2 subsequent time points (see Section 4.2.2). One of the 3 animals receiving 100 ppm methanol
3 and scheduled for sacrifice at 29 months was terminated at 26 months.
4 Histopathologically, no overt degeneration of the retina, optical nerve, cerebral cortex, or
5 other potential target organs (liver and kidney) was reported in the chronic experiment.
6 Regarding the peripheral nervous system, 1/3 monkeys exposed to 100 ppm (131 mg/m3) and 2/3
7 exposed to 1,000 ppm (1,310 mg/m3) for 29 months showed slight but clear changes in the
8 peroneal nerves. There was limited evidence of CNS degeneration inside the nucleus of the
9 thalamus of the brain at exposure to 100 ppm (131 mg/m3) or 1,000 ppm (1,310 mg/m3) for
10 7 months or longer. Abnormal appearance of stellate cells (presumed astroglia) within the
11 cerebral white matter was also observed in a high proportion (7/8 in both mid- and high-exposure
12 groups) of monkeys exposed to 100 ppm and 1,000 ppm for 7 months or more. All monkeys that
13 had degeneration of the inside nucleus of the thalamus also had degeneration of the cerebral
14 white matter. According to NEDO (1987), the stellate cell response was transient and "not
15 characteristic of degeneration." The authors also noted that the stellate cell response was "nearly
16 absent in normal monkeys in the control group" and "in the groups exposed to a large quantity of
17 methanol or for a long time their presence tended to become permanent, so a relation to the long
18 term over which the methanol was inhaled is suspected." However, all control group responses
19 are reported in a single table in the section of the NEDO (1987) report that describes the acute
20 monkey study, with no indication as to when the control group was sacrificed.
21 In the recovery experiment, monkeys were exposed to 1,000, 2,000, 3,000, or 5,000 ppm
22 methanol, followed by recovery periods of various duration. Monkeys exposed to 3,000 ppm for
23 20 days followed by a 6-month recovery period experienced relatively severe fibrosis of
24 responsive stellate cells and lucidation of the medullary sheath. However, resolution of some of
25 the glial responses was noted in the longer duration at lower exposure levels, with no effects
26 observed on the cerebral white matter in monkeys exposed for 7 months to 1,000 ppm methanol
27 followed by a 6-month recovery period. In general, the results from the recovery experiment
28 corroborated results observed in the chronic experiment. NEDO (1987) interpreted the lack of
29 glial effects after a 6-month recovery as an indication of a transient effect. The authors failed to
30 recognize that glial responses to neural damage do not necessarily persist following resolution of
31 neurodegeneration (Aschner and Kimmelberg, 1996).
32 The limited information available from the NEDO (1987) summary report suggests that
33 100 ppm (131 mg/m3) may be an effect level following continuous, chronic exposure to
34 methanol. However, the current report does not indicate a robust dose response for the
35 neurodegenerative changes in the thalamus and glial changes in the white matter. The number of
36 animals at each exposure level for each serial sacrifice also limits statistical power
37 (2-3 monkeys/time point/exposure level). Confidence in this study is also weakened by the lack
38 of documentation for a concurrent control group.
4-54 DRAFT-DO NOT CITE OR QUOTE
-------
1 Weiss et al. (1996) exposed 4 cohorts of pregnant Long-Evans rats (10-12 dams/
2 treatment group/cohort) to 0 or 4,500 ppm (0 and 5,897 mg/m3) methanol vapor (high-
3 performance liquid chromatography [HPLC] grade), 6 hours/day, from GD6 to PND21. Pups
4 were exposed together with the dams during the postnatal period. Average blood methanol levels
5 in pups on PND7 and PND14 were about twice the level observed in dams. However, methanol
6 exposure had no effect on maternal gestational weight gain, litter size, or postnatal pup weight
7 gain up to PND1846. Neurobehavioral tests were conducted in neonatal and adult offspring; the
8 data generated from those tests were evaluated by repeated measures ANOVA. Three
9 neurobehavioral tests conducted in 13-26 neonates/group included a suckling test, conditioned
10 olfactory aversion test, and motor activity test. In contrast to earlier test results reported by
11 Infurna and Weiss (1986), methanol exposure had no effect on suckling and olfactory aversion
12 tests conducted on PND5 and PND10, respectively. Results of motor activity tests in the
13 methanol group were inconsistent, with decreased activity on PND18 and increased activity on
14 PND25. Tests that measured motor function, operant behavior, and cognitive function were
15 conducted in 8-13 adult offspring/group. Some small performance differences were observed
16 between control and treated adult rats in the fixed wheel running test only when findings were
17 evaluated separately by sex and cohort. The test requires the adult rats to run in a wheel and
18 rotate it a certain amount of times in order to receive a food reward. A stochastic spatial
19 discrimination test examined the rats' ability to learn patterns of sequential responses. Methanol
20 exposure had no effect on their ability to learn the first pattern of sequential responses, but
21 methanol-treated rats did not perform as well on the reversal test. The result indicated possible
22 subtle cognitive deficits as a result of methanol exposure. A morphological examination of
23 offspring brains conducted on PND1 and PND21 indicated that methanol exposure had no effect
24 on neuronal migration, numbers of apoptotic cells in the cortex or germinal zones, or
25 myelination. However, neural cell adhesion molecule (NCAM) 140 and NCAM 180 gene
26 expression in treated rats was reduced on PND4 but not 15 months after the last exposure.
27 NCAMs are glycoproteins required for neuron migration, axonal outgrowth, and establishing
28 mature neuronal function patterns.
29 Stanton et al. (1995) exposed 6-7 pregnant female Long-Evans rats/group to 0 or
30 15,000 ppm (0 and 19,656 mg/m3) methanol vapors (> 99.9% purity) for 7 hours/day on GD7-
31 GD19. Mean serum methanol levels at the end of the 1st, 4th, 8th, and 12th days of exposure
32 were, 3,836, 3,764, 3,563, and 3,169 |ig/mL, respectively. As calculated by authors, dams
33 received an estimated methanol dose of 6,100 mg/kg-day. A lower body weight on the first
46 The fact that this level of exposure caused effects in the Sprague-Dawley rats of the NEDO (1987) study but did
not cause a readily apparent maternal effect in Long-Evans rats of this study could be due to diffences in strain
susceptibility.
4-55 DRAFT-DO NOT CITE OR QUOTE
-------
1 2 days of exposure was the only maternal effect; there was no increase in postimplantation loss.
2 Dams were allowed to deliver and nurse litters. Parameters evaluated in pups included mortality,
3 growth, pubertal development, and neurobehavioral function. Examinations of pups revealed
4 that two pups from the same methanol-exposed litter were missing one eye; aberrant visually
5 evoked potentials were observed in those pups. A modest but significant reduction in body
6 weight gain on PND1, PND21, and PND35 was noted in pups from the methanol group. For
7 example, by PND35, male pups of dams exposed to methanol had a mean body weight of
8 129 grams versus 139 grams in controls (p < 0.01). However, postnatal mortality was unaffected
9 by exposure to methanol. The study authors did not consider a 1.7-day delay in vaginal opening
10 in the methanol group to be an adverse effect. Preputial separation was not affected by prenatal
11 methanol exposure. Neurobehavioral status was evaluated using 8 different tests on specific
12 days up to PND160. Tests included motor activity on PND13-PND21, PND30, and PND60,
13 olfactory learning and retention on PND18 and PND25, behavioral thermoregulation on
14 PND20-21, T-maze delayed alternation learning on PND23-PND24, acoustic startle reflex on
15 PND24, reflex modification audiometry on PND61-PND63, passive avoidance on PND73, and
16 visual evoked potentials on PND160. A single pup/sex/litter was examined in most tests, and
17 some animals were subjected to multiple tests. The statistical significance of neurobehavioral
18 testing was assessed by one-way ANOVA, using the litter as the statistical unit. Results of the
19 neurobehavioral testing indicated that methanol exposure had no effect on the sensory, motor, or
20 cognitive function of offspring under the conditions of the experiment. However, given the
21 comparatively small number of animals tested for each response, it is uncertain whether the
22 statistical design had sufficient power to detect small compound-related changes.
23 NEDO (1987, unpublished report) sponsored a teratology study that included an
24 evaluation of postnatal effects in addition to standard prenatal endpoints in Sprague-Dawley rats.
25 Thirty-six pregnant females/group were exposed to 0, 200, 1,000, or 5,000 ppm (0, 262, 1,310,
26 and 6,552 mg/m3) methanol vapors (reagent grade) on GD7-GD17 for 22.7 hours/day. Statistical
27 significance of results was evaluated by t-test, Mann-Whitney U test, Fisher's exact test, and/or
28 Armitage's ^ test.
29 Postnatal effects of methanol inhalation were evaluated in the remaining 12 dams/group
30 that were permitted to deliver and nurse their litters. Effects were only observed in the
31 5,000 ppm. There were no adverse effects on offspring body weight from methanol exposure.
32 However, the weights of some organs (brain, thyroid, thymus, and testes) were reduced in
33 8-week-old offspring following prenatal-only exposure to 5,000 ppm methanol. An unspecified
34 number of offspring were subjected to neurobehavioral testing or necropsy, but results were
35 incompletely reported.
4-56 DRAFT-DO NOT CITE OR QUOTE
-------
1 NEDO (1987, unpublished report) also contains an account of a two-generation
2 reproductive study that evaluated the effects of pre- and postnatal methanol (reagent grade)
3 exposure (20 hours/day) on reproductive and other organ systems of Sprague-Dawley rats and in
4 particular the brain. The FO generation (30 males and 30 females per exposure group)47 was
5 exposed to 0, 10, 100, and 1,000 ppm (0, 13.1, 131, and 1,310 mg/m3) from 8 weeks old to the
6 end of mating (males) or to the end of lactation period (females). The FI generation was exposed
7 to the same concentrations from birth to the end of mating (males) or to weaning of 72 pups
8 21 days after delivery (females). Males and females of the F2 generation were exposed from
9 birth to 21 days old (1 animal/sex/litter was exposed to 8 weeks of age). NEDO (1987) noted
10 reduced brain, pituitary, and thymus weights, in the offspring of FO and FI rats exposed to
11 1,000 ppm methanol. As discussed in the report, NEDO (1987) sought to confirm the possible
12 compound-related effect of methanol on the brain by carrying out an additional study in which
13 Sprague-Dawley rats were exposed to 0, 500, 1,000, and 2,000 ppm (0, 655, 1,310, and
14 2,620 mg/m3) methanol from the first day of gestation through the FI generation. Brain weights
15 were measured in 10-14 offspring/sex/group at 3, 6, and 8 weeks of age. As illustrated in
16 Table 4-14, brain weights were significantly reduced in 3-week-old males and females exposed
17 to > 1,000 ppm. At 6 and 8 weeks of age, brain weights were significantly reduced in males
18 exposed to > 1,000 ppm and females exposed to 2,000 ppm. Due to the toxicological
19 significance of this postnatal effect and the fact that it has not been measured or reported
20 elsewhere in the peer-reviewed methanol literature, the brain weight changes observed by NEDO
21 (1987) following gestational and postnatal exposures and following gestation-only exposure (in
22 the teratology study discussed above) are evaluated quantitatively and discussed in more detail in
23 Section 5 of this review.
47 A second control group of 30 animals/sex was maintained in a separate room to "confirm that environmental conditions inside
the chambers were not unacceptable to the animals."
4-57 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-14. Brain weights of rats exposed to methanol vapors during gestation and
lactation
Offspring
age
3wk
6 wk
8wka
8wkb
Sex
Male
Female
Male
Female
Male
Female
Male
Female
Brain weight (g) (% control) at each exposure level
0 ppm
1.45 ±0.06
1.41 ±0.06
1.78 ±0.07
1.68 ±0.08
1.99 ±0.06
1.85 ±0.05
2.00 ±0.05
1.86 ±0.08
200 ppm
-
-
-
2.01 ±0.08
(100%)
1.91 ±0.06
(103%)
500 ppm
1.46 ±0.08
(101%)
1.41 ±0.07
(100%)
1.74 ±0.09
(98%)
1.71 ±0.08
(102%)
1.98 ±0.09
(99%)
1.83 ±0.07
(99%)
-
1,000 ppm
1.39±0.05C
(96%)
1.33±0.07d
(94%)
1.69±0.06d
(95%)
1.62 ±0.07
(96%)
1.88±0.08d
(94%)
1.80 ±0.08
(97%)
1.99 ±0.07
(100%)
1.90 ±0.08
(102%)
2,000 ppm
1.27±0.06e
(88%)
1.26±0.09e
(89%)
1.52±0.07e
(85%)
1.55±0.05e
(92%)
1.74±0.05e
(87%)
1.67±0.06e
(90%)
-
5,000 ppm
-
-
-
1.81±0.16d
(91%)
1.76 ±1.09
(95%)
aExposed throughout gestation and FI generation.
''Exposed on gestational days 7-17 only.
cp < 0.05, dp < 0.01, ep < 0.001, as calculated by the authors.
Values are means ± S.D.
Source: NEDO(1987).
1 Burbacher et al. (1999a,1999b) carried out toxicokinetic, reproductive, developmental
2 and postnatal neurological and neurobehavioral studies of methanol in M. fascicularis monkeys
3 that were published by HEI in a two-part monograph. Some of the data were subsequently
4 published in the open scientific literature (Burbacher et al., 2004a, 2004b). The experimental
5 protocol featured exposure to 2 cohorts of 12 monkeys/group to low-exposure levels (relative to
6 the previously discussed rodent studies) of 0, 200, 600, or 1,800 ppm (0, 262, 786, and 2,359
7 mg/m3) methanol vapors (99.9% purity), 2.5 hours/day, 7 days/week, during a premating period
8 and mating period (-180 days combined) and throughout the entire gestation period (-168 days).
9 The monkeys were 5.5-13 years old and were a mixture of feral-born and colony-bred animals.
10 The outcome study included an evaluation of maternal reproductive performance (discussed in
11 Section 4.3.2) and tests to assess infant postnatal growth and newborn health, neurological
12 outcomes included reflexes, behavior, and development of visual, sensorimotor, cognitive, and
13 social behavioral function. Blood methanol levels, clearance, and the appearance of formate
14 were also examined and are discussed in Section 3.2. The effects observed were in the absence of
15 appreciable increases in maternal blood formate levels.
4-58
DRAFT-DO NOT CITE OR QUOTE
-------
1 Neurobehavioral function was assessed in 8-9 infants/group during the first 9 months of
2 life (Burbacher et al., 2004a, 1999b). Although results in 7/9 tests were negative, 2 effects were
3 possibly related to methanol exposure. The Visually Directed Reaching (VDR) test is a measure
4 of sensorimotor development and assessed the infants' ability to grasp for a brightly colored
5 object containing an applesauce-covered nipple. Beginning at 2 weeks after birth, infants were
6 tested 5 times/day, 4 days/week. Performance on this test, measured as age from birth at
7 achievement of test criterion (successful object retrieval on 8/10 consecutive trials over 2 testing
8 sessions), was reduced in all treated male infants. The times (days after birth) to achieve the
9 criteria for the VDR test were 23.7 ±4.8 (n = 3), 32.4 ± 4.1 (n = 5), 42.7 ± 8.0 (n = 3), and 40.5 ±
10 12.5 (n = 2) days for males and 34.2 ± 1.8 (n = 5), 33.0 ± 2.9 (n = 4), 27.6 ± 2.7 (n = 5), and 40.0
11 ± 4.0 (n = 7) days for females in the control to 1,800 ppm groups, respectively. Statistical
12 significance was obtained in the 1,800 ppm group when males and females were evaluated
13 together (p = 0.04) and in the 600 ppm (p = 0.007) for males only. However, there was no
14 significant difference between responses and/or variances among the dose levels for males and
15 females combined (p = 0.244), for males only (p = 0.321) and for males only, excluding the high-
16 dose group (p = 0.182). Yet there was a significant dose-response trend for females only
17 (p = 0.0265). The extent to which VDR delays were due to a direct effect of methanol on
18 neurological development or a secondary effect due to the methanol-induced decrease in length
19 of pregnancy and subsequent prematurity is not clear. Studies of reaching behavior have shown
20 that early motor development in pre-term human infants without major developmental disorders
21 differs from that of full-term infants (Fallang et al., 2003). Clinical studies have indicated that
22 the quality of reaching and grasping behavior in pre-term infants is generally less than that in
23 full-term infants (Fallang et al., 2003; Plantinga et al., 1997). For this reason, measures of
24 human infant development generally involve adjustment of a child's "test age" if he or she had a
25 gestational age of fewer than 38 weeks, often by subtracting weeks premature from the age
26 measured from birth (Wilson and Cradock, 2004). When this type of adjustment is made to the
27 Burbacher et al. (2004a, 1999b) VDR data, the dose-response trend for males only becomes
28 worse (p = 0.448) and the dose-response trend for the females only is improved (p = 0.009),
29 though the variance in the data could not be modeled adequately. Thus, only the unadjusted
30 VDR response for females only exhibited a dose response that could be adequately modeled for
31 the purposes of this assessment (see Appendix C).
32 At 190-210 days of age, the Fagan Test of infant intelligence was conducted. The
33 paradigm makes use of the infant's proclivity to direct more visual attention to novel stimuli
34 rather than familiar stimuli. The test measures the time infants spend looking at familiar versus
35 novel items. Deficits in the Fagan task can qualitatively predict deficits in intelligence quotient
36 (IQ) measurements assessed in children at later ages (Fagan and Singer, 1983). Control monkey
4-59 DRAFT-DO NOT CITE OR QUOTE
-------
1 infants in the Burbacher et al. (2004a, 1999b) study spent more than 62% ± 4% (mean for both
2 cohorts) of their time looking at novel versus familiar monkey faces, while none of the treated
3 monkeys displayed a preference for the novel faces (59% ± 2%, 54% ± 2% and 59% ± 2% in
4 200, 600 and 1,800 ppm groups, respectively). Unlike the VDR results discussed previously,
5 results of this test did not appear to be gender specific and were neither statistically significant
6 (ANOVA p = 0.38) nor related to exposure concentration. The findings indicated a cohort effect
7 which appeared to reduce the statistical power of this analysis. The authors' exploratory analysis
8 of differences in outcomes between the 2 cohorts indicated an effect of exposure in the second
9 cohort and not the first cohort due to higher mean performance in controls of cohort 2 (70% +
10 5% versus 55% ± 4% for cohort 1). In addition, this latter finding could reflect the inherent
11 constraints of this endpoint. If the control group performs at the 60% level and the most
12 impaired subjects perform at approximately the 50% chance level (worse than chance
13 performance would not be expected), the range over which a concentration-response relationship
14 can be expressed is limited. Because of the longer latency between assessment and birth, these
15 results would not be confounded with the postulated methanol-induced decrease in gestation
16 length of the exposed groups of this study. Negative results were obtained for the remaining
17 seven tests that evaluated early reflexes, gross motor development, spatial and concept learning
18 and memory, and social behavior. Infant growth and tooth eruption were unaffected by methanol
19 exposure.
4.4.3. Studies Employing In Vitro, S.C. and I.P. Exposures
20 There is some experimental evidence that the presence of methanol can affect the activity
21 of acetylcholinesterase (Tsakiris, et al., 2006). Although these experiments were carried out on
22 erythrocyte membranes in vitro, the apparent compound-related changes may have implications
23 for possible impacts of methanol and/or its metabolites on acetylcholinesterase at other centers,
24 such as the brain. Tsakiris et al. (2006) prepared erythrocyte ghosts from blood samples of
25 healthy human volunteers by repeated freezing-thawing. The ghosts were incubated for 1 hour at
26 37°C in 0, 0.07, 0.14, 0.6 or 0.8 mmol/L methanol and the specific activities of
27 acetylcholinesterase monitored. Respective values (in change of optical density units/minute-mg
28 protein) were 3.11 ± 0.15, 2.90 ± 0.10, 2.41 ± 0.10 (p < 0.05), 2.05 ± 0.11 (p < 0.01), and 1.81 ±
29 0.09 (p < 0.001). More recently, Simintzi et al. (2007) carried out an in vitro experiment to
30 investigate the effects of aspartame metabolites, including methanol, on 1) a pure preparation of
31 acetylcholinesterase, and 2) the same activity in homogenates of frontal cortex prepared from the
32 brains of (both sexes of) Wistar rats. The activities were measured after incubations with 0, 0.14,
33 0.60, or 0.8 mmoles/L (0, 4.5, 19.2, and 25.6 mg/L) methanol, and with methanol mixed with the
34 other components of aspartame metabolism, phenylalanine and aspartic acid. After incubation at
4-60 DRAFT-DO NOT CITE OR QUOTE
-------
1 37°C for 1 hour, the activity of acetylcholinesterase was measured spectrophotometrically. As
2 shown in Table 4-15, the activities of the acetylcholinesterase preparations were reduced dose
3 dependency after incubation in methanol. Similar results were also obtained with the other
4 aspartame metabolites, aspartic acid, and phenylalanine, both individually or as a mixture with
5 methanol. While the implications of this result to the acute neurotoxicity of methanol are
6 uncertain, the authors speculated that methanol may bring about these changes through either
7 interactions with the lipids of rat frontal cortex or perturbation of proteinaceous components.
Table 4-15. Effect of methanol on Wistar rat acetylcholinesterase activities
Methanol concentration
(mmol/L)
Control
0.14
0.60
0.80
Acetylcholinesterase activity (AOD/min-mg)
Frontal cortex
0.269 ±0.010
0.234 ±0.007a
0.223 ± 0.009b
0.204 ± 0.008b
Pure enzyme
1.23 ±0.04
1.18 ±0.06
1.05±0.04b
0.98±0.05b
Values are means ± S.D. for four experiments. The average value of each experiment was derived from three
determinations of each enzyme activity.
><0.01.
V< 0.001.
Source: Simintzi et al. (2007).
8 In another experiment of relevance to neurotoxicity, the impact of repeat methanol
9 exposure on amino acid and neurotransmitter expression in the retina, optic nerve, and brain was
10 examined by Gonzalez-Quevedo et al. (2002). The goal of the study was to determine whether a
11 sustained increase in formate levels, at concentrations below those known to produce toxic
12 effects from acute exposures, can induce biochemical changes in the retina, optical nerve, or
13 certain regions of the brain. Male Sprague-Dawley rats (5-7/group; 100-150 g) were divided
14 into 6 groups and treated for 4 weeks according to the following plan. Four groups of animals
15 received tap water ad libitum as drinking water for 1 week. During the second week, groups 1
16 and 2 (control and methanol respectively) received saline subcutaneously, (s.c.) and groups 3 and
17 4 (methotrexate48 [MTX] and methotrexate-methanol [MTX-methanol], respectively) received
18 MTX s.c. (0.2 mg/kg-day). During the 3rd week, MTX was reduced to 0.1 mg/kg and 20%
19 methanol (2g/kg-day) was given i.p. to groups 2 (methanol) and 4 (MTX-methanol). Groups 1
20 (control) and 3 (MTX) received equivalent volumes of saline administered i.p. The treatment
21 was continued until the end of the fourth week. Groups 5, (taurine49 [Tau]) and 6, (Tau-MTX-
48 Methotrexate depletes folate stores (resulting in an increase in the formate levels of methanol exposed animals) by
interfering with tetrahydrofolate(THF) regeneration (Dorman et al., 1994).
49 Taurine plays and important role in the CNS, especially in the retina and optical nerve, and was administered here
to explore its possible protective effect (Gonzalez-Quevedo et al., 2002)
4-61 DRAFT-DO NOT CITE OR QUOTE
-------
1 methanol) received 2% Tau in their drinking water ad libitum during the first 4 weeks, after
2 which they were treated in the same manner as groups 1 and 4, respectively. Weights were
3 documented weekly on all animals. Blood for formate and amino acid determinations and
4 biopsy samples of retina, optic nerve, hippocampus, and posterior cortex of each animal were
5 collected at the end of the experiment. Formate levels were not affected by Tau alone or MTX
6 alone. While methanol alone increased blood formate levels, MTX-methanol, and Tau-MTX-
7 methanol produced a threefold increase in blood formate levels as compared to controls and a
8 twofold increase as compared to methanol alone. The amino acids aspartate, glutamate,
9 asparagine, serine, histidine, glutamine, threonine, glycine, arginine, alanine, hypotaurine,
10 gamma-aminobutyric acid (which is also a neurotransmitter), and tyrosine were measured in
11 blood, brain, and retinal regions.
12 None of the amino acids measured were altered in the blood of methanol-, MTX-, or
13 MTX-methanol-treated animals. Tau was increased in the blood of animals treated with taurine
14 in the drinking water (Tau and Tau-MTX-methanol) and histidine was increased in the Tau group
15 but not in the Tau-MTX-methanol group.
16 The levels of aspartate, Tau, glutamine, and glutamate were found to be altered by
17 treatment in various areas of the brain. Aspartate was increased in the optic nerve of animals
18 treated with MTX-methanol and Tau-MTX-methanol, indicating a possible relation to formate
19 accumulation. The authors note that L-aspartate is a major excitatory amino acid in the brain and
20 that increased levels of excitatory amino acids can trigger neuronal cell damage and death (Albin
21 and Greenamyre, 1992). Aspartate, glutamine and Tau were found to be increased with respect
22 to controls in the hippocampus of the three groups receiving methanol. Glutamate was
23 significantly increased in the hippocampus in the methanol and the Tau-MTX-methanol groups
24 with respect to controls, but no statistically significant difference was found in the MTX-
25 methanol group when compared to controls, methanol alone, or the Tau-MTX-methanol groups.
26 The authors suggest that increased levels of aspartate and glutamine in the hippocampus could
27 provide an explanation for some of the CNS symptoms observed in methanol poisonings on the
28 basis of their observed impact on cerebral arteries (Huang et al., 1994). The fact that these
29 increases resulted primarily from methanol without MTX is significant in that it indicates
30 methanol can cause excitotoxic effects without formate mediation. The treatments used did not
31 produce any significant changes in amino acid levels in the posterior cortex.
32 The neurotransmitters serotonin (5-HT) and dopamine (DA) and their respective
33 metabolites, 5-hydroxyindolacetic acid (5-HIAA) and dihydroxyphenylacetic acid (DOPAC),
34 were measured in the brain regions described. The levels of these monoamines were not affected
35 by formate accumulation, as the only increases were observed for 5-HT and 5-HIAA following
36 methanol-only exposure. 5-HT was increased in the retina and hippocampus of methanol-only
4-62 DRAFT-DO NOT CITE OR QUOTE
-------
1 treated animals, and the metabolite 5-HIAA was increased in the hippocampus of methanol-only
2 treated animals; DA and DOPAC levels were not altered by the treatments in any of the areas
3 measured. The posterior cortex did not show any changes in monoamine levels for any treatment
4 group.
5 Rajamani et al. (2006) examined several oxidative stress parameters in male Wistar rats
6 following methotrexate-induced folate deficiency. Animals (6/group) were divided into 3e
7 groups: saline controls, methotrexate (MTX) controls, and MTX-methanol treated animals.
8 Animals in the MTX-only group were treated with 0.2 mg/kg-day MTX s.c. injection for 7 days
9 and following confirmation of folate deficiency, received either saline for MTX control and
10 saline controls or a single dose of 3 g/kg methanol (20% w/v in saline) i.p. on day 8. On the 9th
11 day, all animals were sacrificed and blood and tissue samples were collected. The optic nerve,
12 retina, and brain were collected and the brain was dissected into the following regions: cerebral
13 cortex, cerebellum, mid-brain, pons medulla, hippocampus and hypothalamus. Each region was
14 homogenized, then centrifuged at 300 x g for 15 minutes and the supernatant was examined for
15 indicators of oxidative stress including the free radical scavengers superoxide dismutase (SOD),
16 CAT, glutathione peroxidase, and reduced GSH levels. The levels of protein thiols, protein
17 carbonyls, and amount of lipid peroxidation were also measured. Compared to controls the
18 levels of SOD, CAT, GSH peroxidase, oxidized GSH, protein carbonyls and lipid peroxidation
19 were elevated in all of the brain regions where it was measured, with greater increases observed
20 in the MTX-methanol treated animals than in the MTX alone group. The level of GSH and
21 protein thiols was decreased in all regions of the brain, with a greater decrease observed in the
22 MTX-methanol-treated animals than MTX-treated animals. In addition, expression of HSP70, a
23 biomarker of cellular stress, was increased in the hippocampus. Overall, these results suggest that
24 methanol treatment of folate-deficient rats results in increased oxidative stress in the brain, retina
25 and optic nerve.
26 To determine the effects of methanol intoxication on the HPA axis, a combination of
27 oxidative stress, immune and neurobehavioral parameters were observed (Parthasarathy et al.,
28 2006a). Adult male Wistar albino rats (6 animals/group) were treated with either 0 or 2.37g/kg-
29 day methanol i.p. for 1, 15 or 30 days. Oxidative stress parameters examined included SOD,
30 CAT, GSH peroxidase, GSH, and ascorbic acid (Vitamin C). Plasma corticosterone levels were
31 measured, and lipid peroxidation was measured in the hypothalamus and the adrenal gland. An
32 assay for DNA fragmentation was conducted in tissue from the hypothalamus, the adrenal gland
33 and the spleen. Immune function tests conducted included the footpad thickness test for delayed
34 type hypersensitivity (DTH), a leukocyte migration inhibition assay, the hemagglutination assay
35 (measuring antibody titer), the neutrophil adherence test, phagocytosis index, and a nitroblue
36 tetrazolium (NET) reduction and adherence assay used to measure the killing ability of
4-63 DRAFT-DO NOT CITE OR QUOTE
-------
1 polymorphonuclear leukocytes (PMNs). The open field behavior test was used to measure
2 general locomotor and explorative activity during methanol treatment in the 30-day treatment
3 group, with tests conducted on days 1, 4, 8, 12, 16, 20, 24, and 28. All enzymatic (SOD, CAT,
4 and GSH peroxidase) and nonenzymatic antioxidants (GSH and Vitamin C) were significantly
5 increased in the 1-day methanol-exposed group as compared to controls. However, with
6 increasing time of treatment, all of the measured parameters were significantly decreased when
7 compared with control animals. Lipid peroxidation was significantly increased in both the
8 hypothalamus and the adrenal gland at 1, 15, and 30 days, with the 30-day treated animals also
9 significantly increased when compared to the 15-day methanol-treated animals.
10 Leukocyte migration and antibody titer were both significantly increased over controls
11 for all time points, while footpad thickness was significantly decreased in 15- and 30-day treated
12 animals. Neutrophil adherence was significantly decreased after 1 and 30 days of exposure. A
13 significant decrease in the NET reduction and adherence was found when comparing PMNs from
14 the 30-day treated animals with cells from the 15-day methanol-treated group.
15 The open field behavior tests showed a significant decrease in ambulation from the 4th
16 day on and significant decreases in rearing and grooming from the 20th day on. A significant
17 increase was observed in immobilization from the 8th day on and in fecal bolus from the 24th
18 day on in methanol-exposed animals.
19 While corticosterone levels were significantly increased following 1 or 15 days of
20 methanol treatment, they were significantly decreased after 30 days of treatment, as compared to
21 controls. Following 30 days of methanol treatment, DNAfrom the hypothalamus, the adrenal
22 gland, and the spleen showed significant fragmentation. The authors conclude that exposure to
23 methanol-induced oxidative stress, disturbs HPA-axis function, altering corticosterone levels and
24 producing effects in several nonspecific and specific immune responses.
4.5. IMMUNOTOXICITY
25 Parthasarathy et al. (2005a) provided data on the impact of methanol on neutrophil
26 function in an experiment in which 6 male Wistar rats/group were given a single i.p. exposure of
27 2,370 mg/kg methanol mixed 1:1 in saline. Another group of 6 animals provided blood samples
28 that were incubated with methanol in vitro at a methanol concentration equal to that observed in
29 the in vivo-treated animals 30 and 60 minutes postexposure. Total and differential leukocyte
30 counts were measured from these groups in comparison to in vivo and in vitro controls.
31 Neutrophil adhesion was determined by comparing the neutrophil index in the untreated blood
32 samples to those that had been passed down a nylon fiber column. The cells' phagocytic ability
33 was evaluated by their ability to take up heat-killed Candida albicans. In another experiment,
34 neutrophils were assessed for their killing potential by measuring their ability to take up then
4-64 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
convert NET to formazan crystals.50 One hundred neutrophils/slides were counted for their total
and relative percent formazan-positive cells.
The blood methanol concentrations 30 and 60 minutes after dosing were 2,356 ±162 and
2,233 ±146 mg/L, respectively. The mean of these values was taken as the target concentration
for the in vitro methanol incubation. In the in vitro studies, there were no differences in total and
differential leukocyte counts, suggesting that no lysis of the cells had occurred at this methanol
concentration. This finding contrasts with the marked difference in total leukocytes observed as
a result of methanol incubation in vivo, in which, at 60 minutes after exposure, 16,000 ± 1,516
cells/mm3 were observed versus 23,350 ± 941 in controls (p < 0.001). Some differences in
neutrophil function were observed in blood samples treated with methanol in vitro and in vivo.
These differences are illustrated for the 60-minute postexposure samples in Table 4-16.
Table 4-16. Effect of methanol on neutrophil functions in in vitro and in vivo studies in
male Wistar rats
Parameter
Phagocytic index (%)
Avidity index
NET reduction (%)
Adherence (%)
In vitro studies (60 minutes)
Control
89.8 ±3.07
4.53 ±0.6
31.6 ±4.6
50.2 ±5.1
Methanol
81.6±2.2a
4.47 ±0.7
48.6±4.3b
39.8±2.4a
In vivo studies (60 minutes)
Control
66.0 ±4.8
2.4 ±0.1
4.6 ±1.2
49.0 ±4.8
Methanol
84.0 ± 7.0b
3.4±0.3a
27.0 ± 4.6b
34.6±4.0b
12
13
14
15
16
17
18
19
20
21
22
23
Values are means ± S.D. for six animals.
><0.01.
V< 0.001.
Source: Parthasarathy et al. (2005a).
Parthasarathy et al. (2005a) observed differences in the neutrophil functions of cells
exposed to methanol in vitro versus in vivo, most notably in the phagocytic index that was
reduced in vitro but significantly increased in vivo. However, functions such as adherence and
NET reduction showed consistency in the in vitro and in vivo responses. The authors noted that,
by and large, the in vivo effects of methanol on neutrophil function were more marked than those
in cells exposed in vitro.
Another study by Parthasarathy et al. (2005b) also exposed 6 male Wistar rats/group i.p.
to methanol at approximately 1/4 the LD50 (2.4 g/kg). The goal was to further monitor possible
methanol-induced alterations in the activity of isolated neutrophils and other immunological
parameters. The exposure protocol featured daily injections of methanol for up to 30 days in the
presence or absence of sheep RBCs. Blood samples were assessed for total and differential
leukocytes, and isolated neutrophils were monitored for changes in phagocytic and avidity
50 Absence of NET reduction indicates a defect in some of the metabolic pathways involved in intracellular
microbial killing.
4-65
DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
indices, NET reduction, and adherence. In the latter test, blood samples were incubated on a
nylon fiber column, then eluted from the column and rechecked for total and differential
leukocytes. Phagocytosis was monitored by incubating isolated buffy coats from the blood
samples with heat-killed C. albicam. NET reduction capacity examined the conversion of the
dye to formazan crystals within the cytoplasm. The relative percentage of formazan-positive
cells in each blood specimen gave a measure of methanol's capacity to bring about cell death.
As tabulated by the authors, there was a dose-dependent reduction in lymphoid organ weights
(spleen, thymus, and lymph node) in rats exposed to methanol for 15 and 30 days via i.p.
injection, irrespective of the presence of sheep RBCs. Methanol also appeared to result in a
reduction in the total or differential neutrophil count. These and potentially related changes to
neutrophil function are shown in Table 4-17.
Table 4-17. Effect of intraperitoneally injected methanol on total and differential
leukocyte counts and neutrophil function tests in male Wistar rats
Parameter
Without sheep red blood cell treatment
Control
15-day
methanol
30-day
methanol
With sheep red blood cell treatment
Control
15-day
methanol
30-day
methanol
Organ weights (mg)
Spleen
Thymus
Lymph node
1223 ± 54
232 ±12
32 ±2
910±63a
171 ±7a
24±3a
696 ± 83a'b
121 ± 10a'b
16 ± 2a'b
1381 ±27
260 ±9
39 ±2
1032 ±39a
172 ± 10a
28±la
839 ± 35a'b
130 ± 24a'b
23 ± la'b
Leukocyte counts
Total leukocytes
% neutrophils
% Lymphocytes
23,367 ± 946
24 ±8
71±7
16,592 ±1219a
21 ±3
76 ±3
13,283 ±
2553a'b
16±3a
79 ±5
18,633 ± 2057
8±3
89 ±4
16,675 ± 1908
23±4a
78.5 ±4a
14,067 ± 930a'b
15 ± 5a'b
82 ±6
Neutrophil function tests
Phagocytic index
(%)
Avidity index
NET reduction (%)
Adherence (%)
91.0 ±2.0
2.6 ±0.3
6.3 ±2.0
49.0 ±5.0
80.0 ± 4.0a
3.2±0.5a
18.2±2.0a
44.0 ±5.0
79.0±2.0a
3.2±0.1a
15.0±1.0a'b
29.5 ± 5.0a'b
87.0 ±4.0
4.1±0.1
32.0 ±3.3
78.0 ±9.2
68.0±3.0a
2.6±0.3a
22.0±3.0a
52.0±9.0a
63.0±4.0a
2.1±0.3a
19.0±2.4a
30.0±4.3a'b
Values are means ± S.D. (n = 6).
ap < 0.05 from respective control.
bp < 0.05 between 15-and 30-day treatment groups.
Source: Parthasarathy et al. (2005b).
12 The study provided data that showed altered neutrophil functions following repeated
13 daily exposures of rats to methanol for periods up to 30 days. This finding is indicative of a
14 possible effect of methanol on the immunocompetence of an exposed host.
4-66
DRAFT-DO NOT CITE OR QUOTE
-------
1 Parthasarathy et al. (2006a) reported on additional immune system indicators as part of a
2 study to determine the effects of methanol intoxication on the HPA axis. As described in
3 Section 4.4.3. immune function tests conducted included the footpad thickness test for DTK, a
4 leukocyte migration inhibition assay, the hemagglutination assay (measuring antibody titer), the
5 neutrophil adherence test, phagocytosis index, and a NET reduction and adherence assay used to
6 measure the killing ability of PMNs.
7 Leukocyte migration and antibody titer were both significantly increased over controls
8 for all time points, while footpad thickness was significantly deceased in 15- and 30-day treated
9 animals. Neutrophil adherence was significantly decreased after 1 and 30 days of exposure. A
10 significant decrease in the NET reduction and adherence was found when comparing PMNs from
11 the 30-day treated animals with cells from the 15-day methanol-treated group.
12 Parthasarathy et al. (2007) reported the effects of methanol on a number of specific
13 immune functions. As before, 6 male Wistar rats/group were treated with 2,370 mg/kg methanol
14 in a 1:1 mixture in saline administered intraperitoneally for 15 or 30 days. Animals
15 scheduled/designated for termination on day 15 were immunized intraperitoneally with 5 x 109
16 sheep RBCs on the 10th day. Animals scheduled for day 30 termination were immunized on the
17 25th day. Controls were animals that were not exposed to methanol but immunized with sheep
18 RBCs as described above. Blood samples were obtained from all animals at sacrifice and
19 lymphoid organs including the adrenals, spleen, thymus, lymph nodes, and bone marrow were
20 removed. Cell suspensions were counted and adjusted to 1 x 108 cells/mL. Cell-mediated
21 immune responses were assessed using a footpad thickness assay and a leucocyte migration
22 inhibition (LMI) test, while humoral immune responses were determined by a hemagglutination
23 assay, and by monitoring cell counts in spleen, thymus, lymph nodes, femoral bone marrow, and
24 in splenic lymphocyte subsets. Plasma levels of corticosterone were measured along with levels
25 of such cytokines as TNF-a, IFN-y, IL-2, and IL-4. DNA damage in splenocytes and thymocytes
26 was also monitored using the Comet assay.
27 Table 4-18 shows decreases in the animal weight/organ weight ratios for spleen, thymus,
28 lymph nodes and adrenal gland as a result of methanol exposure. However, the splenocyte,
29 thymocyte, lymph node, and bone marrow cell counts were time-dependently lower in methanol-
30 treated animals.
4-67 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-18. Effect of methanol exposure on animal weight/organ weight ratios and on
cell counts in primary and secondary lymphoid organs of male Wistar rats.
Organ
Immunized
Control
15 days
30 days
Animal weight/organ weight ratio
Spleen
Thymus
Lymph node
Adrenal
3.88 ±0.55
1.35 ±0.29
0.10 ±0.01
0.14 ±0.01
2.85±0.36a
0.61±0.06a
0.08±0.01a
0.15 ±0.01
2.58±0.45a
0.63 ±0.04a
0.06 ± 0.02a
0.12±0.01a'b
Cell counts
Splenocytes (x 108)
Thymocytes (x 108)
Lymph node (x 107)
Bone marrow (x 107)
5.08 ±0.06
2.66 ±0.09
3.03 ±0.04
4.67 ±0.03
3.65±0.07a
1.95±0.03a
2.77 ± 0.07a
3.04±0.09a
3.71±0.06a
1.86±0.09a
2.20±0.06a'b
2.11±0.05a'b
Values are means ± six animals.
ap < 0.05 versus control groups.
bp < 0.05 versus 15-day treated group.
Source: Parthasarathy et al. (2007).
1 Parthasarathy et al. (2007) also documented their results on the cell-mediated and
2 humoral immunity induced by methanol. Leucocyte migration was significantly increased
3 compared to control animals, an LMI of 0.82 ± 0.06 being reported in rats exposed to methanol
4 for 30 days. This compares to an LMI of 0.73 ± 0.02 in rats exposed for 15 days and 0.41 ± 0.10
5 in controls. By contrast, footpad thickness and antibody titer were decreased significantly in
6 methanol-exposed animals compared to controls (18.32 ± 1.08, 19.73 ± 1.24, and 26.24 ± 1.68%
7 for footpad thickness; and 6.66 ± 1.21, 6.83 ± 0.40, and 10.83 ± 0.40 for antibody titer in 30-day,
8 15-day exposed rats, and controls, respectively).
9 Parthasarathy et al. (2007) also provided data in a histogram that showed a significant
10 decrease in the absolute numbers of Pan T cells, CD4, macrophage, major histocompatibility
11 complex (MHC) class II molecule expressing cells, and B cells of the methanol-treated group
12 compared to controls. The numbers of CDS cells were unaffected. Additionally, as illustrated in
13 the report, DNA single strand breakage was increased in immunized splenocytes and thymocytes
14 exposed to methanol versus controls. Although some fluctuations were seen in corticosterone
15 levels, the apparently statistically significant change versus controls in 15-day exposed rats was
16 offset by a decrease in 30-day exposed animals. Parthasarathy et al. (2007) also tabulated the
17 impacts of methanol exposure on cytokine levels; these values are shown in Table 4-19.
4-68
DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-19. The effect of methanol on serum cytokine levels in male Wistar rats
Cytokines (pg/mL)
IL-2
IL-4
TNF-a
IFN-y
Immunized
Control
1810 ±63.2
44.8 ±2.0
975 ± 32.7
1380 ±55.1
15 days
1303.3 ±57.1a
74.0 ± 5. la
578.3 ± 42.6a
9616±72.7a
30 days
1088.3 ± 68.8a'b
78.8±4.4a
585 ± 45a
950±59.6a
Values are means ± six animals.
ap < 0.05 versus control groups.
bp < 0.05 versus 15-day treated group.
Source: Parthasarathy et al. (2007).
1 Drawing on the results of DNA single strand breakage in this experiment, the authors
2 speculated that methanol-induced apoptosis could suppress specific immune functions such as
3 those examined in this research report. Methanol appeared to suppress both humoral and cell-
4 mediated immune responses in exposed Wistar rats.
4.6. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MOA
5 While the role of the methanol metabolite, formate, in inducing the toxic consequences of
6 acute exposure to methanol, including ocular toxicity and metabolic acidosis, is well established
7 in humans (see Section 4.1), there is controversy over the possible roles of the parent compound,
8 metabolites, and folate deficiency (potentially associated with methanol metabolism) in the
9 developmental neurotoxicity of methanol. Experiments that have attempted to address these
10 issues are reviewed in the following paragraphs.
4.6.1. Role of Methanol and Metabolites in the Developmental Toxicity of Methanol
11 Dorman et al. (1995) conducted a series of in vitro and in vivo studies that provide
12 information for identifying the proximate teratogen associated with developmental toxicity in
13 CD-I mice. The studies used CD-I ICR BR (CD-I) mice, HPLC grade methanol, and
14 appropriate controls. PK and developmental toxicity parameters were measured in mice exposed
15 to sodium formate (750 mg/kg by gavage), a 6-hour methanol inhalation (10,000 or 15,000 ppm),
16 or methanol gavage (1.5 g/kg). In the in vivo inhalation study, 12-14 dams/ group were exposed
17 to 10,000 ppm methanol for 6 hours on GD8,51 with and without the administration of
18 fomepizole (4-methylpyrazole) to inhibit the metabolism of methanol by ADH1. Dams were
19 sacrificed on GD10, and fetuses were examined for neural tube patency. As shown in
20 Table 4-20, the incidence of fetuses with open neural tubes was significantly increased in the
51 Dorman et al. (1995) state that GD8 was chosen because it encompasses the period of murine neurulation and the
time of greatest vulnerability to methanol-induced neural tube defects.
4-69
DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
methanol group (9.65% in treated versus 0 in control) and numerically but not significantly
increased in the group treated with methanol and fomepizole (7.21% in treated versus 0 in
controls). These data should not be interpreted to suggest that a decrease in methanol
metabolism is protective. As discussed in Section 3.1, rodents metabolize methanol via both
ADH1 and CAT. This fact and the Dorman et al. (1995) observation that maternal formate levels
in blood and decidual swellings (swelling of the uterine lining) did not differ in dams exposed to
methanol alone or methanol and fomepizole suggest that the role of ADH1 relative to CAT and
nonenzymatic methanol clearance is not of great significance in adult rodents.
Table 4-20. Developmental outcome on GD10 following a 6-hour 10,000 ppm
(13,104 mg/m3) methanol inhalation by CD-mice or formate gavage (750 mg/kg) on
GD8
Treatment
Air
Air/fomepizole
Methanol
Methanol/fomepizole
Water
Formate
No. of litters
14
14
12
12
10
14
Open neural tubes (%)
2.29 ±1.01
2.69 ±1.19
9.65±3.13a
7.21 ±2.65
0
2.02 ±1.08
Head length (mm)
3. 15 ±0.03
3.20 ±0.05
3.05 ±0.07
3.01 ±0.05
3.01 ±0.07
2.91 ±0.08
Body length (mm)
5.89 ±0.07
5.95 ±0.09
5.69 ±0.13
5.61 ±0.11
5.64 ±0.11
5.49 ±0.12
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Values are means ± S.D.
a/> < 0.05, as calculated by the authors.
Source: Dorman etal. (1995) (adapted).
The data in Table 4-20 suggest that the formate metabolite is not responsible for the
observed increase in open neural tubes in CD-I mice following methanol exposure. Formate
administered by gavage (750 mg/kg) did not increase this effect despite the fact that this formate
dose produced the same toxicokinetic profile as a 6-hour exposure to 10,000 ppm methanol
vapors (1.05 mM formate in maternal blood and 2.0 mmol formate/kg in decidual swellings).
However, the data are consistent with the hypotheses that the formaldehyde metabolite of
methanol may play a role. Both CAT and ADH1 activity are immature at days past conception
(DPC)8 (Table 4-21). If fetal ADH1 is more mature than fetal CAT, it is conceivable that the
decrease in the open neural tube response observed for methanol combined with fomepizole
(Table 4-20) may be due to fomepizole having a greater effect on the metabolism of fetal
methanol to formaldehyde than is observed in adult rats. Unfortunately, the toxicity studies were
carried out during a period of development where ADH1 expression and activity are just starting
to develop (Table 4-21); therefore, it is uncertain whether any ADH1 was present in the fetus to
be inhibited by fomepizole.
4-70
DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-21. Summary of ontogeny of relevant enzymes in CD-I mice and humans
Somites
CAT
mRNA
activity3
embryo
VYS
ADH1
mRNA
activity
embryo
VYS
ADH3
mRNA
activity
embryo
VYS
CD-I Mouse
Days Past Conception (DPC)
6.5
-
+
7.5
-
+
(8-12)
1
10
320
240
300
500
8.5
-
+
(13-20)
10
15
460
280
490
500
9.5
(21-29)
20
20
+
450
290
+
550
550
Human
Trimesters
1
N/A
+
—
2
N/A
+
—
o
6
N/A
+
+
""Activity of CAT and ADH1 are expressed as nmol/minute/mg and pmol/minute/mg, respectively.
Source: Harris et al. (2003).
1 Dorman et al. (1995) provide additional support for their hypothesis that methanol's
2 developmental effects in CD-I mice are not caused by formate in an in vitro study involving the
3 incubation of GD8 whole CD-I mouse embryos with increasing concentrations of methanol or
4 formate. Developmental anomalies were observed on GD9, including cephalic dysraphism,
5 asymmetry and hypoplasia of the prosencephalon, reductions of brachial arches I and II,
6 scoliosis, vesicles on the walls of the mesencephalon, and hydropericardium (Table 4-22). The
7 concentrations of methanol used for embryo incubation (0-375 mM) were chosen to be broadly
8 equivalent to the peak methanol levels in plasma that have been observed (approximately
9 100 mM) after a single 6-hour inhalation exposure to 10,000 ppm (13,104 mg/m3). As discussed
10 above, these exposure conditions induced an increased incidence of open neural tubes on GD10
11 embryos when pregnant female CD-I mice were exposed on GD8. (Table 4-20). Embryonic
12 lesions such as cephalic dysraphism, prosencephalic lesions, and brachial arch hypoplasia were
13 observed with 250 mM (8,000 mg/L) methanol and 40 mM (1,840 mg/L) formate. The study
14 authors noted that a formate concentration of 40 mM (1,840 mg/L) greatly exceeds blood
15 formate levels in mice inhaling 15,000 ppm methanol (0.75 mM = 35 mg/L), a teratogenic dose.
4-71
DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-22. Dysmorphogenic effect of methanol and formate in neurulating CD-I
mouse embryos in culture (GD8)
Treatment
Vehicle
Methanol
Formate
Concentration
(mM)
62
125
187
250
375
4
8
12
20
40
Live embryos
Total
20
13
14
13
15
12
12
13
9
16
16
No.
abnormal
3
1
5
7
7
7
2
5
5
7
14a
Cephalic dysraphism
Severe
0
0
1
2
2
6a
0
1
0
2
10a
Moderate
2
0
0
4
5
5
0
5
5
5
4
Total
2
0
2
6
7
lla
0
6
5
7
14a
Prosencephalic lesions
Hypoplasia
2
1
2
3
7a
9a
2
4
1
2
3
Asymmetry
0
0
2
1
1
1
0
2
2
1
5a
Total
2
1
4
4
8
10a
2
6
3
3
8
Brachial
arch
hypoplasia
0
0
1
1
6a
8a
1
0
0
1
13a
a/> < 0.05, as calculated by the authors.
Source: Dormanetal. (1995) (adapted).
1 As discussed in Section 4.3.3, a series of studies by Harris et al. (2004, 2003) also
2 provide evidence as to the moieties that may be responsible for methanol-induced developmental
3 toxicity. Harris et al. (2004) have shown that among methanol and its metabolites, viability of
4 cultured rodent embryos is most affected by formate. In contrast, teratogenic endpoints (of
5 interest to this risk assessment) in cultured rodent embryos are more sensitive to methanol and
6 formaldehyde than formate. Data from these studies indicate that developmental toxicity may be
7 more related to formaldehyde than methanol, as formaldehyde-induced teratogenicity occurs at
8 several orders of magnitude lower than methanol (Table 4-14) (Hansen et al., 2005; Harris et al.,
9 2004). It should also be noted that CAT, ADH1, and ADH3 activities are present in both the rat
10 embryo and VYS at stages as early as 6-12 somites (Harris et al., 2003); thus, it is presumable
11 that in these ex vivo studies methanol is metabolized to formaldehyde and formaldehyde is
12 subsequently metabolized to S-formylglutathione.
13 Studies involving GSH also lend support that formaldehyde may be a key proximal
14 teratogen. Inhibition of GSH synthesis with butathione sulfoximine (BSO) has little effect on
15 developmental toxicity endpoints, yet treatment with BSO and methanol or formaldehyde
16 increases developmental toxicity (Harris et al., 2004). Among the enzymes involved in methanol
17 clearance, only ADH3-mediated metabolism of formaldehyde is GSH dependent. This
18 hypothesis that ADH3-mediated metabolism of formaldehyde is important for the amelioration
19 of methanol's developmental toxicity is also supported by the diminished ADH3 activity in the
20 mouse versus rat embryos, which is consistent with the greater sensitivity of the mouse to
4-72 DRAFT-DO NOT CITE OR QUOTE
-------
1 methanol developmental toxicity (Harris et al., 2003) (Section 4.3.3). Similarly reasonable
2 explanations for this greater mouse sensitivity are not readily apparent for the two MO As
3 described below that attribute methanol toxicity to methanol metabolism per se, either through
4 the depletion of folate (Section 4.6.2) or the generation of reactive oxidant species (Section
5 4.6.3). Mouse livers actually have considerably higher hepatic tetrahydrofolate and total folate
6 than rat or monkey liver. Harris et al. (2003) and Johlin et al. (1987) have shown that CAT
7 activity in the embryo and VYS of rats and mice appear similar..
8 Without positive identification of the actual moiety responsible for methanol-induced
9 teratogenicity, MOA remains unclear. If the moiety is methanol, then it is possible that
10 generation of NADH during methanol oxidation creates an imbalance in other enzymatic
11 reactions. Studies have shown that ethanol intake leads to a >100-fold increase in cellular
12 NADH, presumably due to ADHl-mediated reduction of the cofactor NAD+ to NADH
13 (Cronholm, 1987; Smith and Newman, 1959). This is of potential importance because, for
14 example, ethanol intake has been shown to increase the in vivo and in vitro enzymatic reduction
15 of other endogenous compounds (e.g., serotonin) in humans (Svensson et al., 1999; Davis et al.,
16 1967). In rodents, CAT-mediated methanol metabolism may obviate this effect; in humans,
17 however, methanol is primarily metabolized by ADH1.
18 If the teratogenic moiety of methanol is formaldehyde, then reactivity with protein
19 sulfhydryls and nonprotein sulfhydryls (e.g., GSH) or DNA protein cross-links may be involved.
20 Metabolic roles ascribed to ADH3, particularly regulation of S-nitrosothiol biology (Foster and
21 Stamler, 2004), could also be involved in the MOA. Recently, Staab et al. (2008) have shown
22 that formaldehyde alters other ADH3-mediated reactions through cofactor recycling and that
23 formaldehyde alters levels of cellular S-nitrosothiol, which plays a key role in cellular signaling
24 and many cellular functions and pathways (Hess et al., 2005).
25 Studies such as those by Harris et al. (2004, 2003) and Dorman et al. (1995, 1994)
26 suggest that formate is not the metabolite responsible for methanol's teratogenic effects. The
27 former researchers suggest that formaldehyde is the proximate teratogen, and provide evidence
28 in support of that hypothesis. However, questions remain. Researchers in this area have not yet
29 reported using a sufficient array of enzyme inhibitors to conclusively identify formaldehyde as
30 the proximate teratogen. Studies involving other inhibitors or toxicity studies carried out in
31 genetically engineered mice, while not devoid of confounders, might further inform regarding
32 the methanol MOA for developmental toxicity. Even if formaldehyde is ultimately identified as
33 the proximate teratogen, methanol would likely play a prominent role, at least in terms of
34 transport to the target tissue. The high reactivity of formaldehyde would limit its unbound and
35 unaltered transport as free formaldehyde from maternal to fetal blood (Thrasher and Kilburn,
4-73 DRAFT-DO NOT CITE OR QUOTE
-------
1 2001), and, as has been discussed, the capacity for the metabolism of methanol to formaldehyde
2 is likely lower in the fetus and neonate versus adults (Section 3.3).
4.6.2. Role of Folate Deficiency in the Developmental Toxicity of Methanol
3 As discussed in Sections 3.1 and 4.1, humans and other primates are susceptible to the
4 effects of methanol exposure associated with formate accumulation because they have lower
5 levels of hepatic tetrahydrofolate-dependent enzymes that help in formate oxidation.
6 Tetrahydofolate-dependent enzymes and critical pathways that depend on folate, such as purine
7 and pyrimidine synthesis, may also play a role in the developmental toxicity of methanol.
8 Studies of rats and mice fed folate-deficient diets have identified adverse effects on reproductive
9 performance, implantation, fetal growth and developmental defects, and the inhibition of folate
10 cellular transport has been associated with several developmental abnormalities, ranging from
11 neural tube defects to neurocristopathies such as cleft-lip and cleft-palate, cardiacseptal defects,
12 and eye defects (Antony, 2007). Folate deficiency has been shown to exacerbate some aspects of
13 the developmental toxicity of methanol in mice (see discussion of Fu et al., 1996, and Sakanashi
14 et al., 1996, in Section 4.3.1) and rats (see discussion of Aziz et al., 2002, in Section 4.4.1).
15 The studies in mice focused on the influence of FAD on the reproductive and skeletal
16 malformation effects of methanol. Sakanashi et al. (1996) showed that dams exposed to
17 5 g/kg-day methanol on GD6-GD15 experienced a threefold increase in the percentage of litters
18 affected by cleft palate and a 10-fold increase in the percentage of litters affected by exencephaly
19 when fed a FAD (resulting in a 50% decrease in liver folate) versus a FAS diet. They speculated
20 that the increased methanol effect from FAD diet could have been due to an increase in tissue
21 formate or a critical reduction in conceptus folate concentration immediately following the
22 methanol exposure. The latter appears more likely, given the high levels of formate needed to
23 cause embryotoxicity (Section 4.3.3) and the decrease in conceptus folate that is observed within
24 2 hours of GD8 methanol exposure (Dorman et al., 1995). Fu et al. (1996) confirmed the
25 findings of Sakanashi et al. (1996) and also determined that the maternal FAD diet had a much
26 greater impact on fetal liver folate than maternal liver folate levels.
27 The rat study of Aziz et al. (2002) focused on the influence of FAD on the developmental
28 neurotoxicity of methanol. Experiments by Aziz et al. (2002) involving Wistar rat dams and
29 pups exposed to methanol during lactation provide evidence that methanol exposure during this
30 postnatal period affects the developing brain. These effects (increased spontaneous locomotor
31 activity, decreased conditioned avoidance response, disturbances in dopaminergic and
32 cholinergic receptors and increased expression of GAP-43 in the hippocampal region) were more
33 pronounced in FAD as compared to FAS rats. This suggests that folic acid may play a role in
34 methanol-induced neurotoxicity. These results do not implicate any particular proximate
4-74 DRAFT-DO NOT CITE OR QUOTE
-------
1 teratogen, as folate deficiency can increase levels of both methanol, formaldehyde and formate
2 (Medinsky et al., 1997). Further, folic acid is used in a number of critical pathways such as
3 purine and pyrimidine synthesis. Thus, alterations in available folic acid, particularly to the
4 conceptus, could have significant impacts on the developing fetus apart from the influence it is
5 presumed to have on formate removal.
4.6.3. Methanol-Induced Formation of Free Radicals, Lipid Peroxidation, and Protein
Modifications
6 Oxidative stress in mother and offspring has been suggested to be part of the teratogenic
7 mechanism of a related alcohol, ethanol. Certain reproductive and developmental effects (e.g.,
8 resorptions and malformation rates) observed in Sprague-Dawley rats following ethanol
9 exposure were reported to be ameliorated by antioxidant (Vitamin E) treatment (Wentzel et al.,
10 2006; Wentzel and Eriksson, 2006). A number of studies have examined markers of oxidative
11 stress associated with methanol exposure.
12 Skrzydlewska et al. (2005) provided inferential evidence for the effects of methanol on
13 free radical formation, lipid peroxidation, and protein modifications, by studying the protective
14 effects of N-acetyl cysteine and the Vitamin E derivative, U83836E, in the liver of male Wistar
15 rats exposed to the compound via gavage. Forty-two rats/group received a single oral gavage
16 dose of either saline or 50% methanol. This provided a dose of approximately 6,000 mg/kg, as
17 calculated by the authors. Other groups of rats received the same concentration of methanol, but
18 were also injected intraperitoneally with either N-acetylcysteine or U-83836E. N-acetylcysteine
19 and U-83836E controls were also included in the study design. Animals in each group were
20 sacrificed after 6, 14, and 24 hours or after 2, 5, or 7 days. Livers were rapidly excised for
21 electron spin resonance (ESR) analysis, and 10,000 x g supernatants were used to measure GSH,
22 malondialdehyde, a range of protein parameters, including free amino and sulfhydryl groups,
23 protein carbonyls, tryptophan, tyrosine, and bityrosine, and the activity of cathepsin B.
24 Skrzydlewska et al. (2005) provided data that showed an increase in an ESR signal at
25 g = 2.003 in livers harvested 6 and 12 hours after methanol exposure. The signal, thought to be
26 indicative of free radical formation, was opposed by N-acetylcysteine and U83836E. Other
27 compound-related changes included: 1) a significant decrease in GSH levels that was most
28 evident in rats sacrificed 12 and 24 hours after exposure; 2) increased concentrations in the lipid
29 peroxidation product, malondialdehyde (by a maximum of 44% in the livers of animals
30 sacrificed 2 days after exposure); 3) increased specific concentrations of protein carbonyl groups
31 and bityrosine; but 4) reductions in the specific level of tryptophan. Given the ability of N-
32 acetylcysteine and U83836E to oppose these changes, at least in part, the authors speculated that
33 a number of potentially harmful changes may have occurred as a result of methanol exposure.
4-75 DRAFT-DO NOT CITE OR QUOTE
-------
1 These include free radical formation, lipid peroxidation, and disturbances in protein structure.
2 However, it is unclear whether or not the metabolites of methanol, formaldehyde, and/or formate,
3 were involved in any of these changes.
4 Rajamani et al. (2006) examined several oxidative stress parameters in male Wistar rats
5 following methotrexate-induced folate deficiency. Compared to controls, the levels of free
6 radical scavengers SOD, CAT, GSH peroxidase, oxidized GSH, protein carbonyls, and lipid
7 peroxidation were elevated in several regions of the brain, with greater increases observed in the
8 MTX-methanol-treated animals than in the MTX-alone group. The level of GSH and protein
9 thiols was decreased in all regions of the brain, with a greater decrease observed in the MTX-
10 methanol-treated animals than MTX-treated animals.
11 Dudka (2006) measured the total antioxidant status (TAS) in the brain of male Wistar rats
12 exposed to a single oral gavage dose of methanol at 3 g/kg. The animals were kept in a nitrous
13 oxide atmosphere (N2O/O2) throughout the experiment to reduce intrinsic folate levels, and
14 various levels of ethanol and/or fomepizole (as ADH antidotes) were administered i.p. after
15 4 hours. Animals were sacrificed after 16 hours, the brains homogenized, and the TAS
16 determined spectrophotometrically. As illustrated graphically by the author, methanol
17 administration reduced TAS in brain irrespective of the presence of ADH antidotes. The author
18 speculated that, while most methanol is metabolized in the liver, some may also reach the brain.
19 Metabolism to formate might then alter the NADH/NAD+ ratio resulting in an increase in
20 xanthine oxidase activity and the formation of the superoxide anion.
21 Parthasarathy et al. (2006b) investigated the extent of methanol-induced oxidative stress
22 in rat lymphoid organs. Six male Wistar rats/group received 2,370 mg/kg methanol (mixed 1:1
23 with saline) injected i.p. for 1, 15 or 30 days. A control group received a daily i.p. injection of
24 saline for 30 days. At term, lymphoid organs such as the spleen, thymus, lymph nodes, and bone
25 marrow were excised, perfused with saline, then homogenized to obtain supernatants in which
26 such indices of lipid peroxidation as malondialdehyde, and the activities of CAT, SOD, and GSH
27 peroxidase were measured. Parthasarathy et al. (2006b) also measured the concentrations of
28 GSH and ascorbic acid (nonenzymatic antioxidants) and the serum concentrations of a number of
29 indicators of liver and kidney function, such as ALT, AST, blood urea nitrogen (BUN), and
30 creatinine.
31 Table 4-23 shows the time-dependent changes in serum liver and kidney function
32 indicators that resulted from methanol administration. Treatment with methanol for increasing
33 durations resulted in increased serum ALT and AST activities and the concentrations of BUN and
34 creatinine.
4-76 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-23. Time-dependent effects of methanol administration on serum liver and
kidney function, serum ALT, AST, BUN, and creatinine in control and experimental
groups of male Wistar rats
Parameters
ALT (umoles/min-mg)
AST (umoles/min-mg)
BUN (mg/L)
Creatinine (mg/L)
Methanol administration (2,370 mg/kg)
Control
29.0 ±2.5
5.8 ±0.4
301 ±36
4.6 ±0.3
Single dose
31.4±3.3
6.4 ±0.3
332 ±29
4.8 ±0.3
15 days
53.1±2.3a
9.0±1.2a
436±35a
5.6±0.2a
30 days
60.4±2.8a
13.7±1.2a
513±32a
7.0±0.4a
Values are means ± S.D. of 6 animals.
a/> < 0.05 versus controls.
Source: Parthasarathy et al. (2006b) (adapted).
Table 4-24. Effect of methanol administration on male Wistar rats on malondialdehyde
concentration in the lymphoid organs of experimental and control groups and the
effect of methanol on antioxidants in spleen
Parameters
Methanol administration (2,370 mg/kg)
Control
Single dose
15 days
30 days
Malondialdehyde in lymphoid organs
Spleen
Thymus
Lymph nodes
Bone marrow
2.62 ±0.19
3.58 ±0.35
3. 15 ±0.25
3. 14 ±0.33
4.14±0.25a
5.76±0.36a
5.08±0.24a
4.47±0.18a
7.22±0.31a
9.23±0.57a
8.77±0.57a
7.20 ± 0.42a
9.72±0.52a
11.6±0.33a
9.17±0.67a
9.75±0.56a
Antioxidant levels in spleen
SOD (units/mg protein)
CAT (umoles H2O2
consumed/min-mg protein
GPx (ug GSH
consumed/min-mg protein)
GSH (ug/mg protein)
Vit C (ug/mg protein)
2.40 ±0.16
35.8 ±2.77
11.2 ±0.60
2.11±0.11
0.45 ±0.04
4.06±0.19a
52.5±3.86a
20.0±1.0a
3.75±0.15a
0.73±0.05a
1.76±0.09a
19.1±1.55a
7.07±0.83a
1.66±0.09a
0.34±0.18a
1.00±0.07a
10.8±1.10a
5.18±0.45a
0.89±0.04a
0.11±0.03a
Values are means ± S.D. of six animals.
a/> < 0.05, versus controls.
Source: Parthasarathy et al. (2006b) (adapted).
1 Table 4-24 gives the concentration of malondialdehyde in the lymphoid organs of control
2 and experimental groups, and, as an example of all tissue sites examined, the levels of enzymatic
3 and nonenzymatic antioxidants in spleen. The results show that malondialdehyde concentrations
4 were time-dependently increased at each tissue site and that, in spleen as an example of all the
5 lymphoid tissues examined, increasing methanol administration resulted in lower levels of all
6 antioxidants examined compared to controls. Parthasarathy et al. (2006b) concluded that
4-77
DRAFT-DO NOT CITE OR QUOTE
-------
1 exposure to methanol may cause oxidative stress by altering the oxidant/antioxidant balance in
2 lymphoid organs in the rat.
4.6.4. Exogenous Formate Dehydrogenase as a Means of Detoxifying the Formic Acid that
Results from Methanol Exposure
3 In companion reports, Muthuvel et al. (2006a, 2006b) used 6 male Wistar rats/group to
4 test the ability of exogenously-administered formate dehydrogenase (FD) to reduce the serum
5 levels of formate that were formed when 3 g/kg methanol was administered i.p. to rats in saline.
6 In the first experiment, purified FD (from Candida boitinif) was administered by i.v. conjugated
7 to the N-hydroxysuccinimidyl ester of monomethoxy polyethylene glycol propionic acid
8 (PEG-FD) (Muthuvel et al., 2006a). In the second, rats were administered FD-loaded
9 erythrocytes (Muthuvel et al., 2006b). In the former case, some groups of rats were made folate
10 deficient by means of a folate-depleted diet; in the latter, folate deficiency was brought about by
11 i.p. administration of methotrexate. In some groups, the rats received an infusion of an
12 equimolar mixture of carbonate and bicarbonate (each at 0.33 mol/L) to correct the formate-
13 induced acidosis. As illustrated by the authors, methanol-exposed rats receiving a folate-
14 deficient diet showed significantly higher levels of serum formate than those receiving a folate-
15 sufficient diet. However, administration of native or PEG-FD reduced serum formate in
16 methanol-receiving folate-deficient rats to levels seen in animals receiving methanol and the
17 folate-sufficient diet.
18 In the second report, Muthuvel et al. (2006b) carried out some preliminary experiments to
19 show that hematological parameters of normal, reconstituted but unloaded, and reconstituted and
20 FD-loaded erythrocytes, were similar. In addition, they showed that formate levels of serum
21 were reduced in vitro in the presence of FD-loaded erythrocytes. Expressing blood formate
22 concentration in mmol/L at the 1-hour time point after carbonate/bicarbonate and enzyme-loaded
23 erythrocyte infusion via the tail vein, the concentration was reduced from 10.63 ±1.3 (mean ±
24 S.D.) in methanol and methotrexate-receiving controls to 5.83 ± 0.97 (n = 6). This difference
25 was statistically significant at the/? < 0.05 level. However, FD-loaded erythrocytes were less
26 efficient at removing formate in the absence of carbonate/bicarbonate. Effective elimination of
27 formate appears to require an optimum pH for the FD activity in the enzyme-loaded
28 erythrocytes.
4.6.5. Mechanistic Data Related to the Potential Carcinogenicity of Methanol
4.6.5.1. Gen otoxicity
29 The genotoxicity/mutagenicity of methanol has not been extensively studied, but the
30 results of those studies that have thus far have been mostly negative. For example, in a survey of
4-78 DRAFT-DO NOT CITE OR QUOTE
-------
1 the capacity of 71 drinking water contaminants to induce gene reversion in the Ames test,
2 Simmon et al. (1977) listed methanol as one of 45 chemicals that gave negative results with
3 Salmonella typhimurium strains TA98, 100, 1535, 1537, and 1538, irrespective of the presence
4 or absence of metabolic activation (an S9 microsomal fraction). This result was confirmed by
5 DeFlora et al. (1984) and in NEDO (1987) for the same strains of Salmonella. DeFlora et al.
6 (1984) also found methanol to be negative for induction of DNA repair in E. coli strains WP2,
7 WP2 (uvrA, polA~), and CM871 (uvrA, recA, lexA~), again irrespectively of the presence or
8 absence of S9.
9 Abbondandolo et al. (1980) used a ade6-60/radlO-198,h~ strain of Schizosaccharomyces
10 pombe (PI strain) to determine the capacity of methanol and other solvents to induce forward
11 mutations. Negative results were obtained for methanol, irrespective of metabolic activation
12 status. In other genotoxicity/mutagenicity studies of methanol using fungi, Griffiths (1981)
13 reported methanol to be negative for the induction of aneuploidy in Neurospora crassa. By
14 contrast, weakly positive results for the compound were obtained by Crebelli et al. (1989) for the
15 induction of chromosomal malsegregation in the diploid strain PI of Aspergillus nidulans.
16 In an extensive review of the capacity of a wide range of compounds to induce
17 transformation in mammalian cell lines, Heidelberger et al. (1983) reported methanol to be
18 negative in Syrian hamster embryo (SHE) cells. It also did not enhance the transformation of
19 SHE cells by Simian adenovirus. However, McGregor et al. (1985) reported in an abstract that a
20 statistically significant increase in forward mutations in the mouse lymphoma L5178Y tk+/tk" cell
21 line occurred at a concentration of 7.9 mg/mL methanol in the presence of S9.
22 The capacity of methanol to bring about genetic changes in human cell lines was
23 examined by Ohno et al. (2005), who developed a system in which the chemical activation of the
24 p53R2 gene was assessed by the incorporation of a/>5.iR2-dependent luciferase reporter gene
25 into two human cell lines, MCF-7 and HepG2. Methanol, among 80 chemicals tested in this
26 system, gave negative results. NEDO (1987) used Chinese hamster lung (CHL) cells to monitor
27 methanol's capacity to induce 1) forward mutations to azaguanine, 6-thioguanine, and ouabain
28 resistance, and 2) chromosomal aberrations (CA) though with negative results throughout.
29 However, methanol did display some capacity to induce sister chromatid exchanges (SCE) in
30 CHL cells, since the incidence of these lesions at the highest concentration (28.5 mg/mL) was
31 significantly greater than in controls (9.41 ± 0.416 versus 6.42 ± 0.227 [mean ± SE per 100
32 cells]).
33 In an in vivo experiment examining the genotoxicity/mutagenicity of methanol, Campbell
34 et al. (1991) exposed 10 male C57BL/6J mice/group to 0, 800, or 4,000 ppm (0, 1,048, and
35 5,242 mg/m3) methanol, 6 hours/day, for 5 days. At sacrifice, blood cells were examined for the
36 formation of micronuclei (MN). Excised lung cells for SCE, CA and MN, and excised testicular
4-79 DRAFT-DO NOT CITE OR QUOTE
-------
1 germ cells were examined for evidence of synaptonemal damage, in each case with negative
2 results.
3 There was no evidence of methanol-induced formation of MN in the blood of fetuses or
4 pregnant CD-mice when the latter were gavaged twice daily with 2,500 mg/kg methanol on
5 GD6-GD10 (Fu et al., 1996). The presence of marginal or adequate amounts of folic acid in the
6 diet of the dams did not affect MN formation. NEDO (1987) carried out an in vivo MN test in
7 6 male SPF mice/group who received a single gavage dose of 1,050, 2,110, 4,210, and 8,410
8 mg/kg methanol. Twenty-four hours later, 1,000 cells were counted for MN in bone marrow
9 smears. No compound-related effects on MN incidence were observed. Table 4-25 provides a
10 summary of the genotoxicity/mutagenicity studies of methanol.
Table 4-25. Summary of genotoxicity studies of methanol
Test system
Cell/strain
Result
Reference
Comments
In vitro tests
Gene reversion/
S. typhimurium
DNA repair/It, coli
Forward mutations/
S. pombe
Aneuploidy/
N. crassa
Chromosomal
malsegregation/
A. nidulans
Forward
mutations/Mouse
lymphoma cells
Forward
mutations/Chinese
hamster lung cells
Chromosomal
aberrations/Chinese
hamster lung cells
Sister chromatid
exchanges/Chinese
hamster lung cells
TA98; TA100; TA1535,
TA1537, TA1538
TA98; TA100; TA1535,
TA1537, TA1538
TA98; TA100; TA1535,
TA1537, TA1538
WP2, WP2 (uvrA,
polA'),CM&ll(uvrA', recA~,
lexA)
PI
(ade6-60/radl 0-198, K)
(arg-1, ad-3A, ad-3B, nic-
2, to/, C/c, D/d, E/e)
PI (diploid)
L5178Ytk+/tk~
to azaguanine, 6-
thioguanine and ouabain
resistance
- (+S9); - (-S9)
- (+S9); - (-S9)
- (+S9); - (-S9)
- (+S9), - (-S9)
-(+S10), -(-S10)
- (S9 status not
reported)
+ (S9 status not
reported)
+ (+S9), ND (-S9)
- (-S9), ND (+S9)
- (-S9), ND (+S9)
+ (-S9), ND (+S9)
Simmon etal. (1977)
De Flora etal. (1984)
NEDO (1987)
DeFlora etal. (1984)
Abbondandolo et al.
(1980)
Griffiths (1981)
Crebelli etal. (1989)
McGregor et al.
(1985)
NEDO (1987)
NEDO (1987)
NEDO (1987)
Molecular activation
used a 10,000 xg(S10)
supernatant from liver
of induced Swiss mice
Results reported in an
abstract
4-80
DRAFT-DO NOT CITE OR QUOTE
-------
Test system
Genetic activation/
human cell lines
Cell transformation/
Syrian hamster
embryo cells
Cell/strain
MCF-7 and HepG2
containing ap53R2-
dependent luciferase
reporter gene
with/without
transformation by Simian
adenovirus
Result
- (-S9), ND (+S9)
- (-S9), ND (+S9)
- (-S9), ND (+S9)
Reference
Ohno et al. (2005)
Heidelberger et al.
(1983)
Comments
Review
In vivo tests
Mouse/MN
formation
Mouse/SCEs
Mice/CA
Mouse/synaptonem
al damage
Mouse/MN
formation
C57BL/6J
(Blood cells)
C57BL/6J
(Lung cells)
C57BL/6J
(Lung cells)
C57BL/6J
(Lung cells)
C57BL/6J
(Testicular germ cells)
CD-I
(Blood cells)
SPF
(Bone marrow cells)
-
-
-
-
-
-
-
Campbell et al.
(1991)
Campbell et al.
(1991)
Campbell et al.
(1991)
Campbell et al.
(1991)
Campbell et al.
(1991)
Fuetal. (1996)
NEDO (1987)
Molecular activation
not applicable
Molecular activation
not applicable
Molecular activation
not applicable
Molecular activation
not applicable
Molecular activation
not applicable
Molecular activation
not applicable
Molecular activation
not applicable
ND = not determined.
4.6.5.2. Lymphoma Responses Reported in ERF Life span Bioassays of Compounds
Related toMethanol, Including an Analogue (Ethanol), Precursors (Aspartame and Methyl
Tertiary Butyl Ether), and a Metabolite (Formaldehyde)
1 The ERF or the European Foundation of Oncology and Environmental Sciences have
2 conducted nearly 400 experimental bioassays on over 200 compounds/agents, using some
3 148,000 animals over nearly 4 decades. Of the over 200 compounds tested by ERF,52 8 have
4 been associated with an increased incidence of hemolymphoreticular tumors in Sprague-Dawley
5 rats, suggesting that it may be a rare and potentially species/strain-specific finding. These eight
6 chemicals are: methanol, formaldehyde, aspartame, MTBE, DIPE, TAME, mancozeb, and
7 toluene. Methanol, formaldehyde, aspartame, and MTBE share a common metabolite,
8 formaldehyde, and DIPE, TAME, methanol and MTBE are all gasoline-oxygenate additives
9 (Caldwell et al., 2008).
10 With the exception of a positive study for malignant lymphomas in Swiss Webster mice
11 exposed to methanol (Apaja, 1989), lymphoma responses have not been reported by other
12 institutions performing long-term testing of these chemicals in various strains of rats, including
52 While ERF has tested over 200 chemicals in 398 long-term ERF bioassays, only 112 of their bioassays have been
published to date (Caldwell, et al., 2008). The extent to which the unpublished studies are documented varies.
4-81 DRAFT-DO NOT CITE OR QUOTE
-------
1 formaldehyde inhalation studies in F344 (Kerns et al., 1983; Kamata et al., 1987)53 and Sprague-
2 Dawley (Albert et al., 1982; Sellakumar et al., 1985) rats, formaldehyde oral studies in Wistar
3 rats (Til et al., 1989; Tobe et al., 1989), toluene oral studies in F344 rats (NTP, 1990), MTBE
4 inhalation studies in F344 rats (Chun et al., 1992), aspartame oral studies in Wistar (Ishii et al.,
5 1981) and Sprague-Dawley (Molinary, 1984) rats, and methanol inhalation studies in F344 rats
6 (NEDO, 1987, 1985/2008b). Several differences in study design may contribute to the
7 differences in responses observed across institutions, particularly study duration and test animal
8 strain. Fischer-344 rats have a high background of mononuclear cell leukemia (20% in control
9 females)54 and a very low background rate of "lymphoma" (0% in control females) at 104 weeks
10 (NTP, 2006). In contrast, Sprague-Dawley rats from NTP studies exhibit a low background rate
11 of "leukemias" (0.8% in control females) and a higher background rate of "lymphomas" (1.08%
12 in control females) at 104 weeks (NTP, 2006). Similarly, Chandra et al. (1992) report a
13 background level of 1.6% for malignant lymphocytic lymphomas in female control Sprague-
14 Dawley rats for 17 2-year carcinogenicity studies.
15 In lifetime studies of Sprague-Dawley rats at ERF, the overall incidence of
16 lymphomas/leukemias has been reported to be 13.3% (range, 4.0-25.0%) in female historical
17 controls (2,274 rats) and 20.6% (range, 8.0-30.9%) in male historical controls (2,265 rats)
18 (Soffritti et al., 2007). The difference in background rates reported by ERF versus other labs for
19 this tumor type could be due to differences in study duration, differences in tumor classification
20 systems, and/or misdiagnoses due to confounding effects (see discussion in Section 4.2.1.3). A
21 high background incidence can increase the difficulty of detecting chemically related responses
22 (Melnick et al., 2007), and the background rate reported by ERF for this tumor type is considered
23 to be high relative to other tumor types and relative to the background rate for this tumor type in
24 Sprague-Dawley rats from other laboratories (EFSA, 2006; Cruzan, 2009).55 However, it is in a
25 range that can be considered reasonable for studies that employ a large number of animals
26 (Leakey et al., 2003; Caldwell et al., 2009).
27 Thus, with respect to the identification of hemolymphoreticular carcinogenic responses,
28 life span studies of Sprague-Dawley rats performed by ERF may be more sensitive than the
29 2-year studies of Fischer 344 (F344) strain of rats used by NTP (1990) and NEDO (1987,
53 Though Kerns et al. (1983) did not report a positive response for lymphoma, a survival-adjusted analysis of the
data from this study indicates a statistically significant trend in female rat mononuclear cell leukemia (p = 0.0056)
and a nearly significant increase in female mouse lymphoma (p = 0.06). In the Kamata et al., (1987) study, only a
small percentage of the original 32 rats/group survived to the end of the study (28 months) due largely to interim
sacrifices (5/group) at 12, 18 and 24 months.
54 Due to this and other health concerns, NTP transitioned to the use of Wistar rats in 2008, and more recently has
adopted Sprague-Dawley rats as the rat model for NTP studies due to the reproductive capability and size of Wistar
rats (http://ntp.niehs.nih.gov/go/29502).
55 Cruzan (2009) reports that the incidences of total cancers derived from bloodforming cells, designated as
hemolymphoreticular tumors by Ramazzini pathologists, is consistently about four times higher than the incidences
of such tumors in SD rats recorded in the Charles River Laboratory historical database (CRL database).
4-82 DRAFT-DO NOT CITE OR QUOTE
-------
1 1985/2008b). The results of ERF studies of the carcinogenic potential of methanol, MTBE, and
2 formaldehyde and related chemicals, ethanol and aspartame, are summarized in this section.
3 4.6.5.2.1. Ethanol. In a study that was reported in the same article that described the
4 carcinogenic responses of Sprague-Dawley rats to methanol, Soffritti et al. (2002a) exposed
5 110 Sprague-Dawley rats/sex/group to ethanol in drinking water at concentrations of 0 or 10%
6 (v/v) beginning at 39 weeks of age and ending at natural death, and including a single breeding
7 cycle. Various numbers of the offspring (30 male controls, 39 female controls, 49 exposed
8 males, and 55 exposed females) were exposed to ethanol in drinking water at the same
9 concentrations as their parents. The experiment concluded with the death of the last offspring at
10 179 weeks of age. Animals were examined for the same toxicological parameters as those
11 described for methanol, and organs and tissues were grossly and histopathologically examined at
12 necropsy. Soffritti et al. (2002a) reported that food and drinking water intake were lower in
13 exposed animals compared to controls but that body weight changes were similar among the
14 groups.56 There were no compound-related clinical signs of toxicity and no differences in
15 survival rates among the groups. While there were apparently no nononcogenic pathological
16 changes evident on gross inspection or histopathologic examination, a number of benign and
17 malignant tumors were considered by the authors to be compound-related. Compared to
18 controls, these included increased incidences of: 1) total malignant tumors in male and female
19 breeders (145/220 versus 99/220) and offspring (49/69 versus 54/104); 2) total malignant tumors
20 per 100 animals in female breeders (130 versus 60.9) and offspring (164.1 versus 96.4); 3)
21 carcinomas of the head and neck, especially to the oral cavity, lips and tongue in male and
22 female breeders (27/220 versus 5/220) and offspring (26/69 versus 5/104); 4) squamous cell
23 carcinomas of the forestomach in male and female breeders (5/220 versus 0/220) and offspring
24 (2/69 versus 0/104); 5) interstitial cell adenomas of the testis in male breeders (23/110 versus
25 9/110) and offspring (4/30 versus 4/49); 6) Sertoli-Leydig cell tumors in ovaries of female
26 offspring (3/39 versus 1/55); 7) adenocarcinomas of the uterus in female breeders (9/110 versus
27 2/110) and offspring (8/39 versus 6/55); 8) pheochromoblastomas in male and female breeders
28 (13/220 versus 4/220) and offspring (4/69 versus 2/104); and 9) osteosarcomas in male and
29 female breeders ([for the head] 14/220 versus 4/220) and offspring ([for the head] 10/69 versus
30 7/104). Notably, Soffritti et al. (2002a) did not observe increases in any of the lymphoma
31 responses reported in their methanol bioassay. Incidence data for these responses and their
statistical significance compared to controls are shown in Table 4-26.
56 Test animals were likely receiving calories from ethanol exposure.
4-83 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-26. Incidence of carcinogenic responses in Sprague-Dawley rats exposed to
ethanol in drinking water for up to 2 years
Tissues/affected sites
Total malignant tumors
Oral cavity (carcinomas)
Forestomach (squamous
cell carcinomas)
Testis (interstitial cell
adenomas)
Sertoli-Leydig cell tumors
(ovary)
Uterus (adenocarcinomas)
Head (osteosarcomas)
Adrenal gland
(pheochromoblastomas)
Total malignant tumors per
100 animals
Concentration in percent (v/v)
Male (breeders)
0
51/110
3/110
0/110
9/110
0/110
3/110
61.8
10
66/110
15/1 10b
2/110
23/1 10d
8/110
9/110
89.1b
Female (breeders)
0
48/110
2/110
0/110
1/110
2/110
4/110
1/110
60.9
10
79/1 10b
12/110
3/110
2/110
9/1 10C
6/110
4/110
130b
Male (offspring)
0
23/49
2/49
0/49
4/49
4/49
1/49
61.2
10
23/303
10/30b
1/30
4/30
6/30
4/30
136.7b
Female (offspring)
0
31/55
3/55
0/55
0/55
6/55
3/55
1/55
96.4
10
26/39
16/39b
1/39
3/39
8/39
4/39
0/39
164.1b
a/> < 0.05 using the ^ test, as calculated by the authors.
bp < 0.01 using the %2 test, as calculated by the authors.
°p < 0.05 using Fisher's exact test, as calculated by the reviewers.
dp < 0.01 using Fisher's exact test, as calculated by the reviewers.
Source: Soffritti et al. (2002a).
1 4.6.5.2.2. Aspartame. Soffritti et al. (2006, 2005) reported the results of a cancer bioassay on
2 the artificial sweetener aspartame. The study has potential relevance to the carcinogenicity of
3 methanol because aspartame has been shown to be metabolized to aspartic acid, phenylalanine,
4 and methanol in the GI tract prior to absorption into systemic circulation. In the study, aspartame
5 (>98% purity) was given to 100 or 150 Sprague-Dawley rats/sex/group in feed at dietary
6 concentrations of 0, 80, 400, 2,000, 10,000, 50,000, and 100,000 ppm. The authors reported
7 these concentrations to be equivalent to approximate daily doses of 0, 4, 20, 100, 500, 2,500, and
8 5,000 mg/kg-day, respectively, under the conditions of the study. Animals were maintained until
9 their "natural death," with the in-life phase of the experiment concluding with the death of the
10 last animal at 151 weeks. All animals were monitored for body weight, food and water
11 consumption. At death, animals were examined grossly and given a complete histopathological
12 examination.
13 Soffritti et al. (2006, 2005) reported that there were no differences among the groups in
14 mean body weight, survival, or daily water consumption. However, there appeared to be a dose-
15 related reduction in food consumption in both male and female rats.
4-84 DRAFT-DO NOT CITE OR QUOTE
-------
1 The principal histopathological finding was an increased incidence of lymphomas and
2 leukemias in female rats, a response reported by the authors to be statistically significant
3 compared to concurrently exposed controls (Table 4-27) and greater than the range of overall
4 incidence of lymphomas and leukemias in historical controls at the ERF (13.4% [range, 7.0-
5 18.4] in females and 21.8% [range, 8.0-30.9] in males). Among the hemolymphoreticular
6 neoplasms observed, the most frequent type observed was lympho-immunoblastic lymphoma.
7 The authors concluded that aspartame causes a "dose-related statistically significant increase in
8 lymphomas and leukemias in females at dose levels very near those to which humans can be
9 exposed." They postulated that an increase in the incidence of lymphomas and leukemias could
10 be associated with the formation of either methanol or formaldehyde. Other potentially
11 compound-related effects of aspartame were (1) an increase in combined dysplastic hyperplasias,
12 papillomas, and carcinomas of the renal pelvis and ureter, and (2) an increasing trend in the
13 formation of malignant schwannomas in peripheral nerves (Table 4-28).
14 The European Commission asked the European Food Safety Authority (EFSA) to assess
15 the study and review all ERF findings related to aspartame. An EFSA review panel assessed the
16 study and considered additional unpublished data provided to it by the ERF. In their report,
17 EFSA (2006) concluded that the Soffritti et al. (2005, 2006) study had flaws that brought into
18 question the reported findings. The review panel noted the high background of chronic
19 inflammatory changes in the lung and other vital organs. These background inflammatory
20 changes were thought to contribute significant uncertainty to the interpretations of the study. In
21 fact, the review panel concluded that most of the documented changes, in particular, the apparent
22 compound-related increase in lymphomas and leukemias, may have been incidental findings and,
23 therefore, unrelated to aspartame.
Table 4-27. Incidence of lymphomas and leukemias in Sprague-Dawley rats exposed to
aspartame via the diet
Group
I
II
III
IV
V
VI
VII
ppm in feed
0
80
400
2,000
10,000
50,000
100,000
Dose (mg/kg-day)
0
4
20
100
500
2,500
5,000
Lymphomas/leukemias (incidence and %)
Male
31/150(21)
23/150 (15)
25/150 (17)
33/150 (22)
15/100 (15)
20/100 (20)
29/100 (29)
Female
13/150 (9)
22/150 (15)
30/150b (20)
28/150a(19)
19/100a(19)
25/100b (25)
25/100b (25)
*p < 0.05 using the poly-k test.
V < 0.01 using the poly-k test.
Source: Soffritti et al. (2006, 2005).
4-85
DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-28. Incidence of combined dysplastic hyperplasias, papillomas and carcinomas
of the pelvis and ureter and of malignant schwannomas in peripheral nerve in
Sprague-Dawley rats exposed to aspartame via the diet
Group
I
II
III
IV
V
VI
VII
ppm in feed
0
80
400
2,000
10,000
50,000
100,000
Dose
(mg/kg-day)
0
4
20
100
500
2,500
5,000
Incidence and %
Combined hyperplasias,
papillomas and carcinomas
of the pelvis and ureter
Male
1/150 (0.7)
3/149 (2)
5/149 (3)
5/150 (3)
3/100 (3)
3/100 (3)
4/100 (4)
Female
2/150c (1.3)
6/150 (4)
9/1 50a (6)
10/1503 (7)
10/100b (10)
10/99b (10)
15/100b (15)
Peripheral nerve malignant
schwannomas
Male
1/150C (0.7)
1/150 (0.7)
3/150 (2)
2/150(1.3)
2/100 (2)
3/100 (3)
4/100 (4)
Female
0/150 (0)
2/150(1.3)
0/150 (0)
3/150 (2)
1/100 (1)
1/100 (1)
2/100 (2)
ap < 0.05 using the poly-k test.
bp < 0.01 using the poly-k test.
°p < 0.05 using the Cochran-Armitage test for trend.
Source: Soffritti et al. (2006).
1 In their conclusions, the EPS A review panel took note of negative results of 2-year
2 carcinogenic studies of aspartame (Ishii et et al., 1981; Ishii, 1981; NTP, 2003) and of the
3 findings of a recent epidemiological study carried out by the US National Cancer Institute (NCI,
4 2006).
5 In an effort to further clarify these issues, Soffritti et al. (2007) reported the results of
6 another lifetime study of aspartame in which 95 controls and 70 Sprague-Dawley rats/sex/group
7 were exposed, first in utero, then via the diet, to aspartame at concentrations of 0, 400, or
8 2,000 ppm (mg/kg) of feed. The authors assumed an average food consumption of 20 g/day and
9 an average body weight (males and females) of 400 g, thereby deriving average target doses of 0,
10 40, and 200 mg/kg-day. Soffritti et al. (2007) began administering the aspartame-supplemented
11 feed to the dams on GDI2; and offspring received feed containing aspartame at the appropriate
12 concentration from weaning until natural death. Animals were observed three times daily
13 Monday-Friday, and twice daily on Saturdays, Sundays, and holidays. This regimen was both to
14 monitor clinical signs and to reduce the possibility of decedents undergoing autolysis before
15 discovery. As described by the authors, all deceased animals were refrigerated, then necropsied
16 no more than 19 hours after discovery.
17 Food and drinking water consumption was monitored once/day. Beginning at 6 weeks of
18 age, individual body weights were recorded once a week for 13 weeks, then every 2 weeks until
19 natural death. All animals were examined grossly every 2 weeks. After necropsy, tissues and
20 organs were sampled for histopathological processing and microscopic examination (including
4-86 DRAFT-DO NOT CITE OR QUOTE
-------
1 skin and subcutaneous tissue, mammary gland, brain, pituitary, Zymbal's gland, salivary gland,
2 Harderian gland, cranium, tongue, thyroid, parathyroid, pharynx, larynx, thymus and mediastinal
3 lymph nodes, trachea, lung and main stem bronchi, heart, diaphragm, liver, spleen, pancreas,
4 kidney, adrenal gland, esophagus, stomach, intestine, urinary bladder, prostate, vagina, gonads,
5 interscapular brown fat pads, subcutaneous and mesenteric lymph nodes), as were all
6 pathological lesions identified on gross necropsy.
7 There were no differences in food and water consumption or in body weights among the
8 dose groups. As illustrated graphically by the authors, there was little change in overall survival
9 rates. Discussion of the histopathological findings focused exclusively on the cancer outcomes.
10 The incidence of total malignant tumors was increased significantly in high-dose males
11 compared to controls (p < 0.01). The slight increase in the incidence of total malignant tumors in
12 females was not statistically significant (Table 4-29). With regard to the incidence of type- or
13 site-specific neoplasms, Soffritti et al. (2007) reported statistically significant increases
14 (calculated using the Cox regression model) in combined lymphomas and leukemias in both
15 sexes of Sprague-Dawley rats. In males, the most frequently observed histiotypes were
16 lymphoblastic lymphomas involving the lung and mediastinal peripheral nodes, while in females,
17 the most commonly observed lesions were lymphocytic lymphomas and lympho-immunoblastic
18 lymphomas involving the thymus, spleen, lung and peripheral nodes. There was also an increase
19 in the incidence of mammary gland carcinomas in female Sprague-Dawley rats. The incidences
20 of total malignant, mammary, and lymphocytic and leukocytic tumors, in comparison to
21 concurrent and the range of historical controls for combined lymphomas and leukemias and
22 mammary gland tumors observed at the ERF, are shown in Table 4-29.
Table 4-29. Incidence of tumors in Sprague-Dawley rats exposed to aspartame from
GD12 to natural death
Dose
(mg/kg-day)
Malignant tumors
Tumor-bearing
animals (percent)
Tumors/100
animals
Lymphomas/leukemias
Tumor-bearing animals
(percent)
Mammary carcinomas
Tumor-bearing animals
(percent)
Males
0
20
100
Historical controls
23/95 (24.2)
18/70 (25.7)
28/70 (40.0)a
ND
27.4
27.1
44.3
ND
9/95 (9.5)
11/70(15.7)
12/70 (17. l)b
8-31%
0/95 (0)
0/70 (0)
2/70 (2.9)
NR
Females
0
20
100
Historical controls
42/95 (44.2)
31/70(44.3)
37/70 (52.9)
ND
50.5
62.9
85.7
ND
12/95 (12.6)
12/70(17.1)
22/70 (33.4)a
7-18%
5/95 (5.3)
5/70(7.1)
1 1/70(1 5. 7)b
4-14%
4-87
DRAFT-DO NOT CITE OR QUOTE
-------
Dose
(mg/kg-day)
Malignant tumors
Tumor-bearing
animals (percent)
Tumors/100
animals
Lymphomas/leukemias
Tumor-bearing animals
(percent)
Mammary carcinomas
Tumor-bearing animals
(percent)
1
2
3
4
5
6
7
zp < 0.01 versus concurrent controls, as calculated by the authors using the Cox regression model.
bp < 0.05 versus concurrent controls, as calculated by the authors using the Cox regression model.
ND = no data.; NR = not reported.
Source: Soffritti et al. (2007).
With regard to target organs and tissues susceptible to aspartame carcinogenicity, Soffritti
et al. (2007) drew attention to the similar outcome of these results to those reported by Soffritti
et al. (2006, 2005). The authors suggested that the increased incidence in combined lymphomas
and leukemias in female Sprague-Dawley rats as compared to the earlier study was likely due to
the earlier exposure to aspartame experienced by the animals in the Soffritti et al. (2007) study
(prenatal and postnatal versus postnatal only). The authors provided a direct comparison of the
incidence of lymphomas/leukemias between the studies, as summarized in Table 4-30.
Table 4-30. Comparison of the incidence of combined lymphomas and leukemias in
female Sprague-Dawley rats exposed to aspartame in feed for a lifetime, either pre-
and postnatally or postnatally only
Dose (mg/kg-day)
0
20
100
Percent with lymphomas/leukemias
Pre-and postnatal exposure"
12.6
17.1
31.4
Postnatal exposure onlyb
8.7
20.0
18.7
9
10
11
12
13
14
15
16
17
18
19
20
Source: a Soffritti et al. (2007); b Soffritti et al. (2006, 2005).
An EPS A (2009) review of the Sofftirti et al. (2007) study notes that the ratio between the
incidence in the low-dose group and the incidence in the concurrent control in Table 4-30 is
considerably lower in the animals exposed prenatally (1.4:1) compared to those exposed
postnatally (2.3:1). The ratio in the groups exposed to 100 mg/kg bw/day relative to the
respective concurrent controls is only slightly higher in animals exposed prenatally (2.5:1)
compared to those exposed postnatally (2.2:1). EFSA also notes that the incidence of
lymphomas and leukemias in aspartame-receiving males of the Soffritti et al. (2007) study was
within the range of historical controls for these responses in Sprague-Dawley rats at the ERF.
For example, the percent incidence of combined lymphomas and leukemias in males exposed
pre- and postnatally to 100 mg/kg-day aspartame (17.1%) was within the range of historical
controls for this response (8-31%, with an overall mean of 20.6%). Soffritti et al. (2007)
acknowledge this fact, but reason that comparisons of potentially compound-associated
incidences of tumor formation to incidences in concurrent controls provide a more scientifically
DRAFT-DO NOT CITE OR QUOTE
-------
1 valid indicator of the tumorigenic impact of a chemical under investigation than comparisons to
2 historical control data.57 Furthermore, the incidence of combined lymphomas and leukemias in
3 the female rats in the high-dose group is well above the historical control range. Therefore,
4 Soffritti et al. (2007) concluded that their second experiment confirmed the carcinogenic
5 potential of aspartame in Sprague-Dawley rats observed in Soffritti et al. (2006, 2005). The high
6 and variable incidence of this tumor type in ERF controls remains a concern. However, the
7 results provide support for studies suggesting similar effects from methanol (Soffritti et al.,
8 2002a) since methanol is one of the degradation products of aspartame and appears to have
9 carcinogenic potential at some of the same target organs and tissues.
10 4.6.5.2.3. MTBE. In an experiment that also may be relevant to the carcinogenicity of
11 methanol, scientists at the ERF carried out a cancer bioassay on MTBE, in which the compound
12 was administered to 60 Sprague-Dawley rats/sex/group by gavage in olive oil at 0, 250, and
13 1,000 mg/kg-day, 4 days/week, for 104 weeks (Belpoggi et al., 1995). Doses adjusted to daily
14 dose were 0, 143, and 571 mg/kg-day. This experiment and its findings may relate to the
15 carcinogenicity of methanol, since methanol is one of several metabolites of MTBE (ATSDR,
16 1997). At the end of the exposure period, the animals were allowed to live out their "natural"
17 life, the last animal dying 166 weeks after the start of the experiment (at 174 weeks of age).
18 Mean daily feed and drinking water consumption were determined weekly for the first
19 13 weeks of the experiment, then every 2 weeks until 112 weeks of age. Individual body weights
20 were measured according to the same protocol, then every 8 weeks until the end of the
21 experiment. All animals were examined for gross lesions weekly for the first 13 weeks, then
22 every 2 weeks until term. All animals were examined grossly at death, then histopathologically
23 examined for a full suite of organs and tissues.
24 As described by the authors, there were no differences among the groups in body weight
25 and clinical signs of toxicity. Survival was dose-dependently reduced in female rats after
26 16 weeks of exposure. Paradoxically, survival was improved in high-dose males compared to
27 controls after 80 weeks. Although there were no noncarcinogenic effects of MTBE reported, a
28 number of benign and malignant tumors were identified, including tumor types that were not
29 observed in the ERF methanol study such as an increased incidence of testicular Ley dig tumors
30 in high-dose males and as determined by the authors, as well as a dose-related statistically
31 significant increase in lymphomas and leukemias in females. The incidences of these tumors
32 compared to the initial number of animals exposed and compared to those at risk at the time of
33 the first observed tumor formation are shown in Table 4-31.
57 There are also potential problems with the use of historical control information from a colony that has been
maintained for over three decades. Population sensitivity can and does change over time.
4-89 DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-31. Incidence of Leydig cell testicular tumors and combined lymphomas and
leukemias in Sprague-Dawley rats exposed to MTBE via gavage for 104 weeks
Duration-
adjusted dose
0
143
571
Leydig cell tumors
Number of males
Affected
2
2
11"
Initial
60
60
60
At
riskc
26
25
32
Combined lymphomas and leukemias
Number of males
Affected
10
9
7
Initial
60
60
60
At
Riskd
59
59
58
Number of females
Affected
2
6b
12b
Initial
60
60
60
At
Risk6
58
51
47
ap < 0.05 using prevalence analysis
bp < 0.01 using a log-ranked test.
0 Alive male rats at 96 weeks of age, when first Leydig cell tumor was observed.
d Alive male rats at 32 weeks of age, when first leukemia was observed.
e Alive female rats at 56 weeks of age, when first leukemia was observed.
Source: Belpoggietal. (1995).
1 The possible contribution of the metabolite methanol to the reported responses cannot be
2 quantified. It is also possible that the parent compound and/or one or more of MTBE's other
3 metabolites (e.g., tertiary butanol or formaldehyde) may be etiologically linked to the formation
4 of the identified neoplasms (Blancato et al., 2007).
5 4.6.5.2.4. Formaldehyde. Scientists at the ERF have carried out two long-term experiments on
6 the potential carcinogenicity of formaldehyde, which is itself a metabolite of methanol,
7 aspartame and MTBE. While the tumorigenic effects at the portal-of-entry (such as in the oral
8 cavity and GI tract, for oral studies) may lack relevance to the possible effects of metabolites
9 formed in situ following methanol exposure, systemic neoplasms such as lymphomas and
10 leukemias have been described for formaldehyde as well (Soffritti et al., 2002b, 1989). This
11 suggests that formaldehyde metabolized from methanol, aspartame and MTBE may be
12 etiologically important in the formation of lymphomas and leukemias in animals exposed to
13 these compounds.
14 In the first formaldehyde study (designated BT 7001; Soffritti et al., 1989), 50 Sprague-
15 Dawley rats/sex/group (starting at 7 weeks of age) were exposed to formaldehyde in drinking
16 water at concentrations of 10, 50, 500, 1,000, and 1,500 mg/L for 104 weeks. Another 50
17 Sprague-Dawley rats/sex received methanol in drinking water at 15 mg/L, and 100 rats/sex
18 received water only, as controls. Body weight and water and food consumption were monitored
19 weekly for the first 13 weeks, then every 2 weeks thereafter. All animals were allowed to live
20 out their "natural" life, at which point they were subjected to necropsy and a complete
21 histopathological examination.
22 The final results of the BT 7001 Soffritti et al. (1989) experiment were reported by
23 Soffriti et al. (2002b). Water consumption was reduced compared to controls in high-dose males
4-90 DRAFT-DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
and in females at the three highest doses. However, there appeared to be no evidence of
compound-related body weight changes, clinical signs of toxicity among the groups, nor
nononcogenic histopathological effects of formaldehyde. The authors noted statistically
significant increases in the incidence of tumor-bearing males at 1500 ppm (p<0.01) and in total
malignant tumors in females at 100, 1000 and 1500 ppm (p<0.01) and in males at 500 ppm
(p<0.05) and 1500 ppm (p<0.01). They reported statistically significant increases in malignant
mammary tumors in females at 100 ppm (p<0.01) and 1500 ppm (p<0.05), and in testicular
interstitial cell adenomas in the 1000 ppm males (p<0.05). They also noted sporadic incidences
in the treatment groups only (primarily at the highest dose) of leiomyosarcomas of the stomach
and intestine considered to be very rare for the ERF rat colony. As for methanol and the other
compounds discussed in this section, they reported increases in the number of
hemolymphoreticular tumors for both sexes. The incidence of hemolymphoreticular neoplasms
among the dose groups is shown in Table 4-32.
Table 4-32. Incidence of hemolymphoreticular neoplasms on Sprague-Dawley rats
exposed to formaldehyde in drinking water for 104 weeks
Concentration in drinking water (mg/L)
0
0(1 5 mg/L methanol)
10
50
100
500
1,000
1,500
Males
8/100
10/50
4/50
10/50
13/50b
12/50a
ll/50a
23/50b
Females
7/100
5/50
5/50
7/50
8/50
7/50
ll/50a
10/50b
14
15
16
17
18
19
20
21
22
23
24
"p < 0.05 using the tf test; bp < 0.01 using the x test.
Source: Soffritti et al. (2002b).
Soffritti et al. (1989) also described the results of another experiment (BT 7005) in which
approximately 20 Sprague-Dawley rats/sex/group were exposed to either regular drinking water
or 2,500 mg/L formaldehyde, beginning at 25 weeks of age for 104 weeks. These animals were
allowed to mate and approximately 40-60 of the FO pups were likewise exposed to 0 or
2,500 ppm formaldehyde in drinking water (after weaning) for 104 weeks. As before, parents
and progeny lived out their normal life span but then were subjected to a complete
histopathological examination. Incidence of leukemias in exposed breeders and offspring is
shown in Table 4-33. The authors considered this data to indicate a "slight" increase in
leukemias in breeders at 2,500 ppm, but the changes did not achieve statistical significance.
4-91
DRAFT-DO NOT CITE OR QUOTE
-------
Table 4-33. Incidence of leukemias in breeder and offspring Sprague-Dawley rats
exposed to formaldehyde in drinking water for 104 weeks (Test BT 7005)
Concentration (ppm)
0
2,500
Incidence of leukemias
Breeder
Males
0/20
2/18
Females
1/20
2/18
Offspring
Males
3/59
4/36
Females
3/49
0/37
Source: Soffrittietal. (1989).
4.7. SYNTHESIS OF MAJOR NONCANCER EFFECTS
4.7.1. Summary of Key Studies in Methanol Toxicity
1 A substantial body of information exists on the toxicological consequences to humans
2 who consume or are acutely exposed to large amounts of methanol. Neurological and
3 immunological effects have been noted in adult human subjects acutely exposed to as low as
4 200 ppm (262 mg/m3) methanol (Mann et al., 2002; Chuwers et al., 1995). Nasal irritation
5 effects have been reported by adult workers exposed to 459 ppm (601 mg/m3) methanol. Frank
6 effects such as blurred vision and bilateral or unilateral blindness, coma, convulsions/tremors,
7 nausea, headache, abdominal pain, diminished motor skills, acidosis, and dyspnea begin to occur
8 as blood levels approach 200 mg methanol/L, and 800 mg/L appears to be the threshold for
9 lethality. Data for subchronic, chronic or in utero human exposures are very limited.
10 Determinations regarding longer term effects of methanol are based primarily on animal studies.
11 An end-point-by-end-point survey of the primary effects of methanol in experimental
12 animals is given in the following paragraphs. Tabular summaries of the principal toxicological
13 studies that have examined the impacts of methanol when experimental animals were exposed to
14 methanol via the oral or inhalation routes are provided in Tables 4-34 and 4-35. Most studies
15 focused on developmental and reproductive effects. A large number of the available studies were
16 performed by routes of exposure (e.g., i.p.) that are less relevant to the assessment. The data are
17 summarized separate sections that address oral exposure (Section 4.7.1.1) and inhalation
18 exposure (Section 4.7.1.2).
Table 4-34. Summary of studies of methanol toxicity in experimental animals (oral)
Species, strain,
number/sex
Rat
Sprague-Dawley
30/sex/group
Dose/duration
0, 100, 500, and
2,500 mg/kg-day for
13 wk
NOAEL
(mg/kg-day)
500
LOAEL
(mg/kg-day)
2,500
Effect
Reduction of brain
weights, increase in the
serum activity of ALT
and AP. Increased liver
weights
Reference
U.S. EPA
(1986c)
4-92
DRAFT-DO NOT CITE OR QUOTE
-------
Species, strain,
number/sex
Rat
Sprague-Dawley
100/sex/group
Mouse
Swiss
Dose/duration
0, 500, 5,000, or
20,000 ppm (v/v) in
drinking water, for
104 wk. Doses were
approx. 0, 46.6, 466,
and 1,872 mg/kg-day
(male) and 0, 52.9,
529, and 2101 mg/kg-
day (female)
560, 1000 and 2 100
mg/kg/d (female)
and 550, 970, and
1800 mg/kg/d (male),
6 days/wk for life
NOAEL
(mg/kg-day)
466-529
ND
LOAEL
(mg/kg-day)
1,872-2,101
1,800-2,100
Effect
Increased incidence of
ear duct3 carcinomas,
lymphoreticular
tumors, and total
malignant tumors. No
noncancer effects
Increased incidence of
liver parenchymal cell
necrosis and malignant
lymphomas
Reference
Soffritti et al.
(2002a)
Apaja (1980)
Reproductive/developmental toxicity studies
Rat
Long-Evans
10 pregnant
females/group
Mouse
CD-I
8 pregnant
females and 4
controls
0 and 2,500 mg/kg-
day on either GD15-
GD17orGD17-
GD19.
4 g/kg-day in 2 daily
dosesonGD6-GD15
NA
NA
2,500
4,000
Neurobehavioral
deficits (such as
homing behavior,
suckling ability
Increased incidence of
totally resorbed litters,
cleft palate and
exencephaly. A
decrease in the number
of live fetuses/litter
Infurna and
Weiss (1986)
Rogers, et al.
(1993a)
NA = Not applicable; ND = Not determined; M= male, F=female.
aln an NTP evaluation of pathology slides from another bioassay from this laboratory in which a similar ear duct
carcinoma finding was reported (Soffritti et al., 2006, 2005), NTP pathologists interpreted a majority of these ear
duct responses as being hyperplastic, not carcinogenic, in nature (EFSA, 2006; Hailey, 2004).
Table 4-35. Summary of studies of methanol toxicity in experimental animals
(inhalation exposure)
Species, strain,
number/sex
Monkey
M. fascicularis,
lor 2
animals/group
Dog (2)
Rat
Sprague-Dawley
5 males/ group
Dose/duration
0, 3,000, 5,000,
7,000, or 10,000
ppm, 21 hr/day, for
up to 14 days
10,000 ppm for 3
min, 8 times/day for
100 days
0, 200, 2000, or
10,000 ppm, 8
hr/day, 5 days/wk for
up to 6 wk
NOAEL
(ppm)
ND
NA
NA
LOAEL
(ppm)
ND
NA
200
Effect
Clinical signs of toxicity, CNS
changes, including degeneration
of the bilateral putamen, caudate
nucleus, and claustrum. Edema
of cerebral white matter.
None
Transient reduction in plasma
testosterone levels
Reference
NEDO (1987)
Sayers et al.
(1944)
Cameron et al.
(1984)
4-93
DRAFT-DO NOT CITE OR QUOTE
-------
Species, strain,
number/sex
Rat
Sprague-Dawley
5 males/ group
Rat
Sprague-Dawley
5/sex/group
Monkey
M. fascicularis
3/sex/group
Rat
Sprague-Dawley
10/sex/group
Rat
Sprague-Dawley
1 5/sex/group
Monkey
M. fascicularis
2 or 3 animals/
group/time point
Rat
F344
20/sex/group
Mouse
B6C3FJ
30/sex/group
Mouse
B6C3FJ
52-53/sex/group
Rat
F344
52/sex/group
Dose/duration
0, or 200 ppm,
6 hr/day, for either 1
or 7 days
0, 500, 2,000, or
5,000 ppm,
5 days/wk for 4 wk
0, 500, 2,000, or
5,000 ppm,
5 days/wk for 4 wk
0,300, or 3, 000 ppm,
6 hr/day, 5 days/wk
for 4 wk
0 or 2,500 ppm, 6
hr/day, 5 days/wk for
4wk
0, 10, 100, or
1,000 ppm, 21 hr/day
for either 7, 19, or 29
mo
0, 10, 100, or
1,000 ppm,
20 hr/day, for 12 mo
0, 10, 100, or
1000 ppm, 20 hr/day,
for 12 mo
0, 10, 100, or
1,000 ppm,
20 hr/day, for 12 mo
0, 10, 100, or
1,000 ppm,
-20 hr/day for 2 yr
NOAEL
(ppm)
NA
5,000
5,000
NA
NA
ND
ND
NA
NA
100
100
LOAEL
(ppm)
200
NA
NA
300
2,500
ND
ND
NA
NA
1,000
1,000
Effect
Transient reduction in plasma
testosterone levels
No compound-related effects
No compound-related effects
Reduction in size of thyroid
follicles
Reduction of relative spleen
weight in females,
histopathologic changes to the
liver, irritation of the upper
respiratory tract
Limited fibrosis of the liver
Possible myocardial and renal
effects
No compound-related effects
No clear-cut compound-related
effects
Increase in kidney weight,
decrease in testis and spleen
weights
Fluctuations in a number of
urinalysis, hematology, and
clinical chemistry parameters.
Development of pulmonary
adenoma/adenocarcinoma
(males), pheochromocytomas
(females)
Reference
Cameron et al.
(1985)
Andrews et al.
(1987)
Poon et al.
(1994)
Poon et al.
(1995)
NEDO (1987)
Reproductive/developmental toxicity studies
Rat
Sprague-Dawley
15/pregnant
females/group
Rat
Sprague-Dawley
36/pregnant
females/group
0, 5,000, 10,000, or
20,000 ppm, 7 hr/day
on either GDI-GDI 9
orGD7-GD15.
0, 200, 1,000, or
5000 ppm,
22.7 hr/day, on GD7-
GD17
5,000
1,000
10,000
5,000
Reduced fetal body weight,
increased incidence of visceral
and skeletal abnormalities,
including rudimentary and extra
cervical ribs
Late-term resorptions, reduced
fetal viability, increased
frequency of fetal
malformations, variations and
delayed ossifications.
Nelson et al.
(1985)
NEDO (1987)
4-94
DRAFT-DO NOT CITE OR QUOTE
-------
Species, strain,
number/sex
Rat
Sprague-Dawley
FI and F2
generations of a
two-generation
study
Rat
Sprague-Dawley
Follow-up study of
brain weights in FI
generation of 10-
14/sex/group in FI
generation
Mouse
CD-I
30-1 14 pregnant
females/group
Mouse
CD-I
12-17 pregnant
females/group
Rat
Long-Evans
6-7 pregnant
females/group
Rat
Long-Evans
10-12 pregnant
females/group
Monkey
M. fascicularis
12 monkeys/group
Dose/duration
0, 10, 100, or
1000 ppm, 20 hr/day;
FI- birth to end of
mating (M) or
weaning (F); F2-
birth to 8 wks
0, 500, 1,000, and
2,000 ppm; GDO
through F! generation
0, 1,000, 2,000,
5,000, 7,500, 10,000,
or 15,000 ppm,
7 hr/day on GD6-
GD15.
0 and 10,000 ppm on
two consecutive days
during GD6-GD 13
or on a single day
during GD5-GD9
0 or 15,000 ppm,
7 hr/day on GD7-
GD19
0 or 4,500 ppm from
GD10toPND21.
0, 200, 600, or
1800 ppm, 2.5
hr/day, 7 days/wk,
during premating,
mating and gestation
NOAEL
(ppm)
100
500
1,000
NA
NA
NA
ND
LOAEL
(ppm)
1,000
1,000
2,000
10,000
15,000
4,500
NDa
Effect
Reduced weight of brain,
pituitary, and thymus at 8, 16
and 24 wk postnatal in F! and at
8wkinF2
Reduced brain weight at 3 wk
and 6 wk (males only). Reduced
brain and cerebrum weight at 8
wk (males only)
Increased incidence of extra
cervical ribs, cleft palate,
exencephaly; reduced fetal
weight and pup survival,
Delayed ossification
Cleft palate, exencephaly,
skeletal malformations
Reduced pup weight
Subtle cognitive deficits
Shortened period of gestation;
may be related to exposure (no
dose-response),
neurotoxicological deficits
including reduced performance
in the VDR test; may be related
to premature births.
Reference
Rogers et al.
(1993a)
Rogers and
Mole (1997)
Stanton et al.
(1995)
Weiss et al.
(1996)
Burbacher
et al. (20004a,
2004b, 1999a,
1999b)
1
2
3
4
ND = Not determined due to study limitations such as small number of animals /time point/ exposure level
NA = Not applicable.
aGestation resulted in a shorter period of gestation in dams exposed to as low as 200 ppm (263 mg/m3). However,
because of uncertainties associated with these results, including clinical intervention and the lack of a dose-
response, EPA was not able to identify a definitive NOAEL or LOAEL from this study.
4.7.1.1. Oral
There have been very few subchronic, chronic, or in utero experimental studies of oral
methanol toxicity. In one such experiment, an EPA-sponsored 90-day gavage study in Sprague-
Dawley rats suggested a possible effect of the compound on the liver (U.S. EPA, 1986c). In the
absence of gross or histopathologic evidence of toxicity, fluctuations on some clinical chemistry
4-95 DRAFT-DO NOT CITE OR QUOTE
-------
1 markers of liver biochemistry and increases in liver weights at the highest administered dose
2 (2,500 mg/kg-day) justify the selection of the mid-dose level (500 mg/kg-day) as a NOAEL for
3 this effect under the operative experimental conditions. That the bolus effect may have been
4 important in the induction of those few effects that were apparent in the subchronic study is
5 suggested by the outcome of lifetime drinking water study of methanol that was carried out in
6 Sprague-Dawley rats by Soffritti et al. (2002a). According to the authors, no noncancer
7 toxicological effects of methanol were observed at drinking water concentrations of up to
8 20,000 ppm (v/v). Based on default assumptions on drinking water consumption and body
9 weight gain assumptions, the high concentration was equivalent to a dose of 1,780 mg/kg-day in
10 males and 2,177 mg/kg-day in females. In the stated absence of any changes to parameters
11 reflective of liver toxicity in the Soffritti et al. (2002a) study, the slight impacts to the liver
12 observed in the subchronic study at 2,500 mg/kg-day suggest the latter dose to be a minimal
13 LOAEL. Logically, the true but unknown threshold would at the high end of the range from 500
14 (the default NOAEL) to 2,500 mg/kg-day for liver toxicity via oral gavage.
15 Two studies have pointed to the likelihood that oral exposure to methanol is associated
16 with developmental neurotoxicity or developmental deficits. When Infurna and Weiss (1986)
17 exposed pregnant Long-Evans rats to 2% methanol in drinking water (providing a dose of
18 approximately 2,500 mg/kg-day), they observed no reproductive or developmental sequelae
19 other than from 2 tests within a battery of fetal behavioral tests (deficits in suckling ability and
20 homing behavior). In the oral section of the Rogers et al. (1993a) study, such teratological
21 effects as cleft palate and exencephaly and skeletal malformations were observed in fetuses of
22 pregnant female mice exposed to daily gavage doses of 4,000 mg/kg methanol during GD6-
23 GDIS. Likewise, an increase in totally resorbed litters and a decrease in the number of live
24 fetuses/litter appear likely to have been an effect of the compound. Similar skeletal
25 malformations were observed by Rogers and Mole (1997), Rogers et al. (1993a), and Nelson
26 et al. (1985) following inhalation exposure.
4.7.1.2. Inhalation
27 Some clinical signs, gross pathology, and histopathological effects of methanol have been
28 seen in experimental animals including adult nonhuman primates exposed to methanol vapor.
29 Results from an unpublished study (NEDO, 1987) of M. fasciculoris monkeys, chronically
30 exposed to concentrations as low as 10 ppm for up to 29 months, resulted in histopathological
31 effects in the liver, kidney, brain and peripheral nervous system. These results were generally
32 reported as subtle and do not support a robust dose response over the range of exposure levels
33 used. Confidence in the methanol-induced findings of effects in adult nonhuman primates is
34 limited because this study utilized a small number (2-3) of animals/dose level/time of sacrifice
35 and inadequately reporting of results (i.e., lack of clear documentation of a concurrent control
4-96 DRAFT-DO NOT CITE OR QUOTE
-------
1 group). In addition, the monkeys used in this study were all wild-caught. All of these concerns
2 limit the study's utility in derivation of an RfC.
3 A number of studies have examined the potential toxicity of methanol to the male
4 reproductive system (Lee et al., 1991; Cameron et al., 1985, 1984). The data from Cameron
5 et al. (1985, 1984) showed a transient but not necessarily dose-related decrease in serum
6 testosterone levels of male Sprague-Dawley rats. Lee et al. (1991) reported the appearance of
7 testicular lesions in 18-month-old male Long-Evans rats that were exposed to methanol for
8 13 weeks and maintained on folate-deficient diets. Taken together, the Lee et al. (1991) and
9 Cameron et al. (1985, 1984) study results could indicate chemically-related strain on the rat
10 system as it attempts to maintain hormone homeostasis. However, the available data are
11 insufficient to definitively characterize methanol as a toxicant to the male reproductive system.
12 When Sprague-Dawley rats were exposed to methanol, 6 hours/day for 4 weeks, there
13 were some signs of irritation to the eyes and nose. Mild changes to the upper respiratory tract
14 were also described in Sprague-Dawley rats that were exposed for 4 weeks to up to 300 ppm
15 methanol (Poon et al., 1995). Other possible effects of methanol in rats included a reduction in
16 size of thyroid follicles (Poon et al., 1994), panlobular vacuolation of the liver, and a decrease in
17 spleen weight (Poon et al., 1995). NEDO (1987) reported dose-related increases in moderate
18 fatty degeneration in hepatocytes of male mice exposed via inhalation for 12 months, but this
19 finding was not observed in the NEDO (1987) 18-month mouse inhalation study. Nodes
20 reported in the liver of mice from the 18-month study may have been precancerous, but the
21 18-month study duration was not of sufficient duration to make a determination.
22 One of the most definitive and quantifiable toxicological impacts of methanol when
23 administered to experimental animals via inhalation is related to the induction of developmental
24 abnormalities in fetuses exposed to the compound in utero. Developmental effects have been
25 demonstrated in a number of species, including monkeys, but particularly rats and mice. Most
26 developmental teratological effects appear to be more severe in the latter species. For example,
27 in the study of Rogers et al. (1993 a) in which pregnant female CD-I mice were exposed to
28 methanol vapors on GD6-GD15 at a range of concentrations, reproductive and fetal effects
29 included an increase in the number of resorbed litters, a reduction in the number of live pups, and
30 increased incidence of exencephaly, cleft palate, and the number of cervical ribs. While the
31 biological significance of the cervical rib effect has been the subject of much debate (See
32 discussion of Chernoff and Rogers [2004] in Section 5), it appears to be the most sensitive
33 indicator of developmental toxicity from this study, with a NOAEL of 1,000 ppm (1,310 mg/m3).
34 In rats, however, the most sensitive developmental effect, as reported in the NEDO (1987) two-
35 generation inhalation studies, was a postnatal reduction in brain weight at 3, 6 and 8 weeks
36 postnatally, which was significantly lower than controls when pups and their dams were exposed
4-97 DRAFT-DO NOT CITE OR QUOTE
-------
1 to 1,000 ppm (1,310 mg/m3) during gestation and throughout lactation. The NOAEL reported in
2 this study was 500 ppm (655 mg/m3).
3 Rogers and Mole (1997) addressed the question of which period of gestation was most
4 critical for the adverse developmental effects of methanol in CD-I rats. Such malformations and
5 anomalies as cleft palate, exencephaly, and a range of skeletal defects, appeared to be induced
6 with a greater incidence when the dams were exposed on or around GD6. These findings were
7 taken to indicate that methanol is most toxic to embryos during gastrulation and in the early
8 stages of organogenesis. However, NEDO (1987) gestation-only and two-generation studies
9 showed that significant reductions in brain weight were observed at a lower exposure levels
10 when pups and their dams were exposed during lactation as well as gestation, indicating that
11 exposure during the later stages of organogenesis, including postnatal development, can
12 significantly contribute to the severity of the effects in this late-developing organ system.
13 In comparing the toxicity (NOAELs and LOAELs) for the onset of developmental effects
14 in mice and rats exposed in utero, there is suggestive evidence from the above studies that mice
15 may be more susceptible to methanol than rats. Supporting evidence for this proposition has
16 come from in vitro studies in which rat and mouse embryos were exposed to methanol in culture
17 (Andrews et al., 1993). Further evidence for species-by-species variations in the susceptibility of
18 experimental animals to methanol during organogenesis has come from experiments on monkeys
19 (Burbacher et al., 2004a, 2004b, 1999a,1999b). In these studies, exposure of monkeys to
20 methanol during premating, mating, and throughout gestation resulted in a shorter period of
21 gestation in dams exposed to as low as 200 ppm (263 mg/m3). The shortened gestation period
22 was largely the result of C-sections performed in the methanol-exposure groups "in response to
23 signs of possible difficulty in the maintenance of pregnancy," including vaginal bleeding.
24 Though statistically significant, the finding of a shortened gestation length may be of limited
25 biological significance. Gestational age, birth weight and infant size observations in all exposure
26 groups were within normal ranges for M. fascicularis monkeys, and vaginal bleeding 1-4 days
27 prior to delivery of a healthy infant does not necessarily imply a risk to the fetus (as cited in
28 CERHR, 2004). An ultrasound examination could have substantiated fetal or placental problems
29 arising from presumptive pregnancy duress (see Section 4.3.2). As discussed in Section 4.4.2,
30 there is also evidence from this study that methanol caused neurobehavioral effects in exposed
31 monkey infants that may be related to the gestational exposure. However, the data are not
32 conclusive,and a dose-response trend is not robust. There is insufficient evidence to determine if
33 the primate fetus is more or less sensitive than rodents to methanol teratogenesis. Several other
34 uncertainties contributed to decreased confidence in the use of this primate in quantitative
35 estimates of risk. These included: a mixture of wild- and colony-derived monkey mothers used
36 in the study; the use of a cohort design necessitated by the complexity of this study also
4-98 DRAFT-DO NOT CITE OR QUOTE
-------
1 seemingly resulted in limitations in power to detect effects (e.g., Pagan test results for controls);
2 and no apparent adjustment in statistical analysis for results from the neurobehavioral battery of
3 tests employed leading to concern about inflation of type 1 error. Because of the uncertainties
4 associated with these results, including the fact that the decrease in gestational length was not
5 exacerbated with increasing methanol exposure, EPA was not able to identify a definitive
6 NOAEL or LOAEL from this study. This study does support the weight of evidence for
7 developmental neurotoxicity in the hazard characterization of low-level methanol exposure.
8 Weiss et al. (1996) and Stanton et al. (1995) evaluated the developmental and
9 developmental neurotoxicological effects of methanol exposure on pregnant female Long-Evans
10 rats and their progeny. In the former study, exposure of dams to 15,000 ppm (19,656 mg/m3),
11 7 hours/day on GD7-GD19 resulted in reduced weight gain in pups, but produced little other
12 evidence of adverse developmental effects. The authors subjected the pups to a number of
13 neurobehavioral tests that gave little if any indication of compound-related changes. This study,
14 while using high exposure levels, was limited in its power to detect effects due to the small
15 number of animals used. In the Weiss et al. (1996) study, exposure of pregnant female Long-
16 Evans rats to 0 or 4,500 (0 and 5,897 mg/m3) methanol from GD6 to PND21 likewise provided
17 fluctuating and inconsistent results in a number of neurobehavioral tests that did not necessarily
18 indicate any compound-related impacts. The finding of this study indicated subtle cognitive
19 defects not on the learning of an operant task but in the reversal learning. This study also
20 reported exposure-related changes in neurodevelopmental markers of NCAMs on PND4.
21 NCAMs are a family of glycoproteins that is needed for migration, axonal outgrowth, and
22 establishment of the pattern for mature neuronal function.
23 Taking all of these findings into consideration reinforces the conclusion that the most
24 appropriate endpoints for use in the derivation of an RfC for methanol are associated with
25 developmental neurotoxicity and developmental toxicity. Among an array of findings indicating
26 developmental neurotoxicity and developmental malformations and anomalies that have been
27 observed in the fetuses and pups of exposed dams, an increase in the incidence of cervical ribs of
28 gestationally exposed mice (Rogers et al., 1993a) and a decrease in the brain weights of
29 gestationally and lactationally exposed rats (NEDO, 1987) appear to be the most robust and most
30 sensitive effects.
4.8. NONCANCER MO A INFORMATION
31 A review by Jacobsen and McMartin (1986) has provided a comprehensive summary of
32 the mechanism by which methanol brings about its acute toxic effects. Overwhelmingly, the
33 evidence points to methanol poisoning being a consequence of formate accumulation. This
34 compound is formed from formaldehyde under the action of ADH3. Formaldehyde itself is
4-99 DRAFT-DO NOT CITE OR QUOTE
-------
1 formed from methanol under the action of ADH1. Evidence for the involvement of formate
2 comes from the delay in the onset of harmful symptoms, detection of formate in the blood
3 stream, and the profound acidosis that develops 12-24 hours after exposure to methanol.
4 Treatments for methanol poisoning include the i.v. administration of buffer to correct the
5 acidosis, hemodialysis to remove methanol from the blood stream, and i.v. administration of
6 either ethanol or fomepizole to inhibit the activity of ADH1. Therapies to increase endogenous
7 levels of folate may enhance the activity of THF synthetase, an enzyme that catalyzes the
8 oxidation of formate to CC>2. Jacobsen and McMartin (1986) have drawn attention to the
9 accumulation of lactate in advanced stages of severe methanol poisoning, a possible consequence
10 of formate inhibition of mitochondrial respiration and tissue hypoxia. The additional decrease in
11 blood pH is likely to enhance the nonionic diffusion of formic acid across cell membranes, with
12 resulting CNS-depression, hypotension, and further lactate production.
13 Jacobsen and McMartin (1986) summarized a body of evidence that also points to the
14 formate-related acidosis as the etiologically important factor in ocular damage. The hypothesis
15 suggests that ocular toxicity is due to the inhibition of cytochrome oxidase in the optic nerve by
16 formate. This would cause inhibition of ATP formation and consequent disruption of optic nerve
17 function.
18 While it is well established that the toxic consequences of acute methanol poisoning arise
19 from the action of formate, there is less certainty on how the toxicological impacts of longer-
20 term exposure to lower levels of methanol are brought about. For example, since developmental
21 effects in experimental animals appear to be significant adverse effects associated with in utero
22 methanol exposure, it is important to determine potential MO As for how these specific effects
23 are brought about.
24 As described in Section 4.6.1, data from experiments carried out by Dorman et al. (1995),
25 formate is not the probable proximate teratogen in pregnant CD-I mice exposed to high
26 concentrations of methanol vapor. This conclusion is based on the fact that there appeared to be
27 little, if any, accumulation of formate in the blood of methanol-exposed mice, and exencephaly
28 did not occur until formate levels were grossly elevated. Another line of argument is based on
29 the observation that treatment of pregnant mice with a high oral dose of formate did not induce
30 neural tube closure defects at media concentrations comparable to those observed in uterine
31 decidual swelling after maternal exposure to methanol. Lastly, methanol- but not formate-
32 induced neural tube closure defects in mouse embryos in vitro at media concentrations
33 comparable to the levels of methanol detected in blood after a teratogenic exposure.
34 Harris et. al (Hansen et al., 2005; Harris et al., 2004, 2003) carried out a series of
35 physiological and biochemical experiments on mouse and rat embryos exposed to methanol,
36 formaldehyde and formate, concluding that the etiologically important substance for embryo
4-100 DRAFT-DO NOT CITE OR QUOTE
-------
1 dysmorphogenesis and embryolethality was likely to be formaldehyde rather than the parent
2 compound or formate. Specific activities for enzymes involved in methanol metabolism were
3 determined in rat and mouse embryos during the organogenesis period of 8-25 somites (Harris
4 et al., 2003). The experiment was based on the concept that differences in the metabolism of
5 methanol to formaldehyde and formic acid by the enzymes ADH1, ADH3, and CAT may
6 contribute to hypothesized differences in species sensitivity that were apparent in toxicological
7 studies. A key finding was that the activity of ADH3 (converting formaldehyde to formate) was
8 lower in mouse VYS than that of rats throughout organogenesis, consistent with the greater
9 sensitivity of the mouse to the developmental effects of methanol exposure. Another study
10 (Harris et al., 2004) which showed that the inhibition of GSH synthesis increases the
11 developmental toxicity of methanol also lends support to this hypothesis because ADH3-
12 mediated metabolism of formaldehyde is the only enzyme involved in methanol clearance that is
13 GSH-dependent. These findings provide inferential evidence for the proposition that
14 formaldehyde may be the ultimate teratogen through diminished ADH3 activity. This concept is
15 further supported by the demonstration that the LOAELs for the embryotoxic effects of
16 formaldehyde in rat and mouse embryos were much lower than those for formate and methanol
17 (Hansen et al., 2005). Taking findings from both sets of experiments together, Harris et. al.
18 (Hansen et al., 2005; Harris et al., 2004, 2003) concluded that the demonstrable lower capacity of
19 mouse embryos to transform formaldehyde to formate (by ADH3) could explain the increased
20 susceptibility of mouse versus rat embryos to the toxic effects of methanol.
21 While studies such as those by Harris et al. (2004, 2003) and Dorman et al. (1995, 1994)
22 strongly suggest that formate is not the metabolite responsible for methanol's teratogenic effects,
23 there are still questions regarding the relative involvement of methanol versus formaldehyde. In
24 vitro evidence suggests that formaldehyde is the more embryotoxic moiety, but methanol would
25 likely play a prominent role, at least in terms of transport to the target tissue. The high reactivity
26 of formaldehyde would limit its unbound and unaltered transport as free formaldehyde from
27 maternal to fetal blood (Thrasher and Kilburn, 2001), and the capacity for the metabolism of
28 methanol to formaldehyde is likely lower in the fetus and neonate versus adults (see discussion
29 in Section 3.3)
30 In humans, metabolism of methanol occurs primarily through ADH1, whereas in rodents
31 methanol clearance involves primarily CAT, as well as ADH1. There are no known studies that
32 compare enzyme activities of human ADH1 and rodent CAT. Assuming that relative expression
33 and activity of ADH1 is comparable across species, rodents are expected to clear methanol more
34 rapidly than humans due to involvement of CAT. In fact, even among rodents the metabolism of
35 methanol may be quite different, as one study has demonstrated that the rate of methanol
36 oxidation in mice is twice the rate in rats, as well as nonhuman primates (Mannering et al.,
4-101 DRAFT-DO NOT CITE OR QUOTE
-------
1 1969). Despite a faster rate of methanol metabolism, mice have consistently shown higher blood
2 methanol levels than rats following exposure to equivalent concentrations (Tables 3-4 and 3-5).
3 A faster respiration rate and increased fraction of absorption by mice is thought to be the reason
4 for the higher blood methanol levels compared to rats (Perkins et al., 1995a). Using the exposure
5 conditions of Horton et al. (1992) for rats, when the respiration rate scaling coefficient (QPC)
6 was increased from the rat value of 16.4 to the mouse value of 25.4 while holding all other
7 parameters constant, peak blood concentrations were predicted by the PBPK model to increase
8 by 1.4-fold at 200 ppm and 1.8-fold at 2,000 ppm (where metabolism is becoming saturated).
9 Because smaller species generally have faster breathing rates than larger species (in the PBPK
10 model, the respiration rate/BW is 3 times slower in humans versus rats and almost 10 times
11 slower versus mice), humans would be expected to accumulate less methanol than rats or mice
12 inhaling equivalent concentrations and given the same metabolism rate. However, Horton et al.
13 (1992) measured a blood concentration in rats exposed to 200 ppm methanol of about 3.7 mg/L
14 after 6 hours of exposure while Sedevic et al. (1981) measured around 5.5 mg/L in human
15 volunteers after 6 hours of exposure to 231 ppm. Correcting for the higher exposure, human
16 blood concentrations would be around 4.8 mg/L if exposed at 200 ppm. Simulations with the
17 mouse model predict a blood level of 5.7 mg/L after 6 hours of exposure to 200 ppm, only 20%
18 higher than this interpolated human value. Thus the slower inhalation rate in humans is offset by
19 the slower metabolic rate, leading to equivalent blood concentrations. (If the same rate of
20 metabolism/BW as mice is used, human blood concentrations are predicted to decrease by
21 approximately fivefold.). These differences are considered in Section 5 for the characterization
22 of human and rodent PBPK models used for the derivation of human equivalent concentrations
23 (HECs).
4.9. EVALUATION OF CARCINOGENICITY
4.9.1. Summary of Overall Weight-of-Evidence
24 Under the current Guidelines for Carcinogen Risk Assessment (U.S. EPA 2005a, 2005b),
25 methanol is likely to be carcinogenic to humans by all routes of exposure based on dose-
26 dependent trends in multiple tumors in both sexes of two strains of rats, by inhalation and oral
27 routes of exposure and increases in malignant lymphoma in both sexes of Swiss mice by oral
28 exposure. Specifically, EPA's analysis of the Soffritti et al. (2002a) lifespan study of Sprague-
29 Dawley rats exposed to methanol in drinking water for 104 weeks indicates a statistically
30 significant increase in the incidence of lymphoma58 in lung and other organs at the two highest
58 Combining lymphoblastic lymphomas, lymphocytic lymphomas, lympho-immunoblastic lymphomas and/or
lymphoblastic leukemias as malignant lymphomas but excluding myeloid leukemias, histocytic sarcomas and
monocytic leukemia as tumors of different origin (Cruzan, 2009; Hailey, 2004; McConnell et al., 1986).
4-102 DRAFT-DO NOT CITE OR QUOTE
-------
1 doses for males and across all doses for females (Fisher's exact, p< 0.05) and a statistically
2 significant increase in relatively rare hepatocellular carcinomas in males compared to historical
3 controls (n=407)59 (Fisher's exact/? < 0.05 for all doses and/? < 0.01 for the high-dose group).
4 Statistically significant increases in the incidence of malignant lymphomas relative to historical
5 controls (Fisher's exact,/? < 0.05) have also been observed in another rodent species, Swiss
6 mice, following similar mg/kg-day exposures to methanol in drinking water for life (Apaja,
7 1980). The only available chronic inhalation studies of methanol (NEDO, 1985/2008a, 2008b)
8 reported slight but statistically significant tumor responses in F344 rats at 24 months, and no
9 evidence of carcinogenicity in B6C3Fi mice at 18 months. EPA's analysis of the NEDO
10 (1985/2008b) inhalation study of F344 rats indicates a dose-response trend (Cochrane-Armitage
11 p< 0.05) and an increased incidence over concurrent controls at the high dose (Fisher's exact
12 p< 0.05) of pulmonary adenomas/adenocarcinomas in male rats. This analysis also indicates a
13 statistically significant dose-response trend (Cochrane-Armitage/? < 0.05) and a statistically
14 significant increased incidence over NTP historical controls at the high-dose (Fisher's exact
15 p< 0.05) of pheochromocytomas in female rats.
16 This WOE conclusion is supported by the results of other studies performed by ERF that
17 have shown tumorigenic responses similar to that of methanol in male and female Sprague-
18 Dawley rats exposed to formaldehyde (via drinking water), a metabolite of methanol, and to
19 aspartame (via feed) and MTBE (via olive oil gavage), substances that hydrolyze to release
20 methanol and formaldehyde. Confidence in the designation of methanol as a likely human
21 carcinogen is strengthened by the fact that methanol is metabolized to formaldehyde, a chemical
22 that has been classified as carcinogenic to humans (group 1) (IARC, 2004). As discussed below
23 and in Section 5.4.3, there are uncertainties in the interpretation of these findings. All of the key
24 studies have design and reporting limitations. EPA has reanalyzed the reported data from both
25 the ERF (Soffritti et al., 2002a) and NEDO (1987, 1985/2008a, 2008b) studies. In reassessing
26 the ERF study data, EPA decided to combine only those lymphomas considered to have
27 originated from the same cell type. In the case of the NEDO data, the significance of the tumor
28 findings were incompletely reported in the original NEDO (1987) summary. Hence, EPA used
29 translations of the original, detailed Japanese study reports provided by NEDO and the Methanol
30 Institute (NEDO, 1985/2008a,2008b) and reanalyzed the individual animal data.
4.9.2. Synthesis of Human, Animal, and Other Supporting Evidence
31 Evidence of the carcinogenic potential of methanol arises from drinking water studies in
32 Sprague-Dawley rats (Soffritti et al., 2002a) and in Eppley Swiss Webster mice (Apaja, 1980),
59 Obtained by combining control data from ERF studies of methanol, formaldehyde, aspartame, MTBE, and
TAME.available from the ERF website at http://www.ramazzini.it/fondazione/foundation.aspX
4-103 DRAFT-DO NOT CITE OR QUOTE
-------
1 and an inhalation study in F344 rats (NEDO, 1985/2008b), with no information available in
2 humans. As is described in Section 4.2.1.3 (Table 4-2), Soffritti et al. (2002a) reported a number
3 of tumors in methanol-exposed Sprague-Dawley rats. EPA reanalyzed the tumor findings from
4 this study using individual animal pathology available from the ERF website (see Section
5 5.4.1.1).60 As indicated above, the increase in a relatively rare hepatocellular carcinoma in males
6 compared to historical controls (Fisher's exact/? < 0.05 for all doses andp < 0.01 for the high-
7 dose group) is potentially related to methanol dosing. A significant increase in the incidence of
8 ear duct carcinoma was also reported by Soffritti et al. (2002a). However, the high incidence for
9 this tumor in controls of the Soffritti et al. (2002a) study relative to other studies of Sprague-
10 Dawley rats (Cruzan, 2009) and the results of an NTP evaluation of pathology slides from
11 another bioassay (EFSA, 2006; Hailey, 2004) raise questions about the ear duct pathological
12 determinations of Soffritti et al. (2002a).61
13 As is described in Section 4.2.1.3 (Table 4-3), Apaja (1980) found an increase in
14 malignant lymphomas in mid-dose (p = 0.06) and high-dose (p < 0.05) female and mid-dose
15 (p < 0.05) male Eppley Swiss Webster mice exposed for life via drinking water. The lack of a
16 concurrent unexposed control data limit the confidence that can be placed on the relevance of the
17 increased lymphoma responses noted in this study. However, while controls were not
18 concurrent, they were from proximate (within 3 years) generations of the same mouse colony,
19 lymphomas were evaluated via the same classification criteria and, in the case of the Hinderer
20 (1979) controls, the histopathological analysis was performed by the same author (Apaja, 1980).
21 In addition, this is a late developing tumor, as noted by the author, suggesting the possibility of a
22 higher tumor response in the females of all exposure groups had their survival not been
23 significantly lower than untreated historical controls (see dose-response analysis in Appendix E).
24 Further, additional support for these study results comes from the fact that, as described above,
25 Soffritti et al. (2002a) subsequently reported an increase in lymphomas following similar levels
26 of mg/kg-day methanol drinking water doses to Sprague-Dawley rats, another rodent species
27 with a high spontaneous lymphoma rate.
28 Chronic inhalation bioassays have been conducted in monkeys, mice, and F344 rats
29 (NEDO, 1987, 1985/2008a; 1985/2008b). No exposure-related carcinogenic responses were
30 observed in the monkey or mouse studies. As is described in Section 4.2.2.3, individual tumor
31 responses from the rat study were not significantly increased over concurrent controls, but the
60 ERF provided the EPA with the detailed, individual animal data via reports available through their web portal
(http://www.ramazzini.it/fondazione/foundation.aspX This allowed the EPA to combine lymphomas of similar
histopathological origins and confirm the tumor incidences reported in the Soffriti et al. (2002a) paper.
61 In an NTP evaluation of pathology slides from another bioassay from this laboratory in which a similar ear duct
carcinoma finding was reported (Soffritti et al., 2006, 2005), NTP pathologists interpreted a majority of these ear
duct responses as being hyperplastic, not carcinogenic, in nature (EFSA, 2006; Hailey, 2004).
4-104 DRAFT-DO NOT CITE OR QUOTE
-------
1 response in the high-dose (1,000 ppm) group for pulmonary adenomas/adenocarcinomas in male
2 rats was increased over concurrent controls (Fisher's exact/? < 0.05), and the dose-response for
3 both pulmonary adenomas/adenocarcinomas in male rats and pheochromocytomas in female rats
4 represent increasing trends (Cochran-Armitage trend test/? < 05). Further, the high-dose
5 responses for both of these tumor types were elevated (p < 0.05) over historical control
6 incidences within their respective sex and strain. As can be seen from Table 4-5, the severity and
7 combined incidence of effects reported in the alveolar epithelium of male rat lungs (epithethial
8 swelling, adenomatosis, pulmonary adenoma and pulmonary adenocarcinoma) and the adrenal
9 glands of female rats (hyperplasia and pheochromocytoma) were increased over controls and
10 lower exposure groups. This pathology and the appearance of a rare adenocarcinoma in the
11 high-dose group are suggestive of a progressive effect associated with methanol exposure. The
12 increased pheochromocytoma response in female rats is considered to be potentially treatment
13 related because this is a historically rare tumor type for female F344 rats (NTP, 2007, 1999;
14 Haseman et al., 1998)62 and because, when viewed in conjunction with the increased medullary
15 hyperplasia observed in the mid-exposure (100 ppm) group females, it is indicative of a
16 proliferative change with increasing methanol exposure.
17 Additional support for the designation of methanol as a likely carcinogen is provided by
18 the fact that methanol is metabolized to formaldehyde, which has been associated with increased
19 incidences of lymphoma and leukemia in humans (IARC, 2004). Furthermore, lymphomas
20 similar to those noted in Sprague-Dawley rats following exposure to methanol in drinking water
21 and following a similar dose-response pattern were noted in a bioassay for formaldehyde in
22 drinking water conducted by the same laboratory (Soffritti et al., 2002b, 1989) (Section 4.9.3).
23 These shared endpoints suggests that the carcinogenic effects of methanol may result from its
24 conversion to formaldehyde, though the moiety and MOA responsible for methanol-associated
25 tumor formation have not been identified.
26 Significant increases in the incidence of lymphoreticular tumors have also been reported
27 for other chemicals that convert in the body to methanol and/or formaldehyde including
28 aspartame (Soffritti et al., 2007, 2006, 2005) and MTBE (Belpoggi et al., 1997, 1995). In
29 contrast, no such tumors have been reported in a similar study conducted with a structurally
30 similar alcohol, ethanol (Soffritti et al., 2002a). In addition, epidemiological studies have
31 associated formaldehyde exposure with increases in the incidence of related
32 lymphohematopoietic tumors. While lymphomas are a rare finding in chronic laboratory
33 bioassays, NCI (Hauptmann et al., 2003) and NIOSH (Pinkerton et al., 2004) have reported
34 increased lymphohematopoietic cancer risk, principally leukemia, in humans from occupational
ft1")
Haseman et al. (1998) report rates for spontaneous pheochromocytomas in 2-year NTP bioassays of 5.7%
(benign) and 0.3% (malignant) in male F344 rats and 0.3% (benign) and 0.1% (malignant) in female (n = 1517)
F344 rats.
4-105 DRAFT-DO NOT CITE OR QUOTE
-------
1 exposure to formaldehyde.63 The similarities in tumor response across these chemicals, as well
2 as a similar shape in the dose-response curve, supports the hypothesis that the common
3 carcinogenic moiety for these compounds is the generation or presence of formaldehyde. The
4 dose-response analysis discussed in Section 5 provides additional evidence supporting a role for
5 the formaldehyde metabolite of methanol. When "total metabolites in blood" predicted by a
6 PBPK model was used as the dose metric, model fit to the dose-response data was significantly
7 improved.
8 As discussed in Section 4.6.5.2, there are challenges relative to the interpretation of the
9 observed lymphoreticular tumors because there is no indication that ERF used specific pathogen-
10 free (SPF) rats (Schoeb et al., 2009), and the protocol for the studies conducted by the ERF
11 (Soffritti et al., 2002c) is different from 2-year bioassays conducted by NTP and NEDO. A
12 distinct characteristic of the protocol for long-term bioassays conducted by the ERF is to
13 maintain animals until spontaneous death, rather than sacrificing them at the end of exposure at
14 104 weeks. This difference in protocol may have an impact on the tumors observed compared to
15 a 2-year bioassay (Melnick et al., 2007). The ERF methanol and ethanol studies (Soffritti et al.,
16 2002a), as well as the aspartame studies (Soffritti et al., 2007, 2006) described in Section 4.6.5.2,
17 employed a large number of animals (100 or more per dose group) compared to a typical (e.g.,
18 NTP) cancer bioassay. In addition, the Sprague-Dawley rats used by ERF appear to have
19 increased sensitivity to certain lymphoma responses relative to F344 rats that have been typically
20 used in NTP studies (Caldwell et al., 2008).64 According to Soffritti et al. (2007, 2006), the
21 overall incidence of lymphomas/leukemias in ERF studies is 13.3% (range, 4.0-25.0%) in
22 female historical controls (2,274 rats) and 20.6% (range, 8.0-30.9%) in male historical controls
23 (2,265 rats). This background rate is considered to be high relative to other tumor types and
24 relative to the background rate for this tumor type in Sprague-Dawley rats from other
25 laboratories (Cruzan, 2009; EFSA, 2006),65 However, it is in a range that can be considered
26 reasonable for studies that employ a large number of animals (Caldwell et al., 2009; Leakey et
27 al., 2003). These characteristics of ERF studies (i.e., lifetime observation, large number of
28 animals, and test strain sensitive to endpoint but with a relatively low control background rate
29 and mortality) may give them the sensitivity needed to detect a chemically related lymphoma
30 response.
63IARC (2004) concluded that there was sufficient epidemiological evidence that formaldehyde causes
nasopharyngeal cancer in humans but, also, that there was strong evidence for a causal association between
formaldehyde and the development of leukemia in humans.
64 F344 rats have a high mortality rate due to late-developing mononuclear cell leukemia, but the lymphoblastic and
immunoblastic lymphomas reported in the Sprague-Dawley rat by ERF following methanol, MTBE, formaldehyde
and aspartame administration are rarely diagnosed in the F344 rat (Caldwell et al., 2008).
65 Cruzan (2009) reports that the incidences of total cancers derived from bloodforming cells, designated as
hemolymphoreticular tumors by Ramazzini pathologists, is consistently about four times higher than the incidences
of such tumors in SD rats recorded in the Charles River Laboratory historical database (CRL database).
4-106 DRAFT-DO NOT CITE OR QUOTE
-------
1 Other aspects of ERF studies may impede their ability to reliably detect a chemically
2 related response (EFSA, 2009, 2006). Chronic inflammatory responses have been reported in
3 test animals of some ERF studies (EFSA, 2009, 2006), which may be the result of infections in
4 test animals resulting from a bioassay design that does not employ SPF rats (Schoeb et al., 2009)
5 and allows the rats to live out their "natural life span" in the absence of disease barriers (e.g.,
6 fully enclosed cages). In fact, the ERF has acknowledged that the primary cause of spontaneous
7 death in their rats is respiratory infection (Caldwell et al., 2008; Ramazzini Foundation, 2006;
8 Soffritti et al., 2006). Cruzan (2009) has suggested that respiratory infections in test animals of
9 the Soffritti et al. (2002a) methanol study were not specific to older rats, as findings of lung
10 pathology were reported as often in rats dying prior to 18 months as in rats dying at or after
11 24 months.66
12 In their reviews of the recently published ERF studies on aspartame (Soffritti et al., 2007,
13 2006), the European Food Safety Authority (EFSA) have suggested that the increased incidence
14 of lymphomas/leukemias reported in treated rats was related to chronic respiratory disease in the
15 rat colony (EFSA, 2009, 2006), which they suggest was caused by aMycoplasmapulmonis
16 (M pulmonis) infection. EFSA felt that the increased incidence of these tumors was unrelated to
17 aspartame, given the high background incidence of chronic inflammatory changes such as
18 bronchopneumonia in the lungs of treated and untreated rats, and the concern that such tumors
19 might arise as a result of abundant lymphoid hyperplasia in the lungs of rats suffering from
20 chronic respiratory disease. The scientific evidence to support the EFSA opinion that
21 lymphomas/leukemias can result from chronic infection is limited (Schoeb et al., 2009; Caldwell
22 et al., 2008). Epithelial hyperplasias and lymphoid accumulations are commonly found in the
23 larynx and trachea of rats infected with M. pulmonis, but induction of lymphoma has not been
24 noted (Everitt and Richter, 1990; Lindsey et al., 1985). Further, the lung, not the larynx or
25 trachea, has been reported as the site of respiratory tract hemolymphoreticular tumors in ERF
26 studies of MTBE (Belpoggi et al., 1998, 1995) and methanol (Soffritti et al., 2002a).67 In their
27 review of the molecular biology and pathogenicity of M pulmonis, Razin et al. (1998) note that
28 further study is needed before any conclusion can be reached regarding a relationship between
29 M. pulmonis and neoplasia. In addition, if the increased incidence of lymphoreticular tumors in
30 the ERF methanol study was strictly the consequence of an incipient respiratory infection in the
31 ERF rat colony, one would expect this to be a common finding across ERF studies. However, as
32 discussed in Section 4.6.5.2, of the 200 compounds tested by ERF, only 8, which includes
66 The infection rate did not have a significant impact on survival, however. The 2-year survival rate was 40-50% in
the ERF methanol bioassay (see Appendix E, Figures E-l and E-2), which is above the average 2-year NTP study
survival rate of 41.5% for Sprague-Dawley rats (Caldwell et al., 2008).
67 ERF provided EPA with the detailed, individual animal data for the Soffriti et al. (2000a) via reports available
through their web portal (http://www.ramazzini.it/fondazione/foundation.asp).
4-107 DRAFT-DO NOT CITE OR QUOTE
-------
1 methanol, have been associated with an increased incidence of hemolymphoreticular tumors.
2 Further, the chemicals for which hemolymphoreticular tumors have been reported have chemical
3 characteristics or physical properties in common,68 consistent with the hypothesis that the
4 increased response is chemical-related.
5 While evidence for a causal association between respiratory infections and lymphomas is
6 limited, there is evidence that respiratory infections may have confounded the interpretation of
7 lung lesions in the ERF studies. Schoeb et al. (2009) state that lymphomas illustrated in two
8 ERF studies (Figure 10 of Soffritti et al., 2005 and Figures 1-5 of Belpoggi et al., 1999) do not
9 demonstrate the lymphoma type, cellular morphology, and organ distribution typical of
10 lymphoma in rats, but are consistent with "lymphocyte and plasma cell accumulation in the lung
11 that is characteristic of M pulmonis disease." They suggest that, because M pulmonis disease
12 can be exacerbated by chemical treatment, a plausible alternative explanation for the dose-related
13 response reported in the MTBE, aspartame and methanol ERF bioassays is that the studies were
14 confounded by M. pulmonis disease and that lesions of the disease were interpreted as
15 lymphoma. However, several ERF lymphoma diagnoses in multiple rat organ systems, including
16 the lung, have been confirmed by an independent panel of six NIEHS pathologists (Hailey,
17 2004). Further, 60% of the lymphoma incidences reported in the ERF methanol study involved
18 organ systems other than the lungs (Schoeb et al., 2009). The incidence of "lung-only"
19 lymphomas is evenly distributed across the control and dose groups of the methanol study such
20 that removing "lung-only" lympho-immunoblastic lymphomas from consideration (i.e., using
21 only lymphomas from other organ systems) does not significantly alter the dose-response for this
22 lesion (see Section 5.4.3.2).
23 Based on the NEDO (1987) summary report, IPCS (1997) concluded that "no evidence of
24 carcinogenicity was found in either species [F344 rats and B6C3Fi mice]." This determination
25 was made based on Fisher's exact test results which indicated that the reported high-dose
26 pulmonary adenoma response in male rats and the high-dose pheochromocytoma response in
27 female rats were not statistically significant. However, IPCS did not have translations of the
28 original NEDO mouse and rat chronic studies (NEDO, 1985/2008a; 1985/2008b), which
29 provided additional detail for EPA's analysis and reported combined lung adenoma and
30 adenocarcinoma results for high-dose male rats. In addition, IPCS did not consider trend test
31 results or historical tumor data for F344 rats, both of which indicate a positive result for lung
32 adenoma/adenocarcinoma (males) and pheochromocytomas (females) from the NEDO rat study.
68 Methanol, formaldehyde, aspartame, and MTBE, have common metabolites (e.g., formaldehyde); DIPE, TAME,
methanol, and MTBE are all gasoline-oxygenate additives.
4-108 DRAFT-DO NOT CITE OR QUOTE
-------
4.9.3. MOA Information
1 As discussed in Section 4.6.5.1, the results of genotoxicity/mutagenicity studies have
2 been largely negative, irrespective of the presence or absence of metabolic activation (an S9
3 microsomal fraction). Studies that investigate the MOA for methanol, particularly with respect
4 to its developmental effects, have been discussed extensively in Sections 4.6. and 4.8. Studies
5 such as those by Harris et al. (2004, 2003) suggest that formaldehyde is the proximate teratogen
6 and provide evidence in support of that hypothesis. It is reasonable to hypothesize that the
7 highly reactive molecule, formaldehyde, has a role in the carcinogenicity of methanol, given the
8 ability of formaldehyde to bind to proteins and DNA, induce DNA-protein cross-links, and
9 possibly participate in reactions leading to free radical formation and the formation of lipid
10 peroxidation products. As discussed in Section 4.6.3, evidence of oxidative stress following
11 methanol exposure has been reported in several organ systems. Studies of Wistar rats suggest
12 that methanol exposure can cause the production of free radical formation, lipid peroxidation,
13 and protein modifications in the liver (Skrzydlewska et al., 2005) and brain (Rajamani et al.,
14 2006), and adversely impact the oxidant/antioxidant balance in the brain (Dudka, 2006) and
15 lymphoid organs (Parthasarathy et al., 2006b).
16 As discussed in Section 4.6.5.2, ERF studies of a number of compounds that have
17 formaldehyde as a metabolic product have been reported to cause lymphomas in
18 Sprague-Dawley rats. As described in Section 4.6.5.2.4, the ERF has conducted a formaldehyde
19 drinking water study (Soffritti et al., 2002b, 1989) that is comparable in its design to the
20 methanol drinking water study of Soffritti et al. (2002a). The mg/kg-day doses of metabolized
21 methanol in Sprague-Dawley rats from the ERF methanol study estimated from the PBPK model
22 described in Section 3.4 and mg/kg-day doses of formaldehyde reported in the ERF
23 formaldehyde study were plotted together versus the hemolymphoreticular neoplasm incidences
24 in their respective studies (Figure 4-1). Separate linear models were fit to the male and female
25 rat data from these studies. The model fits shown in Figure 4-1 demonstrate that when
26 metabolized methanol is used as the dose metric for the methanol study data, the dose-response
27 data from these two studies can be adequately fit by two separate linear dose-response functions
28 for the combined male (R2 = 0.6722) and combined female (R2 = 0.779) responses. Even if it is
29 true that formaldehyde is the common moiety responsible for these tumors, one would not expect
30 this approach to result in perfect dose-response alignment because the metabolized methanol
31 estimate is not an accurate representation of formaldehyde distribution, and formaldehyde from
32 methanol administration would not be expected to distribute the same as orally administered
33 formaldehyde. However, the similarities in the dose-response data for male and female rats from
34 these studies are consistent with the hypothesis that formaldehyde is key to methanol's
35 carcinogenic MOA.
4-109 DRAFT-DO NOT CITE OR QUOTE
-------
0.5
0.45 -
I
tfl
<0 0.4
o.
o
0>
Z 0.35 -
J2
.1 0.3 -
•5 0.15-1
o>
! c
s
o
— 0.05 -
y = 0.0492Ln(x) + 0.0932
R2 = 0.6832
10.00 100.00
mg formaldehyde/kg/d or mg metabolized methanol/kg/d
n Sofritti et al. (2002b), formaldehyde - males
• Sofiritti et al. (2002a), methanol - females
• Sof itti et al. (2002a), methanol - males
o Sofrritti et al. (2002b), formaldehyde - females
Figure 4-1. Hemolymphoreticular neoplasms in male and female Sprague-Dawley rats
in formaldehyde and methanol drinking water studies versus mg formaldehyde/kg/day
or mg metabolized methanol/kg/day (predicted by EPA PBPK model).
Source: Soffritti et al. (2002b).
1 As discussed above, methanol is metabolized to formaldehyde, which is deemed to be
2 carcinogenic to humans (IARC, 2004) by both the oral and inhalation routes, and there are
3 readily apparent similarities between the dose-response data from oral studies of rats exposed to
4 formaldehyde and methanol. In addition, the dose-response model fit for the lymphoma
5 response observed in the Soffritti et al. (2002a) study is improved when predicted total
6 metabolites is used as the dose-metric (Section 5.4.1.2). However, the database of information
7 available concerning methanol's carcinogenic MOA is limited, and the extent to which the parent
8 or a metabolite such as formaldehyde is responsible for the carcinogenic effects observed in the
9 studies conducted by Soffritti et al. (2002a) or NEDO (1987, 1985/2008b) is not clear.
4.10. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.10.1. Possible Childhood Susceptibility
10 Studies in animals have identified the fetus as being more sensitive than adults to the
11 toxic effects of methanol; the greatest susceptibility occurs during gastrulation and early
12 organogenesis (CERHR, 2004). Table 4-21 summarizes some of the data regarding the relative
13 ontogeny of CAT, ADH1, and ADH3 in humans and mice. Human fetuses have limited ability to
14 metabolize methanol as ADH1 activity in 2-month-old and 4-5 month-old fetuses is 3-4% and
4-110 DRAFT-DO NOT CITE OR QUOTE
-------
1 10% of adult activity, respectively (Pikkarainen and Raiha, 1967). ADH1 activity in 9-22 week
2 old fetal livers was found to be 30% of adult activity (Smith et al., 1971). Likewise, ADH1
3 activity is -20-50% of adult activity during infancy (Smith et al., 1971; Pikkarainen and Raiha,
4 1967). Activity continues to increase until reaching adult levels at 5 years of age (Pikkarainen
5 and Raiha, 1967). However, no difference between blood methanol levels in 1-year-old infants
6 and adults was observed following ingesting the same doses of aspartame, which releases 10%
7 methanol by weight during metabolism (Stegink et al., 1983). Given that the exposure was
8 aspartame as opposed to methanol, it is difficult to draw any conclusions from this study vis-a-
9 vis ontogeny data and potential influences of age differences in aspartame disposition. With
10 regard to inhalation exposure, increased breathing rates relative to adults may result in higher
11 blood methanol levels in children compared to adults (CERHR, 2004). It is also possible that
12 metabolic variations resulting in increased methanol blood levels in pregnant women could
13 increase the fetus' risk from exposure to methanol. In all, unresolved issues regarding the
14 identification of the toxic moiety increase the uncertainty with regards to the extent and
15 pathologic basis for early life susceptibility to methanol exposure.
16 The prevalence of folic acid deficiency has decreased since the United States and Canada
17 introduced a mandatory folic acid food fortification program in November 1998. However,
18 folate deficiency is still a concern among pregnant and lactating women, and factors such as
19 smoking, a poor quality diet, alcohol intake, and folic antagonist medications can enhance
20 deficiency (CERHR, 2004). Folate deficiency could affect a pregnant woman's ability to clear
21 formate, which has also been demonstrated to produce developmental toxicity in rodent in in
22 vitro studies at high-doses (Dorman et al., 1995). It is not known if folate-deficient humans have
23 higher levels of blood formate than individuals with adequate folate levels. A limited study in
24 folate-deficient monkeys demonstrated no formate accumulation following an endotracheal
25 exposure of anesthetized monkeys to 900 ppm methanol for 2 hours (Dorman et al., 1994). The
26 situation is obscured by the fact that folic acid deficiency during pregnancy by itself is thought to
27 contribute to the development of severe congenital malformations (Pitkin, 2007).
4.10.2. Possible Gender Differences
28 There is limited information on potential differences in susceptibility to the toxic effects
29 of methanol according to gender. However, one study reported a higher background blood
30 methanol level in human females versus males (Batterman and Franzblau, 1997). In rodents,
31 fetuses exposed in utero were found to be the most sensitive subpopulation. One study suggested
32 a possible increased sensitivity of male versus female rat fetuses and pups. When rats were
33 exposed to methanol pre- and postnatally, 6- and 8-week-old male progeny had significantly
34 lower brain weights at 1,000 ppm, compared to those in females that demonstrated the same
4-111 DRAFT-DO NOT CITE OR QUOTE
-------
1 effect only at 2,000 ppm (NEDO, 1987). In general, there is little evidence for substantial
2 disparity in the level or degree of toxic response to methanol in male versus female experimental
3 animals or humans. However, it is possible that the compound-related deficits in fetal brain
4 weight that were evident in the pups of FI generation Sprague-Dawley rats exposed to methanol
5 in the NEDO (1987) study may reflect a threshold neurotoxicological response to methanol. It is
6 currently unknown whether higher levels of exposure would result in brain sequelae comparable
7 to those observed in acutely exposed humans.
4.10.3. Genetic Susceptibility
8 Polymorphisms in enzymes involved in methanol metabolism may affect the sensitivity
9 of some individuals to methanol. For example, as discussed in Chapter 3, data summarized in
10 reviews by Agarwal (2001), Burnell et al. (1989), Bosron and Li (1986), and Pietruszko (1980)
11 discuss genetic polymorphisms for ADH. Class IADH, the primary ADH in human liver, is a
12 dimer composed of randomly associated polypeptide units encoded by three genetic loci
13 (ADH1 A, ADH1B, and ADH1C). Polymorphisms are observed at the ADH1B (ADH1B*2,
14 ADH1B*3) and ADH1C (ADH1C*2) loci. The ADH1B*2 phenotype is estimated to occur in
15 -15% of Caucasians of European descent, 85% of Asians, and less that 5% of African
16 Americans. Fifteen percent of African Americans have the ADH1B*3 phenotype, while it is
17 found in less than 5% of Caucasian Europeans and Asians. The only reported polymorphisms in
18 ADH3 occur in the promoter region, one of which reduces the transcriptional activity in vitro
19 nearly twofold (Hedberg et al., 2001). While polymorphisms in ADH3 are described in more
20 than one report (Cichoz-Lach et al., 2007; Hedberg et al., 2001), the functional consequence(s)
21 for these polymorphisms remains unclear.
4-112 DRAFT-DO NOT CITE OR QUOTE
-------
5. DOSE-RESPONSE ASSESSMENT AND CHARACTERIZATION
5.1. INHALATION RfC FOR EFFECTS OTHER THAN CANCER69
1 In general, the RfC is an estimate of a daily exposure of the human population (including
2 susceptible subgroups) that is likely to be without an appreciable risk of adverse health effects
3 over a lifetime. It is derived from a POD, generally the statistical lower confidence limit on the
4 BMCL or BMDL, with uncertainty/variability factors applied to reflect limitations of the data
5 used. The inhalation RfC considers toxic effects for both the respiratory system (portal-of-entry)
6 effects and systems peripheral to the respiratory system (extra-respiratory or systemic effects). It
7 is generally expressed in mg/m3. EPA performed an IRIS assessment of methanol in 1991 and
8 determined that the database was inadequate for derivation of an RfC. While some limitations
9 still exist in the database (see Sections 5.1.3.2 and 5.3), the experimental toxicity database has
10 expanded and newer methods and models have been developed to analyze the results. In this
11 update, the PBPK model, described in Section 3.4, was developed by EPA and is used to estimate
12 HECs and HEDs from inhalation study data for the derivation of both the RfC and RfD. In both
13 cases, the use of a PBPK model replaces part of the UF adjustments traditionally used for
14 species-to-species extrapolation.
15 Additionally, this assessment uses the BMD method in its derivation of the POD.70 The
16 suitability of these methods to derive a POD is dependent on the nature of the toxicity database
17 for a specific chemical. Details of the BMD analyses are found in Appendix C. The use of the
18 BMD approach for determining the POD improves the assessment by including consideration of
19 shape of the dose-response curve, independence from experimental doses, and estimation of the
20 uncertainty pertaining to the calculated dose response. However, the methanol database still has
21 limitations and uncertainties associated with it, in particular, those uncertainties associated with
22 human variability, animal-to-human differences, and limitations in the database influence
23 derivation of the RfC.
5.1.1. Choice of Principal Study and Critical Effect(s)
5.1.1.1. Key Inhalation Studies
24 While a substantial body of information exists on the toxicological consequences to
25 humans exposed to large amounts of methanol, no human studies exist that would allow for
69 The RfC discussion precedes the RfD discussion in this assessment because the inhalation database ultimately
serves as the basis for the RfD. The RfD development would be difficult to follow without prior discussion of
inhalation database and PK models used for the route-to-route extrapolation.
70 Use of BMD methods involves fitting mathematical models to dose-response data and using the results to select a
POD that is associated with a predetermined benchmark response (BMR), such as a 10% increase in the incidence of
a particular lesion or a 10% decrease in body weight gain (see Section 5.1.2.2).
5-1 DRAFT—DO NOT CITE OR QUOTE
-------
1 quantification of sub chronic, chronic, or in utero effects of methanol exposure. Table 4-35
2 summarizes available experimental animal inhalation studies of methanol. Several of these
3 studies, including the monkey chronic (NEDO, 1987) and developmental (Burbacher et al.,
4 2004a, 2004b, 1999a, 1999b) studies, the male rat reproductive studies (Lee et al., 1991;
5 Cameron et al., 1985, 1984), and the 4-week rat studies (Poon et al., 1994), are lacking in key
6 attributes (e.g., documented dose response, documented controls, and duration of exposure)
7 necessary for their direct use in the quantification of a chronic RfC. These studies will be
8 considered in this chapter for their contributions to the overall RfC uncertainty. Several
9 inhalation reproductive or developmental studies were adequately documented and are of
10 appropriate size and design for quantification and derivation of an RfC. These studies are
11 considered for use in the derivation of an RfC and are summarized below.
5.1.1.2. Selection of Critical Effect(s)
12 Developmental effects have been assessed in a number of toxicological studies of
13 monkeys, rats, and mice. The findings of Rogers and Mole (1997) indicate that methanol is toxic
14 to mouse embryos in the early stages of organogenesis, on or around GD7. In the study of
15 Rogers et al. (1993a), in which pregnant female CD-I mice were exposed to methanol vapors
16 (1,000, 2,000, and 5,000 ppm) on GD6-GD15, reproductive and fetal effects included an
17 increase in the number of resorbed litters, a reduction in the number of live pups, and increased
18 incidences of exencephaly, cleft palate, and the number of cervical ribs. They reported a
19 NOAEL for cervical rib malformations at 1,000 ppm (1,310 mg/m3) and a LOAEL of 2,000 ppm
20 (2,620 mg/m3, 49.6% per litter versus 28.0% per litter in the control group). Increased incidence
21 of cervical ribs was also observed in the rat organogenesis study (NEDO, 1987) in the 5,000 ppm
22 dose group (65.2% per litter versus 0% in the control group), indicating that the endpoint is
23 significant across species.
24 The biological significance of the cervical rib endpoint within the regulatory arena has
25 been the subject of much debate (Chernoff and Rogers, 2004). Previous studies have classified
26 this endpoint as either a malformation (birth defect of major importance) or a variation
27 (morphological alternation of minor significance). There is evidence that incidence of
28 supernumerary ribs (including cervical ribs) is not just the addition of extraneous, single ribs but
29 rather is related to a general alteration in the development and architecture of the axial skeleton
30 as a whole. In CD-I mice exposed during gestation to various types of stress, food and water
31 deprivation, and the herbicide dinoseb, supernumerary ribs were consistently associated with
32 increases in length of the 13th rib (Branch et al., 1996). This relationship was present in all fetal
33 ages examined in the study. The authors concluded that these findings are consistent with
34 supernumerary ribs being one manifestation of a basic alteration in the differentiation of the
35 thoraco-lumbar border of the axial skeleton. The biological significance of this endpoint is
36 further strengthened by the association of supernumerary ribs with adverse health effects in
5-2 DRAFT—DO NOT CITE OR QUOTE
-------
1 humans. The most common effect produced by the presence of cervical ribs is thoracic outlet
2 disease (Nguyen et al., 1997; Fernandez Noda et al., 1996; Henderson, 1914). Thoracic outlet
3 disease is characterized by numbness and/or pain in the shoulder, arm, or hands. Vascular effects
4 associated with this syndrome include cerebral and distal embolism (Beam et al., 1993; Connell
5 et al., 1980; Short, 1975), while neurological symptoms include extreme pain, migraine, and
6 symptoms similar to Parkinson's (Evans, 1999; Saxton et al., 1999; Fernandez Noda et al., 1996).
7 Schumacher et al. (1992) observed 242 rib anomalies in 218 children with tumors (21.8%) and
8 11 (5.5%) in children without malignancy, a statistically significant (p < 0.001) difference that
9 indicates a strong association between the presence of cervical ribs and childhood cancers.
10 Specific cancers associated with statistically significant increases in anomalous ribs included
11 leukemia, brain, tumor, neuroblastoma, soft tissue sarcoma, and Wilm's tumor.
12 A number of rat studies have confirmed the toxicity of methanol to embryos during
13 organogenesis (Weiss et al., 1996; NEDO, 1987; Nelson et al., 1985). NEDO (1987) reported
14 reduced brain, pituitary, and thymus weights in FI and F2 generation Sprague-Dawley rats at
15 1,000 ppm methanol. In a follow-up study of the FI generation brain weight effects, NEDO
16 (1987) reported decreased brain, cerebellum, and cerebrum weights in FI males exposed at
17 1,000 ppm methanol from GDO through the FI generation. The exposure levels used in these
18 studies are difficult to interpret because dams were exposed prior to gestation, and dams and
19 pups were exposed during gestation and lactation. However, it is clear that postnatal exposure
20 increases the severity of brain weight reduction. In another experiment in which NEDO (1987)
21 exposed rats only during organogenesis (GD7-GD17), the observed decreases in brain weights in
22 offspring at 8 weeks of age were less severe than in the studies for which exposure was
23 continued postnatally. This finding is not unexpected, given that the brain undergoes tremendous
24 growth beginning early in gestation and continuing in the postnatal period. Rats are considered
25 altricial (i.e., born at relatively underdeveloped stages), and many of their neurogenic events
26 occur postnatally (Clancy et al., 2007). Brain effects from postnatal exposure are also relevant to
27 humans given that, in humans, gross measures of brain growth increase for at least 2-3 years
28 after birth, with the growth rate peaking approximately 4 months after birth (Rice and Barone,
29 2000).
30 A change in brain weight is considered to be a biologically significant effect (U.S. EPA,
31 1998). This is true regardless of changes in body weight because brain weight is generally
32 protected during malnutrition or weight loss, unlike many other organs or tissues
33 (U.S. EPA, 1998). Thus, change in absolute brain weight is an appropriate measure of effects on
34 this critical organ system. Decreases in brain weight have been associated with simultaneous
35 deficits in neurobehavioral and cognitive parameters in animals exposed during gestation to
36 various solvents, including toluene and ethanol (Gibson, 2000; Coleman et al., 1999; Hass,
5-3 DRAFT—DO NOT CITE OR QUOTE
-------
1 1995). NEDO (1987) reports that brain, cerebellum, and cerebrum weights decrease in a dose-
2 dependant manner in male rats exposed to methanol throughout gestation and the FI generation.
3 Developmental neurobehavioral effects associated with methanol inhalation exposure
4 have been investigated in monkeys. Burbacher et al. (2004a, 2004b, 1999a, 1999b) exposedM
5 fascicularis monkeys to 0, 262, 786, and 2,359 mg/m3 methanol, 2.5 hours/day, 7 days/week
6 during premating/mating and throughout gestation (approximately 168 days). In these studies,
7 exposure of monkeys to up to 1,800 ppm (2,359 mg/m3) methanol during premating, mating, and
8 throughout gestation resulted in no changes in reproductive parameters other than a shorter
9 period of gestation in all exposure groups that did not appear to be dose related. The shortened
10 gestation period was largely the result of C-sections performed in the methanol exposure groups
11 "in response to signs of possible difficulty in the maintenance of pregnancy," including vaginal
12 bleeding. As discussed in Section 4.7.1.2, though statistically significant, the shortened gestation
13 finding may be of limited biological significance given questions concerning its relation to the
14 methanol exposure. Developmental parameters, such as fetal crown-rump length and head
15 circumference, were unaffected, but there appeared to be neurotoxicological deficits in methanol -
16 exposed pups. VDR was significantly reduced in the 786 mg/m3 group for males and the 2,359
17 mg/m3 group for both sexes. However, a dose-response trend for this endpoint was only
18 exhibited for females. In fact, this is the only effect reported in the Burbacher et al. (2004a,
19 2004b, 1999a, 1999b) studies for which a significant dose-response trend is evident. As
20 discussed in Section 4.4.2, confidence may have been increased by statistical analyses to adjust
21 for multiple testing (CERHR, 2004). Yet it is worth noting that the dose-response trend for VDR
22 in females remained significant with (p = 0.009) and without (p = 0.0265) an adjustment for the
23 shortened gestational periods, and it is a measure of functional deficits in sensorimotor
24 development that is consistent with early developmental CNS effects (brain weight changes
25 discussed above) that have been observed in rats.
26 Another test, the Fagan test of infant intelligence, indicated small but not significant
27 deficits of performance (time spent looking at novel faces versus familiar faces) in treated
28 monkeys. Although not statistically significant and not quantifiable, the results of this test are
29 also important when considered in conjunction with the brain weight changes noted in the NEDO
30 (1987) rat study. As discussed in Section 4.7.1.2, the monkey data are not conclusive, and there
31 is insufficient evidence to determine if the primate fetus is more or less sensitive than rodents to
32 methanol teratogenesis. Taken together, however, the NEDO (1987) rat study and the Burbacher
33 et al. (2004a, 2004b, 1999a, 1999b) monkey study suggest that prenatal exposure to methanol
34 can result in adverse effects on developmental neurology pathology and function, which can be
35 exacerbated by continued postnatal exposure.
36 A number of studies described in Section 4.3.2 and summarized in Section 4.7.1.2 have
37 examined the potential toxicity of methanol to the male reproductive system (Lee et al., 1991;
5-4 DRAFT—DO NOT CITE OR QUOTE
-------
1 Cameron et al., 1985, 1984). Some of the observed effects, including a transient decrease in
2 testosterone levels, could be the result of chemically related strain on the rat system as it attempts
3 to maintain hormone homeostasis. However, the data are insufficient to definitively characterize
4 methanol as a toxicant to the male reproductive system.
5 The studies considered for use in the derivation of an RfC are summarized in Table 5-1.
6 As discussed in Sections 5.1.3.1 and 5.3, there is uncertainty associated with the selection of an
7 effect endpoint from the methanol database for use in the derivation of an RfC. Taking into
8 account the limitations of the studies available for quantification purposes, decreased brain
9 weight at 6 weeks in male Sprague-Dawley rats exposed throughout gestation and the postnatal
10 period (NEDO, 1987) was chosen as the critical effect for the purposes of this dose-response
11 assessment as it can be reliably quantified and represents both a sensitive organ system and a key
12 period of development. RfC derivations utilizing alternative endpoints (e.g., cervical rib effects
13 in mice and delayed sensorimotor development in monkeys) and alternative methods (e.g., use of
14 different BMRLs) are summarized in Appendix C and in Section 5.1.3.1.
Table 5-1. Summary of studies considered most appropriate for use in derivation of an
RfC
Reference
NEDO (1987)
Two-generation
study
NEDO (1987)
Follow-up study
of FI generation
NEDO (1987)
Teratology study
Nelson et al.
(1985)
Species
(strain)
Rat
Sprague-
Dawley
Rat
Sprague-
Dawley
Rat
Sprague-
Dawley
Sex
M,F
M,F
F
Number/
dose group
Not specified -
F! and F2
generation
10-147 sex/
group- F!
generation
10-12/sex/
group
15 pregnant
dams/group
Exposure
Duration
Fj- Birth to end
of mating (M)
or weaning (F);
F2-birth to 8
wk
GDO through
F! generation
GD7-GD17
GDI-GDI 9 or
GD7-GD15
Critical Effect
Reduced weight of
brain, pituitary, and
thymus at 8, 16, and
24 wk postnatal in F!
and at 8 wk in F2
Reduced brain weight
at 3 wk and 6 wk
(males only).
Reduced brain and
cerebrum weight at 8
wk (males only)
Reduced brain,
pituitary, thyroid,
thymus, and testis
weights at 8 wk
postnatal.
Reduced fetal body
weight, increased
incidence of visceral
and skeletal
abnormalities,
including rudimentary
and extra cervical ribs
NOAEL
(ppm)
100
500
1,000
5,000
LOAEL
(ppm)
1,000
1,000
5,000
10,000
5-5
DRAFT—DO NOT CITE OR QUOTE
-------
Reference
Rogers et al.
(1993a)
Burbacher et al.
(20004a, 2004b,
1999a, 1999b)
Species
(strain)
Mouse
CD-I
M.
fascicularis
Sex
F
Number/
dose group
30-114
pregnant dams/
group
12 pregnant
monkeys/group
Exposure
Duration
GD6-15
2.5 hr/day,
7 days/wk,
during
premating,
mating and
gestation
Critical Effect
Increased incidence
of extra cervical ribs,
cleft palate,
exencephaly; reduced
fetal weight and pup
survival, delayed
ossification
Shortened period of
gestation; may be
related to exposure
(no dose response),
neurotox. deficits
including reduced
performance in the
VDR test
NOAEL
(ppm)
1,000
LOAEL
(ppm)
2,000
b
"Animals were dosed 20-21 hr/day. NS = Not Specified
bGestational exposure resulted in a shorter period of gestation in dams exposed to as low as 200 ppm (263 mg/m3).
However, because of uncertainties associated with these results, including clinical intervention and the lack of a
dose-response, EPA was not able to identify a definitive NOAEL or LOAEL from this study.
5.1.2. Methods of Analysis for the POD—Application of PBPK and BMD Models
1 Potential PODs for the RfC derivation, described here and in Appendix C, have been
2 calculated via the use of monkey, rat and mouse PBPK models, described in Section 3.4. First,
3 the doses used in an experimental bioassay were converted to an internal dose metric that is most
4 appropriate for the endpoint being assessed. The PBPK models are capable of calculating
5 several measures of dose for methanol, including the following:
6 • Cmax- The peak concentration of methanol in the blood during the exposure period;
7 *A UC - Area under the curve, which represents the cumulative product of
8 concentration and time for methanol in the blood; and
9 • Total metabolism - The production of metabolites of methanol, namely formaldehyde
10 and formate.
11 As described in Section 3.4, the focus of model development is on obtaining accurate
12 predictions of increased body burdens over background. The PBPK models do not describe or
13 account for background levels of methanol, formaldehyde or formate.
14 Although there remains uncertainty surrounding the identification of the proximate
15 teratogen of importance (methanol, formaldehyde, or formate), the dose metric chosen for
16 derivation of an RfC was based on blood methanol levels. This decision was primarily based on
17 evidence that the toxic moiety is not likely to be the formate metabolite of methanol (CERHR,
18 2004) and evidence that levels of the formaldehyde metabolite following methanol maternal
19 and/or neonate exposure would be much lower in the fetus and neonate than in adults. While
20 recent in vitro evidence indicates that formaldehyde is more embryotoxic than methanol and
21 formate, the high reactivity of formaldehyde would limit its unbound and unaltered transport as
5-6 DRAFT—DO NOT CITE OR QUOTE
-------
1 free formaldehyde from maternal to fetal blood (Thrasher and Kilburn, 2001), and the capacity
2 for the metabolism of methanol to formaldehyde is likely lower in the fetus and neonate versus
3 adults (see discussion in Section 3.3). Thus, even if formaldehyde is identified as the proximate
4 teratogen, methanol would likely play a prominent role, at least in terms of transport to the target
5 tissue. Further discussions of methanol metabolism, dose metric selection, and MOA issues are
6 covered in Sections 3.3, 4.6, 4.8 and 4.9.2.
7 A BMDL was then derived in terms of the internal dose metric utilized. Finally, the
8 BMDL values were converted to HECs via the use of a PBPK model parameterized for humans.
9 The next section describes the rationale for and application of the benchmark modeling
10 methodology for the RfC derivation.
5.1.2.1. Application of the BMD/BMDL Approach
11 Several developments over the past few years impact the derivation of the RfC: 1) EPA
12 has developed draft BMD assessment methods (U.S. EPA, 2000b, 1995) and supporting software
13 (Appendix C) to improve upon the previous NOAEL/LOAEL approach; 2) MOA studies have
14 been carried out that can give more insight into methanol toxicity; and 3) EPA has refined PBPK
15 models for methanol on the basis of the work of Ward et al. (1997) (see Section 3.4. for
16 description of the EPA model). The EPA PBPK model provides estimates of HECs from rodent
17 exposures that are supported by pharmacokinetic information available for rodents and humans.
18 The following sections describe how the BMD/BMDL approach, along with the EPA PBPK
19 model, is used to obtain a POD for use in the derivation of an RfC for methanol in accordance
20 with current draft BMD technical guidance (U.S. EPA, 2000b).
21 The BMD approach attempts to fit models to the dose-response data for a given endpoint.
22 It has the advantage of taking more of the dose-response data into account when determining the
23 POD, as well as estimating the dose for which an effect may have a specific probability of
24 occurring. The BMD approach also accounts, in part, for the quality of the study (e.g., study
25 size) by estimating a BMDL, the 95% lower bound confidence limit on the BMD. The BMDL is
26 closer to the BMD (higher) for large studies and further away from the BMD (lower) for small
27 studies. Because the BMDL approach will account, in part, for a study's power, dose spacing,
28 and the steepness of the dose-response curve, it is generally preferred over the NOAEL
29 approach.
30 When possible, all experimental data points are included in this assessment to ensure
31 adequate fit of a BMD model and derivation of a BMDL. A summary of the POD values
32 determined by BMD analysis for the critical endpoint (as well as other considered endpoints)
33 (see Appendix C for modeling results), application of UFs, and conversion to HECs using the
34 BMD and PBPK approach, is included in Section 5.1.3.1.
5-7 DRAFT—DO NOT CITE OR QUOTE
-------
1 Use of the BMD approach has uncertainty associated with it. An element of the BMD
2 approach is the use of several models to determine which best fits the data.71 In the absence of
3 an established MOA or a theoretical basis for why one model should be used over another, model
4 selection is based on best fit to the experimental data selection. Model fit was determined by
5 statistics (AIC and ^ residuals of individual dose groups) and visual inspection recommended by
6 EPA (2000b)72
7 The PBPK model developed by EPA for methanol (described in Section 3.4) was applied
8 for the estimation of methanol blood levels in the exposed dams (NEDO, 1987). When using
9 PBPK models, it is very important to determine what estimate of internal dose (i.e., dose metric)
10 can serve as the most appropriate dose metric for the health effects under consideration.
11 The results of NEDO (1987), shown in Table 4-8, indicate that there is not a cumulative
12 effect of ongoing exposure on brain-weight decrements in rats exposed postnatally; i.e., the dose
13 response in terms of percent of control is about the same at 3 weeks postnatal as at 8 weeks
14 postnatal in rats exposed throughout gestation and the FI generation. However, there does
15 appear to be a greater brain-weight effect in rats exposed postnatally versus rats exposed only
16 during organogenesis (GD7-GD17). In male rats exposed during organogenesis only, there is no
17 statistically significant decrease in brain weight at 8 weeks after birth at the 1,000 ppm exposure
18 level. Conversely, in male rats exposed to the same level of methanol throughout gestation and
19 the FI generation, there was an approximately 5% decrease in brain weights (statistically
20 significant at the/? < 0.01 level). The fact that male rats exposed to 5,000 ppm methanol only
21 during organogenesis experienced a decrease in brain weight of 10% at 8 weeks postnatal
22 indicates that postnatal exposure is not necessary for the observation of persistent postnatal
23 effects. However, the fact that this decrease was less than the 13% decrease observed in male
24 rats exposed to 2,000 ppm methanol throughout gestation and the 8 week postnatal period
25 indicates that both exposure concentration and duration are important components of the ability
26 of methanol to cause this effect. The extent to which the observation of the increased effect is
27 due to a cumulative effect in rats exposed postnatally versus recovery in rats for which exposure
28 was discontinued at birth is not clear.
29 The fact that brain weight is susceptible to both the level and duration of exposure
30 suggests that a dose metric that incorporates a time component would be the most appropriate
31 metric to use. For these reasons, and because it is more typically used in internal-dose-based
71USEPA's BMDS 2.1 was used for this assessment as it provides data management tools for running multiple
models on the same dose-response data set. At this time, BMDS offers over 30 different models that are appropriate
for the analysis of dichotomous, continuous, nested dichotomous and time-dependent lexicological data. Results
from all models include a reiteration of the model formula and model run options chosen by the user, goodness-of-fit
information, the BMD, and the estimate of the lower-bound confidence limit on the BMD (BMDL).
72Akaike's Information Criterion (AIC) (Akaike, 1973) is used for model selection and is defined as -2L + 2P where
L is the log-likelihood at the maximum likelihood estimates for the parameters and P is the number of model degrees
of freedom.
5-8 DRAFT—DO NOT CITE OR QUOTE
-------
1 assessments and better reflects total exposure within a given day, daily AUC (measured for
2 22 hours exposure/day) was chosen as the most appropriate dose metric for modeling the effects
3 of methanol exposure on brain weights in rats exposed throughout gestation and continuing into
4 the FI generation.
5 Application of the EPA methanol PBPK model (described in Section 3.4) to the NEDO
6 (1987) study, in which developing rats were exposed during gestation and the postnatal period,
7 presents complications that need to be discussed. The neonatal rats in this study were exposed to
8 methanol gestationally before parturition as well as lactationally and inhalationally after
9 parturition. The PBPK model developed by EPA only estimates internal dose metrics for
10 methanol exposure in NP adult mice and rats. Experimental data indicate that inhalation-route
11 blood methanol kinetics in NP mice and pregnant mice on GD6-GD10 are similar (Dorman
12 et al., 1995; Perkins et al., 1995a,1995b; Rogers et al., 1993a, 1993b). In addition, experimental
13 data indicate that the maternal blood:fetal partition coefficient for mice is approximately 1 (see
14 Section 3.4.1.2). Assuming that these findings apply for rats, the data indicate that PBPK
15 estimates of PK and blood dose metrics for NP rats are better predictors of fetal exposure during
16 gestation than would be obtained from default extrapolations from external exposure
17 concentrations. However, as is discussed to a greater extent in Section 5.3, the additional routes
18 of exposure presented to the pups in this study (lactation and inhalation) present uncertainties
19 that suggest the average blood levels in pups in the NEDO (1987) report might be greater than
20 those of the dam. The assumption made in this assessment is that, if such differences exist
21 between human mothers and their offspring, they are not expected to be significantly greater than
22 that which has been postulated for rats. Thus, the PBPK model-estimated adult blood methanol
23 level is considered to be an appropriate dose metric for the purpose of this analysis and HEC
24 derivation.
5.1.2.2. BMD Approach Applied to Brain Weight Data in Rats
25 The NEDO (1987) study reported decreases in brain weights in developing rats exposed
26 during gestation only (GD7-GD17) or during gestation and the postnatal period, up to 8 weeks.
27 Because of the biological significance of decreases in brain weight as an endpoint in the
28 developing rat and because this endpoint was not evaluated in other peer-reviewed studies, BMD
29 analysis was performed using these data. For the purposes of deriving an RfC for methanol from
30 developmental endpoints using the BMD method and rat data, decreases in brain weight at
31 6 weeks of age in the more sensitive gender, males, exposed throughout gestation and continuing
32 into the FI generation (both through lactation and inhalation routes) were utilized. Decreases in
33 brain weight at 6 weeks (gestation and postnatal exposure), rather than those seen at 3 and 8
34 weeks, were chosen as the basis for the RfC derivation because they resulted in lower estimated
35 BMDs and BMDLs. Decreased brain weights in male rats at 8 weeks age after gestation-only
5-9 DRAFT—DO NOT CITE OR QUOTE
-------
1 exposure were not utilized because they were less severe at the same dose level (1,000 ppm)
2 compared to gestation and postnatal exposure.
3 The first step in the current BMD analysis is to convert the inhalation doses, given
4 as ppm values from the studies, to an internal dose metric using the EPA PBPK model (see
5 Section 3.4). For decreased brain weight in male rats, AUC of methanol in blood (hr x mg/L) is
6 chosen as the appropriate internal dose metric for the reasons discussed in Section 5.1.2.1.
7 Predicted AUC values for methanol in the blood of rats are summarized in Table 5-2. These
8 AUC values are then used as the dose metric for the BMD analysis of decreased brain weight in
9 male rats.73 The full details of this analysis are reported in Appendix C. More details concerning
10 the PBPK modeling were presented in Section 3.4.
Table 5-2. The EPA PBPK model estimates of methanol blood levels (AUC) in rats
following inhalation exposures
Exposure level (ppm)
500
1,000
2,000
Methanol in blood AUC
(hr x mg/L)a in Rats
79.2
226.7
967.8
aAUC values were obtained by simulating 22 hr/day exposures for 5 days and calculated for the last 24 hours of that
period.
11 The current draft BMD technical guidance (U.S. EPA, 2000b) suggests that, in the
12 absence of knowledge as to what level of response to consider adverse, a change in the mean
13 equal to one S.D. from the control mean can be used as a BMR for continuous endpoints.
14 However, it has been suggested that other BMRs, such as 5% change relative to estimated
15 control mean, are also appropriate when performing BMD analyses on fetal weight change as a
16 developmental endpoint (Kavlock et al., 1995). Therefore, both a one S.D. change from the
17 control mean and a 5% change relative to estimated control mean were considered (see Appendix
18 C for RfC derivations using alternative BMRs). For this endpoint, a one S.D. change from the
19 control mean returned the lowest BMDL estimates and was considered the most suitable BMR
20 for use in the RfC derivation. All models were fit using restrictions and option settings
21 suggested in the draft EPA BMD technical guidance document (U.S. EPA, 2000b).
22 A summary of the results most relevant to the development of a POD using the BMD
23 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 6 weeks in male
24 rats exposed to methanol throughout gestation and continuing into the FI generation is provided
25 in Table 5-3. BMDL values in Table 5-3 represent the 95% lower-bound confidence limit on the
26 AUC estimated to result in a mean that is one S.D. from the control mean. There is a 2.5-fold
All BMD assessments in this review were performed using BMDS version 2.1. A copy of the BMDS can be
obtained at: .
5-10 DRAFT—DO NOT CITE OR QUOTE
-------
1 range of BMDL estimates from adequately fitting models, indicating considerable model
2 dependence. In addition, the fit of the Hill and more complex Exponential models is better than
3 the other models in the dose region of interest as indicated by a lower scaled residual at the dose
4 group closest to the BMD (0.09 versus -0.67 or -0.77) and visual inspection. In accordance with
5 draft EPA BMD Technical Guidance (EPA, 2000b), the BMDL from the Hill model (bolded), is
6 selected as the most approriate basis for an RfC derivation because it results in the lowest BMDL
7 from among a broad range of BMDLs and provides a superior fit in the low dose region nearest
8 the BMD. The Hill model dose-response curve for decreased brain weight in male rats is
9 presented in Figure 5-1, with response plotted against the chosen internal dose metric of AUC of
10 methanol in rats. The BMDLiso was determined to be 90.9 hr x mg/L using the 95% lower
11 confidence limit of the dose-response curve expressed in terms of the AUC for methanol in
12 blood.
Table 5-3. Comparison of benchmark dose modeling results for decreased brain weight
in male rats at 6 weeks of age using modeled AUC of methanol as a dose metric
Model
Linear
2nd degree polynomial
3rd degree polynomial
Power
Hillb
Exponential 2
Exponential 3
Exponential 4
Exponential 5
BMD1SD (AUC,
hr x mg/L) a
278.30
278.30
278.30
278.30
170.57
260.94
260.94
172.08
172.08
BMDL1SD
(AUC,
hr x mg/L)a
225.30
225.30
225.30
225.30
90.93
209.10
209.10
96.93
96.93
/7-value
0.5376
0.5376
0.5376
0.5376
0.8366
0.612
0.612
0.8205
0.8205
AICC
-203.84
-203.84
-203.84
-203.84
-203.04
-204.10
-204.10
-203.10
-203.10
Scaled residual"1
-0.77
-0.77
-0.77
-0.77
0.09
-0.67
-0.67
0.09
0.09
"The BMDL is the 95% lower confidence limit on the AUC estimated to decrease brain weight by 1 control mean
S.D. using BMDS model options and restrictions suggested by EPA BMD technical guidance (U.S. EPA, 2000b).
bln accordance with draft EPA BMD Technical Guidance (EPA, 2000), the BMDL from the Hill model (bolded) is
chosen for us in an RfC derivation because it is the lowest of a broad range of BMDL estimates from adequately
fitting models and because the Hill model provides good fit in the dose region of interest as indicated by a relatively
low scaled residual at the dose group closest to the BMD (0.09 versus -0.67 or -0.77).
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
Vd residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest to
the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD. Residuals
that exceed 2.0 in absolute value should cause one to question model fit in this region.
Source: NEDO(1987).
5-11
DRAFT—DO NOT CITE OR QUOTE
-------
Hill Model with 0.95 Confidence Level
8-
0)
or
1.85
1.8
1.75
1.7
1.65
1.6
1.55
1.5
1.45
09:05 08/24 2009
200
400 600
dose
800
1000
Figure 5-1. Hill model BMD plot of decreased brain weight in male rats at 6 weeks age
using modeled AUC of methanol in blood as the dose metric, 1 control mean S.D.
1
2 Once the BMDLiso was obtained in units of hr x mg/L, it was used to derive a chronic
3 RfC. The first step is to calculate the HEC using the PBPK model described in Appendix B. An
4 algebraic equation is provided (Equation 1 of Appendix B) that describes the relationship
5 between predicted methanol AUC and the human equivalent inhalation exposure concentration
6 (HEC) in ppm.
7 BMDLHEC (ppm) = 0.0224*BMDLlSD+(1334*BMDLlSD)/(794+ BMDL1SD)
8 BMDLHEC (ppm) = 0.0224*90.9+( 1334*90.9)7(794+ 90.9) = 139 ppm
9 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
10 converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
11
HEC (mg/m3) = 1.31 x 139 ppm = 182 mg/m3
5-12
DRAFT—DO NOT CITE OR QUOTE
-------
5.1.3. RfC Derivation - Including Application of Uncertainty Factors
5.1.3.1. Comparison Bet ween En dp oin ts an d BMDL Modeling Approa ch es
1 A summary of the PODs for the various developmental endpoints and BMD modeling
2 approaches considered for the derivation of an RfC, along with the UFs applied74 and the
3 conversion to an HEC, are presented in Table 5-4 and graphically compared in Figure 5-2 (see
4 Appendix C for details). Information is presented that compares the use of different endpoints
5 (i.e., cervical rib, decreased brain weight, and increased latency of VDR) and different methods
6 (i.e., different BMR levels) for estimating the POD. These comparisons are presented to inform
7 the analysis of uncertainty surrounding these choices. Each approach considered for the
8 determination of the POD has strengths and limitations, but when considered together for
9 comparative purposes they allow for a more informed determination for the POD for the
10 methanol RfC.
11 A 10% extra risk BMR is adequate for most traditional bioassays using 50 animals per
12 dose group. A smaller BMR of 5% extra risk can sometimes be justified for developmental
13 studies (e.g., Rogers et al., 1993a), because they generally involve a larger number of subjects.
14 Reference values estimated for cervical rib incidence in mice using Cmax as the dose metric were
15 13.6 and 10.4 mg/m3 using BMDLio and BMDLos PODs, respectively (see Appendix D for
16 discussion of choice of Cmax as the appropriate dose metric for incidence of cervical rib in mice).
17 The reference value estimated for alterations in sensorimotor development and performance as
18 measured by the VDR test in female monkeys using AUC as the dose metric was 1.7 mg/m3
19 using the BMDLSo as the POD. As discussed in Section 4.4.2, confidence in this endpoint is
20 reduced by a marginal dose-response trend in one sex (females) and a limited sample size.
21 Although the VDR test demonstrates that prenatal and continuing postnatal exposure to methanol
22 can result in neurotoxicity, the use of such statistically borderline results is not warranted in the
23 derivation of the RfC, given the availability of better dose-response data in other species.
24 Decreases in brain weight at 6 weeks of age in male rats exposed during gestation and
25 throughout the FI generation using AUC as the dose metric yield the reference values of 1.8 and
26 2.4 mg/m3 for BMRs of one S.D. from the control mean and 5% change relative to control mean,
27 respectively. Because decreases in brain weight in male rats at 6 weeks postbirth resulted in a
28 clear dose response and returned RfC estimates lower than or approximate to the other endpoints
29 considered, it was chosen as the critical endpoint. One S.D. from the control mean was chosen
30 as the appropriate level of response (BMR) for the calculation of the RfC because it is the
31 standard recommended by EPA's draft technical guidance (U.S. EPA, 2000b) and yields a lower
32 BMDL than 5% relative deviance for this data set. Thus, the RfC is:
74 The rationale for the selection of these UFs is discussed later in Section 5.1.3.
5-13 DRAFT—DO NOT CITE OR QUOTE
-------
RfC = PODHEc - UF = 182 mg/m3 + 100 = 2 mg/m3 (rounded to one significant figure)
Table 5-4. Summary of PODs for critical endpoints, application of UFs and conversion
to HEC values using BMD and PBPK modeling
BMDL
HEC (mg/m3)c
UFHd
UFAe
UFD
UFS
UFL
UF-TOTAL
RfC (mg/m3)
Rogers et al. (1993a)
BMDL10 mouse
cervical rib Cmax
94.3 mg/L
1360
10
3
3
1
1
100
13.6
BMD L05 mouse
cervical rib Cmax
44.7 mg/L
1036
10
3
3
1
1
100
10.4
Burbacher et al.
(1999a,1999b)
BMDL1SD
female monkey
VDRa AUC
81.7 hrxmg/L
165
10
3
3
1
1
100
1.7
NEDO (1987)
BMDLos
rat brain wt. b
AUC
123. 9 hrxmg/L
240
10
3
3
1
1
100
2.4
BMDL1SD
rat brain wt. b
AUC
90.9 hrxmg/L
182
10
3
3
1
1
100
1.8
aVDR = test of sensorimotor development as measured by age from birth at achievement of test criterion for
grasping a brightly colored object.
bBrain weight at 6 weeks postbirth, multiple routes of exposure (whole gestation, lactation, inhalation)
cThe PBPK model used for this HEC estimate is described in Appendix B. An algebraic equation (Equation 1 of
Appendix B) describes the relationship between predicted methanol AUC and the human equivalent inhalation
exposure concentration (HEC) in ppm. This equation can also be used to estimate model predictions for HECs
from Cmax values because Cmax values and AUC values were estimated at steady-state for constant 24 hours
exposures (i.e., AUC = 24 xCmax). The ppm HEC estimate is then converted to mg/m3 by multiplying by 1.31.
dThe rationale for the selection of these UFs is discussed in Section 5.1.3 below.
These uncertainty factor (UF) acronyms are defined in Sections 5.1.2.1.1 to 5.1.2.1.4.
5-14
DRAFT—DO NOT CITE OR QUOTE
-------
10000
1000
£ 100
Rogers etal., 1993a;GD Rogers et al., 1993a; GD Burbacher et al., 1999a,b; NEDO, 1987; 1-gen repro NEDO, 1987; 1-gen repro
6-15 mouse study; 6-15 mouse study; whole gestation monkey study; 6 week male brain study; 6 week male brain
cervical rib; cervical rib BMDL(HEC)05 study; visually directed weight; BWDL(HEC)05 weight; BMDL(HEC)1SD
BMDL(HEC)10 reaching; BMDL(HEC)1SD
POD I I Interspecies Extrapolation
Inter-individual Variation I I Database
RfC
Figure 5-2. PODs (in mg/m3) for selected endpoints with corresponding applied UFs
(chosen RfC value is circled)
5.1.3.2. Applica tion of UFs
1 UFs are applied to the POD, identified from the rodent data, to account for recognized
2 uncertainties in extrapolation from experimental conditions to the assumed human scenario (i.e.,
3 chronic exposure over a lifetime). A composite UF of 100-fold (10-fold for interindividual
4 variation, 3-fold for residual toxicodynamic differences associated with animal-to-human
5 extrapolation, and 3-fold for database uncertainty) was applied to the POD for the derivation of
6 the RfC, as described below.
7 5.1.3.2.1. Interindividual Variation UFH. A factor of 10 was applied to account for variation in
8 sensitivity within the human population (UFH). The UF of 10 is commonly considered to be
9 appropriate in the absence of convincing data to the contrary. The data from which to determine
10 the potential extent of variation in how humans respond to chronic exposure to methanol are
11 limited, given the complex nature of the developmental endpoint employed and uncertainties
12 surrounding the importance of metabolism to the observed teratogenic effects. Susceptibility to
13 methanol is likely to involve intrinsic and extrinsic factors. Some factors may include alteration
14 of the body burden of methanol or its metabolites, sensitization of an individual to methanol
5-15
DRAFT—DO NOT CITE OR QUOTE
-------
1 effects, or augmentation of underlying conditions or changes in processes that share common
2 features with methanol effects. Additionally, inherent differences in an individual's genetic
3 make-up, diet, gender, age, or disease state may affect the pharmacokinetics and
4 pharmacodynamics of methanol, influencing susceptibility intrinsically. Co-exposure to a
5 pollutant that alters metabolism or clearance or that adds to background levels of metabolites
6 may also affect the pharmacokinetics and pharmacodynamics of methanol, influencing
7 susceptibility extrinsically (see Section 4.9). The determination of the UF for human variation is
8 supported by several types of information, including information concerning background levels
9 of methanol in humans, variation in pharmacokinetics revealed through human studies and from
10 PBPK modeling, variation of methanol metabolism in human tissues, and information on
11 physiologic factors (including gender and age), or acquired factors (including diet and
12 environment) that may affect methanol exposure and toxicity.
13 In using the AUC of methanol in blood as the dose metric for derivation of health
14 benchmarks for methanol, the assumption is made that concentrations of methanol in blood over
15 time are related to its toxicity, either through the actions of the parent or it subsequent
16 metabolism. However, the formation of methanol's metabolites has been shown in humans to be
17 carried out by enzymes that are inducible, highly variable in activity, polymorphic, and to also be
18 involved in the metabolism of other drugs and environmental pollutants. Hence, differences in
19 the metabolism of methanol that are specific for target tissue, gender, age, route of
20 administration, and prior exposure to other environmental chemicals may give a different pattern
21 of methanol toxicity if metabolism is required for that toxicity. Eighty-five percent of Asians
22 carry an atypical phenotype of ADH that may affect their ability to metabolize methanol
23 (Agarwal, 2001; Bosron and Li, 1986; Pietruszko, 1980). Also, polymorphisms in ADH3
24 occurring in the promoter region reduce the transcriptional activity in vitro nearly twofold,
25 although no studies have reported differences in ADH3 enzyme activity in humans (Hedberg
26 etal.,2001).
27 Although data on the specific potential for increased susceptibility to methanol are
28 lacking, there is information on PK and pharmacodynamic factors suggesting that children may
29 have differential susceptibility to methanol toxicity (see Section 4.10.1). Thus, there is
30 uncertainty in children's responses to methanol that should be taken into consideration for
31 derivation of the UF for human variation that is not available from either measured human data
32 or PBPK modeling analyses. The enzyme primarily responsible for metabolism of methanol in
33 humans, ADH, has been reported to be reduced in activity in newborns. Differences in
34 pharmacokinetics include potentially greater pollutant intake due to greater ventilation rates,
35 activity, and greater intake of liquids in children. In terms of differences in susceptibility to
36 methanol due to pharmacodynamic considerations, the substantial anatomical, physiologic, and
37 biochemical changes that occur during infancy, childhood, and puberty suggest that there are
5-16 DRAFT—DO NOT CITE OR QUOTE
-------
1 developmental periods in which the endocrine, reproductive, immune, audiovisual, nervous, and
2 other organ systems may be especially sensitive.
3 There are some limited data from short-term exposure studies in humans and animal
4 experiments that suggest differential susceptibility to methanol on the basis of gender. Gender
5 can provide not only different potential targets for methanol toxicity but also differences in
6 methanol pharmacokinetics and pharmacodynamics. NEDO (1987) reported that in rats exposed
7 to methanol pre- and postnatally, 6- and 8-week-old male progeny had significantly lower brain
8 weights at 1,000 ppm, whereas females only showed decreases at 2,000 ppm. In general, gender-
9 related differences in distribution and clearance of methanol may result from the greater muscle
10 mass, larger body size, decreased body fat, and increased volumes of distribution in males
11 compared to females.
12 5.1.3.2.2. Animal-to-Human Extrapolation UFA. Afactor of 3 was applied to account for
13 uncertainties in extrapolating from rodents to humans. Application of a full UF of 10 would
14 depend on two areas of uncertainty: toxicokinetic and toxicodynamic uncertainty. In this
15 assessment, the toxicokinetic component is largely addressed by the determination of a HEC
16 through the use of PBPK modeling. Given the chosen dose metric (AUC for methanol blood),
17 uncertainties in the PBPK modeling of methanol are not expected to be greater for one species
18 than another. The analysis of parameter uncertainty for the PBPK modeling performed for
19 human, mouse, and rat data gave similar results as to how well the model fit the available data.
20 Thus, the human and rodent PBPK model performed similarly using this dose metric for
21 comparisons between species. As discussed in Section 5.3 below, uncertainty does exist
22 regarding the relation of maternal blood levels estimated by the model to fetal and neonatal
23 blood levels that would be obtained under the (gestational, postnatal and lactational) exposure
24 scenario employed in the critical study. However, at environmentally relevant exposure levels, it
25 is assumed that the ratio of the difference in blood concentrations between a human infant and
26 mother would be similar to and not significantly greater than the difference between a rat dam
27 and its fetus. Key parameters and factors which determine the ratio of fetal or neonatal human
28 versus mother methanol blood levels either do not change significantly with age (partition
29 coefficients, relative blood flows) or scale in a way that is common across species
30 (allometrically). For this reason and because EPA has confidence in the ability of the PBPK
31 model to accurately predict adult blood levels of methanol, the PK uncertainty is reduced and a
32 value of 1 was applied. Rodent-to-human pharmacodynamic uncertainty is covered by a factor
33 of 3, as is the practice for deriving RfCs (U.S. EPA, 1994b). Therefore, a factor of 3 is used for
34 interspecies uncertainty.
35 5.1.3.2.3. Database UFD. A database UF of 3 was applied to account for deficiencies in the
36 toxicity database. The database for methanol toxicity is quite extensive: there are chronic and
5-17 DRAFT—DO NOT CITE OR QUOTE
-------
1 developmental toxicity studies in rats, mice, and monkeys, a two-generation reproductive
2 toxicity study in rats, and neurotoxicity and immunotoxicity studies. However, there is
3 uncertainty regarding which test species is most relevant to humans. In addition, limitations of
4 the developmental toxicity database employed in this assessment include gaps in testing and
5 imperfect study design, reporting, and analyses. Developmental studies were conducted at levels
6 inducing maternal toxicity, a full developmental neurotoxicity test (DNT) in rodents has not been
7 performed and is warranted given the critical effect of decreased brain weight, there are no
8 chronic oral studies in mice, and chronic and developmental studies in monkeys were generally
9 inadequate for quantification purposes, for reasons discussed in Section 5.1.1.1. Problems of
10 interpretation of developmental and reproductive studies also arise given the dose spacing
11 between lowest and next highest level. For these reasons, an UF of 3 was applied to account for
12 deficiencies in the database.
13 5.1.3.2.4. Extrapolation from Subchronic to Chronic andLOAEL-toNOAEL
14 Extrapolation UFs. AUF was not necessary to account for extrapolation from less than chronic
15 results because developmental toxicity (cervical rib and decreased brain weight) was used as the
16 critical effect. The developmental period is recognized as a susceptible lifestage where exposure
17 during certain time windows is more relevant to the induction of developmental effects than
18 lifetime exposure (U.S. EPA, 1991).
19 A UF for LOAEL-to-NOAEL extrapolation was not applied because BMD analysis was
20 used to determine the POD, and this factor was addressed as one of the considerations in
21 selecting the BMR. In this case, a BMR of one S.D. from the control mean in the critical effect
22 was selected based on the assumption that it represents a minimum biologically significant
23 change.
5.1.4. Previous RfC Assessment
24 The health effects data for methanol were assessed for the IRIS database in 1991 and
25 were determined to be inadequate for derivation of an RfC.
5.2. ORAL RfD
26 In general, the RfD is an estimate of a daily exposure to the human population (including
27 susceptible subgroups) that is likely to be without an appreciable risk of adverse health effects
28 over a lifetime. It is derived from a POD, generally the statistical lower confidence limit on the
29 BMDL, with uncertainty/variability factors applied to reflect limitations of the data used. The
30 RfD is expressed in terms of mg/kg-day of exposure to an agent and is derived by a similar
31 methodology as is the RfC. Ideally, studies with the greatest duration of exposure and conducted
32 via the oral route of exposure give the most confidence for derivation of an RfD. For methanol,
33 the oral database is currently more limited than the inhalation database. With the development of
5-18 DRAFT—DO NOT CITE OR QUOTE
-------
1 PBPK models for methanol, the inhalation database has been used to help bridge data gaps in the
2 oral database to derive an RfD.
5.2.1. Choice of Principal Study and Critical Effect-with Rationale and Justification
3 No studies have been reported in which humans have been exposed subchronically or
4 chronically to methanol by the oral route of exposure and thus, would be suitable for derivation
5 of an oral RfD. Data exist regarding effects from oral exposure in experimental animals, but
6 they are more limited than data from the inhalation route of exposure (see Sections 4.2, 4.3, and
7 4.4).
8 Only 2 oral studies of 90-days duration or longer in animals have been reported (Soffritti
9 et al., 2002a; U.S. EPA, 1986c) for methanol. EPA (1986c) reported that there were no
10 differences in body weight gain, food consumption, or gross or microscopic evaluations in
11 Sprague-Dawley rats gavaged with 100, 500, or 2,500 mg/kg-day versus control animals. Liver
12 weights in both male and female rats were increased, although not significantly, at the 2,500
13 mg/kg-day dose level, suggesting a treatment-related response despite the absence of
14 histopathologic lesions in the liver. Brain weights of high-dose group males and females were
15 significantly less than control animals at terminal (90 days) sacrifice. The data were not reported
16 in adequate detail for dose-response modeling and BMD estimation. Based primarily on the
17 qualitative findings presented in this study, the 500 mg/kg-day dose was deemed to be a
18 NOAEL.75
19 The only lifetime oral study available was conducted by Soffiritti et al. (2002a) in
20 Sprague-Dawley rats exposed to 0, 500, 5,000, 20,000 ppm (v/v) methanol, provided ad libitum
21 in drinking water. Based on default, time-weighted average body weight estimates for Sprague-
22 Dawley rats (U.S. EPA, 1988), average daily doses of 0, 46.6, 466, and 1,872 mg/kg-day for
23 males and 0, 52.9, 529, 2,101 mg/kg-day for females were reported by the study authors. All rats
24 were exposed for up to 104 weeks, and then maintained until natural death. The authors report
25 no substantial changes in survival nor was there any pattern of compound-related clinical signs
26 of toxicity. The authors did not report noncancer lesions, and there were no reported compound-
27 related signs of gross pathology or histopathologic lesions indicative of noncancer toxicological
28 effects in response to methanol.
29 Five oral studies investigated the reproductive and developmental effects of methanol in
30 rodents (Aziz et al., 2002; Fu et al., 1996; Sakanashi et al., 1996; Rogers et al., 1993a; Infurna
31 and Weiss, 1986), including three studies that investigated the influence of FAD diets on the
32 effects of methanol exposures (Aziz et al., 2002; Fu et al., 1996; Sakanashi et al., 1996). Infurna
33 and Weiss (1986) exposed pregnant Long-Evans rats to 2,500 mg/kg-day in drinking water on
34 either GDIS-GDI7 or GD17-GD19. Litter size, pup birth weight, pup postnatal weight gain,
75 EPA (1986c) did not report details required for a BMD analysis such as standard deviations for mean responses.
5-19 DRAFT—DO NOT CITE OR QUOTE
-------
1 postnatal mortality, and day of eye opening were not different in treated animals versus controls.
2 Mean latency for nipple attachment and homing behavior (ability to detect home nesting
3 material) were different in both methanol treated groups. These differences were significantly
4 different from controls. Rogers et al. (1993a) exposed pregnant CD-I mice via gavage to 4 g/kg-
5 day methanol, given in 2 equal daily doses. Incidence of cleft palate and exencephaly was
6 increased following maternal exposure to methanol. Also, an increase in totally resorbed litters
7 and a decrease in the number of live fetuses per litter were observed.
8 Aziz et al. (2002), Fu et al. (1996), and Sakanashi et al. (1996) investigated the role of
9 folic acid in methanol-induced developmental neurotoxicity. Like Rogers et al. (1993a), the
10 former 2 studies observed that an oral gavage dose of 4-5 g/kg-day during GD6-GD15 or GD6-
11 GD10 resulted in an increase in cleft palate in mice fed sufficient folic acid diets, as well as an
12 increase in resorptions and a decrease in live fetuses per litter. Fu et al. (1996) also observed an
13 increase in exencephaly in the FAS group. Both studies found that an approximately 50%
14 reduction in maternal liver folate concentration resulted in an increase in the percentage of litters
15 affected by cleft palate (as much as threefold) and an increase in the percentage of litters affected
16 by exencephaly (as much as 10-fold). Aziz et al. (2002) exposed rat dams throughout their
17 lactation period to 0, 1, 2, or 4% v/v methanol via the drinking water, equivalent to
18 approximately 480, 960 and 1,920 mg/kg-day.76 Pups were exposed to methanol via lactation
19 from PND1-PND21. Methanol treatment at 2% and 4% was associated with significant
20 increases in activity (measured as distance traveled in a spontaneous locomotor activity test) in
21 the FAS group (13 and 39%, respectively) and most notably, in the FAD group (33 and 66%,
22 respectively) when compared to their respective controls. At PND45, the CAR in FAD rats
23 exposed to 2% and 4% methanol was significantly decreased by 48% and 52%, respectively,
24 relative to nonexposed controls. In the FAS group, the CAR was only significantly decreased in
25 the 4% methanol-exposed animals and only by 22% as compared to their respective controls.
5.2.1.1. Expansion of the Oral Database by Route-to-Route Extrapolation
26 Given the oral database limitations, including the limited reporting of noncancer findings
27 in the subchronic (U.S. EPA, 1986c) and chronic studies (Soffritti et al., 2002a) of rats and the
28 high-dose levels used in the two rodent developmental studies, EPA has derived an RfD by using
29 relevant inhalation data and route-to-route extrapolation with the aid of the EPA PBPK model
30 (see Sections 3.4 and 5.1). Several other factors support use of route-to-route extrapolation for
31 methanol. The limited data for oral administration indicate similar effects as reported via
32 inhalation exposure (e.g., the brain and fetal skeletal system are targets of toxicity). Methanol
33 has been shown to be rapidly and well-absorbed by both the oral and inhalation routes of
76 Assuming that Wistar rat drinking water consumption is 60 mL/kg-day (Rogers et al., 2002), 1% methanol in
drinking water would be equivalent to 1% x 0.8 g/mL x 60 mL/kg-day = 0.48 g/kg-day = 480 mg/kg-day.
5-20 DRAFT—DO NOT CITE OR QUOTE
-------
1 exposure (CERHR, 2004; Kavet and Nauss, 1990). Once absorbed, methanol distributes rapidly
2 to all organs and tissues according to water content, regardless of route of exposure.
3 As with the species-to-species extrapolation used in the development of the RfC, the dose
4 metric used for species-to-species and route-to-route extrapolation of inhalation data to oral data
5 is the AUC of methanol in blood. Simulations for human oral methanol exposure were
6 conducted using the model parameters as previously described for human inhalation exposures,
7 with human oral kinetic/absorption parameters from Sultatos et al. (2004) (i.e., KAS = 0.2,
8 KSI = 3.17, and KAI = 3.28). Human oral exposures were assumed to occur during six drinking
9 episodes during the day, at times 0, 3, 5, 8, 11, and 15 hours from the first ingestion of the day.
10 For example, if first ingestion occurred at 7 am, these would be at 7 am, 10 am, 12 noon, 3 pm, 6
11 pm, and 10 pm. Each ingestion event was treated as occurring over 3 minutes, during which the
12 corresponding fraction of the daily dose was infused into the stomach lumen compartment. The
13 fraction of the total ingested methanol simulated at each of these times was 25%, 10%, 25%,
14 10%, 25%, and 5%, respectively. Six days of exposure were simulated to allow for any
15 accumulation (visual inspection of plots showed this to be finished by the 2nd or 3rd day), and
16 the results for the last 24 hours were used. Dividing the exposure into more and smaller episodes
17 would decrease the estimated peak concentration but have little effect on AUC. This dose metric
18 was used for dose-response modeling to derive the POD, expressed as a BMDL. The BMDL
19 was then back-calculated using the EPA PBPK model to obtain an equivalent oral drinking water
20 dose in terms of mg/kg-day.
5.2.2. RfD Derivation-Including Application ofUFs
5.2.1.2. Consider a tion oflnhala tion Da ta
21 Inhalation studies considered for derivation of the RfC are used to supplement the oral
22 database using the route-to-route extrapolation, as previously described. BMD approaches were
23 applied to the existing inhalation database, and the EPA PBPK model was used for species-to-
24 species extrapolations. The rationale and approach for determining the RfC is described above
25 (Section 5.1), and the data used to support the derivation of the RfC were extrapolated using the
26 EPA PBPK model to provide an oral equivalent POD.
5.2.1.3. Selection of Critical Effect(s) from Inhalation Data
27 Methanol-induced effects on the brain in rats (weight decrease) and fetal axial skeletal
28 system in mice (cervical ribs and cleft palate) were consistently observed at lower levels, than
29 other targets, in the oral and inhalation databases. Analysis of inhalation developmental toxicity
30 studies shows lower BMDLs for decreased male brain weight in rats exposed throughout
31 gestation and the FI generation (NEDO, 1987) than BMDLs associated with the fetal axial
32 skeletal system in mice. Therefore, as was noted in Section 5.1.3.1, the BMDL for decreases in
5-21 DRAFT—DO NOT CITE OR QUOTE
-------
1 brain weight in male rats is chosen to serve as the basis for the route-to-route extrapolation and
2 cal cul ati on of the RfD.
5.2.1.4. Selection ofth e POD
3 The BMDL chosen for the RfC is used to determine the POD for the RfD. This value is
4 based on a developmental toxicity dataset that includes in utero and postnatal exposures and is
5 below the range of estimates for other developmental datasets consisting of exposure only
6 throughout organogenesis. The neonatal brain is the target organ chosen for derivation of the
7 RfC. The BMDL for the RfC (AUC of 90.9 hr x mg/L methanol in blood) is converted using the
8 EPA model to a human equivalent oral exposure of 38.5 mg/kg-day.77
5.2.2. RfD Derivation-Application of UFs
9 In an approach consistent with the RfC derivation, UFs are applied to the oral POD of
10 38.5 mg/kg-day to address interspecies extrapolation, intraspecies variability, and database
11 uncertainties for the RfD. Because the same dataset, endpoint, and PBPK model used to derive
12 the RfC were also used to calculate the oral POD, the total UF of 100 is applied to the BMDL of
13 38.5 mg/kg-day to yield an RfD of 1.12 mg/kg-day for methanol.
14 RfD = 38.5 mg/kg-day + 100 = 0.4 mg/kg-day (rounded to one significant figure)
5.2.3. Previous RfD Assessment
15 The previous IRIS assessment for methanol included an RfD of 0.5 mg/kg-day that was
16 derived from a EPA (1986c) subchronic oral study in which Sprague-Dawley rats (30/sex/dose)
17 were gavaged daily with 0, 100, 500, or 2,500 mg/kg-day of methanol. There were no
18 differences between dosed animals and controls in body weight gain, food consumption, gross or
19 microscopic evaluations. Elevated levels of SGPT, serum alkaline phosphatase (SAP), and
20 increased but not statistically significant liver weights in both male and female rats suggest
21 possible treatment-related effects in rats dosed with 2,500 mg methanol/kg-day, despite the
22 absence of supportive histopathologic lesions in the liver. Brain weights of both high-dose group
23 males and females were significantly less than those of the control group. Based on these
24 findings, 500 mg/kg-day of methanol was considered a NOAEL in this rat study. Application of
25 a 1,000-fold UF (interspecies extrapolation, susceptible human subpopulations, and subchronic
26 to chronic extrapolation) yielded an RfD of 0.5 mg/kg-day.
77 The PBPK model used for this HEC estimate is described in Appendix B. An algebraic equation is provided
(Equation 2) that describes the relationship between predicted methanol AUC and the HED in mg/kg-day.
5-22 DRAFT—DO NOT CITE OR QUOTE
-------
5.3. UNCERTANTIES IN THE INHALATION RfC AND ORAL RfD
1 The following is a more extensive discussion of the uncertainties associated with the RfC
2 and RfD for methanol beyond that which is addressed quantitatively in Sections 5.1.2, 5.1.3, and
3 5.2.2. A summary of these uncertainties is presented in Table 5-5.
Table 5-5. Summary of uncertainties in methanol noncancer risk assessment
Consideration
Choice of endpoint
Choice of dose metric
Choice of model for
BMDL derivation
Choice of animal-to-
human extrapolation
method
Statistical uncertainty
at POD (sampling
variability due to
bioassay size)
Choice of bioassay
Choice of
species/gender
Human population
variability
Potential Impact
Use of other endpoint
could t RfC by up to
~5-fold (see Table 5-4
and Section 5. 3.1)
Alternatives could t
or | RfC/D (e.g., use
of Cmax increased RfC
by -20%)
Use of a linear model
could t RfC by -2.5-
fold (see Table 5-3)
Alternatives could t
or | RfC/D (e.g., use
of standard dosimetry
assumption would t
RfC by -2-fold; see
Section 5. 3. 4)
POD would be -90%
higher if BMD were
used
Alternatives could t
RfC/D
RfC would be t or |
if based on another
species/gender
RfC could | or t if
another value of the
UF was used
Decision
RfC is based on the
most sensitive and
quantifiable endpoint,
decreased brain weight
in male rats exposed
pre- and postnatally
AUC for methanol in
arterial blood
Hill model used
A PBPK model was
used to extrapolate
animal to human
concentrations
A BMDL was used as
the POD
NEDO (1987)
RfC is based on the
most sensitive and
quantifiable endpoint
(I brain weight) in the
most sensitive species
and gender adequately
evaluated (male rats).
10-fold uncertainty
factor applied to derive
the RfC/RfD values
Justification
Chosen endpoint is considered the most
relevant due to its biological significance,
and consistency across a developmental
and a subchronic study in rats and with the
observation of other developmental
neurotoxicities reported in monkeys.
AUC was selected as the most appropriate
dose metric because it incorporates time
(brain weight is sensitive to both the level
and duration of exposure) and better
reflects exposure within a given day.
Hill model gave lowest of a broad range of
BMDL estimates from adequate models
and provides good fit in low dose region.
Use of a PBPK model reduced uncertainty
associated with the animal to human
extrapolation. AUC blood levels of
methanol is an appropriate dose metric and
a peer-reviewed PBPK model that
estimates this metric was verified by EPA
using established (U.S. EPA, 2006a)
methods and procedures
Lower bound is 95% CI of administered
exposure
Alternative bioassays were available, but
the chosen bioassay was adequately
conducted and reported and resulted in the
most sensitive and reliable BMDL for
derivation of the RfC.
Choice of female rats would have resulted
in a higher RfC/D. Effects in mice also
yield higher RfCs. Qualitative evidence
from NEDO (1987) and Burbacher et al.
(2004a, 2004b) suggest that monkeys may
be a more sensitive species, but data are not
as reliable for quantification.
10-fold UF is applied because of limited
data on human variability or potential
susceptible subpopulations, particularly
pregnant mothers and their neonates.
5-23
DRAFT—DO NOT CITE OR QUOTE
-------
5.3.1. Choice of Endpoint
1 The impact of endpoint selection on the derivation of the RfC and RfD was discussed in
2 Sections 5.1.3.1 and 5.2.2.2. Potential RfC values considered ranged from 1.7 to 13.6 mg/m3,
3 depending on whether neurobehavioral function in male monkeys, brain weight decrease in male
4 rats, or cervical ribs incidence in mice was chosen as the critical effect for derivation of the POD,
5 with the former endpoint representing the lower end of the RfC range. The use of other
6 endpoints, particularly pre-term births identified in the Burbacher et al. (2004a, 2004b, 1999a,
7 1999b) monkey study, would potentially result in lower reference values, but significant
8 uncertainties associated with those studies preclude their use as the basis for an RfC.
9 Burbacher et al. (2004a, 2004b, 1999a, 1999b) exposedM fascicularis monkeys to 0,
10 262, 786, and 2,359 mg/m3 methanol 2.5 hours/day, 7 days/week during premating/mating and
11 throughout gestation (approximately 168 days). They observed a slight but statistically
12 significant gestation period shortening in all exposure groups that was largely due to C-sections
13 performed in the methanol exposure groups "in response to signs of possible difficulty in the
14 maintenance of pregnancy," including vaginal bleeding. As discussed in Sections 4.3.2 and
15 5.1.1.2, there are questions concerning this effect and its relationship to methanol exposure. An
16 ultrasound was not done to confirm the existence of real fetal or placental problems.
17 Neurobehavioral function was assessed in infants during the first 9 months of life. Two tests out
18 of nine, returned positive results possibly related to methanol exposure. VDR performance was
19 reduced in all treated male infants, and was significantly reduced in the 2,359 mg/m3 group for
20 both sexes and the 786 mg/m3 group for males. However, an overall dose-response trend for this
21 endpoint was only observed in females. As discussed in Section 4.4.2, confidence in this
22 endpoint may have been increased by statistical analyses to adjust for multiple testing (CERHR,
23 2004), but it is a measure of functional deficits in sensorimotor development that is consistent
24 with early developmental CNS effects (brain weight changes discussed above) that have been
25 observed in rats. The Pagan test of infant intelligence indicated small but not significant deficits
26 of performance (time spent looking a novel faces versus familiar faces) in treated infants.
27 Although these results indicate that prenatal and continuing postnatal exposure to methanol can
28 result in neurotoxicity to the offspring, especially when considered in conjunction with the gross
29 morphological effects noted in NEDO (1987), the use of such statistically borderline results is
30 not warranted in the derivation of the RfC, given the availability of better dose-response data in
31 other species.
32 NEDO (1987) also examined the chronic neurotoxicity of methanol inM fascicularis
33 monkeys exposed to 13.1, 131, or 1,310 mg/m3 for up to 29 months. Multiple effects were noted
34 at 131 mg/ m3, including slight myocardial effects (negative changes in the T wave on an EKG),
35 degeneration of the inside nucleus of the thalamus, and abnormal pathology within the cerebral
36 white tissue in the brain. The results support the identification of 13.1 mg/m3 as the NOAEL for
5-24 DRAFT—DO NOT CITE OR QUOTE
-------
1 neurotoxic effects in monkeys exposed chronically to inhaled methanol. However, as discussed
2 in Section 4.2.2.3, there exists significant uncertainty in the interpretation of these results and
3 their utility in deriving an RfC for methanol. These uncertainties include lack of appropriate
4 control group data, limited nature of the reporting of the neurotoxic effects observed, and use of
5 wild-caught monkeys in the study. Thus, while the NEDO (1987) study suggests that monkeys
6 may be a more sensitive species to the neurotoxic effects of chronic methanol exposure than
7 rodents, the substantial deficits in the reporting of data preclude the quantification of data from
8 this study for the derivation of an RfC.
9 The increased incidence of cervical ribs was identified as a biologically significant,
10 potential co-critical effect based on the findings of Rogers et al. (1993a). Mice were exposed to
11 1,000, 2,000, or 5,000 ppm, and incidence of cervical ribs was statistically increased at
12 2,000 ppm. However, given that the reference values for the increased incidence of cervical ribs
13 are estimated to be approximately five times higher than the reference values calculated using
14 decreases in brain weight in male rats (NEDO, 1987), decreased brain weight was chosen as the
15 basis for the derivation of the RfC.
5.3.2. Choice of Dose Metric
16 A recent review of the reproductive and developmental toxicity of methanol by a panel of
17 experts concluded that methanol, not its metabolite formate, is likely to be the proximate
18 teratogen and that blood methanol level is a useful biomarker of exposure (CERHR, 2002;
19 Dormanetal., 1995). The CERHR Expert Panel based their assessment of potential methanol
20 toxicity on an assessment of circulating blood levels (CERHR, 2002). In contrast to the
21 conclusions of the NTP-CERHR panel, in vitro data from Harris et al. (2004, 2003) suggest that
22 the etiologically important substance for embryo dysmorphogenesis and embryolethality was
23 likely to be formaldehyde rather than the parent compound or formate. Although there remains
24 uncertainty surrounding the identification of the proximate teratogen of importance (methanol,
25 formaldehyde, or formate), the dose metric chosen for derivation of an RfC was based on blood
26 methanol levels. This decision was primarily based on evidence that the toxic moiety is not
27 likely to be the formate metabolite of methanol (CERHR, 2004), and evidence that levels of the
28 formaldehyde metabolite following methanol maternal and/or neonate exposure would be lower
29 in the fetus and neonate than in adults. While recent in vitro evidence indicates that
30 formaldehyde is more embryotoxic than methanol and formate, the high reactivity of
31 formaldehyde would limit its unbound and unaltered transport as free formaldehyde from
32 maternal to fetal blood (Thrasher and Kilburn, 2001) (see discussion in Section 3.3). Thus, even
33 if formaldehyde is ultimately identified as the proximate teratogen, methanol would likely play a
34 prominent role, at least in terms of transport to the target tissue. Further discussions of methanol
35 metabolism, dose metric selection, and MOAissues are in Sections 3.3, 4.6, 4.8 and 4.9.2.
5-25 DRAFT—DO NOT CITE OR QUOTE
-------
1 There exists some concern in using the FI generation NEDO (1987) rat study as the basis
2 from which to derive the RfC. This concern mainly arises from issues related to the low
3 confidence that the PBPK model is accurately predicting dose metrics for neonates exposed
4 through multiple and simultaneous routes. The PBPK model was structured to predict internal
5 dose metrics for adult NP animals and was optimized using adult metabolic and physiological
6 parameters. Young animals have very different metabolic and physiological profiles than adults
7 (enzyme activities, respiration rates, etc.). This fact, coupled with multiple routes of exposure,
8 make it likely that the PBPK did not accurately predict the internal dose metrics for the offspring.
9 Stern et al. (1996) reported that when rat pups and dams were exposed together during lactation
10 to 4,500 ppm methanol in air, methanol blood levels in pups from GD6-PND21 were
11 approximately 2.25 times greater than those of dams. This discrepancy persisted until PND48,
12 when postnatal exposure continued to PND52. It is logical to assume that similar differences in
13 blood methanol levels would also be observed in the NEDO (1987) FI study, as the exposure
14 scenario is similar to that of Stern et al. (1996). Differences between pup and dam blood
15 methanol levels might be expected to be slightly greater than twofold in the NEDO (1987) FI
16 study as the exposure was continuous (versus 6 hours/day in the Stern et al. [1996] paper) and
17 lasted for a longer duration (-64 days versus 37). Under a similar scenario, human newborns
18 may experience higher blood levels than their mothers as a result of breast feeding. As has been
19 discussed in Chapter 3, children have a limited capacity to metabolize methanol via ADH;
20 however, there is some evidence that human infants are able to efficiently eliminate methanol at
21 high-exposure levels, possibly via CAT (Tran et al., 2007). At environmentally relevant
22 exposure levels, it is assumed that the ratio of the difference in blood concentrations between
23 infant and mother would not be significantly greater than the twofold difference that has been
24 observed in rats.78 For this reason and because EPA has confidence in the ability of the PBPK
25 model to accurately predict adult blood levels of methanol, the maternal blood methanol levels
26 for the estimation of HECs from the NEDO (1987) study were used as the dose metric.
5.3.3. Choice of Model for BMDL Derivations
27 The Hill model adequately fit the dataset (goodness-of-fit^-value = 0.84). Data points
28 were well predicted near the BMD (scaled residual = 0.09) (see Figure 5-1). There is a 2.5-fold
29 range of BMDL estimates from adequately fitting models, indicating considerable model
30 dependence. The BMDL from the Hill model was selected, in accordance with EPA BMD
31 Technical Guidance (EPA, 2000b), because it results in the lowest BMDL from among a broad
32 range of BMDLs and provides a superior fit in the low dose region nearest the BMD.
78 Key parameters and factors which determine the ratio of fetal or neonatal human versus mother methanol blood
levels either do not change significantly with age (partition coefficients, relative blood flows) or scale in a way that
is common across species (allometrically).
5-26 DRAFT—DO NOT CITE OR QUOTE
-------
5.3.4. Choice of Animal-to-Human Extrapolation Method
1 APBPK model developed by the EPA, adapted from Ward et al. (1997), was used to
2 extrapolate animal-to-human concentrations. An AUC blood level of methanol (90.9 hr x mg/L)
3 associated with a one S.D. change from the control mean for brain weights in rats was estimated
4 using the rat PBPK model. Then the human PBPK model was used to convert back to a human
5 equivalent exposure concentration or a BMCLHEc/isD of 182 mg/m3. If no PBPK models were
6 available, a BMCLHEc/isD of 424 mg/m3 would have been derived by adjusting the 556.5 mg/m3
7 BMCLiso for external exposure concentration for duration and the animal-to-human standard
8 adjustment factor for systemic effects (the ratio of animal and human blood:air partition
9 coefficients). This value is approximately 2-fold higher than the value derived using the PBPK
10 model. However, as discussed above, use of PBPK-estimated maternal blood methanol levels
11 for the estimation of HECs allows for the use of data-derived extrapolations rather than standard
12 methods for extrapolations from external exposure levels.
13 As discussed in Section 3.4, the PBPK models do not describe or account for background
14 levels of methanol, formaldehyde or formate, and background levels were subtracted from the
15 reported data before use in model fitting or validation (if not already subtracted by study
16 authors), as described below. This approach was taken because the relationship between
17 background doses and background responses is not known, because the primary purpose of this
18 assessment is for the determination of noncancer and cancer risk associated with increases in the
19 levels of methanol or its metabolites (e.g., formate, formaldehyde) over background, and because
20 the subtraction of background levels is not expected to have a significant impact on PBPK model
21 parameter estimates as background levels of methanol and its metabolites are low relative to
22 exposure levels used in methanol bioassays.
5.3.5. Route-to-Route Extrapolation
23 To estimate an oral dose POD for decrease in brain weight in rats, a route-to-route
24 extrapolation was performed on the inhalation exposure POD used to derive the RfC. One way
25 to characterize the uncertainty associated with this approach is to compare risk levels (BMDL
26 values) using the dose metric, AUC methanol, for developmental decreases in brain weight
27 derived from 1) an existing oral subchronic study and 2) from a model estimating this metric
28 from an existing inhalation subchronic study. There are currently no oral developmental studies
29 investigating decreases in brain weight available to compare to the risk values estimated using
30 the second procedure. However, the fact that the oral BMDL of 38.5 mg/kg-day estimated in this
31 assessment from the NEDO (1987) inhalation study of neonate rats via a PBPK model is lower
32 than the NOAEL of 500 mg/kg-day identified in EPA (1986c) methanol study of adult rats is
33 consistent with other studies which suggest that fetal/neonatal organisms are a sensitive
34 subpopulation.
5-27 DRAFT—DO NOT CITE OR QUOTE
-------
5.3.6. Statistical Uncertainty at the POD
1 There is uncertainty in the selection of the BMR level. For decreased brain weight in
2 rats, no established standard exists, so a BMR of one S.D. change from the control mean was
3 used. Parameter uncertainty can be assessed through CIs. Each description of parameter
4 uncertainty assumes that the underlying model and associated assumptions are valid. For the
5 Hill model applied to the data for decreased brain weight in rats, there is a degree of uncertainty
6 at the one S.D. level (the POD for derivation of the RfC), with the 95% one-sided lower
7 confidence limit (BMDL) being -50% below the maximum likelihood estimate of the BMD.
5.3.7. Choice of Bioassay
8 The NEDO (1987) study was used for development of the RfC and RfD because it
9 resulted in the lowest BMDL. It was also a well-designed study, conducted in a relevant species
10 with an adequate number of animals per dose group, and with examination of appropriate
11 developmental toxicological endpoints. Developmental (Burbacher et al., 2004a, 2004b, 1999a,
12 1999b) and chronic studies (NEDO, 1987) of methanol have been performed in monkeys. As
13 discussed above in Section 5.3.1 and other sections of this assessment, while the monkey may be
14 a sensitive species for use in the determination of human risk, reporting deficits and study
15 uncertainties preclude their use in the derivation of an RfC.
5.3.8. Choice of Species/Gender
16 The RfC and RfD were based on decreased brain weight at 6 weeks postbirth in male rats
17 (the gender most sensitive to this effect) (NEDO, 1987). This decrease in brain weight also
18 occurs in female rats; however, if the decreased brain weight in female rats had been used, higher
19 RfC and RfD values would have been derived (approximately 66% higher than the male derived
20 values).
5.3.9. Human Population Variability
21 The extent of interindividual variation of methanol metabolism in humans has not been
22 well characterized. As discussed in Section 4.9, there are a number of issues that may lead to
23 sensitive human subpopulations. Potentially sensitive subpopulations would include individuals
24 with polymorphisms in the enzymes involved in the metabolism of methanol and individuals
25 with significant folate deficiencies. Sensitive lifestages would include children and neonates, as
26 they have increased respiration rates compared to adults, which may increase their methanol
27 blood levels compared to adults. Also, children have been shown to have decreased ADH
28 activity relative to adults, thus decreasing their ability to metabolize and eliminate methanol. As
29 demonstrated by these examples, there exists considerable uncertainty pertaining to human
5-28 DRAFT—DO NOT CITE OR QUOTE
-------
1 population variability in methanol metabolism, which provides justification for the 10-fold
2 intraspecies UF used to derive the RfC and RfD.
5.4.5.4. CANCER ASSESSMENT
5.4.1. Oral Exposure
5.4.1.1. Choice of Study/Data—with Rationale and Justification
3 No human data exist that would allow for quantification of the cancer risk of chronic
4 methanol exposure. Table 4-34 summarizes the available experimental animal oral exposure
5 studies of methanol. The Soffritti et al. (2002a) and Apaja (1980) oral studies report effects that
6 show a statistically significant increase in incidence of cancer endpoints in the treated groups
7 versus the control group (pair-wise comparison). As detailed in Section 4.2.1.3, Soffritti et al.
8 (2002a) exposed Sprague-Dawley rats via drinking water to 500-20,000 ppm methanol for
9 104 weeks. Exposure ended at 104 weeks, but the animals were not euthanized and were
10 followed until their natural death. Increased lymphoma responses in multiple organs of male and
11 female rats were the only carcinogenic effects reported in the Soffritti et al. (2002a) methanol
12 drinking water study that are considered dose related and quantifiable. Hepatocellular
13 carcinomas observed in male rats are considered potentially dose related (relative to historical
14 controls) but are not quantifiable due to the lack of a statistically significant dose-response trend.
15 Significant increases reported for head and ear duct carcinomas in male rats were not used
16 because NTP pathologists interpreted a majority of these ear duct responses as being
17 hyperplastic, not carcinogenic, in nature (EFSA, 2006; Hailey, 2004). Apaja (1980) observed
18 significant increases in malignant lymphomas relative to untreated, historical controls in Swiss
19 mice exposed to methanol in drinking water for life. Due to the lack of a concurrent control, the
20 Apaja (1980) study was not considered adequate for derivation of an oral slope factor, but a
21 quantitative analysis of the dose-response data from this study is included in Appendix E for
22 comparison purposes.
5.4.1.2. Dose-Response Data
23 The tumor incidence data selected for modeling were the lympho-immunoblastic
24 lymphomas and the combined lympho-immunoblastic, lymphoblastic and lymphocytic
25 lymphomas in both male and female rats of the Soffritti et al (2002a) study. These lymphomas
26 were combined at the recommendation of NTP pathologists due to their similar histological
27 origin. The incidence of histiocytic sarcomas and myeloid leukemias was not significantly
28 increased in either sex, and the data for these tumors was not combined with the lymphoblastic
29 lymphomas because they are of a different cell line and the combination is not typically
30 evaluated either for statistical significance or dose-response modeling (Hailey, 2004; McConnell,
31 et al., 1986). Table 5-6 gives the lymphoma incidence data from the study which differs slightly
5-29 DRAFT—DO NOT CITE OR QUOTE
-------
1 from the data reported in Soffritti et al. (2002a) in the incidence of lympho-immunoblastic
2 lymphomas in the male 5,000 ppm group.79
Table 5-6. Incidence data for lymphoma, lympho-immunoblastic, and all lymphomas
in male and female Sprague-Dawley rats
Dose (ppm)
Dose
(mg/kg-day)
Number of animals
examined
Lymphoma lympho-
immunoblastic
All lymphomas
combined
Female rats
0
500
5,000
20,000
0
66.0
624.1
2,177
100
100
100
100
9
17
19a
21a
9
19a
20a
22b
Male rats
0
500
5,000
20,000
0
53.2
524
1,780
100
100
100
99
16
24
28a
37b
17
27
29a
38b
Statistically significant by Fisher's Exact test: *p < 0.05, > < 0.01
Source: Soffritti et al. (2002a) and ERF web portal (http://www.ramazzini.it/fondazione/foundation.aspX
5.4.1.3. Dose A djustm en ts an d Ext rap ola tion Meth od
3 As with the extrapolations used in the development of the RfC and RfD, the PBPK model
4 was used for species-to-species extrapolation of the doses to be used in the cancer dose-response
5 analysis. Three dose metrics were considered for use in the dose-response analysis: total
6 metabolized methanol; maximum blood concentration of the parent (Cmax); and area under the
7 blood concentration time curve (AUC) for the parent. Internal dose estimates (above
8 background) corresponding to the administered doses from the animal bioassay were determined
9 for each of these metrics with the PBPK model (see Appendix E, Table E-5). To help inform the
10 selection of the most appropriate dose metric, dose-response analyses were performed using
11 these PBPK model results to assess which dose metric best corresponded to the observed
12 incidence data in Table 5-6 (see Appendix E, Tables E-6 and E-7). Figures 5-3, 5-4, and 5-5
13 show the fit of the multistage model to the all lymphoma incidence data for female and males,
14 using each dose metric as the dose input.
79 EPA obtained detailed, individual animal data via an interagency agreement with NIEHS which supported the
development of reports made available through the Ramazzini Foundation (ERF) web portal
(http://www.ramazzini.it/fondazione/foundation.asp). This allowed EPA to combine lymphomas of similar
histopathological origin and confirm the tumor incidences reported in the Soffriti et al. (2000a) paper.
5-30 DRAFT—DO NOT CITE OR QUOTE
-------
Multistage Cancer Model
"Multistage Cat
Female rats (p = 0.18)
0 20 40 60
100 120 140
22:46 07/23 2009
Multistage Cancer Model
< 0.3
Male rats (p = 0.2)
50 100 150 200
22:51 07/23 2009
Figure 5-3. All lymphomas versus methanol metabolized (mg/day) for female and male
rats
Multistage Cancer Model
Female rats (p = 0.077)
22:3507/232009
2000 3000
dose
4000 5000
0.45
0.4
0.35
0.3
0.25
0.2
0.15
Multistage Cancer Model
Multistage Cancer
Male rats (p = 0.15)
0 500 1000 1500 2000 2500 3000 3500 4000 4500
Figure 5-4. All lymphomas versus Cmax (mg/L) for female and male rats.
Multistage Cancer Model Multistage Cancer Model
Multistage Cancer
Female rats (p = 0.075)
0 20000 40000 60000 80000 100000
22:41 07/232009
22:44 07/23 2009
Male rats (p = 0.15)
20000 40000 60000 80000 100000
dose
Figure 5-5. All lymphomas versus AUC (hr x mg/L) for male and female rats
5-31 DRAFT—DO NOT CITE OR QUOTE
-------
1 The dose-response modeling suggests that total metabolized methanol is a better dose
2 metric than the parent compound metrics as indicated by improved model fit to responses
3 reported by Soffritti et al. (2002a) for both lympho-immunoblastic lymphoma (see Appendix E,
4 Table E-7) and all lymphoma (see Figures 5-3 to 5-5 and Appendix E, Table E-7), and also for
5 malignant lymphoma responses reported by Apaja (1989) (see Appendix E, Table E-17). Chi-
6 square p values for the total metabolite dose metric ranged from 0.18 to 0.55 and were
7 consistently higher than for the other dose metrics. This could be an indication of the importance
8 of metabolite formation, which is likely to be more rapid at low doses, to the carcinogenic
9 response. The total metabolized methanol dose metric was selected as the dose metric for use in
10 the dose-response assessment to derive the POD because it provided the best fit to the response
11 data. The estimated BMDL for the total methanol metabolized dose metric was then back-
12 calculated using the EPAPBPK model to obtain a human equivalent oral drinking water dose in
13 terms of mg/kg-day (see Appendix E, Table E-8).
14 Multistage and multistage Weibull time-to-tumor models were applied to the lymphoma
15 data obtained from ERF for the Sofffritti et al. (2002a) drinking water study and considered for
16 determining the POD to be used in the derivation of the oral cancer slope factor (see Appendix E,
17 Table 3-8). Appendix E gives the details and justification for the various approaches used. As
18 described in Appendix E, time-to-tumor modeling and multistage quantal modeling gave similar
19 results, and the tumor responses modeled did not exhibit significant time dependence on dose.
20 The EPA multistage cancer model fit the response data adequately and was used to derive the
21 oral cancer slope factor (CSF) (see Appendix E, Tables E-7, E-8, and Figure E-10).
22 BMDs and BMDLs were estimated for the combined lymphomas in male and female rats.
23 The BMR selected was the standard value of 10% extra risk recommended for dichotomous
24 models (U.S. EPA, 2000b).80 The 95% one-sided lower confidence limit defined the BMDL.
25 The dose terms in the fitting were set equal to the estimated total metabolized doses derived
26 using the PBPK model for methanol for each of the administered doses in the bioassay.
27 Application of the multistage model to the incidence data for all lymphomas in male rats
28 (Table 5-6) resulted in the BMD and BMDLio values presented in Table 5-7. The results for the
29 male rat were used in the derivation of the CSF because the female data for this endpoint yielded
30 slightly higher values (see Appendix E, Tables E-7 and E-8). Assuming that metabolized
31 methanol distributes in the body according to body weight to the % power, the male rat mg-day
32 BMDLio was converted to a human mg-day BMDLio.81 The human PBPK model (Appendix B)
33 was then used to convert this human mg-day value for total methanol metabolized back to a
80 The use of lower BMR values was determined not to have a significant impact on the CSF derivation.
81 Male rat mg/day was converted to human mg/day by multiplying by (BWhuman)yV(BWrat)y4 =(70 kg)yY(0.33 kg f'=
55.6
5-32 DRAFT—DO NOT CITE OR QUOTE
-------
1 human equivalent methanol oral dose HED(BMDLio) of 51.5 mg/kg-day for lymphomas in the
2 male rat (see Appendix E, Table E-8).82
Table 5-7. BMD results and oral CSF using all lymphoma in male rats
Amount metabolized (mg/d)
BMD10
(mg/d)
101.7
Rat BMDL10
(mg/d)
63.9
Estimated human
BMDL10 (mg/d)
3,553
Human equivalent
BMDL10
(mg/kg-day)
51.5
Oral CSF
(mg/kg-day) 1
1.9E-03
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
Source: Soffritti et al. (2002a).
In the case of methanol, there is no information to inform the MOA for carcinogenicity.
As recommended in the Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a), "when
the weight of evidence evaluation of all available data is insufficient to establish the MOA for a
tumor site and when scientifically plausible based on the available data, linear extrapolation is
used as a default approach." Accordingly, for the derivation of a quantitative estimate of cancer
risk for ingested methanol, a linear extrapolation was performed to determine the CSF.
5.4.1.4. Oral Slope Fa ctor
The oral slope factor was derived based on a linear extrapolation from this POD
(BMDLio/HED of 55.4 mg/kg-day for lymphomas in the male rat) to the estimated background
response level:
O.I/HED(BMDLio) = 0.1/51.5 mg/kg-day = 2E-03 (mg/kg-day)'1
(rounded to one significant figure)
5.4.2. Inhalation Exposure
5.4.2.1. Choice of Study/Data-with Rationale and Justification
No human data exist that would allow for quantification of the cancer risk associated with
chronic methanol exposure. Table 4-35 summarizes the available experimental animal inhalation
exposure studies of methanol. The NEDO (1987, 1985/2008b) 24-month rat study is the only
inhalation bioassay available that reports an increase in incidence of any cancer endpoints (see
Section 4.2.2.3). This NEDO (1987, 1985/2008b) study was of high quality and was based on
standard OECD guidelines (OECD, 2007). F344 rats were exposed for 104 weeks to air
concentrations of 0, 10, 100 and 1,000 ppm methanol. Rats were sacrificed and necropsied at the
82 The following algebraic equation is provided in Appendix B (Equation 4) to describe the relationship between
predicted human mg-day methanol cleared and the human equivalent oral dose (HED) in mg/kg-day:
HED = (3.3*3,553 mg/d)/(19,282.9 - 3, 553 mg/d) + (0.014287*3, 553 mg-d) = 51.5 mg/kg-d
5-33
DRAFT—DO NOT CITE OR QUOTE
-------
1 end of the 104-week exposure period. The NEDO (1987, 1985/2008b) study reports increased
2 pulmonary adenomas/adenocarcinomas and pheochromocytomas in high-dose (1,000 ppm) male
3 and female rats, respectively. The combined incidence of pulmonary adenomas and
4 adenocarcinomas was significantly increased in the high-dose males (see Tables 4-4 and 5-8),
5 and both tumor types were considerably elevated at the high-dose over historical control
6 incidences within their respective sex and strain (see discussion in Section 4.2.2.3). As shown in
7 Table 4-5, the severity and combined incidence of potential precursor effects in the alveolar
8 epithelium of male rat lungs (epithethial swelling, adenomatosis, pulmonary adenoma, and
9 pulmonary adenocarcinoma) and the adrenal glands of female rats (hyperplasia and
10 pheochromocytoma) were increased in the higher exposure groups compared with the controls
11 and lower exposure groups. While the incidence of male rat pulmonary adenomas was also high
12 in the lowest (10 ppm) exposure group, the appearance of a rare adenocarcinoma in the high-
13 dose group is suggestive of a progressive effect associated with methanol exposure. While the
14 increased pheochromocytoma response in female rats is not statistically increased over controls,
15 it is considered to be potentially treatment related because this is a historically rare tumor type
16 for female F344 rats (NTP, 2007, 1999; Haseman et al., 1998),83 and when viewed in conjunction
17 with the increased medullary hyperplasia observed in the mid-exposure (100 ppm) group
18 females, it is suggestive of a proliferative change with increasing methanol exposure.
5.4.2.2. Dose-Response Data
19 The tumor incidence data selected for modeling are the NEDO (1987, 1985/2008b)
20 reported incidences of adenoma/adenocarcinoma in male rats and pheochromocytoma in female
21 rats. These data are presented in Table 5-8.
Table 5-8. Incidence data for tumor responses in male and female F344 rats
Dose (ppm)
0
10
100
1000
0
10
100
1000
Number of animals affected/number examined
Pheochromocytoma
Pulmonary adenoma/adenocarcinoma
Female rats
2/50
3/51
2/49
7/51a'b
2/52
0/19
0/20
0/52
Male rats
7/52
2/16
2/10
4/51
1/52
5/50
2/52
7/52b,c
83
Haseman et al., (1998) report rates for spontaneous pheochromocytomas in 2-year NTP bioassays of 5.7%
(benign) and 0.3% (malignant) in male F344 rats and 0.3% (benign) and 0.1% (malignant) in female (n=1517) F344
rats.
5-34 DRAFT—DO NOT CITE OR QUOTE
-------
I Dose (ppm) | Number of animals affected/number examined
ap < 0.05 over NTP historical controls for total (benign, complex and malignant) pheochromocytomas using the
Fisher's Exact test
bp < 0.05 for Cochrane-Armitage test of overall dose-response trend.
cp < 0.05 over concurrent controls using the Fisher's Exact test.
Source: NEDO (1987, 1985/2008b).
5.4.2.3. Dose A djustm en ts an d Ext rap ola tion Met A od
1 As with the extrapolations used in the development of the RfC and RfD, the PBPK model
2 was used for species-to-species extrapolation of the doses to be used in the cancer dose-response
3 analysis. Three dose metrics were considered for use in the dose-response analysis: total
4 metabolized methanol, maximum blood concentration of the parent (Cmax), and area under the
5 blood concentration time curve (AUC) for the parent. Each of the dose metrics corresponding to
6 the administered dose from the animal bioassay was determined with the PBPK model (see
7 Appendix E, Table E-10). To help inform the selection of the most appropriate dose metric,
8 dose-response analyses were performed using these PBPK model results to assess which dose
9 metric best corresponded to the observed incidence data in Table 5-8 (see Table E-l 1). All of
10 the dose metrics resulted in similar fit to the incidence data for both endpoints, with the total
11 metabolites metric providing a slightly improved fit to the female pheochromocytoma response
12 data.
13 Unlike the oral data discussed in Section 5.4.1.2, dose-response modeling of the
14 inhalation data from NEDO (1987, 1985/2008b) does not suggest the use of any one dose metric
15 over the other. However, since the pheochromocytoma response likely involves systemically
16 distributed, metabolized methanol, and to be consistent with the oral CSF, analysis the total
17 methanol metabolized dose metric is selected as the dose metric for use in the dose-response
18 assessment to derive the inhalation POD. The estimated BMDL for the methanol metabolized
19 dose metric was then back-calculated using the EPA PBPK model to obtain a human equivalent
20 air exposure concentration in terms of mg/m3 (see Table E-l3).
21 The EPA multistage model was applied to the data in Table 5-8 obtained from the NEDO
22 (1987, 1985/2008b) inhalation study and considered for determining POD to be used in the
23 derivation of the inhalation cancer unit risk (Table E-l 1). Appendix E gives the details and
24 justification for the various approaches used. As described in Appendix E, time-to-tumor and
25 quantal modeling gave similar results, and the tumor responses modeled did not exhibit
26 significant time dependence on dose. The EPA multistage cancer model fit the response data
27 adequately and was used to derive the IUR (Tables E-l 1, E-12, and Figure E-13).
28 BMDs and BMDLs were estimated for tumor responses in male and female rats as shown
29 in Table 5-8. The BMR selected was the standard value of 10% extra risk recommended for
5-35 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
84
quantal models (U.S. EPA, 2000b). The 95% one-sided lower confidence limit defined the
BMDL. The dose terms in the fitting were set equal to the estimated total metabolized doses
derived using the PBPK model for methanol for each of the administered doses in the bioassay.
Application of the multistage model to the incidence data for pheochromocytomas in
female rats (Table 5-8) resulted in the BMD and BMDLio values presented in Table 5-9. The
results for the female rat were used because the female data for pheochromocytoma yielded
slightly lower BMDL values (Tables E-ll and E-13). Assuming that metabolized methanol
distributes in the body according to body weight to the 3/4 power, the female rat mg-day BMDLio
was converted to a human mg-day BMDLio.85 The human PBPK model (Appendix B) was then
used to convert this human mg-day value for total methanol metabolized back to a human
equivalent methanol inhalation concentration, HEC(BMCLio), of 81.9 mg/m3 or 81,900 ug/m3
for pheochromocytomas in the female rat (see Appendix E, Table E-13).86
Table 5-9. BMD results and IUR using pheochromocytoma in female rats
Amount metabolized (mg/day)
BMD10
(mg/day)
29.5
BMDL10
(mg/day)
14.6
Estimated human
BMDL10 (mg/day)
971
Human equivalent
BMCL10
(mg/m3)
81.9
IUR
(jig/m3)-1
1.2E-06
Source: NEDO (1987, 1985/2008b).
5.4.2.4. IUR
13 The IUR in terms of (ug/m3)-1 was then derived based on a linear extrapolation from this
14 POD to the estimated background response level:
15
16
17
18
0.1/HEC(BMCLio) = 0.17(81,900 ug/m3) = 1E-06 (ug/m3)-1
(rounded to one significant figure)
5.4.3. Uncertainties in Cancer Risk Assessment
The following is a discussion of the uncertainties associated with the cancer potency
estimate for methanol beyond that which can be addressed with the quantitative approach
applied. A summary of these uncertainties is presented in Table 5-10.
84 The use of lower BMR values was determined not to have a significant impact on the IUR derivation.
85 Female rat mg/day is converted to human mg/day by multiplying by (BWhuman)/4/(BWrat)y4 = (70 kg)y7(0.26 kg f' =
66.5
86 The following algebraic equation is provided in Appendix B (Equation 3) to describe the relationship between
predicted human mg-day methanol cleared and the human equivalent inhalation concentration (HEC) in mg/m3:
HED = (14.69 x 971 mg/d)/(19282.9 - 971 mg/d) + (0.063554 x 971 mg/d) = 81.9 mg/ m3
5-36 DRAFT—DO NOT CITE OR QUOTE
-------
Table 5-10. Summary of uncertainty in the methanol cancer risk assessment
Consideration
Quality of the
studies relied
upon for the
determination of
the PODs
Interpretation of
results from
study relied upon
for the
determination of
the POD
Consistency of
results across
chronic studies
Choice of
endpoint for
POD derivation
Choice of
species/gender
Choice of model
for POD
derivation
Choice of
animal-to-human
extrapolation
method
Potential impact
Key chronic studies
not always well
reported; could lead
to t or I of risks
Differences in tumor
classification can
lead to over or
underestimate of risk
If effects not relevant
to humans, risk is
overestimated.
Route-to-route
extrapolation from
Soffriti et al. (2002a)
study would t
inhalation POD by
about 4-fold
CSF and IUR would
be I if based on
another gender
Use of other models
could t or | POD,
but not significantly
Traditional method
could t the
HEC(BMCL10)
estimate by 4-fold.
Decision
Utilize re-analyses of
the Soffritti et al.
(2002a) and NEDO
(1985/2008a, 2008b)
chronic studies
Derive POD based on
incidence of combined
lymphomas as
suggested by NTP
pathologists; Assume
proper classification of
lung and adrenal
tumors by NEDO
Derive PODs based on
Soffritti et al. (2002a)
and NEDO
(1985/2008b).
Derive oral CSF from
lymphoma data and
inhalation CSF from
adrenal effects.
CSF and IUR are
based on the most
sensitive and reliably
quantifiable
species/gender
Derive cancer potency
factor based on
multistage model.
A PBPK model was
used to extrapolate
animal-to-human
concentrations.
Justification
Consideration of all available information resulted
in the determination that the Soffritti et al. (2002a)
and NEDO (1985/2008a, 2008b) chronic studies
are adequate (see discussion of individual studies
in Sections 4.2.1.3 and 4.2.2.3 and summaries in
Sections 4.9. land 4.9.2).
Both NTP and EFSA recommend that only
lymphomas of the same cellular origin be
combined for dose-response analysis. With
respect to lung and adrenal tumors, examination of
concurrent alveolar and adrenal noncancer
hyperplastic endpoints supports a proliferative
change in these organ systems consistent with the
appearance of carcinogenic responses.
Though tissue concordance across species, strains
and routes of exposure is not assumed, lymphomas
have been observed in more than one species by
oral route. Also there is evidence that the observed
lymphomas are relevant to humans (see
discussions in Section 4.9. concerning human
studies of methanol metabolite formaldehyde).
Oral POD was based on the only tumor type from
Soffritti et al. (2002a) drinking water study
significantly increased (all lymphomas); Inhalation
POD based on most sensitive tumor response from
NEDO (1985/2008b) study, increased
pheochromocytoma in female rats.
Choice of female rat lymphoma and male rat
adenoma/adenocarcinoma would have resulted in
lower CSF and IUR values, respectively. Use of
the Apaja (1989) mouse data would have resulted
in a higher, but less reliable CSF due to study
problems, including a lack of concurrent controls
Use of the multistage model is consistent with
EPA guidance (U.S. EPA, 2005a). The multistage
model provides adequate fit to the data, which is
not improved by a time-to-tumor modeling.
Use of a PBPK model reduces uncertainty
associated with the animal to human extrapolation.
Total metabolites normalized by body weight is an
appropriate dose metric and a verified PBPK
model exists that estimates this metric.
5.4.3.1. Quality of Studies that are the Basis for the PODs
1 The protocols used at the laboratories that have performed cancer bioassays of methanol,
2 particularly those of the ERF, differ from the more commonly used (e.g., NTP) protocols. The
3 unique features of the ERF study design and their implications to a methanol cancer risk
5-37 DRAFT—DO NOT CITE OR QUOTE
-------
1 assessment are discussed in Section 4.9.2. Separate from these experimental design issues are
2 considerations relative to the quality of the cancer bioassays and any associate uncertainties.
3 The number of animals per dose group in ERF studies is often higher than the 50 animals
4 per sex per dose group typically used in EPA and NTP studies, increasing the statistical power of
5 the ERF cancer bioassay. However, ERF sometimes shares controls across concurrent studies
6 (Cruzan, 2009; Belpoggi et al., 1995). In contrast, EPA requires (U.S. EPA, 1998) and NTP
7 generally uses (Melnick et al., 2007) concurrent, matched controls for each carcinogen bioassay.
8 The use shared controls does not necessarily compromise a study, but the use of a concurrent,
9 matched control is generally preferred as a means of further avoiding confounding factors and
10 increasing the reliability of a study regarding the interpretation of findings in treated animals.
11 The published report of the methanol bioassay (Soffritti et al., 2002a) indicates that the
12 experiment was performed according to good laboratory practice (GLP) and standard operating
13 procedures (SOP) of the ERF. Further, an independent review of ERF (Huff, 2002) suggests that
14 quality control procedures associated with GLP were in place. However, questions have been
15 raised about the quality of studies at the ERF by European Food Safety Authority (EFSA, 2006)
16 in regards to the aspartame bioassays conducted by the ERF (Soffritti et al., 2006); by extension
17 this EFSA report has raised issues for consideration in regards to methanol. EFSA (2006) has
18 suggested that an inspection by the Italian GLP compliance monitoring authority (Ministry of
19 Health) necessary to confirm GLP had not been conducted.87 The EFSA (2006) report also
20 identifies specific deviations from OECD guidelines (OECD, 2007), including a lack of a
21 complete analysis of the test substance, no clear information on the stability of the substance, a
22 lack of clinical observations or macroscopic changes, a lack of hematological assays, a lack of
23 serology (e.g., to confirm the presence of infection) and limited histopathology reports. While
24 these details may be recorded internally by the ERF as part of their standard protocol, because
25 there is no documentation of these details available for consideration, there remains some
26 uncertainty regarding the level at which they were performed. There is limited evidence,
27 however, that these factors had a significant impact on the adequacy of the study for assessing
28 carcinogenic potential (see Sections 4.2.1.3, 4.9.2 and 5.4.3.2).
29 EFSA (2006) also expresses concern over the possibility of compromised pathological
30 diagnosis in the ERF aspartame study (Soffritti et al., 2006) due to extensive autolysis. ERF
31 performs pathological examinations on "dying animals undergoing necropsy" (Soffritti et al.,
32 2002a). This creates difficulties in pathological examinations associated with cell autolysis that
33 can occur when pathology slides are prepared after natural death. The NTP (Hailey, 2004)
34 commented on the increased prevalence of autolysis in slides from the ERF (Soffritti et al., 2006)
35 aspartame study. EPA conducted a detailed analysis of the individual animal tissue data obtained
87 Since the publication of the EFSA (2006) report, the EPA has confirmed through communication with the ERF
laboratory (Knowles, 2008) that ERF is currently in the process of obtaining this certification.
5-38 DRAFT—DO NOT CITE OR QUOTE
-------
1 from ERF for their chronic methanol, MTBE, formaldehyde, and aspartame studies, and
2 determined that autolysis and other causes of tissue loss did not substantially impact tissue
3 denominators. For most of the tissues evaluated there were more than 96 (individual animal)
4 samples available for microscopic evaluation, and for all sites and dose groups, denominators
5 were larger than for routine NTP bioassays (i.e., >50). Thus, missing tissues does not appear to
6 have been a serious problem in the methanol study. While this analysis does not completely rule
7 out the possibility that pathology slides and diagnoses were impacted by autolysis, it does
8 indicate that this possibility would be offset by the large group size (response denominator)
9 employed. Further, even if autolysis was a confounding factor, its presence would not negate
10 positive cancer findings as autolysis would tend to decrease, not increase, the power to observe
11 an effect.
12 There were no differences in survival among the methanol dose groups of the Soffritti et
13 al. (2002a) study. However, Cruzan (2009) has suggested that "While the survival at 104 weeks
14 was within the normal, but widespread, range for Sprague-Dawley rats, there was significant
15 early mortality among all groups, including the controls" and that "the control group from an
16 inhalation study (Cruzan et al., 1998) had much better survival through 104 weeks than seen in
17 the RF methanol study." Yet, according to Table 12 of the Cruzan (2009) article, 104-week
18 survival in male (-40%) and female (-50%) control rats of the Cruzan et al. (1998) study was
19 not discernibly different from the 104 week survival of male (-40%) and female (-50%) control
20 rats of the Soffritti et al. (2002a) methanol study (see Appendix E, Figures E-l and E-2).
21 Survival of male and female Sprague-Dawley rats in the Soffritti et al. (2002a) study at
22 104 weeks was greater than 40% in all but the female 5,000 ppm group. Further, at the NTP
23 (NTP, 2006), the 104-week survival of 353 control female Sprague-Dawley rats was 41.5%
24 (range of 28.3%-51% in 7 studies) using NTP's new diet and corn oil gavage, indicating that
25 survival in the Soffritti et al. (2002a) methanol study was not low.
26 Studies such as the Soffritti et al. (2002a) methanol study that allow test animals to live a
27 full life span can be difficult to interpret due to the need to distinguish between age-related and
28 chemical-related effects. Full life-span studies may have advantages. Huff et al. (2009) note that
29 "studies truncated after 2 years of exposure do not allow sufficient latency periods for late-
30 developing tumors, such as the 80% of all human cancers that occur after 60 years of age."
31 Several recent publications have noted deficiencies with the 2-year study design used at the NTP
32 and have recommended extending the duration of rodent studies to increase the sensitivity of
33 their bioassays (Huff et al., 2009; Huff et al., 2007; Maronpot et al., 2004; Bucher et al., 2002).
34 While arguments have been documented related to the possible confounding influence of
35 infection and autolysis on the results obtained from the ERF, available evidence does not indicate
36 that these factors significantly influenced the observed lymphoma/leukemia response in the
37 methanol or other bioassays conducted at ERF. In addition, for the purposes of this assessment
5-39 DRAFT—DO NOT CITE OR QUOTE
-------
1 and at the request of the EPA, the ERF and NEDO have provided additional study details beyond
2 that which is normally available from published journal articles, including quality assurance
3 reports and individual animal data. Based on a review of this information, consideration of the
4 issues, and absent additional data to the contrary, EPA has determined that both studies were
5 sufficient for use in the assessment of risk from methanol exposure.
5.4.3.2. Interpretation of Results of the Studies that are the Basis for thePODs
6 There are a number of uncertainties regarding the interpretation of both the lymphoma
7 response in male Sprague-Dawley rats (Soffritti et al., 2002a) that forms the basis of the oral
8 CSF and the pheochromocytoma response in F344 rats (NEDO, 1985/2008b) that forms the basis
9 for the inhalation CSF.
10 There is also a wide range in the background incidence of hemolymphoreticular tumors
11 reported in control groups of ERF studies. Between 1984 and 1997, incidence rates of
12 hemolymphoreticular neoplasms in control rats at ERF increased by 38% among male rats and
13 decreased by 12% among female rats (Caldwell et al., 2008). Soffritti et al. (2007, 2006) reports
14 that among 2,265 untreated males and 2,274 untreated females the average incidence of
15 lymphomas and leukemias is 20.6% (range, 8.0-30.9%) in males and 13.3% (range, 4.0-25%) in
16 females. Caldwell et al. (2008) noted that for the incidences of these lesions for the ERF colony
17 are relatively low and stable across studies. EFSA (2006) and Cruzan (2009) consider it to be
18 high relative to other tumor types and relative to the background rate for this tumor type in
19 Sprague-Dawley rats from other laboratories (see Section 4.9.2 for further discussion).
20 In the ERF bioassays, including methanol, hemolymphoreticular neoplasms were divided
21 into specific histological types (lymphoblastic lymphoma, lymphoblastic leukemia, lymphocytic
22 lymphoma, lympho-immunoblastic lymphoma, myeloid leukemia, histocytic sarcoma, and
23 monocytic leukemia) for identification purposes. Upon examining slides from the aspartame
24 study conducted by the ERF, a PWG of the NTP at the NIEHS (Hailey, 2004) found that "The
25 diagnoses of lymphatic and histocytic neoplasms in the cases reviewed were generally
26 confirmed. The NTP does not routinely subdivide lymphomas into specific histological types as
27 was done by the ERF, however the PWG accepted their more specific diagnosis if the lesion was
28 considered to be consistent with a neoplasm of lymphocytic, histocytic, monocytic, and/or
29 myeloid origin." The NTP PWG also noted that while lymphoblastic lymphomas, lymphocytic
30 lymphomas, lympho-immunoblastic lymphomas and lymphoblastic leukemias as malignant
31 lymphomas can be combined, myeloid leukemias, histocytic sarcomas and monocytic leukemia
32 should be treated as separate malignancies and not combined with the other lymphomas for
33 statistical evaluation since they are of different cellular origin (Hailey, 2004). Other researchers
34 have also noted this distinction between myeloid leukemias and histiocytic sarcomas and other
35 lymphomas (McConnell et al., 1986). To decrease the uncertainty in the combination of tumors
36 relied upon for dose-response modeling, the current dose-response modeling conducted for
5-40 DRAFT—DO NOT CITE OR QUOTE
-------
1 methanol did not include myeloid leukemia, histocytic sarcoma, monocytic leukemia, alone or in
2 combination with lymphoblastic lymphoma, lymphoblastic leukemia, lymphocytic lymphoma,
3 and lympho-immunoblastic lymphoma.
4 As expressed by EFSA (2006) and others (Cruzan, 2009; Schoeb et al., 2009), there is
5 concern that the lympho-immunoblastic lymphoma response in the ERF aspartame study
6 (Soffritti et al., 2006, 2005) was caused by or confused with sequalae of aM pulmonis infection.
7 Infection of the ERF colony with M. pulmonis has not been confirmed (Caldwell et al., 2008)
8 and, as discussed in Section 4.9.2, a link between M. pulmonis infection and induction of
9 lymphoma in rats has not been established in the literature. As noted in Section 4.9.2, there is
10 evidence suggesting that respiratory infections may have confounded the interpretation of lung
11 lesions in ERF studies. Lymphoma illustrations in 2 ERF studies (Figure 10 of Soffritti et al.
12 [2005] and Figures 1-5 of Belpoggi et al. [1999]), suggest that ERF MTBE and aspartame
13 bioassays may have been confounded by a respiratory infection such as M pulmonis and that
14 lesions associated with this infection may have been interpreted as lymphoma (Schoeb et al.,
15 2009). However, other ERF lymphoma diagnoses in multiple rat organ systems, including the
16 lung, have been confirmed by an independent working group panel of six NIEHS pathologists
17 (Hailey, 2004). In addition, the incidence of "lung-only" lympho-immunoblastic lymphomas
18 was evenly distributed across the control and 0, 500 5000, 20,000 ppm dose groups for male (9,
19 10, 14 and 13) and female (3, 5, 6 and 7) rats of the Soffritti et al. (2002a) methanol study
20 (Schoeb et al., 2009). Consequently, removing "lung-only" lympho-immunoblastic lymphomas
21 from consideration and using only lymphomas from organ systems not likely to be confounded
22 by a respiratory infection (i.e., subtracting the lung-only incidence from lympho-immunoblastic
23 lymphomas reported in Table 5-6) still results in significant dose-response curves for this lesion
24 with p values of 0.20 and 0.42 for males and females, respectively (see Figure 5-6). The
25 BMDLio estimates of 99 (males) and 110 (females) mg metabolized methanol/day are about 50%
26 higher than metabolized methanol BMDLio estimates for this endpoint when "lung-only"
27 responses are included (Appendix E, Table E-7).
28
29
5-41 DRAFT—DO NOT CITE OR QUOTE
-------
Multistage Cancer Model with 0.95 Confidence Level
;r Model with 0.95 Confidence Level
0.35
0.3
0.25
0.2
0.15
0.1
0.05
Multistage Cancer — ' ' -
Males (p= 0.20)
-^
^^^^___^— —
^-~^^
^^^^^
^^—~~
BHDL BHD
1
Multistage Cancer
Linear extrapolation
Females (p=
.42)
08:35 09/04 2009
100
dose
08:40 09/04 2009
Figure 5-6. Lympho-immunoblastic Lymphoma minus "lung-only" response in rats of
Soffritti et al. (2002a) methanol study versus methanol metabolized (mg/day).
2 For increases in 2 other tumor types (ear duct and head/oral cavity tumors) reported in the
3 ERF methanol study (Table 4-2), an independent review of the 75 pathology slides from the ERF
4 aspartame study has suggested differences in interpretation. After reviewing these slides, the
5 NTP PWG noted that "about half of hyperplastic and neoplastic lesions in the ear duct or oral
6 cavity were more severely classified by ERF study pathologists, compared with diagnosis from
7 the PWG (EFSA, 2006). Though a similar review has not been conducted of the Soffritti et al.
8 (2002a) ERF methanol bioassay results, there is uncertainty regarding the ERF interpretation of
9 these lesions. For this reason, these lesions were not considered in the derivation of the oral
10 CSF.
11 Another uncertainty that confounds the interpretation of some ERF studies is the
12 possibility of litter effects in ERF test groups. Bucher (2002) has reported that ERF does not
13 randomize the assignment of animals to treatment groups, but generally "assigns all animals
14 from a given litter to the same treatment group, recording which litter each animal came from."
15 This approach would make it more difficult to distinguish the relative contributions of the
16 chemical and genetics to the effect of concern. In response to an EPA query regarding this matter
17 (Knowles, 2008), ERF has stated that "the assignment of test animals to dose groups will vary in
18 ERF studies according to the experimental protocol and aims of the research" and "in the case of
19 experiments in which exposure begins at 6-8 weeks of age (for example BT960, methanol),
20 randomization is performed so as to have no more than one female and one male from each litter
21 in each experimental group." For other experiments in which exposure begins during prenatal
22 life,88 "randomization is performed on the breeders," but the offspring are not randomized across
23 dose groups in order to "..simulate as much as possible the human situation in which all
24 descendents are part of a population."
88 For some chemicals such as vinyl chloride (Maltoni and Cotti, 1988) and aspartame (Soffriti et al., 2006), ERF
has started exposure as early as in utero, This early exposure study design can markedly increase the sensitivity of a
cancer bioassay (Maltoni and Cotti, 1988; Soffriti et al., 2006; Melnick et al., 2007).
5-42 DRAFT—DO NOT CITE OR QUOTE
-------
1 There is uncertainty regarding the pheochromocytoma response observed in the NEDO
2 (1985/2008b) study with respect to both its relation to exposure and to its pathology. As
3 discussed in Section 4.2.2.3, the incidence of pheochromocytomas in female rats exhibited a
4 dose-response trend (Cochrane-Armitage/? < 0.05). While the incidence of 13.7% (7/51) in the
5 high-dose group was significantly elevated (p < 0.05) over NTP historical controls, it was not
6 significantly elevated over the concurrent control rate of 4% (2/50). Much of the uncertainty is
7 inherent in difficulties associated with the characterization of this lesion. According to NTP
8 (2000), the primary criterion used to distinguish pheochromocytomas from nonneoplastic adrenal
9 medullary hyperplasia, the presence of mild-to-moderate compression of the adjacent tissue, can
10 be a difficult determination. Further, while NEDO (1985/2008b) reported adrenal effects as
11 "hyperplasia of medullary cells" and "pheochromocytoma," NTP pathologists categorize
12 pheochromocytomas into three types: benign, complex and malignant (NTP, 2007, 1999). This
13 is an important distinction as complex and malignant pheochromocytomas are a rare tumor type,
14 occurring spontaneously in female F344 rats at rates ranging from 0.1% to 0.7% (NTP, 2007,
15 1999; Haseman et al., 1998) and with cell proliferation activity that is much higher than benign
16 pheochromocytomas (Pace et al., 2002). Thus, severity of the pheochromocytoma response
17 reported by NEDO (1985/2008b) is uncertain, potentially ranging from mischaracterized
18 hyperplasia to highly proliferative and potentially metastatic malignancies. Finally, the NEDO
19 study was a two-year study, and these lesions, which include diffuse hyperplasia, nodular
20 hyperplasia, and pheochromocytoma, progress with age. Thus, it is possible that a life-span
21 study would have detected a more severe carcinogenic response (e.g., progression of the mid-
22 dose group hyperplastic responses as reported in Table 4-5).
5.4.3.3. Consistency across ChronicBioassays for Methanol
23 The observation of a lymphoma response in rats (Soffritti et al., 2002a) and mice (Apaja,
24 1980), along with reported associations between lymphomas and human exposure to methanol's
25 metabolite, formaldehyde (see Section 4.9), contributes significantly to the cancer weight-of-
26 evidence determination. This was the only carcinogenic response that was observed in more than
27 one animal bioassay. However, tissue concordance across species, strains and routes of exposure
28 is not assumed, and a lack of such concordance does not negate the validity of individual
29 findings nor the potential relevance of such findings to humans.
30 Besides the Soffritti et al. (2002a) drinking water study of rats (2 years) reported by ERF,
31 the only other chronic studies available are the Apaja (1980) lifespan drinking water study study
32 in Swiss mice which did not include a concurrent untreated control group, and those reported in
33 the Japanese NEDO (1987, 1985/2008a, 2008b) study of monkeys (7-29 months), mice
34 (18 months), and F344 rats (2 years). None of the NEDO studies involved lifetime evaluations,
35 and only the rat study can be said to cover a significant portion of the test species life span. The
36 only organ system that exhibited an increase in effects in both inhalation and drinking water
5-43 DRAFT—DO NOT CITE OR QUOTE
-------
1 studies was the testes. High-dose rats of the Soffritti et al. (2002a) methanol study exhibited an
2 increase in testicular interstitial cell adenomas, and the incidence of testicular hyperplasia was
3 increased in high-dose rats of the NEDO (1987, 1985/2008b) study. However, neither effect was
4 statistically increased over controls, and both effects were well within their historical control
5 ranges. An increase in lymphomas was the only carcinogenic effect reported in the Soffritti et al.
6 (2002a) and Apaja (1980) drinking water studies that is considered dose related and quantifiable,
7 and male pulmonary adenoma/adenocarcinoma and female pheochromocytoma were the only
8 carcinogenic effects from the NEDO (1987, 1985/2008b) inhalation study that are considered
9 exposure related and quantifiable.
10 As discussed in Section 4.9.2, there are several differences between the NEDO (1987,
11 1985/2008b) and Soffritti et al. (2002a) studies conducted in rats which limit their direct
12 comparison and may explain some of the differences in response. These include route of
13 exposure, lifetime evaluation period, and strain of species used. Pulmonary adenomas observed
14 in the NEDO inhalation study could be portal-of-entry effects associated with the inhalation
15 route of exposure. Differences in systemic effects observed in the two studies such as the
16 pheochromocytoma response in the NEDO (1987, 1985/2008b) study and the lymphoma
17 responses observed in the Soffritti et al. (2002a) study are systemic responses and differences
18 would not be expected based on route of exposure. Differences in the evaluation period between
19 the two studies may have contributed to the lack of endpoint concordance. In the Soffritti et al.
20 (2002a) study, the animals were administered methanol for 104 weeks and then followed until
21 the completion of their natural lifetime. The average life span for these animals was 94 and 96
22 weeks for males and females, respectively, with animals living as long as 153 weeks (female).
23 However, this does not explain the difference in lymphoma response between the studies as
24 many of the lympho-immunoblastic lymphomas (most common type observed) were observed
25 prior to 104 weeks (control - 5/9; 500 ppm - 7/17; 5,000 ppm - 13/19; 20,000 ppm - 11/21). The
26 NEDO (1987, 1985/2008b) study was conducted in F344 rats versus the Sprague-Dawley rats
27 used in the Soffritti et al. (2002a) ERF study. More importantly, the background rates of selected
28 types of lymphomas and leukemias are very different between the two strains of rats. In the
29 F344 rat, there is a high background rate of mononuclear cell leukemia, while there is a much
30 lower background rate of this leukemia type in the Sprague-Dawley rat (Caldwell et al., 2008).
31 The types of lymphoma reported in the Sprague-Dawley rat by Soffritti et al. (2002a) following
32 methanol administration are rarely diagnosed in the F344 rat. Thus, the strain difference
33 between the two studies is a likely explanation for the fact that lymphomas were only increased
34 in the Soffritti et al. (2002a) rat study. NTP (2007, 1999) reports do not suggest a significant
35 difference between F344 and Sprague-Dawley rats with respect to pulmonary adenomas and
36 pheochromocytomas. Another possible explanation for this difference includes a different
37 profile of circulating metabolites associated with oral first-pass liver metabolism.
5-44 DRAFT—DO NOT CITE OR QUOTE
-------
5.4.3.4. Ch oice of En dp oin t for POD Deriva tion
1 Keeping in mind the aforementioned uncertainties associated with the interpretation of
2 the Soffritti et al. (2002a) and NEDO (1985/2008b) study results, the choice of tumors to use for
3 the derivation of the oral and inhalation CSFs was made on the basis of the appearance of a dose-
4 related increase in response rates for the selected tumor categories. Analysis of lymphomas used
5 a pair-wise statistical comparison (Fisher's Exact test) and a trend test (Cochran Armitage trend
6 test) to determine whether the slope of that trend was statistically significantly greater than 0.
7 The Fisher's Exact result for the male rat high-dose group and the results of the Cochran
8 Armitage Trend Test were both/? < 0.01. The decision not to include the liver tumors in the
9 dose-response analysis was made based on a lack of dose response according to pair-wise and a
10 trend test results versus concurrent controls This is not to say that the slight increase in the
11 incidence of this tumor type observed in all dose groups of male rats was not related to methanol
12 exposure (this is a relatively rare Sprague-Dawley rat tumor), only that the increase was not
13 statistically significant and did not contribute significantly to the overall tumor response.
14 As discussed in Sections 4.2.2.3 and 4.9.2, the high-dose incidences for pulmonary
15 adenomas/adenocarcinomas were increased over concurrent controls (p < 0.05). While the high-
16 dose incidences of pheochromocytomas in the NEDO (1985/2008b) study were not statistically
17 increased over concurrent controls, the dose-response for both tumor types represents increasing
18 trends (Cochran Armitage trend test; p < 0.05), and in both cases, the high-dose response
19 incidences were considerably elevated over historical control incidences (p < 0.05) within their
20 respective sex and strain. Further, both tumor responses are accompanied by proliferative
21 changes (e.g., hyperplastic responses) in their respective cell types that suggest tumor
22 progression.
5.4.3.5. Ch oice ofSp ecies/Gen der
23 The oral CSF was based on male rat lymphomas rather than female rat lymphomas. The
24 inhalation IUR was based on female rat pheochromocytomas rather than male rat
25 adenoma/carcinomas. In both cases, the gender that exhibited the steeper dose response and the
26 higher risk estimate was chosen.
27 Both the CSF and IUR were based on rat studies. Use of the Apaja (1989) mouse data
28 would have resulted in a 5-fold higher, but less reliable oral CSF due to a high level of
29 uncertainty associated with the Apaja (1989) study, which contained limited experimental detail
30 and did not include a concurrent control group (see Section 4.2.1.3).
5.4.3.6. Ch oice of Mod el for POD Deriva tion
31 The multistage-cancer model contained in EPA's BMDS version 2.1 was used to derive
32 both the CSF and IUR estimates for methanol. When no biologically-based cancer model exists
33 and evidence for a nonlinear cancer MOA is lacking, as is the case for methanol, the preference
5-45 DRAFT—DO NOT CITE OR QUOTE
-------
1 within the EPA's IRIS program is for the use of a multistage model. There is uncertainty
2 associated with whether the multistage model is the most appropriate choice, but in the absence
3 of a biologically based model, dose-response modeling is largely a curve-fitting exercise, and the
4 multistage model is sufficiently flexible for most cancer bioassay data. In the case of the oral
5 CSF, individual animal response data was obtained from the authors of the principal study
6 (Sofffritti et al., 2002a) and a multistage-Weibull time-to-tumor model was applied to determine
7 whether the lifespan study design of the study had an appreciable impact on the dose-response
8 analysis. As described in Appendix E, time-to-tumor modeling and multistage quantal modeling
9 gave similar results, and the tumor responses modeled did not exhibit significant time
10 dependence on dose.
11
5.4.3.7. Ch oice of Dose Metric
12 The total methanol metabolized was selected over AUC or the Cmax as the most
13 appropriate dose metric for derivation of the oral cancer slope factor and inhalation unit risk
14 primarily because it provided the best fit to response data, particularly lymphoma incidence from
15 Soffritti et al. (2002a) (see Figures 5-3 through 5-5 and Table E-7 of Appendix E) and Apaja
16 (1989) (see Table E-17 of Appendix E). Also, lymphomas and respiratory effects have been
17 observed in studies conducted with formaldehyde, and lymphomas have been observed in
18 chronic bioassays conducted with other compounds that convert to formaldehyde (i.e., MTBE
19 and aspartame). As discussed in Section 4.9.3, metabolites of methanol, particularly
20 formaldehyde, may play a role in the MOA.
21 In considering the dose-response relationship for methanol-induced carcinogenesis, a key
22 factor is the saturation of metabolism since metabolic transformation to formaldehyde and
23 generation of oxidative stress are considered likely candidates in the mode of action. Cruzan
24 (2009) indicates that saturation occurs at dose of "600 mg/kg," but saturation depends on the
25 dose rate, not the total administered dose. For example, if 600 mg/kg is given in a single bolus,
26 the internal concentration immediately following that bolus could well be high enough to
27 saturate metabolism, while the same total dose ingested over the course of a day might not.
28 To aid in interpretation of the Soffretti et al. (2002a) bioassay, water ingestion in rats was
29 assumed to shift between nocturnal (high activity) and diurnal (low activity) periods, each lasting
30 12 hours. Rats were assumed to consume 20% of their daily water ingestion during the diurnal
31 period and 80% during the nocturnal period. Ingestion in each period was assumed to occur in
32 "bouts" which were treated as periods of continuous (zero-order) infusion to the stomach.
33 During the nocturnal period each bout was assumed to last 45 minutes, followed by 45 minutes
34 without ingestion (overall period is 1.5 hours) and during the diurnal period the bout was
35 assumed to last on 3 hours followed by 2.5 hours without ingestion (overall period is 3 hours).
5-46 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
10
11
12
13
14
15
16
17
18
19
20
21
Given this exposure pattern, the total amount metabolized per day (after periodicity is
reached) in a 420 g rat (average weight in Soffretti et al., 2002a) was estimated using the PBPK
model, with the results shown in Figure 5-7. The amount increases almost linearly with
exposure until ~ 400 mg/kg/d, but continues to increase above that point, becoming almost
completely saturated by 2,000 mg/kg/d. This pattern occurs in part because of the circadian
ingestion pattern. The more rapid ingestion rate during the dark cycle leads to the highest
internal concentrations and hence the initial metabolic saturation during that part of the day. But
the lower ingestion (light) period, internal concentrations drop, allowing for an exposure range
(400-1600 mg/kg/d) where nocturnal metabolism is saturated but diurnal metabolism is not.
ISO-
cn
T)
o
"
160-
140-
120-
100
80-
0)
£ 60-
!s
O 40-1
20-
Total methanol metabo ism
from drinking water exposure
200
400
600
800
1,000 1,200 1,400 1,600 1,800 2,000 2,200
Ingestion (mg/kg/d)
Figure 5-7. Total amount metabolized per day (after periodicity is reached) in a 420 g
rat
While, based on this exposure-dose pattern, one might expect a similar exposure-
response relationship, this pattern does not include detoxification mechanisms. If such
mechanisms also saturate, then it is possible for the slower increase in total metabolism above
400 mg/kg/d to result in a significant increase in effect, though full metabolic saturation at
-2,000 mg/kg/d would still be expected to result in a maximal effect at that exposure level.
5.4.3.8. Ch oice ofAnim al-to-Hum an Ext rap ola tion Meth od
A PBPK model was used to extrapolate animal-to-human concentrations. The estimated
methanol metabolized for each dose administered to the animals in NEDO (1985/2008b) and
Soffritti et al. (2002a) were determined using the animal PBPK model, and then the BMDLio
determined by the methods described previously (Section 5.4.2.1). Assuming that metabolized
methanol distributes in the body according to body weight to the 3/4 power, the rat mg-day
BMDLio was converted to a human mg-day BMDLio. The human PBPK model (Appendix B)
5-47
DRAFT—DO NOT CITE OR QUOTE
-------
1 was then used to convert this human mg-day values for total methanol metabolized back to a
2 human equivalent methanol oral dose HED(BMDLio) of 51.5 mg/kg-day for lymphomas in the
3 male rat, and a human equivalent methanol inhalation concentration HEC(BMCLio) of
4 81.9 mg/m3 for pheochromocytomas in the female rat. If traditional dosimetry assumptions are
5 used, the HED(BMDLio) and HEC(BMCLio) estimates would have been approximately 4-fold
6 higher than the value derived using the PBPK model.
7 As discussed in Sections 3.4 and 5.3.4, the PBPK models do not describe or account for
8 background levels of methanol, formaldehyde or formate, and background levels were subtracted
9 from the reported data before use in model fitting or validation (if not already subtracted by
10 study authors), as described below. This approach was taken because the relationship between
11 background doses and background responses is not known, because the primary purpose of this
12 assessment is for the determination of noncancer and cancer risk associated with increases in the
13 levels of methanol or its metabolites (e.g., formate, formaldehyde) over background, and because
14 the subtraction of background levels is not expected to have a significant impact on PBPK model
15 parameter estimates as background levels of methanol and its metabolites are low relative to
16 exposure levels used in methanol bioassays. Further, while it is possible that background levels
17 of methanol or its metabolites contribute to background responses for some adverse effects, the
18 results of dose-response modeling of cancer endpoints using "background dose" models suggest
19 that this contribution is relatively small (see discussion in Appendix E, Section E.4).
20
5.4.3.9. Human Relevance of Cancer Responses Observed in Rats and Mice
21 As discussed in Sections 4.9.2, there is human evidence for the association of lymphomas
22 with a metabolite of methanol, formaldehyde. However, there is no information available in the
23 literature regarding the observation of cancer in humans following chronic administration of
24 methanol. The only observations in animals were noted in the chronic studies of methanol
25 conducted by Apaja (1980), Soffritti et al. (2002a) and NEDO (1985/2008b) and there is
26 uncertainty associated with the interpretation of the tumor responses reported in these studies.
27 As a consequence, the overall WOE, while convincing, is not strong.
5-48 DRAFT—DO NOT CITE OR QUOTE
-------
6. MAJOR CONCLUSIONS IN THE CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE
6.1. HUMAN HAZARD POTENTIAL
1 Methanol is the smallest member of the family of aliphatic alcohols. Also known as
2 methyl alcohol or wood alcohol, among other synonyms, it is a colorless, very volatile, and
3 flammable liquid that is widely used as a solvent in many commercial and consumer products. It
4 is freely miscible with water and other short-chain aliphatic alcohols but has little tendency to
5 distribute into lipophilic media. Methanol can be formed in the mammalian organism as a
6 metabolic byproduct and can be ingested with foodstuffs, such as fruits or vegetables. A
7 potential for human exposure exists today in the form of the artificial sweetener, aspartame,
8 which is a methyl ester of the dipeptide aspartyl-phenylalanine. Methanol is the major anti-
9 freeze constituent of windshield washer fluid. Its use as a fuel additive for internal combustion
10 engines is, as yet, limited by its corrosive properties.
11 Because of its very low oil:water partition coefficient, methanol is taken up efficiently by
12 the lung or the intestinal tract and distributes freely in body water without any tendency to
13 accumulate in fatty tissues. It can be metabolized completely to CO2, but may also, as a regular
14 byproduct of metabolism, enter the Ci-pool and become incorporated into biomolecules. Animal
15 studies indicate that blood methanol levels increase with the breathing rate and that metabolism
16 becomes saturated at high exposure levels. Because of its volatility it can also be excreted
17 unchanged via urine or exhaled air.
18 The acute toxicity in laboratory animals in response to high levels of exposure results
19 from CNS depression. NEDO (1987) reported that methanol blood levels around 5,000 mg/L
20 were necessary to cause clinical signs and CNS changes in monkeys. In humans, however, acute
21 toxicity is caused by metabolic acidosis that appears to affect predominantly the nervous system,
22 with potentially lasting effects such as blindness, Parkinson-like symptoms, and cognitive
23 impairment. These effects can be observed in humans when blood methanol levels exceed
24 200 mg/L. The species differences in toxicity from acute exposures appear to be the result of a
25 limited ability of humans to metabolize formic acid.
26 Despite the existence of many case reports on acute human exposures, the knowledge
27 base for long-term, low-level exposure of humans to methanol is limited. The current TLV for
28 methanol is 200 ppm (262 mg/m3) (ACGIH, 2000). Controlled experiments with human
29 volunteers indicate that only minor neurobehavioral changes occur following 4-hour exposure to
30 this concentration. A limited study on self-reported health effects in 66 persons exposed to
31 methanol at levels that came close to or exceeded the NIOSH short-term ceiling of 800 ppm
32 (1048 mg/m3), in comparison with an age-matched group of 66 less or not exposed persons,
33 suggested a statistically significant increase in the incidence of CNS-related symptoms, such as
34 dizziness, nausea, headache, and blurred vision (Frederick et al., 1984). Impaired vision and
6-1 DRAFT -DO NOT CITE OR QUOTE
-------
1 nasal irritation were observed in a study of 33 methanol-exposed workers (Kawai et al., 1991).
2 None of the case reports or human studies have investigated cancer as a potential outcome of
3 methanol exposure.
4 A number of reproductive, developmental, subchronic and chronic exposure duration
5 studies have been conducted in mice, rats, and monkeys. This summary will focus primarily on
6 reproductive and developmental toxicity, and cancer as the main endpoints of concern. Sections
7 4.7, 5.1.1 and 5.2.1 contain more extensive summaries that consider the dose-related effects that
8 have been observed in other organ systems following subchronic or chronic exposure.
9 Although there is no evidence in humans, methanol has shown to be a reproductive and
10 developmental toxicant in several animal studies. No studies have been reported in which
11 humans have been exposed subchronically or chronically to methanol by the oral route of
12 exposure, and thus would be suitable for derivation of an oral RfD. Data exist regarding effects
13 from oral exposure in experimental animals, but they are more limited than data from the
14 inhalation route of exposure (see Sections 4.2, 4.3, and 4.4). Two oral studies in rats (Soffritti
15 et al., 2002a; U.S. EPA, 1986), one oral study in mice (Apaja, 1980) and several inhalation
16 studies in monkeys, rats and mice (NEDO, 1987, 1985/2008a, 2008b) of 90-days duration or
17 longer have been reported. While some noncancer effects of methanol exposure were noted in
18 these studies, principally in the liver and brain, they were either not quantifiable due to study
19 limitations or occurred at high doses relative to reproductive/developmental effects. As
20 discussed below, the results of inhalation reproductive/developmental toxicity studies in rats
21 (NEDO, 1987), mice (Rogers et al., 1993a), and monkeys (Burbacher et al., 2004a, 2004b,
22 1999a, 1999b) are the principal considerations for both the RfD and RfC values derived in this
23 assessment.
24 A larger number of studies have used the inhalation route to assess the potential of
25 reproductive or developmental toxicity of methanol in mice, rats, and monkeys, with
26 concentrations ranging from 200 to 20,000 ppm (blood levels reaching as high as 8.65 mg/mL).
27 To sum up the findings, rat dams survived even the highest doses without gross signs of toxicity,
28 but their offspring were severely affected (Nelson et al., 1985). Two more inhalation studies,
29 Rogers et al. (1993a,1993b) and Rogers and Mole (1997), confirmed that methanol causes
30 exencephaly and cleft palate in mice, the most sensitive days being GD6 and GD7 (i.e., early
31 organogenesis). These severe malformations were observed at exposure concentrations of
32 5,000 ppm or above. Nelson et al. (1985) and Rogers et al. (1993a) also observed an increased
33 occurrence of ossification disturbances and skeletal anomalies at methanol concentrations
34 > 2,000 ppm, of which cervical ribs in mouse fetuses is considered the critical effect for toxicity
35 value derivation in this review. A study conducted in pregnant cynomolgus monkeys that were
36 exposed to 200-600 ppm methanol for 2.5 hours/day throughout premating, mating, and
37 gestation showed no signs of maternal or fetal toxicity. The potential compound-related effects
6-2 DRAFT—DO NOT CITE OR QUOTE
-------
1 noted were a shortening of the gestation period by less than 5% and developmental neurotoxicity,
2 particularly delayed sensorimotor development (Burbacher et al., 2004a, 2004b, 1999a, 1999b).
3 While all of the above studies were conducted with exposure durations of 7 hours/day or
4 less, NEDO (1987) conducted a series of developmental/reproductive studies in rats that used
5 exposure times of 20 hours/day or more at concentrations between 500 and 5,000 ppm. Atwo-
6 generation study by these researchers that exposed the dams throughout pregnancy and the pups
7 through 8 weeks of age, demonstrated dose-dependent reductions in brain weights that forms the
8 basis for the RfC derived in this review.
9 Carcinogenic effects following methanol exposure were observed in a chronic drinking
10 water study in Eppley Swiss Webster mice (Apaja, 1980) and two chronic rat studies: a drinking
11 water study of Sprague-Dawley rats (Soffritti et al., 2002a) and an inhalation study of F344 rats
12 (NEDO, 1985/2008b). Following administration via drinking water, both Apaja (1980) and
13 Soffritti et al. (2002a) observed positive dose-response trends for increases in the incidence of
14 lymphomas in both test animal genders. Soffritti et al. (2002a) characterized the lymphomas in
15 their study as lymphoreticular, principally lympho-immunoblastic. EPA re-analyzed the
16 lymphoma data from the Soffritti et al. (2002a) study for quantification purposes, combining
17 only tumors of the same cell type origin. There was a slight increase in hepatocellular
18 carcinomas in male rats of all exposure groups of this study that was not statistically elevated
19 over controls in any group, but potentially this tumor is related to methanol exposure given the
20 low historical background rate for this tumor in this rat strain. As discussed in Section 5.4.1.1,
21 the other tumor increases reported by Soffritti et al. (2002a) are not quantifiable or were
22 considered hyperplastic rather than carcinogenic following a review by NTP pathologists (EFSA,
23 2006; Hailey, 2004). No tumor responses were significantly increased over controls in the
24 chronic inhalation bioassays performed by NEDO (1987) in monkeys, and mice, but the high-
25 dose incidences for pulmonary adenomas/adenocarcinomas in male rats was elevated over
26 concurrent controls and pheochromocytomas in female rats were significantly elevated over
27 historical control incidences for these tumor types within their respective sex and strain. The
28 dose response for both of these tumor types represents increasing trends (Cochran Armitage
29 trend test; p < 0.05). Further, both tumor responses are accompanied by proliferative changes
30 (e.g., hyperplastic responses) in their respective cell types.
6.2. DOSE RESPONSE
6.2.1. Noncancer/Inhalation
31 Clearly defined toxic endpoints at moderate exposure levels have been observed only in
32 reproductive and developmental toxicity studies. Three endpoints from developmental toxicity
33 studies were considered for derivation of the RfC: formation of cervical ribs in CD-I mice
34 exposed to methanol during organogenesis (Rogers et al., 1993a), deficits in sensorimotor
6-3 DRAFT—DO NOT CITE OR QUOTE
-------
1 development as measured by VDR tests administered to monkeys exposed to methanol
2 (Burbacher et al., 2004a, 2004b, 1999a, 1999b), and reduced brain weights in rats exposed to
3 methanol from early gestation through 8 weeks of postnatal life (NEDO, 1987). For the purpose
4 of comparability and to better illustrate methodological uncertainty, reference values were
5 derived for all of these endpoints using a BMD modeling approach which evaluated several
6 models and various measures of risk. In the present review, mostly because of a paucity of
7 adequate long-term or developmental oral studies and the existence of several inhalation studies
8 that examined sensitive subpopulations (pregnant mothers, developing fetuses and neonates) in
9 various species, it was decided to use the critical effect from an inhalation study to derive an
10 RfD. Thus, the criteria and rationales on which the RfC assessment is based also form the basis
11 for the RfD derivation.
12 The Rogers et al. (1993a) inhalation study is a multidose developmental study that was
13 considered for use in the derivation of a reference value. The exposure concentrations in this
14 study were 0, 1,000, 2,000, and 5,000 ppm administered for 7 hours/day on GD7-GD17. The
15 BMD evaluation, based on the nested log-logistic model of BMDS version 2.1, produced
16 BMD/BMDL values in terms of internal peak blood methanol (Cmax). PBPK modeling was used
17 to convert the internal animal dose metrics to HECs, and a UF of 100 was applied to yield RfCs
18 of 10.4 mg/m3 and 13.6 mg/m3 for 5 and 10% extra risk, respectively.
19 Reproductive and developmental neurobehavioral effects observed in monkeys following
20 methanol inhalation exposure (Burbacher et al., 2004a, 2004b, 1999a, 1999b ) were also
21 considered for use in the derivation of a reference value. M. fascicularis monkeys were exposed
22 to 0, 262, 786, and 2,359 mg/m3 methanol 2.5 hours/day, 7 days/week during premating/mating
23 and throughout gestation (approximately 168 days). Delayed sensorimotor development as
24 measured by a VDR test was the only effect in this study that exhibited a dose-response and is a
25 measure of a functional deficit that is consistent with early developmental CNS effects (e.g.,
26 brain weight changes) that have been observed in rats (NEDO, 1987). Though there is
27 uncertainty associated with this effect and its relation to methanol exposure, a BMD analysis was
28 performed for comparative purposes. BMD/BMDL values for the VDR endpoint were estimated
29 using AUCs derived from a monkey PBPK model of blood methanol data reported in the
30 Burbacher et al. (1999a) study. A human methanol PBPK modeling was then used to convert the
31 internal AUC BMDL to an FIEC, and a UF of 100 was applied to yield a reference value estimate
32 of 1.7 mg/m3.
33 Reduced brain weight was evaluated based on the results of a two-generation study by
34 NEDO (1987) in which fetal rats and their dams were exposed from the first day of gestation
35 until 8 weeks of age, and brain weights were determined at 3, 6, and 8 weeks of age. To obtain
36 reference value estimates from these studies, a rat PBPK model was used to predict PODs in
37 terms of internal doses, which were divided by UFs and converted to HEC reference values via a
6-4 DRAFT—DO NOT CITE OR QUOTE
-------
1 human PBPK model (see Table 5-4). BMD modeling was executed using two different BMRs,
2 one S.D. (as is usual with continuous data) and 5% relative (to control response) risk. The
3 resulting reference value estimates were 2.4 and 1.8 mg/m3 (5% relative risk and 1 S.D.,
4 respectively) for reduced brain weight at 6 weeks of age following gestational and postnatal
5 exposure.
6 Despite the variety of approaches, different critical effects, and different data sources, all
7 reference value estimates fell within a narrow range. The reference value associated with the
8 BMD estimate of the dose corresponding to a one S.D. decrease in brain weight in male rats at 6
9 weeks post-birth observed in the NEDO (1987) developmental toxicity study is considered most
10 suitable for derivation of the methanol chronic RfC due to the relevance of the exposure
11 scenario/study design (see Sections 5.1.2.2 and 5.3), and endpoint (see Section 5.3) to the
12 potential for developmental effects in neonatal humans, the relative robustness of the dose-
13 response data and because it resulted in one of the lowest reference values of the BMD
14 derivations (see Table 5-4). Thus, the proposed chronic RfC for exposure to methanol is 2
15 mg/m3, an evaluation that includes a UFH of 10 for intraspecies variability, a UFA of 3 to address
16 the pharmacodynamic component of interspecies variability, and a UFD of 3 for database
17 uncertainty.
18 The confidence in this RfC is medium to high. Confidence in the NEDO (1987)
19 developmental studies is medium to high. While there are issues with the lack of reporting
20 detail, the critical effect (brain weight reduction) has been reproduced in an oral study of adult
21 rats (U.S. EPA, 1986c), and the exposure regimen involving pre- and postnatal exposures
22 addresses a potentially sensitive human subpopulation. Confidence in the database is medium.
23 Despite the fact that skeletal and brain effects have been demonstrated and corroborated in
24 multiple animal studies in rats, mice, and monkeys, some study results were not quantifiable,
25 there is uncertainty regarding which is the most relevant test species, and there is limited data
26 regarding reproductive or developmental toxicity of methanol in humans. There is also
27 uncertainty regarding the potential active agent—the parent compound, methanol, formaldehyde,
28 or formic acid. There are deficiencies in our knowledge of the metabolic pathways of methanol
29 in the human fetus during early organogenesis, when the critical effects can be induced in
30 animals. Thus, the medium-to-high confidence in the critical study and the medium confidence
31 in the database together warrant an overall confidence descriptor of medium to high.
6.2.2. Noncancer/Oral
32 There is a paucity of scientific data regarding the outcomes of chronic oral exposure to
33 methanol. No data exist for long-term methanol exposure of humans. A subchronic (90-day)
34 oral study in Sprague-Dawley rats reported brain and liver weight changes, with some evidence
35 for minor liver damage at 2,500 mg/kg-day that was not supported by histopathologic findings
6-5 DRAFT—DO NOT CITE OR QUOTE
-------
1 (U.S. EPA, 1986c). Liver necrosis was reported in Eppley Swiss Webster mice that consumed
2 approximately 2000 mg/kg-day (Apaja, 1980). In the only other study that administered
3 methanol chronically to animals by the oral route, Soffritti et al. (2002a) reported that, overall,
4 there was no pattern of compound-related clinical signs of toxicity in Sprague-Dawley rats
5 exposed to up to approximately 2,000 mg/kg-day. The authors further reported that there were
6 no compound-related signs of gross pathology nor histopathologic lesions indicative of
7 noncancer toxicological effects in response to methanol; however, they did not provide any
8 detailed data to illustrate these findings.
9 As discussed above and in Section 5.1.1, reproductive and developmental effects are
10 considered the most sensitive and quantifiable effects reported in studies of methanol. Oral
11 reproductive and developmental studies employed single doses that were too high to be of use.
12 In the absence of suitable reproductive or developmental data from oral exposure studies, it was
13 decided to conduct a route-to-route extrapolation and to use the critical effect from the inhalation
14 study (brain weight) to derive an RfD. Thus, the POD (in terms of AUC methanol in blood) used
15 for the derivation of the RfC was also used for the derivation of the RfD. This POD was divided
16 by a UF of 100, and a human PBPK model was used to obtain an RfD value of 0.4 mg/kg-day.
17 As for the RfC, the 100-fold UF includes a UFn of 10 for intraspecies variability, a UFA of 3 to
18 address pharmacodynamic uncertainty, and a UFD of 3 for database uncertainty.
19 The confidence in the RfD is medium to high. Despite the relatively high confidence in
20 the critical studies, all limitations to confidence as presented for the RfC also apply to the RfD.
21 Confidence in the RfD is slightly lower than for the RfC due to the lack of adequate oral studies
22 for the RfD derivation, necessitating a route-to-route extrapolation.
6.2.3. Cancer/Oral and Inhalation
23 Under the current Guidelines for Carcinogen Risk Assessment (U.S. EPA 2005a, 2005b),
24 methanol fulfils the criteria to be described as likely to be a human carcinogen by all routes of
25 exposure. This descriptor is based principally on findings of dose-related, statistically significant
26 increases in the incidence of: lymphoreticular tumors in lifetime studies of both sexes of Eppley
27 Swiss Webster mice (Apaja, 1980) and both sexes of Sprague-Dawley rats (Soffritt et al., 2002a),
28 a slight but significant (compared to historical controls) increase in relatively rare hepatocellular
29 carcinomas in male Sprague-Dawley rats following oral exposure (Soffritti et al., 2002a), and
30 dose-related occurrences of pulmonary adenomas/adenocarcinomas and pheochromocytomas in
31 F344 rats by inhalation exposure (NEDO, 1985/2008b). This determination is supported by the
32 results of other studies that have shown tumorigenic responses similar to those observed by
33 Soffritti et al. (2002a) in rats exposed to formaldehyde, a metabolite of methanol, and to the
34 metabolic precursors of methanol and formaldehyde, aspartame and MTBE. In addition,
35 epidemiological studies have associated exposure to formaldehyde with increases in the
6-6 DRAFT—DO NOT CITE OR QUOTE
-------
1 incidence of both leukemias and lymphomas (IARC, 2004). However, the key studies, Soffritti
2 et al. (2002a), NEDO (1985/2008b) and Apaja (1980), have associated uncertainties (see below
3 and discussions in Sections 4.9.2 and 5.4.3) that reduce confidence in the chosen descriptor.
4 The statistically significant increase in the incidence of lymphoreticular tumors observed
5 in the Soffritti et al. (2002a) drinking water study of Sprague-Dawley rats was used in the
6 determination of the POD for estimating the methanol oral CSF. APBPK model was developed,
7 and several model predictions of internal dose metrics were considered for use in the dose-
8 response analysis and derivation of the human equivalent dose. Methanol metabolized was
9 selected as the dose metric best suited for derivation of the oral POD because of its superior fit to
10 the response data and consistency with the hypothesis that formaldehyde may be the
11 carcinogenic agent associated with methanol exposure. The EPA multistage cancer model was
12 used to derive a BMDLio for the male rat in terms of mg methanol metabolized/day. Assuming
13 that metabolized methanol distributes in the body according to body weight to the 3/4 power, the
14 rat BMDLio of 63.9 mg-day was converted to a human BMDLio of 3,553 mg-day. The human
15 PBPK model was then used to convert this human mg-day value for total methanol metabolized
16 back to a human equivalent methanol oral dose HED(BMDLio) of 51.5 mg/kg-day for
17 lymphomas in the male rat. The oral CSF of 2E-03 (mg/kg-day)"1 was then derived based on a
18 linear extrapolation from this POD to estimated background levels.
19 Pulmonary adenomas/adenocarcinomas in male F344 rats and pheochromocytomas in
20 female F344 rats observed in the chronic inhalation study of NEDO (1985/2008b) were
21 considered in the determination of the POD for estimating the inhalation CSF. In this case, all
22 dose metrics estimated by the PBPK model provided a similar acceptable fit to the tumor
23 response data. Methanol metabolized was selected as the dose metric for derivation of the
24 inhalation POD for consistency with the approach used for the derivation of the oral POD and
25 with the hypothesis that formaldehyde may be the carcinogenic agent associated with methanol
26 exposure. As for the oral POD, the EPA multistage cancer model was used to derive a BMDLio
27 for the rat in terms of mg methanol metabolized/day. Assuming that metabolized methanol
28 distributes in the body according to body weight to the % power, the rat BMDLio of 15 mg-day
29 was converted to a human BMDLio of 971 mg-day. The human PBPK model was then used to
30 convert this human mg-day value for total methanol metabolized back to a human equivalent
31 methanol inhalation concentration F£EC(BMCLio) of 81,900 ug/m3 for pheochromocytomas in
32 the female rat. The inhalation cancer unit risk of 1E-06 (ug/m3)"1 was then derived based on a
33 linear extrapolation from this POD to estimated background levels.
34 Section 5.4.3 of this assessment documents several uncertainties with the quantification
35 of cancer risk. The main uncertainties can be grouped into issues related to study quality, the
36 interpretation of study results, and the consistency of the results with other laboratories. Other
37 uncertainties discussed in Section 5.4.3 include the choice of tumor endpoint, the choice of dose-
6-7 DRAFT—DO NOT CITE OR QUOTE
-------
1 response model, the PBPK model and dose metric used for the animal to human extrapolations,
2 and the human relevance of the carcinogenic responses in rats and mice.
6-8 DRAFT—DO NOT CITE OR QUOTE
-------
7. REFERENCES
1 Abbondandolo, A; Bonatti, S; Corsi, C; et al. (1980) The use of organic solvents in
2 mutagenicity testing. Mutat Res 79:141-150.
O
4 ACGIH (American Conference of Governmental Industrial Hygienists). (2000) Threshold
5 Limit Values for Chemical Substances and Physical Agents and Biological Exposure Indices.
6 American Conference of Governmental Industrial Hygienists. Cincinnati, OH.
7
8 Adanir, J; Ozkalkanti, MY; Aksun, M. (2005) Percutaneous methanol intoxication: Case
9 report. Eur J Anaesthiol 22:560-561.
10
11 Agarwal, DP. (2001) Genetic polymorphisms of alcohol metabolizing enzymes. Pathol
12 Biol (Paris) 49(9):703-709.
13
14 Airas, L; Paavilainen, T; Marttila, RJ; et al. (2008) Methanol intoxication-induced
15 nigrostriatal dysfunction detected using 6-[18F]fluoro-L-dopa PET. Neurotoxicol 29:671-674.
16
17 Akaike, H. (1973) Information theory and an extension of the maximum likelihood
18 principle. Petrov, BN; Csaki, F; eds. Second International Symposium on Information Theory,
19 September 1971, Tsahkadsor, Armenia, USSR. Budapest, Hungary: Akademia Kiado.
20 pp.267-281.
21
22 Albert RE, Sellakumar AR, Laskin S, Kuschner M, Nelson N, Snyder CA. 1982.
23 Gaseous formaldehyde and hydrogen chloride induction of nasal cancer in the rat. J Natl Cancer
24 Inst. 68(4):597-603.
25
26 Albin, RL and Greenamyre, JT. (1992) Alternative excitotoxic hypotheses. Neurology
27 42:733-738.
28
29 Allen, BC; Kavlock, RJ; Kimmel, CA; et al. (1994) Dose-response assessment for
30 developmental toxicity II: comparison of generic benchmark dose estimates with no observed
31 adverse effect levels. Fundam Appl Toxicol 23:487-495.
32
33 Andresen, H; Schmoldt, H; Matschke, J; et al. (2008) Fatal methanol intoxication with
34 different survival times -morphological findings and postmortem methanol distribution. Forensic
35 Sci Intl 179:206-210.
36
7-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 Andrews, JE; Ebron-McCoy, M; Logsdon, TR; et al. (1993) Developmental toxicity of
2 methanol in whole embryo culture: a comparative study with mouse and rat embryos. Toxicology
3 81(3):205-215.
4
5 Andrews, JE; Ebron-McCoy, M; Kavlock, RJ; et al. (1995) Developmental toxicity of
6 formate and formic acid in whole embryo culture: a comparative study with mouse and rat
7 embryos. Teratology 51:243-251.
8
9 Andrews, JE; Ebron-McCoy, M; Schmid, JE; et al. (1998) Effects of combinations of
10 methanol and formic acid on rat embryos in culture. Teratology 58(2):54-61.
11
12 Andrews, LS; Clary, JJ; Terrill, JB; et al. (1987) Subchronic inhalation toxicity of
13 methanol. J Toxicol Environ Health 20:117-124.
14
15 Ang, HL; Deltour, L; Hayamizu, TF; et al. (1996) Retinoic acid synthesis in mouse
16 embryos during gastrulation and craniofacial development linked to class IV alcohol
17 dehydrogenase gene expression. J Biol Chem 271:9526-9534.
18
19 Antony, AC. (2007) In utero physiology: role of folic acid in nutrient delivery and fetal
20 development. Am J Clin Nutr 85(suppl):598S- 603S.
21
22 Apaja, M. (1980) Evaluation of toxicity and carcinogenicity of malonaldehyde. Acta
23 Universitatis Ouluensis Series D Medica 55 Anatomica, Pathologica, Microbiologica No. 8, 1-
24 59.
25
26 Arora, V; Nijjar, IS; Thukral, H; et al. (2005) Bilateral putaminal necrosis caused by
27 methanol intoxication- a case report. Ind J Radiol Imag 15:341-342.
28
29 Arora, V; Nijjar, IBS; Multani, AS; et al. (2007) MRI findings in methanol intoxication: a
30 report of two cases. Brit J Radiol 80:e243-e246.
31
32 Aschner, M; Kimmelberg, HK. (1996) The Role of Glia in Neurotoxicity, Published by
33 CRC Press, p.365.
34
35 ATSDR (Agency for Toxic Substances and Disease Registry). (1997) Toxicological
36 review of methyl tertiary butyl ether. Agency for Toxic Substances and Disease Registry,
37 Department of Health, Atlanta, GA.
38
7-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 ATSDR (Agency for Toxic Substances and Disease Registry). (1999) Toxicological
2 review of formaldehyde. Agency for Toxic Substances and Disease Registry, Department of
3 Health, Atlanta, GA.
4
5 Aufderheide, TP; White, SM; Brady, WJ; et al. (1993) Inhalational and percutaneous
6 methanol toxicity in two firefighters. Ann Emergency Med 22(12): 1916-1918.
7
8 Aziz, MH; Agrawal, AK; Adhami, VM; et al. (2002) Methanol-induced neurotoxicity in
9 pups exposed during lactation through mother: role of folic acid. Neurotoxicol Teratol 24(4):
10 519-527.
11
12 Azmak, D. (2006) Methanol related deaths in Edirne. Legal Med 8:35-42.
13
14 Barceloux, DG; Bond, GR; Krenzelok, EP; et al. (2002) American Academy of Clinical
15 Toxicology practice guidelines on the treatment of methanol poisoning. J Toxicol Clin Toxicol
16 40(4):415-446.
17
18 Battelle. (1981) Final report on a chronic inhalation toxicology study in rats and mice
19 exposed to formaldehyde. Prepared for the Chemical Industry Institute of Toxicology. Prepared
20 by Battelle Columbus Laboratories, Columbus, OH. CUT Docket # 10922.
21
22 Batterman, SA; Franzblau, A. (1997) Time-resolved cutaneous absorption and permeation
23 rates of methanol in human volunteers. Int Arch Occup Environ Health 70(5):341-351.
24
25 Batterman, SA; Franzblau, A; D'Arcy, JB; et al. (1998) Breath, urine, and blood
26 measurements as biological exposure indices of short-term inhalation exposure to methanol.
27 Int Arch Occup Environ Health 71:325-335.
28
29 Beam, P; Patel, J; O'Flynn, WR. (1993) Cervical ribs: a cause of distal and cerebral
30 embolism. Postgrad Med J 69:65-68.
31
32 Bebarta, VS; Heard, K; Dart, RC. (2006) Inhalational abuse of methanol products:
33 elevated methanol and formate levels without vision loss. Am J Emerg Med 24:725-728.
34
35 Belpoggi F; Soffritti M; Guarino M; et al. (2002a) Results of long-term experimental
36 studies on the carcinogenicity of Ethylene-bis-Dithiocarbamate (Mancozeb) in rats. Ann NY
37 AcadSci 982:123-136.
38
7-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 Belpoggi F; Soffritti M; Minardi F; et al. (2002b) Results of long-term carcinogenicity
2 bioassays on tert-amyl-methyl-ether (TAME) and di-isopropyl-ether (DIPE) in rats. Ann N Y
3 Acad Sci 982:70-86.
4
5 Belpoggi, F; Soffritti, M; Maltoni, C. (1995) Methyl-tertiary-butyl ether (MTBE) - a
6 gasoline additive - causes testicular and lymphohaematopoietic cancers in rats. Toxicol Ind
7 Health 11:119-149.
8
9 Belpoggi, F; Soffritti, M; Filippini, F; et al. (1997) Results of long-term experimental
10 studies on the carcinogenicity of methyl tert-butyl ether. Ann NY Acad Sci 837:77-95.
11
12 Belpoggi, F; Soffritti, M; Maltoni, C. (1998) Pathological characterization of testicular
13 tumours and lymphomas-leukaemias, and of their precursors observed in Sprague-Dawley rats
14 exposed to methy-tertiary-butyl-ether (MTBE). Eur J Oncol 3(3):201-206.
15
16 Belpoggi, F; Soffritti, M; Maltoni, C. (1999) Immunoblastic lymphomas in Sprague-
17 Dawley rats following exposure to the gasoline oxygenated additives Methyl-Tertiary-Butyl
18 Ether (MTBE) and Ethyl-Tertiary-Butyl Ether (ETBE): early observations on their natural
19 history. Eur J Oncol 4:563-572.
20
21 Bennett, IL, Jr; Gary, FH; Mitchell, GL, Jr; et al. (1953) Acute methyl alcohol poisoning:
22 a review based on experiences in an outbreak of 323 cases. Medicine 32:431-463.
23
24 Benton, CD; Calhoun, FP Jr. (1952) The ocular effects of methyl alcohol poisoning:
25 report of a catastrophe involving three hundred and twenty persons. Trans Am Acad Ophthalmol
26 Otolaryngol 56(6):875-885.
27
28 Bessell-Browne, RJ; Bynevelt, M. (2007) Two cases of methanol poisoning: CT and MRI
29 features. Aust Radiol 51:175-178.
30
31 Bhatia, KP; Marsden, CD. (1994) The behavioral and motor consequences of focal
32 lesions of the basal ganglia in man. Brain 117(Pt 4): 859-876.
33
34 Black, KA; Eells, JT; Noker, PE; et al. (1985) Role of hepatic tetrahydrofolate in the
35 species difference in methanol toxicity. Proc Natl Acad Sci 82:3854-3858.
36
7-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 Blancato, JN; Evans, MV; Power, FW; et al. (2007) Development and use of PBPK
2 modeling and the impact of metabolism on variability in dose metrics for the risk assessment of
3 methyl tertiary butyl ether (MTBE). J Environ Protect Sci 1:29-51.
4
5 Blanco, M; Casado, R; Vasquez; et al. (2006) CT and MR imaging findings in methanol
6 intoxication. Am J Neuroradiol 27:452-454.
7
8 Bosron, WF; Li, TK. (1986) Genetic polymorphism of human liver alcohol and aldehyde
9 dehydrogenases, and their relationship to alcohol metabolism and alcoholism. Hepatology
10 6(3):502-510.
11
12 Bouchard, M; Brunet, RC; Droz, PO; et al. (2001) A biologically based dynamic model
13 for predicting the disposition of methanol and its metabolites in animals and humans. Toxicol Sci
14 64:169-184.
15
16 Boutwell, RK. (1964) Some biological aspects of skin carcinogenesis. Prog Exp Tumor
17 Res 4:207-250.
18
19 Braden GL; Strayhorn CH; Germain MJ; et al. (1993) Increased osmolal gap in alcoholic
20 acidosis. Arch Intern Med 153(20):2377-80.
21
22 Brahmi, N; Blel, Y; Abidi, N; et al. (2007) Methanol poisoning in Tunisia: report of 16
23 cases. Clin Toxicol 45(6):717-20.
24
25 Branch, S; Rogers, JM; Brownie, CF; et al. (1996) Supernumerary lumbar rib:
26 manifestation of basic alterations in embryonic development of ribs. J. Appl. Toxicol.
27 16:115-119.
28
29 Brent, J; Lucas, M; Kulig, K; et al. (1991) Methanol poisoning in a six-week-old infant.
30 JPediatr 118:644-646.
31
32 Brien, JF; Clarke, DW; Richardson, B; et al. (1985) Disposition of ethanol in maternal
33 blood, fetal blood, and amniotic fluid of third-trimester pregnant ewes. Am J Obstet Gynecol
34 1985 Jul 1;152(5):583-90.
35
36 Brown, RP; Delp, MD; Lindstedt, SL; et al. (1997) Physiological parameters values for
37 physiologically based pharmacokinetic models. Toxicol Ind Health 13:407-484.
38
7-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 Bucher, JR. (2002) The National Toxicology Program rodent bioassay: designs,
2 interpretations, and scientific contributions. Ann N Y Acad Sci. Dec; 982:198-207.
3
4 Burbacher, TM. (1993) Neurotoxic effects of gasoline and gasoline consituents. Environ
5 Health Perspect 101:133-141.
6
7 Burbacher, T; Shen, D; Grant, K; et al. (1999a) Methanol disposition and reproductive
8 toxicity in adult females. Reproductive and offspring developmental effects following maternal
9 inhalation exposure to methanol in nonhuman primates. HEI Research Report Number 89: Part I:
10 9-68.
11
12 Burbacher, T; Grant, K; Shen, D; et al. (1999b) Developmental effects in infants exposed
13 prenatally to methanol. Reproductive and offspring developmental effects following maternal
14 inhalation exposure to methanol in nonhuman primates. HEI Research Report Number 89: Part I:
15 69-117.
16
17 Burbacher, TM; Grant, KS; Shen, DD; et al. (2004a) Chronic maternal methanol
18 inhalation in nonhuman primates (M. fascicularis): reproductive performance and birth outcome.
19 Neurotoxicol Teratol 26:639-650.
20
21 Burbacher, TM; Shen, DD; Lalovic, B; et al. (2004b) Chronic methanol inhalation in
22 nonhuman primates {M. fascicularis}: exposure and toxicokinetics prior to and during pregnancy.
23 Neurotoxicol Teratol 26:201-221.
24
25 Burnell, JC; Li, T-K; Bosron, WE (1989) Purification and steady-state kinetic
26 characterization of human liver b3b3 alcohol dehydrogenase. Biochemistry 28:6810-6815.
27
28 Burwell RD, Whealin J, Gallagher M. (1992) Effects of aging on the diurnal pattern of
29 water intake in rats. Behav Neural Biol. Nov;58(3): 196-203. PubMed PMID: 1456941.
30
31 Butchko, HH; Stargel, WW; Comer, CP; et al. (2002) Aspartame: review of safety. Regul
32 Toxicol Pharmacol 35:81-93.
33
34 Calabrese, EJ. (2001) Assessing the default assumption that children are always at risk.
35 Hum Ecol Risk Assess 7(l):37-59.
36
37 Caldwell, J; Jinot, J; Devoney, D; et al. (2008) Evaluation of evidence for infection as a
38 MOAfor induction of rat lymphoma. Environ Mol Mutagen 49(2): 155-164.
7-6 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Caldwell, J; Jinot, J; Devoney, D; et al. (2009) Author reply to comment on "Evaluation
3 of evidence for infection as the mode of action for induction of rat lymphoma." Environ Mol
4 Mutagen. Jan;50(l): 6-9.
5
6 Cameron, AM; Nilsen, OG; Haug, E; et al. (1984) Circulating concentrations of
7 testosterone, luteinizing hormone and follicle stimulating hormone in male rats after inhalation
8 of methanol. Arch Toxicol 7(Suppl):441-443.
9
10 Cameron, AM; Zahlsen, K; Haug, E; et al. (1985) Circulating steroids in male rats
11 following inhalation of n-alcohols. Arch Toxicol 8(Suppl):422-424.
12
13 Campbell, JA; Howard, DR; Backer, LC; et al. (1991) Evidence that methanol inhalation
14 does not induce chromosome damage in mice. Mutat Res 260:257-264.
15
16 Carson, BL; McCann, JL; Ells, JV III; et al. (1981) Methanol Health Effects. Prepared by
17 the Midwest Research Institute. Prepared for the Office of Mobile Source Air Pollution Control,
18 U. S. Environmental Protection Agency, Ann Arbor, MI. PB82-160797.
19
20 Caspi, R; Foerster, H; Fulcher, CA; et al. (2006) MetaCyc: a multiorganism database of
21 metabolic pathways and enzymes. Nucleic Acids Res 34:D511-6.
22
23 CERHR (Center for the Evaluation of Risks to Human Reproduction). (2004)
24 NTP-CERHR Expert Panel report on the reproductive and developmental toxicity of methanol.
25 Reprod Toxicol 18:303-390.
26
27 Chemfinder. (2002) Methanol. Available at http://chemfmder.cambridgesoft.com/.
28
29 Chen, JC; Schneiderman, JF; Wortzman, G. (1991) Methanol poisoning: bilateral
30 putaminal and cerebellar cortical lesions on CT and MR. J Comput Assist Tomogr 15(3):
31 522-524.
32
33 Chen-Tsi, C. (1959) Materials on the hygienic standardization of the maximally
34 permissible concentration of methanol vapors in the atmosphere. Gig Sanit 24:7-12 (as cited in
35 Carson, et al., 1981).
36
7-7 DRAFT - DO NOT CITE OR QUOTE
-------
1 Chernoff, N; Roger, JM. (2004) Supernumerary ribs in developmental toxicity bioassays
2 and in human populations: incidence and biological significance. J Toxicol Environ Health
3 7:437_449.
4
5 Chuwers, P; Osterloh, J; Kelly, T; et al. (1995) Neurobehavioral effects of low-level
6 methanol vapor exposure in healthy human volunteers. Environ Res 71(2): 141-150.
7
8 Cichoz-Lach, H; Partycka, J; Nesina, I; et al. (2007) Genetic polymorphism of alcohol
9 dehydrogenase 3 in digestive tract alcohol damage. Hepatogastroenterology 54(76): 1222-7.
10
11 Clarke, DW; Steenaart, NA; Brien, F. (1986) Disposition of ethanol and activity of
12 hepatic and placental alcohol dehydrogenase and aldehyde dehydrogenases in the third-trimester
13 pregnant guinea pig for single and short-term oral ethanol administration. Alcohol Clin Exp Res
14 10(3):330-6.
15
16 Clancy, B; Finlay, BL; Darlington, RB; et al. (2007) Extrapolating brain development
17 from experimental species to humans. Neurotoxicology 28(5):931-7.
18
19 Clary, JJ. (2003) Methanol, is it a developmental risk to humans? Regul Toxicol
20 Pharmacol 37:83-91.
21
22 Clewell, HJ III; Gentry, PR; Gearhart, JM; et al. (2001) Development of a physiologically
23 based pharmacokinetic model of isopropanol and its metabolite acetone. Toxicol Sci 63:160-172.
24
25 Cogliano, VJ; Grosse, Y; Baan, RA; et al. (2005) Meeting report: summary of IAEC
26 monographs on formaldehyde, 2-butoxyethanol, and l-tert-butoxy-2-propanol. Environ Health
27 Perspect 113:1205-1208.
28
29 Coleman, CN; Mason, T; Hooker, ER; et al. (1999) Developmental effects of intermittent
30 prenatal exposure to 1,1,1-trichloroethane in the rat. Neurotoxicol Teratol 21(6):699-708.
31
32 Connell, JL; Doyle, JC; Gurry, JF. (1980) The vascular complications of cervical ribs.
33 AustNZJSurg 50:125-130.
34
35 Cook, MR; Bergman, FJ; Cohen, HD; et al. (1991) Effects of methanol vapor on human
36 neurobehavioral measures. HEI Research Report Number 42.
37
7-8 DRAFT - DO NOT CITE OR QUOTE
-------
1 Cook, RJ; Champion, KM; Giometti, CS. (2001) Methanol toxicity and formate oxidation
2 in NEUT2 mice. Arch Biochem Biophys 393:192-198.
3
4 Cooper, RL; Mole, ML; Rehnberg, GL; et al. (1992) Effect of inhaled methanol on
5 pituitary and testicular hormones in chamber acclimated and nonacclimated rats. Toxicology
6 71:69-81.
7
8 Corley, RA; Barmett, GA; Ghanayem, BI. (1994) Physiologically based
9 pharmacokinetics of 2-butoxyethanol and its major metabolite, 2-butoxyacetic acid, in rats and
10 humans. Toxicol Appl Pharmacol 129:61-79.
11
12 Corley, RA; Mast, TJ; Carney, EW; et al. (2003) Evaluation of physiologically based
13 models of pregnancy and lactation for their application in children's health risk assessments. Crit
14 Rev Toxicol 33:137-211.
15
16 Crebelli, R; Conti, G; Conti, L; et al. (1989) A comparative study on ethanol and
17 acetaldehyde as inducers of chromosome malsegregation in Aspergillus nidulans. Mutat Res
18 215:187-195.
19
20 Cronholm, T. (1987) Effect of ethanol on the redox state of the coenzyme bound to
21 alcohol dehydrogenase studied in isolated hepatocytes. Biochem J 248:567-572.
22
23 Crump, KS. (1984) A new method for determining allowable daily intakes. Fundam Appl
24 Toxicol 4:854-871.
25
26 Crump, KS. (1995) Calculation of benchmark doses from continuous data. Risk Anal
27 15:79-89.
28
29 Cruzan, G. (2009) Assessment of the cancer potential of methanol. Critical Reviews in
30 Toxicology, 2009; 39(4): 347-363.
31
32 Cruzan, G; Cushman, JR; Andrews, LS; Granville, GC; Johnson, KA; Hardy, CJ;
33 Coombes, DW; Mullins, PA; Brown, WR. (1998) Chronic toxicity/oncogenicity study of styrene
34 in CD rats by inhalation exposure for 104 weeks. Toxicol Sci 46:266-281.
35
36 Cumming, ME; Ong, BY; Wade, JG; et al. (1984) Maternal and fetal ethanol
37 pharmacokinetics and cardiovascular responses in near-term pregnant sheep. Can J Physiol
38 Pharmacol 62(12): 1435-9.
7-9 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Davis, VE; Brown, H; Huff, JA; et al. (1967) The alteration of serotonin metabolism to
3 5-hydroxytrypophol by ethanol ingestion in man. J Lab Clin Med 69:132-140.
4
5 Davoli, E; Cappellini, L; Airoldi, L; et al. (1986) Serum methanol concentrations in rats
6 and in men after a single dose of aspartame. Food Chem Toxicol 24:187-189.
7
8 D'Alessandro, A; Osterloh, JD; Chuwers; et al. (1994) Formate in serum and urine after
9 controlled methanol exposure at the threshold limit value. Environ Health Perspect 102:178.
10
11 De Brabander, N; Wojciechowski, M; De Decker, K; et al. (2005) Fomepizole as a
12 therapeutic strategy in paediatric methanol poisoning. Eur J Pediatr 164:158-161.
13
14 De Flora, S; Zanacchi, P; Camoirano, A; et al. (1984) Genotoxic activity and potency of
15 135 compounds in the Ames reversion test and in a bacterial DNA-repair test. Mutat Res
16 133:161-198.
17
18 Degitz, SJ; Zucker, RM; Kawanishi, CY; et al. (2004a) Pathogenesis of methanol-induced
19 craniofacial defects in C57BL/6J mice. Birth Defects Res. A 70:172-178.
20
21 Degitz, SJ; Rogers, JM; Zucker, RM; et al. (2004b) Developmental toxicity of methanol:
22 pathogenesis in CD-I and C57BL/6J mice exposed in whole embryo culture. Birth Defects Res
23 A70:179-184.
24
25 Deltour, L; Foglio, MH; Duester, G. (1999) Metabolic deficiencies in alcohol
26 dehydrogenase ADH1, ADH3, and ADH4 null mutant. Overlapping roles of ADH1 and ADH4 in
27 ethanol clearance and metabolism of retinol to retinoic acid. J Biol Chem 274:16796-16801.
28
29 Dethlefs, R; Naraqi, S. (1978) Ocular manifestations and complications of acute methyl
30 alcohol intoxication. Med J Aust 2(10):483-485.
31
32 Devore, JL. (1995) Probability and statistics for engineering and the sciences (4th ed).
33 Belmont, CA: Duxbury.
34
35 Dicker, E; Cedebaum, AL (1986) Inhibition of the low Km mitachondrial aldehyde
36 dehydrogenase by diethyl maleate and phorone in vivo and in vitro implications for
37 formaldehyde metabolism. Biochem J 240:821-828.
38
7-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 Dikalova, AE; Kadiiska, MB; Mason, RP (2001) An in vivo ESR spin-trapping study:
2 free radical generation in rats from formate intoxication-role of the Fenton reaction. Proc Natl
3 AcadSci 98:13549-13553.
4
5 Dorman, DC; Moss, OR; Farris, GM; et al. (1994) Pharmacokinetics of inhaled (14C)
6 methanol and methanol-derived 14-C formate in normal and folate-deficient cynomolgus
7 monkeys. Toxicol Appl Pharmacol 128(2):229-238.
8
9 Dorman, DC; Bolon, B; Struve, MF; et al. (1995) Role of formate in methanol-induced
10 exencephaly in CD-I mice. Teratology 52(1):30-40.
11
12 Downie, A; Khattab, TM; Malik, MI; et al. (1992) A case of percutaneous industrial
13 methanol toxicity. Occup Med (London) 42(1): 47-49.
14
15 Dudka, J. (2006) The total antioxidant status in the brain after ethanol or 4-methyl-
16 pyrazole administration to rats intoxicated with methanol. Exp Toxicol Pathol 57:445-448.
17
18 Dutkiewicz, B; Konczalik, J; Karwacki, W. (1980) Skin absorption and per os
19 administration of methanol in men. Int Arch Occup Environ Health 47(l):81-88.
20
21 Eells, JT; Black, KA; Tedford, CE; et al. (1983) Methanol toxicity in the monkey: effects
22 of nitrous oxide and methionine. J Pharmacol Exp Ther 227:349-353.
23
24 EFSA (European Food Safety Authority). (2006) Opinion of the scientific panel on food
25 additives, flavourings, processing aids and materials in contact with food (AFC) on a request
26 from the commission related to a new long-term carcinogenicity study on aspartame. The EFSA
27 Journal 356:1-44.
28
29 EFSA (European Food Safety Authority). (2009) Updated Scientific Opinion of the Panel
30 on Food Additives and Nutrient Sources added to Food on a request from the European
31 Commission related to the 2nd ERF carcinogenicity study on aspartame, taking into consideration
32 study data submitted by the Ramazzini Foundation in February 2009. The EFSA Journal
33 1015:1-18.
34
35 Ernstgard, L; Shibata, E; Johanson, G. (2005) Uptake and disposition of inhaled methanol
36 vapor in humans. Toxicol Sci 88:30-38.
37
7-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 Ernstgard, L. (2005) Personal communication. E-mail correspondance from Lena
2 Ernstgard to Torka Poet dated March 23, 2005 [see Section B.3.5 of Appendix B].
3
4 Estonius, M; Svensson, S; H00g, JO. (1996) Alcohol dehydrogenase in human tissues:
5 localization of transcripts coding for five classes of the enzyme. FEES Lett 397:338-342.
6
7 Evans, AL. (1999) Pseudoseizures as a complication of painful cervical ribs. Dev Med
8 Child Neurol 41:840-842.
9
10 Everitt JI, Richter CB. (1990) Infectious diseases of the upper respiratory tract:
11 implication for toxicology studies. Environ Health Perspect 85:239-247.
12
13 Pagan, JF; Singer, LT. (1983) Infant recognition memory as a measure of intelligence. In:
14 Advances in Infancy Research, Ed, Lipsitt, LP Ablex, New York, NY (as cited in Burbacher
15 etal., 1999a)
16
17 Fallang, B; Saugstad, OD; Gragaard, J; et al. (2003) Kinematic quality of reaching
18 movements in preterm infants. Pediatr Res 53:836-842.
19
20 Feany, MB; Anthony, DC; Frosch, MP; et al. (2001) August 2000: two cases with
21 necrosis and hemorrhage in the putamen and white matter. Brain Pathol 11(1): 121-122.
22
23 Fernandez Noda, El; Nunez-Arguelles, J; Perez Fernandez, J; et al. (1996) Neck and
24 brain transitory vascular compression causing neurological complications, results of surgical
25 treatment on 1300 patients. J Cardiovasc Surg 37:155-166.
26
27 Finkelstein, Y; Vardi, J. (2002) Progressive Parkinsonism in a young, experimental
28 physicist following long-term exposure to methanol. Neurotoxicol 23:521-525.
29
30 Fiserova-Bergerova, V; Diaz, ML. (1986) Determination and prediction of tissue-gas
31 partition coefficients. Int Arch Occup Environ Health 58:75-87.
32
33 Fisher, JW; Dorman, DC; Medinsky, MA; et al. (2000) Analysis of respiratory exchange
34 of methanol in the lung of the monkey using a physiological model. Toxicol Sci 53:185-193.
35
36 Fontenot, AP; Pelak, VS. (2002) Development of neurologic symptoms in a 26-year-old
37 woman following recovery from methanol intoxication. Chest 122(4): 1436-1438.
38
7-12 DRAFT - DO NOT CITE OR QUOTE
-------
1 Foster, MW; Stamler, J. (2004) New insights into protein s-nitrosylation-
2 mitochondria as a model system. 279(24):25891-25897.
3
4 Frederick, LJ; Schulte, PA; Apol, A. (1984) Investigation and control of occupational
5 hazards associated with the use of spirit duplicators. Am Ind Hyg Assoc J 45:51-55.
6
7 Fu, SS; Sakanashi, TM; Rogers, JM; et al. (1996) Influence of dietary folic acid on the
8 developmental toxicity of methanol and the frequency of chromosomal breakage in the CD-I
9 mouse. Reprod Toxicol 10:455-463.
10
11 Gaiter, D; Carmine, A; Buervenich, S; et al. (2003) Distribution of class I, III, and IV
12 alcohol dehydrogenase mRNAs in the adult rat, mouse and human brain. Eur J Biochem
13 270:1316-1326.
14
15 Gannon KS, Smith JC, Henderson R, Hendrick P. (1992) A system for studying the
16 microstructure of ingestive behavior in mice. Physiol Behav. Mar;51(3):515-21. PubMed PMID:
17 1523228.
18
19 Gaul, HP; Wallace, CJ; Auer, RN et al. (1995) MR findings in methanol intoxication. Am
20 JNeuroradiol 16:1783-1786.
21
22 Gentry, PR; Covington, TR; Andersen, ME; et al. (2002) Application of a physiologically
23 based pharmacokinetic model for isopropanol in the derivation of a reference dose and reference
24 concentration. Regul Toxicol Pharmacol 36:51-68.
25
26 Gentry, PR; Covington, TR; Clewell, HJ III. (2003) Evaluation of the potential impact of
27 pharmacokinetic differences on tissue dosimetry in offspring during pregnancy and lactation.
28 Regul Toxicol Pharmacol 38:1-16.
29
30 Gibson, MAS; Butters, NS; Reynolds, JN; et al. (2000) Effects of chronic prenatal
31 ethanol exposure on locomotor activity, and hippocampal weight, neurons, and nitric oxide
32 synthase activity of the young postnatal guinea pig. Neurotoxicol Teratol 22:183-192.
33
34 Gonzalez-Quevedo, A; Obregon, F; Urbina, M; et al. (2002) Effect of chronic methanol
35 administration on amino acids and monoamines in retina, optic nerve, and brain of the rat.
36 Toxicol Appl Pharmacol 185(2):77-84.
37
7-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 Griffiths, AJF. (1981) Neurospora and environmentally induced aneuploidy. In Stich, HF;
2 San, RHC. eds. Short-term Tests for Chemical Carcinogens. Springer-Verlag, pp. 187-199.
3
4 Guerri, C and Sanchis, R. (1985) Acetaldehyde and alcohol levels in pregnant rats and
5 their fetuses. Alcohol 2(2):267-70.
6
7 Guggenheim, MA; Couch, JR; Weinberg, W. (1971) Motor dysfunction as a permanent
8 complication of methanol ingestion. Arch Neurol 24:550-554.
9
10 Hailey, JR. (2004) Chairperson's report: lifetime study in rats conducted by the
11 Ramazzini Foundation. Prepared by J.R. Hailey (pathology working group chair) NIEHS and
12 submitted to F. Belpoggi, November 30, 2004.
13
14 Hansen, JM; Contreras, KM; Harris, C. (2005) Methanol, formaldehyde, and sodium
15 formate exposure in rat and mouse conceptuses: a potential role of the visceral yolk sac in
16 embryotoxicity. Birth Defects Res (Part A) 73:72-82.
17
18 Hantson, P-E. (2005) Acute methanol intoxication: physiopathology, prognosis and
19 treatment. Bull Mem Acad R Med Belg 160:294-300.
20
21 Hantson, P; Duprez, T; Mahieu, P. (1997a) Neurotoxicity to the basal ganglia shown by
22 magnetic resonance imaging (MRI) following poisoning by methanol and other substances.
23 J Toxicol Clin Toxicol 35(2): 151-161.
24
25 Hantson, P; Lambermont, JY; Mahieu, P. (1997b) Methanol poisoning during late
26 pregnancy. J Toxicol Clin Toxicol 35(2): 187-191.
27
28 Hanzlik, RP; Fowler, SC; Eells, JT. (2005) Absorption and elimination of formate
29 following oral administration of calcium formate in female human subjects. Drug Metab Dispos
30 33:282-286.
31
32 Hassanian-Moghaddam, H; Pajoumand, A; Dadgar, SM; et al.. (2007) Prognostic factors
33 in methanol poisoning. Human Exp Toxicol 26(7):583-6.
34
35 Harris, C; Wang, S-W; Lauchu, JJ; et al. (2003) Methanol metabolism and
36 embryotoxicity in rat and mouse conceptuses: comparison of alcohol dehydrogenase (ADH1),
37 formaldehyde dehydrogenase (ADH3), and CAT. Reprod Toxicol 17:349-357.
38
7-14 DRAFT - DO NOT CITE OR QUOTE
-------
1 Harris, C; Dixon, M; Hansen, JM. (2004) Glutathione depletion modulates methanol,
2 formaldehyde, and formate toxicity in cultured rat conceptuses. Cell Biol Toxicol 20:133-145.
3
4 Harris, N; Jaffe, E; Diebold, J; et al. (1999) World Health Organization classification of
5 neoplastic diseases of the hematopoietic and lymphoid tissues: report of the clinical advisory
6 committee meeting-Airlie house, Virginia, November 1997. J Clin Oncol 17(12):3835-3849.
7
8 Haseman JK, Hailey JR, Morris RW (1998) Spontaneous neoplasm incidences in Fischer
9 344 rats and B6C3F1 mice in two-year carcinogenicity studies: ANational Toxicology Program
10 update. Toxicol Pathol 26:428-441.
11
12 Hass, U; Lund, SP; Simonsen, L; et al. (1995) Effects of prenatal exposure to xylene on
13 postnatal development and behavior in rats. Neurotoxicol Teratol 17(3):341-349.
14
15 Hauptmann, M; Lubin, JH; Stewart, PA; et al. (2003) Mortality from
16 lymphohematopoietic malignancies among workers in formaldehyde industries. J Natl Cancer
17 Inst 95:1615-1623.
18
19 Health Review Committee. (1999) Commentary. HEI Research Report Number 89:
20 pp.119-133.
21
22 Heck, H; Casanova, M. (2004) The implausibility of leukemia induction by
23 formaldehyde: a critical review of the biological evidence on distant-site toxicity. Regul Toxicol
24 Pharmacol 40:92-106.
25
26 Hedberg, JJ; Backlund, M; Stromberg, P; et al. (2001) Functional polymorphisms in the
27 alcohol dehydrogenase 3 (ADH3) promoter. Pharmacogenetics ll(9):815-824.
28
29 HEI. (1987) Special Report: Automotive methanol vapors and human health: An
30 evaluation of existing scientific information and issues for further research.
31
32 Heidelberger, C; Freeman, AE; Pienta, RJ; et al. (1983) Cell transformation by chemical
33 agents — a review and analysis of the literature. Areport of the U.S. Environmental Protection
34 Agency Gene-Tox Program. Mutat Res 114:283-385.
35
36 Henderson, MS. (1914) Cervical rib. Report of thirty-one cases. Am J Orthop Surg
37 11:408-430.
38
7-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 Henderson, WR; Brubacher, J. (2002) Methanol and ethylene glycol poisoning: a case
2 study and review of current literature. Can J Emerg Med 4:34-40.
3
4 Hess, DT; Matsumoto, A; Kim, SO; et al. (2005) Protein S-nitrosylation: purview and
5 parameters. Nat Rev Mol Cell Biol 6:150-166
6
7 Hinderer, RK; Johnson, MN, Seppala, A; Jacobs, MM; Apaja, M; Patil, KD; Shubik, P.
8 (1979) Toxicity and carcinogenicity of Agerite Resin D (ARD), trimethyldihydroquinoline
9 polymer. Submitted for publication to J. Env. Path. Toxicol. (as cited in Apaja, 1980)
10
11 Hines, RN; McCarver, DG. (2002) The ontogeny of human drug-metabolizing enzymes:
12 phase I oxidative enzymes. J Pharmacol Exp Ther 300:355-360.
13
14 Horton, VL; Higuchi, MA; Rickert, DE. (1992) Physiologically based pharmacokinetic
15 model for methanol in rats, monkeys, and humans. Toxicol Appl Pharmacol 117:26-36.
16
17 Hovda, KE; Hunderi, OH; Tafjord, A-B; et al. (2005) Methanol outbreak in Norway
18 2002-2004; epidemiology, clinical features and prognostic signs. J Int Med 258:181-190.
19
20 Hovda, KE; Mundal, H; Urdal, P; et al. (2007) Extremely slow formate elimination in
21 severe methanol poisoning: a fatal case report. Clin Toxicol 45:516-521.
22
23 HSDB (Hazardous Substances Data Bank). (2002) Available at
24 http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen7HSDB.
25
26 Hsu, HH; Chen, CY; Chen, FH et al. (1997) Optic atrophy and cerebral infarcts caused by
27 methanol intoxication. Neuroradiol 39:192-194.
28
29 Huang, QF; Gebrewold, A; Zhang, A; et al. (1994) Role of excitatory amino acids in
30 regulation of rat pial microvasculature. Am J Physiol 266:R158-R163.
31
32 Huff, J. (2002) Chemicals studied and evaluated in long-term carcinogenesis bioassays
33 by both the Ramazzini Foundation and the National Toxicology Program: In tribute to Cesare
34 Maltoni and David Rail. Ann NY Acad Sci 982208-982230.
35
36 Huff, J; LaDou, J. (2007) Aspartame bioassay findings portend human cancer hazards. Int
37 J Occup Environ HealthOct-Dec;13(4):446-8.
38
7-16 DRAFT - DO NOT CITE OR QUOTE
-------
1 Huff, J; Jacobson, MF and Davis, DL. (2009) The Limits of Two-Year Bioassay Exposure
2 Regimens for Identifying Chemical Carcinogens. Environ Health Perspect 116: 1439-1442.
3
4 Hunderi, OH; Hovda, KE; Jacobsen, D. (2006) Use of the osmolal gap to guide the start
5 and duration of dialysis in methanol poisoning. Scand J Urol Nephrol 40:70-74.
6
7 IARC (International Agency for Research on Cancer). (2004) Formaldehyde,
8 2-butoxyethanol and l-tert-butoxy-2-propanol. Summary of data reported and evaluation. IARC
9 Monographs on the Evaluation of Carcinogenic Risks to Humans 88:1-7.
10
11 Im, H; Mun, J; Khim, JY; et al. (2006) Evaluation of toxicological monitoring markers
12 using proteomic analysis in rats exposed to formaldehyde. J Proteome Res 5:1354-1366.
13
14 Infurna, R; Weiss, B. (1986) Neonatal behavioral toxicity in rats following prenatal
15 exposure to methanol. Teratology 33:259-265.
16
17 Ingemansson, SO. (1984) Clinical observations on ten cases of methanol poisoning with
18 respect to ocular manifestations. Acta Opthalmolog (Copenh) 62(1): 15-24.
19
20 IPCS (International Programme on Chemical Safety). (1997) Environmental Health
21 Criteria 196 - Methanol 180.
22
23 Ishii, H. (1981) Incidence of brain tumors in rats fed aspartame. Toxicol Lett 7:433^137
24 (as cited in EFSA, 2006).
25
26 Ishii, H; Koshimizu, T; Usami, S; et al. (1981) Toxicity of aspartame and its
27 diketopiperazine for Wistar rats by dietary administration for 104 weeks. Toxicol Lett 21:91-94
28 (as cited in EFSA, 2006)
29
30 Jacobsen, D; McMartin, KE. (1986) Methanol and ethylene glycol poisonings
31 Mechanism of toxicity, clinical course, diagnosis and treatment. Med Toxicol 1:309-334.
32
33 Jacobsen, D; Webb, R; Collins, TD; et al. (1988) Methanol and formate kinetics in late
34 diagnosed methanol intoxication. Med Toxicol 3:418-423.
35
36 Johanson, G; Kronborg, H; Naslund, PH; et al. (1986) Toxicokinetics of inhaled
37 2 butoxyethanol (ethylene glycol monobutyl ether) in man. Scand J Work Environ Health 12(6):
38 594-602.
7-17 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Johlin, FC; Fortman, CS; Nghiem, DD; et al. (1987) Studies on the role of folic acid and
3 folate-dependent enzymes in human methanol poisoning. Mol Pharmacol 31(5):557-61.
4
5 Kahn, A; Blum, D (1979) Methyl alcohol poisoning in an 8-month-old boy: an unusual
6 route of intoxication. J Pediatrics 94(5): 841-843.
7
8 Kamata E, Nakadate M, Uchida O, Ogawa Y, Suzuki S, Kaneko T, Saito M, Kurokawa Y.
9 1997. Results of a 28-month chronic inhalation toxicity study of formaldehyde in male Fisher-
10 344 rats. J Toxicol Sci. 22(3):239-54.
11
12 Kavet, R; Nauss, K. (1990) The toxicity of inhaled methanol vapors. Crit Rev Toxicol
13 21(1):21-50.
14
15 Kavlock, RJ; Allen, BC; Faustman, EM; et al. (1995) Dose-response assessments for
16 developmental toxicity. IV: Benchmark doses for fetal weight changes. Fund Appl Toxicol 26:
17 211-222.
18
19 Kawai, T; Yasugi, T; Mizunma, K; et al. (1991) Methanol in urine as a biological
20 indicator of occupational exposure to methanol vapor. Int Arch Occup Environ Health
21 63:311-318.
22
23 Keles, GT; Orgii9, S; Toprak, B; et al. (2007) Methanol poisoning with necrosis corpus
24 callosum. Clin Toxicol 45:307-308.
25
26 Kerns, WD; Pavkov, KL; Donofrio, DJ; et al. (1983) Carcinogenicity of formaldehyde in
27 rats and mice after long-term inhalation exposure. Cancer Res 43:4382-4392.
28
29 Kerns, WII; Tomaszewski, C; McMartin, K; et al. META Study Group. (2002) Formate
30 kinetics in methanol poisoning. J Toxicol Clin Toxicol 40:137-143.
31
32 Keys DA, Schultz IR, Mahle DA, Fisher JW. (2004) A quantitative description of suicide
33 inhibition of dichloroacetic acid in rats and mice. Toxicol Sci. Dec;82(2):381-93. Sep 16.
34 PubMedPMID: 15375292.
35
36 Kim, SW; Jang, YJ; Chang, JW; et al. (2003) Degeneration of the nigrostriatal pathway
37 and induction of motor deficit by tetrahydrobiopterin: an in vivo model relevant to Parkinson's
38 disease. Neurobiology of Disease 13:167-176.
7-18 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Knowles, K. (2008) Personal communication. E-mail correspondance from Kathryn
3 Knowles to Jeff Gift dated June 14, 2008.
4
5 Kraut, JA; Kurtz, I. (2008) Toxic alcohol ingestions: clinical features, diagnosis, and
6 management. Clin J Am Soc Nephrol 3(l):208-225.
7
8 Kuteifan, K; Oesterle, H; Tajahmady, T; et al., (1998) Necrosis and hemorrhage of the
9 putamen in methanol poisoning shown on MRI. Neuroradiol 40:158-160.
10
11 Leakey JEA, Seng JE, Allaben WT. 2003. Body weight considerations in the B6C3F1
12 mouse and the use of dietary control to standardize background tumor incidence in chronic
13 bioassays. Toxicol Appl Pharmacol 193:237-265.
14
15 Lee, EW; Brady, AN; Brabec, MJ; et al. (1991) Effects of methanol vapors on
16 testosterone production and testis morphology in rats. Toxicol Ind Health 7(4):261-275.
17
18 Lee, EW; Terzo, TS; D'Arcy, JB; et al. (1992) Lack of blood formate accumulation in
19 humans following exposure to methanol vapor at the current permissible exposure limit of
20 200 ppm. Am IndHyg Assoc J 53:99-104.
21
22 Lee, EW; Garner, CD; Terzo, TS. (1994) Animal model for the study of methanol
23 toxicity: comparison of folate-reduced rat responses with published monkey data. J Toxicol
24 Environ Health 41:71-82.
25
26 Lewis, RJ Jr. (1992) Sax's dangerous properties of industrial materials. 8th edition Van
27 Nostrand Reinhold, New York, NY.
28
29 Ley, CO; Gali, FG (1983) Parkinsonian syndrome after methanol intoxication. Eur
30 Neurol 22:405-409.
31
32 Lilly, MZ; Gold, KW; Rogers, RR. (1995) Integrating QA prinicples with basic elements
33 of a research program promotes quality science in a nonGLP research laboratory. In vitro Toxicol
34 8:191-198.
35
36 Lindsey, JR; Davidson, MK; Schoeb, TR; Cassell, GH. (1985)Mycoplasmapulmonis-
37 host relationships in a breeding colony of Sprague-Dawley rats with enzootic murine respiratory
38 mycoplasmosis. Lab Anim Sci 35:597-608.
7-19 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Liu, JJ; Daya, MR; Carrasquillo, O; et al. (1998). Prognostic factors in patients with
3 methanol poisoning. J Toxicol Clin Toxicol 36(3):175-181.
4
5 Liu, L; Hausladen, A; Zeng, M; et al. (2001) A metabolic enzyme for S-nitrosothiol
6 conserved from bacteria to humans. Nature 410:490-494.
7
8 Lorente, C; Cordier, S; Bergeret, A; et al. (2000) Maternal occupational risk factors for
9 oral clefts. Scand J Work Environ Health 26(2): 137-145.
10
11 Mahieu, P; Hassoun, A; Lauwerys, R; et al. (1989) Predictors of methanol intoxication
12 with unfavorable outcome. Hum Toxicol 8(2): 135-137.
13
14 Makar, AB; Tephly, TR; Mannering, GJ. (1968) Methanol metabolism in the monkey.
15 MolPharmacol 4:471-483.
16
17 Maltoni, C; Lefemine, G; Cotti, G (1986) Experimental research on trichlorothylene
18 carcinogenesis. Princeton: Princeton Scientific Publishing Co.
19
20 Maltoni C, Ciliberti A, Cotti G, Perino G. 1988. Long-term carcinogenicity bioassays on
21 acrylonitrile administered by inhalation and by ingestion to Sprague-Dawley rats. Ann N Y Acad
22 Sci. 534:179-202.
23
24 Maltoni, C; Ciliberti, A; Pinto, C; et al. (1997) Results of long-term experimental
25 carcinogenicity studies of the effects of gasoline, correlated fuels, and major gasoline aromatics
26 on rats. Ann NY Acad Sci 837:15-52.
27
28 Mann, WJ; Muttray, A; Schaefer, D; et al. (2002) Exposure to 200 ppm of methanol
29 increases the concentrations of interleukin-lbeta and interleukin-8 in nasal secretions of healthy
30 volunteers. Ann Otol Rhinol Laryngol 111:633-638.
31
32 Mannering, GJ; Van Harken, DR; Makar, AB; et al. (1969) Role of the intracellular
33 distribution of hepatic CAT in the peroxidative oxidation of methanol. Ann NY Acad Sci
34 168:265-280.
35
36 Maronpot, RR; Flake, G; Huff, J. (2004) Relevance of animal carcinogenesis findings to
37 human cancer predictions and prevention. Toxicol Pathol. Mar-Apr;32 Suppl 1:40-8. Review.
38
7-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 McGregor, DB; Martin, R; Riach, CG; et al. (1985) Optimization of a metabolic
2 activation system for use in the mouse lymphoma L5178Y tk+tk" system. Environ Mutagen
3 7(Suppl3):10.
4
5 McConnell, E; Solleveld, H; Swenberg, J; et al. (1986) Guidelines for combining
6 neoplasms for evaluation of rodent carcinogenesis studies. J Natl Cancer Inst 76(2):283-289.
7
8 McKellar, MJ; Hidajat, RR; Elder, MJ. (1997) Acute ocular toxicity: clinical and
9 electrophysiological features Aust NZ J Opthalmol. 25(3):225-230.
10
11 Medinsky, MA; Dorman, D; Bond, J; et al. (1997) Pharmacokinetics of methanol and
12 formate in female cynomolgus monkeys exposed to methanol vapors. HEI Research Report
13 Number 77:1-38.
14
15 Melnick RL, Thayer KA, and Bucker JR. 2007. Conflicting views on chemical
16 carcinogenesis arising from the design and evaluation of rodent carcinogenicity studies. Environ
17 Health Perspect 116:130-135; doi:10.1289/ehp.9989 (availableatdx.doi.org).
18
19 Meister, A; Anderson, ME. (1983) Glutathione. Ann Rev Biochem 52:711-760.
20
21 Methanol Institute. (2006) Methanol Institute Milestones. Washington, D.C. Available at
22 http ://www.methanol. org
23
24 Meyer, RJ; Beard, ME; Ardagh, MW; et al. (2000) Methanol poisoning. NZ Med J
25 113(1102):11-13.
26
27 Molinary, S.V. 1984. Preclinical studies of aspartame in nonprimate animals. In
28 Aspartame: Physiology and Biochemistry (L.D. Stegink and LJ. Filer, Jr., Eds.), pp. 289-306.
29 Marcel Dekker, Inc., New York.
30
31 Molotkov, A; Deltour, L; Foglio, MH; et al. (2002) Distinct retinoid metabolic functions
32 for alcohol dehydrogenase genes ADH1 and ADH4 in protection against vitamin A toxicity or
33 deficiency revealed in double null mutant mice. J Biol Chem 277:13804-13811.
34
35 Montserrat, CA; Field, MS; Perry, C; Ghandour, H; Chiang, E; Selhub, J; Shane, B; and
36 Stover, PJ. (2006) Regulation of Folate-mediated One-carbon Metabolism by 10-
37 Formyltetrahydrofolate Dehydrogenase. The Journal of Biological Chemistry 281(27): 18335-
38 18342.
7-21 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Mooney, SM; Miller, MW. (2001) Episodic exposure to ethanol during development
3 differentially effects brainstem nuclei in the macaque. J Neurocytol 30:973-982.
4
5 Muthuvel, A; Rajamani, R; Senthilvelan, M; et al. (2006a) Modification of allergenicity
6 and immunogenicity of formate dehydrogenase by conjugation with linear mono methoxy poly
7 ethylene glycol: improvement in detoxification of formate in methanol poisoning. Clin Chim
8 Acta 374:122-128.
9
10 Muthuvel, A; Rajamani, R; Manikandan, S; et al. (2006b) Detoxification of formate by
11 formate dehydrogenase-loaded erythrocytes and carbicarb in folate-deficient methanol-
12 intoxicated rats. Clin Chim Acta 367:162-169.
13
14 Nagasawa, H; Wada, M; Koyama, S; et al. (2005) A case of methanol intoxication with
15 optic neuropathy visualized on STIR sequence of MR images. Clin Neurol 45:527-530.
16
17 Naraqi, S; Dethlefs, RF; Slobodniuk, RA; et al. (1979) An outbreak of acute methyl
18 alcohol intoxication Aust N Z J Med 9(l):65-68.
19
20 NCI (National Cancer Institute). (2006) Prospective study of aspartame-containing
21 beverages and risk of hematopoietic and brain cancers. Abstract No. 4010. Available from the
22 American Association for Cancer Research (97th AACR Annual Meeting, Washington, DC) at
23 http://www.aacr.org (as cited in EFSA, 2006).
24
25 NEDO (New Energy Development Organization). (1987) Toxicological research of
26 methanol as a fuel for power station. Summary report on tests with monkeys, rats and mice. New
27 Energy Development Organization, Tokyo, Japan. Unpublished report.
28
29 NEDO. (1985/2008a) Test Report: 18-month inhalation carcinogenicity study on
30 methanol in B6C3F1 mice (Test No. 4A-223). Eight volume translation (consisting of report
31 text, summary tables, individual animal tables and copies of pathology slides) of original report
32 prepared by the Mitsubishi Kasei Institute of Toxicology and Environmental Sciences (currently
33 Mitsubishi Chemical Safety Insitute, Ltd.) for the New Energy Development Organization,
34 Tokyo, Japan. (Original report prepared in 1985; Translation submitted to the EPA IRIS hotline
35 by the Methanol Insitute on March 28, 2008 and certified as accurate and complete by NEDO in
36 letter dated May 12, 2008). Unpublished report.
37
7-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 NEDO. (1985/2008b) Test Report: 24-month inhalation carcinogenicity study on
2 methanol in Fischer rats (Test No. 5A-268). Eight volume translation (consisting of report text,
3 summary tables, individual animal tables and copies of pathology slides) of original report
4 prepared by the Mitsubishi Kasei Institute of Toxicology and Environmental Sciences (currently
5 Mitsubishi Chemical Safety Insitute, Ltd.) for the New Energy Development Organization,
6 Tokyo, Japan. (Original report prepared in 1985; Translation submitted to the EPA IRIS hotline
7 by the Methanol Insitute on March 28, 2008 and certified as accurate and complete by NEDO in
8 letter dated May 12, 2008). Unpublished report.
9
10 Nelson, BK; Brightwell, WS; MacKenzie, DR; et al. (1985) Teratological assessment of
11 methanol and ethanol at high inhalation levels in rats. Fund Appl Toxicol 5:727-736.
12
13 Nelson, J. (1967) Respiratory infections of rats and mice with emphasis on indigenous
14 mycoplasms. In: Pathology of laboratory rats and mice. Cotchin, E; Rose, F, eds. Blackwell
15 Scientific Publications, Oxford and Edinburgh, G.B. pp. 259-294.
16
17 Nguyen, T; Baumgartner, F; Nelems, B. (1997) Bilateral rudimentary first ribs as a cause
18 of thoracic outlet syndrome. J Natl Med Assoc 89:69-73.
19
20 Nilsson, JA; Hedberg, JJ; Vondracek, M; et al. (2004) Alcohol dehydrogenase 3
21 transcription associates with proliferation of human oral keratinocytes. Cell Mol Life Sci
22 61:610-617.
23
24 Noker, PE; Eells, JT; Tephly, TR; et al. (1980) Methanol toxicity: treatment with folic
25 acid and 5-formyl tetrahydrofolic acid. Alcohol Clin Exp Res 4:378-383.
26
27 NRC (National Research Council). (1994) Science and judgment in risk assessment.
28 Prepared by the Committee on Risk Assessment of Hazardous Air Pollutants, Board on
29 Environmental Studies and Toxicology, Commission on Life Sciences. National Academy Press,
30 Washington, D.C.
31
32 NRC (National Research Council). (1983) Risk assessment in the federal government:
33 managing the process. Washington, DC: National Academy Press.
34
35 NTP (National Toxicology Program). (1990) Toxicology and Carcinogenesis Studies of
36 Toluene (CAS No. 108-88-3) in F344/N Rats and B6C3F1 Mice (Inhalation Studies). Public
37 Health Service, U.S. Department of Health and Human Services; NTP TR 371; Available from
7-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 the National Institute of Environmental Health Sciences, Research Triangle Park, NC and online
2 at http://ntp.mehs.nih.gov/ntp/htdocs/LT_rpts/tr371 .pdf.
3
4 NTP (National Toxicology Program). (1999) Toxicology Data Management System,
5 Historical Control Tumor Incidence Summary. - Inhalation Studies of Fischer 344 Rats. Report
6 updated 12/99. Available at http://ntp.niehs.nih.gov.
7
8 NTP (National Toxicology Program). (2000) NTP technical report on the toxicology and
9 carcinogenesis studies of 2 butoxyethanol (CAS No. Ill 76 2) in F344/N rats and B6C3F1 mice
10 (inhalation studies). Public Health Service, U.S. Department of Health and Human Services;
11 NTP TR 484; NIH Publ. No. 00-3974. Available from the National Institute of Environmental
12 Health Sciences, Research Triangle Park, NC and online at
13 http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr484.pdf.
14
15 NTP (National Toxicology Program). (2003) NTP Technical Report. Toxicity studies of
16 aspartame in FVB/N-TgN(v-Ha-ras)Led (Tg.AC) hemizygous mice and carcinogenicity studies
17 of aspartame in B6.129-Trp53m&Brd (N5) haploinsufficient mice. NTP GMM 1, Research
18 Triangle Park, NC (as cited in EFSA, 2006).
19
20 NTP (National Toxicology Program). (2006) Specifications for the conduct of studies to
21 evaluate the toxic and carcinogenic potential of chemical, biological and physical agents in
22 laboratory animals for the national toxicology program (NTP). Available at
23 http://ntp.Niehs.Nih.Gov/files/specifications 2006octl.Pdf
24
25 NTP (National Toxicology Program). (2007) NTP Historical Controls Report, All Routes
26 and Vehicles, Rats. Report prepared October, 2007. Available at http://ntp.niehs.nih.gov.
27
28 OECD (Organization for Economic Co-operation and Development ). (2007)
29 Establishment and control of archives that operate in compliance with the principles of GLP
30 OECD Environment, Health and Safety Publications. No. 15. OECD Environment Directorate,
31 Paris, France.
32
33 Ohno, K; Tanaka-Azuma, Y; Yoneda, Y; et al. (2005) Genotoxicity test system based on
34 p53R2 gene expression in human cells: Examination with 80 chemicals. Mutat Res 588:47-57.
35
36 Onder, F; Ilker, S; Kansu, T; et al. (1999) Acute blindness and putaminal necrosis in
37 methanol intoxication. Int Opthalmol 22:81-84.
38
7-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 Osterloh, JD; D'Alessandro, A; Chuwers, P; et al. (1996) Serum concentrations of
2 methanol after inhalation at 200 ppm. J Occup Environ Med 38:571-576.
3
4 Pace, V; Perentes, E; Germann, PG. (2002) Pheochromocytomas and ganglioneuromas in
5 the aging rats: morphological and immunohistochemical characterization. Toxicol Pathol
6 30(4):492-500.
7
8 Parthasarathy, NJ; Kumar, RS; Karthikeyan, P; et al. (2005a) In vitro and in vivo study of
9 neutrophil functions after acute methanol intoxication in albino rats. Toxicol Environ Chem
10 87:559-566.
11
12 Parthasarathy, NJ; Kumar, RS; Devi, RS. (2005b) Effect of methanol intoxication on rat
13 neutrophil functions. J Immunotoxicol 2:115-211.
14
15 Parthasarathy, NJ; Kumar, RS; Manikandan, S; et al. (2006a) Effect of methanol-induced
16 oxidative stress on the neuroimmune system of experimental rats. Chemico-Biological Inter
17 161:14-25.
18
19 Parthasarathy, NJ; Kumar, RS; Manikandan, S; et al. (2006b) Methanol-induced oxidative
20 stress in rat lymphoid organs. J Occup Health 48:20-27.
21
22 Parthasarathy, NJ; Kumar, RS; Manikandan, S; et al. (2007) Effect of methanol
23 intoxication on specific immune functions of albino rats. Cell Biol Toxicol 23:177-187.
24
25 Pastino, GM; Conolly, RB. (2000) Application of a physiologically based pharmaco-
26 kinetic model to estimate the bioavailability of ethanol in male rats: distinction between gastric
27 and hepatic pathways of metabolic clearance. Toxicol Sci 55(2):25665.
28
29 Patankar, T; Bichile, L; Karnad, D; et al. (1999) Methanol poisoning: computed
30 tomography scan findings in four patients. Austral Radiol 43(4):526-528.
31
32 Pelletier, JU; Habib, MH; Khalil, R; et al. (1992) Putaminal necrosis after methanol
33 intoxication. J Neurol Neurosurg Psychiatry 55(3): 234-235.
34
35 Peng, MT; Chen, YT; Hung, SH; Yaung, CL. (1990) Circadian rhythms of feeding and
36 drinking behavior of rats aged from 3 to 21 months. Proc Natl Sci Counc Repub China B.
37 Apr;14(2):98-104. PubMed PMID: 2247537.
38
7-25 DRAFT - DO NOT CITE OR QUOTE
-------
1 Perkins, RA; Ward, KW; Pollack, GM. (1995a) Comparative toxicokinetics of inhaled
2 methanol in the female CD-I mouse and Sprague-Dawley rat. Fundam Appl Toxicol 28:
3 245-254.
4
5 Perkins, RA; Ward, KW; Pollack, GM. (1995b) A pharmacokinetic model of inhaled
6 methanol in humans and comparison to methanol disposition in mice and rats. Environ Health
7 Perspect 103:726-73 3.
8
9 Perkins, RA; Ward, KW; Pollack, GM. (1996) Methanol inhalation: site and other factors
10 influencing absorption, and an inhalation toxicokinetic model for the rat. Pharm Res 13:749-755.
11
12 Phang, PT; Passerini, L; Mielke, B; et al. (1988) Brain hemorrhage associated with
13 methanol poisoning. Crit Care Med 16:137-140.
14
15 Pietruszko, R. (1980) Alcohol and aldehyde dehydrogenase isozymes from mammalian
16 liver-their structural and functional differences. Isozymes Curr Top Biol Med Res 4:107-130.
17
18 Pikkarainen, PH; Raiha, NCR. (1967) Development of alcohol dehydrogenase activity in
19 the human liver. Pediatr Res 1(3): 165-168.
20
21 Pinkerton, LE; Hein, MJ; Stayner, LT. (2004) Mortality among a cohort of garment
22 workers exposed to formaldehyde: an update. Occup Environ Med 61:193-200.
23
24 Pitkin, RM. (2007) Folate and neural tube defects. Am J Clin Nutr 85:2858-2888.
25
26 Plantinga, Y; Perdock, J; de Groot, L. (1997) Hand function in low-risk preterm infants:
27 its relation to muscle power regulation. Dev Med Child Neurol 39:6-11.
28
29 Poet, TS; Teeguarden, JG; Hinderliter, PM. (2006) Final Report: Development,
30 calibration and application of a methanol PBPK model. Prepared for the National Center for
31 Environmental Assessment, U.S. EPA. Prepared by the Center for Biological Monitoring and
32 Modeling, Battelle Pacific Northwest National Laboratory, Richland, WA.
33
34 Pollack, GM; Brouwer, KL. (1996) HEI Research Report Number 74: Maternal-Fetal
35 Pharmacokinetics of Methanol. 58.
36
37 Pollack, GM; Brouwer, KL; Kawagoe, JL. (1993) Toxicokinetics of intravenous methanol
38 in the female rat. Fundam Appl Toxicol 21:105-110.
7-26 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Pollack, GM; Kawagoe, JL. (1991) Determination of methanol in whole blood by
3 capillary gas chromatography with direct on-column injection. J Chromatogr B Biomed Appl
4 570:406-11.
5
6 Poon, R; Chu, IH; Bjarnason, S; et al. (1994) Inhalation toxicity study of methanol,
7 toluene and methanol/toluene mixtures in rats: effects of 28-day exposure. Toxicol Ind Health
8 10(3):231-245.
9
10 Poon, R; Chu, IH; Bjarnason, S; et al. (1995) Short-term inhalation toxicity of methanol,
11 gasoline and methanol/gasoline in the rat. Toxicol Ind Health 11(3):343-361.
12
13 Prabhakaran, V; Ettler, H; Mills, A; et al. (1993) Methanol poisoning: two cases with
14 similar plasma methanol concentrations but different outcomes. Can Med Assoc J 148(6):
15 981-984.
16
17 Que, LG; Yan, Y; Whitehead, GS; et al. (2005) Protection from experimental asthma by
18 an endogenous bronchodilator. Science 308:1618-1621.
19
20 Rajamani, R; Muthuvel, A; Senthilvelan, M; et al. (2006) Oxidative stress induced by
21 methotrexate alone and in the presence of methanol in discrete regions of the rodent brain, retina
22 and optic nerve. Toxicol Lett 165:265-273.
23
24 Ramazzini Foundation. 2006. European Ramazzini Foundation stands behind aspartame
25 study results, announces ongoing research on artificial sweeteners. Available at
26 http://www.ramzzini/it/eng/fondazione/eventidettagli.asp?id5292
27
28 Ramsey, JC; Andersen, ME. (1984) A physiologically based description of the inhaled
29 pharmacokinetics of styrene in rats and humans. Toxicol Appl Pharmacol 73:159-175.
30
31 Rappaport, H. (1966) Tumors of the hematopoietic system. In: Atlas of Tumor Pathology,
32 Sect III. Fasc. 9: Armed Forces Institute of Pathology. Wahsington, D.C., pp. 49-64 [as cited in
33 Apaja, 1989]
34
35 Razin, S; Yogev, D; Naot, Y. (1998) Molecular biology and pathogenicity of
36 mycoplasmas. Microbiol Mol Biol Rev. Dec; 62(4): 1094-156. Review.
37
7-27 DRAFT - DO NOT CITE OR QUOTE
-------
1 Rice, D; Barone, S Jr. (2000) Critical periods of vulnerability for the developing nervous
2 system: evidence from humans and animal models. Environ Health Perspect 108(Suppl 3):
3 511-33.
4
5 Riegel, H; Wolf, G. (1966) Severe neurological deficiencies as a consequence of methyl
6 alcohol poisoning. Fortschr Neurol Psychiat 34:346-351.
7
8 Rogers, JM; Barbee, BD; Mole, ML. (1995) Exposure concentration and time (CxT)
9 relationships in the developmental toxicity of methanol in mice. Toxicologist 15:164.
10
11 Rogers, JM; Barbee, BD; Rhenberg, BF. (1993b) Critical periods of sensitivity for the
12 developmental toxicity of inhaled methanol. Teratology 47:395.
13
14 Rogers, JM; Brannen, KC; Barbee, BD; et al. (2004) Methanol exposure during
15 gastrulation causes holoprosencephaly, facial dysgenesis, and cervical vertebral malformations in
16 C57BL/6J mice. Birth Defects Res B 71:80-88.
17
18 Rogers, JM; Mole, ML. (1997) Critical periods of sensitivity to the developmental
19 toxicity of inhaled methanol in the CD-I mouse. Teratology 55(6):364-372.
20
21 Rogers, JM; Mole, ML; Chernoff, N; et al. (1993a) The developmental toxicity of inhaled
22 methanol in the CD-I mouse, with quantitative dose-response modeling for estimation of
23 benchmark doses. Teratology 47(3): 175-88.
24
25 Rogers, VV; Wickstrom, M; Liber, K; et al. (2002) Acute and Subchronic Mammalian
26 Toxicity of Naphthalenic Acids from Oil Sands Tailings. Toxicol Sci 66:347-355.
27
28 Rotenstreich, Y; Assia, El; Kesler, A. (1997) Late treatment of methanol blindness. Brit J
29 Opthalmol 81:415-420.
30
31 Rubinstein, D; Escott, E; Kelly, JP (1995) Methanol intoxication with putaminal and
32 white matter necrosis: MR and CT findings. Am J Neuroradiol 16(7): 1492-1494.
33
34 Sakanashi, TM; Rogers, JM; Fu, SS; et al. (1996) Influence of maternal folate status on
35 the developmental toxicity of methanol in the CD-I mouse. Teratology 54:198-206.
36
37 Salzman, M. (2006) Methanol neurotoxicity. Clin Toxicol 44:89-90.
38
7-28 DRAFT - DO NOT CITE OR QUOTE
-------
1 Saxton, EH; Miller, TQ; Collins, JD. (1999) Migraine complicated by brachial
2 plexopathy as displayed by MRI and MRA: aberrant subclavian artery and cervical ribs. J Natl
3 MedAssoc 91:333-341.
4
5 Sayers, RR; Yant, WP; Schrenk, HH; et al. (1944) Methanol poisoning: II. Exposure of
6 dogs for brief periods eight times daily to high concentrations of high methanol vapor in air.
7 J. Ind Hyg Toxicol 28:255-259.
8
9 Schaefer, PW; Grant, PE; Gonzalez, RG (2000) Diffusion-weighted MR imaging of the
10 brain. Radiol 217(2):331-345.
11
12 Schoeb, TR; Mcconnell, EE; Juliana, MM; Davis, JK; Davidson, MK; and Lindsey, JR.
13 (2009) Mycoplasmapulmonis and Lymphoma in Bioassays in Rats. Vet Pathol 46.
14
15 Schulte, PA; Burnett, CA; Boeniger, MF; et al. (1996) Neurodegenerative diseases:
16 occupational occurrence and potential risk factors, 1982-1991. Am J Public Health 86:
17 1281-1288.
18
19 Schumacher, R; Mai, A; Gutjahr, P. (1992) Association of rib abnormalities and
20 malignancy in childhood. Eur J Pediatr 151:432-434.
21
22 Scrimgeour, EM; Dethlefs, RF; Kevau, I. (1982) Delayed recovery of vision after
23 blindness caused by methanol poisoning. Med J Aust 2(10):481-483.
24
25 Sedivec, V; Mraz, M; Flek, J. (1981) Biological monitoring of persons exposed to
26 methanol vapors. Int Arch Occup Environ Health 48:257-271.
27
28 Sefidbakht, S; Rasekhi, AR; Kamali, K; et al. (2007) Methanol poisoning: acute MR and
29 CT findings in nine patients. Neuroradiol 49:427-435.
30
31 Sellakumar AR, Snyder CA, Solomon JJ, Albert RE. 1985. Carcinogenicity of
32 formaldehyde and hydrogen chloride in rats. Toxicol Appl Pharmacol. 81(3 Pt l):401-6.
33
34 Short, DW. 1975. The subclavian artery in 16 patients with complete cervical ribs.
35 J Cardiovasc Surg 16:135-141.
36
7-29 DRAFT - DO NOT CITE OR QUOTE
-------
1 Simintzi, I; Schulpis, KH; Angelogianni, P; et al. (2007) The effect of aspartame
2 metabolites on the suckling rat frontal cortex acetylcholinesterase. An in vitro study. Food Chem
3 Toxicol 45:2397-2401.
4
5 Simmon, VF; Kauhanen, K; Tardiff, RG. (1977) Mutagenic activity of chemicals
6 identified in drinking water. Scott, D; Bridges, B; Sobel, F. eds. Progress in Genetic Toxicology.
7 Vol 2. Elsevier/North Holland Press; pp. 249-268.
8
9 Sinkeldam, E; Kuper, C; Beems, R; et al. (1991) Combined chronic toxicity and
10 carcinogenicity study with acesulfame-k in rats. Acesulfame-k (Mayer, D; Kemper, F, eds.). New
11 York: Marcel Dekker; pp. 43-58.
12
13 Siragusa, RJ; Cerda, JJ; Baig, MM; et al. (1988) Methanol production from the
14 degradation of pectin by human colonic bacteria. Am J Clin Nutr. 47:848-51
15
16 Skrzydlewska, E; Elas, M; Ostrowska, J. (2005) Protective effects of N-acetylcysteine
17 and vitamin E derivative U83836E on proteins modifications induced by methanol intoxication.
18 Toxicol Mechanisms Methods 15:263-270.
19
20 Smith, ME; Newman, HW. (1959) The rate of ethanol metabolism in fed and fasting
21 animals. J Biol Chem 234:1544-1549.
22
23 Smith, M; Hopkinson, DA; Harris, H. (1971) Developmental changes and polymorphism
24 in human alcohol dehydrogenase. Ann Hum Genet 34:251-271.
25
26 Soffritti, M; Maltoni, C; Maffei, F; et al. (1989) Formaldehyde: an experimental
27 multipotential carcinogen. Toxicol Ind Health 5:699-730.
28
29 Soffritti, M; Belpoggi, F; Cevolani, D; et al. (2002a) Results of long-term experimental
30 studies on the carcinogenicity of methyl alcohol and ethyl alcohol in rats. Ann NY Acad Sci
31 982:46-69.
32
33 Soffritti, M; Belpoggi, F; Lambertini, L; et al. (2002b) Results of long-term experimental
34 studies on the carcinogenicity of formaldehyde. Ann NY Acad Sci 982:87-105.
35
36 Soffritti, M; Belpoggi, F; Minardi, F; et al. (2002c) Ramazzini Foundation cancer
37 program: history and major projects, life span carcinogenicity bioassay design, chemicals
38 studied, and results. Ann NY Acad Sci 982:26-45.
7-30 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Soffritti, M; Belpoggi, F; Degli Esposti, D; et al. (2005) Aspartame induces lymphomas
3 and leukemias in rats. Eur J Oncol 10:107-116.
4
5 Soffritti, M; Belpoggi, F; Degli Esposti, D; et al. (2006) First experimental demonstration
6 of the multipotential carcinogenic effects of aspartame administered in the feed to Sprague-
7 Dawley rats. Environ Health Perspect 114:379-385.
8
9 Soffritti, M; Belpoggi, F; Tibaldi, E; et al. (2007) Life span exposure to low doses of
10 aspartame beginnig during prenatal life increases cancer effects in rats. Environ Health Perspect
11 115:1293-1297.
12
13 Speit, G. (2006) The implausibility of systemic genotoxic effects measured by the comet
14 assay in rats exposed to formaldehyde. J Proteosome Res 5:2523-2524.
15
16 Spiteri NJ. (1982) Circadian patterning of feeding, drinking and activity during diurnal
17 food access in rats. Physiol Behav. Jan;28(l): 139-47. PubMed PMID:7200613.
18
19 Staab, Ca; Lander, J; Brandt, M; et al. (2008) Reduction of S-nitrosoglutathione by
20 alcohol dehydrogenase 3 is facilitated by substrate alcohols via direct cofactor recycling and
21 leads to GSH-controlled formation of glutathione transferase inhibitors. Biochem. J. 413:
22 493-504
23
24 Staats, DA; Fisher, JW; Conolly, RB. (1991) GI absorption of xenobiotics in
25 physiologically based pharmacokinetic models: a two-compartment description. Drug Metab
26 Disp 19(1): 144-8.
27
28 Stanton, ME; Crofton, KM; Gray, LE; et al. (1995) Assessment of offspring development
29 and behavior following gestational exposure to inhaled methanol in the rat. Fundam Appl
30 Toxicol 28(1): 100-110.
31
32 Starr, TB; Festa, JL. (2003) A proposed inhalation reference concentration for methanol.
33 Regul Toxicol Pharmacol 38:224-231.
34
35 Stegink, LD; Brummel, MC; McMartin, KE; et al. (1981) Blood methanol concentrations
36 in normal adult subjects administered abuse doses of aspartame. J Toxicol Environ Health 7:
37 281-290.
7-31 DRAFT - DO NOT CITE OR QUOTE
-------
1 Stegink, LD; Brummel, MC; Filer, LJ; et al. (1983) Blood methanol concentrations in
2 one-year-old infants administered graded doses of aspartame. J Nutr 113:1600-1606.
3
4 Stegink, LD; Filer, LJ Jr; Bell, EF; et al. (1989) Effect of repeated ingestion of aspartame-
5 sweetened beverage on plasma amino acid, blood methanol, and blood formate concentrations in
6 normal adults. Metabolism 38(4):357-63.
7
8 Steiner, EC; Rey, TD; McCroskey, PS. (1990) Reference guide for Sumusolv. Dow
9 Chemical Co., Midland, MI.
10
11 Stern, S; Reuhl, K; Soderholm, S; et al. (1996) Perinatal methanol exposure in the rat I:
12 Blood methanol concentration and neural cell adhesion molecules. Fundam Appl Toxicol 34:
13 36-46.
14
15 Sultatos, LG; Pastino, GM; Rosenfeld, CA; et al. (2004) Incorporation of the genetic
16 control of alcohol dehydrogenase into a physiologically based pharmacokinetic model for
17 ethanol in humans. Toxicol Sci. 2004 Mar; 78(1):20-31.
18
19 Svensson, S; Some, M; Lundsj0, A; et al. (1999) Activities of human alcohol
20 dehydrogenase in the metabolic pathways of ethanol and serotonin. Eur J Biochem 262:324-329.
21
22 Tanner, CM. (1992) Occupational and environmental causes of Parkinsonism. Occup
23 Med 7:503-513.
24
25 Teeguarden, JG; Deisinger, PJ; Poet, TS; et al. (2005) Derivation of a human equivalent
26 concentration for n-butanol using a physiologically based pharmacokinetic model for n-butyl
27 acetate and metabolites n-butanol and n-butyric acid. Toxicol Sci 85:429-446.
28
29 Teng, S; Beard, K; Pourahmad, J; et al. (2001) The formaldehyde metabolic
30 detoxification enzyme systems and molecular cytotoxic mechanism in isolated rat hepatocytes.
31 Chem Biol Interact 130-132:285-296.
32
33 Tephly, TR; McMartin, KE. (1984) Methanol metabolism and toxicity. Food Sci Technol
34 12:111-140.
35
36 Tiboni, GM; Giampietro, F; Di Giulio, C. (2003) The nitro oxide synthesis N-omega-
37 nitro-L-arginine methyl ester (L-NAME) causes limb defects in mouse fetuses: protective effect
38 of acute hyperoxia. Pediatr Res 54:69-76.
7-32 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Til, HP; Woutersen, RA; Feron, VJ; et al. (1989) Two-year drinking water study of
3 formaldehyde in rats. Food Chem Toxicol 27:77-87'.
4
5 Thomas, DG; Breslow, N; Gart, JJ. (1997) Trend and Homogeneity Analyses of
6 Proportions and Life Table Data. Version 2.1. Computers and Biomedical Research 10:373-381.
7
8 Thrasher, JD; Kilburn, KH. (2001) Embryo toxicity and teratogenicity of formaldehyde.
9 Arch Environ Health. 2001 Jul-Aug;56(4):300-ll.
10
11 Tobe, M; Naito, K; Kurokawa, Y. (1989) Chronic toxicity study on formaldehyde
12 administered orally to rats. Toxicology 56:79-86.
13
14 Toth, BA; Wallcave, L; Patil, K; Schmeltz, I; and Hoffmann, D. (1977) Induction of
15 tumors in mice with the herbicide succinic acid 2,2-dimethylhydrazide. Cancer Research
16 37:3497-3500.
17
18 Tran, MN; Wu, AH; Hill, DW. (2007) Alcohol dehydrogenase and CAT content in
19 perinatal infant and adult livers: potential influence on neonatal alcohol metabolism. Toxicol
20 Lett. 30;169(3):245-52.
21
22 Trump, BF; McDowell, EM; Harris, CC. (1984) Chemical carcinogenesis in the
23 tracheobronchial epithelium. Environ Health Perspect 55:77-84.
24
25 Tsakiris, T; Giannoulia-Karantana, A; Simintzi, I; et al. (2006) The effect of aspartame
26 metabolites on human erythrocyte membrane acetylcholinesterase activity. Pharmacol Res
27 53:1-5.
28
29 Tsutsui, M; Shimokawa, H; Morishita, T; et al. (2006) Development of genetically
30 engineered mice lacking all three nitric oxide synthases. J Pharmacol Sci 102:147-154.
31
32 Turner, C; Spanel, P; and Smith, D. (2006) A longitudinal study of methanol in the
33 exhaled breath of 30 healthy volunteers using selected ion flow tube mass spectrometry, SIFT-
34 MS Physiol Maes. 27:637-648.
35
36 U.S. EPA (Environmental Protection Agency). (1986a) Guidelines for the health risk
37 assessment of chemical mixtures. Federal Register 51(185):34014-34025. Available from:
38 http://www.epa.gov/iris/backgr-d.htm.
7-33 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 U.S. EPA (Environmental Protection Agency). (1986b) Guidelines for mutagenicity risk
3 assessment. Federal Register 51(185):34006-34012. Available from:
4 http://www.epa.gov/iris/backgr-d.htm.
5
6 U.S. EPA (Environmental Protection Agency). (1986c) Rat oral subchronic toxicity study
7 with methanol. Conducted by Toxicity Research Laboratories, Ltd. TRL No. 032-005.
8
9 U.S. EPA (Environmental Protection Agency). (1988) Recommendations for and
10 documentation of biological values for use in risk assessment. Prepared by the Environmental
11 Criteria and Assessment Office, Office of Health and Environmental Assessment, Cincinnati, OH
12 for the Office of Solid Waste and Emergency Response, Washington, DC; EPA 600/6-87/008.
13 Available from: http://www.epa.gov/iris/backgr-d.htm.
14
15 U.S. EPA (Environmental Protection Agency). (1991) Guidelines for developmental
16 toxicity risk assessment. Federal Register 56(234):63798-63826. Available from:
17 http://www.epa.gov/iris/backgr-d.htm.
18
19 U.S. EPA (Environmental Protection Agency). (1994a) Interim policy for particle size
20 and limit concentration issues in inhalation toxicity studies. Federal Register 59(206):53799.
21 Available from: http://www.epa.gov/iris/backgr-d.htm.
22
23 U.S. EPA (Environmental Protection Agency). (1994b) Methods for derivation of
24 inhalation reference concentrations and application of inhalation dosimetry. Office of Research
25 and Development, Washington, DC; EPA/600/8-90/066F. Available from:
26 http://www.epa.gov/iris/backgr-d.htm.
27
28 U.S. EPA (Environmental Protection Agency). (1995) Use of the benchmark dose
29 approach in health risk assessment. Risk Assessment Forum, Washington, DC; EPA/630/R-
30 94/007. Available from: http://cfpub.epa.gov/ncea/raf/recordisplay.cfm?deid=42601.
31
32 U.S. EPA (Environmental Protection Agency). (1996) Guidelines for reproductive
33 toxicity risk assessment. Federal Register 61(212):56274-56322. Available from:
34 http://www.epa.gov/iris/backgr-d.htm.
35
36 U.S. EPA (Environmental Protection Agency). (1997) Guidance on cumulative risk
37 assessment, Part 1 Planning and Scoping. Science Policy Council, Washington, D.C.
38
7-34 DRAFT - DO NOT CITE OR QUOTE
-------
1 U.S. EPA (Environmental Protection Agency). (1998) Guidelines for neurotoxicity risk
2 assessment. Federal Register 63(93):26926-26954. Available from:
3 http://www.epa.gov/iris/backgr-d.htm.
4
5 U.S. EPA (Environmental Protection Agency). (2000a) Science policy council handbook:
6 risk characterization. Office of Science Policy, Office of Research and Development,
7 Washington, DC; EPA 100-B-00-002. Available from: http://www.epa.gov/iris/backgr-d.htm.
8
9 U.S. EPA (Environmental Protection Agency). (2000b) Benchmark dose technical
10 guidance document [external review draft]. Risk Assessment Forum, Washington, DC;
11 EPA/630/R-OO/OO1. Available from: http://www.epa.gov/iris/backgr-d.htm.
12
13 U.S. EPA (Environmental Protection Agency). (2000c) Supplementary guidance for
14 conducting for health risk assessment of chemical mixtures. Risk Assessment Forum,
15 Washington, DC; EPA/63O/R-00/002. Available from: http://www.epa.gov/iris/backgr-d.htm.
16
17 U.S. EPA (Environmental Protection Agency) (2000d) Toxicological review of vinyl
18 chloride. EPA.635R-00/004.
19
20 U.S. EPA (Environmental Protection Agency). (2002a) A review of the reference dose
21 and reference concentration processes. Risk Assessment Forum, Washington, DC; EPA/630/P-
22 02/0002F. Available from: http://www.epa.gov/iris/backgr-d.htm.
23
24 U.S. EPA (Environmental Protection Agency). (2002b) Health assessment of 1,3-
25 butadiene. EPA/600/P-98/001F.
26
27 U.S. EPA (Enviornmental Protection Agency) (2004). Integrated risk information system:
28 vinyl chloride. Available from www.epa.gov/iris/subst/1001 .htm.
29
30 U.S. EPA (Environmental Protection Agency). (2005a) Guidelines for carcinogen risk
31 assessment. Risk Assessment Forum, Washington, DC; EPA/630/P-03/001B. Available from:
32 http://www.epa.gov/iris/backgr-d.htm.
33
34 U.S. EPA (Environmental Protection Agency). (2005b) Supplemental guidance for
35 assessing susceptibility from early-life exposure to carcinogens. Risk Assessment Forum,
36 Washington, DC; EPA/630/R-03/003F. Available from: http://www.epa.gov/iris/backgr-d.htm.
37
7-3 5 DRAFT - DO NOT CITE OR QUOTE
-------
1 U.S. EPA (Environmental Protection Agency). (2006a) Science policy council handbook:
2 peer review. Third edition. Office of Science Policy, Office of Research and Development,
3 Washington, DC; EPA/1 OO/B-06/002. Available from: http://www.epa.gov/iris/backgr-d.htm.
4
5 U. S. EPA (Environmental Protection Agency). (2006b) A Framework for Assessing
6 Health Risk of Environmental Exposures to Children. National Center for Environmental
7 Assessment, Washington, DC, EPA/600/R-05/093E Available from:
8 http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid= 158363.
9
10 U.S. EPA (Environmental Protection Agency). (2006c) Toxic Release Inventory (TRI)
11 On-site and Off-site Reported Disposed of or Otherwise Released (in pounds), for facilities in All
12 Industries, for All Chemicals, U.S., 2003. Available from: http://www.epa.gov/tri. Washington,
13 D.C.
14
15 Vara-Castrodeza, A; Perez-Castrillon, JL; Duefias-Laita, A. (2007) Magnetic resonance
16 imaging in methanol poisoning. Clin Toxicol 45:429-430.
17
18 Ward, KW; Pollack, GM. (1996) Comparative toxicokinetics of methanol in pregnant and
19 NP rodents. Drug Metab Dispos 24:1062-1070.
20
21 Ward, KW; Perkins, RA; Kawagoe, et al. (1995) Comparative toxicokinetics of methanol
22 in the female mouse and rat. Fundam Appl Toxicol 26:258-264.
23
24 Ward, KW; Blumenthal, GM; Welsch, F; et al. (1997) Development of a physiologically
25 based pharmacokinetic model to describe the disposition of methanol in pregnant rats and mice.
26 Toxicol Appl Pharmacol 145:311-322.
27
28 Weiss, B; Stern, S; Soderholm, SC; et al. (1996) Developmental neurotoxicity of
29 methanol exposure by inhalation in rats. HEI Research Report Number 73.
30
31 Wentzel, P; Rydberg, U; Eriksson, UJ. (2006) Antioxidative treatment diminishes
32 ethanol-induced congenital malformations in the rat. Alcohol Clin Exp Res. 30(10): 1752-60.
33
34 Wentzel, P; Eriksson, UJ. (2006) Ethanol-induced fetal dysmorphogenesis in the mouse
35 is diminished by high antioxidative capacity of the mother. Toxicol Sci 2(2):416-22.
36
37 White, LR; Marthinsen, ABL; Richards, RJ; et al. (1983) Biochemical and cytological
3 8 studies of rat lung after inhalation of methanol vapour. Toxicol Lett 17:1-5.
7-36 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 Wilson, SL; Cradock, MM. (2004) Review: Accounting for prematurity in developmental
3 assessment and the use of age-adjusted scores. J Pediatr Pschol 29:641-649.
4
5 Wu, AH; Kelly, T; McKay, C; et al. (1995) Definitive identification of an exceptionally
6 high methanol concentration in an intoxication of a surviving infant: methanol metabolism by
7 first-order elimination kinetics. J Forensic Sci 40(2):315-320.
8
9 Yuan J. (1993) Modeling blood/plasma concentrations in dosed feed and dosed drinking
10 water toxicology studies. Toxicol Appl Pharmacol. Mar; 119(1):131-41. PubMed PMID:
11 8470117.
12
13 Zorzano, A; Herrera, E. (1989) Disposition of ethanol and acetaldehyde in late pregnant
14 rats and their fetuses. Pediatr Res 25(1): 102-6.
7-37 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION
[Page intentionally left blank]
A-1 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX B. DEVELOPMENT, CALIBRATION AND APPLICATION OF A
METHANOL PBPK MODEL
1 This appendix is adapted from a report prepared for the U. S. EPA under contract by the
2 Center for Biological Monitoring and Modeling, Battelle Northwest Laboratories (WA 09,
3 Battelle Project No. 48746, January 31, 2007).
B.I. SUMMARY
4 This appendix describes the development, calibration, and approach for application of
5 mouse, rat, and human PBPK models to extrapolate mouse and rat methanol inhalation-route
6 internal dose metrics to human inhalation exposure concentrations that result in the same internal
7 dose (HEC). The human oral methanol dose(s) yielding internal dose(s) equivalent to the mouse
8 or rat internal dose at the (HED) is also presented.
9 A PBPK model was developed to describe the blood kinetics of methanol (MeOH) in
10 mice and humans. The model includes compartments for lung/blood MeOH exchange, liver, fat,
11 and the rest of the body. To describe blood MeOH kinetics, the model employs two saturable
12 descriptions of MeOH metabolic clearance in mice and rats, and one first-order metabolic
13 clearance, and a first-order description of renal clearance (from blood) in humans. Renal
14 clearance is a minor pathway and does not appreciably affect MeOH blood kinetics.
15 This model is a revision of the model reported by Ward et al. (1997), reflecting
16 significant simplifications (removal of compartments for placentae, embryo/fetus, and
17 extraembrionic fluid) and two elaborations (addition of an intestine lumen compartment to the
18 existing stomach lumen compartment and addition of a bladder compartment which impacts
19 simulations for human urinary excretion.), while maintaining the ability to describe MeOH blood
20 kinetics. The model reported here uses a single consistent set of parameters; the Ward et al.
21 model employed a number of data-set specific parameters. Other biokinetic MeOH models that
22 were considered as starting points for the current model also used varied parameters by dataset to
23 achieve model fits to the data. For example, the model of Bouchard et al. (2001) used different
24 respiratory rates and fractional inhalation absorbed for different human exposures.
25 The mouse model was calibrated against inhalation-route blood MeOH kinetic data and
26 verified using intravenous-route blood MeOH kinetic data. The rat model was calibrated against
27 low-dose intravenous data and validated with inhalation-route data. The human model was
28 calibrated against inhalation-route MeOH kinetic data. The models accurately described the
29 inhalation route pharmacokinetics of MeOH. Mouse model simulations of oral- and i.v.-route
30 kinetics compare well to some but not all the experimental data.
31 The MeOH HECs predicted by the model (based on 1,000 ppm inhalation exposure in
32 mice) were >1,000 ppm using either blood AUC or Cmax as the dose metrics. The MeOH HED
B-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 derived by cross-route extrapolation of this inhalation-route HEC was 110 mg/kg-day, based on
2 MeOH blood AUC following zero order uptake of MeOH (a constant rate of delivery). Because
3 of the lack of human data from high-dose exposures, it was not possible to calibrate the model
4 for inhalation exposures above 1,000 ppm or oral exposures above 110 mg/kg-day. However,
5 because the BMD approach was used to estimate an internal experimental animal dose and UFs
6 were applied to the internal dose, the human model can be used to back-calculate that internal
7 dose to an RfC and an RfD below 1,000 ppm or 110 mg/kg-day, respectively.
B.2. MODEL DEVELOPMENT
B.2.1. Model Structure
8 This model is a revision of the model reported by Ward et al. (1997), reflecting
9 significant simplifications and two elaborations, while maintaining the ability to describe MeOH
10 blood kinetics in mice, rats, and humans (Figure B-l). The kidney, pregnancy and the fetal
11 compartment have been removed. The kidney was lumped with the body compartment because
12 the blood:tissue partition coefficients for these tissues were similar. The elaborate time-
13 dependent descriptions of pregnancy were removed because analysis of the available
14 pharmacokinetic data indicates that blood MeOH kinetics in NP and pregnant mice are not
15 different enough to warrant separate descriptions. Because the maternal blood:fetal blood
16 partition coefficients were near 1, there was no need to explicitly model fetal kinetics; they will
17 be equivalent to maternal blood kinetics. Further supporting data exist for ethanol, which is
18 quite similar to MeOH in its partitioning and transport properties. In rats (Zorzano and Herrara,
19 1989; Guerri and Sanchis 1985), sheep (Brien et al., 1985; Cumming et al., 1984), and guinea
20 pigs (Clarke et al., 1986), fetal and maternal blood concentrations of ethanol are virtually
21 superimposable; maternal to fetal blood ratios are very close to 1, including during late gestation.
22 Also, fetal brain concentrations in guinea pigs (Clarke et al., 1986) were also very similar to the
23 mothers'.
24 In addition to the absolute maternal-fetal concentration similarity noted above, it is
25 common practice to use blood concentrations as an appropriate metric for risk extrapolation via
26 PBPK modeling for effects in various tissues, based on the reasonable expectation that any
27 tissue:blood differences will be similar in both the test species and humans. For example, even if
28 the brain:blood ratio was around 1.2 in the mouse or rat, the similar biochemical make-up of
29 brain tissue and blood in rats and humans leads to the expectation that the brain:blood levels in
30 humans (which depend on the biochemical make-up) will also be close to 1.2, and so the relative
31 "error" that might occur by using blood instead of brain concentration in evaluating the dose-
32 response in rats will be cancelled out by using blood instead of brain concentration in the human.
33 The fact that measured fetal blood levels are virtually identical to maternal levels for methanol
34 (and ethanol) tells us that the rate of metabolism in the fetus is not sufficient to significantly
B-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 reduce the fetal concentration versus maternal, and use of a PBPK model to predict maternal
2 levels will give a better estimate of fetal exposure than use of the applied dose or exposure,
3 because there are animal-human differences in adult PK of MeOH for which the model accounts,
4 based on PK data from humans as well as rodents.
Inhalation
Exposure
V.
Bladder
KB [1 (human
Urine |>
KLLor
Vma
Km.
I/
JlataKrklita
only)
X-
Vmax
Km2
c
Venous Blood
f
-^
±=^-
2
Kfec^J
Alveolar Air
Lung Blood
Body
Fat
Liver
^Exha
Arterial Blood
tied Air
TKAI \KASorVmAS.KMAS
ntestine^Xr^
-------
1 fit to data from inhalation and IV exposure, a small rate of elimination from the intestine (lumen)
2 compartment to feces. (The mouse data could be adequately fit with this rate set to zero,
3 corresponding to 100% absorption; humans were assumed to have zero fecal elimination, like the
4 mouse.) The final models thus include compartments for fat, liver, the rest of the body, bladder
5 (only used for humans), and lung. The mouse and rat models describe inhalation, oral, and
6 intravenous route dosing and the human model describes inhalation and oral route dosing and the
7 rat model includes a non-zero rate of fecal elimination. Although there is an endogenous
8 background level of both MeOH and formate (See Section 3.3), the model does not explicitly
9 describe or account for background levels of MeOH or formate. In this analysis, when non-zero
10 background levels have been measured (in blood), that background was simply subtracted from
11 the concentrations measured during exposure. However, a zero-order rate of infusion could be
12 added to the liver, blood, or stomach compartments to mimic background levels if that was
13 considered necessary.
14 MeOH is well absorbed by the inhalation and oral routes, and is readily metabolized to
15 formaldehyde, which is rapidly converted to formate in both rodents and humans. Although the
16 enzymes responsible for metabolizing formaldehyde are different in rodents (CAT) and humans
17 (ALD) the metabolite, formate, is the same, and the metabolic rates are similar (Clary, 2003).
18 Most of the published rodent kinetic models for MeOH describe the metabolism of MeOH to
19 formaldehyde as a saturable process but differ in the handling of formate metabolism and
20 excretion (Bouchard et al., 2001; Fisher et al., 2000; Ward et al., 1997; Horton et al., 1992).
21 Ward et al. (1997) used one saturable and one first-order pathway for mice, and Horton et al.
22 (1992) applied two saturable pathways of metabolism to describe MeOH elimination in rats.
23 Bouchard et al. (2001) employed one metabolic pathway and a second pathway described as
24 urinary elimination in rats and humans. The need for two saturable metabolic pathways in the
25 mouse model was confirmed through simulation and optimization. High exposure (>2,000 ppm
26 MeOH) and low exposure (1,000 ppm MeOH) blood data could not be adequately fit either
27 visually or by more formal optimization without the second saturable metabolic pathway. The
28 optimization approach and results are found below and in the Additional Materials at the end of
29 this appendix.
30 While the PPK model explicitly describes the concentration of methanol, it only
31 describes the rate of metabolism or conversion of MeOH to its metabolites. Distribution and
32 metabolism of formaldehyde is not considered by the model, and this model tracks neither
33 formate nor formaldehyde. (The data that would be needed to parameterize or validate a specific
34 description of either of these metabolites is not available). Since the metabolic conversion of
35 formaldehyde to formate is rapid (< 1 minute) in all species (Kavet and Nauss, 1990), the MeOH
36 clearance rate may approximate a formate production rate, though this has not been verified.
37 Thus the rate of MeOH metabolism predicted by the model can be used as a dose metric for
38 either or both of these metabolites, but scaling of that metabolic rate metric to humans would
B-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 require use of an inter-species scaling factor, (BWhUman/BWrodent) ' , to account for the general
2 expectation of slower clearance of the metabolites in humans.
3 The model was initially coded in acslXtreme vl .4 and updated in acslXtreme v 2.3
4 (Aegis Technologies, Huntsville, AL). Most procedures used to generate this report, except those
5 for the optimization, may be run by executing the corresponding .m files. The model code
6 (acslXtreme .csl file) and supporting .m files are available electronically and as text in the
7 Additional Materials at the end of this appendix. A key identifying .m files associated with
8 figures and tables in this report is also provided in the Additional Materials.
B.2.2. Model Parameters
9 Physiological parameters such as tissue volumes, blood flows, and ventilation rates were
10 obtained from the open literature (Table B-l). Parameters for blood flow, ventilation, and
11 metabolic capacity were scaled as a function of body weight raised to the 0.75 power, according
12 to the methods of Ramsey and Andersen (1984).
Table B-l. Parameters used in the mouse and human PBPK models
Body weight (kg)
Mouse
0.03"
Rat
SD | F344
0.2756
Human
70
Source
Measured/estimated
Tissue volume (% body weight)
Liver
Blood arterial
venous
Fat
Lung
Rest of body
5.5
1.23
3.68
7.0
0.73
72.9
3.7
1.85
4.43
7.0
0.50
73.9
2.6
1.98
5.93
21.4
0.8
58.3
Brown etal. 1997
Calculatedc
Flows (L/hr/kg0'75)
Alveolar ventillation''
Cardiac output
25.4
25.4
16.4
16.4
16.5
24.0
Perkins et al. 1995a; Brown et al. 1997;
U.S. EPA, 2004
Percentage of cardiac output
Liver
Fat
Rest of body
25.0
5.0
70.0
Biochemical constants8
VmaxC (mg/hr/kg075)
Km (mg/L)
Vmax2C (mg/hr/kg075)
Km2 (mg/L)
K1C (BW025/hr)
KLLC (BW025/hr/
19
5.2
3.2
660
NA
NA
25.0
7.0
68
5.0 0
6.3 NA
8.4 22.3
65 100
NA
NA
22.7
5.2
72.1
1s*
, saturable
order
NA 33.1
NA 23.7
NA
NA
0.0373 0.0342
95.7 NA
Brown etal. 1997
Calculated
Fitted
B-5
DRAFT - DO NOT CITE OR QUOTE
-------
Oral absorption
VmASC (mg/hr/kg075)
KMASC (mg/kg)
KSI (hr1)
KAI (hr"1)
Kfec (hr1)
1830
620
2.2
0.33
0
5570
620
7.4
0.051
0.029
377
620
3.17
3.28
0
Mouse and rat fitted (mouse and human
KMASC assumed = rat); other human
values are those for ethanol from Sultatos et
al. (2004), with VmASC set so that for a
70-kg person VmAS/KM = the first-order
constant of Sultatos et al.
Partition coefficients
Liver:Blood
Fat:Blood
Blood:Air
Body:Blood
Lung:Blood
Bladder time-constant
(KBL, hr-1/
Inhalation fractional
availability (%)
1.06
0.083
1350'
0.66
1
1.06
0.083
1350
0.66
1
NA
0.665
0.20
0.583*
0.142
1626
0.805
1.07
0.564 0.612
0.866*
Ward et al., 1997; Fiserova-Bergerova and
Diaz, 1986
Horton et al., 1992; Fiserova-Bergerova and
Diaz, 1986
Rodent: estimated; human: Fiserova-
Bergerova and Diaz, 1986 (human "body"
assumed = muscle)
Fitted (human)
Rodent: fitted; human
Ernstgard et al., 2005
1
2
3
4
5
6
7
NA - Not applicable for that species
"Both sources of mouse data report body weights of approximately 30 g
*The midpoints of rat weights reported for each study was used and ranged from 0.22 to 0.33 kg
The volume of the other tissues was subtracted from 91% (whole body minus a bone volume of approximately 9%) to get
the volume of the remaining tissues
''Minute ventilation was measured and reported for much of the data from Perkins et al. (1996) and the average alveolar
ventilation (estimated as 2/3 minute ventilation) for each exposure concentration was used in the model. When ventilation
rates were not available, a mouse QPC (Alveolar Ventilation/BW°75) of 25.4 was used (average from Perkins et al., 1995a).
The QPC used to fit the human data was obtained from U.S. EPA (2004). This QPC was somewhat higher than calculated
from Brown etal. (1997) (-13 L/hr/kg075)
eVmax, Km, and Vmax2, Km2 represent the two saturable metabolic clearance processes assumed to occur solely in the liver.
The Vmax used in the model = VmaxC (mg/kg°75-hr) XBW075. K1C is the first-order loss from the blood for human
simulations that represents urinary elimination. Allometric scaling for first-order clearance processes was done as previously
described (Teeguarden et al., 2005); The Kl used in the model= K1C / BW°25
fKLLC - alternate human first-order metabolism rate (used only when VmaxC = Vmax2C = 0)
sHuman oral simulations used a zero order dose rate equal to the mg/kg.day dose
*Human liverblood estimated from correlation to (measured) fatblood, based on data from 28 other solvents
7Rat partition coefficient used for mice as done by Ward et al. (1997)
JKBL - a first-order rate constant for clearance from the bladder compartment, used to account for the difference between
blood kinetics and urinary excretion data as observed in humans
*For human exposures, the fractional availability was from Sedivec et al. (1981), corrected for the fact that alveolar
ventilation is 2/3 of total respiration rate
Mouse model partition coefficients were used as reported (liver, fat, blood:air) or
estimated (lung, body). The mouse body compartment partition coefficient was set
approximately equal to the measured value for muscle (Ward et al., 1997). The mouse lung
partition coefficient was assumed to be 1.0, similar to the liver partition coefficient. This
parameter has no numerically significant impact on modeled blood dose metrics.
Human partition coefficients were reported by Horton et al. (1992), but were in fact
measured in rat tissues. The reported rat fat partition coefficient was considerably closer to unity
than reported for MeOH or ethanol by other researchers (Ward et al., 1997; Pastino and Conolly,
B-6
DRAFT - DO NOT CITE OR QUOTE
-------
1 2000) and assumed to be in error. Human partition coefficients were obtained from Fiserova-
2 Bergerova and Diaz (1986).
B.2.3. Mouse Model Calibration
B. 2.3.1. Inhalation-Route Calibration
3 For purposes of conducting interspecies extrapolations of MeOH dosimetry, the
4 inhalation route was the most important route requiring calibration for the mouse model. The
5 critical endpoint and NOEL, which are the basis for the HEC estimation, are from inhalation-
6 route studies. The ability to predict blood MeOH concentrations from inhalation exposures was
7 therefore a priority. Pharmacokinetic data from other routes, i.v. and oral, were used to verify
8 clearance terms derived by fitting to the inhalation data or to estimate a MeOH oral uptake rate
9 constants. Holding other parameters constant, the mouse PBPK model was calibrated against
10 inhalation-route blood pharmacokinetic data (Figure B-2) by fitting five parameters: Michaelis-
11 Menten constants for one high affinity/low capacity and one low-affinity high-capacity enzyme
12 and the inhalation fractional availability term.
10,000
1,000
I
o
Oj
100
10
— 1000 pprn sim
— 2000 pprn sim
— 5000 ppm sim
— 7500 ppm sim
— 10000 ppm sim
15000 pprn sim
A 1000 pprn data (GD6 & 10)
• 2000 ppm data (GD6 &. 10)
• 5000 ppm data (GD6 & 10)
+ 7500 ppm data (GD6 & 10)
A 10000 pprn data (GD6 & 10)
V 10000 ppm data (GD7)
n 15000 ppm data (GD6 & 10)
12 15
Time (hr)
18
21
24
27
13
14
15
Figure B-2. Model fits to data sets from GD6, GD7, and GD10 mice for
7-hour inhalation exposures to 1,000-15,000 ppm MeOH. Maximum
concentrations are from Table 2 in Rogers et al. (1993). The complete data set for
GD7 mice exposed to 10,000 ppm is from Rogers et al. (1997) and personal
communication (Additional Materials). Symbols are concentration means of a
minimum of n = 4 mice/concentration. Default ventilation rates (Table B-l) were
used to simulate these data.
For these mouse simulations, pulmonary ventilation was set to 25.4 (L/hr/kg0'75), the
average value measured by Perkins et al. (1995a), which is similar to the value of 29 (L/hr/kg0'7
reported in Brown et al. (1997). Where ventilation rates were reported for individual exposure
B-7
DRAFT - DO NOT CITE OR QUOTE
-------
1 concentrations by Perkins et al. (1995a), they were used directly in the model and a notation was
2 made in the figure legend. Reported ventilation rates varied from 592 to 857 L/kg x 8 hr,
3 depending on exposure concentration (Perkins et al., 1995a). Adjusting these values to 2/3 total
4 (for alveolar ventilation) and allometrically scaling by BW°'75, values used in the model for these
5 exposures ranged from 20.5 to 29.7 (L/hr/kg0'75) (See Table B-l). A fractional availability of
6 73% of alveolar ventilation was visually optimized to best describe the inhalation-route blood
7 MeOH pharmacokinetic data. This percentage of uptake for inhalation exposures is similar to
8 values reported for other alcohols in rodents (Teeguarden et al., 2005), but considerably lower
9 than the value reported by Perkins et al. (1995a) of 126% of alveolar ventilation (85% of total
10 ventilation).
11 The calibrated model predicted blood MeOH concentration time-course agreed well with
12 measured values in adult mice in the inhalation studies of Rogers et al. (1997, 1993) (Figure B-
13 2), and Perkins et al. (1995a), as well as in NP and early gestation (GD8) mice of Dorman et al.
14 (1995) (Figure B-3). Parameter values used in the calibrated model are given in Table B-l.
8,000-
7,000-
J 6,000-
£ 5,000-
O 4,000-
4)
3,000-
•o
0
5 2'°°°-
1,000-
0-
+ 2,500 ppm NP mouse
A 5,000 ppm NP mouse
o 10,000 ppm NP mouse
D 6 hr, 15,000 ppm GD 8 mouse (Dorman) J
0
15 20
Time (hr)
30
35
Figure B-3. Simulation of inhalation exposures to MeOH in NP mice from Perkins
15 et al. (1995a) (8-hour exposures) and Dorman et al. (1995), (6-hour exposures). Data
16 points represent measured blood MeOH concentrations and lines represent PBPK model
17 simulations. Note: data was obtained using Digitizlt (Sharlt! Inc. Greensburg, PA) to
18 digitize data from Figure 2 of Perkins et al. (1995a) and Figure B-2 from Dorman et al.,
19 (1995). Default ventilation rates (Table B-l) were used to simulate the Dorman data. The
20 alveolar ventilation rate for each data set from Perkins et al. (1995) was set equal to the
21 measured value reported in that manuscript. For the 2,500, 5,000, and 10,000 ppm exposure
22 groups, the alveolar ventilation rates were 29, 24, and 21 (L/hr/kgO.75), respectively. The
23 cardiac output for these simulations was set equal to the alveolar ventilation rate.
DRAFT - DO NOT CITE OR QUOTE
-------
B. 2.3.1. (should be B. 2.3.2) Oral-Route Calibration
1 The mouse model was calibrated for the oral route by fitting the rate constants for oral
2 uptake of MeOH. Calibration of the oral route was not required for interpretation of the critical
3 toxicology studies. This exercise was undertaken to estimate the rate constants for oral uptake so
4 it could be used to make dose-route extrapolations for calculating human oral-route exposures
5 equivalent to mouse exposures at the NOEL.
6 Ward et al. (1997) described MeOH uptake as the sum of a fast and slow process (two
7 rate constants), with a fraction of the administered dose attributed to each process. The rate
8 constants and the fraction of the dose attributed to each process were varied to describe oral-
9 route blood MeOH kinetics for each GD. For instance, the fraction of the total oral dose
10 assigned to the fast absorption process varied from 54 to 71%, depending on the data set. An
11 alternative approach with uptake attributed to stomach and intestine, which allows for greater
12 flexibility in fitting the data (Staats et al., 1991), was compared to a simpler one utilizing a single
13 rate of uptake. In both the current model and the model of Ward et al. (1997), orally ingested
14 MeOH was assumed to be 100 % absorbed.
15 Initially, a single oral absorption rate constant (KAS, hr"1) was fitted to oral-route blood
16 MeOH kinetics reported by Ward et al. (1997, 1995). Using these data, an average KAS
17 (0.62 hr"1) was estimated that provides adequate fits to MeOH blood kinetics following 2,500
18 mg/kg dose in NP and GDIS mice and 1,500 mg/kg in GD8 mice up to ~8 hours. At later time
19 points, however, a model using a single oral uptake rate constant consistently under predicts
20 blood concentrations of MeOH (results not shown). Fits were improved by using the two
21 compartment GI tract model (Figure B-4). However, when fitting the oral data in rats, it was
22 found that the fits were significantly improved if the uptake from the stomach was treated as a
23 saturable process. Vmax (VMASC) was scaled as BW°'75, as is done for other Vmaxs, and the Km
24 (KMASC) was scaled as BW1 to reflect that the variable used is the total amount in the stomach,
25 whose volume is expected to scale with BW1. For the mouse, model fits were not significantly
26 improved when KMASC was allowed to vary (change from the value fitted to rats, 1830 mg/kg),
27 so it was kept at the rat value.
28 Using the two-compartment oral absorption model and adjusting only the absorption
29 parameters resulted in a good fit to the lower oral dose (1,500 mg/kg) (Dorman et al., 1995), but
30 consistently under-prediction of the 2,500 mg/kg oral dosing blood levels (Ward et al., 1997).
31 When the metabolic constants (VmaxC values) were decreased, the data from the higher dose
32 were fit, but the fit of the data for the 1,500 mg/kg dose was lost (see Additional Materials,
33 Figure B-19). Also, when using the lower clearance required to fit the data of Ward et al. (1997),
34 the inhalation data of Rogers et al. (1993) could no longer be fit by the model (see Additional
35 Materials, Figure B-20). The two-compartment GI tract approach (with parameters that better fit
36 the low dose data) was retained in the model and used for all final mouse oral route simulations.
37
B-9 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
o
O
m
3,500-
3,000-
2,500-
2,000-
1,500-
o 1,000-
OD
500-
0-1
• Oral 1500 mg/kg GD 8 mouse
• Oral 2500 mg/kg NP mouse
n Oral 2500 mg/kg GD IS mouse
— 1500 mg/kg simulation
— 2500 mg/kg simulation
10 15
Time (hr)
20
25
Figure B-4. Oral exposures to MeOH in pregnant mice on GD8 (Dorman et al., 1995)
or NP and GDIS. Data points represent measured blood concentrations and lines represent
PBPK model estimations for NP mice.
Source: Ward et al., (1997).
B. 2.3.1. (B2.3.3) Intravenous Route Simulation
The parameterization of MeOH clearance (high-and-low affinity metabolic pathways)
was verified by simulation of data sets describing the intravenous-route pharmacokinetics of
MeOH. MeOH blood kinetics data in NP mice are only available for a single i.v. dose of
2,500 mg/kg (Ward et al., 1997). MeOH blood kinetics are also reported in GDIS mice
following administration of a broader range of doses: 100, 500, and 2,500 mg/kg. Because
MeOH kinetics appear similar for NP and pregnant mice after administration of 2,500 mg/kg
prior to 20 hours, the model is expected to fit data for both pregnant and NP mice using the same
set of parameters, and hence, data for both life stages were used to verify metabolic clearance of
MeOH.
Initial blood concentrations of MeOH following i.v. administration were not proportional
to administered dose in the data from Ward et al. (1997). Initial concentrations were
approximately 1.5-fold lower in the 100 mg/kg dose group than expected if a dose-independent
volume of distribution (VD) is assumed (Figure B-5A). Initial blood concentrations were,
however, proportional to administered dose between 2,500 and 500 mg/kg. Two possible
explanations were then considered:
1) the VD, which is not impacted by any other PBPK parameters and is only determined by
the biochemical partitioning properties of MeOH, is twofold lower at 100 mg/kg than at
the higher concentrations, while the VD at 500 and 2,500 mg/kg are exactly as predicted
by the PBPK model without adjustment; or
B-10
DRAFT - DO NOT CITE OR QUOTE
-------
1 2) a dilution error occurred during preparation of the 100 mg/kg dosing solution used by
2 Ward etal. (1997).
10,000
~ 1,000
_l
~~-
01
E
o
Oj
•o
o
0
100
10
1,000
100
O
a
o 10
- IV 2500 simulation
- IV 500 simulation
— IV 100 simulation
• IV 2500 NP mouse
a IV 2500 GD 8 mouse
D IV 2500 GD IS mouse
i IV 500 GD 18 mouse
• IV 100 GD 18 mouse
10
15
Time (hr)
A
20
'-— IV 200 simulation }
— IV 100 simulation
IV 200 GD 18 mouse I
25
B
0.5
1,5
2 2,5
Time (hr)
3.5
4.5
3
4
5
6
7
8
9
10
11
Figure B-5. Mouse intravenous route MeOH blood kinetics. A) MeOH was
infused over 1.5 minutes into female CD-I mice at target doses of 100 (circles),
500 (triangles) or 2,500 (squares) mg/kg. Mice were NP, GD9 or GDI8 at the
time of dosing. Data points represent measured blood concentrations and lines
represent PBPK model simulations. B) Comparison of the 100 mg/kg dose data
(points) and PBPK model simulations assuming a 100 mg/kg dose as reported
(solid line) or to a presumed 200 mg/kg dose (dashed lines). Note: the 24-hour
time point data from the 500 mg/kg NP and GD 9 mice are below the reported
detection limit (2 |ig/ml) and so are not shown.
Source: Adapted from Ward et al. (1997).
3 To account for this unexpected nonproportionality, Ward et al. (1997) used higher
4 partition coefficients for placenta and embryonic fluid and lower Vmax for the metabolism of
B-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 MeOH for the 100 and 500 mg/kg doses than for the 2,500 mg/kg dose. These adjustments to
2 partition coefficients effectively change the volume of distribution. However, the PBPK model
3 obtained with measured partition coefficients and otherwise calibrated to inhalation data, as
4 described above, was capable of simulating both the 500 and 2,500 mg/kg data without adjust-
5 ment or varying parameters between those 2 doses. Further, the data at the nominal dose of 100
6 mg/kg could also be adequately fit without other parameter adjustment simply by simulating a
7 dose of 200 mg/kg (dotted line, Figure B-5B). If the relative difference in the VD at 100 mg/kg
8 was not a round number and/or the apparent value at 500 mg/kg was intermediate between 100
9 and 2,500 mg/kg, the possibility of a dose-related variation in this or other parameters would be
10 given more weight. Also it seems unlikely that the volumes of distribution would be different in
11 these animals solely based on differing exposure concentrations. But the far simpler and more
12 likely explanation for the observed pattern of VD seems to be a dosing error (2). Therefore, a
13 single set of parameters was retained which describe the 2,500 and 500 mg/kg doses rather than
14 adjust parameters to fit a data set (the 100 mg/kg dose group) that appeared inconsistent and may
15 be the result of a simple experimental error rather than attributable to a dose dependence.
16 Thus, high- (2,500 mg/kg) and mid-dose (500 mg/kg) intravenous-route pharmacokinetic
17 data were used to validate the parameters calibrated from the inhalation studies for the metabolic
18 clearance of MeOH. Metabolic constants reasonably predict blood MeOH kinetics following a
19 2,500 mg/kg dose in NP animals until 12 hours postexposure, but under predict blood MeOH in
20 GD9 and GDIS mice at 8 hours of exposure and beyond, and under-predict levels in both NP and
21 pregnant mice at 15 hours and beyond. At this high-dose, where blood kinetics of MeOH were
22 reported in NP, GD9, and GDIS mice, the data for the GDIS mice was inconsistent with the
23 GD9 and NP animals. The GD9 data at 12 hours appears inconsistent with the NP data, but then
24 the 2 are nearly identical again at 15 hours, so it is not clear if that difference at 12 hours is real
25 or just due to experimental variability. Blood levels of MeOH were -500 mg/L in GDIS mice at
26 24 hours, but were nondetectable after 18 hours in the other groups (detection limit 2 mg/L).
27 Blood concentrations were accurately predicted following administration of 500 mg/kg MeOH
28 (Figure B-5A). The model predictions did not match the 100 mg/kg data unless one assumed an
29 error in dose preparation, as described above (Figure B-5B). The calibration of the MeOH
30 PBPK model is consistent with both the available inhalation and oral-route data.
B.2.3.1. (B2.3.4)TotalMethanolMetabolic Clearance
31 Quantifying production of formaldehyde following MeOH exposure for use as an
32 alternative dose metric is of particular interest because formaldehyde is also undergoing toxicity
33 assessment. However, it is important to understand that because the model was developed to
34 describe blood MeOH kinetics, metabolism of MeOH to neither formaldehyde nor formate is
35 specifically described; the model tracks neither of these metabolites. While the metabolic
36 clearance of MeOH described by the model may be presumed to equate with formaldehyde
B-12 DRAFT - DO NOT CITE OR QUOTE
-------
1 production, this metabolic flux simply leaves the computational model system without specific
2 attribtution. Since the metabolic conversion of formaldehyde to formate is rapid in all species
3 (< 1 minute) (Kavet and Nauss, 1990), the MeOH clearance rate may approximate a formate
4 production as well as a formaldehyde production rate, though this has not been verified.
5 Thus, production of formaldehyde or formate following exposure to MeOH can only be
6 estimated by summing the total amount of MeOH cleared by metabolic processes. If used, this
7 metric of formaldehyde or formate dose should be adjusted by an inter-species scaling factor
8 (SF = [BWhuman/BWrodent]0'25) to adjust for expected species differences in the clearance of these
9 two metabolites (this is a default factor based on the generally accepted assumption that total
10 metabolism scales as BW°75 and hence clearance per BW scales as 1/BW0'25). The rate of
11 MeOH clearance may roughly be equated to the total amount of metabolites produced. Values of
12 total MeOH clearance as a function of exposure in mice and humans are presented in the
13 Additonal Materials (Tables 7-9).
B. 2.3.1. (B.2.3.5) Formal Optimization of Mouse Model Parameters
14 Formal optimization of five parameters (inhalation fractional availability and the Vmax
15 and Kms for high and low affinity MeOH metabolism) was attempted using optimization
16 routines in acslXtreme v2.01.1.2. Under the best circumstances, formal optimizations offer the
17 benefit of repeatability and confirmation that global optima have not been significantly missed
18 by user-guided visual optimization. Incorporating judgments regarding the value of specific data
19 sets, while possible when visually fitting, is more difficult when using optimization routines.
20 This is an important distinction between these approaches for this modeling exercise.
21 The mouse inhalation route NOEL was less than 1,000 ppm MeOH. The model is
22 calibrated against inhalation-route data because of the importance of this exposure route in the
23 assessment. Unfortunately, the vast majority of the MeOH data came from much higher
24 exposure concentrations. As expected, various attempts at formal optimization lead to improved
25 fits for some but never all data sets. This is to be expected when there is significant variability in
26 the underlying data. Various data-weighting schemes were included to improve overall
27 optimization while maintaining a good fit to the lowest concentration (1,000 ppm) data. In the
28 end, formal optimization provided no significant improvement over the fractional availability
29 and metabolic parameter values obtained by visual optimization, so these were retained in the
30 final version of the model.
31 Further details on the approach and results from the formal optimization are found in the
32 Additional Materials in outline format with supporting figures. More complete documentation
33 was not developed because the products of the optimizations were not used in the final model.
34 The documentation is intended only to demonstrate that appropriate optimizations were
35 conducted and what the results of those optimizations were.
B-13 DRAFT - DO NOT CITE OR QUOTE
-------
B.2.4. Mouse Model Sensitivity Analysis
1 An evaluation of the importance of selected parameters on mouse model estimates of
2 blood MeOH AUC was performed by conducting a sensitivity analysis using the subroutines
3 within acslXtreme. Files for reproducing the sensitivity analysis are available in the model as
4 described in the additional materials. The analysis was conducted by measuring the change in
5 model output corresponding to a 1% change in a given model parameter when all other
6 parameters were held fixed. A normalized sensitivity coefficient of 1 indicates that there is a
7 one-to-one relationship between the fractional change in the parameter and model output; values
8 close to zero indicate a small effect on model output. A positive value for the normalized
9 sensitivity coefficient indicates that the output and the corresponding model parameter are
10 directly related while a negative value indicates they are inversely related.
11 Sensitivity analyses were conducted for the inhalation and oral routes. The
12 inhalation-route analysis was conducted under the exposure conditions of Rogers and Mole
13 (1997) and Rogers et al. (1993), 7-hour inhalation exposures at the NOEL concentration of
14 1,000 ppm. The oral route sensitivity analysis was conducted for an oral dose of 1,000 mg/kg.
15 Parameters with sensitivity coefficients less than 0.1 are not reported. The parameters
16 with the largest sensitivity coefficients for the inhalation route at 1,000 ppm were pulmonary
17 ventilation, VmaxC, and partitioning to the body compartment (Figures B-6 [metabolism] and B-7
18 [flows and partition coefficients]). MeOH AUC was also sensitive to KM2 and VmaxC. The
19 sensitivity coefficient for pulmonary ventilation increases from 1 to -1.75 during the exposure
20 period as metabolism begins to saturate. The sensitivity coefficient is 1 for concentrations 100
21 ppm MeOH or less or when hepatic clearance is nonlimiting.
22 Oral-route mouse blood MeOH AUC was sensitive to the rate constants for uptake.
23 Blood AUC was most sensitive to the first-order rate constant for uptake from the stomach, KAS,
24 during the first hour after exposure, becoming less important over time (Figure B-8). Blood
25 MeOH AUC was also modestly sensitive to KAI, and KSI, the rate constants for uptake from the
26 intestine and transfer rates between compartments, respectively.
27
B-14 DRAFT - DO NOT CITE OR QUOTE
-------
c
d>
0
o
'fn
0)
Oc
.O
*i t;
2C
.0
o
-o
3C
^A
lfm
r^~ r\m«_ ~x
r /
\
%x
\
V
* V, %
"" * ^
«^_ ./—»*—• VmaxC • - -
6 8 10
Time (hr)
12
14
16
Figure B-6. Mouse model inhalation route sensitivity coefficients for metabolic
1 parameters. Sensitivity coefficients calculated for an exposure of 1,000 ppm MeOH are
2 reported for blood MeOH AUG. Note: Km, Vmax refer to the high-affinity, low-capacity
3 pathway and Km2, Vmax2 refer to the low-affinity, high-capacity pathway.
d>
0
it
0)
o
o
'fn
o>
t
3C
.0
2C
•\ <^
0
U.O
n
n
-0.5
(
S *~r*.
*"
QPC"
s
s
/
\
i QLC-J"^
D 2 4 6 8 10 12 14 1
Time (hr)
1
2
Figure B-7. Mouse model inhalation route sensitivity coefficients for flow rates (QCC:
cardiac output; QPC: alveolar ventilation), and partitioning to the body (PR)
compartment are reported for blood MeOH AUC. Sensitivity coefficients calculated for
an exposure to 1,000 ppm MeOH.
B-15
DRAFT - DO NOT CITE OR QUOTE
-------
0)
'o
1
8
(A
0)
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
\
VmASC
KAI
KSI
KMASC
6 8 10
Time (hr)
12
14
16
1 Figure B-8. Mouse model sensitivity coefficients for oral exposures to MeOH. The
2 sensitivity of blood MeOH AUC to oral absorption rate constants (KAS: stomach; KAI:
3 intestine; KSI: transfer between compartments) is reported.
1
2
3
4
5
6
7
8
9
10
11
B.2.5. Mouse Drinking Water Ingestion Pattern
To simulate exposures of mice via drinking water under bioassay conditions, an ingestion
pattern first used by Keys et al. (2004), based on data from Yuan (1993) was used. The pattern
specifies a fraction of percent of total daily ingestion consumed in each half-hour interval. The
first interval was shifted to correspond to the beginning of the active (dark) period, for
consistency with patterns used for humans and rats. A Table function was used in acslXtreme to
interpolate an instantaneous rate between the measured (30-min) values, with normalization so
that the 24-hour integral equals 100%. The daily pattern is shown in Figure B-9A and the
resulting blood concentration for a mouse exposed for 6 days per week (2100 mg/kg) is shown in
Figure B-9B.
B-16
DRAFT - DO NOT CITE OR QUOTE
-------
^-, 1,500-,
_l
cn
E 1,200-
c
o
S3
ID
b
c
•D
U
C
o
U
~0
o
_0
CD
900-
600-
300-
i— —r
18 20
22 24
Time (h)
B
24
48
72 96
Time (h)
120
144
168
Figure B-9. Mouse daily drinking water ingestion pattern (A) and resulting
predicted blood concentration for a 6 d/wk exposure (B). Mouse drinking
water exposures were simulated by multiplying the fractional rate (1/h) as a
function of clock time by the daily total dose ingested (mg) to obtain a rate of
addition of methanol into the stomach lumen compartment (mg/h).
Source: Yuan (1993); Keys et al. (2004)
B-17
DRAFT - DO NOT CITE OR QUOTE
-------
B.2.6. Rat Model Calibration
1 The model was initially calibrated-to-fit data from intravenous, inhalation, and oral
2 exposures in Sprague-Dawley (SD) rats using the 100 and 2500 mg/kg intravenous (IV) data
3 provided in the command file of Ward et al. (1997). Holding other parameters constant, the rat
4 PBPK model was calibrated against the Ward et al. (1997) IV-route blood pharmacokinetic data
5 (Figure B-10) by fitting Michaelis-Menten constants for one high affinity/low capacity and one
6 low-affinity, high-capacity enzyme, using the optimization routines in acslXtreme v2.3. Also
7 shown for comparison in Figure B-10A are the 100 mg/kg IV data of Horton et al. (1992),
8 obtained using Fischer 344 (F344) rats (data extracted from figures using Digitizlt), with a model
9 simulation (heavy red line) which differs from that for the SD rat only due to the predicted effect
10 of know body weight differences. While the fit to the Ward et al. data for SD rats is excellent,
11 especially for the lower dose, the rate of clearance is over-predicted for the F344 rat when
12 parameters fit to SD rat data are used. The 100 mg/kg IV data, with an alternate simulation for
13 the F344 rat obtained with distinct parameters (see below) is expanded in Figur B-10B,
14 emphasizing the difference in clearance between the two strains.
15 We then attempted to fit the model to the inhalation data of Horton et al. (1992) by
16 adjusting only the inhalation fractional uptake (FRACIN). The results, shown in Figure B-11A,
17 are clearly poor. While the model does match the uptake portion of the inhalation data for the
18 1200 and 2000 ppm exposures, it under-predicts the peak concentration reached at 200 ppm.
19 Further, the post-exposure clearance predicted by the model is much more rapid than indicated
20 by the data, as occurred with the IV kinetics (Figure B-10). (Since the peak concentration for the
21 2000 ppm inhalation exposure actually occurred at 7 hr, we also simulated a 7-hr exposure,
22 shown by the thin black line. The result indicates that the data are more consistent with and
23 better predicted by the longer exposure duration, but clearance is still over-predicted post-
24 exposure.) Therefore we concluded that the data show a clear strain difference in metabolism,
25 and should support at least a partially independent set of parameters for SD and F344 rats.
26 We then combined the 100 mg/kg IV and inhalation data of Horton et al. (for F344 rats)
27 and attempted to simultaneously identify the four metabolic parameters (Vmax and Km for two
28 pathways) and FRACIN. However when this was done the resulting values for the two Km's
29 were ~ 90 ± 50 mg/L and 70 ± 40 mg/L (Km and Km2, respectively), which are clearly
30 indistinguishable from a statistical standpoint. If instead the Km's were fixed at the more distinct
31 values identified from the SD rat IV data (6.3 and 65 mg/L), the optimization routine tended to
32 set the Vmax associated with the lower Km to zero. Thus the F344 rat data of Horton et al.
33 (1992) appear to be most consistent with a single metabolic pathway, even though the observed
34 concentrations spanned almost 2 orders of magnitude. Therefore those data (including the 100
35 mg/kg IV data) were simultaneously fit by adjusting a single Vmax and Km, along with the
B-18 DRAFT - DO NOT CITE OR QUOTE
-------
1 inhalation fraction, FRACIN, with the second metabolic pathway set to zero (Figures B-10B and
2 B-11B).
10,000
en
I
O
-------
100
10
I
o
v
•o
o
o
0.1
— 2000 ppm simulation — 1200 ppm simulation
— 200 ppm simulation o 2000 ppm data
* 1200 ppm data A 200 ppm data
— 2000 ppm, 7-hr simulation
6 8 10
Time (hr)
12
14 16
100
10
I
O
0!
o
o
0.1
B
l IN.
f— 2000 ppm simulation — 1200 ppm simulation
— 200 ppm simulation D 2000 ppm data
*• 1200 ppm data A 200 ppm data
(j- 2000 ppm, 7-hr simulation
8
Time (hr)
10
14 16
Figure B-ll. Model fits to data sets from inhalation exposures to 200
(triangles), 1,200 (diamonds), or 2,000 (squares) ppm MeOH in male F-344
rats. (A) Model fits with metabolic parameters set to values obtaine from IV data
for Sprague-Dawley rats, with only the inhaled fraction (FRACIN) adjusted. (B)
Model fits obtained by fitting metabolic parameters (Vmax and Km) for a single
pathway, along with FRACIN, to these data as well as the 100 mg/kg IV data
from F-344 rats (Figure B-9B).
Symbols are concentrations obtained using Digitizlt!. Lines represent PBPK
model fits. As the 7-hour data point at 2,000 ppm is higher than the 6-hour data
point (more evident on a linear scale) and appears more consistent with a 7-hour
exposure, a model simulation for a 7-hour exposure at 2,000 ppm is also shown
(lighter line).
Source: Horton et al. (1992)
B-20
DRAFT - DO NOT CITE OR QUOTE
-------
1 Model simulations of the F344 rat data with the F344-specific parameters are shown in
2 Figure B-10B (heavy red line) and Figure B-lIB. Unfortunately we were unable to
3 simultaneously fit the inhalation data for all exposure levels, although a wide range of metabolic
4 saturation (Km) values were tested. We could obtain a better fit of the high-concentration data
5 by constraining FRACIN to a higher value, for example, but then the fits to the lower
6 concentration data were compromised (not shown). Examining Figure 2 of Horton et al. (1992),
7 the experimental variability (indicated by the error bars) on the 2000 ppm data was much larger
8 than the 200 or 1200 ppm data, and as indicated by the simulations in Figure B-ll here, there is
9 at least the appearance that the exposure was actually for 7 hr instead of 6 hr. (To be clear, the
10 2000 ppm data were used in the optimization with the duration of inhalation set to 6 hr, but the
11 routine selected parameters which only poorly fit those data.) Since our greatest concern is in
12 predicting dosimetry at lower exposure levels, near to the points of departure, we decided to
13 retain the fits shown here. The corresponding parameters are listed in Table B-l. The fractional
14 absorption (20%) was lower than that estimated for mice (66.5%), but Perkins et al. (1995) also
15 found lower fractional absorption of inhaled methanol in rats vs. mice.
16 Finally, first-order oral absorption parameters were first fit to the lower dose (100 mg/kg)
17 oral absorption data reported by Ward et al. (1997), using the optimization routines in
18 acslXtreme v2.3 (Figure B-12, heavy/solid lines). (Since the animals used were SD rats, the SD-
19 specific metabolic parameters were used.) While the fit to the low-concentration data was quite
20 good (Figure B-12, lower panel), the fit to the 2500 mg/kg data (Figure B-12, upper panel)
21 exhibited a much faster and higher peak than shown by the data. Even when the model was fit to
22 both the high- and low-concentration data simultaneously, the fit to the high-concentration data
23 could not be significantly improved without completely degrading the low-concentration fit (not
24 shown). Also note that the 2500 mg/kg linear simulation completely over-estimates all the data
25 points; i.e., the area-under-the-curve for this dose is higher than indicated by the data, indicating
26 that the assumption of 100% absorption is not valid. Therefore, an alternative model using a
27 saturable (Michaelis-Menten) equation for absorption from the stomach and fecal elimination
28 (linear term) from the intestine was considered (thin lines) and found to significantly improve the
29 high-concentration simulation, with a nearly identical fit the low-concentration data. While
30 methanol absorption is not known to be regulated by transporters or other processes that would
31 give rise to rate saturation, it is clear form the discrepancy between the linear model and the 2500
32 mg/kg data that uptake is slower than predicted by such a model and its use would lead to an
33 over-prediction of internal concentrations. Therefore parameters for the saturable uptake model
34 are reported in Table B-l and the KMASC applied to mice and humans. Note that since the
35 saturation constant corresponds to a fairly large dose (620 mg/kg), the model is still effectively
36 linear at low- to moderate dose rates.
B-21 DRAFT - DO NOT CITE OR QUOTE
-------
E
3E
2,700
2,400 -
2,100-
1,800
1,500-
1,200-
900-
600-
300-
0-
2500 mg/kg linear uptake simulation
2500 mg/kg saturable uptake simulation
2500 mg/kg data
100 mg/kg linear uptake simulation
100 mg/kg saturable uptake simulation
100 mg/kg data
0
12
18
Time (hr)
24
30
36
o
E
••_••
O
Qj
100 H
80
60
40
3 20
— 100 mg/kg linear uptake simulation
100 mg/kg saturable uptake simulation
n 100 mg/kg data
012345678
Time (hr)
Figure B-12. Model fits to data sets from oral exposures to 100 (squares) or 2,500
1 (diamonds) mg/kg MeOH in female Sprague-Dawley rat (Expanded scale in lower
2 panel). Symbols are concentrations obtained from the command file. The thick lines
3 represent PBPK model fits using a linear (first-order) equation for absorption from
4 the stomach compartment with no fecal elimination, while the thin lines use a
5 Michaelis-Menton equation with a small fraction eliminated in the feces. All other
6 Gl rates, including absorption from the intestine, are first order.
Source: Ward etal.( 1997).
B-22
DRAFT - DO NOT CITE OR QUOTE
-------
B..2.7. Rat Model Simulations
1 A range of adverse developmental effects was noted in rat pups exposed to methanol
2 throughout embryogenesis (NEDO, 1987). SD rats were exposed in utero over different periods
3 of pregnancy and as neonates via inhalation or in drinking water. Inhalation exposures to
4 methanol were carried out for 18-22 hours, depending on the exposure group. Simulations of
5 predicted Cmax, AUC, and total metabolized from 22-hour exposures to 500, 1,000, and 5,000
6 ppmMeOHare shown in Figure B-13. Simulations of oral exposures of SD rats to 65.9, 624.1,
7 or 2,177 mg/kg-day (500, 5,000, 20,000 ppm in drinking water), daily dose estimations from the
8 study of Soffritti et al. (2002), based on measured water consumption, kindly provided by
9 Cynthia Van Landingham, Environ International, Ruston, Louisiana, are shown in Figure B-13.
10 Although the exposures in these studies are to rats over long periods and in some cases exposures
11 of the newborn pups, the model simulations are to NP adult rats only, using the dose-group
12 specific average body weights of 0.33-0.34 kg BW from the study of Soffritti et al. (2002) and do
13 not take into account changes is body weight or composition. These simulated values are
14 presumed to be a better surrogate for and predictor of target-tissue concentrations in developing
15 rats, and the corresponding estimated human concentrations a better predictor of developmental
16 risk in humans than would be obtained using the applied concentration or dose and default
17 extrapolations. The logic here is simply that the ratio of actual target tissue concentration (in the
18 developing rat pup or human) to the simulated concentration in the NP adult is expected to be the
19 same in both species and hence, that proportionality drops out in calculating a HEC.
20 Figure B-13 depicts simulations run to determine internal doses for 22 hours/day
21 inhalation exposures at 500, 1,000, or 2,000 ppm. Simulation results for continuous inhalation
22 exposures are shown for contrast. The simulations show that for all but the highest dose (2,000
23 ppm) steady state is reached within 22 hours, and that "periodicity," where the concentration
24 time course is the same for each subsequent day, is reached by the 2nd day of exposure. At
25 2,000 ppm, however, steady state is not reached until after 48 hours for the continous exposure.
26 Therefore, the Cmax, 24-hour AUC and amount metabolized per day (AMET) were by simulating
27 22 hours/day exposures for 5 days and calculating values of AUC and AMET over the last day
28 (24 hours) of that period.
29
B-23 DRAFT - DO NOT CITE OR QUOTE
-------
CD
12 24 36 48 60
Time (hr)
72
84
96
Exposure
concentration
(ppm)
500
1,000
2000
^max
(mg/L)
3.6
10.6
48.5
AUC
(hrmg/L
)
79
227
968
AMET
(mgEq)
17.6
34.8
67.2
Figure B-13. Simulated Sprague-Dawley rat inhalation exposures to 500,1,000, or
1 2,000 ppm MeOH. Rat BW was set to 0.33 kg. Simulations are shown for both
2 continuous (thin, dashed/dotted lines in plot) and 22 hours/day exposures (thick, solid lines
3 in plot). Cmax, AUC, and amount metabolized (AMET) are determined from the 22
4 hour/day simulations, run for a total of 5 days (120 hours), with the AUC and AMET
5 calculated for the last 24 hours of the simulation.
1
2 Figure B-14 depicts simulations run to mimic a single oral exposure, treated as a
3 continuous infusion for 12 hours (assuming 12-hour period when rats are awake and active).
4 Total AUC and AMET and AUC24 and AMET24 for the first 24 hours after start of exposure
5 were calculated.
B-24
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
4
5
6
7
8
9
10
11
12
13
14
15
16
10,000-a
1,000-
G~
"Si 100-
X
o
01
o
_o
CD
10-
1
0.1-J
0.01
0.001
0
40
Time (hr)
60
80
Exposure
dose
(mg/kg-day)
66.0
624.1
2177
Body
weight (kg)
0.33
0.33
0.34
^max
(mg/L)
9.3
631.6
2,832.4
AUC
(hrmg/L)
104.8
9,525
72,617
AUC24
(hrmg/L)
104.8
8,817
45,138
AMET
(mgEq)
21.2
155.6
347.4
AMET24
(mgEq)
21.2
122.6
138.4
Figure B-14. Simulated rat oral exposures of Sprague-Dawley rats to 65.9, 624.2, or
2,177 mg MeOH/kg/day. Dosing was simulated as a 12-hour, zero-order infusion to the
liver compartment. The AUC and total amount metabolized are given for a period
sufficient for the MeOH to clear (84 hours), and the AUC24 and AMET24 values represent
just the first 24 hours of exposure. (Results shown for illustrative purposes. Dosimetry used
in assessment was simulated using a more realistic water ingestion pattern.)
To simulate ingestion of methanol in drinking water by rats under bioassay conditions, an
ingestion pattern based on the observations of Spiteri (1982) and Peng et al. (1990). While mice
ingest water in frequent, small bouts (Gannon et al., 1992) that are reasonably described as a
continuous delivery to the stomach, rats exhibit clear periods of ingestion alternating with
periods where no ingestion occurs (Spiteri, 1982; Peng et al., 1990). Based on those data a
reasonable representation of rat water ingestion can be described as serious of pulses. During the
dark/active period of each day (first 12 hr) each bout of drinking was assumed to last 45 min
followed by 45 min without ingestion (total of 8 bouts). During the light/inactive period (next 12
hr) drinking bouts were assumed to last only 30 min followed by 2.5 hr (150 min) without
drinking (4 bouts). An equal amount was assumed to be consumed in each bout within the dark
period, likewise within each light-period bout, with the respective amounts adjusted such that
80% of the total ingestion occurs during the dark and 20% during the light (Burwell et al., 1992).
The resulting absorption pattern is shown in Figure B-15A and a simulated blood concentration
time-curve (for 50 mg/kg/day dosing) is shown in Figure B-15B.
B-25
DRAFT - DO NOT CITE OR QUOTE
-------
0,12-
I i
10
*- 0,1-
c
B 0,08-
Q.
0 0,06-
_Q
to 0,04-
C 0,02-
o
4-J
f 1
—
—
A
-
-
10 12 14 16
Time (hr)
18 20 22
24
B
12
18 24 30
Time (hr)
36
42
48
Figure B-15. Rat daily drinking water ingestion pattern (A) and resulting predicted
blood concentration for a 2-day exposure (B).Rat drinking water exposures were
simulated by multiplying the fractional absorption rate (1/hr) as a function of clock
time by the daily total dose ingested (mg) to obtain a rate of addition of methanol
into the stomach lumen compartment (mg/h).
B.2.8. Human Model Calibration
B. 2.8.1. Inhalation Route
4 The mouse model was scaled to human body weight (70 kg or study-specific average),
5 using human tissue compartment volumes and blood flows, and calibrated to fit the human
6 inhalation-exposure data available from the open literature, which comprised data from four
7 publications (Ernstgard et al., 2005; Batterman et al., 1998; Osterloh et al., 1996; Sedivec et al.,
8 1981).
9 A first-order rate of loss of MeOH from the blood, K1C, and a first-order bladder
10 compartment time constant, KBL, were used to provide an estimate of urinary MeOH
B-26 DRAFT - DO NOT CITE OR QUOTE
-------
1 elimination. The inhalation-route urinary MeOH kinetic data described by Sedivec et al. (1981)
2 (Figure B-16) were used to inform these parameters. The urine MeOH concentration data
3 reported by the authors were converted to amount in urine by assuming 0.5 mL/hr/kg total
4 urinary output (Horton et al., 1992). Sedivec et al. (1981) measured a fractional uptake of
5 57.7%, based on total amount inhaled. Since the PBPK model uses alveolar rather than total
6 ventilation and this is typically assumed to be 2/3 of total ventilation the fractional uptake of
7 Sedivec et al. (1981) was corrected by dividing by 2/3 to obtain a value for FRACIN of 0.8655.
8 The resulting values of K1C and KBL, shown in Table B-l, differ somewhat depending on
9 whether first-order or saturable liver metabolism is used. These are only calibrated against a
10 small data set and should be considered an estimate. Urine is a minor route of MeOH clearance
11 with little impact on blood MeOH kinetics.
12 Although the high-doses used in the mouse studies warrant the use of a second metabolic
13 pathway with a high Km, the human exposure data all represent lower concentrations and may
14 not require or allow for accurate calibration of a second metabolic pathway. Horton et al. (1992)
15 employed two sets of metabolic rate constants to describe human MeOH disposition, similar to
16 the description used for rats and mice, but in vitro studies using monkey tissues with non-MeOH
17 substrates were used as justification for this approach. Although Bouchard et al. (2001)
18 described their metabolism using Michaelis-Menten metabolism, Starr and Festa (2003) reduced
19 that to an effective first-order equation and showed adequate fits. Perkins et al. (1995) estimated
20 a Km of 320 + 1273 mg/L (mean + S.E.) by fitting a one-compartment model to data from a
21 single oral poisoning to an estimated dose. In addition to the extremely high standard error, the
22 large standard error for the associated Vmax (93 + 87 mg/kg/hr) indicates that the set of
23 Michaelis-Menten constants was not uniquely identifiable using this data. Other Michaelis-
24 Menten constants that have been used to describe MeOH metabolism in various models for
25 primates are given in Table B-2. Because the Km calculated by Perkins et al. (1995) from the
26 high-dose oral exposure is 320 mg/L, while the highest observed concentration in the data sets
27 considered here is 14 mg/L (Batterman et al., 1998), forcing the model to use this higher Km
28 would simply result in fits that are effectively indistinguishable from the linear model. A simple,
29 linear model is preferred over the use of a Km value that high.
B-27 DRAFT - DO NOT CITE OR QUOTE
-------
10
9 -
8 -
O)
3 -
2
1 -
Sedivec et al., urine
concentration
231 ppm saturable sim
157 ppm saturable sim
78 ppm saturable sim
231 ppm linear sim
157 linear sim
78 ppm linear sim
• 231 ppm data
A 157 ppm data
O 78 ppm data
20
24
2.5 -
Sedivec et al.,
total urinary excretion
20
24
Figure B-16. Urinary MeOH elimination concentration (upper panel) and cumulative
1 amount (lower panel), following inhalation exposures to MeOH in human volunteers.
2 Data points in lower panel represent estimated total urinary MeOH elimination from
3 humans exposed to 78 (diamonds), 157 (triangles), and 231 (circles) ppm MeOH for 8
4 hours, and lines represent PBPK model simulations. Solid lines are model results with the
5 saturable equation for hepatic metabolism while dashed lines show results for liner
6 metabolism. Data digitized from Sedivec et al. (1981) and provided for modeling by the
7 EPA.
Source: Sedivec et al. (1981).
B-28
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-2. Primate kms reported in the literature
Km(mg/L)
320 +1273
716 + 489
278
252+116
33.9
0.66
23.7 + 8.7"
Reference
Perkins et al, 1995
Perkins et al, 1995
Perkins et al., 1995
Perkins et al., 1995
Hortonetal., 1992
Fisher etal., 2000
(This analysis.)
Note
Human: oral poisoning, estimated dose
Cynomolgus monkey: 2 g/kg dose
Rhesus monkey: 0.05-1 mg/kg dose
Cynomolgus monkey: 1 g/kg dose
PBPK model: adapted from rat Km
PBPK model, Cynomolgus monkey: 10-900 ppm
PBPK model, human: 100-800 ppm
Note- the values from Perkins et al. (1995b), are + S.E.
aMean + S.D. This Km was optimized while also varying Vmax, K1C, and KBL, from all of the at-rest human
inhalation data as a part of this project. The S.D. given for this analysis is based on the Optimize function of
acslXtreme, which assumes all data points are discrete and not from sets of data obtained over time and
therefore a true S.D. would be a higher value. The final value reported in Table B-l (21 mg/L) was obtained by
sequentially rounding and fixing these parameters, then re-optimizing the remaining ones. For more detail, see
text and Table B-3.
1 To estimate both the Michaelis-Menten and first-order rates, all human data under
2 nonworking conditions (Batterman et al., 1998; Osterloh et al., 1996; Sedivec et al., 1981) were
3 used. Before discussing the parameter estimation, however, adjustments were made to one of
4 these data sets (Osterloh et al., 1996). Batterman et al., (1998) and Sedivec et al. (1981) both
5 subtracted background levels before reporting their results. However, Osterloh et al. (1996)
6 measured and reported (plotted) blood methanol in nonexposed controls (data shown in
7 Figure B-17). The data for Osterloh et al. (1996) clearly show a time-dependent trend which is
8 close to linear, and a linear regression is also included. However, the blood concentration
9 (average) in the exposed group of that study was -1.2 mg/L, whereas the data and regression in
10 Figure B-17 indicate a value of- 0.9 mg/L. Therefore, the exposure data for Osterloh et al
11 (1996) were corrected by subtracting time-zero value for the exposed group plus a time-
12 dependent factor obtained by multiplying the slope of this regression (0.093 mg/L-hr) by the
13 measurement time.
14 The metabolic (first-order or saturable) and urinary elimination constants were
15 numerically fit to the nonworking human data sets while holding the value for FRACIN at
16 0.8655 (estimated from the results of Sedivec et al. as described above) and holding the
17 ventilation rate constant at 16.5 L/hr/kg0'75 and QPC at 24 L/hr/kg0'75 (values used by EPA
18 [2000d] for modeling the inhalation-route kinetics vinyl chloride). Other human-specific
19 physiological parameters were set as reported in Table B-l.
B-29
DRAFT - DO NOT CITE OR QUOTE
-------
1.8
1.6
J 1.4
oi
5.1.2
I 1
| 0.8
§ 0.6
<§ 0.4
0.2
0
Control blood concentrations
(Osterlohetal., 1996) •
y = 0.093x + 0.9094
R2 = 0.9391
Time (hr)
Figure B-17. Control (nonexposed) blood methanol concentrations
Source: Ernstgard et al. (2005); Osterloh et al. (1996).
1 Either (a) the set of VmaxC, Km, K1C, and KBL were simultaneously varied while fitting
2 the entire data set or (b) KLLC, K1C, and KBL were so varied and fit. Thus the two model fits
3 are separated by a single degree of freedom (one additional parameter in case [a]). Statistical
4 results given in Tables B-2 and B-3 are from these global fitting exercises. Final fitted
5 parameters that have been used in the model for the risk assessment are given in Table B-l. The
6 resulting fits of the two parameterizations (1st order or optimized Km/Vmax) are shown in
7 Figures B-l6 and B-l8.
8 Use of a first-order rate has the advantage of resulting in one fewer variable in the model
9 and results in an adequate fit to the data, but the saturable model clearly fits some of the data
10 better (Figures B-l6 and B-l8). To discriminate the goodness-of-fit resulting from the inclusion
11 of an additional variable necessary to describe saturable metabolism versus using a single first-
12 order rate, a likelihood ratio test was performed. Models are considered to be nested when the
13 basic model structures are identical except for the addition of complexity, such as the added
14 metabolic rate. Under these conditions, the likelihood ratio can be used to statistically compare
15 the relative ability of the two different metabolism scenarios to describe the same data, as
16 described by Collins et al. (1999). The hypothesis that one metabolic description is better than
17 another is calculated using the likelihood functions evaluated at the maximum likelihood
18 estimates. Since the parameters are optimized in the model using the maximum LLF, the
19 resultant LLF is used for the statistical comparison of the models. The equation states that two
20 times the log of the likelihood ratio follows a %2 distribution with r degrees of freedom:
21 -2[log(/l(modell)//l(model2))] =-2[log/l(modell)-log/l(model2)] = X2r
B-30
DRAFT - DO NOT CITE OR QUOTE
-------
1 The likelihood ratio test states that if twice the difference between the maximum LLF of the two
2 different descriptions of metabolism is greater than the ^ distribution, then the model fit has
3 been improved (Devore, 1995; Collins et al., 1999; Steiner et al., 1990).
Batterman et al., blood concentration
(800 ppm exposures)
-Saturable metabolsim
First-order metabolism
2 h data
1 h data
30 min data
Osterloh et al.
(200 ppm inhalation)
O Data
Saturable metabolsim
First-order metabolism
Time (hr)
Figure B-18. Data showing the visual quality of the fit using optimized first-order or
4 Michaelis-Menten kinetics to describe the metabolism of MeOH in humans. The rate
5 constants used for each simulation are given in Table B-3.
Source: Batterman et al. (1998: top); Osterloh et al. (1996: bottom).
Table B-3. Parameter estimate results obtained using acslXtreme to fit all human data
using either saturable or first-order metabolism
Parameters
Michaelis-Menten (optimized)
Km
Optimized value
T3 O
23.8
S.D.
8.8
Correlation matrix
-0.994
LLF
-24.1
B-31
DRAFT - DO NOT CITE OR QUOTE
-------
vmaxc
First Order
KLLC
33.2
95.7
10.1
5.4
NA
-31.0
Note: The S.D.s are based on the Optimize function of acslXtreme, which assumes all data points are discrete
and not from sets of data obtained over time and therefore a true S.D. would be a higher value.
1 At greater than a 99.95% confidence level, using 2 metabolic rate constants (Km and
2 VmaxC) is preferred over utilizing a single rate constant (Table B-4). While the correlation
3 coefficients (Table B-3) indicate that Vmax and Km are highly correlated, that is not unexpected,
4 and the S.D.s (Table B-3) indicate that each is reasonably bounded. If the data were
5 indistinguishable from a linear system, Km in particular would not be so bounded from above,
6 since the Michaels-Menten model becomes indistinguishable from a linear model as VmaxC and
7 Km tend to infinity. Moreover, the internal dose candidate POD, for example the BMDLio for
8 the inhalation-induced brain-weight changes from NEDO (1987), with methanol blood AUC as
9 the metric, is 374.67 mg-hr/L, which corresponds to an average blood concentration of 15.6
10 mg/L. Therefore the Michaelis-Menten metabolism rate equation appears to be sufficiently
11 supported by the existing data, and its use is expected to improve the accuracy of the HEC
12 calculations, since those are being conducted in a concentration range in which the nonlinearity
13 has an impact.
Table B-4. Comparison of LLF for Michaelis-Menten and first-order metabolism
LLF(logI)forM-M
-24.1
LLF (logl) for 1st
order
-31.0
LLF
1st versus M-Ma
34.1
xl (99% confidence)13
13.8
X2r (99.95%
confidence) b
12.22
Note: The models were optimized for all of the human data sets under non working conditions. M-M:
Michaelis-Menten
"obtained using this equation: - 2[log /l(model l) - log /l(model 2)]
^significance level at r = 1 degree of freedom.
14 While the use of Michaelis-Menten kinetics might allow predictions across a wide
15 exposure range (into the nonlinear region), extrapolation above 1,000 ppm is not suggested since
16 the highest human exposure data are for 800 ppm. Extrapolations to higher concentrations are
17 potentially misleading since the nonlinearity in the exposure-internal-dose relationship for
18 humans is uncertain above this point. The use of a BMD or internally applied UFs should place
19 the exposure concentrations within the range of the model.
20 The data from Ernstgard et al. (2005) was used to assess the use of the first-order
21 metabolic rate constant to a dataset collected under conditions of light work. Historical measures
22 of QPC (52.6 L/hr/kg0'75) and QCC (26 L/hr/kg0'75) for individuals exposed under conditions of
23 50 w of work from that laboratory (52.6 L/hr/kg0'7) (Ernstgard, personal communication; Corley
B-32
DRAFT - DO NOT CITE OR QUOTE
-------
1 et al., 1994; Johanson et al., 1986) were used for the 2-hour exposure period (Figure B-19).
2 Otherwise, there were no changes in the model parameters (no fitting to these data). The results
3 are remarkably good, given the lack of parameter adjustment to data collected in a different
4 laboratory, using different human subjects than those to which the model was calibrated.
Ernstgard et al. (2005)
5
6
7
8
9
10
5
6
7
8
9
10
11
12
13
14
15
16
17
18
OQ
4 6
Time (hr)
10
Figure B-19. Inhalation exposures to MeOH in human volunteers. Data points
represent measured blood MeOH concentrations from humans (4 males and 4 females)
exposed to 100 ppm (open symbols) or 200 ppm (filled symbols) for 2 hours during light
physical activity. Solid lines represent PBPK model simulations with no fitting of model
parameters. For the first 2 hours, a QPC of 52.6 L/hr/kgO.75 (Johansen et al. 1986), and a
QCC of 26 L/hr/kgO.75 (Corley et al., 1994) were used by the model.
Source: Ernstgard et al.(2005).
B.2.8.1. (Should be B.2.8.2) Oral Route
There were no human data available for calibration or validation of the oral route for the
human model. In the absence of data to estimate rate constants for oral uptake, the 'humanset.m'
file which sets parameters for human simulations applies the KMAS for the mouse with the other
absorption parameters set to match those identified for ethanol in humans by Sultatos et al.
(2004); VmASC was set such that for a 70-kg person, VMAS/KMAS matches the first-order
uptake constant of Sultatos et al. (2004) (0.21 hr"1). While Sultatos et al. include a term for
ethanol metabolism in the stomach, no such term is included here and the rate of fecal
elimination is set to zero, corresponding to 100% absorption. However zero-order ingestion, a
continuous infusion at a constant rate into the stomach lumen equal to the daily dose/24 hours,
was assumed for all human simulations. Since absorption was assumed to be 100% of
administered MeOH, at steady state the rate of uptake from the stomach and intestine
compartments (combined) must equal the rate of infusion to the stomach. Since Cmax is driven by
the oral absorption rate, which was assumed rather than fitted and verified, Cmax was not used as
a dose metric for human oral route simulations. AUC, which is less dependent on rate of uptake,
B-33
DRAFT - DO NOT CITE OR QUOTE
-------
1 was used as the dose metric and for estimation of HEDs. Since the AUC was computed for a
2 continuous oral exposure, its value is just 24-hours times the steady-state blood concentration at
3 a given oral uptake rate.
B. 2.8.1. (B. 2.8.3) Inhalation Route HECs and Oral Route HEDs
4 The atmospheric MeOH concentration resulting in a human daily average blood MeOH
5 AUC (hrxmg/L) or Cmax (mg/L) equal to that occurring in experimental animals following
6 exposure at the POD concentration is termed the HEC. Similarly, the oral dose (rate) resulting in
7 human daily average blood MeOH AUC (hrxmg/L) equivalent to that occurring in an
8 experimental animal at the POD concentration is termed the HED.
9 To determine the HEC for specific exposures in mice, the mouse PBPK model is first
10 used to determine the daily blood MeOH 24-hours AUC and Cmax associated with 7 hour/day
11 inhalation exposures. Mice were exposed each day for 10 days, so the full 10-day exposure was
12 simulated and an average 24-hours AUC calculated over that time, so no other duration
13 adjustment was needed. The human AUC was determined for the last 24 hours of a continuous
14 1,000-hour exposure, to assure steady state was achieved. The human Cmax was determined at
15 steady state and so is equivalent to the steady state blood MeOH concentration. Results are given
16 in Table B-5 and for inhalation shown in Figure B-20.
17 For example, for a 1,000 ppm exposure this resulted in model-predicted peak blood of
18 133 mg/L and an AUC of 770 (hrxmg/L). The human model can then be used to determine the
19 human MeOH exposure concentration leading to the same daily average AUC or Cmax under
20 continuous exposure conditions. Based on AUC, the HEC of the 1,000-ppm exposure is
21 684 ppm, while based on peak (human steady-state) concentration, the HEC is predicted to be
22 1110 ppm. The parameters used in the human model for these simulations are listed in Table B-l
23 for saturable kinetics.
24 The HED was calculated by using the human model to find the oral dose (mg/kg-day) that
25 gave a blood MeOH AUC equivalent to the mouse AUC following an exposure at the POD.
26 Zero-order absorption was assumed. For example, the human oral exposure equivalent to a
27 1,000-ppm inhalation exposure in mice (i.e., with an AUC of 770 mg-hr/L) is 163 mg/kg-day.
28 Since a 200 mg/kg-day oral exposure gave a human AUC of 1,320 mg-hr/L, which falls between
29 the values predicted for inhalation exposures at 800 ppm (1,090 mg-hr/L) and 1,000 ppm (2,090
30 mg-hr/L), this oral exposure rate was taken to be the upper end for the model to accurately
31 estimate an HED.
B-34 DRAFT - DO NOT CITE OR QUOTE
-------
Table B-5. PBPK model predicted Cn
exposed to MeOH
and 24-hour AUC for mice and humans
Exposure
concentration
(ppm)
1
10
50
100
250
500
1,000
2000
5000
1,0000
Inhalation Route
Mouse"
AUC
(mg-hr/L)
0.15
1.52
7.98
17.0
53.4
170
770
3310
17300
51200
Cmax (mg/L)
0.021
0.22
1.14
2.45
7.77
26.1
133
524
2000
4710
Humanb
AUC
(mg-hr/L)
0.59
5.97
30.6
63.3
177
447
2090
31100
147000
341,000
Css (mg/L)
0.025
0.25
1.28
2.64
7.36
18.6
87.2
1300
6130
14200
Oral Route
Human
Dose
(mg/kg-day)
0.1
1
10
50
100
200
500
1,000
2000
5000
AUC
(mg-hr/L)
0.204
2.05
21.2
124
315
1320
39400
125000
297000
814000
"The mouse 24-hour average AUC were calculated under the conditions of the bioassay: 10 days of exposure
with 7 hours of exposure during each 24 hour period.
bHuman simulation results are considered unreliable above 1,000 ppm (inhalation) or 200 mg/kgday (oral), but are
included for comparison
1 Again, since the available human exposure data is to, at most, 800 ppm, the model could
2 not be calibrated for higher exposures that approximate most of the mouse and rat exposure
3 concentrations. The AUC in humans for similar exposure levels is ~3 times greater than in the
4 mouse, primarily because human exposure estimates are expected to result from 24-hour
5 exposures and the mice were exposed for 7 hours.
B-35
DRAFT - DO NOT CITE OR QUOTE
-------
1000000
100000 -
10000 -
•?
O)
E.
O
<
1000 -
100 -
10 -
1 -
0.1
•Mouse
Human
Human-linear
1 10 100 1000 10000
Concentration (ppm)
10000
10 100 1000 10000
Concentration (ppm)
10000
"8 1000
I iT100 H
°
o
<
i 10 -
1 -
0.1
Human
Human-linear
10 100 1000
Concentration (ppm)
10000
Figure B-20. Predicted 24-hour AUC (upper left), Cmax (upper right), and amount
metabolized (lower) for MeOH inhalation exposures in the mouse (average over a 10-
day exposure at 7 hours/day) and humans (steady-state values for a continuous
exposure). Cmax for human exposures is equal to the steady-state blood concentration. For
humans, the long-dashed lines are model predictions using Michaelis-Menten
metabolism (optimized Km of 23.8 mg/L) and the short-dashed lines are model
predictions using first-order kinetics. Amount metabolized normalized to BW° 75 to
reflect cross-species scaling (Human simulation results above 1,000 ppm are not
considered reliable but are shown for comparison).
1 While the PBPK computational code can be used in the future to derive HECs or HEDs
2 for other exposures, an alternative approach was developed that allows non-PBPK model users to
3 estimate MeOH HECs and HEDs from benchmark doses in the form of AUCs. This approach
4 uses algebraic equations describing the relationship between predicted MeOH 24-hour AUC or
5 total amount metabolized in the liver (MET, mg [per day]) (constant 24-hour exposure) and the
6 inhalation exposure level (i.e., an HEC in ppm) (Equations 1, Ib, 3 or 3b) or oral exposure rate
7 (i.e., an HED in mg/kg-day) (Equations 2, 2b, 4 or 4b). To use the equations to derive an HEC or
8 HED, the target human AUC is applied to the appropriate equation. Since these relationships are
9 for continuous exposures, blood concentration is constant, and hence extrapolation for a Cmax is
10 obtained by simply using AUC = 24*Cmax.
B-36
DRAFT - DO NOT CITE OR QUOTE
-------
4
5
6
7
JTLC,^ ^PJJIH
Rf C (ppm)
HED (mg/kg/
RfD (mg/kg/d£
HEC(ppm
-------
1 species, and hence that using NP maternal blood levels in place of fetal concentrations will not
2 lead to a systematic error when extrapolating risks. Thus, a full representation of pregnancy and
3 the fetal/conceptus compartment appears to be unnecessary.
4 While lactational exposure is less direct than fetal exposure and blood or target-tissue
5 levels in the breast-feeding infant or pup are likely to differ more from maternal levels, the
6 health-effects data indicate that most of the effects of concern are due to fetal exposure, with
7 only a small influence due to postbirth exposures. Separating out the contribution of postbirth
8 exposure from pre-birth exposure to a given endpoint in a way that would allow the risk to be
9 estimated from estimates of both exposure levels would be extremely difficult, even if one had a
10 lactation/child PBPK model that allowed for prediction of blood (or target-tissue) levels in the
11 offspring. And one would still expect the target-tissue concentrations in the offspring to be
12 closely related to maternal blood levels (which depend on ambient exposure and determine the
13 amount delivered through breast milk), with the relationship between maternal levels and those
14 in the offspring being similar across species.
15 Therefore, the development of a lactation/child PBPK model appears not to be supported,
16 given the minimal change that is likely to result in risk extrapolations and use of (NP) maternal
17 blood levels as a measure of risk in the offspring is still considered preferable over use of default
18 extrapolation methods. In particular, the existing human data allow for accurate predictions of
19 maternal blood levels, which depend strongly on the rate of maternal methanol clearance.
20 Failing to use the existing data (via PBPK modeling) on human methanol clearance (versus that
21 in other species) would be to ignore this very important determinant of exposure to breast-fed
22 infants. And since bottle-fed infants do not receive methanol from their mothers, they are
23 expected to have lower or, at most, similar overall exposures for a given ambient concentration
24 than the breast-fed infant, so that use of maternal blood levels for risk estimation should also be
25 adequately protective for that group.
26 During model development, several inconsistencies between experimental blood MeOH
27 kinetic data embedded in the Ward et al. model (1995) and the published figures first reporting
28 these data were discovered. Therefore, data were digitized from the published literature when a
29 figure was available, and the digitized data was compared to the provided data. When the
30 digitized data and the data embedded in the computational files (i.e., provided to Battelle under
31 contract from the EPA) were within 3% of each other, the provided data was used; when the
32 difference was greater than 3%, the digitized data was used. Often, using the published figures
33 as a data source resulted in substantial improvements of the fit to the data in the cases where the
34 published figures were different from the embedded data.
35 The final MeOH PBPK model fits well inhalation-route blood kinetic data from separate
36 laboratories in rodents and humans. Intravenous-route blood MeOH kinetic data in NP mice
37 were only available for a single i.v. dose of 2,500 mg/kg, but were available for GDIS mice
38 following administration of a broader range of doses: 100, 500, and 2,500 mg/kg. Up to
B-3 8 DRAFT - DO NOT CITE OR QUOTE
-------
1 20 hours postexposure, blood MeOH kinetics appear similar for NP and pregnant mice after
2 administration of 2,500 mg/kg. The intravenous pharmacokinetic data in GDIS mice showed an
3 unexpected dose-dependent nonlinearity in initial blood concentrations, suggesting either a dose
4 dependence on the volume distribution, which is unlikely, or some source of experimental
5 variability. To account for this nonlinearity, Ward et al. (1997) used dose-specific partition
6 coefficients for placenta and embryonic fluid and Vmax for the metabolism of MeOH. The
7 current model uses a consistent set of parameters that are not varied by dose and therefore does
8 not fit these 100 mg/kg dose intravenous data. The model does fit the 500 and 2,500 mg/kg
9 doses, and if a presumed i.v. dose of 200 mg/kg (twice the reported 100 mg/kg) is employed, is
10 able to predict initial blood concentrations for the lowest dose data, as expected. The i.v. data
11 from the Ward et al. (1995) model does match the corresponding published figures.
12 The model fits to the mouse oral-route MeOH kinetic data using a consistent set of
13 parameters (Figure B-4) are reasonably good but not as good as fits to the inhalation data. The
14 model consistently underpredicts the amount of blood MeOH reported in two studies (Ward
15 et al., 1997, 1995). Ward et al. (1997) utilized a different Vmax for each oral absorption data set.
16 In the report by Ward et al. (1997) the GDIS and the GD8 data from Dorman et al. (1995) were
17 both fit using a Vmax of-80 mg/kg/hr (body weights were not listed, the model assumed that
18 GD8 and GDIS mice were both 30 g; Ward et al. (1997) did not scaled by body weight), but
19 lower partition coefficients for placenta (1.63 versus 3.28) and embryonic fluid (0.0037 versus
20 0.77). The current model adequately fits the oral pharmacokinetic data using a single set of
21 parameters that is not varied by dose or source of data.
22 The fits of the rat model to the limited dataset readily available were quite good. The low-
23 dose exposures of all routes were emphasized in model optimization since they were the doses
24 most relevant to risk assessment. Based on a rat inhalation exposure to 500 ppm, the human HEC
25 would be 301 ppm (by applying an AUC of 226 [Figure B-12] to Equation 1).
26 The mouse, rat, and human models fit multiple datasets from multiple research groups
27 using consistent parameters that are representative of each species, but are not varied within
28 species. Using the model, it will be possible to ascertain chronic human exposure concentrations
29 that are likely to be without an appreciable risk of deleterious effects.
B.3. ADDITIONAL MATERIAL
30 • Results from Optimizations
31 • acslXtreme Program (.csl) File (Electronic Attachment)
32 • acslXtreme procedure (.cmd) file
33 • Key to .m files for reproducing the results in this report
34 • Code for .m files
B-39 DRAFT - DO NOT CITE OR QUOTE
-------
1 • Personal communication from Lena Ernstgard regarding human exposures reported in the
2 Ernstgard and Johanson, 2005 SOT poster
3 • Personal communication from Dr. Rogers regarding mouse exposures to MeOH
4 • Data and simulations for MeOH Metabolic Clearance/Total Metabolites Produced
5 • Multiple daily oral dosing for humans
B.3.1. Results From Optimizations
B.3.1.1. Approach for and Results of the Optimization of Metabolic Parameters and
Inhalation Route Fractional Availability in Mice
6 The approach and results are presented below in outline format with supporting figures. More
7 complete documentation was not developed because the products of the optimizations were not
8 used in the final model. The documentation here is intended only to demonstrate that appropriate
9 optimizations were conducted and what the results of those optimizations were.
10 1. The Vmax for the low affinity pathway was set to 0 and the remaining VmaxC, Km, and
11 fractional availability were optimized using inhalation data only.
12 a. The optimizer was unable to find a value for Km that was greater than 0.
13 b. The resulting metabolic parameters essentially represented a zero order loss process.
14 2. The Vmax for the low affinity pathway was set to 0 and the remaining VmaxC and Km were
15 optimized using all (oral, intravenous and inhalation) data.
16 a. The optimized single Km, 135 mg/L, was equal to the average of the 2 original Kms.
17 b. Fits to the MeOH blood levels following inhalation exposures > 2,000 ppm are slightly
18 improved, but the model fits to the 1,000 ppm exposure concentration overpredict
19 reported values by 20%.
20 3. Parameters for both metabolic pathways were optimized using all (oral, intravenous and
21 inhalation) data.
22 a. The fit to the high-dose intravenous data from Ward etal. (1997) (2,500 mg/kg) was
23 improved (Figure B-21).
24 b. The fit to the high-dose oral data, also from Ward etal. (1997), (2,500 mg/kg) was
25 improved (Figure B-22).
26 c. The fit to the mid-dose i.v. data (500 mg/kg) dose was not as good as using the visually
27 fit parameters (Figure B-21)
28 d. The fit to the low-dose oral data (1,500 mg/kg) was not as good when the visually fit
29 parameters were used (Figure B-22). The low-dose data was from Dorman et al. (1995).
B-40 DRAFT - DO NOT CITE OR QUOTE
-------
e. Neither set of parameters resulted in an adequate fit to the low-dose intravenous data
(100 mg/kg; Figure B-21).
f. Fits to the inhalation data following exposures to < 5,000 ppm MeOH were substantially
worse than when using the visually fit parameters (Figure B-23)
10000
1000
o
-------
Figure B-22. Fit of the model to oral data using different clearance and uptake
1 parameter optimizations. Solid blue lines - visually optimized; dashed red lines -
2 clearance parameters (Km, Km2, VmaxC, Vmax2C) optimized using all inhalation data sets;
3 dash/dot green lines - clearance parameters optimized using all data sets (inhalation, oral,
4 and intravenous).
Source: Ward et al. (1997); Dorman et al. (1995).
Figure B-23. Fit of the model to inhalation data using different clearance and uptake
1 parameter optimizations. Dotted black lines - model optimized fractional inhalation, solid
2 blue lines - visually optimized; dashed red lines - clearance parameters (Km, Km2, VmaxC,
3 Vmax2C) optimized using all inhalation data sets; dash/dot green lines - clearance
4 parameters optimized using all data sets (inhalation, oral, and intravenous).
Source: Rogers et al. (1997).
B-42
DRAFT - DO NOT CITE OR QUOTE
-------
B. 3.1.1. [B3.1.2.J Conclusion
1 Under the best circumstances, formal optimizations offer the benefit of repeatability and
2 confirmation that global optima have not been missed by user-guided visual optimization.
3 Incorporating judgments regarding the value of specific data sets while easy when visually
4 fitting, is difficult at best when using optimization routines. This is an important distinction
5 between these approaches for this modeling exercise.
6 The mouse NOEL was 1,000 ppm MeOH. Fitting the blood MeOH concentration data at
7 this exposure drove our modeling exercises because of the importance of this exposure group in
8 the risk assessment. Unfortunately, the vast majority of the blood MeOH data came from much
9 higher exposures. As expected, our various attempts at optimization led to fits that were better
10 for some, but never all, data sets. This is to be expected when there is clearly significant
11 variability in the underlying data. Various data weighting schemes were included to improve
12 overall optimization while maintaining a good fit to the 1,000 ppm data. In the end, optimization
13 offered no significant improvement over the fractional uptake and metabolic parameter values
14 obtained by visual optimization, so these were retained in the final version of the model.
B.3.2. acslXtrerne Program (.csl)File
15 PROGRAM MeOH - PBPK Model for Methanol
16 PROGRAM MeOH - PBPK Model for Methanol
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44 INITIAL
45 ! Initialize some Variables before start
Based on MeOH Model by Ward et al 1997 with these revisions:
TS Poet, P Hinderliter and J Teeguarden,
Center for Biological Monitoring and Modeling 4/16/05
Pacific Northwest National Laboratory
Model contains inhalation, iv, and oral (multiple patterns).
1) Removed fetal compartment and other tissues that could be lumped
based on similarity of partition coefficients or did not need to be
specified directly (Bone, mammary tissue) for the modeling purposes here.
2) Changed day to hr.
3) Flows (scaled to BWor BW**0.75), Metabolism (BW**0.75) and
tissue volumes (BW) are scaled in the model.
Final has somach and intestine compartments which provide fast and
slow absorption rates, respectively.
4) Bladder compartment (for human simulations) added by Paul
Schlosser, U.S. EPA, Oct. 2008
5) "Sipping" drinking water exposure code for rats, to match data
from Pengetal. (1990)
6) Time-variable drinking pattern for mice from Keys et al. (2004)
added by Paul Schlosser, U.S. EPA, Aug. 2009
Version is final version used for simulations
***** MODEL UNITS***
Concentration, mg/L
Mass of Chemical, mg
Volume, L
Flow, L/hr
Body Weight Kg
=72 Character Line=
B-43 DRAFT - DO NOT CITE OR QUOTE
-------
1 Integer IDS, MULIE
2 REAL DRT(6), DRP(6) Istore drink water times, percents in array!
O
4 CONSTANT BW = 0.030 ! Body weight (kg)
5 CONSTANT QPC = 15. ! Alveolar ventilation (L/hr/kg**0.75)
6
7 ! Blood Flows (fraction of cardiac output)
8 CONSTANT QCC=15.0 ! Cardiac output (L/hr/kg**0.75)
9 CONSTANT QFC = 0.05 ! Fat
10 CONSTANT QLC = 0.25 ! Liver
11 ! Blood flow to rest of body Calculated by Flow Balance!
12 QRC= 1.0-(QFC + QLC)
13 QC = QCC*BW**0.75
14 OP = QPC*BW**0.75
15
16 ! Tissue Volumes for mice (fraction of body weight)
17 CONSTANT VAC = 0.0123 ! Arterial blood
18 CONSTANT VFC = 0.07 ! Fat
19 CONSTANT VLC = 0.055 ! Liver
20 CONSTANT VLuC = 0.0073 ! Lung tissue
21 CONSTANT VVBC = 0.0368 ! Venous blood
22 VRC = 0.91 - (VAC+VFC+VLC+VLuC+VVBC)
23
24 ! Partition Coefficients (Mouse values from Ward et al used as default)
25 CONSTANT PB = 1350 ! MeOH Blood:Air; Use Morton value!
26 CONSTANT PF = 0.08 ! MeOH Fat:Blood
27 CONSTANT PL = 1.1 ! MeOH Liver:Blood
28 CONSTANT PLU = 1.0 ! MeOH Lung:Blood, compartment for dosing only
29 CONSTANT PR = 0.8 ! MeOH Rest of body:Blood
30
31 ! Hepatic Metabolism of MeOH
32 CONSTANT KM = 45.0 ! mg/L
33 CONSTANT VMAXC = 15.0 ! mg/hr/BW**0.75
34 VMAX = VMAXC*BW**0.75 ! mg/hr
35 CONSTANT VMAX2C = 15.0! 2nd saturable pathway
36 VMAX2 = VMAX2C*BW**0.75
37 CONSTANT KM2= 45.0
38 CONSTANT KLLC = 0.0 ! First-order metabolism
39 ! Set VMAXC = VMAX2C = 0, when KLLC > 0
40 KLL = KLLC/BW**0.25
41
42 ! MeOH Clearance from Blood!
43 CONSTANT K1C = 0.01 ! First-order clearance, BW**0.25/hr
44 K1 =K1C/BW**0.25 ! Scaled blood elimination, hr-1
45 ! This lumped term was used in the WARD model an accounted for
46 ! renal elimination and "additional" non-hepatic metabolism of
47 ! MeOH associated only with high dose i.v. data.
48 ! A 1st-order term should not be used to represent two processes
49 ! with different dose-dependencies.
50 ! This has not been used for mouse data (set=0), but was used to
51 ! approximate human urinary data!
52
53 ! Bladder compartment added (Paul Schlosser, U.S. EPA, 10/2008)
54 CONSTANT KBL=0.0 ! Bladder constant, 1/hr
55
56 ! Fractional Absorption of MeOH
57 CONSTANT FRACin = 0.85 ! Inhalation, value from Perkins et al
B-44 DRAFT - DO NOT CITE OR QUOTE
-------
1 CONSTANT KFEC = 0.0 ! Fecal elimination constant, 1/hr
2 ! KFEC determines oral bioavailability
O
4 ! Molecular Weight of MeOH
5 CONSTANT MWMe = 32.0 ! mol wt, g/mol!
6
7 ! Closed Chamber Parameters
8 CONSTANT VChC = 100.0 ! Volume of closed chamber (L)
9 CONSTANT Rats = 0.0 ! Number of rats in chamber
10 CONSTANT kLoss = 0.0 ! Chamber loss rate /hr
11 ! Set RATS = 0.0 and KLOSS = 0.0 for open chamber
12
13 ! Blood Flows (L/hr)
14 OF = QFC*QC ! Fat
15 QL = QLC*QC ! Liver
16 OR = QRC*QC ! Rest of Body
17
18 ! Tissue Volumes (mL)
19 VAB = VAC*BW! Arterial blood volume
20 VF = VFC*BW ! Fat
21 VL = VLC*BW ! Liver
22 VLu = VLuC*BW ! Lung
23 VR = VRC*BW ! Rest of the body
24 VVB = VVBC*BW ! Venous blood
25 VBL = VAB + VVB ! Total blood
26
27 ! Timing commands !
28 CONSTANT TCHNG = 6.0 ! End of exposure!
29 CONSTANT TSTOP = 24.0 ! End of experiment/simulation!
30 CONSTANT POINTS = 1000.0 ! No. points for simulation output!
31 CONSTANT REST = 100000.0 ! End of work period for human exercise
32 CONSTANT WORK = 100000.0 ! Start of work period for human exercise
33 SCHEDULE DS1 .AT.REST ! Change from work to rest conditions
34 SCHEDULE DS2.AT.WORK ! Change from rest to work conditions
35 ! Human Rest/Work (changes in blood-flow fractions to fat/liver not currently usedO
36 CONSTANT QPCHR=15.0, QCCHR=15.0, QLCHR=0.25, QFCHR=0.05 ! Rest
37 CONSTANT QPCHW=52.0, QCCHW=26.0, QLCHW=0.16, QFCHW=0.06 ! Work
38
39 | Simulation Control !
40 ! Exposure Conditions Based on User Defined Initial Amounts of
41 ! Chemical (mg)
42 CONSTANT CONCppm = 0.0 ! Air Concentration in ppm
43 VCh = VChC-(Rats*BW) ! Volume of Occupied Chamber
44 CONCmg = CONCppm*MWMe/24451 ! Convert ppm to mg/Liter!
45 ACHO = CONCmg*VCH ! Init Amt in Chamber, mg!
46
47 ! Oral dosing
48 CONSTANT KAS = 0.1 ! 1st order oral abs, hr-1
49 CONSTANT KMASC = 550 ! Saturable oral abs Kmasc [=] mg/kg
50 KMAS = KMASC*BW
51 CONSTANT VASC = 1740 ! Saturable oral ab VmaxC, mg/hr/kgA0.75
52 VAS = VASC*BW**0.75 ! Saturable oral ab Vmax, mg/hr
53 CONSTANT KAI = 0.1 ! 1st order oral abs from intestine, hr-1
54 CONSTANT KSI = 0.5 ! 1st order transfer stom to intes hr-1
55 CONSTANT DOSE = 0.0 ! Oral dose in mg/kg BW
56 CONSTANT ODS = 0.0 ! Switch for zero order oral uptake
57 ! (Set to 1 for zero order, set to 0 for first order)
B-45 DRAFT - DO NOT CITE OR QUOTE
-------
1 ODOSE = DOSE*BW*(1.0-ODS) ! Convert mg/kg to mg total (oral)
2 RAOZ = DOSE*BW*ODS/24.0 ! mg/hr for zero order dosing
O
4 ! Daily dose for steady drinking water by "sipping" (by rats)
5 CONSTANT DWDOSE = 0 ! mg/kg/d by periodic sipping
6 CONSTANT PER1 = 1.5 ! Period between sipping episodes (hr) during dark
7 ! "Between" means from the start of 1 to the start of the next episode
8 CONSTANT DUR1 = 0.75 ! Duration of sipping episodes during dark (hr)
9 CONSTANT PER2 = 3.0 ! Period during light (hr) between sipping episodes
10 CONSTANT DUR2 = 0.5 ! Duration of sipping episodes during light (hr)
11 CONSTANT FNIGHT = 0.8 ! Fraction of drinking during night
12 constant days = 7.0 ! days/week of oral exposure
13 constant metd = 7.0 ! number of days at end over which AUCBF and AMETF
14 ! are averaged
15 tmetf = metd*24.0
16 dayon=24.0*days
17 ! Night sipping rate (mg/h) during episodes
18 DWRNIGHT = DWDOSE*BW*FNIGHT*PER1/(12.0*DUR1)
19 ! Day sipping rate (mg/h) during episodes
20 DWRDAY= DWDOSE*BW*(1-FNIGHT)*PER2/(12.0*DUR2)
21 IDOSE=0
22 ! Above assumes 12-hr each for day/night
23
24 ! Drinking Table from Deborah Keys for mice, as used in
25 ! A quantitative description of suicide inhibition of dichloroacetic acid in rats and mice.
26 ! Keys DA, Schultz IR, Mahle DA, Fisher JW.
27 ! Toxicol Sci. 2004 Dec;82(2):381-93.
28 ! Based on data of Yuan, J. (1993). Modeling blood/plasma concentrations in dosed feed and dosed
29 ! drinking water toxicology studies. Toxicol. Appl. Pharmacol. 119, 131-141.
30 constant rdrink =1.0! Default for use of sipping w/ DWDOSE abopve
31 ! set rdrink = 0.0 to use pattern below
32 table mdrinkp,1,49 / 0., .5, 1., 1.5, 2., 2.5, 3., 3.5, &
33 4., 4.5, 5., 5.5, 6., 6.5, 7., 7.5, &
34 8., 8.5, 9., 9.5, 10., 10.5, 11., 11.5, &
35 12., 12.5, 13., 13.5, 14., 14.5, 15., 15.5, &
36 16., 16.5, 17., 17.5, 18., 18.5, 19., 19.5, &
37 20., 20.5, 21, 21.5, 22., 22.5, 23., 23.5, 24.0, &
38 0.12 , 0.9, 1.6, 1.8, 1.9, 2.9, 4.0, 4.5, 4.9, 4.9, &
39 4.8,4.4,4.0,5.0,5.9,5.3,4.5,3.9, &
40 3.2, 3.0, 2.7, 2.5, 2.3, 2.3, 2.3, 1.9, &
41 1.4, 1.4, 1.3, 1.3, 1.3, 1.1,0.8,0.8, &
42 0.8, 0.6, 0.5, 0.7, 0.8, 0.6, 0.4, 0.2, &
43 0.05, 0.08, 0.14, 0.07, 0.06, 0.08, 0.12 /
44
45 ! Larger bolus dosing
46 CONSTANT DRDOSE=0.0 ! Total dose by drinking water in boluses, mg/kg day
47 ! Times for multiple oral drinks/day *after* 0
48 ! Must be ascending, 0 <= times < 24 hr
49 ! CONSTANT DRT=0, 2,4,6,8,10 ! Rat values
50 Constant DRT = 0.0, 3.0, 5.0, 8.0, 11.0, 15.0 ! Human values
51 ! DRTIME(1) assumed = 0 and not used
52 ! Fraction consumed by drinking at those times
53 CONSTANT DRP = 0.25, 0.1, 0.25, 0.1, 0.25, 0.05
54
55 !Total oral bolus dose; initial value given at t=0 via initial condition
56 TODOSE = DRP(1)*DRDOSE*BW*(1.0-ODS) + ODOSE
57
B-46 DRAFT - DO NOT CITE OR QUOTE
-------
1 ! IV dosing
2 CONSTANT IVDOSE = 0.0 ! IV dose, mg/kg
3 CONSTANT TINF = 0.025 ! Length of exposure (hrs), default = 1.5 min (bolus)
4 ! 1.5 min reported by Ward and Pollack, DMD 1996
5 TIV = IVDOSE*BW ! Expected amt infused, mg
6 IV1 = TIV/TINF ! Rate of infusion, mg/hg
7
8 ! For I.V. Runs, control step size if necessary by changing MaxT, not POINTS or CINT
9 MAXT = 1.0 ! Maximum Step Size, Hours
10 !IF (IVDOSE.GE.1.0E-4) MAXT = 1.0E-4
11
12 ! Liver infusion
13 CONSTANT LIVRO = 0.0 ! Zero-order liver total, mg/kg/day
14 RLIVO = LIVRO*BW/TCHNG ! Rate in mg/hr
15
16 ! Dose Scheduling
17 CONSTANT MULTE=0 ! Default is *no* repeated dosing/inhalation
18 CIZON E = 1.0 ! Start with inhalation on
19 IVZON E = 1.0 ! Start with IV on
20 SCHEDULE OFF.AT.TCHNG ! Turn off exposure at TCHNG
21 DAY = 0;
22 NEWDAY = 0; IDS = 2 ! First dose given as initial condition
23 IF (MULTE) SCHEDULE ORALDOSE.AT.DRT(2)
24 ALGORITHM IALG = 2 ! Gear algorithm
25 END ! END OF INITIAL
26
27 DYNAMIC
28
29 DERIVATIVE
3Q ,.„„„„„„.„.„„„„ MeOH .........................
31 IVR = IVZONE*IV1 ! IV dosing; IVR = ate of infusion, mg/hg
32 ! Oral Dosing
33 OWING = ((DWRNIGHT*PULSE(0.0,PER1,DUR1)*PULSE(0.0,24.0,12.0) + &
34 DWRDAY*PULSE(0.0,PER2,DUR2)*(1-PULSE(0.0,24.0,12.0)) )*rdrink + &
35 (1-rdrink)*mdrinkp(mod(T,24.0))*0.02*DWDOSE*BW)*PULSE(0.0,168,dayon)
36 RAS = KAS*STOM + VAS*STOM/(KMAS+STOM)
37 RSTOM = OWING+ RAOZ-RAS-KSI*STOM ! Change in stomach (mg/hr)
38 RINT = KSI*STOM - RFEC - KAI*AINTEST ! Change in intestines (mg/hr)
39 RLZ = RLIVO*CIZONE ! Zero-order to liver
40 RAO = RAS + KAI*AINTEST + RLZ ! Oral absorption (mg/hr)
41 RFEC = KFEC*AINTEST
42 FEC = INTEG(RFEC, 0.0)
43 STOM = INTEG(RSTOM, TODOSE) ! Amt in stomach (mg)
44 AINTEST = INTEG(RINT, 0.0) ! Amt in intestines (mg)
45 OralDoseCheck = INTEG(RAO, 0.0)
46
47 ! Arterial Blood
48 RAAB = QC*(CVLU - CAB)
49 AAB = INTEG(RAAB, 0.0) ! Amount, mg
50 CAB = AAB/VAB ! Concentration, mg/L
51 AAUCB = INTEG(CAB, 0.0) ! AUC, hr*mg/L
52
53 ! Fat
54 RF = QF*(CAB - CVF)
55 AF = INTEG(RF, 0.0) ! Amount, mg
56 CF = AF/VF ! Concentration, mg/L
57 CVF = CF/PF ! AUC, hr*mg/L
B-47 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 ! Liver
3 RAL = QL*(CAB - CVL) + RAO - RMETL - RMETL2 - RMETL3
4 AL = INTEG(RAL, 0.0) ! Amount, mg
5 CL = AL/VL ! Concentration, mg/L
6 CVL = CL/PL ! Concentration, mg/L
7 AUCL = INTEG(CL, 0.0)! ADC, hr*mg/L
8
9 ! Liver Metabolism
10 RM ETL = VMAX*CVL/(KM + CVL)
11 METL = INTEG(RMETL, 0.0)
12 RMETL2 = VMAX2*CVL/(KM2 + CVL)
13 METL2 = INTEG(RMETL2, 0.0)
14 RMETL3 = KLL*CVL
15 METL3 = INTEG(RMETL3, 0.0)
16
17 ! Total Amount Metabolized (Formate and Formaldehyde)
18 ! Does not include K1C for human MeOH excretion estimate
19 AMET = METL + METL2 + METL3
20 AMET24 = AMET*24.0/TSTOP
21 ! Total amount metabolized in last tmetf hr of exposure, averaged per day
22 AMETF = INTEG((RMETL+RMETL2+RMETL3)*PULSE(TSTOP-tmetf,TSTOP,tmetf),0.0)* &
23 24.0/tmetf ! (tmetf = 24.0*metd)
24 ! Chamber concentration (mg/L)
25 RACh = (Rats*QP*CLEx) - (FRACin*Rats*QP*CCh) - (kLoss*ACh)
26 ACh = INTEG(RACh, AChO)
27
28 ! The following calculation yields an air concentration equal to the
29 ! closed chamber value if a closed chamber run is in place and a
30 ! specified constant air concentration if an open chamber run is in place
31 CCh = ACh*Cizone/VCh
32 CCPPM = CCh*24451/MWMe
33 CLoss = INTEG(kLoss*ACh, 0.0)
34
35 !Lungs
36 RALu = QP*(FRACin*CCh - CLEx) + QC*(CVB - CVLu)
37 ALu = INTEG(RALu, 0.0)
38 CLu = ALu/VLu ! Concentration, mg/L
39 CVLu = CLu/PLu ! Exiting Concentration, mg/L
40
41 ! Amount Inhaled
42 Rlnh = FRACin*QP*CCh
43 Alnh = INTEG(Rlnh, 0.0) ! mg per rat
44 AlnhC = Alnh*Rats ! mg for a group of rats
45
46 ! Amount Exhaled
47 CLEx = CVB/PB ! Concentration, mg/L
48 RAEx = QP*CLEx
49 AEx = INTEG(RAEx, 0.0)*PULSE(0,TCHNG,TSTOP) ! Amount, mg per rat
50 AExC = AEx*Rats ! Amount, mg, for a group of rats
51 AxF = INTEG(RAEx*PULSE(TCHNG,24,24), 0.0) ! Amount exhaled post-exposure
52
53 ! Rest of Body
54 RAR = QR*(CAB - CVR)
55 AR = INTEG(RAR, 0.0) ! Amount, mg
56 CR = AR/VR ! Concentration, mg/L
57 CVR = CR/PR ! Exiting Venous Concentration, mg/L
B-48 DRAFT - DO NOT CITE OR QUOTE
-------
1 AUCR = INTEG(CR, 0.0) ! AUC, hr*mg/L
2
3 ! Venous Blood (mg)
4 RURB = K1*CVB*VVB ! Lumped Clearance from Blood
5 RAVB = QF*CVF + QL*CVL + QR*CVR + IVR - QC*CVB - RURB
6 AVB = INTEG(RAVB, 0.0) ! Amount, mg
7 CVB = AVB / VVB ! Concentration, mg/L
8 AUCB = INTEG(CVB, 0.0) ! AUC, hr*mg/L (total over entire exposure)
9 AUCBB = AUCB*24.0/TSTOP ! Average over exposure, hr*mg/(L*day)
10 AUCBF = INTEG(CVB*PULSE(TSTOP-tmetf,TSTOP,tmetf),0)*24.0/tmetf
11 ! AUCBF = Last tmetf AUC averaged/day (tmetf = 24.0*metd)
12 ! For "steady state" AUC in blood over a day, set exposures to
13 ! several weeks to reach "periodicity", then use AUCBF w/ metd = 7
14
15 ! Bladder compartment, added by PS, U.S. EPA, 10/2008
16 RBL = KBL*ABL ! Rate of clearance from bladder (mg/hr)
17 ABL = INTEG((RURB-RBL),0.0)! Amount in bladder (mg)
18 RUR= RBL/(BW*0.5e-3)! Urine concentration = rate/[BW*(0.5e-3 L/h/kg BW)]
19 URB = INTEG(RBL, 0.0) ! Amount cleared to urine, mg
20 URBF = INTEG(RURB*PULSE(TSTOP-tmetf,TSTOP,tmetf),0)*24.0/tmetf
21 ! Amount cleared to urine in last tmetf averaged/day (tmetf = 24.0*metd)
22
23 |************************* Magg Balance **************************
24 Tbody = AAB + AF + AL + ALU + AR + AVB + ABL + STOM + AINTEST
25 MetabORCIrd = METB + METL + METL2 + METL3 + AEX + FEC
26 TMass = Tbody + MetabORCIrd
27 TDose = AinH + INTEG(IVR+DWING+RAOZ+RLZ,0.0) + TODOSE
28 MassBal=100*(TDose-TMass)/(TMass+1e-12)
29 Icompare to TIV, ODOSE, or AINHC
30 ! Check Blood Flows
31 QTOT = QF + QL + QR
32 QRECOV= 100.0*QTOT/QC
33 END ! End of Derivative
34 TERMT(T.GE.TStop)
35
36 ! Exposure Control
37 DISCRETE ORALDOSE ! Stom is amount in stomach
38 IDOSE = DRP(IDS)*DRDOSE*BW
39 STOM = STOM + IDOSE ! Drinking percent
40 TODOSE = TODOSE + IDOSE
41 IF(IDS.EQ.1)THEN
42 STOM = STOM + ODOSE
43 TODOSE = TODOSE + ODOSE
44 ENDIF
45 IDS = IDS+1
46 IF(IDS.EQ.7)THEN ! For 6 doses
47 IDS = 1
48 N EWDAY = N EWDAY + 24
49 SCHEDULE ORALDOSE.AT.NEWDAY ! Go to start of the next day
50 ELSE
51 SCHEDULE ORALDOSE.AT.(NEWDAY+DRT(IDS)) ! Go to next drink time
52 ENDIF
53 END ! OF DISCRETE ORALDOSE
54
55 DISCRETE OFF ! Turn IN HAL exposure off
56 CIZONE = 0.0
57 IVZONE = 0.0
B-49 DRAFT - DO NOT CITE OR QUOTE
-------
1 DAY=DAY+1
2 IF (MULTE) SCHEDULE ON.AT.(DAY*24.0)
3 END ! OF DISCRETE OFF
4
5 DISCRETE ON
6 CIZONE=1.0
7 SCHEDULE OFF.AT.(T+TCHNG)
8 END ! OF DISCRETE ON
9
10 DISCRETE DS1 ! Human at rest
11 ! Equations scheduled for change during simulation repeated here
12 QC = QCCHR*BW**0.75
13 OP = QPCHR*BW**0.75
14 OF = QFC*QC ! QFCHR*QC ! Equations for alternate flow fractions
15 QL = QLC*QC ! QLCH R*QC ! But QFC and QLC taken to be 'at rest' values
16 QRC= 1.0-(QFC + QLC)
17 OR = QRC*QC
18 END ! OF DISCRETE DS1
19
20 DISCRETE DS2 ! Human at work (50VV)
21 ! Equations scheduled for change during simulation repeated here
22 QC = QCCHW*BW**0.75
23 OP = QPCHW*BW**0.75
24 OF = QFC*QC ! QFCHW*QC ! Equations for alternate flow fractions
25 QL = QLC*QC ! QLCHW*QC ! But don't seem to work (fit data) well
26 QRC= 1.0-(QFC + QLC)
27 QR = QRC*QC
28 END ! OF DISCRETE DS2
29
30 END ! End of Dynamic
31 END ! End of Program
B.3.3. acslXtreme procedure (.cmd) file
File MEOHCBMMfinal.CMD - FOR PBPK MODEL FOR METHANOL
taken from .cmd file from Ward et al., Edited by KWW - 06/02/96
Developed for this (CBMM) model - 4/15/15
Final with Digitized Data - 5/25/05
Final Version has fast and slow rates of oral absorption
Version 4 is final version used for simulations
Final Version 1.10.06
Beyond this comment, this file is left "as is" for archival purposes. But most if not all of
the functions and data sets defined here are replicated and/or replaced in the .m files below.
Only use these when there is no corresponding .m file.
- Paul Schlosser, U.S. EPA, Oct. 2008
32
33
34
35
36
37
38
39
40
41
42
43
44 PREPARE T,CVB,MetB
45
46 ! Procedural blocks for general mouse/rat data
47 PROCED CDMICE ! Anatomic/physiologic data for mice
48 SETBW=0.03, TSTOP=1.5
49 SET IVDose=0, DOSE=0, CONCppm=0
50 SET PL=1.06, PF=0.083, PR=0.66, PB=1350
B-50 DRAFT - DO NOT CITE OR QUOTE
-------
1 SET QPC=25.4,QCC=25.4,fracin=0.73
2 SET QLC = 0.25,QFC=0.05
3 SET KM=12,VmaxC=14.3,KLC=0.0,KAS=2
4 SET Vmax2c=19,km2=210,KAI=0.22,KSI=l.l
5 SET VAC = 0.0123,VFC = 0.07,VLC = 0.055
6 SET VLuC = 0.0073, VVBC = 0.0368
7 ! Volumes from Brown et al
8 !Mouse QPC avg from Brown 29, 24 used in Corley et al and others
9 ! AVG of measured vent rates by Perkins et al 25.4 L/hr/kgA0.75
10 IBlood volume 4.9% total. As per Brown 25:75 split art:ven
11 IMetab originally from Ward et al - KldC for mice =0
12 END
13
14 PROCED HUMAN
15 SETBW=70
16 SET IVDose=0, DOSE=0, CONCppm=0
17 SET PL=1.06, PR=0.66, fracin=0.75
18 SET VFC=0.214, VLC=0.026,VLUC=0.008
19 SETVAC=0.0198,VVBC=0.0593
20 SET QPC=18.5, QCC=18.5, QLC=0.227, QFC=0.052
21 SET KM=12,VmaxC=l 1,KLC=0.044,KAS=2.0
22 SET KAI=0.22,KSI=1.1
23 SET PB = 1626, PF=0.14
24 SET Vmax2c=0
25 ! Volumes from Brown et al
26 ! QPC from Brown, upper end 13.4 L/hr/kgA0.75
27 INeed higher for data, 15 L/hr/kgA0.75 used in several published human models
28 IBlood volume 7.9% total. As per Brown 25:75 split art:ven
29 IFrac absorbed from Ernstgard SOT poster + personal communication
30 !Human Partition Coef. equal to mice. Horton et al. used rat
31 lExcept Human Partition Coef blood and fat - from Fiserova-Bergerova and Diaz, 1986
32 ! - but rat values are inconsistent with expected fat partitioning for an alcohol like this
33 ! - for example Pastino and Conolly EtOH model, fat PC =0.1
34 END
35
36 PROCED SDRAT ! Anatomic/physiologic data for rats
37 SETBW=0.3, TSTOP=1.5
38 SET IVDose=0, DOSE=0, CONCppm=0
39 SETPL=1.6, PF=0.1,PR=1.3
40 SET KM=45,VmaxC=15,KLC=0.1,KAS=5
41 SET VAC = 0.0185, VFC=0.07, VLC= 0.034, VLuC=0.005, VVBC=0.0555
42 ! Volumes from Brown et al
43 !PC from horton et al., PF reduced to 0.1 from Horton's 1.1
44 IBlood volume 7.4% total. As per Brown 25:75 split art:ven
45 IMetab originally from Ward et al - KldC for mice =0
46 I Rat model not calibrated
47 END
48
49 PROCED PREG
B-51 DRAFT - DO NOT CITE OR QUOTE
-------
1 !For GD 18 mice, BW increased as estimated from Roger's et al
2 ! Increased VFC as per Corley CRT development review
3 ! This just to give a WAG as to how data might change from BW and different volume of
4 distribution
5 !Not invoked for any PROCs below as the default
6 ! Liver to 140% of NP
7 SET BW = 0.055, VFC=0.08,VLC=0.11,VVBC=0.05
8 END
9
10 PROCED CLEARIT
11 SET IVDose=0, DOSE=0, CONCppm=0
12 END
13
14 PROCED SHOWIT
15 display Vmaxc,km,klc,pb,pf,pr,pl,kas,fracin
16 END
17
18 [Procedural blocks for all non-pregnant mouse data
19 ! IV
20 PROCED MWARDIV25
21 !Ward etal., TAP 1997
22 ! Figure 2, data from Ward model cmd
23 IData was checked via digitizit - within +1-5% of cmd file
24 CLEARIT
25 CDMICE
26 SET TSTOP=24.0
27 SET IVDOSE=2500., tchng=0.025
28 END
29
30 PROCED PMWARDIV25
31 PLOT /D=MWARDIV25, CVB
32 END
33
34 DATA MWARDIV25(T,CVB)
35 0.08 4481.8
36 0.25 4132.2
37 0.5 3888
38 1.00 3164.8
39 2.0 2303.5
40 4.00 1921.5
41 6 1883.8
42 8 1620
43 12 838
44 18 454.7
45 24 NaN
46 END
47
48 PROCED MWARD95IV25
49 ! Ward etal., FAT 1995
B-52 DRAFT - DO NOT CITE OR QUOTE
-------
0.53
1.06
1.54
3.07
4.07
5.02
6.02
7.02
8.02
9.03
10.03
END
3299.60
3244.54
3190.71
2803.13
2544.36
2237.77
2063.59
1873.10
1521.92
1670.30
1423.12
1 ! Figure 2
2 !Data via digitizit
3 CLEARIT
4 CDMICE
5 SET TSTOP=24.0
6 SET IVDOSE=2500., tchng=0.025
7 END
8
9 PROCED PMWARD95IV25
10 PLOT /D=MWARD95IV25, CVB
11 END
12
13 DATA MWARD95IV25(T,CVB)
14
15
16
17
18
19
20
21
22
23
24
25
26
27 ! Procs for pegnant IV below: MWARDGD9IV25, MWARDGD18IV25, MWARDGD18IV5,
28 MWARDGD18IV1
29 ! Oral
30 PROCED MWARDPO25
31 !Ward etal., FAT 1995
32 ! Figure 2, data from Ward model cmd
33 IData was checked via digitizit - within +1-5% of cmd file
34 CLEARIT
35 CDMICE
36 SET TSTOP=24, DOSE=2500
37 END
38
39 PROCED PMWARDPO25
40 PLOT /D=MWARDPO25, CVB
41 END
42
43 DATA MWARDPO25(T,CVB)
44 0.504 2370
45 0.96 2645
46 1.44 2705
47 1.992 2719
48 2.208 2781
49 3 2704
B-53 DRAFT - DO NOT CITE OR QUOTE
-------
1 4.008 2370
2 4.992 2617
3 6 2516
4 7.008 2635
5 7.992 2213
6 9 2370
7 10.008 2028
8 10.992 1916
9 12 1347
10 13.008 1467
11 13.9921354
12 15 1175
13 16.008 864.3
14 16.992745.2
15 18 422.4
16 19.01 428
17 21 243
18 24 136
19 END
20
21 IProcs for pregnant Oral below: MDORGD8PO15, MWARDGD18PO25
22 ! Inhalation
23 ! QPC set to measured as in Perkins et al., FAT, 1995 for each concentration
24
25 PROCED MPERKIN25
26 !Perkins etal., FAT, 1995
27 !Fig. 2 data in Ward cmd file
28 CLEARIT
29 CDMICE
30 SET TSTOP=24, CONCppm=2500, vchc=5000
31 SET QPC = 29., QCC=29.
32 SET TCHNG=8
33 END
34
35 PROCED PMPERKIN25
36 PLOT /D=MPERKIN25, CVB
37 END
38
39 IThis data from Digitizlt
40 DATA MPERKIN25(T,CVB)
41 2.0 414.0
42 4.0 453.0
43 6.0 586.0
44 8.25 694.0
45 12 282.0
46 16 0.6
47 END
48
49 IThis data from cmd file
B-54 DRAFT - DO NOT CITE OR QUOTE
-------
1 !DATAMPERKIN25(T,CVB)
2 11.99 386.49
3 14.01 617.57
4 16.00 816.22
5 18.26 970.27
6 112.00 393.24
7 116.0 13.51
8 END
9
10 PROCED MPERKIN50
11 !Perkins et al., FAT, 1995
12 !Fig. 2, data in Ward cmd file
13 IData in command file higher than appears in figure
14 CLEARIT
15 CDMICE
16 SET TSTOP=24, CONCppm=5000, vchc=5000
17 SET TCHNG=8, qpc=24.,qcc=24.
18 END
19
20 PROCED PMPERKIN50
21 PLOT /D=MPERKIN5 0, CVB
22 END
23
24 Ithis from Digitizit, Fig 2 Perkins et al
25 DATA MPERKIN50(T,CVB)
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
1
2
O
4
6
8.25
12
16
20
24
END
IThis
644.00
877.00
1340.00
1450.00
2040.00
2290.0
1410.0
583.0
271.0
9.7
data from cmd file
IDATA MPERKIN50(T,CVB)
!1.0
!2.0
!3.0
!4.0
!6.0
!8.3
112.0
116.0
120.0
124.0
906.76
1202.7
1828.38
1986.49
2800
3125.68
1914.86
806.76
367.57
10.81
B-55 DRAFT - DO NOT CITE OR QUOTE
-------
1 !END
2
3 PROCEDMPERKIN100
4 !Perkins etal., FAT, 1995
5 !Fig. 2 data in Ward cmd file
6 INote, Table 6 in Ward paper - max value of 3260 +/- 151
7 CLEARIT
8 CDMICE
9 SET TCHNG=8, CONCppm= 1,0000, tstop=36,vchc=5000
10 SET QPC=21,qcc=21
11 END
12
13 PROCED PMPERKIN100
14 PLOT /D=MPERKIN 100, CVB
15 END
16
17 Ithis from Digitizit, Fig 2 Perkins et al
18 DATAMPERKIN100(T,CVB)
19 2.0 2080.0
20 4.0 2530.0
21 6.0 3350.0
22 8.25 3350.0
23 12 2370.0
24 16 1830.0
25 20 1080.0
26 24 591.0
27 28 44.6
28 END
29
30 !DATAMPERKIN100(T,CVB)
31 ! This from original cmd file
32 !2.0 2809.46
33 !4.0 3405.4
34 !6.0 4528.38
35 !8.3 4524.32
36 112.0 3212.16
37 116.0 2456.76
38 120.0 1439.19
39 124.0 798.65
40 128.0 55.4
41 !END
42
43 ! Procs for Preg mouse Inhalaiton date below: MDOR8IN10,MDOR8IN15
44 !and:MROGGD7IN10, MROGGD6IN1, MROGGD6IN2, MROGGD6IN5,
45 MROGGD6IN10
46
47 ! pregnant mice
48 ! IV
49 PROCED MWARDGD9IV25
B-56 DRAFT - DO NOT CITE OR QUOTE
-------
1 !Ward etal.,DMD, 1996
2 !Not used in the manuscript, only in cmd file
3 CLEARIT
4 CDMICE
5 SET TSTOP=24
6 SET IVDOSE=2500., TINF=0.025
7 END
8
9 PROCED PMWARDGD9IV25
10 PLOT /D=MWARDGD9IV25, CVB
11 END
12
13 DATA MWARDGD9IV25(T,CVB)
14 0.0833 4606.2
15 0.25 4079.5
16 0.5 3489.3
17 1 2939.6
18 2 3447.6
19 4 2605.0
20 6 2690.5
21 8 2574.9
22 12 1506.1
23 18 498.6
24 24. NaN
25 END
26
27 PROCED PROCED MWARDGD18IV25
28 !Ward etal.,DMD, 1996
29 !Note, Table 6 in Ward paper - max value of 3521+/- 492
30 CLEARIT
31 CDMICE
32 SET TSTOP=24
33 SET IVDOSE=2500., TINF=0.025
34 END
35
36 PROCED PMWARDGD18IV25
37 PLOT /D=MWARDGD18IV25, CVB
38 END
39
40 DATAMWARDGD18IV25(T,CVB)
41 0.0833 4250.0
42 0.25 3445.1
43 0.5 2936.8
44 1.0 2470.5
45 2.0 2528.1
46 4.0 2292.3
47 6.0 2269.4
48 8.0 2057.0
49 12 1805.9
B-57 DRAFT - DO NOT CITE OR QUOTE
-------
1 18 1482.2
2 24.0 496.1
3 END
4
5 PROCED MWARDGD18IV5
6 !Ward etal.,DMD, 1996
7 INote, Table 6 in Ward paper - max value of 868.8 +/- 53.9
8 CLEARIT
9 CDMICE
10 SET TSTOP=6
11 SET IVDOSE=500., TINF=0.025
12 END
13
14 PROCED PMWARDGD181V5
15 PLOT /D=MWARDGD18IV5, CVB
16 END
17
18 DATA MWARDGD 18IV5(T,CVB)
19 0.25 854.7
20 0.5 720.2
21 1.0 624.1
22 2.0 453.2
23 3.0 307.6
24 4.0 217.7
25 4.5 202.6
26 END
27
28 PROCED MWARDGD 18IV1
29 !Ward etal.,DMD, 1996
30 ! Ward Proc GD8, but must be 18 as per PBPK manuscript
31 INote, Table 6 in Ward paper - max value of 252 +/- 12.9
32 liable matches file
33 CLEARIT
34 CDMICE
35 SET TSTOP=4
36 SETIVDOSE=100.
37
3 8 PROCED PMWARDGD 18IV1
39 PLOT /D=MWARDGD18IV1, CVB
40 END
41
42 DATAMWARDGD18IV1(T,CVB)
43 0.25 252
44 0.52 242.2
45 1.0 222.7
46 2 176.4
47 3 134.2
48 3.5 94.41
49 END
B-58 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 ! Oral
3
4 PROCED MDORGD8PO15
5 IWard et al., cmd file
6 INote, Table 6 in Ward paper - max value of 1610 +/- 704
7 ! Table and file match w/in round off
8 I Data must be from Dorman
9 IDorman Teratology, 1995, Fig. 1
10 ! within error for Digitiz data the same
11 CLEARIT
12 CDMICE
13 SETTSTOP=24, DOSE=1500
14 END
15
16 PROCED PMDORGD8PO15
17 PLOT /D=MDORGD8PO15, CVB
18 END
19
20 DATAMDORGD8PO15(T,CVB)
21 1 1609.6
22 2 1331.2
23 4 1241.6
24 8 707.2
25 16 160.0
26 24 38.4
27 END
28
29 PROCED MWARDGD18PO25
30 !Ward etal.,DMD, 1996
31 INote, Table 6 in Ward paper - max value of 3205 +/- 291
32 CLEARIT
33 CDMICE
34 SET TSTOP=24, DOSE=2500
35 END
36
37 PROCED PMWARDGD18PO25
38 PLOT /D=MWARDGD18PO25, CVB
39 END
40
41 I from cmd file, replaced with digitized
42 IDATA MWARDGD 18PO25(T,CVB)
43 10.25 2770.
44 !0.5 3299.
45 !1 3336.
46 12 3502.
47 14 3217.
48 16 2999.
49 !10 2036.
B-59 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
!12 1832.
!15 949.1
!18 403.5
!21 40.47
!24. 16.03
!END
IDigitizit data
DATA MWARDGD18PO25(T,CVB)
0.5 2024
1 2554
2 3193
4 3002
6 2933
10 1976
12 1922
15 1339
18 1033
21 832
24 580
END
! Inhalation
PROCED MDOR8IN10
! Ward etal., TAP 1997
INote, Table 6 in Ward paper - max value
!Fig 7? Table 6 attributes to Dorman
of 2080
IDigitizit of Dorman Fig 2 matches cmd file
! actual exposure ppm 9900
CLEARIT
CDMICE
SET TCHNG=6, CONCppm=9900, tstop=36
END
PROCED PMDOR8IN10
PLOT /D=MDOR8IN10, CVB
END
DATA MDOR8IN10(T,CVB)
1 771.2
2 1017.6
4 1788.8
6 2076.8
8 2281.6
16 1152.0
24 268.8
END
/-800
B-60 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
4
6
8
16
24
END
1475.2
2486.4
4588.8
7123.2
5888.0
3456.0
1446.4
1 PROCEDMDOR8IN15
2 ! Ward etal., TAP 1997
3 INote, Table 6 in Ward paper - max value of 7136 +/- 736
4 !Fig 7? Table 6 attributes to Dorman
5 IDigitizit of Dorman Fig 2 matches cmd file
6 CLEARIT
7 CDMICE
8 SET TCHNG=6, CONCppm=15000, tstop=36
9 SET vchc=5000000000
10 END
11
12 PROCEDPMDOR8IN15
13 PLOT /D=MDOR8IN15, CVB
14 END
15
16 DATAMDOR8IN15(T,CVB)
17
18
19
20
21
22
23
24
25
26 !Files above provided in cmd file from Ward PBPK model, TAP 1997
27 IFiles below added for this evaluation,
28 ! sources described in proc files and in notebook
29
30 PROCEDMROGGD7IN10
31 ! Rogers et al., Teratology, 1997
32 ! Actual Values kindly Provided by Rogers
33 CLEARIT
34 CDMICE
35 SET TCHNG=7, CONCppm= 1,0000, tstop=36
36 SET vchc=500000000,bw=0.032
37 END
38
39 PROCEDPMROGGD7IN10
40 PLOT /D=MROGGD7IN10, CVB
41 END
42
43 DATA MROGGD7IN10(T,CVB)
44 1 930
45 4 2800
46 6 3360
47 7 3990
48 7.5 3980
49 8 4120
B-61 DRAFT - DO NOT CITE OR QUOTE
-------
1 9 3270
2 12 2630
3 16 1690
4 26 60
5 END
6
7 PROCED MROGGD6IN1
8 CLEARIT
9 IRogers et al., Teratology, 1993
10 IRogers data from GD 6 and 10
11 I In Table 2
12 CDMICE
13 SET TCHNG=7, CONCppm= 1,000, tstop=36,vchc=500000000,bw=0.032
14 END
15
16 PROCED PMROGGD6IN1
17 PLOT /D=MROGGD6IN1, CVB
18 END
19
20 DATA MROGGD6IN1 (T,CVB)
21 7 63
22 7 131
23 END
24
25
26 PROCED MROGGD6IN2
27 I Rogers et al., Teratology, 1993
28 IRogers data from GD 6 and 10
29 I In Table 2
30 CLEARIT
31 CDMICE
32 SET TCHNG=7, CONCppm=2000, tstop=36, vchc=500000000,bw=0.032
33 END
34
35 PROCED PMROGGD6IN2
36 PLOT /D=MROGGD6IN2, CVB
37 END
38
39 DATA MROGGD6IN2(T,CVB)
40 7 487
41 7 641
42 END
43
44 PROCED MROGGD6IN5
45 I Rogers et al., Teratology, 1993
46 IRogers data from GD 6 and 10
47 I In Table 2
48 CLEARIT
49 CDMICE
B-62 DRAFT - DO NOT CITE OR QUOTE
-------
1 SET TCHNG=7, CONCppm=5000, tstop=36,vchc=500000000,bw=0.032
2 END
3
4 PROCED PMROGGD6IN5
5 PLOT /D=MROGGD6IN5, CVB
6 END
7
8 DATAMROGGD6IN5(T,CVB)
9 7 2126
10 7 1593
11 END
12
13 PROCED MROGGD6IN10
14 ! Rogers et al., Teratology, 1993
15 ! Rogers data from GD 6, 10, 15
16 ! In Table 2
17 CLEARIT
18 CDMICE
19 SET TCHNG=7, CONCppm= 1,0000, tstop=36,vchc=500000000,bw=0.032
20 END
21
22 PROCED PMROGGD6IN10
23 PLOT /D=MROGGD6IN10, CVB
24 END
25
26 DATA MROGGD6IN10(T,CVB)
27 7 4653
28 7 4304
29 7 3655
30 END
31
32 IHuman inhalation dta
33
34 PROCED HJOHIN1
35 lErnstgard et al. SOT poster 200 ppm human
36 ! Digitized from Fig 2
37 ! Also personal communication - Ernstgard
38 !QPC from Johanson et al. Scand J. Work Env. 86 =52.6
39 !If Assume value = alveolar, similar to Astrand '83 value of 56 L/hr/krA0.75
40 IFracin - 50% of total (from poster) -76%
41 !QCC from Corley et al TAP 129, 1994
42 CLEARIT
43 HUMAN
44 SET TCHNG=2, CONCppm= 100, tstop= 16
45 SET QPC=52.6,qcc=26,vchc=500000000
46 END
47
48 PROCED PHJOHIN1
49 PLOT /D=HJOHIN1, CVB
B-63 DRAFT - DO NOT CITE OR QUOTE
-------
1 END
2
3 DATAHJOHIN1(T,CVB)
4 0.20 0.87
5 0.46 1.50
6 0.97 2.31
7 1.46 3.24
8 1.91 3.65
9 2.17 3.52
10 2.50 2.55
11 2.91 2.23
12 3.51 1.59
13 4.01 1.72
14 5.02 0.41
15 6.00 0.50
16 9.24 0.12
17 END
18
19 PROCED HJOHIN2
20 lErnstgard et al. SOT poster 200 ppm human
21 ! Digitized from Fig 2
22 ! Also personal communication - Ernstgard
23 !QPC from Johanson et al. Scand J. Work Env. 86 =52.6
24 ! If Assume value = alveolar, similar to Astrand '83 value of 56 L/hr/krA0.75
25 IFracin - 50% of total (from poster) -75%
26 !QCC from Corley et al TAP 129, 1994
27 CLEARIT
28 HUMAN
29 SET TCHNG=2, CONCppm=200, tstop= 16
30 SET QPC=52.6,qcc=26,vchc=500000000
31 END
32
33 PROCED PHJOHIN2
34 PLOT /D=HJOHIN2, CVB
35 END
36
37 DATA HJOHIN2(T,CVB)
38 0.22 1.63
39 0.49 2.92
40 0.92 4.76
41 1.47 6.30
42 1.90 7.65
43 2.16 6.20
44 2.47 5.49
45 2.91 4.96
46 3.50 3.64
47 4.00 3.43
48 4.99 1.94
49 5.97 1.03
B-64 DRAFT - DO NOT CITE OR QUOTE
-------
1 8.90 0.21
2 END
3
4 PROCED HOSTERIN2
5 ! Osterloh et al., JOEM 1996
6 ! Digitized data provided by EPA
7 ! Subtracted background from exposure blood levels
8 CLEARIT
9 HUMAN
10 SET TCHNG=4, CONCppm=200, tstop= 16
11 SET vchc=500000000, BW=78.2
12 END
13
14 PROCED PHOSTERIN2
15 PLOT/D=HOSTERIN2, CVB
16 END
17
18 DATAHOSTERIN2(T,CVB)
19 0.05 0.54
20 0.25 1.39
21 0.50 1.82
22 0.75 2.28
23 1.00 2.42
24 1.50 2.94
25 2.00 3.37
26 2.50 3.90
27 3.00 4.21
28 3.50 4.61
29 4.00 4.82
30 5.00 2.99
31 6.00 2.30
32 7.00 1.40
33 7.95 1.07
34 END
35
36 PROCED HBATIN82
37 IBatterman et al., Int Arch Occ Health 1998
38 ! Digitized Data
39 CLEARIT
40 HUMAN
41 SET TCHNG=2, CONCppm=800, tstop= 16
42 SET vchc=500000000
43 END
44
45 PROCED PHBATIN82
46 PLOT /D=HBATIN82, CVB
47 END
48
49 DATA HBATIN82(T,CVB)
B-65 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
2.223 13.658
2.495 13.282
2.742 11.928
3.230 9.456
4.231 6.197
5.247 3.953
6.262 2.325
7.251 1.551
8.216 1.176
END
PROCEDHBATIN81
IBatterman et al., Int Arch Occ Health
! Digitized Data
CLEARIT
HUMAN
1998
SET TCHNG=1, CONCppm=800, tstop=16
SET vchc=500000000
END
PROCEDPHBATIN81
PLOT /D=HBATIN81, CVB
END
DATA HBATIN81(T,CVB)
1.096 6.477
1.398 6.136
1.644 5.345
2.143 4.270
3.178 2.661
4.188 1.307
5.199 0.732
6.266 0.552
7.292 0.356
8.209 0.093
END
PROCED HBATIN830
IBatterman et al., Int Arch Occ Health
! Digitized Data
Ibody weight not provided
CLEARIT
HUMAN
1998
SET TCHNG=0.5, CONCppm=800, tstop=16
SET vchc=500000000
END
PROCED PHBATIN830
PLOT /D=HBATIN830, CVB
B-66 DRAFT - DO NOT CITE OR QUOTE
-------
1 END
2
3 DATAHBATIN830(T,CVB)
4 0.579 4.608
5 0.857 4.685
6 1.137 4.870
7 1.650 3.452
8 2.650 2.082
9 3.662 0.910
10 4.693 0.316
11 5.713 0.320
12 6.643 0.292
13 7.696 0.547
14 END
15
16 PROCED HSEDIN231
17 ! Sedivec et al., Int Arch Occ Health 1981
18 ! Digitized Data
19 INote, urine volumes not given, these are estimates
20 lurine production of 0.75 mg/hr, this for info purposes only!!!
21 CLEARIT
22 HUMAN
23 SET TCHNG=8, CONCppm=231, tstop=24
24 SET vchc=500000000
25 END
26
27 PROCED PHSEDIN231
28 PLOT /D=HSEDIN231, Metb
29 END
30
31 DATA HSEDIN231 (T,Metb)
32 0.043 0.0042
33 2.174 0.33
34 4.478 0.87
35 6.478 1.46
36 8.522 2.15
37 10.3482.63
38 12.1302.91
39 14.0443.07
40 18.8703.32
41 23.6963.52
42 END
43
44 PROCED HSEDIN157
45 ! Sedivec et al., Int Arch Occ Health 1981
46 ! Digitized Data
47 INote, urine volumes not given, these are estimates
48 lurine production of 0.75 mg/hr, this for info purposes only 11ICLEARIT
49 HUMAN
B-67 DRAFT - DO NOT CITE OR QUOTE
-------
1 SET TCHNG=8, CONCppm=157, tstop=24
2 SET vchc=500000000
3 END
4
5 PROCEDPHSEDIN157
6 PLOT /D=HSEDIN157, Metb
7 END
8
9 DATA HSEDIN157(T,Metb)
10 0.126 0.0038
11 2.204 0.228
12 4.242 0.576
13 6.196 0.975
14 8.326 1.47
15 10.163 1.81
16 12.0942.00
17 14.0162.12
18 18.8966 2.34
19 23.7762.53
20 END
21
22 PROCED HSEDIN78
23 ! Sedivec et al., Int Arch Occ Health 1981
24 ! Digitized Data
25 INote, urine volumes not given, these are estimates
26 lurine production of 0.75 mg/hr, this for info purposes only!!!
27 CLEARIT
28 HUMAN
29 SET TCHNG=8, CONCppm=78, tstop=24
30 SET vchc=500000000
31 END
32
33 PROCED PHSEDIN78
34 PLOT /D=HSEDIN78, Metb
35 END
36
37 DATA HSEDIN78(T,Metb)
38 0.03 0.013
39 2.06 0.189
40 3.96 0.397
41 6.09 0.652
42 8.09 0.820
43 10.11 0.933
44 11.93 1.02
45 13.92 1.09
46 18.89 1.27
47 END
48
49 ! AUC, Cmax estimation procedures
B-68 DRAFT - DO NOT CITE OR QUOTE
-------
1 Proced mousin
2 !To determine AUC for 7 hr exposure in mice
3 CLEARIT
4 CDMICE
5 SET TCHNG=7, tstop=24
6 SET vchc=50000000000
7 SET CONCppm=l
8 start /nc
9 d concppm,AUCB,amet,cvb
10 SET CONCppm=5
11 start /nc
12 d concppm,AUCB,amet,cvb
13 SET CONCppm=10
14 start/nc
15 d concppm,AUCB,amet,cvb
16 SET CONCppm=25
17 start/nc
18 d concppm,AUCB,amet,cvb
19 SET CONCppm=50
20 start /nc
21 d concppm,AUCB,amet,cvb
22 SET CONCppm=75
23 start /nc
24 d concppm,AUCB,amet,cvb
25 SET CONCppm=100
26 start /nc
27 d concppm,AUCB,amet,cvb
28 SET CONCppm=175
29 start /nc
30 d concppm,AUCB,amet,cvb
31 SET CONCppm=208.3
32 start /nc
33 d concppm,AUCB,amet,cvb
34 SET CONCppm=250
35 start/nc
36 d concppm,AUCB,amet,cvb
37 SET CONCppm=325
38 start/nc
39 d concppm,AUCB,amet,cvb
40 SET CONCppm=500
41 start/nc
42 d concppm,AUCB,amet,cvb
43 SET CONCppm=750
44 start /nc
45 d concppm,AUCB,amet,cvb
46 SET CONCppm=l,000
47 start /nc
48 d concppm,AUCB,amet,cvb
49 SET CONCppm=2000
B-69 DRAFT - DO NOT CITE OR QUOTE
-------
1 start /nc
2 d concppm,AUCB,amet,cvb
3 SET CONCppm=2500
4 start /nc
5 d concppm,AUCB,amet,cvb
6 SET CONCppm=5000
7 start /nc
8 d concppm,AUCB,amet,cvb
9 SET CONCppm=l,0000
10 start/nc
11 d concppm,AUCB,amet,cvb
12 SET CONCppm=50000
13 start/nc
14 d concppm,AUCB,amet,cvb
15 END
16
17 Proced mousinC
18 !To determine 7 hr Cmax, note - not at SS
19 CLEARIT
20 CDMICE
21 SET TCHNG=7, tstop=7,VCHC=50000000000
22 SET CONCppm=l
23 start /nc
24 d cone ppm,cvb
25 SET CONCppm=10
26 start /nc
27 d concppm,CVB
28 SET CONCppm=50
29 start /nc
30 d concppm,CVB
31 SET CONCppm=100
32 start /nc
33 d concppm,CVB
34 SET CONCppm=250
35 start/nc
36 d concppm,CVB
37 SET CONCppm=500
38 start/nc
39 d concppm,CVB
40 SET CONCppm=l,000
41 start/nc
42 d concppm,CVB
43 SET CONCppm=2000
44 start /nc
45 d concppm,CVB
46 SET CONCppm=2500
47 start /nc
48 d concppm,CVB
49 SET CONCppm=5000
B-70 DRAFT - DO NOT CITE OR QUOTE
-------
1 start /nc
2 d concppm,CVB
3 SET CONCppm=l,0000
4 start /nc
5 d concppm,CVB
6 SET CONCppm=50000
7 start /nc
8 d concppm,CVB
9 END
10
11 Proced humin
12 ! To determine 24 hr AUC,Cmax at SS for human
13 CLEARIT
14 human
15 SET TCHNG=360, tstop=l,000
16 SET vchc=5000000000, points=48
17 SET Concppm=l
18 Start/nc
19 d concppm,aucBb,cvb
20 SET CONCppm=10
21 start/nc
22 d concppm,aucBb,cvb
23 SET CONCppm=50
24 start /nc
25 d concppm,aucBb,cvb
26 SET CONCppm=100
27 start /nc
28 d concppm,aucBb,cvb
29 SET CONCppm=250
30 start/nc
31 d concppm,aucBb,cvb
32 SET CONCppm=500
33 start/nc
34 d concppm,aucBb,cvb
35 SET CONCppm=625
36 start/nc
37 d concppm,aucBb,cvb
38 SET CONCppm=750
39 start/nc
40 d concppm,aucBb,cvb
41 SET CONCppm=875
42 start /nc
43 d concppm,aucBb,cvb
44 SET CONCppm=l,000
45 start /nc
46 d concppm,aucBb,cvb
47 SET CONCppm=2000
48 start /nc
49 d concppm,aucBb,cvb
B-71 DRAFT - DO NOT CITE OR QUOTE
-------
1 SET CONCppm=2500
2 start /nc
3 d concppm,aucBb,cvb
4 SET CONCppm=5000
5 start /nc
6 d concppm,aucBb,cvb
7 SET CONCppm=l,0000
8 start /nc
9 d concppm,aucBb,cvb
10 SET CONCppm=50000
11 start /nc
12 d concppm,aucBb,cvb
13 END
14
15 Proced humor
16 ! To determine 24 hr AUC
17 ! Oral exposure
18 CLEARIT
19 human
20 SET TCHNG= 1,000, tstop= 1,000
21 SET vchc=5000000000, points=48
22 SETdose=0.1
23 SET ODS=1
24 start /nc
25 d dose,aucBb
26 SET dose=l
27 Start /nc
28 d dose,aucBb
29 SET dose=5
30 start/nc
31 d dose,aucBb
32 SET dose=10
33 start/nc
34 d dose,aucBb
35 SET dose=50
36 start/nc
37 d dose,aucBb
38 SET dose=100
39 start/nc
40 d dose,aucBb
41 SET dose=250
42 start /nc
43 d dose,aucBb
44 SET dose=350
45 start /nc
46 d dose,aucBb
47 SET dose=500
48 start /nc
49 d dose,aucBb
B-72
DRAFT - DO NOT CITE OR QUOTE
-------
1 SET dose=750
2 start /nc
3 d dose,aucBb
4 SET dose=l,000
5 start /nc
6 d dose,aucBb
7 SET dose=2500
8 start /nc
9 d dose,aucBb
10 SET dose=5000
11 start /nc
12 d dose,aucBb
13 S ODS=0
14 END
15
16 !Procedural blocks for all non-pregnant rat data
17 !Not calibrated!!!!!!
18 !Procs from Ward CMD file
19
20 PROCED WARDIV25
21 CLEARIT
22 SDRAT
23 SET TSTOP=48.
24 SET IVDOSE=2500.
25 END
26
27 PROCED PWARDIV25
28 PLOT /D=WARDIV25, CVB
29 END
30
31 DATA WARDIV25(T, CVB)
32 0.072 4849
33 0.168 3926
34 0.24 2965
35 0.504 2836
36 1.008 3248
37 1.992 2589
38 3 2619
39 4.008 2514
40 7.008 2315
41 19.992 1495
42 22.992 1272
43 24 1214
44 25.992 982
45 28.008 957
46 30 860
47 37.992 238
48 39 200
49 40.008 150
B-73 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
40.992 167
43.008 77
END
PROCED WARDIV1
CLEARIT
SDRAT
SET TSTOP=8
SET IVDOSE=100., tchng=0.016
END
PROCED PWARDIV1
PLOT /D=RGOIV1, CVB
END
DATA WARDIV1 (T,CVB)
0.072 141.7
0.168 121.8
0.24 111.6
0.504 99.7
0.744 97.4
1.008 86.3
1.488 80.3
1.992 58
3 44.4
4.008 22.8
4.992 10.9
6 3.8
7.008 1.4
END
PROCED WARDPO25
CLEARIT
cdmice
SET BW=0.3
SET TSTOP=48
SET DOSE=2500
END
PROCED PWARDPO25
PLOT /D=WARDPO25, CVB
END
DATA WARDPO25(T,CVB)
0.072 862.7
0.168 1243
0.24 1356
0.504 1621
1.008 1641
B-74
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
1.992 1611
3 1869
4.008 1896
7.008 2181
24 1365
25.992 1081
28.008 921
30 958.4
31.008969.8
45 42.9
46.00827.1
46.992 16.4
48 23.9
49.00841.9
49.992 13.1
52.0082.3
52.992 1
END
PROCEDWARDPO1
CLEARIT
cdmice
SET BW=0.3
SET DOSE=100, tstop=8
END
PROCED PWARDPO1
PLOT /D=WARDPO1, CVB
END
DATA WARDPO1(T,CVB)
0.072
0.168
0.24
0.504
0.744
1.008
1.488
1.992
O
4.008
4.992
6
7.008
END
85.5
95.6
95.5
91.1
86.6
80.6
71.3
61.1
45.1
27.4
16.4
8.9
4.2
B-75
DRAFT - DO NOT CITE OR QUOTE
-------
B.3.4. Procedural .m files for reproducing the results in Appendix B and Chapter 3
B.3.4.1. Key to ACSL Extreme v2.5.0.6.mfiles
1 Found in the Runtime Files Folder
2 CDmice.m Sets parameters for (CD) mouse simulations
3 Rogers-mouse-inhal.m Figure B-2 - Simulations of mouse inhalation exposures from GD 6,
4 7, 8 and 10 mice from Rogers et al., 1993.
5 PerkinsDorm-mouse-inh.m Figure B-3 - Simulations of inhalation exposures to MeOH in NP
6 mice from Perkins et al. 1995 (8 hr exposures) and GD 8 mice from
7 Dorman et al. 1995 (6 hr exposures)
8 Ward_mouse_GD18.m Figure B-4 - Oral exposures to MeOH in pregnant and non-pregnant
9 mice Data from Dorman et al., 1995 and Ward et al., 1997
10 Ward-mouse-iv.m Figure B-5 - Simulations of mouse IV exposures to MeOH from
11 Wardetal., 1997
12 Apaja-mouse-drink.m Calculates internal doses for mice in Apaja (1980)
13
14 SDrat.m Sets parameters for Sprague-Dawley (SD) rat simulations
15 F344rat.m Sets parameters for F344 rat simulations
16 Ward-rat-iv.m Figure B-10 - Simulations rat IV exposures from Ward et al., 1997 and
17 Hortonetal., 1992
18 Horton-rat-inhal.m Figure B-l 1 - Simulations rat inhalation exposures from Horton et al., 1992
19 Ward-rat-oral.m Figure B-12 - Simulations rat oral exposures from Ward et al., 1997
20 Nedo-rat-inhal-devpmt-rat.m Figure B-13 - Simulations rat inhalation (bioassay) exposures
21 (200, 500, 1,000, 2000, & 5000 ppm)
22 Nedo-rat-inhal-cancer.m Simulations for NEDO F344 rat cancer inhalation study
23 rat-infu-sims.m Figure B-14 - Simulations rat "oral" exposures (bioassay doses, but using
24 liver infusion; for illustration only)
25
26 humanset.m Sets human MeOH PBPK parameters
27 Sedivec_human_inh.m Figure B-16 - Simulation of human urinary MeOH elimination
28 following Inhalation exposures from Sedivec et al. 1981
29 Batterman_human_inh.m Figure B-17 (upper panel) - Simulations of human inhalation exposure
30 data of Batterman et al. 1998
31 Osterloh_human_inh.m Figure B-17 (lower panel) - Simulations of human inhalation exposure
32 data of Osterloh et al. 1996
33 Ernstgard_human_inh.m Figure B-l 8 - Simulations of human inhalation exposures to MeOH
34 from Ernstgard et al. 2005
35
36 mouse_inh_sim.m Produces data for Table B-5, mouse inhalation exposures
37 human_inh_sim.m Produces data for Table B-5, human inhalation expsoures
38 human_oral_sim.m Produces data for Table B-5, human oral exposures
39 human_drink_compare.m Figure B-24 and Table B-9 (altenate drinking pattern comparison)
40
41 Found in the Sensitivity Analysis Files Folder
42 Fig_B-6 Sensitivity of the mouse model to metabolic parameters (e.g., Km and Vmax)
43 for the inhalation route
44 Fig_B-7 Sensitivity of the mouse model to flow parameters (e.g., blood flow to liver)
45 and to the rest-of-body partition coefficient for the inhalation route
B-76 DRAFT - DO NOT CITE OR QUOTE
-------
1 Fig_B-8 Sensitivity analysis of the rat model to oral absorption parameters for a bolus
2 oral exposure (1,000 mg/kg)
B.3.4.1. [B.3.4.2.] Code for .m files
3 % File CDmice.m
4 % Sets parameters for mouse simulations, MeOH PBPK model
5 CONCPPM=10; WESITG=0; WEDITG=0; CINT-0.1;
6 start @nocallback
7 BW=0.03; TSTOP=24; TCHNG=7; REST=20000; WORK=20000;
8 IVDOSE=0; DOSE=0; DRDOSE=0; RATS=0; KLOSS=0; LIVRO=0;
9 PL=1.06; PF=0.083; PR=0.66; PLU=1; PB=1350;
10 QPC=25.4; QCC=25.4; FRACIN=0.73; KFEC=0;
11 QLC=0.25; QFC=0.05;
12 KM=12; VMAXC=14.3; K1C=0.0; KAS=0.0; KLLC=0;
13 VMAX2C=19; KM2=210; KAI=0.5; KSI=5.0;
14 VAC=0.0123; VFC=0.07; VLC=0.055; VLUC=0.0073; VVBC=0.0368;
15 CONCPPM=0.0; IVDOSE=0.0; DOSE=0.0; DWDOSE=0; MULTE=0; RDRINK=1;
16 % Volumes from Brown et al
17 % Mouse QPC avg from Brown 29; 24 used in Corley et al and others
18 % AVG of measured vent rates by Perkins et al 25.4 L/hr/kgA0.75
19 % Blood volume 4.9% total. As per Brown 25:75 split art:ven
20 % Metab originally from Ward et al - KldC for mice = 0
21
22 % use mouselNH_fit-params.m % File contents copied below
23 % Updated parameters as obtained by Paul Schlosser, U.S. EPA
24 % August 11, 2009 [this file updated]
25
26 % Values generated through parameter estimation script 'mouselNH_fit.m'
27 VMAX2C = 3.222500e+00; KM2 = 660; VMAXC = 19; KM = 5.2; FRACIN = 6.650939e-01;
28
29 % Values generated through parameter estimation script 'mouseor_fit.m'
30 VASC = 1.833246e+03; KSI = 2.2; KAI = 0.33; MASC = 620;
31
32 % File Rogersjnousejnhal.m (Figure B-2)
33 % Produces MeOH PBPK figures for Rogers' mouse inhalation exposures
34 % Variables in the plot command are case sensitive
35 use CDmice
36 % set mouse parameters
37 % DATA BLOCKS
38 % These data blocks taken directly from MeOH CBMMv3.cmd
39 % Data for are T (hours), CV (mg/L)
40 % semicolons (";") creates a new line in a data file
41
42 % Rogers et al., Teratology, 1997
43 D7IN10 = [1, 930; 4, 2800; 6, 3360; 7, 3990; 7.5, 3980;
44 8, 4120; 9, 3270; 12, 2630; 16, 1690; 26, 60];
45
46 % Rogers et al., Teratology, 1993
47 D6IN1 = [7, 63; 7, 131]; D6IN2 = [7, 487; 7, 641];
48 D6IN5 = [7, 2126; 7, 1593]; D6IN7p5 = [7, 2801; 7, 3455];
49 D6IN10 = [7, 4653; 7, 4304]; D6IN15 = [7, 7720; 7, 7394];
50
51 % RUN MODEL
52 RATS=0.0; KLOSS=0.0; % -> open chamber
53 TCHNG=7; CONCPPM=10000; TSTOP=27.0; MULTE=0; BW=0.032;
54 CINT=TSTOP/1000; cs=[]; prepare ©clear T CVB
B-77 DRAFT - DO NOT CITE OR QUOTE
-------
1 forCONCPPM=[1, 2, 5, 7.5, 10, 15]*1000
2 start @nocallback
3 cs=[cs,_cvb];
4 % Since TSTOP & CINT not changing, assume _t also the same.
5 end
6
7 % PLOT COM MAN DS
8 % The rogers.aps file will retain changes made using the plot
9 % editor as long as the editor is called by clicking the
10 % words EDIT PLOT PROPERTIES not the little icon in the
11 % properties dialogue box
12 plot(_t,cs(:,1), _t,cs(:,2), _t,cs(:,3), _t,cs(:,4), _t,cs(:,5), _t,cs(:,6), ...
13 D6IN1(:,1),D6IN1(:,2),D6IN2(:,1),D6IN2(:,2),D6IN5(:,1),D6IN5(:,2), ...
14 D6IN7p5(:,1),D6IN7p5(:,2),D6IN10(:,1),D6IN10(:,2), ...
15 D7IN10(:,1),D7IN10(:,2),D6IN15(:,1),D6IN15(:,2), 'rogers.aps')
16
17 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
18 % Cannot save data with different # of rows to the same table.
19 cs=[_t,cs];
20 save cs @file='Rogersplotdata.csv' @format=ASCII @separator=ascii
21
22 % File: PerkinsDorm-mouse-inh.m (Figure B-3)
23 % Produces MeOH PBPK simulations Perkins 1995 inhalation exposures,
24 % and Ward 1997 (pregnant) and Dorman 1995 for comparison)
25 % Includes all nonpregnant and "early" GD ( open chamber
57 MULTE=0; TSTOP=24; CONCPPM=2500; QPC = 29; QCC=29; TCHNG=8;
B-78 DRAFT - DO NOT CITE OR QUOTE
-------
1 start @nocallback
2 Cs25 = _cvb; Ts25 = _t;
3 CONCPPM=5000; QPC=24; QCC=24; start @nocallback
4 Cs50 = _cvb; Ts50 = _t;
5 CONCPPM=10000; TSTOP=36; QPC=21; QCC=21; start @nocallback
6 Cs100 = _cvb;Ts100 = _t;
7 use CDmice
8 RATS=0; KLOSS=0; % -> open chamber
9 TCHNG=6; CONCPPM=15000; TSTOP=36; start @nocallback
10
11 % PLOT COM MAN DS
12 % The .aps file will retain changes made using the plot
13 % editor as long as the editor is called by clicking the
14 % words EDIT PLOT PROPERTIES not the little icon in the
15 % properties dialogue box
16 plot(Ts25, Cs25, Ts50, Cs50, TslOO, Cs100,_t, _cvb, ...
17 Perk25(:,1), Perk25(:,2), Perk50(:,1), Perk50(:,2), ...
18 Perk100(:,1), Perk100(:,2),Dor815(:,1), Dor815(:,2), 'inhalation.aps')
19
20 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
21 % Can't save data with different # of rows to the same table.
22 mytablel = [Ts25, Cs25, Ts50, Cs50, TslOO, Cs100,_t, _cvb];
23 save mytablel @file='PerkinDormanplotdata.csv' @format=ASCII @separtor=comma
24
25 %FileWardGD18.m
26 % Creates Figure B-4, including Ward et al 1997 NP and GD 18 mouse data
27 % and Dorman et al 1995 GD 8 mouse data.
28 CDmice
29 TSTOP=25; DOSE=1500; CONCPPM=0; MULTE=0;
30 prepare ©clear T CVB
31 start @nocallback
32 T1=_t;P1=_cvb;
33 DOSE=2500; start @nocallback
34 D15=[1, 1609.6; 2, 1331.2; 4, 1241.6;
35 8, 707.2; 16, 160; 24, 38.4];
36 D25a=[0.5, 2370; 0.96, 2645; 1.44, 2705; 2, 2719;
37 2.2, 2781; 3, 2704; 4, 2370; 5, 2617; 6, 2516;
38 7, 2635; 8, 2213; 9, 2370; 10, 2028; 11, 1916;
39 12, 1347; 13, 1467; 14, 1354; 15, 1175; 16, 864.3;
40 17, 745.2; 18, 422.4; 19, 428; 21, 243; 24, 136];
41 D25b=[0.5, 2024; 1, 2554; 2, 3193; 4, 3002; 6, 2933;
42 10, 1976; 12, 1922; 15, 1339; 18, 1033; 21, 832; 24, 580];
43
44 plot(D15(:,1),D15(:,2),D25a(:,1),D25a(:,2),D25b(:,1),D25b(:,2),...
45 T1,P1,_t,_cvb,"wardgd18plot.aps")
46 % File Ward-mouse-iv.m
47 % M File for reproducing MeOH PBPK Figure B-5 For WARD iv mouse exposures
48 % (also Ward Pregnant Includes all nonpregnant and Pregnant)
49
50 % DATA BLOCKS
51 %Taken directly from MeOH CBMMv3.cmd, values are [T (hours), CV (mg/L)]
52 % Ward etal, FAT 1995
53 NPIV25=[0.08, 4481.8; 0.25, 4132.2; 0.5, 3888; 1, 3164.8; 2, 2303.5;
54 4, 1921.5; 6, 1883.8; 8, 1620; 12, 838; 18, 454.7; 24, 1.41];
55 %PROCED MWARDGD8IV25
56 GD8IV25=[0.0833, 4606.2; 0.25, 4079.5; 0.5, 3489.3; 1, 2939.6; 2, 3447.6;
57 4, 2605.0; 6, 2690.5; 8, 2574.9; 12, 1506.1; 18, 498.6; 24, 0.554];
B-79 DRAFT - DO NOT CITE OR QUOTE
-------
1 %!Wardetal., DMD, 1996
2 GD18IV25=[0.0833, 4250.0; 0.25, 3445.1; 0.5, 2936.8; 1, 2470.5; 2, 2528.1;
3 4, 2292.3; 6, 2269.4; 8, 2057.0; 12, 1805.9; 18, 1482.2; 24.0, 496.1];
4 %Wardetal., DMD, 1996
5 GD18IV5=[0.25, 854.7; 0.5, 720.2; 1, 624.1;
6 2, 453.2; 3, 307.6; 4, 217.7; 4.5, 202.6];
7 %Wardetal., DMD, 1996
8 GD18IV1=[0.25, 252; 0.52, 242.2; 1, 222.7; 2, 176.4; 3, 134.2; 3.5, 94.41];
9
10 % RUN MODEL
11 useCDMICE
12 TSTOP=24.0; IVDOSE=2500; TCHNG=0.025; start @nocallback
13 CVs25 = _cvb; Ts25 = _t; TSTOP=6; IVDOSE=500; start @nocallback
14 CVs5 = _cvb; Ts5 = _t; TSTOP=4; IVDOSE=100; start @nocallback
15 CVs1 = _cvb; Ts1 = _t; IVDOSE=200; start @nocallback
16
17 % PLOT COM MAN DS
18 % The .aps file will retain changes made using the plot
19 % editor as long as the editor is called by clicking the
20 % words EDIT PLOT PROPERTIES not the little icon in the
21 % properties dialogue box
22 plot(Ts25, CVs25, Ts5, CVs5, Ts1, CVs1, NPIV25(:,1), NPIV25(:,2),...
23 GD8IV25(:,1), GD8IV25(:,2), GD18IV25(:,1), GD18IV25(:,2), ...
24 GD18IV5(:,1), GD18IV5(:,2), GD18IV1(:,1), GD18IV1(:,2), 'iv.aps')
25 plot(_t, _cvb, Ts1, CVs1, GD18IV1(:,1), GD18IV1(:,2), 'ivb.aps')
26
27 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
28 % Cant save data with different # of rows to the same table.
29 mytablel = [Ts25, CVs25, Ts5, CVs5, _t, _cvb, Ts1, CVs1];
30 save mytablel @file='WardlV.csv' @format=ASCII @separator=comma
31
32 % File Apaja-mouse-drink.m
33 % Calculates internal doses for mice in Apaja (1980)
34 use CDmice
35 DWDOSE=1; start @nocallback
36 DWDS=[0.045, 550; 0.045, 970; 0.045, 1800;
37 0.040, 560; 0.040, 1000; 0.040. 2100];
38 % Above are BWs and doses for males, then females, from Apaja (1980)
39 ODS=1; TSTOP=24*3*7; MULTE=1; DAYS=6.0; simres=[]; LIVRO=0;
40 PER1=1.5; DUR1=0.75; PER2=3.0; DUR2=0.5; FNIGHT=0.8; CINT=0.01;
41 prepare @clear T CVB STOM
42 for RDRINK =0 %[1, 0] % 0 -> mouse drinking pattern
43 fori=1:length(DWDS)
44 BW=DWDS(i,1); DWDOSE=DWDS(i,2); start @nocallback
45 simres=[simres;[TDOSE*(24/TSTOP)/BW,BW,AUCBF,max(_cvb),AMF]];
46 end
47 plot(_t,_cvb)
48 end
49 simres=[simres(:,1)*7/6, simres];
50 simres/100 % Print values to screen (/100)
51 TDOSE*(24/TSTOP)/BW % Check that final total dose/day is correct
52 save simres @file='Apaja_mouse_drink_sims.csv' @format=ascii @separator=comma
53
54 % File SDrat.m
55 % Sets parameters for rat simulations, MeOH PBPK model
56 CONCPPM=10; WESITG=0; WEDITG=0; TSTOP=24; TCHNG=6; MULTE=0;
57 REST=20000; WORK=20000;
B-80 DRAFT - DO NOT CITE OR QUOTE
-------
1 start @nocallback
2 BW=0.275; TSTOP=24; FRACIN=0.2;
3 IVDOSE=0; DOSE=0; CONCPPM=0; DRDOSE=0; DWDOSE=0; ODS=0; LIVRO=0;
4 QCC=16.4; QPC=16.4; QFC=0.07; QLC=0.25;
5 PL=1.06; PF=0.083; PR=0.66; PB=1350;
6 KM=6.3; VMAXC=5.0; VMAX2C=8.4; KM2=65;
7 KLLC=0.0; K1C=0.0;
8 VAC=0.0185; VFC=0.07; VLC=0.037; VLUC=0.005; VVBC=0.0443;
9
10 KAS=10.9; KSI=6.8; KAI=0.039; KFEC=0.0; VASC=0;
11 % Just above are linear absorption params fit to 100 mg/kg
12 % oral data, w/ no fecal elimination
13
14 KAS=0.0; % Below are forsaturable uptake model
15 % Values generated through parameter estimation script 'ratoral_fit.m'
16 KSI = 7.4; KAI = 0.051; KFEC = 0.029; VASC = 5.573125e+03; KMASC = 620;
17
18 % File F344rat.m: parameters specific to F344 rat
19 % Created by Paul Schlosser, U.S. EPA, Aug. 2009
20 use SDrat
21 VMAXC=0;VMAX2C=22.3; KM2=100;
22
23 % File: Ward-rat-iv.m
24 % Creates Figure B-9; rat MeOH PBPK model, to simulate
25 % Ward '97 rat iv 2500 & 100 mg/kg (BW=275, SD)
26 % and Morton '92 iv 100 mg/kg (BW=100, F344)
27 rwi25 =[0.072, 4849; 0.168, 3926; 0.24, 2965; 0.5, 2836; 1, 3248;
28 2, 2589; 3, 2619; 4, 2514; 7, 2315; 20, 1495; 23, 1272; 24, 1214;
29 26, 982; 28, 957; 30, 860; 38, 238; 39, 200; 40, 150; 41, 167; 43, 77];
30
31 % ward '97 rat iv 100 mg/kg (BW 275, SD)
32 rwi1=[0.072, 141.7; 0.168, 121.8; 0.24, 111.6; 0.5, 99.7; 0.744, 97.4;
33 1, 86.3; 1.488, 80.3; 2, 58; 3, 44.4; 4, 22.8; 5, 10.9; 6, 3.8];
34
35 % horton '92 rat iv 100 mg/kg (BW 220, F344)
36 rhi1=[0.24, 91.13; 0.52, 80.14; 0.90, 70.29; 1.38, 61.22;
37 1.86, 60.63; 2.79, 39.40; 4.30, 26.05; 5.81, 12.51];
38
39 use SDrat
40 prepare @clear T CVB
41 TSTOP=48; IVDOSE=2500; TCHNG=0.016; BW=0.275; CINT=0.1; start @nocallback
42 Twi25 = _t; Cwi25 = _cvb;
43 IVDOSE=100; start @nocallback
44 Twi1 = _t; Cwi1 = _cvb;
45 BW=0.22; start @nocallback
46
47 plot(Twi25, Cwi25, Twi1, Cwi1, _t, _cvb, rwi25(:,1), rwi25(:,2), ...
48 rwM(:,1), rwi1(:,2), rhi1(:,1), rhi1(:,2), 'rwi2500.aps')
49
50 use F344rat
51 IVDOSE=100; BW=0.22; TCHNG=0.016; CINT=0.1; start @nocallback
52 plot(Twi25, Cwi25, Twi1, Cwi1, _t, _cvb, rwi25(:,1), rwi25(:,2), ...
53 rwi1(:,1), rwi1(:,2), rhi1(:,1), rhi1(:,2), 'rwMOO.aps')
54
55
56
57 % Creates Figure B-10
B-81 DRAFT - DO NOT CITE OR QUOTE
-------
1 hi20=[0.46, 18.70; 1, 23.76; 3, 59.73; 6, 80.12
2 7, 83.25; 9, 53.49; 12, 16.54; 15, 0.91];
3 hi12=[0.46, 4.89; 1, 8.02; 3, 20.57; 6, 26.63;
4 7, 16.12; 8, 9.28; 9, 5.23; 10.5, 2.93; 12, 0.98];
5 hi2=[0.48, 1.2; 3, 3.1; 6, 3.7; 6.47, 2.7;
6 7, 2.0; 8, 1.6; 9, 1.2];
7
8 use F344rat
9 prepare ©clear T CVB
10 TSTOP=16; CONCPPM=2000; TCHNG=6; BW=0.22; CINT=0.1; start ©nocallback
11 120 = _t; c20 = _cvb;
12 CONCPPM=1200; TSTOP=13; start ©nocallback
13 t12 = _t;c12 = _cvb;
14 CONCPPM=200; TSTOP=10; start ©nocallback
15 t2 = _t; c2 = _cvb;
16 TSTOP=16; CONCPPM=2000; TCHNG=7; start ©nocallback
17
18 plot(t20, c20, t12, c12,12, c2, hi20(:,1), hi20(:,2), ...
19 hi12(:,1), hi12(:,2), hi2(:,1), hi2(:,2), _t, _cvb, 'hi2000.aps')
20
21 % File: Ward-rat-oral.m
22 % MeOH PBPK model rat simulations for Ward '97 rat oral data
23 % Creates Figre B-11
24 use SDRat
25 KAS=10.9; KSI=6.8; KAI=0.039; KFEC=0.0; VASC=0;
26 BW=0.3; ODS=0; prepare ©clear T CVB
27 DOSE=100; TSTOP=10; start ©nocallback
28 t1=_t;c1=_cvb;
29 DOSE=2500; TSTOP=36; start ©nocallback
30 t2=_t;c2=_cvb;
31
32 KAS=0; use ratoral_fit-params
33 DOSE=100; TSTOP=10; start ©nocallback
34 t1A=_t;c1A=_cvb;
35 DOSE=2500; TSTOP=36; start ©nocallback
36 t2A=_t;c2A=_cvb;
37
38 d1=[0.072, 85.5; 0.168, 95.6; 0.24, 95.5; 0.504, 91.1;
39 0.744, 86.6; 1.008, 80.6; 1.488, 71.3; 1.992, 61.1;
40 3, 45.1; 4.008, 27.4; 4.992, 16.4; 6, 8.9; 7.008, 4.2];
41
42 d25=[0.072, 862.7; 0.168, 1243; 0.24, 1356; 0.504, 1621;
43 1.008, 1641; 1.992, 1611; 3, 1869; 4.008, 1896; 7.008, 2181;
44 24, 1365; 25.992, 1081; 28.008, 921; 30, 958.4; 31.008, 969.8];
45
46 plot(t2,c2, t2A,c2A, d25(:,1),d25(:,2), ...
47 t1,c1, t1A,c1A, d1(:,1),d1(:,2),"wardratoralplot.aps")
48 plot(t2,c2, t2A,c2A, d25(:,1),d25(:,2), ...
49 t1,c1, t1A,c1A, d1(:,1),d1(:,2),"wardratoralplotb.aps")
50
51 % File: Nedo-rat-inhal-devpmt.m
52 % MeOH PBPK model rat simulations for rat inhalation exposures
53 % 200, 500, 1000, 2000, and 5000 ppm
54 % Internal doses for NEDO developmental inhalation exposures, Sprague-Dawley rats
55 % Creates Figure B-13 ('simres' is tabulated results)
56 use SDRat
57 prepare ©clear T CVB
B-82 DRAFT - DO NOT CITE OR QUOTE
-------
1
2 simres=[]; ts=[]; cs=[]; ts2=[]; cs2=[]; TSTOP=24*2*7; CINT=1;
3 for CONCPPM=[200, 500, 1000, 2000, 5000]
4 TCHNG=22; MULTE=1; start @nocallback
5 res=[CONCPPM,max(_cvb),AUCBF,AMF];
6 ts=[ts,_t]; cs=[cs,_cvb];
7 TCHNG=TSTOP; MULTE=0; %CONCPPM=22*cp/24;
8 start @nocallback
9 simres=[simres;[res,CONCPPM,max(_cvb),AUCBF,AMF]];
10 ts2=[ts2,_t]; cs2=[cs2,_cvb];
11 end
12
13 simres
14 plot(ts(:,2),cs(:,2),ts(:,3),cs(:,3),ts(:,4),cs(:,4), ...
15 ts2(:,2),cs2(:,2),ts2(:,3),cs2(:,3),ts2(:,4),cs2(:,4), ...
16 [2424],[0,95],'fig13.aps')
17 save simres @file='Nedo_devpomt_rat_inhal_sims.csv' @format=ascii @separator=comma
18 cs=[ts(:,1),cs,cs2]; save cs @file='Fig13_sims.csv' @format=ascii @separator=comma
19
20 % File Nedo-rat-inhal-cancer.m
21 % Simulations for NEDO F344 rat cancer inhalation study
22 use F344rat
23 TCHNG=22.7; TSTOP=5*7*24; MULTE=1;
24 res=[]; CONCPPM=200; prepare ©clear T CVB
25 start @nocallback
26 TCHNG=19.5; ODS=1; cppm=[0,10,100,1000];
27 bwm=[422.1, 418.3, 417.7, 410.0];
28 bwf=[268.7, 270.6, 267.0, 264.9];
29 forsdose=[0,25.4863]
30 fori=1:length(cppm)
31 CONCPPM=cppm(i);
32 BW=bwm(i)/1000;
33 DOSE=sdose/(BWA0.25);
34 start @nocallback
35 res=[res;[CONCPPM,TCHNG,BW,AUCBF,max(_cvb),AMF]]
36 BW=bwf(i)/1000;
37 DOSE=sdose/(BWA0.25);
38 start @nocallback
39 res=[res;[CONCPPM,TCHNG,BW,AUCBF,max(_cvb),AMF]]
40 end
41 end
42
43 save res @file='Nedo_rat_cancer_sims.csv' @format=ascii @separator=comma
44
45 % File: rat-infu-sims.m
46 % MeOH PBPK model rat simulations for zero-order liver infusions
47 % Creates Figre B-14
48 use SDRat
49 Iv0=[0.33, 65.9; 0.33, 624.1; 0.34, 2177;
50 0.49, 53.2; 0.50, 524;
51 0.54 1780];
52 % Above are BWs and doses from Soffretti et al. 2002a
53 prepare ©clear T CVB
54 TCHNG=12; MULTE=0; simres=[]; ts=[]; cs=[];
55 fori=1:3
56 BW=lvO(i,1); LIVRO=lvO(i,2); TSTOP=24; start @nocallback
57 res=[LIVRO,BW,max(_cvb),0,AUCB,0,AMET];
B-83 DRAFT - DO NOT CITE OR QUOTE
-------
1 TSTOP=84; start @nocallback
2 res(4)=AUCB; res(6)=AMET; simres=[simres;res];
3 ts=[ts,_t]; cs=[cs,_cvb];
4 end
5 simres/100
6 plot(ts(:,1),cs(:,1),ts(:,2),cs(:,2),ts(:,3),cs(:,3),'fig14b.aps')
7 save simres @file='rat_liver-infusion_sims.csv' @format=ascii @separator=comma
8
9 % File: humanset.m
10 % Sets parameters for human simulations. Expects the user to define
11 % metabf = "linear" to use 1st-order metabolism parameters; otherwise
12 % metabf set to "non-linear" and Michaelis-Menten parameters used.
13 BW = 70; FRACIN = 0.8655; IVDOSE=0; DOSE=0; CONCPPM=0; LIVRO=0;
14 PB = 1626; PL = 0.583; %1.06;
15 PF = 0.142; PR = 0.805; %0.66;
16 PLU=1.07;%1.0;
17 VFC = 0.214; VLC = 0.026; VLUC = 0.008;
18 VAC = 0.0198; WBC = 0.0593;
19 QPC = 18.5; QCC = 18.5; QLC = 0.227; QFC = 0.052;
20 KM = 12.76; VMAXC = 0; %VMAXC=11.72; %low KM optimum
21 KM2 = 460; VMAX2C = 0; %VMAX2C=304.5; %high KM optimum
22 KLLC = 60.7; %linear liver metabolism optimum
23 K1C = 0.0397; KAI = 0.22; KSI = 1.1; KAS = 2.0;
24 RATS=0; KLOSS=0; % constant exposure/no chamber losses
25 QPC=24.0; QCC=16.5; REST=3000; WORK=3000;
26
27 % Below are optimal values for Michaelis-Menten liver metabolism
28 K1C = 0.0342; KLLC = 0.0; KBL=0.612;
29 KM = 23.7; VMAXC = 33.1;
30
31 % Mouse oral uptake KMASC; others set to match ethanol values
32 % for humans from Sultatos et al. (2004), with VASC set so that
33 % VASC/KMAS = 0.21/h, the Sultatos et al. 1st-order constant,
34 % and KFEC = 0 corresponding to assumed 100% absorption.
35 VASC = 377; KSI = 3.17; KAI = 3.28; KMASC = 620; KFEC=0;
36
37 exist metabf; % check if metabf defined
38 if-ans % If not...
39 metabf = "non-linear"
40 end
41 if metabf=="linear"
42 % Below are optimal values for 1st-order liver metabolism
43 K1C = 0.0373; KLLC = 95.7; KBL=0.564; VMAXC=0.0;
44 % Below are 'no bladder values; uncomment next line to use
45 %K1C = 0.0278; KLLC = 70.3; KBL=1000.0; VMAXC=0.0;
46 else metabf="non-linear";
47 end
48 disp(['Simulation for ',ctot(metabf),' human kinetics']);
49
50 % File: Sedivec_human_inh.m
51 % Creates MeOH PBPK Figure B-16
52 % For human inhalation exposures, w/ data of Sedivec et al
53
54 o/0 DATA BLOCKS
55 % These data blocks taken directly from MeOH CBMMv3.cmd
56 % Data are T (hours), CV (mg/L), cumulative urinary clearance (mg)
57 % Rounded to 3-4 sig figs
B-84 DRAFT - DO NOT CITE OR QUOTE
-------
1 % Sedivec etal., Int Arch Occ Health 1981, urine
2 HS231 = [2, 3.338, 0.1168; 4, 5.776, 0.4358; 6, 7.371, 0.8960;
3 8, 8.581, 1.454; 10, 6.576, 1.985; 12, 3.243, 2.328;
4 14, 1.32 , 2.488; 19, 0.333, 2.632; 24, 0, 2.661];
5
6 HS157 = [2, 2.185, 0.0765; 4, 3.941, 0.291; 6, 4.896, 0.600;
7 8, 5.708, 0.971; 10, 4.360, 1.324; 12, 1.738, 1.537;
8 14, 0.776, 1.625; 19, 0.231, 1.713; 24, 0, 1.733];
9
10 HS78 = [2, 0.881, 0.0308; 4, 1.648, 0.1193; 6, 2.285, 0.2570;
11 8, 2.551, 0.4263; 10, 1.515, 0.5686; 12, 0.708, 0.6464;
12 14, 0.430, 0.6863; 19, 0.094, 0.7321; 24, 0, 0.7403];
13
14 % RUN MODEL
15 use humanset
16 TCHNG=8; TSTOP=24; CONCPPM=231;
17 prepare @clear T RU R M ETB
18 start @nocallback
19 ur1 = _metb; t1 = _t; cu1=_rur;, % Save time series for urine MeOHc
20 CONCPPM=157; start @nocallback
21 ur2 = _metb; t2 = _t; cu2= _rur;
22 CONCPPM=78; start @nocallback
23
24 % PLOT COM MAN DS
25 plot(t1,ur1,t2,ur2,_t,_metb,HS231(:,1),HS231(:,3),...
26 HS157(:,1),HS157(:,3),HS78(:,1),HS78(:,3), 'sedivic.aps')
27 plot(t1 ,cu1 ,t2,cu2,_t,_rur,HS231 (:,1),HS231 (:,2),...
28 HS157(:,1),HS157(:,2),HS78(:,1),HS78(:,2), 'sedivic2.aps')
29
30 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
31 % Cant save data with different # of rows to the same table.
32 mytablel = [t1 ,ur1 ,cu1 ,t2,ur2,cu2,_t,_metb,_rur];
33 eval(['save mytablel @file=Sedv_fit_KLLC.',num2str(round(KLLC)),'.csv@format=ascii
34 @separator=comma']);
35
36 % File: Batterman_human_inh.m
37 % Creates MeOH PBPK Figure B-17 (upper panel)
38 % For human inhalation exposures of Batterman et al 1998
39
40 %These data blocks taken directly from MeOH CBMMv3.cmd
41 %Data are T (hours), CV (mg/L)
42 % Batterman et al., Int Arch Occ Health 1998
43 HB82=[2, 13.6; 2.25, 13.4; 2.5, 12; 3, 9.6;
44 4, 6.4; 5, 4.1; 6, 2.6; 7, 1.8; 8, 1.4];
45 HB81=[1, 6.5; 1.25, 6.2; 1.5, 5.4; 2, 4.3; 3, 2.8;
46 4, 1.5; 5, 0.94; 6, 0.72; 7, 0.52; 8, 0.23];
47 HB830=[0.5, 4.6; 0.75, 4.7; 1, 4.9; 1.5, 3.5; 2.5, 2.2;
48 3.5, 1; 4.5, 0.52; 5.5, 0.51; 6.5, 0.47; 7.5, 0.68];
49
50 use humanset
51 prepare ©clear T CVB
52 TCHNG=2; CONCPPM=800; TSTOP=16; start @nocallback
53 t2=_t; c2=_cvb; TCHNG=1; start @nocallback
54 t1=_t; c1=_cvb; TCHNG=0.5; start @nocallback
55 t30=_t; c30=_cvb;
56
57 % PLOT COMMANDS
B-85 DRAFT - DO NOT CITE OR QUOTE
-------
1 plot(t2,c2, t1,d, t30,c30, HB82(:,1),HB82(:,2), ...
2 HB81(:,1),HB81(:,2),HB830(:,1),HB830(:,2), 'batterman.aps')
O
4 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
5 % Cant save data with different # of rows to the same table.
6 le=1 :min([length(t1),length(t2),length(t30)]);
7 mytablel = [t2(le),c2(le),t1 (Ie),c1 (Ie),t30(le),c30(le)];
8 eval(['save mytablel @file=Batter_fit_KLLC.',num2str(round(KLLC)),'.csv '...
9 '@format=ascii @separator=comma']);
10
11 % File: Osterloh_human_inh.m
12 % Creates Fig B-17 (lower panel)
13 % Data from Osterloh et al., JOEM 1996
14 % Digitized data provided by EPA (datl)
15 % Subtracted background from exposure blood levels
16 % by Paul Schlosser, U.S. EPA
17 use humanset
18 BW=78.2;
19 dat1=[0.05, 0.54; 0.25, 1.39; 0.50, 1.82; 0.75, 2.28; 1, 2.42;
20 1.5, 2.94; 2, 3.37; 2.5, 3.90; 3, 4.21; 3.5, 4.61;
21 4, 4.82; 5, 2.99; 6, 2.30; 7, 1.40; 8, 1.07];
22 dat=[0.25, 1.183; 0.5, 1.526; 0.75, 1.948; 1, 2.073; 1.5, 2.741;
23 2, 3.118; 2.5, 3.495; 3, 3.998; 3.5, 4.181; 4, 4.48;
24 5, 2.790; 6, 1.943; 7, 1.106; 8, 0.687];
25
26 prepare ©clear T CVB
27 TCHNG=4; CONCPPM=200; TSTOP=16; start @nocallback
28 plot(_t,_cvb,dat(:,1),dat(:,2), 'osterloh.aps')
29 mytablel =[_t,_cvb];
30 eval(['save mytablel @file=Oster_fit_KLLC.',num2str(round(KLLC)),'.txt @format=ascii']);
31
32 % File: Ernstgard_human_inh.m
33 % Creates MeOH PBPK Figure B-18, w/ data of Ernstgard et al 2005a,b
34 % For human inhalation exposures w/ exercise
35 o/0 DATA BLOCKS
36 %These data blocks taken directly from MeOH CBMMv3.cmd
37 %Data are T (hours), CV (mg/L)
38 % Ernstgard et al. SOT poster, 100 ppm & 200 ppm human
39 ernl =[0.20, 0.87; 0.46, 1.50; 0.97, 2.31; 1.46, 3.24;
40 1.91, 3.65; 2.17, 3.52; 2.50, 2.55; 2.91, 2.23; 3.51, 1.59;
41 4.01, 1.72; 5.02, 0.41; 6.00, 0.50; 9.24, 0.12];
42 ern2 =[0.22, 1.63; 0.49, 2.92; 0.92, 4.76; 1.47, 6.30;
43 1.90, 7.65; 2.16, 6.20; 2.47, 5.49; 2.91, 4.96; 3.50, 3.64;
44 4.00, 3.43; 4.99, 1.94; 5.97, 1.03; 8.90, 0.21];
45
46 % RUN MODEL
47 use humanset
48 QPCHR=QPC; QCCHR=QCC; REST=2.0; WORK=0.0;
49 TCHNG=2.0; CONCPPM=100; TSTOP=10.0; QPCHW=52.6; QCCHW=26.0;
50 %FRACIN=0.9509; %FRACIN=0.9324;
51 prepare T CVB OP QC
52 start @nocallback,
53 cv1 = _cvb; t1 = _t; CONCPPM=200; start @nocallback
54
55 % PLOT COMMANDS
56 plot(t1,cv1,_t, _cvb,...
57 ern1(:,1),ern1(:,2),ern2(:,1),ern2(:,2), 'ernstgard.aps')
B-86 DRAFT - DO NOT CITE OR QUOTE
-------
1 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
2 % Cant save data with different # of rows to the same table.
3 mytablel = [t1, cv1, _t, _cvb];
4 eval(['save mytablel @file=Ernst_nofit_KLLC.',num2str(round(KLLC)),'.csv'...
5 '@format=ascii @separator=comma']);
6
7 % File: mouse_inh_sim.m
8 % Runs simulations for Table B-5, mouse internal-dose calculations
9 % from inhalation exposure, over the concentration range specified
10 % in the 'for' statement below.
11 % Results saved to file 'MouselnhalSims.csv'.
12 use CDMice
13 BW=0.03; TCHNG=7; % 7 hr/day exposures
14 TSTOP=240; MULTE=1; % Run for 10 days; multi-day exposure 'on'
15 RATS=0.0; KLOSS=0.0; % -> open chamber
16 prepare ©clear T CVB
17 CONCPPM=10000; CINT=0.02; start @nocallback
18 % plot(_t,_cvb) % uncomment to see/check that periodicity reached by TSTOP
19 inhres=[]; CINT=0.2;
20 forCONCPPM=[1, 10, 50, 100, 250, 500, 1000, 2000, 5000, 10000]
21 start @nocallback
22 inhres=[inhres;[CONCPPM, AUCBB, max(_cvb), AMET24/(BWA0.75)]];
23 end
24 save inhres @file=MouselnhalSims.csv @format=ASCII @separator=comma
25
26 % File: human_inh_sim.m
27 % Runs simulations for Table B-5, human internal-dose calculations
28 % from inhalation exposure, over the concentration range specified
29 % in the 'for' statement below.
30 % Results saved to file 'HumanlnhSims_KLLC.#.csv', where # is
31 % value of KLLC used (0 if non-linear/Michaelis-Menten kinetics).
32 % If metab="linear", 1st order kinetics used; otherwise non-linear.
33 use humanset
34 WESITG=0; WEDITG=0; simres=[]; MULTE=0; CINT=1.0; RATS=0.0; KLOSS=0.0;
35 CONCPPM=0; TSTOP=1000; TCHNG=1000; DOSE=0; DWDOSE=0; ODS=1;
36 prepare @clear T CVB STOM
37 start @nocallback
38 forCONCPPM=[1, 10, 50, 100, 250, 500, 1000, 2000, 5000, 10000]
39 start @nocallback
40 simres=[simres;[CONCPPM,AUCBF,max(_cvb),AMF/(BWA0.75)]]
41 end
42 disp(['Simulation for ',ctot(metabf),' human kinetics']);
43 eval(['save simres @file=HumanlnhSims_KLLC.',num2str(round(KLLC)), ...
44 '.csv @format=ascii @separator=comma']);
45
46 % File: human_oral_sim.m
47 % Runs simulations for Table B-5, human internal-dose calculations from
48 % oral exposure, over the exposure range specified in the 'for' statement below.
49 % Results saved to file 'Hum_DW_Sims_KLLC.#.csv', where # is value of KLLC
50 % used (0 if non-linear/Michaelis-Menten kinetics).
51 % If metab="linear", 1st order kinetics used; otherwise non-linear.
52 use humanset
53 WESITG=0; WEDITG=0; MULTE=0; CINT=0.1; RATS=0.0; KLOSS=0.0;
54 CONCPPM=0; TSTOP=1000; DWDOSE=0; DOSE=1; ODS=1; DRDOSE=0;
55 prepare @clear T CVB STOM
56 start @nocallback
57 simres=[];
B-87 DRAFT - DO NOT CITE OR QUOTE
-------
1 forDOSE=[0.1, 1, 10,50, 100,250,500, 1000,2000,5000]
2 start @nocallback
3 simres=[simres;[DOSE,AUCBF,max(_cvb)],AMF];
4 end
5 disp(['Simulation for ',ctot(metabf),' human kinetics']);
6 eval(['save simres @file=Hum_DW_Sims_KLLC.',num2str(round(KLLC)),'.csv'...
7 '@format=ascii @separator=comma']);
8
9 % file human_drink_compare.m
10 % creates Figure B-24 and Table B-9
11 % Created by Paul Schlosser, U.S. EPA, 8/26/09
12 use humanset
13 WESITG=0; WEDITG=0; MULTE=0; CINT=0.1; TSTOP=48; DOSE=0.1; ODS=1; DRDOSE=0;
14 prepare ©clear T CVB
15 start @nocallback
16 T1=_t;C1=_cvb; DOSE=0; LIVRO=0.1; TCHNG=12; MULTE=1; start @nocallback
17 T2=_t; C2=_cvb; LIVRO=0; ODS=0; DOSE=0.1; start @nocallback
18 T3=_t; C3=_cvb; DOSE=0; DRDOSE=0.1; start @nocallback
19 plot(T1 ,C1 ,T2,C2,T3,C3,_t,_cvb,'humoralsim.aps')
20
21 tbl=[];metd = 1.0;
22 fordse=[0.1, 1.0, 10, 100, 250, 500]
23 row=[]; LIVRO=0; DOSE=0; DRDOSE=dse; start @nocallback
24 row=[dse,max(_cvb),AUCBF,AMETF];
25 LIVRO=dse; DOSE=0; DRDOSE=0; start @nocallback
26 row=[row,max(_cvb),AUCBF,AMETF];
27 LIVRO=0; DOSE=dse; DRDOSE=0; start @nocallback
28 tbl=[tbl;[row,max(_cvb),AUCBF,AMETF]];
29 end
30 tbl
31
B.3.5. Personal Communication from Lena Ernstgard Regarding Human Exposures
Reported in the Ernstgard and Johanson, 2005 SOT Poster
32 From: Lena Ernstgard [Lena.Ernstgard@imm.ki.se]
33 Sent: Wednesday, March 23, 2005 12:39 AM
34 To: Poet, Torka S
3 5 Subject: RE: Human MeOH poster
36 Hi,
3 7 We measured the ventilation rate and they ought to be similar to those reported by Dr. Johanson at the same
3 8 workload.
39 Sincerly,
40 Lena Ernstgard
41
42 At 18:41 2005-03-22, you wrote:
43
44 Thank you very much. Your net uptake is what we thought. Did you measure ventilation rates?
45
46 Thanks again,
47 Torka
48
49 Torka Poet, PhD
5 0 Center for Biological Monitoring and Modeling
51 Pacific Northwest National Laboratories
52 902 Battelle Blvd.
53 P.O. Box 999, MSIN P7-59
B-88 DRAFT - DO NOT CITE OR QUOTE
-------
1 Richland, WA 99352
2 ph: (509)376-7740
3 fax: (509)376-9449
4 e-mail: Torka.poet@pnl.gov
5 (Express Mail Delivery: 790 Sixth Street, Zip Code 99354)
6
7 From: Lena Ernstgard [mailto :Lena.Ernstgard(@,imm.ki. sel
8 Sent: Sunday, March 20, 2005 11:21 PM
9 To: Poet, Torka S
10 Subject: Re: Human MeOH poster
11 Hi,
12 The manuscript has not been submitted yet, but it will be soon I hope. I will save your mail and send you a
13 copy as soon as possible.
14 When I say % of net uptake, i mean the relative uptake. It is calculated as: cone in exposure chamber -
15 (minus) exhaled cone / (divided by) cone in exposure chamber. I hope you understand how we have done.
16 Sincerly,
17 Lena Ernstgard
B.3.6. Personal Communication from Dr. Rogers Regarding Mouse Exposures
18 Jeff Gift, Ph.D.
19 National Center for Environmental Assessment EPA (B243-01) RTF, NC 27711
20 919-541-4828
21 919-541-0245 (fax)
22 gift.jeff@epa.gov
23
24 -— Forwarded by Jeff Gift/RTP/USEPA/US on 04/04/2005 04:31 PM -—
25 John Rogers
26 To: JeffGift/RTP/USEPA/US@EPA
27 04/04/2005 03:50PM
28 Subject: Re: report(Document link: Jeff Gift)
29
30 Hi Jeff:
31 It's easier just to give you the numbers from that figure in tabular form:
32
33
34 (exposure Blood [MeOH] SEM
35 ends @ 7 hr) (mg/ml)
36
37
38
39
40
41
42
43
44
45
46
47 Whew! Surprised I could find these numbers that fast, I get a little worried when someone asks
48 for 12 year old data, I'm not as organized as I'd like to be,to say the least.
49
B-89 DRAFT - DO NOT CITE OR QUOTE
Time (hr)
(exposure
ends @ 7
1
4
6
7
7.5
8
9
12
16
26
Blood [Me
hr)
0.93
2.80
3.36
3.99
3.98
4.12
3.27
2.63
1.69
0.06
0.05
0.20
0.08
0.13
0.21
0.07
0.16
0.21
0.08
0.02
-------
1
2
3
4
5
6
7
John
John M. Rogers, Ph.D.
Chief, Developmental Biology Branch (MD-67) Reproductive Toxicology Division National
Health and Environmental Effects Research Laboratory Research Triangle Park, NC 27711
T: (919)541-5177
F:(919)541-4017
e-mail: rogers.john@epa.gov
B.3.7. Total MeOH Metabolic Clearance/Metabolites Produced
Table B-6. Mouse total MeOH metabolic clearance/metabolites produced
following inhalation exposures3
Exposure concentration
(ppm)
1
10
50
100
250
500
525
550
575
600
625
675
750
875
1,000
2000
5000
1,0000
AUC (mg/L-hr)
1.51E-01
1.53E+00
8.03E+00
1.72E+01
5.38E+01
1.72E+02
1.89E+02
2.09E+02
2.29E+02
2.51E+02
2.74E+02
3.24E+02
4.09E+02
5.77E+02
7.76E+02
5.12E+03
1.73E+04
4.98E+04
Cmax (mg/L)
2.16E-02
2.18E-01
1.15E+00
2.46E+00
7.83E+00
2.64E+01
2.94E+01
3.26E+01
3.62E+01
3.99E+01
4.40E+01
5.30E+01
6.84E+01
9.88E+01
1.34E+02
7.57E+02
2.00E+03
4.60E+03
Total MeOH metabolically cleared
(mg)
1.20E-02
1.20E-01
6.01E-01
1.20E+00
2.99E+00
5.89E+00
6.17E+00
6.45E+00
6.73E+00
7.01E+00
7.28E+00
7.83E+00
8.63E+00
9.93E+00
1.12E+01
2.37E+01
3.77E+01
5.50E+01
aTotal over a 36-hour period during which mice were exposed for 7 hours to MeOH according to the conditions of
the dose-response study.
B-90
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-7. Human total MeOH metabolic clearance/metabolites produced
from inhalation exposures3
Exposure concentration (ppm)
1
10
50
100
250
500
625
750
875
1,000
AUC
(mg/L-hr)
0.7142
7.142
35.71
71.42
178.6
357.1
446.4
535.7
625.0
714.2
Cmax (mg/L)
0.0300
0.300
1.498
2.997
7.491
14.98
18.73
22.47
26.22
29.97
Total MeOH metabolically cleared
(mg)
10.23
102.3
511.7
1023
2559
5117
6396
7676
8955
10234
aTotal over a 24-hour period during which humans were exposed continuously to MeOH.
Table B-8. Human total MeOH metabolic clearance/metabolites produced
following oral exposures3
Exposure concentration (ms/ks.day)
0.1
1
5
10
50
100
250
AUC (mg/L-hr)
0.3795
3.7954
18.977
37.954
189.8
379.5
948.8
Total MeOH metabolically cleared (mg)
6.2152
62.152
310.8
621.5
3108
6215
15538
aTotal over a 24-hour period during which humans were exposed continuously to MeOH.
Note: MeOH in the model is eliminated via exhalation, metabolism, and urinary excretion (human only). Total
MeOH metabolically cleared approximates total production of down stream metabolites, but as a dose metric is not
equivalent to formaldehyde or formate concentration.
B-91
DRAFT - DO NOT CITE OR QUOTE
-------
B.3.8. Multiple Daily Oral Dosing for Humans
1 Current mode simulations of oral exposures to humans use a constant rate of infusion to
2 the stomach lumen. This approach results in a steady rate of absorption from the stomach equal
3 to the exposure rate irregardeless of the oral uptake rate constants (assumed equal to the mouse),
4 hence avoids the difficulty that independent values of these constants are not available for
5 humans due to a lack of human oral PK data.. A more likely drinking scenario was tested by
6 using additional code within the model to simulate a 6-times/day drinking schedule, over the
7 course of 15 h (see code below). The schedule is still an approximation, as it assumes 6 episodes
8 of drinking, each considered to be a bolus. Specifically, it was assumed that humans drank at 0,
9 3, 5, 8, 11, and 15 hours from the first ingestion of each day, with the respective fractions of daily
10 consumption being 25, 10, 25, 10, 25, and 5% at those times. The predicted blood
11 concentrations resulting from simulations of six daily boluses, once/day boluses, 12 h/d infusion
12 (zero order), or constant (zero order) are shown in Figure B-24 for a total dose of 0.1 mg/kg.
13 Table B-9 shows PBPK model predicted Cmax, AUC, and Amet (for the last 24 hours of repeated
14 exposures) for humans exposed to MeOH via six daily boluses, 12 h/d infusion, or a single daily
15 gavage.
0.08-
< 0.07-
•S 0.06 H
c
.0 0,05-1
4-J
eo
-b 0,04-1
cp
u 0.03-1
o
u 0,02^
13
_g 0,01-
CQ
0-
Continucrus infusion
12 h/d infusion
Single daily bolus
Six-bolus pattern
10
20 30
Time (h)
40
Figure B-24. Simulated human oral exposures to 0.1 mg MeOH/kg/-day
comparing the first few days for four exposure scenarios: continuous (zero-
order) infusion; 12 h/d infusion, a single daily bolus, and a pattern of six boluses
per day (see text).
B-92
DRAFT - DO NOT CITE OR QUOTE
-------
Table B-9. Repeated daily oral dosing of humans with MeOH"
Dose
(mg/kg)
0.1
1
10
100
250
500
Six daily boluses
^max
(mg/L)
0.0204
0.205
2.16
33.4
182
746
AUC
(mg-h/L)
0.0581
0.584
6.17
109
720
3,290
Amet
(mg)
1.97
19.7
197
1,950
4,400
5,180
12 h/d infusion
^max
(mg/L)
0.0170
0.171
1.83
41.4
237
866
AUC
(mg-h/L)
0.0583
0.586
6.21
125
857
3,540
Amet
(mg)
1.97
19.8
197
1,950
4,420
5,190
Sinsle-dailv bolus
^max
(mg/L)
0.0569
0.579
6.64
95.2
296
903
AUC
(mg-h/L)
0.0584
0.594
7.01
204
1,150
4,250
Amet
(mg)
1.98
19.8
197
1,920
4,440
5,270
*AUC in blood and Amet (amount metabolized) computed from 24-48 h
B-93
DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX C. RfC DERIVATION OPTIONS
C.I. RFC DERIVATIONS USING THE NEDO METHANOL REPORT (NEDO, 1987)
1 The BMD approach was utilized in the derivation of potential chronic inhalation
2 reference values. In the application of the BMD approach, continuous models in the EPA's
3 BMDS, version 2.1, were fit to datasets for decreased brain weight in male rats exposed
4 throughout gestation and the postnatal period to 6 weeks and male rats exposed during gestation
5 on days 7-17 only (NEDO, 1987). Although there remains uncertainty surrounding the
6 identification of the proximate teratogen of importance (methanol, formaldehyde, or formate),
7 the dose metrics chosen for the derivation of RfCs were based on blood methanol levels. This
8 decision was primarily based on evidence that the toxic moiety is not likely to be the formate
9 metabolite of methanol (CERHR, 2004), and evidence that levels of the formaldehyde metabolite
10 following methanol maternal and/or neonate exposure would be much lower in the fetus and
11 neonate than in adults. While recent in vitro evidence indicates that formaldehyde is more
12 embryotoxic than methanol and formate, the high reactivity of formaldehyde would significantly
13 limit its transport from maternal to fetal blood, and the capacity for the metabolism of methanol
14 to formaldehyde is lower in the fetus and neonate versus adults. Further discussions of methanol
15 metabolism, dose metric selection and MOAissues are covered in Sections 3.3, 4.5 and 4.6.
C.I.I. Decreased Brain Weight in Male Rats Exposed throughout Gestation and into the
Postnatal Period (NEDO, 1987)
16 The results of NEDO (1987), shown in Table 4-14, indicate that there is not a cumulative
17 effect of ongoing exposure on brain-weight decrements in rats exposed postnatally; i.e., the dose
18 response in terms of percent of control is about the same at 3 weeks postnatal as at 8 weeks
19 postnatal in rats exposed throughout gestation and the FI generation. However, there does
20 appear to be a greater brain-weight effect in rats exposed postnatally versus rats exposed only
21 during organogenesis (GD7-GD17). In male rats exposed during organogenesis only, there is no
22 statistically significant decrease in brain weight at 8 week after birth at the 1,000 ppm exposure
23 level. Conversely, in male rats exposed to the same level of methanol throughout gestation and
24 the FI generation, there was an approximately a 5% decrease in brain weights (statistically
25 significant at the/? < 0.01 level). The extent to which this observation is due to recovery in rats
26 for which exposure was discontinued at birth versus a cumulative effect in rats exposed
27 postnatally is not clear. The fact that male rats exposed to 5,000 ppm methanol only during
28 organogenesis experienced a decrease of brain weight of 10% at 8 weeks postnatal indicates that
29 postnatal exposure is not necessary for the observation of persistent postnatal effects. However,
30 the fact that this decrease was less than the 13% decrease observed in male rats exposed to
C-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 2,000 ppm methanol throughout gestation, and the 8 week postnatal period indicates that the
2 absence of postnatal exposure allows for some measure of recovery.
3 It appears that once methanol exposure is discontinued, continuous biological processes
4 that are disrupted by exposure, manifesting as decreased brain weight, undergo some recovery
5 and brain weights begin to return to normal values. This indicates that brain weight is
6 susceptible to both the level and duration of exposure. Therefore, a dose metric that incorporates
7 a time component would be the most appropriate metric to use. For these reasons and because it
8 is more typically used in internal-dose-based assessments and better reflects total exposure
9 within a given day, daily AUC (measured for 22-hour exposure/day) was chosen as the most
10 appropriate dose metric for modeling the effects of methanol exposure on brain weights in rats
11 exposed throughout gestation and continuing into the FI generation.
12 Application of the EPA methanol PBPK model (described in Section 3.4) to the NEDO
13 (1987) study in which developing rats were exposed during gestation and the postnatal period
14 presents complications that need to be discussed. The neonatal rats in this study were exposed to
15 methanol gestationally before parturition, as well as lactationally and inhalationally after
16 parturition. The PBPK model developed by the EPA only estimates internal dose metrics for
17 methanol exposure in NP adult mice and rats. Experimental data indicate that inhalation-route
18 blood methanol kinetics in NP mice and pregnant mice on GD6-GD10 are similar (Dorman
19 et al., 1995; Perkins et al., 1995a,b; Rogers et al., 1993a,b). In addition, experimental data
20 indicate that the maternal blood:fetal partition coefficient for mice is approximately 1 (see
21 Section 3.4.1.2). Assuming that these findings apply for rats, they indicate that pharmacokinetic
22 and blood dose metrics for NP rats are appropriate surrogates for fetal exposure during early
23 gestation. However, as is discussed to a greater extent in Section 5.3, the additional routes of
24 exposure presented to the pups in this study (lactation and inhalation) present uncertainties that
25 make it reasonable to assume that average blood levels in pups in the NEDO report are also
26 greater than those of the dam. However, it is also reasonable to assume that any differences seen
27 between the pups and dams would also be seen in mothers and human offspring. Therefore, the
28 presumed differences between pup and dam blood methanol levels are deemed relatively
29 inconsequential, and the PBPK model-estimated adult blood methanol levels are assumed to be
30 appropriate dose metrics for the purpose of this analysis.
31 The first step in the current analysis is to convert the inhalation doses, given as ppm
32 values from the studies, to an internal dose surrogate or dose metric using the EPA PBPK model
33 (see Section 3.4). Predicted AUC values for methanol in the blood of rats and humans are
34 summarized in Table C-l.
35
C-2 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-l. EPA's PBPK model estimates of methanol blood levels (AUC) in rats
following inhalation exposures
Exposure level (ppm)
500
1,000
2000
Methanol in blood AUC (hr x mg/L)a
in rats
79.2
226.7
967.8
aAUC values were obtained by simulating 22 hour/day exposures for 5 days and calculated for the last 24 hours of
that period.
1 The current BMD technical guidance (U.S. EPA, 2000b) suggests that in the absence of
2 knowledge as to what level of response to consider adverse, a change in the mean equal to
3 1 control S.D. from the control mean can be used as a BMR for continuous endpoints. However,
4 it has been suggested that other BMRs, such as 5% change relative to estimated control mean,
5 are also appropriate when performing BMD analyses on fetal weight change as a developmental
6 endpoint (Kavlock et al., 1995). Therefore, in this assessment, both a 1 control mean S.D.
7 change and a 5% change relative to estimated control mean were considered. All models were fit
8 using restrictions and option settings suggested in the EPA BMD technical guidance document
9 (U.S. EPA, 2000b).
C.I.2. BMD Approach with a BMR of 1 Control Mean S.D. -Gestation and into the
Postnatal Period (NEDO, 1987)
10 A summary of the results most relevant to the development of a POD using the BMD
11 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 6 weeks in male
12 rats exposed to methanol throughout gestation and continuing into the FI generation, with a
13 BMR of 1 control mean S.D, is provided in Table C-2. The 6 week male brain weight responses
14 were chosen because they resulted in lower BMD and BMDL estimates than male responses at 3
15 and 8 weeks and female responses at any time point (data not shown). Model fit and was
16 determined by statistics (AIC and $ residuals of individual dose groups) and visual inspection,
17 as recommended by EPA (2000b). There is a 2.5-fold range of BMDL estimates from adequately
18 fitting models, indicating considerable model dependence. In addition, the fit of the Hill and
19 more complex Exponential models is better than the other models in the dose region of interest
20 as indicated by a lower scaled residual at the dose group closest to the BMD (0.09 versus -0.67
21 or -0.77) and visual inspection. In accordance with EPA BMD Technical Guidance (EPA,
22 2000b), the BMDL from the Hill model (bolded), is selected as the most approriate basis for an
23 RfC derivation because it results in the lowest BMDL from among a broad range of BMDLs and
24 provides a superior fit in the low dose region nearest the BMD. The BMDLiso was determined
25 to be 90.9 hr x mg/L using the 95% lower confidence limit of the dose-response curve expressed
26 in terms of the AUC for methanol in blood.
C-3 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-2. Comparison of BMDiSD results for decreased brain weight in male rats at 6
weeks of age using modeled AUC of methanol as a dose metric
Model
Linear
2nd degree polynomial
3rd degree polynomial
Power
Hillb
Exponential 2
Exponential 3
Exponential 4
Exponential 5
BMD1SD (AUC,
hr x mg/L) a
278.30
278.30
278.30
278.30
170.57
260.94
260.94
172.08
172.08
BMDL1SD
(AUC,
hr x mg/L)a
225.30
225.30
225.30
225.30
90.93
209.10
209.10
96.93
96.93
^-value
0.5376
0.5376
0.5376
0.5376
0.8366
0.612
0.612
0.8205
0.8205
AICC
-203.84
-203.84
-203.84
-203.84
-203.04
-204.10
-204.10
-203.10
-203.10
Scaled residual"1
-0.77
-0.77
-0.77
-0.77
0.09
-0.67
-0.67
0.09
0.09
aThe BMDL is the 95% lower confidence limit on the AUC estimated to decrease brain weight by 1 control mean
S.D. using BMDS model options and restrictions suggested by EPA BMD technical guidance (U.S. EPA, 2000b).
bThere is a 2.5-fold range of BMDL estimates from adequately fitting models, indicating considerable model
dependence. In addition, the fit of the Hill and more complex Exponential models is better in the dose region of
interest as indicated by a lower scaled residual at the dose group closest to the BMD (0.09 versus -0.67 or -0.77)
and visual inspection. Thus, in accordance with EPA BMD Technical Guidance (EPA, 2000), the BMDL from the
Hill model (bolded) is considered the most appropriate POD for us in an RfC derivation.
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
d%2d residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest to
the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD. Residuals
that exceed 2.0 in absolute value should cause one to question model fit in this region.
Source: NEDO(1987).
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hilm-6wk-brwHil-
Restrict.(d)
Gnuplot Plotting File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hilm-6wk-
brwHil-Restrict.pit
Mon Aug 24 13:08:09 2009
BMDS Model Run
The form of the response function is:
Y[dose] = intercept + v*dose/xn/(k^n + dose^n)
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
v is set to -0.5
Power parameter restricted to be greater than 1
A constant variance model is fit
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
C-4
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
T^ /
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 0.00539333
rho = 0 Specified
intercept = 1.78
v = -0.5 Specified
n = 1.2699
k = 924.206
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter (s) -rho -v
have been estimated at a boundary point, or have
the user,
and do not appear in the correlation matrix )
alpha intercept n k
alpha 1 1.7e-008 -1.7e-008 -2.9e-009
intercept 1.7e-008 1 -0.56 -0.73
n -1.7e-008 -0.56 1 0.15
k -2.9e-009 -0.73 0.15 1
Parameter Estimates
95.0% Wald
Variable Estimate Std. Err. Lower Conf. Limit
alpha 0.00496114 0.0010023 0.00299667
intercept 1.77806 0.0193152 1.7402
n 1.07212 0.217805 0.645228
k 915.874 179.818 563.437
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev
0 12 1.78 1.78 0.07 0.0704
79.2 12 1.74 1.74 0.09 0.0704
226.7 11 1.69 1.69 0.06 0.0704
967.8 14 1.52 1.52 0.07 0.0704
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma/x2
Model A2 : Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma (i) ^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma/x2
Model A3 uses any fixed variance parameters that
been specified by
Confidence Interval
Upper Conf. Limit
0.00692562
1.81591
1.49901
1268.31
Scaled Res.
0.0956
-0.21
0.159
-0.0355
C-5
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
were specified by the user
Model R: Yi = Mu + e(i)
Var{e (i) } = Signal
Likelihoods of Interest
Model Log (likelihood) #
Al 105.539862
A2 106.570724
A3 105.539862
fitted 105.499930
R 77.428662
Explanation of Tests
Param' s AIC
5 -201.079724
8 -197.141449
5 -201.079724
4 -202.999859
2 -150.857324
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit?
(Note: When rho=0 the results of Test 3
Tests of Interest
Test -2*log (Likelihood Ratio) Test
Test 1 58.2841 6
Test 2 2.06173 3
Test 3 2.06173 3
Test 4 0.0798644 1
The p-value for Test 1 is less than .05.
(A3 vs. fitted)
and Test 2 will be the same.)
df p-value
<.0001
0.5597
0.5597
0.7775
There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .1.
model appears to be appropriate here
The p-value for Test 3 is greater than .1.
to be appropriate here
The p-value for Test 4 is greater than .1.
to adeguately describe the data
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard
Confidence level = 0.95
BMD = 169.597
A homogeneous variance
The modeled variance appears
The model chosen seems
deviations from the control mean
C-6
DRAFT - DO NOT CITE OR QUOTE
-------
BMDL =
103.912
Hill Model with 0.95 Confidence Level
o
Q.
o:
c
(0
OJ
1.85
1.8
1.75
1.7
1.65
1.6
1.55
1.5
1.45
Hill
BMDL BMD
200
400 600
dose
800
1000
09:05 08/24 2009
Figure C-l. Hill model, BMR of 1 Control Mean S.D. - Decreased Brain weight in
male rats at 6 weeks age versus AUC, Fl Generation inhalational study
Source: NEDO(1987).
1 Once the BMDLiso was obtained in units of hr x mg/L, it was used to derive a chronic
2 inhalation reference value. The first step is to calculate the HEC using the PBPK model
3 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
4 describes the relationship between predicted methanol AUC and the human equivalent inhalation
5 exposure concentration (HEC) in ppm.
6
7
BMDLHEC (ppm)= 0.0224*BMDLisD+(1334*BMDLisD)
BMDLHEC (ppm)= 0.0224*374.7+(1334*374.7)/(794+ 90.9) = 139 ppm
Next, because RfCs are typically expressed in units of mg/m , the HEC value in ppm
was converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
C-7
DRAFT - DO NOT CITE OR QUOTE
-------
HEC (mg/m3) = 1.31 x 139 ppm = 182 mg/m3
2 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
3 associated with animal to human differences, 10 for consideration of human variability, and 3 for
4 database deficiencies) to obtain the chronic inhalation reference value:
5 RfC (mg/m3) = 182 mg/m3 -100 = 1.8 mg/m3
C.I.3. BMD Approach with a BMR of 0.05 Change Relative to Estimated Control Mean -
Gestation and into the Postnatal Period (NEDO, 1987)
6 A summary of the results most relevant to the development of a POD using the BMD
7 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 6 weeks in male
8 rats exposed to methanol throughout gestation and continuing into the FI generation, with a
9 BMR of 0.05 change relative to estimated control mean, is provided in Table C-3. The 6 week
10 male brain weight responses were chosen because they resulted in lower BMD and BMDL
11 estimates than male responses at 3 and 8 weeks and female responses at any time point (data not
12 shown). Model fit was determined by statistics (AIC and $ residuals of individual dose groups)
13 and visual inspection, as recommended by U.S. EPA (2000b). There is a 2.4-fold range of
14 BMDL estimates from adequately fitting models, indicating considerable model dependence. In
15 addition, the fit of the Hill and more complex Exponential models is better than the other models
16 in the dose region of interest as indicated by a lower scaled residual at the dose group closest to
17 the BMD (0.09 versus -0.67 or -0.77) and visual inspection. In accordance with EPA BMD
18 Technical Guidance (EPA, 2000b), the BMDL from the Hill model (bolded), is selected as the
19 most approriate basis for an RfC derivation because it results in the lowest BMDL from among a
20 broad range of BMDLs and provides a superior fit in the low dose region nearest the BMD. The
21 BMDLos was determined to be 123.9 hr x mg/L, using the 95% lower confidence limit of the
22 dose-response curve expressed in terms of the AUC for methanol in blood.
C-8 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-3. Comparison of BMD0s results for decreased brain weight in male rats at 6
weeks of age using modeled AUC of methanol as a dose metric
Model
Linearb
2nd degree polynomial
3rd degree polynomial
Power
Hill
BMD05 (AUC,
hr x mg/L)a
344.49
344.49
344.49
344.49
223.18
BMDLos (AUC,
hr x mg/L)a
297.92
297.92
297.92
297.92
123.87
/7-value
0.5376
0.5376
0.5376
0.5376
0.8366
AICC
-203.84
-203.84
-203.84
-203.84
-203.04
Scaled Residual"1
-0.77
-0.77
-0.77
-0.77
-0.09
Exponential! 325.82 278.26 0.612 -204.10 -0.67
Exponentials 325.82 278.26 0.612 -204.10 -0.67
Exponential 4 223.94 129.97 0.8205 -203.10 0.09
Exponentials 223.94 129.97 0.8205 -203.10 0.09
The BMDL is the 95% lower confidence limit on the AUC estimated to decrease brain weight by 5% using BMDS
model options and restrictions suggested by EPA BMD technical guidance (U.S. EPA, 2000b).
bThere is a 2.4-fold range of BMDL estimates from adequately fitting models, indicating considerable model
dependence. In addition, the fit of the Hill and more complex Exponential models is better in the dose region of
interest as indicated by a lower scaled residual at the dose group closest to the BMD (0.09 versus -0.67 or -0.77)
and visual inspection. Thus, in accordance with EPA BMD Technical Guidance (EPA, 2000), the BMDL from the
Hill model (bolded) is considered the most appropriate POD for us in an RfC derivation.
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
dj^d residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest to
the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD. Residuals
that exceed 2.0 in absolute value should cause one to question model fit in this region.
Source: NEDO(1987).
1
2
O
4
5
6
7
c
O
9
10
1 1
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hilm-6wk-brwHil-
Restrict. (d)
Gnuplot Plotting File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hilm-6wk-
brwHil-Restrict .pit
Mon Aug 24
BMDS Model Run
The form of the response function is:
Y[dose] = intercept + v*dose/xn/ (k^n + dose^n)
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
Power parameter restricted to be greater than 1
A constant variance model is fit
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
09:40:25 2009
C-9
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
Default Initial Parameter Values
alpha
rho
intercept
v
n
k
0.00539333
0
1.78
-0.26
1.08282
401.076
Specified
the user,
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho -n
have been estimated at a boundary point, or have been specified by
and do not appear in the correlation matrix )
k
6.2e-009
-0.64
-0.99
1
alpha
intercept
V
k
alpha
1
4e-009
-1.2e-008
6.2e-009
intercept
4e-009
1
0.54
-0.64
V
-1.2e-008
0.54
1
-0.99
Parameter Estimates
95.0% Wald Confidence Interval
Variable
alpha
intercept
v
n
k
Estimate
0.00495736
1.77823
-0.600699
1
1284.65
Std. Err.
0.00100154
0.0184949
0.340055
NA
1317.9
Lower Conf. Limit
0.00299438
1.74198
-1.2672
-1298.39
Upper Conf. Limit
0.00692034
1.81448
0.0657974
3867.68
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
0
79.2
226.7
967.8
12
12
11
14
1.78
1.74
1.69
1.52
1.78
1.74
1. 69
1.52
0.07
0.09
0.06
0.07
0.0704
0.0704
0.0704
0.0704
0.0872
-0.165
0.0884
-0.00677
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma/x2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
C-10
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = Signal
Model A3 uses any fixed variance parameters that
were specified by the user
Model R: Yi = Mu + e(i)
Var{e (i) } = Signal
Likelihoods of Interest
Model Log(likelihood)
Al 105.539862
A2 106.570724
A3 105.539862
fitted 105.518584
R 77.428662
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
# Param' s
5
8
5
4
2
AIC
-201.079724
-197.141449
-201.079724
-203.037168
-150.857324
Test
Test 1
Test 2
Test 3
Test 4
Tests of Interest
-2*log(Likelihood Ratio) Test df
58.2841
2.06173
2.06173
0.0425562
p-value
<.0001
0.5597
0.5597
0.8366
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .1. A homogeneous variance
model appears to be appropriate here
The p-value for Test 3 is greater than .1. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems
to adeguately describe the data
Benchmark Dose Computation
Specified effect = 0.05
Risk Type = Relative risk
Confidence level = 0.95
BMD = 223.178
C-ll
DRAFT - DO NOT CITE OR QUOTE
-------
BMDL =
123.872
Hill Model with 0.95 Confidence Level
o
Q.
o:
c
(0
OJ
1.85
1.8
1.75
1.7
1.65
1.6
1.55
1.5
1.45
Hill
BMDL
BMD
200
400 600
dose
800
1000
09:40 08/24 2009
Figure C-2. Hill model, BMR of 0.05 relative risk - Decreased Brain weight in male
rats at 6 weeks age versus AUC, FI Generation inhalational study.
Source: NEDO(1987).
1 Once the BMDLos was obtained in units of hr x mg/L, it was used to derive a chronic
2 inhalation reference value. The first step is to calculate the HEC using the PBPK model
3 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
4 describes the relationship between predicted methanol AUC and the human equivalent inhalation
5 exposure concentration (HEC) in ppm.
6 BMDLHEC (ppm)= 0.0224*BMDL05+(1334*BMDLo5)/(794+ BMDL05)
7 BMDLHEC (ppm)= 0.0224*503.0 + ((1334*123.9)7(794+ 123.9)) = 183 ppm
8 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in
9 ppm was converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
C-12
DRAFT - DO NOT CITE OR QUOTE
-------
HEC (mg/m3) =1.31 x 183 ppm = 240 mg/m3
2 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
3 associated with animal to human differences, 10 for consideration of human variability, and 3 for
4 database deficiencies) to obtain the chronic inhalation reference value:
5 RfC (mg/m3) = 240 mg/m3 - 100 = 2.4 mg/m3
C.2. DECREASED BRAIN WEIGHT IN MALE RATS EXPOSED DURING
GESTATION ONLY (GD7-GD17) (NEDO, 1987)
6 Cmax, as calculated by the EPAs PBPK model, was selected as the dose metric for this
7 exposure scenario, in concordance with the choice of this dose metric for the increased incidence
8 of cervical rib in mice in the Rogers et al. study (1993). Exposures occurred only during the
9 major period of organogenesis in both studies. As there is evidence that Cmaxis a better predictor
10 of response than AUC for incidence of cervical rib (see Appendix D), it was assumed appropriate
11 to consider Cmax the better predictor for decreased brain weight as well.
12 The first step in the current analysis is to convert the inhalation doses, given as ppm
13 values from the studies, to an internal dose surrogate or dose metric using the EPA PBPK model
14 (see Section 3.4). Predicted Cmax values for methanol in the blood of rats are summarized in
15 Table C-4.
Table C-4. EPA's PBPK model estimates of methanol blood levels (Cmax) in rats
following inhalation exposures
Exposure level (ppm)
200
1,000
5000
Methanol in lood Cmax (mg /L)a in
rats
1.2
10.6
630.5
aCmax values were obtained by simulating 22 hr/day exposures
16 The current BMD technical guidance (U.S. EPA, 2000b) suggests that in the absence of
17 knowledge as to what level of response to consider adverse, a change in the mean equal to
18 1 control S.D. from the control mean can be used as a BMR for continuous endpoints. However,
19 it has been suggested that other BMRs, such as 5% change relative to estimated control mean,
20 are also appropriate when performing BMD analyses on fetal weight change as a developmental
21 endpoint (Kavlock et al., 1995). Therefore, in this assessment, both a 1 control mean S.D.
22 change and a 5% change relative to estimated control mean were considered. All models were fit
23 using restrictions and option settings suggested in the EPAs BMD technical guidance document
24 (U.S. EPA, 2000b).
C-13 DRAFT - DO NOT CITE OR QUOTE
-------
C.2.1. BMD Approach with a BMR of 1 Control Mean S.D. (GD7-GD17) (NEDO, 1987)
1 A summary of the results most relevant to the development of a POD using the BMD
2 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 8 weeks in male
3 rats exposed to methanol during gestation from days 7-17, with a BMR of 1 control mean S.D, is
4 provided in Table C-5. Male brain weight responses were chosen because they resulted in lower
5 BMD and BMDL estimates than female responses (data not shown). Model fit was determined
6 by statistics (AIC and j^ residuals of individual dose groups) and visual inspection, as
7 recommended by EPA (2000b). The polynomial and power models reduced to linear form and
8 returned identical modeling results. In contrast, the more complex Hill and Exponential
9 models, which estimate a response "plateau" or asymptote, returned similar, markedly nonlinear
10 results. This is because these models approximated the response "plateau" to be near the
11 maximum drop in brain weight observed in the study (approximately 10% at the high dose),
12 resulting in a distinctly "L" shaped dose-response curve.89 In this case, the only PBPK model
13 estimated Cmax dose that is associated with a significant response over controls, the high-dose, is
14 60-fold higher than the mid-dose Cmax estimate. Thus, there are many plausible curve shapes
15 and,consequently, a wide range of BMDL estimates. Per EPA (2000b) guidance and to err on the
16 side of public health protection, the lowest BMDLiso of 10.26 mg methanol/L in blood estimated
17 from adequate and plausible models was chosen for use in the RfC derivation. However, it
18 should be noted that there is a great deal of uncertainty and model dependence associated with
19 these dose-response data.
Table C-5. Comparison of BMDisD results for decreased brain weight in male rats at 8
weeks of age using modeled Cmax of methanol as a dose metric
Model
Linear
2nd degree polynomial
3rd degree polynomial
Power
BMD1SD (Cmax,
mg/L)a
207.18
207.18
207.18
207.18
BMDL1SD (Cmax,
mg/L)a
135.22
135.22
135.22
135.22
/7-value
0.7881
0.7881
0.7881
0.7881
AIC
-173.12
-173.12
-173.12
-173.12
Scaled residuald
-0.43
-0.43
-0.43
-0.43
Hillb 43.08 10.26 0.9602 -171.59 -0.10
Exponential! 199.98 127.55 0.9494 -173.13 -0.42
Exponentials 199.98 127.55 0.9494 -173.13 -0.42
Exponential 4b 39.53 10.26 Not reported -171.59 0.10
89 The extent of the "L" shape is dependent on the asymptote term, or "plateau" level, estimated for the data. If the
asymptote term (v) in the Hill model is set to -.4 (representing a 20% drop from the control brain weight of 2
grams), the model result is more linear and the BMD and BMDL estimates are approximately 4-fold higher.
C-14 DRAFT - DO NOT CITE OR QUOTE
-------
aThe BMDL is the 95% lower confidence limit on the Cmax estimated to decrease brain weight by 1 control mean
S.D. using BMDS model options and restrictions suggested by EPA BMD technical guidance (U.S. EPA, 2000b).
bPer EPA (2000b) guidance and to err on the side of public health protection, the lowest BMDL1SD of 10.26 mg
methanol/L in blood estimated from adequate and plausible models was chosen for use in the RfC derivation
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
d% d residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest
to the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD.
Residuals that exceed 2.0 in absolute value should cause one to question model fit in this region.
Source: NEDO (1987)
1 ====================================================================
2 Hill Model. (Version: 2.14; Date: 06/26/2008)
3 Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-only\hilm-8wk-
4 brwHil-Restrict.(d)
5 Gnuplot Plotting File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-only\hilm-
6 8wk-brwHil-Restrict.plt
7 Tue Aug 25 12:40:30 2009
8 ====================================================================
9
10 BMDS Model Run
j j
12
13 The form of the response function is:
14
15 Y[dose] = intercept + v*dose/xn/(k^n + dose^n)
16
17
18 Dependent variable = Mean
19 Independent variable = Dose
20 Power parameter restricted to be greater than 1
21 The variance is to be modeled as Var(i) = exp(lalpha + rho * ln(mean(i)))
22
23 Total number of dose groups = 4
24 Total number of records with missing values = 0
25 Maximum number of iterations = 250
26 Relative Function Convergence has been set to: le-008
27 Parameter Convergence has been set to: le-008
28
29
30
31 Default Initial Parameter Values
32 lalpha = -4.68678
33 rho = 0
34 intercept = 2
35 v = -0.19
36 n = 0.861776
37 k = 303.331
38
39
40 Asymptotic Correlation Matrix of Parameter Estimates
41
42 ( *** The model parameter(s) -n
43 have been estimated at a boundary point, or have been specified by the user,
44 and do not appear in the correlation matrix )
45
46 lalpha rho intercept v k
47
48 lalpha 1 -1 -0.083 0.6 -0.18
49
50 rho -1 1 0.096 -0.6 0.18
51
C-15 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
intercept
v
k
-0.083
0.6
-0.18
0.096
-0. 6
0.18
1
0.19
-0.55
0.19
1
-0.73
-0.55
-0.73
1
Parameter Estimates
Variable
lalpha
rho
intercept
v
n
k
Estimate
7.03732
-18.1432
2.0068
-0.232906
1
121.949
Std. Err.
4.98399
7.32604
0.0134454
0.0881494
NA
194.687
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
-2.73112 16.8058
-32.502 -3.78448
1.98045 2.03316
-0.405676 -0.0601362
-259.631
503.529
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
0
1.2
10.6
630.5
11
11
12
10
2
2.01
1.99
1.81
2.01
2
1.99
1.81
0.047
0.075
0.072
0.161
0.0608
0.0614
0.0662
0.154
-0.371
0.295
0.0954
-0.0338
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma/x2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + rho*ln(Mu(i)))
Model A3 uses any fixed variance parameters that
were specified by the user
Model R: Yi = Mu + e(i)
Var{e (i) } = Sigma/x2
Likelihoods of Interest
Model
Al
A2
A3
fitted
R
Log (likelihood)
83.205960
92.060485
90.797178
90.795933
70.761857
# Param' s
5
8
6
5
2
AIC
-156.411920
-168.120970
-169.594356
-171.591867
-137.523714
Explanation of Tests
C-16
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Test
Test 1
Test 2
Test 3
Test 4
Tests of Interest
-2*log(Likelihood Ratio) Test df
42.5973
17.7091
2.52661
0.00248896
p-value
<.0001
0.000505
0.2827
0.9602
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .1.
model appears to be appropriate
A non-homogeneous variance
The p-value for Test 3 is greater than .1.
to be appropriate here
The p-value for Test 4 is greater than .1.
to adeguately describe the data
The modeled variance appears
The model chosen seems
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 43.0842
BMDL = 10.2551
C-17
DRAFT - DO NOT CITE OR QUOTE
-------
Hill Model with 0.95 Confidence Level
2.05
o
Q.
(/)
0)
o:
c
(0
0)
1.7
600
12:4008/252009
Figure C-3. Hill model, BMR of 1 Control Mean S.D. - Decreased Brain weight in
male rats at 8 weeks age versus Cmax, Gestation only inhalational study.
Source: NEDO(1987).
1 Once the BMDLiso was obtained in units of mg/L, it was used to derive a chronic
2 inhalation reference value. The first step is to calculate the HEC using the PBPK model
3 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
4 describes the relationship between predicted methanol AUC and the human equivalent inhalation
5 exposure concentration (HEC) in ppm. This equation can also be used to estimate model
6 predictions for HECs from Cmax values because Cmax values and AUC values, were estimated at
7 steady-state for constant 24-hour exposures (i.e., AUC = 24 x Cmax).
8 BMDLHEC (ppm) = 0.0224*BMDLisD*24+(1334*BMDLisD*24)/(794+ BMDLiSD*24)
9 BMDLHEC (ppm) = 0.0224*10.3*24 + ((1334*10.3*24)7(794+ 10.3*24)) = 322 ppm
C-18
DRAFT - DO NOT CITE OR QUOTE
-------
1 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
2 converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
3 HEC (mg/m3) = 1.31 x 322 ppm = 422 mg/m3
3
4 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
5 associated with animal to human differences, 10 for consideration of human variability, and 3 for
6 database deficiencies) to obtain the chronic inhalation reference value:
7 RfC (mg/m3) = 422 mg/m3 - 100 = 4.2 mg/m3
C.2.2. BMD Approach with a BMR of 0.05 Change Relative to Control Mean (GD7-GD17)
(NEDO, 1987)
8 A summary of the results most relevant to the development of a POD using the BMD
9 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 8 weeks in male
10 rats exposed to methanol during gestation from days 7 to 17, with a BMR of 0.05 change relative
11 to estimated control mean, is provided in Table C-6. Model fit was determined by statistics (AIC
12 and ^ residuals of individual dose groups) and visual inspection, as recommended by EPA
13 (2000b). Modeling considerations and uncertainties for this dataset were discussed in C.2.1 and,
14 as was done for the BMR of 1 S.D., the lowest BMDLos of 21.07 mg methanol/L in blood
15 estimated from adequate and plausible models was chosen for use in the RfC derivation.
C-19 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-6. Comparison of BMD0s modeling results for decreased brain weight in male
rats at 8 weeks of age using modeled Cmax of methanol as a common dose metric
Model
Linearb
2nd degree polynomial
3rd degree polynomial
Power
BMD05 (Cmax,
mg/L)a
328.84
328.84
328.84
328.84
BMDL05 (Cmax,
mg/L)a
226.08
226.08
226.08
226.08
^-value
0.7881
0.7881
0.7881
0.9446
AICC
-173.12
-173.12
-173.12
-173.12
Scaled residual11
0.02
0.02
0.02
0.02
Hillb 92.30 Not reported 0.9602 -171.59 0.10
Exponential! 320.62 215.13 0.9494 -173.13 0.02
Exponentials 320.62 215.13 0.9494 -173.13 0.02
Exponential 4b 76.36 21.07 Not reported -171.59 0.10
aThe BMDL is the 95% lower confidence limit on the Cmax estimated to decrease brain weight by 5% using BMDS
model options and restrictions suggested by EPA BMD technical guidance (U.S. EPA, 2000b).
bPer EPA (2000b) guidance and to err on the side of public health protection, the lowest BMDL05 of 21.07 mg
methanol/L in blood estimated from adequate and plausible models was chosen for use in the RfC derivation.
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
dj^d residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest to
the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD. Residuals
that exceed 2.0 in absolute value should cause one to question model fit in this region.
Source: NEDO (1987).
1
2
O
4
5
6
7
/
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Exponential Model. (Version: 1.61; Date: 7/24/2009)
Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-only\expm-8wk
brwSetting. (d)
Gnuplot Plotting File:
Tue Aug 25 14:15:15 2009
BMDS Model
The form
Model
Model
Model
Model
Run
„
of the response function by Model:
2:
3:
4:
5:
Note: Y[dose]
sign =
sign =
Model
Model
Model
Dependent
2 is
3 is
4 is
Y[dose] = a
Y[dose] = a
Yfdose] = a
Yfdose] = a
* expfsign * b * dose}
* expfsign * (b * dosej^d}
* [c-(c-l) * exp{-b * dose}]
* [c-(c-l) * exp{-(b * dose)Ad}]
is the median response for exposure = dose;
+1 for increasing trend in data;
-1 for decreasing trend.
nested within
nested within
nested within
Models 3 and 4.
Model 5 .
Model 5.
variable = Mean
Independent variable = Dose
Data are
Variance
assumed to be distributed: normally
Model
: exp(lnalpha
+rho *ln(Y[dose] ) )
The variance is to be modeled as Var(i) = exp(lalpha + log(mean(i)) * rho)
C-20
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
MLE solution provided: Exact
Variable
Inalpha
rho
a
b
c
d
Initial Parameter Values
Model 2 Model 3 Model 4
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
7.32457
-18.5236
2.1105
0.00239093
0.816778
Variable
Inalpha
rho
a
b
c
d
Parameter Estimates by Model
Model 2 Model 3 Model 4
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
7.03418
-18.1386
2.00677
0.00941775
0.902498
NC = No Convergence
Table of Stats From Input Data
Dose N Obs Mean Obs Std Dev
0
1.2
10. 6
630.5
Model
4
11
11
12
10
Do:
1,
10,
630,
36
0
.2
.6
.5
2
2.01
1.99
1.81
Estimated
Est
2
2
1
1
Values
Mean
.007
.005
.988
.812
0
0
0
0
.047
.075
.072
.161
of Interest
Est Std
0.06082
0.06142
0.06617
0.1538
Scaled Residual
-0.3692
0.2932
0.09527
-0.03335
Model 5
NC
NC
NC
NC
NC
NC
Model 5
NC
NC
NC
NC
NC
NC
Other models for which likelihoods are calculated:
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:
Yij = Mu(i) + e(ij)
C-21
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
Var{e(ij)} = exp(lalpha + log(mean(i)
Model R: Yij = Mu + e(i)
Var{e(ij)} = Signal
rho)
Model
Likelihoods of Interest
Log (likelihood) DF
AIC
Al
A2
A3
R
4
83.20596
92.06049
90.61606
70.76186
90.79579
5
8
6
2
5
-156.4119
-168.121
-169.2321
-137.5237
-171.5916
Additive constant for all log-likelihoods = -40.43. This constant added to the
above values gives the log-likelihood including the term that does not
depend on the model parameters.
Explanation of Tests
Test 1: Does response and/or variances differ among Dose levels? (A2 vs. R)
Test 2: Are Variances Homogeneous? (A2 vs. Al)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 6a: Does Model 4 fit the data? (A3 vs 4)
Test
Test 1
Test 2
Test 3
Test 6a
Tests of Interest
-2*log(Likelihood Ratio)
42.6
17.71
2.889
-0.3595
D. F.
6
3
2
1
p-value
< 0.0001
0.000505
0.2359
N/A
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose
levels, it seems appropriate to model the data.
The p-value for Test 2 is less than .1. A non-homogeneous
variance model appears to be appropriate.
The p-value for Test 3 is greater than .1. The modeled
variance appears to be appropriate here.
The p-value for Test 6a is less than .1. Model 4 may not adeguately
describe the data; you may want to consider another model.
Benchmark Dose Computations:
Specified Effect = 0.050000
Risk Type = Relative deviation
Confidence Level = 0.950000
Model
BMD and BMDL by Model
BMD
BMDL
C-22
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
o
o
76.3561
0
0
0
21.0664
0
Not computed
Not computed
Not computed
o
Q.
2.05
1.95
1.9
1.85
1.8
1.75
1.7
Exponential Model 4 with 0.95 Confidence Level
Exponential
BMDL
BMD
0
100
200
300
dose
400
500
600
14:1508/252009
Figure C-4. Exponential model, BMR of 0.05 relative risk - Decreased Brain weight
in male rats at 8 weeks age versus Cmax, Gestation only inhalational study.
Source: NEDO(1987).
6 Once the BMDLos was obtained in units of mg/L, it was used to derive a chronic
7 inhalation reference value. The first step is to calculate the HEC using the PBPK model
8 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
9 describes the relationship between predicted methanol AUC and the human equivalent inhalation
10 exposure concentration (HEC) in ppm. This equation can also be used to estimate model
11 predictions for HECs from Cmax values because Cmax values, and AUC values were estimated at
12 steady-state for constant 24-hour exposures (i.e., AUC = 24 x Cmax).
13 BMDLHEC (ppm) = 0.0224*BMDL05*24+(1334*BMDLo5*24)/(794+ BMDL05*24)
14 BMDLHEC (ppm) = 0.0224*21.07*24 + ((1334*21.07*24)7(794+ 21.07*24)) = 530 ppm
C-23
DRAFT - DO NOT CITE OR QUOTE
-------
1 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
2 converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
HEC (mg/m3) =1.31 x 530 ppm = 695 mg/m3
4 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
5 associated with animal to human differences, 10 for consideration of human variability, and 3 for
6 database deficiencies) to obtain the chronic inhalation reference value:
3
3
7 RfC (mg/m") = 695 mg/m" - 100 = 7.0 mg/m"
C.3. RfC DERIVATIONS USING ROGERS ET AL. (1993)
8 For the purposes of deriving an RfC for methanol from developmental endpoints using
9 the BMD method and mouse data, cervical rib incidence data were evaluated from Rogers et al.
10 (1993). In this paper, Rogers et al. (1993) also utilized a BMD methodology, examining the
11 dosimetric threshold for cervical ribs and other developmental impacts by applying a log-logistic
12 maximum likelihood model to the dose-response data. Using air exposure concentrations (ppm)
13 as their dose metric, a value for the lower 95% confidence limit on the benchmark dose for 5%
14 additional risk in mice was 305 ppm (400 mg/m3), using the log-logistic model. Although the
15 teratology portion of the NEDO study (1987) also reported increases in cervical rib incidence in
16 Sprague-Dawley rats, the Rogers et al. (1993) study was chosen for dose-response modeling
17 because effects were seen at lower doses, it was peer-reviewed and published in the open
18 literature, and data on individual animals were available for a more statistically robust analysis
19 utilizing nested models available in BMDS.
20 The first step in the current BMD analysis is to convert the inhalation doses, given as
21 ppm values from the studies, to an internal dose surrogate or dose metric using the EPA's PBPK
22 model (see Section 3.4). For cervical rib malformations, Cmax of methanol in blood (mg/L) is
23 chosen as the appropriate internal dose metric metric (see Appendix D for further explanation).
24 Predicted Cmax values for methanol in the blood of mice are summarized in Table C-7.
Table C-7. EPA's PBPK model estimates of methanol blood levels (Cmax) in mice
following inhalation exposures
Exposure concentration (ppm)
1
10
50
100
250
Methanol in blood Cmax (mg/L)a
in mice
0.0216
0.218
1.14
2.46
7.83
C-24
DRAFT - DO NOT CITE OR QUOTE
-------
Exposure concentration (ppm)
500
1,000
Methanol in blood Cmax (mg/L)a in mice
26.4
134
aRounded to three significant figures.
1 These Cmax values are then used as the dose metric for the BMD analysis of cervical rib
2 incidence.
3 A 10% BMR level is the value typically calculated for comparisons across chemicals and
4 endpoints for dichotomous responses because this level is near the low end of the observable
5 range for many types of toxicity studies. However, reproductive and developmental studies
6 having a nested design often have a greater sensitivity, and a 5% BMR is typically appropriate
7 for determination of a POD (U.S. EPA, 2000b; Allen et al., 1994). Rogers et al. (1993) utilized a
8 5% added risk for the BMR in the original study. This assessment utilizes both a 10% and 5%
9 extra risk level as a BMR for the determination of a POD.90 The nested suite of models available
10 in BMDS was used to model the cervical rib data. In general, data from developmental toxicity
11 studies are best modeled using nested models, as these models account for any intralitter
12 correlation (i.e., the tendency of littermates to respond similarly to one another relative to other
13 litters in a dose group). All models were fit using restrictions and option settings suggested in
14 the EPAs BMD technical guidance document (U.S. EPA, 2000b).
C.3.1. BMD Approach with a BMR of 0.10 Extra Risk (Rogers et al., 1993)
15 A summary of the results most relevant to the development of a POD using the BMD
16 approach (BMD, BMDL, and model fit statistics) for increased incidence of cervical rib in mice
17 exposed to methanol during gestation from days 6 to 15, with a BMR of 0.10 extra risk, is
18 provided in Table C-8. Model fit was determined by statistics (AIC and $ residuals of individual
19 dose groups) and visual inspection, as recommended by U.S. EPA (2000b). The best model fit to
20 these data (from visual inspection and comparison of AIC values) was obtained using the Nested
21 Logistic (NLogistic) model. The BMDLio was determined to be 94.3 mg/L using the 95% lower
22 confidence limit of the dose-response curve expressed in terms of the Cmax for methanol in blood.
Table C-8. Comparison of BMD modeling results for cervical rib incidence in mice
using modeled Cmax of methanol as a common dose metric
Model
NLogisticb
NCTR
BMD10
(Cmax, mg/L)a
141.492
207.945
BMDL10 (Cmax,
mg/L)a
94.264
103.972
/7-value
0.293
0.241
AIC
1046.84
1048.92
Scaled residual"1
0.649
0.662
90 Starr and Festa (2003) have argued that the Rogers et al. (1993) study's experimental design lacked the statistical
power to detect a 5% risk and that a 5% level lay below the observable response data. However, EPA's BMD
guidance (U.S. EPA, 2000b) does not preclude the use of a BMR that is below observable response data and EPA
has deemed that the Rogers et al. (1993) is adequate for the consideration of a 5% BMR.
C-25
DRAFT - DO NOT CITE OR QUOTE
-------
Model
Rai and Van Ryzin
BMD10
(Cmax, mg/L)a
221.509
BMDL10 (Cmax,
mg/L)a
110.754
/7-value
0.163
AICC
1051.65
Scaled residuald
0.661
aDaily Cmax was estimated using a mouse PBPK model as described in section 3.4 of the methanol lexicological
review; the BMDL is the 95% lower confidence limit on the Cmax for a 10% extra risk (dichotomous endpoints)
estimated by the model using the likelihood profile method (U.S. EPA, 2000b).
bModel choice based on adequate p value (> 0.1), visual inspection, low AIC, and low (absolute) scaled residual.
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
dj^d residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest to
the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD. Residuals
exceeding 2.0 in absolute value should cause one to question model fit in this region.
Source: Rogers etal. (1993).
1 ====================================================================
2 NLogistic Model.
3 (Version: 2.13; Date: 02/20/2007)
4 Input Data File: U:\Methanol\BMDS\CervicalRib\Cmax\NLog_Cmax_10_default.(d)
5 Wed Nov 07 15:45:40 2007
6 ====================================================================
7
8 BMD Method for RfC: Incidence of Cervical Rib in Mice versus Cmax Methanol, GD 6-15
9 inhalational study (Rogers,et al., 1993)
JQ
11 The probability function is:
12
13 Prob. = alpha + thetal*Rij + [1 - alpha - thetal*Rij]/
14
15 [l+exp(-beta-theta2*Rij-rho*log(Dose) ) ] ,
16
17 where Rij is the litter specific covariate.
18
19 Restrict Power rho >= 1.
20
21 Total number of observations = 166
22 Total number of records with missing values = 0
23 Total number of parameters in model = 9
24 Total number of specified parameters = 0
25
26 Maximum number of iterations = 250
27 Relative Function Convergence has been set to: le-008
28 Parameter Convergence has been set to: le-008
29
30 Default Initial Parameter Values
31 alpha = 0.297863
32 beta = -7.94313
33 thetal = 0
34 theta2 = 0
35 rho = 1.09876
36 phil = 0.213134
37 phi2 = 0.309556
38 phi3 = 0.220142
39 phi4 = 0.370587
40
41 Parameter Estimates
42
43 Variable Estimate Std. Err.
44 alpha 0.102434 *
45 beta -4.80338 *
46 thetal 0.0325457 *
47 theta2 -0.436115 *
C-26 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
rho 1 *
phil 0.200733 *
phi2 0.307656 *
phi3 0.212754 *
phi4 0.368426 *
* - Indicates that this value is not calculated.
Log-likelihood: -515.422 AIC: 1046.84
Litter Data
Lit. -Spec. Litter
Dose Cov. Est. Prob. Size Expected Observed
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
1.
1.
2.
2.
2.
2.
2.
3.
3.
3.
3.
4.
4.
4.
4.
5.
5.
5.
5.
5.
5.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.135
.135
.168
.168
.168
.168
.168
.200
.200
.200
.200
.233
.233
.233
.233
.265
.265
.265
.265
.265
.265
.265
.265
.265
.265
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.330
.330
.330
.330
.330
.330
.330
.330
.330
.330
.330
1
1
2
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
7
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
,135
,135
,335
,335
,335
,335
,335
,600
, 600
, 600
,600
,930
,930
,930
,930
,326
,326
,326
,326
,326
,326
,326
,326
,326
,326
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,312
,312
,312
,312
,312
,312
,312
,312
,312
,312
,312
0
0
0
1
0
2
0
0
0
1
1
0
1
0
1
0
1
3
1
0
1
1
0
0
1
3
6
1
0
0
1
2
0
2
3
5
0
3
3
3
5
0
1
2
3
2
3
5
0
2
5
1
Scaled
Residual
-0.3950
-0.3950
-0.5790
1.1490
-0.5790
2.8770
-0.5790
-0.7317
-0.7317
0.4874
0.4874
-0.8699
0.0650
-0.8699
0.0650
-1.0004
-0.2458
1.2632
-0.2458
-1.0004
-0.2458
-0.2458
-1.0004
-1.0004
-0.2458
0.7656
2.6578
-0.4959
-1.1267
-1.1267
-0.4959
0.1348
-1.1267
0.1348
0.7656
2.0271
-1.1267
0.7656
0.7656
0.7656
2.0271
-1.2513
-0.7100
-0.1688
0.3725
-0.1688
0.3725
1.4551
-1.2513
-0.1688
1.4551
-0.7100
C-27
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
0.
0.
0.
0.
0.
0.
0.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
7.
7.
8.
8.
8.
8.
8.
1.
1.
2.
2.
3.
3.
3.
3.
4.
4.
5.
5.
5.
5.
5.
6.
6.
7.
7.
7.
7.
7.
7.
8.
8.
8.
2.
3.
4.
4.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
7.
7.
7.
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.330
.330
.363
.363
.363
.363
.363
.494
.494
.430
.430
.383
.383
.383
.383
.356
.356
.346
.346
.346
.346
.346
.350
.350
.363
.363
.363
.363
.363
.363
.383
.383
.383
.703
. 631
.562
.562
.506
.506
.506
.506
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.444
.444
.444
.444
.444
.444
.444
.444
7
7
8
8
8
8
8
1
1
2
2
3
3
3
3
4
4
5
5
5
5
5
6
6
7
7
7
7
7
7
8
8
8
2
3
4
4
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
2.
2.
2.
2.
2.
2.
2.
0.
0.
0.
0.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
3.
3.
3.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
3.
3.
3.
3.
3.
3.
3.
3.
,312
,312
,902
,902
,902
,902
,902
,494
,494
,859
,859
,150
,150
,150
,150
,425
,425
,732
,732
,732
,732
,732
,099
,099
,543
,543
,543
,543
,543
,543
,068
,068
,068
,406
,892
,250
,250
,530
,530
,530
,530
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,105
,105
,105
,105
,105
,105
,105
,105
2
1
1
4
3
8
2
0
0
0
2
3
1
2
1
3
0
0
4
0
1
0
3
2
3
2
2
2
2
0
2
0
8
2
3
2
1
3
5
3
1
3
3
3
5
6
5
2
0
2
0
5
4
3
2
4
2
0
4
5
1
4
1
5
3
-0.1688
-0.7100
-0.9020
0.5204
0.0463
2.4170
-0.4279
-0.9887
-0.9887
-1.0732
1.4251
1.7287
-0.1400
0.7944
-0.1400
1.1858
-1.0729
-1.0898
1.4275
-1.0898
-0.4604
-1.0898
0.4839
-0.0534
0.2128
-0.2530
-0.2530
-0.2530
-0.2530
-1.1847
-0.4373
-1.2562
2.0195
0.8346
1.1101
-0.1967
-0.9842
0.3091
1.6241
0.3091
-1.0058
0.1162
0.1162
0.1162
1.2556
1.8253
1.2556
-0.4534
-1.5928
-0.4534
-1.5928
1.2556
0.6859
0.1162
-0.4534
0.6859
-0.4534
-1.5658
0.4511
0.9554
-1.0615
0.4511
-1.0615
0.9554
-0.0531
C-28
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
J ~s
60
61
62
63
64
65
66
67
526
526
526
526
526
526
526
526
526
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
7.
7.
7.
7.
8.
8.
8.
9.
9.
1.
1.
1.
2.
2.
3.
3.
3.
3.
4.
4.
4.
4.
4.
4.
4.
5.
5.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
8.
Combine litters
within dose
the
fit of
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0.444
0.444
0.444
0.444
0.437
0.437
0.437
0.443
0.443
0.926
0.926
0.926
0.894
0.894
0.853
0.853
0.853
0.853
0.802
0.802
0.802
0.802
0.802
0.802
0.802
0.743
0.743
0.743
0.743
0.743
0.743
0. 681
0.681
0.681
0. 681
0. 681
0.681
0.681
0. 681
0. 681
0.681
0.681
0. 681
0. 623
0. 623
0.623
0.623
0. 623
0.576
with adjacent
groups until
the
7
7
7
7
8
8
8
9
9
1
1
1
2
2
3
3
3
3
4
4
4
4
4
4
4
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
8
levels of the
the expected count
3.
3.
3.
3.
3.
3.
3.
3.
3.
0.
0.
0.
1.
1.
2.
2.
2.
2.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
,105
,105
,105
,105
,496
,496
,496
,985
,985
,926
,926
,926
,789
,789
,559
,559
,559
,559
,208
,208
,208
,208
,208
,208
,208
,714
,714
,714
,714
,714
,714
,086
,086
,086
,086
,086
,086
,086
,086
,086
,086
,086
,086
,361
,361
,361
,361
,361
, 606
4
1
3
3
0
7
5
0
6
1
1
1
1
2
3
1
1
3
4
4
4
2
3
4
4
1
3
5
5
4
4
6
2
4
5
6
5
4
5
3
6
0
0
7
5
5
7
6
0
0
-1
-0
-0
-1
1
0
-1
0
0
0
0
-1
0
0
-1
-1
0
0
0
0
-1
-0
0
0
-1
-0
0
0
0
0
0
-1
-0
0
0
0
-0
0
-0
0
-2
-2
1
0
0
1
0
-1
.4511
.0615
.0531
.0531
.5793
.5832
.6796
.6270
.8225
.2834
.2834
.2834
.5502
.4157
.5454
.9294
.9294
.5454
.6851
.6851
.6851
.0440
.1795
.6851
.6851
.7660
.4648
.8364
.8364
.1858
.1858
.9945
.0836
.0445
.4750
.9945
.4750
.0445
.4750
.5641
.9945
.1227
.1227
.1486
.2781
.2781
.1486
.7133
.7419
litter-specific covariate
exceeds 3
.0, to help
improve
X^2 statistic to chi-square.
Grouped Data
0
0
0
0
0
0
0
0
Dose
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
Lit
Mean
.-Spec. Cov.
1.0000
2.0000
3.0000
4.0000
5.0000
5.0000
5.0000
5.0000
Expected Observed
0.270
1. 675
2.401
3.722
3.977
3.977
3.977
1.326
0
3
2
2
4
2
1
1
Scaled
Residual
-0.5586
1.0237
-0.2443
-0.8049
0.0098
-0.8614
-1.2970
-0.2458
C-29
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
2005.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
1.0000
2.0000
3.0000
3.0000
4.0000
5.0000
5.0000
5.0000
6.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
2.0000
3.0000
4.0000
5.0000
5.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
9.0000
9.0000
1.0000
3.573
3.573
3.573
3.573
3.573
3.573
3.573
3.573
4.624
4.624
4. 624
4. 624
4.624
4.624
2.312
5.805
5.805
2.902
0.989
1.718
3.449
1.150
2.850
3.463
3.463
1.732
4.199
5.086
5.086
5.086
3.068
3.068
3.068
1.406
1.892
4.500
5.060
5.060
5.592
5.592
5.592
5.592
5.592
5.592
5.592
5.592
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.496
3.496
3.496
3.985
3.985
9
1
1
2
5
5
6
8
1
5
5
5
7
3
1
5
11
2
0
2
6
1
3
4
1
0
5
5
4
2
2
0
8
2
3
3
8
4
6
8
11
2
2
9
5
6
0
4
5
1
4
1
5
3
4
1
3
3
0
7
5
0
6
2
-1
-1
-0
0
0
1
1
-1
0
0
0
0
-0
-0
-0
1
-0
-1
0
1
-0
0
0
-1
-1
0
-0
-0
-1
-0
-1
2
0
1
-0
1
-0
0
0
2
-1
-1
1
-0
0
-1
0
0
-1
0
-1
0
-0
0
-1
-0
-0
-1
1
0
-1
0
.4207
.1474
.1474
.7013
.6367
.6367
.0827
.9747
.3869
.1441
.1441
.1441
.9096
.6214
.7100
.2698
.7418
.4279
.3982
.2488
.3759
.1400
.0799
.2388
.0962
.0898
.3044
.0284
.3578
.0166
.4373
.2562
.0195
.8346
.1101
.8351
.3670
.4926
.1644
.9700
.1785
.4469
.4469
.3729
.2384
.1644
.5658
.4511
.9554
.0615
.4511
.0615
.9554
.0531
.4511
.0615
.0531
.0531
.5793
.5832
.6796
.6270
.8225
2.777
C-30
0.4909
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
2005.0000
Chi-square =
To calculate
at the mean
2.0000
3.0000
3.0000
4.0000
4.0000
4.0000
4.0000
4.0000
4.0000
4.0000
5.0000
5.0000
5.0000
5.0000
5.0000
5.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
7.0000
7.0000
7.0000
7.0000
7.0000
8.0000
105.13 DF
the BMD and BMDL
litter specific
3.577
5.118
5.118
3.208
3.208
3.208
3.208
3.208
3.208
3.208
3.714
3.714
3.714
3.714
3.714
3.714
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.361
4.361
4.361
4.361
4.361
4.606
3
4
4
4
4
4
2
3
4
4
1
3
5
5
4
4
6
2
4
5
6
5
4
5
3
6
0
0
7
5
5
7
6
0
= 98 P-value = 0.
, the litter
covariate of
specific
all the
-0.8022
-0.9786
-0.9786
0.6851
0.6851
0.6851
-1.0440
-0.1795
0.6851
0.6851
-1.7660
-0.4648
0.8364
0.8364
0.1858
0.1858
0.9945
-1.0836
-0.0445
0.4750
0.9945
0.4750
-0.0445
0.4750
-0.5641
0.9945
-2.1227
-2.1227
1.1486
0.2781
0.2781
1.1486
0.7133
-1.7419
2930
covariate is fixed
data: 5.379518
C-31
DRAFT - DO NOT CITE OR QUOTE
-------
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.1
Extra risk
0.95
141.492
94.264
Nested Logistic Model with 0.95 Confidence Level
I
0.8
0.7
0.6
0.5
0.4
0.3
0.2
13:20 11/132007
2000
Figure C-5. Nested Logistic Model, 0.1 Extra Risk - Incidence of Cervical Rib in Mice
versus Cmax Methanol, GD 6-15 inhalational study.
Source: Rogers et al. (1993).
1 Once the BMDLio was obtained in units of mg/L, it was used to derive a chronic
2 inhalation reference value. The first step is to calculate the HEC using the PBPK model
3 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
4 describes the relationship between predicted methanol AUC and the human equivalent inhalation
5 exposure concentration (HEC) in ppm. This equation can also be used to estimate model
6 predictions for HECs from Cmax values because Cmax values and AUC values were estimated at
7 steady-state for constant 24-hour exposures (i.e., AUC = 24 x Cmax).
8
9
BMDLHEC (ppm) = 0.0224*BMDLi0*24+(1334*BMDLi0*24)/(794+ BMDLi0*24)
BMDLHEC (ppm) = 0.0224*94.3*24 + ((1334*94.3*24)7(794+ 94.3*24)) = 1038 ppm
10 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
1 1 converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
12
HEC (mg/m3) = 1.31 x 1038 ppm = 1360 mg/m3
C-32
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
associated with animal to human differences, 10 for consideration of human variability, and 3 for
database deficiencies) to obtain the chronic inhalation reference value:
RfC (mg/m3) =1360 mg/m3 - 100 = 13.6 mg/m3
C.3.2. HMD Approach with a BMR of 0.05 Extra Risk (Rogers et al., 1993)
A summary of the results most relevant to the development of a POD using the BMD
approach (BMD, BMDL, and model fit statistics) for increased incidence of cervical rib in mice
exposed to methanol during gestation from days 6 to 15, with a BMR of 0.05 extra risk, is
provided in Table C-9. Model fit was determined by statistics (AIC and $ residuals of individual
dose groups) and visual inspection, as recommended by U.S. EPA (2000b). The best model fit to
these data (from visual inspection and comparison of AIC values) was obtained using the
NLogistic model. The BMDL05 was determined to be 44.7 mg/L using the 95% lower
confidence limit of the dose-response curve expressed in terms of the Cmaxfor methanol in blood.
Table C-9. Comparison of BMD modeling results for cervical rib incidence in mice
using modeled Cmax of methanol as a common dose metric
Model
NLogisticb
NCTR
Rai and Van
Ryzin
BMDOS
(Cmax,mg/L)a
67.022
101.235
107.838
BMDLos
(Cmax,mg/L)a
44.651
50.618
53.919
/7-value
0.293
0.241
0.163
AICC
1046.84
1048.92
1051.65
Scaled residual"1
0.649
0.662
0.661
aDaily Cma^ was estimated using a mouse PBPK model as described in section 3.4 of the methanol
toxicological review; the BMDL is the 95% lower confidence limit on the Cmax for a 5% extra risk
(dichotomous endpoints) estimated by the model using the likelihood profile method (U.S. EPA,
2000b).
bModel choice based on adequate p value (> 0.1), visual inspection, low AIC, and low (absolute)
scaled residual.
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum
likelihood estimates for the parameters, and P is the number of modeled degrees of freedom (usually
the number of parameters estimated).
Yd residual (measure of how model-predicted responses deviate from the actual data) for the dose
group closest to the BMD scaled by an estimate of its S.D. Provides a comparative measure of model
fit near the BMD. Residuals exceeding 2.0 in absolute value should cause one to question model fit
in this region.
Source: Rogers et al. (1993).
NLogistic Model.
(Version: 2.13; Date: 02/20/2007)
Input Data File: U:\Methanol\BMDS\CervicalRib\Cmax\NLog_Cmax_10_default.(d)
Wed Nov 07 15:45:40 2007
C-33
DRAFT - DO NOT CITE OR QUOTE
-------
BMD Method for RfC: Incidence of Cervical Rib in Mice versus Cmax Methanol, GD 6-15
inhalational study (Rogers, et al., 1993)
J
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
55
56
57
58
59
60
61
62
63
64
65
66
The probability function is:
Prob. = alpha + thetal*Rij + [1 - alpha - thetal*Rij]/
[1+exp (-beta-theta2*Rij-rho*log(Dose) ) ] ,
where Rij is the litter specific covariate.
Restrict Power rho >= 1.
Total number of observations =166
Total number of records with missing values = 0
Total number of parameters in model = 9
Total number of specified parameters = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha = 0.297863
beta = -7.94313
thetal = 0
theta2 = 0
rho = 1.09876
phil = 0.213134
phi2 = 0.309556
phi3 = 0.220142
phi4 = 0.370587
Parameter Estimates
Variable Estimate Std. Err.
alpha 0.102434 *
beta -4.80338 *
thetal 0.0325457 *
theta2 -0.436115 *
rho 1 *
phil 0.200733 *
phi2 0.307656 *
phi3 0.212754 *
phi4 0.368426 *
* - Indicates that this value is not calculated.
Log-likelihood: -515.422 AIC: 1046.84
Litter Data
Lit. -Spec. Litter
Dose Cov. Est. Prob. Size Expected Observed
0.0000 1.0000 0.135 1 0.135 0
0.0000 1.0000 0.135 1 0.135 0
0.0000 2.0000 0.168 2 0.335 0
0.0000 2.0000 0.168 2 0.335 1
0.0000 2.0000 0.168 2 0.335 0
0.0000 2.0000 0.168 2 0.335 2
0.0000 2.0000 0.168 2 0.335 0
0.0000 3.0000 0.200 3 0.600 0
0.0000 3.0000 0.200 3 0.600 0
0.0000 3.0000 0.200 3 0.600 1
0.0000 3.0000 0.200 3 0.600 1
0.0000 4.0000 0.233 4 0.930 0
Scaled
Residual
-0.3950
-0.3950
-0.5790
1.1490
-0.5790
2.8770
-0.5790
-0.7317
-0.7317
0.4874
0.4874
-0.8699
C-34
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
134.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
4.
4.
4.
5.
5.
5.
5.
5.
5.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
8.
8.
8.
8.
8.
1.
1.
2.
2.
3.
3.
3.
3.
4.
4.
5.
5.
5.
5.
5.
6.
6.
7.
7.
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.233
.233
.233
.265
.265
.265
.265
.265
.265
.265
.265
.265
.265
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.298
.330
.330
.330
.330
.330
.330
.330
.330
.330
.330
.330
.330
.330
.363
.363
.363
.363
.363
.494
.494
.430
.430
.383
.383
.383
.383
.356
.356
.346
.346
.346
.346
.346
.350
.350
.363
.363
4
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
1
1
2
2
3
3
3
3
4
4
5
5
5
5
5
6
6
7
7
0.
0.
0.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
0.
0.
0.
0.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
1.
2.
2.
2.
2.
,930
,930
,930
,326
,326
,326
,326
,326
,326
,326
,326
,326
,326
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,786
,312
,312
,312
,312
,312
,312
,312
,312
,312
,312
,312
,312
,312
,902
,902
,902
,902
,902
,494
,494
,859
,859
,150
,150
,150
,150
,425
,425
,732
,732
,732
,732
,732
,099
,099
,543
,543
1
0
1
0
1
3
1
0
1
1
0
0
1
3
6
1
0
0
1
2
0
2
3
5
0
3
3
3
5
0
1
2
3
2
3
5
0
2
5
1
2
1
1
4
3
8
2
0
0
0
2
3
1
2
1
3
0
0
4
0
1
0
3
2
3
2
0.0650
-0.8699
0.0650
-1.0004
-0.2458
1.2632
-0.2458
-1.0004
-0.2458
-0.2458
-1.0004
-1.0004
-0.2458
0.7656
2.6578
-0.4959
-1.1267
-1.1267
-0.4959
0.1348
-1.1267
0.1348
0.7656
2.0271
-1.1267
0.7656
0.7656
0.7656
2.0271
-1.2513
-0.7100
-0.1688
0.3725
-0.1688
0.3725
1.4551
-1.2513
-0.1688
1.4551
-0.7100
-0.1688
-0.7100
-0.9020
0.5204
0.0463
2.4170
-0.4279
-0.9887
-0.9887
-1.0732
1.4251
1.7287
-0.1400
0.7944
-0.1400
1.1858
-1.0729
-1.0898
1.4275
-1.0898
-0.4604
-1.0898
0.4839
-0.0534
0.2128
-0.2530
C-35
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
134.
134.
134.
134.
134.
134.
134.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
7.
7.
7.
7.
8.
8.
8.
2.
3.
4.
4.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
7.
8.
8.
8.
9.
9.
1.
1.
1.
2.
2.
3.
3.
3.
3.
4.
4.
4.
4.
4.
4.
4.
5.
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.363
.363
.363
.363
.383
.383
.383
.703
.631
.562
.562
.506
.506
.506
.506
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.466
.444
.444
.444
.444
.444
.444
.444
.444
.444
.444
.444
.444
.437
.437
.437
.443
.443
.926
.926
.926
.894
.894
.853
.853
.853
.853
.802
.802
.802
.802
.802
.802
.802
.743
7
7
7
7
8
8
8
2
3
4
4
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
7
7
8
8
8
9
9
1
1
1
2
2
3
3
3
3
4
4
4
4
4
4
4
5
2.
2.
2.
2.
3.
3.
3.
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
0.
0.
0.
1.
1.
2.
2.
2.
2.
3.
3.
3.
3.
3.
3.
3.
3.
,543
,543
,543
,543
,068
,068
,068
,406
,892
,250
,250
,530
,530
,530
,530
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,796
,105
,105
,105
,105
,105
,105
,105
,105
,105
,105
,105
,105
,496
,496
,496
,985
,985
,926
,926
,926
,789
,789
,559
,559
,559
,559
,208
,208
,208
,208
,208
,208
,208
,714
2
2
2
0
2
0
8
2
3
2
1
3
5
3
1
3
3
3
5
6
5
2
0
2
0
5
4
3
2
4
2
0
4
5
1
4
1
5
3
4
1
3
3
0
7
5
0
6
1
1
1
1
2
3
1
1
3
4
4
4
2
3
4
4
1
-0.2530
-0.2530
-0.2530
-1.1847
-0.4373
-1.2562
2.0195
0.8346
1.1101
-0.1967
-0.9842
0.3091
1.6241
0.3091
-1.0058
0.1162
0.1162
0.1162
1.2556
1.8253
1.2556
-0.4534
-1.5928
-0.4534
-1.5928
1.2556
0.6859
0.1162
-0.4534
0.6859
-0.4534
-1.5658
0.4511
0.9554
-1.0615
0.4511
-1.0615
0.9554
-0.0531
0.4511
-1.0615
-0.0531
-0.0531
-1.5793
1.5832
0.6796
-1.6270
0.8225
0.2834
0.2834
0.2834
-1.5502
0.4157
0.5454
-1.9294
-1.9294
0.5454
0.6851
0.6851
0.6851
-1.0440
-0.1795
0.6851
0.6851
-1.7660
C-36
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
2005
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
5.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
8.
Combine litters
within dose
the
fit of
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0.743
0.743
0.743
0.743
0.743
0.681
0. 681
0. 681
0.681
0.681
0. 681
0. 681
0.681
0.681
0. 681
0. 681
0.681
0.623
0. 623
0. 623
0. 623
0.623
0.576
with adjacent levels
groups until
the
the expected
XA2 statistic to
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
8
of the
count
3.714
3.714
3.714
3.714
3.714
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.361
4.361
4.361
4.361
4.361
4.606
3
5
5
4
4
6
2
4
5
6
5
4
5
3
6
0
0
7
5
5
7
6
0
-0
0
0
0
0
0
-1
-0
0
0
0
-0
0
-0
0
-2
-2
1
0
0
1
0
-1
.4648
.8364
.8364
.1858
.1858
.9945
.0836
.0445
.4750
.9945
.4750
.0445
.4750
.5641
.9945
.1227
.1227
.1486
.2781
.2781
.1486
.7133
.7419
litter-specific covariate
exceeds 3
.0, to help
improve
chi- square.
Grouped Data
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
134
134
134
134
134
134
134
134
Dose
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
Lit
Mean
.-Spec. Cov.
1.0000
2.0000
3.0000
4.0000
5.0000
5.0000
5.0000
5.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
1.0000
2.0000
3.0000
3.0000
4.0000
5.0000
5.0000
5.0000
Expected Observed
0.
1.
2.
3.
3.
3.
3.
1.
3.
3.
3.
3.
3.
3.
3.
3.
4.
4.
4.
4.
4.
4.
2.
5.
5.
2.
0.
1.
3.
1.
2.
3.
3.
1.
270
675
401
722
977
977
977
326
573
573
573
573
573
573
573
573
624
624
624
624
624
624
312
805
805
902
989
718
449
150
850
463
463
732
0
3
2
2
4
2
1
1
9
1
1
2
5
5
6
8
1
5
5
5
7
3
1
5
11
2
0
2
6
1
3
4
1
0
Scaled
Residual
-0.5586
1.0237
-0.2443
-0.8049
0.0098
-0.8614
-1.2970
-0.2458
2.4207
-1.1474
-1.1474
-0.7013
0.6367
0.6367
1.0827
1.9747
-1.3869
0.1441
0.1441
0.1441
0.9096
-0.6214
-0.7100
-0.2698
1.7418
-0.4279
-1.3982
0.2488
1.3759
-0.1400
0.0799
0.2388
-1.0962
-1.0898
C-37
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
134.
134.
134.
134.
134.
134.
134.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
526.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
2005.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
6.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
2.0000
3.0000
4.0000
5.0000
5.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
9.0000
9.0000
1.0000
2.0000
3.0000
3.0000
4.0000
4.0000
4.0000
4.0000
4.0000
4.0000
4.0000
5.0000
5.0000
5.0000
5.0000
5.0000
5.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
6.0000
4.199
5.086
5.086
5.086
3.068
3.068
3.068
1.406
1.892
4.500
5.060
5.060
5.592
5.592
5.592
5.592
5.592
5.592
5.592
5.592
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.105
3.496
3.496
3.496
3.985
3.985
2.777
3.577
5.118
5.118
3.208
3.208
3.208
3.208
3.208
3.208
3.208
3.714
3.714
3.714
3.714
3.714
3.714
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
4.086
5
5
4
2
2
0
8
2
3
3
8
4
6
8
11
2
2
9
5
6
0
4
5
1
4
1
5
3
4
1
3
3
0
7
5
0
6
3
3
4
4
4
4
4
2
3
4
4
1
3
5
5
4
4
6
2
4
5
6
5
4
5
3
6
0
0
-0
-0
-1
-0
-1
2
0
1
-0
1
-0
0
0
2
-1
-1
1
-0
0
-1
0
0
-1
0
-1
0
-0
0
-1
-0
-0
-1
1
0
-1
0
0
-0
-0
-0
0
0
0
-1
-0
0
0
-1
-0
0
0
0
0
0
-1
-0
0
0
0
-0
0
-0
0
-2
.3044
.0284
.3578
.0166
.4373
.2562
.0195
.8346
.1101
.8351
.3670
.4926
.1644
.9700
.1785
.4469
.4469
.3729
.2384
.1644
.5658
.4511
.9554
.0615
.4511
.0615
.9554
.0531
.4511
.0615
.0531
.0531
.5793
.5832
.6796
.6270
.8225
.4909
.8022
.9786
.9786
.6851
.6851
.6851
.0440
.1795
.6851
.6851
.7660
.4648
.8364
.8364
.1858
.1858
.9945
.0836
.0445
.4750
.9945
.4750
.0445
.4750
.5641
.9945
.1227
C-38
DRAFT - DO NOT CITE OR QUOTE
-------
1 2005.0000 6.0000 4.086 0 -2.1227
2 2005.0000 7.0000 4.361 7 1.1486
3 2005.0000 7.0000 4.361 5 0.2781
4 2005.0000 7.0000 4.361 5 0.2781
5 2005.0000 7.0000 4.361 7 1.1486
6 2005.0000 7.0000 4.361 6 0.7133
7 2005.0000 8.0000 4.606 0 -1.7419
9 Chi-square = 105.13 DF = 98 P-value = 0.2930
10
11 To calculate the BMD and BMDL, the litter specific covariate is fixed
12 at the mean litter specific covariate of all the data: 5.379518
13
i -j
Specified effect = 0.05
Risk Type = Extra risk
Confidence level = 0.95
BMD = 67.0227
BMDL = 44.6514
0.8
0.7
1
g 0.5
t3
ro
0.4
0.3
0.2
Nested Logistic Model with 0.95 Confidence Level
Nr'-trri I nni~tir
T -
^^^-—^
T /K^^^ ^
"I'
BMDL BMD
0 500 1000 1500 2000
dose
11:42 11/162007
Figure C-6. Nested Logistic Model, 0.05 Extra Risk - Incidence of Cervical Rib in
Mice versus Cmax Methanol, GD 6-15 inhalational study.
Source: Rogers et al. (1993).
14 Once the BMDLos was obtained in units of mg/L, it was used to derive a chronic
15 inhalation reference value. The first step is to calculate the HEC using the PBPK model
16 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
17 describes the relationship between predicted methanol AUC and the human equivalent inhalation
18 exposure concentration (HEC) in ppm. This equation can also be used to estimate model
19 predictions for HECs from Cmax values because Cmax values and AUC values were estimated at
20 steady-state for constant 24-hour exposures (i.e., AUC = 24 x Cmax).
21 BMDLHEC (ppm) = 0.0224*BMDL05*24+(1334*BMDLo5*24)/(794+ BMDL05*24)
C-39 DRAFT - DO NOT CITE OR QUOTE
-------
1 BMDLHEC (ppm) = 0.0224*44.7*24 + ((1334*44.7*24)7(794+ 44.7*24)) = 791 ppm
2 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
3 converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
4 HEC (mg/m3) = 1.31 x 791 ppm = 1036 mg/m3
5 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
6 associated with animal to human differences, 10 for consideration of human variability, and 3 for
7 database deficiencies) to obtain the chronic inhalation reference value:
8 RfC (mg/m3) = 1036 mg/m3 - 100 = 10.4 mg/m3
C.4. RFC DERIVATIONS USING BURBACHER ET AL. (1999A,B)
9 The BMD approach was utilized in the derivation of potential chronic inhalation
10 reference values from effects seen in monkeys due to prenatal methanol exposure. Deficits in
11 VDR were evaluated from Burbacher et al. (1999a, 1999b). In the application of the BMD
12 approach, continuous models in the EPAs BMDS, version 1.4. Ic, were fit to the dataset for
13 increased latency in VDR in neonatal monkeys. As the EPAs PBPK model was not
14 parameterized for monkeys, external concentration (ppm) was used as the dose metric.
15 The VDR test, which assesses time (from birth) it takes for an infant to grasp for a
16 brightly colored object containing an applesauce-covered nipple, is a measure of sensorimotor
17 development. Beginning at 2 weeks after birth, infants were tested 5 times/day, 4 days/week.
18 Performance on that test, measured as age from birth at achievement of test criterion (successful
19 object retrieval on 8/10 consecutive trials over 2 testing sessions), was reduced in all treated male
20 infants. The times (days after birth) to achieve the criteria for the VDR test were 23.7 ±4.8
21 (n = 3), 32.4 ± 4.1 (n = 5), 42.7 ± 8.0 (n = 3), and 40.5 ± 12.5 (n = 2) days for males and
22 34.2 ± 1.8 (n = 5), 33.0 ± 2.9 (n = 4), 27.6 ± 2.7 (n = 5), and 40.0 ± 4.0 (n = 7) days for females
23 in the control to 1800 ppm groups, respectively. As discussed in Section 4.3.2, this type of
24 response data is sometimes adjusted to account for premature births by subtracting time (days)
25 premature from the time (days from birth) needed to meet the test criteria (Wilson and Cradock,
26 2004). When this type of adjustment is applied, the times (days after birth or, if shorter, days
27 after control mean gestation length) to achieve the criteria for VDR test were 22.0 ± 9.54 (n = 3),
28 26.2 ± 8.61 (n = 5), 33.3 ± 10.0 (n = 3), and 39.5 ± 16.3 (n = 2) days for males and 32.0 ± 4.3
29 (n = 5), 21.8 ± 5.6 (n = 4), 24.0 ± 5.7 (n = 5), and 32.0 ± 14.8 (n = 7) days for females in the
30 control to 1800 ppm groups, respectively. When these data were modeled within BMDS
31 (version 2.1), there was no significant difference between unadjusted responses and/or variances
32 among the dose levels for males and females combined (p = 0.244), for males only (p = 0.321)
C-40 DRAFT - DO NOT CITE OR QUOTE
-------
1 and for males only with the high-dose group excluded (p = 0.182), or for adjusted responses of
2 males and females combined (p = 0.12), males only (p = 0.448) and males only with the high-
3 dose group excluded (p = 0.586).91 The only data that offered a significant dose-response trend
4 was that for unadjusted (p = 0.0265) and adjusted (p = 0.009) female responses, but the model
5 fits for the adjusted female response data were unacceptable. Only the unadjusted female VDR
6 response data offered both a dose-response trend and acceptable model fits. The modeling
7 results for this data set are presented in Table C-10.
8 The current BMD technical guidance (U.S. EPA, 2000b) suggests that in the absence of
9 knowledge as to what level of response to consider adverse, a change in the mean equal to
10 1 control S.D. from the control mean can be used as a BMR for continuous endpoints. A
11 summary of the results most relevant to the development of a POD using the BMD approach
12 (BMD, BMDL, and model fit statistics) for increased latency of VDR in female neonatal
13 monkeys exposed to methanol with a BMR of 1 control mean S.D. is provided in Table C-10.
14 Model fit was determined by statistics (AIC and $ residuals of individual dose groups) and
15 visual inspection, as recommended by EPA (2000b). The 3rd degree polynomial model returned
16 a lower AIC than the other models.92 The BMDLiso was determined to be 81.7 hrxmg/L, using
17 the 95% lower confidence limit of the dose-response curve expressed in terms of the ppm of
18 external methanol concentration.
91 BMDS continuous models contain a test for dose-response trend, test 1, which compares a model that fits a
distinct mean and variance for each dose group to a model that contains a single mean and variance. The dose
response is considered to be significant if this comparison returns ap value < 0.05.
92 A detailed analysis of this dose response revealed that modeling results, particularly the BMDL estimation, are
very sensitive to the high-dose response. There is no data to inform the shape of the curve between the mid- and
high-exposure levels, making the derivation of a BMDL very uncertain. The data were analyzed without the high
dose to determine if the downward trend in the low- and mid-exposure groups is significant. It was not, so
nonnegative restriction on the (3 coefficients of the poly models was retained.
C-41 DRAFT - DO NOT CITE OR QUOTE
-------
Table C-10. Comparison of BMD modeling results for VDR in female monkeys using
AUC blood methanol as the dose metric
Model
Linear
2nd degree polynomial
3rd degree polynomial
Powerb
Hill
BMD1SD (AUC,
hr x mg/L)a
119.058
114.094
120.176
133.517
132.283
BMDL1SD
(AUC,
hr x mg/L)a
51.9876
59.6412
81.6513
63.0615
-
p-value
0.1440
0.2388
0.2718
0.1112
NA
AICC
110.4492
109.43782
109.17894
111.11010
113.11010
Scaled residual11
0.5380
0.0994
0.0199
0.0000
0.0000
1
2
O
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
aAUC was estimated using a rat PBPK model as described in section 3.4 of the methanol lexicological review; the
BMDL is the 95% lower confidence limit on the AUC of a decrease of 1 control mean S.D. estimated by the model
using the likelihood profile method (U.S. EPA, 2000b).
bModel choice based on adequate p value (> 0.1), visual inspection, low AIC, and low (absolute) scaled residual.
°AIC = Akaike Information Criterion = -2L + 2P, where L is the log-likelihood at the maximum likelihood
estimates for the parameters, and P is the number of modeled degrees of freedom (usually the number of parameters
estimated).
dj^d residual (measure of how model-predicted responses deviate from the actual data) for the dose group closest to
the BMD scaled by an estimate of its S.D. Provides a comparative measure of model fit near the BMD. Residuals
that exceed 2.0 in absolute value should cause one to question model fit in this region.
Source: Burbacheretal. (1999a).
Polynomial Model.
(Version: 2.13; Date: 04/08/2008)
Input Data File: C:\USEPA\BMDS2\Data\Burbacher\PolfemSet.(d)
Gnuplot Plotting File: C:\USEPA\BMDS2\Data\Burbacher\PolfemSet.plt
Fri Dec 12 15:30:29 2008
VDR in female monkeys using AUC blood methanol as the dose metric
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*dose/"2 + ...
Dependent variable = F_VDR
Independent variable = F_Dose
The polynomial coefficients are restricted to be positive
The variance is to be modeled as Var(i) = exp(lalpha + log(mean(i)) * rho)
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
lalpha = 4.07254
rho = 0
beta_0 = 34.2
beta_l = 0
beta_2 = 0
beta_3 = 0
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -beta_l -beta_2
C-42
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
JT
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
have been estimated at a boundary po
the user,
and do not appear in the correlation
lalpha rho beta 0
lalpha 1 -1 -0.0076
rho -1 1 0.0076
beta 0 -0.0076 0.0076 1
beta_3 0.018 -0.018 -0.37
Parameter Estimates
Variable Estimate Std. Err.
Limit
lalpha -13.5062 9.81148
5.72395
rho 4.90831 2.77841
10.3539
beta 0 31.5013 1.49057
34.4228
beta 1 8.36431e-025 NA
beta 2 0 NA
beta 3 3.19775e-006 1.53534e-006
006
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev
0 5 34.2 31.5 4.09
6.73 4 33 31.5 5.83
28.28 5 27.6 31.6 5.94
138.1 7 40 39.9 10.7
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma/x2
Model A2 : Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma (i) ^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + rho*ln (Mu (i) ) )
Model A3 uses any fixed variance parameters that
were specified by the user
Model R: Yi = Mu + e(i)
Var{e (i) } = Sigma/x2
Likelihoods of Interest
Model Log (likelihood) # Param' s
Al -51.042924 5
A2 -47.867444 8
A3 -49.286738 6
fitted -50.589469 4
R -55.013527 2
Explanation of Tests
Test 1: Do responses and/or variances differ among
(A2 versus R)
int, or have been
matrix )
beta_3
0.018
-0.018
-0.37
1
95.0% Wald CI
Lower Conf. Limit
-32.7363
-0.537284
28.5798
1.88544e-007
Est Std Dev Sea
5.55
5.55
5.58
9.93 0
AIC
112.085848
111.734888
110.573475
109.178938
114.027055
Dose levels?
specified by
Upper Conf.
6.20695e-
led Res.
1.09
0.54
-1.59
.0199
C-43
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Test 2: Are Variances Homogeneous? (Al vs A2)
Test 3: Are variances adequately modeled? (A2 versus A3)
Test 4: Does the Model for the Mean Fit? (A3 versus fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Test
Test 1
Test 2
Test 3
Test 4
Tests of Interest
-2*log(Likelihood Ratio) Test df
14.2922
6.35096
2.83859
2.60546
p-value
0.02654
0.09573
0.2419
0.2718
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .1. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is greater than .1. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems
to adequately describe the data
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
Estimated S.D.s from the control mean
0.95
120.176
81.6513
Polynomial Model with 0.95 Confidence Level
50
45
40
35
30
25
20
Polynomial
BMDL
BMD
20
40
60
80
100
120
140
dose
15:30 12/122008
Figure C-7. 3rd Degree Polynomial Model, BMR of 1 Control Mean S.D. - VDR in
female monkeys using AUC blood methanol as the dose metric.
C-44
DRAFT - DO NOT CITE OR QUOTE
-------
Source: Burbacher et al. (1999a, 1999b)
1 Once the BMDLiso was obtained in units of ppm, it was used to derive a chronic
2 inhalation reference value. The first step is to calculate the HEC using the PBPK model
3 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
4 describes the relationship between predicted methanol AUC and the human equivalent inhalation
5 exposure concentration (HEC) in ppm.
6 BMDLHEC (ppm) = 0.0224*BMDLisD+(1334*BMDLisD)/(794+ BMDLiso)
7 BMDLHEC (ppm) = 0.0224*81.7+(1334*81.7)7(794+ 81.7) = 126.3 ppm
8 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
9 converted using the conversion factor specific to methanol of 1 ppm =1.31 mg/m3:
10 HEC (mg/m3) = 1.31 x 126.3 ppm = 165 mg/m3
11 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
12 associated with animal to human differences, 10 for consideration of human variability, and 3 for
13 database deficiencies) to obtain the chronic inhalation reference value:
14 RfC (mg/m3) = 165 mg/m3 -100 = 1.7 mg/m3
C-45 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIXD. RfC DERIVATION- COMPARISON OF DOSE METRICS
D.I. METHODS
D. 1.1. Dose Metric Comparisons
1 Three potential dose metrics were evaluated for possible use in risk extrapolation of
2 methanol-induced developmental effects: AUC of methanol in the blood; Cmax of methanol in the
3 blood; and total metabolism of methanol. The latter metric was considered because
4 developmental effects may be caused by metabolites of methanol, particularly formaldehyde, and
5 formate. These three metrics were evaluated by determining how well they were able to explain
6 the variation in response for incidence of cervical ribs (CR) and supernumerary ribs (SNR) in a
7 concentration-time bioassay by Rogers et al. (1995, raw data obtained from personal
8 communication). In particular, pregnant CD-I mice were exposed to 2,000, 5,000, 10,000, or
9 15,000 ppm methanol for 1, 2, 3, 5, or 7 hours on GD7 and developmental effects evaluated at
10 GDI 7. This endpoint was selected because it was the most sensitive of those examined and gave
11 a reasonable dose-response relationship overall.
12 Initially, the fraction of pups within each litter carrying either or both CR and SNR was
13 calculated, and then the average across all litters in each concentration-time combination was
14 computed. However, as shown in Figure D-l, the resulting data appear to be nonmonotonic,
15 with the responses from 5-hour exposures exceeding those from 7-hour exposures, and the
16 responses from 2-hour exposures exceeding those from 3-hour exposures. It was noted that the
17 study was done with a block-design, where the dams/litters for some concentration-time
18 combination were divided between multiple blocks and the average CR + SNR incidence in
19 controls varied from 30-52% among the 8 blocks. Therefore block-control response (percent)
20 was subtracted from each exposed litter's response (percent) before calculating an average
21 response among litters in a given concentration-time combination. The resulting data are
22 presented in Figure D-1.
D-1 DRAFT - DO NOT CITE OR QUOTE
-------
•\ n^ -T——^^
90
5? 75
OL
W 60
+
O 45
30
A: Raw data starting with ^
grand-mean of controls A- — ^^-'•^'-A
/^xrT o
/- ^ // , -
-^ /^^ r' / ~~ y<.
^ 9f f '
^ / ,''/'
, ./ f'f/
^^^v^^ z — ^
• 7-hr data
— A^ 5-hr data
— • — 3-hr data
0 2-hr data
X 1-hrdata
15 -t 1 1 1 '
0 5000 10000 15000
Concentration (ppm)
75
60
5? 45
tt
W 30
+
0 15
0 C
-15
C
• 7-hr data
— A — 5-hr data
— • — 3-hr data
0 2-hr data
X 1-hrdata
/ *^ O
/X r//' -m
H-?^iK— —mf B: Data minus bkgd in
O each block, then summarized
) 5000 10000 15000
Concentration (ppm)
1
2
3
4
5
Figure D-l. Exposure-response data for methanol-induced CR plus SNR
malformations in mice at various concentration-time combinations. The percent
response in each litter was first calculated, with direct averages shown in the
first panel relative to the grand-mean for the controls. In the second panel,
the percent response in controls for each block of exposures in the study was
first subtracted from each litter's response in that block before taking
averages across litters.
Source: Rogers et al. (1995).
While the correction for background differences does not completely correct the apparent
nonmonotonic dose, the 2-hour response is now less than or below the 3-hour response at 5,000
and 10,000 ppm, and the strong disparity that appeared between the 5- and 7-hour data at 2,000
ppm is eliminated. Overall, the data show a more consistent dependence on duration of
exposure, except for the response to 3 hours of 15,000 ppm methanol. Therefore these
D-2
DRAFT - DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
background-corrected response measures will be used to evaluate the 3 dose metrics, with the
exception that the 3-hour 15,000 ppm data point will be dropped as an outlier. In particular, the
dose-response relationship based on these data will be plotted against each of the dose metrics to
determine which provides the most consistent overall dose-response relationship.
D.2. RESULTS
D.2.1. Dose Metric Comparisons
The average incidence of CR plus SNR from the concentration-time developmental
bioassay of Rogers et al. (1995), with block-specific control values subtracted from each litter
average before calculating overall average responses, is plotted in Figure D-2 against three dose
metrics: AUC, Cmax, and total amount metabolized of methanol (The volume units for Cmax and
AUC were adjusted to put all three data sets on approximately the same scale for comparison).
ID
60
45
E
oTso
-------
1 other two metrics. Thus, Cmax appears to be a better predictor of response than AUC or amount
2 metabolized. Looking at the exposure-response data in Figure D-l, one can see that 2- and
3 3-hour exposures at 5,000 ppm elicit no increase over control, while 5- and 7-hour exposures at
4 this level do.
5 If AUC or amount metabolized were true measures of risk, then one would expect a
6 graded response, where the 2- and 3-hour exposures were intermediate between controls and 5-
7 7-hour exposures. But the lack of response at those shorter times indicates that the concentration
8 (Cmax) has not risen high enough in such a short exposure to cause a response, while it has at the
9 longer durations. From Figure D-2, it appears that a Cmax of 11 mg/cL (1,100 mg/L) is a
10 NOAEL, with a linear increase in CR + SNR from that level to 38 mg/cL, after which the
11 response begins to plateau. Note that while the plot is of response above background, the plateau
12 is effectively at 100% total incidence: the highest points in Figure D-l are from the 7-hour
13 exposures at 15,000 ppm, where actual incidence was 98% (30% in controls); and the next
14 highest points are from the 5-hour 15,000 and 10,000 ppm exposures, where the incidences were
15 94% and 93%, respectively (32% in controls; both from the same block).
D-4 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX E. EVALUATION OF THE CANCER POTENCY OF METHANOL
E.I. INTRODUCTION
1 Two studies were selected for the evaluation of the cancer potency of methanol (Soffritti
2 et al., 2002a, 2000b, 2002c; NEDO, 1987, 1987/2008b). The Soffritti et al. (2002a) study is the
3 only oral study available with effects that show a statistically significant increase in incidence of
4 any cancer endpoints in the treated groups versus the concurrent control group (pair-wise
5 comparison) and is used to derive the POD for deriving an oral cancer slope factor. The NEDO
6 (1987, 1987/2008b) 24-month rat study is the only inhalation study available with effects that
7 show a statistically significant increase in incidence of any cancer endpoints and was used to
8 derive the POD for the inhalation cancer unit risk. A third study, Apaja (1980), reported
9 statistically significant increases in malignant lymphomas in Eppley Swiss Webster mice over
10 historical controls (pair-wise comparison) following drinking water exposure to methanol.
11 Because this study did not involve a concurrent control group it is not used for the derivation of a
12 cancer oral slope factor, but its dose-response is evaluated here for comparative purposes.
E.I.I. Oral CSF POD
13 The Soffritti et al. (2002a) study, conducted by the Ramazzini Foundation, presents a
14 number of challenges if these data are to be used in dose-response modeling to assess the
15 carcinogenic potency of methanol. One challenge, determining the appropriate HED, is best
16 addressed using a PBPK model to derive an HED dose that considers the kinetic differences in
17 humans and the animal model, i.e., species extrapolation. Such a model was developed by the
18 EPA and is not addressed in this appendix; however, the dose metrics derived from that PBPK
19 model are used in the modeling of the data.
20 The other major challenge, which is addressed in this appendix, is how to model the
21 nonstandard protocol by which methanol was tested, as reported in Soffritti et al. (2002a). In
22 most oncogenicity studies, typified by those conducted by the NTP, animals are dosed for
23 104 weeks, with a scheduled sacrifice of all surviving animals at the end of treatment. In the
24 study for methanol reported by Soffritti et al. (2002a), while the animals were treated with
25 methanol for 104 weeks, animals were not euthanized and examined on a specified schedule but
26 were followed until their natural death. It is well known that the incidence of background tumors
27 in a number of organs is different between that seen at a scheduled sacrifice at 104 or 105 weeks
28 and in the same sex/strain that is followed for a lifetime. A higher background incidence can
29 increase the difficulty of detecting chemically related responses (Melnick et al., 2007). Further,
30 performing pathological examinations on tissues collected after natural death can create
E-1 DRAFT -DO NOT CITE OR QUOTE
-------
1 difficulties associated with cell autolysis.93 At the same time, the shorter duration of the 2-year
2 bioassays used at the NTP misses about two thirds of the life span of the rodent, potentially
3 missing late stage or late appearing chemically related tumor responses (Melnick et al., 2007).
4 ERF believes that "cutting short an experiment after two years may mask a possible carcinogenic
5 response," but ERF further suggests that all chronic cancer studies "should continue until
6 spontaneous animal death" (Soffritti et al., 2002a). Soffritti et al. (2002a) cite ERF studies of
7 benzene, xylenes, mancozeb, and vinyl acetate monomer as examples for which carcinogenic
8 responses were observed after the 2-year treatment period.
9 These data were evaluated using three different approaches and two different dose-
10 response models (EPA's multistage cancer and a multistage Weibull time-to-tumor model).
11 These approaches involved using the administered dose and models that would rely upon the
12 published data (Option 1), while another would rely on unpublished information that would be
13 provided by the Ramazzini Foundation (Options 1 and 2), while a third option (Option 3) would
14 also rely upon this unpublished information but would also incorporate results from PBPK
15 modeling for methanol developed by the EPA.
E.I.2. Selection of the Data Modeled and HED
16 The individual animal data from the Ramazzini Foundation study was provided to EPA in
17 the standard NTP format in which the number of days on study, the tissues examined and the
18 tumor types found were given for each animal. The tumors with incidences that were
19 statistically significantly increased or were considered to be rare tumors and considered for dose-
20 response modeling were the incidence of hepatocellular carcinoma in male rats and the incidence
21 of hemolymphoreticular neoplasms in both male and female rats. The incidence of lympho-
22 immunoblastic lymphomas was modeled separately, and the combined incidence of all the
23 lymphomas was considered for dose-response modeling. Table E-2 provides the incidence of
24 these neoplasms reported in each dose group. The incidence of histiocytic sarcomas and myeloid
25 leukemias were not significantly increased in either sex. The incidence of these tumors was not
26 combined with the lymphoblastic lymphomas because they are of a different cell line and the
27 combination is not typically evaluated either for statistical significance or dose-response
28 modeling (McConnell et al., 1986).
29 The drinking water concentrations provided in the Soffritti et al. (2002a) study were
30 converted to doses in mg/kg-day. Initially, an attempt was made to estimate the dose of methanol
31 to individual animals for development of an average dose; however, water consumption
32 information was available only on a cage-by-cage basis. Based on the available information, the
33 average water consumption for each treatment group was calculated using the available data
93 Autolysis may develop in carcasses of animals if they are not processed immediately after death or if the animals
become severely moribund prior to death. These types of changes can compromise pathological diagnosis and
subclassification of neoplasms.
E-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 reported for weeks 1-104. Although individual body weights were available, the corresponding
2 intake was not available. The average body weight over the period of dosing for the experiment
3 (using measurements taken on day 1-day 736) was calculated for each dosed group. A weighted
4 average was calculated for the body weights using the number of animals for which body
5 weights were recorded at each time point. The average body weight and the average water
6 consumption in (mL/day) were used to calculate the mg/kg-day doses. The equation used for
7 this calculation is:
_^ , ., , x Dose(ppm)x WaterConsumption(mL/day)
Dose(mg / kg - day) = rt^ ' —.
1000 x Body Weight(kg)
8 Table E-l provides the values used to obtain the mg/kg-day doses, as well as the resulting
9 mg/kg-day doses. In addition, the average and median times of death were calculated for each
10 group (both dosed and control), for the only the dosed groups combined (excluding the controls),
11 and for all the groups in the study combined (including control). These values were obtained
12 using the reported weeks on study for each animal. One male rat (ID # 129) in the 20,000 ppm
13 group was not examined microscopically and was excluded from the time of death calculations
14 and all modeling. If this animal was included in the calculations for the average and median
15 times of death, the median time of death for all male rat dosed groups would increase from 97-
16 98 weeks; all of the other average and median times of death that include the 20,000 ppm group
17 do not change.
18 In the absence of a kinetic model, extrapolation from animal to human was based on the
19 default assumption of body weight374. This extrapolation was applied to the animal POD
20 estimates to obtain the HEDs reported in Table E-l. This extrapolation was calculated using the
21 average body weight of the dosed animals excluding controls (0.33 kg for the female rats and
22 0.51 kg for the male rats) over the dosing period of the study (through day 736) and 70 kg for the
23 human body weight. The equation used for the body weight4 extrapolation is
[ Animal Body Weight (kg)
.1/4
^ Human Body Weight (kg) j
24 and results in a value of 0.26 for the female rats and 0.29 for the male rats.
E. 1.2.1. Dose-Response Modeling
25 Option 1 - Quantal Dose-Response Modeling. Under this option, the standard default
26 modeling approach outlined in the Cancer Guidelines (U.S. EPA, 2005a) was applied. The
27 cancer bioassay data were fit using a multistage model, which is the current model preferred by
28 EPAs IRIS program for cancer dose-response modeling. The PODs for use in the derivation of
E-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 CSFs were calculated using the multistage-cancer model available in the BMDS program
2 (http ://www. epa. gov/NCEA/bmds) .
3 BMDS was used to estimate BMDs and 95% lower bounds on the BMDs or BMDLs
4 associated with a 10% extra risk (BMDLio). For this assessment, the multistage model was
5 determinded to be an appropriate model for characterization of the dose-response curve in the
6 observable range. At this time, the MOA for the tumors observed following exposure to
7 methanol is not known; therefore, linear extrapolation was conducted to estimate a CSF.
8 Option 2 - Time-to- Tumor Dose-Response Modeling. This option is similar to
9 Option 1; however, rather than the use of quantal models, a time-to-tumor model was applied to
10 the selected datasets. Data for this analysis was provided by the Ramazzini Foundation and can
11 be obtained from their web site (http://www.ramazzini.it/fondazione/study.asp). The same
12 assumptions regarding the FED and low-dose extrapolation were applied. Because BMDS does
13 not include time-to-tumor modeling, the QRISK portion of Statox was relied upon for dose-
14 response modeling. Statox is an internal EPA program that is used for gathering and analyzing
15 animal bioassay data and contains the QRISK component for dose-response modeling. The
16 QRISK component of Statox Version 5.5 fits a multistage Weibull model to the data. The
17 multistage Weibull model is multistage in dose and Weibull in time and essentially assesses the
1 8 probability that a tumor would have been identified at time t. The multistage Weibull model has
19 the form:
p(d t} = 1 - e~v
20 with dose (d) and time (i) as the variables. The parameters estimated by fitting the model to the
21 data are the dose parameters go through q^ the induction time (to) and the power term for
22 time (c).
23 If t0 is interpreted as the time (assumed to be the same for all animals) from when a tumor
24 is observable (i.e., capable of being detected if the animal were to be sacrificed and a necropsy
25 performed) to the time the tumor causes the death of the animal, then these models can be
26 applied to data on incidental and fatal tumors simultaneously. Note that t and t0 only appear in
27 the model in the form of t-t0. To make this explicit, we write P(d,t) = F(d,t-t0). The probability
28 of an incidental tumor by time t is taken to be F(d,t) (t0 = 0) and the probability of a fatal tumor
29 by time t is taken to be F(d,t-t0). There are three possible types of incidence contexts for each
30 animal which contribute separately to the likelihood function for this model. These are:
31 • Censored response - animal died without having the tumor(s) being modeled
32 • Incidental response - the animal died with the tumor(s) but the death was not caused by
33 the tumor(s) (i.e., the time to death from those tumors would have been later than the
34 actual death time); and
E-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 • Fatal incidence - the tumor(s) being modeled was the cause of death.
2 The contribution of each animal to the likelihood is then defined for its time of death ft).
3 The complete likelihood is defined as:
g
Incidence(iJ)=Censored
Inc idenc e (i, j)=Inc idental
n
Incidence(i,j)=Fatal
4 where g is the number of dose groups in the study, including the control group, and /' varies from
5 1 to the total number of animals in the study examined for the tumor type(s) being modeled.
6 As with the quantal modeling, the lower bound on a dose at an extra risk of 10% was
7 estimated. Goodness-of-fit was determined by visually inspecting graphical output of the
8 modeling. AIC values were also calculated for the time-to-tumor model fit.
9 This option was proposed because time-to-tumor modeling is typically applied to account
10 for differences in survival among treated and control groups. However, in this case there were
11 no differences detected in the survival times. Figures E-l and E-2 are graphs of the proportion
12 surviving versus the weeks on study for the female rats and male rats, respectively. In addition,
13 the Life Table program (Thomas et al., 1997) was run on the data. None of the statistical tests in
14 this program indicated a difference in survival between the control and the dosed groups.
15 The protocol used by the Ramazzini Foundation was different from that typically
16 employed in chronic rat bioassays. In typical rodent bioassays, a compound is administered to
17 the animals for approximately 104 weeks, and the animals sacrificed within a short period (days)
18 following the end of treatment. In the Ramazzini Foundation bioassays (i.e., for methanol,
19 formaldehyde, MTBE, and aspartame), the animals were administered compounds for 104 weeks
20 but were allowed to live until a natural death, which, in some animals, occurred months after the
21 completion of chemical administration. This can have an impact on the tumor incidence and
22 therefore, the potential risk of tumor development associated with administration of a given
23 compound. Time-to-tumor modeling was used in this case to attempt to adjust for the extended
24 life span of the some of the animals in this study.
25 For time-to-tumor modeling, the POD must consider a specific time as well as a specific
26 risk level. Since this study was not a standard study with a fixed study length, several
27 assumptions can be made with regard to the time to be used. For this modeling exercise, Two
28 different possible approaches were considered.
29 For the first approach, the model was fit to animal data using all the times reported up to
30 the last death time or 153 weeks for the female rats and 148 weeks for the male rats. Every
31 tumor observed was assumed to be a fatal tumor or the cause of death in the animal. While the
32 animals lived longer than 104 or 105 weeks, the POD was calculated at 105 weeks, since it was
33 assumed that an animal life span of 148-153 weeks would not correspond to the average 70-year
34 human life span.
E-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 For the second approach, an attempt was made to simulate what might have occurred if
2 the study had been a standard 2-year protocol that was terminated at 105 weeks, with all
3 surviving animals sacrificed at 105 weeks. It was assumed that the tumors discovered in the
4 animals that survived longer than 105 weeks would have been present and found at necropsy.
5 Therefore, all tumors in animals that died in weeks 105 and earlier were assumed to be fatal or
6 the cause of death in the animals, and all tumors that would have been discovered at the necropsy
7 of animals were assumed to be incidental or not the cause of death. The life span assumed for
8 this analysis was 105 weeks in the rat and the POD was calculated for a 10% risk to a human at
9 105 weeks. This approach was conducted mainly for comparative purposes to evaluate the
10 potential impact on the POD if the study duration was shortened and for a more direct
11 comparison to the quantal PODs and serves only as bounding exercise for the risk.
12 Option 3 - Dose-Response Modeling using PK Dose Metrics. For the third approach,
13 PK dose metrics obtained from the PBPK model (Section 3.4) were used as the doses. Both
14 time-to-tumor modeling and quantal modeling was done. The Statox program was used to
15 estimate MLEs and lower bounds on dose associated with a 10% extra risk (LEDlOs), and
16 BMDS multistage model was used to estimate BMDs and 95% lower bounds on the BMDs or
17 BMDLs associated with a 10% extra risk (BMDL10). Each of the dose metrics, provided in
18 Table E-5, were used in this option of the dose-response modeling.
E. 1.2.1. Results for Oral Slope Factor POD
19 Quantal Dose-Response Modeling (Option 1). Quantal dose-response modeling was
20 conducted using estimated mg/kg-day doses and the incidence of lympho-immunoblastic
21 neoplasms and the combined lymphomas for the female rat and the hepatocellular carcinomas,
22 the lympho-immunoblastic neoplasms, and the combined lymphomas for the male rat. The
23 results of the quantal modeling for this option are given in Table E-3. The multistage model gave
24 an adequate fit (p value > 0.05) and was able to derive BMDLio values for all of the lymphoma
25 data. However, for the male rat hepatocellular carcinomas, the multistage model failed to
26 estimate a BMDLio. Human equivalent BMDLio values94 and CSF are also provided in
27 Table E-3.
28 The POD values calculated from the female rat data ranged from 259-405 mg/kg-day for
29 the lympho-immunoblastic neoplasms and 251-400 mg/kg-day for the combined lymphomas.
30 For the HED POD values calculated from the male rat data, the range of values for the liver
31 hepatocellular carcinomas was 537-567 mg/kg-day; however, for 3 of the models (Multistage,
32 Quantal Linear and Log-Probit), the BMDS program failed to estimate a BMDLio. The lympho-
33 immunoblastic endpoint in the male rats produced POD values that ranged from 110-223 mg/kg-
34 day, and the all lymphomas incidence data gave POD values of 113-227 mg/kg-day. Figures E-3
94 Computed from the animal values by multiplying by the body weight'7' animal-to-human extrapolation value (0.26
for females and 0.29 for males)
E-6 DRAFT - DO NOT CITE OR QUOTE
-------
1 through E-7 show graphs of the quantal models that give the lowest POD for each of the
2 endpoints modeled.
3 Time-to-Tumor Dose-Response Modeling (Option 2). Results of the time-to-tumor
4 modeling using estimated mg/kg-day doses with POD and CSF values for each approach
5 described above are given in Table E-4. Approach 2 gives smaller POD estimates than
6 Approach 1. With Approach 2, an artificial end of the study is assumed of 105 weeks and the
7 designation of approximately half of the total tumors changed from fatal to incidental. This
8 approach evaluates the potential impact on the POD of terminating the study at 105 weeks, rather
9 than allowing the animals to live until their natural death, assuming that the same animals
10 bearing tumors would be "observed" at 105 weeks, rather than at later time points. For
11 Approach 1, the model was fit to the actual observed weeks-on-study, and a time of 105 weeks
12 was used in calculating the POD. For the female rat, the PODs based on the lympho-
13 immunoblastic neoplasms were 349 mg/kg-day for approach 1 and 179 mg/kg-day
14 forApproach 2. The PODs for the combined lymphomas ranged were 321 and 198 mg/kg-day
15 for the 2 appraoches. For the PODs calculated from the male rat data, the values were 783 and
16 612 mg/kg-day for the hepatocellular carcinomas, 174 and 91 mg/kg-day for the lympho-
17 immunoblastic neoplasms, and 192 and 92 mg/kg-day for the combined lymphomas. Figures E-
18 3 through E-7 show the modeling results for the time-to-tumor modeling using Approache 1
19 where the time is fixed at 105 weeks and the doses are allowed to vary. Figures E-8 and E-9
20 show Kaplan-Meier curves versus the model fit to the combined lymphoma data for the females
21 and males, respectively. In these graphs each line corresponds to a specific dose, and time is
22 allowed to vary up to the study end of 153 or 148 weeks. For the male rat combined lymphomas
23 (Figure E-9), the multistage Weibull predicted values more closely match the Kaplan-Meier at
24 105 weeks than at the average life span (94 weeks) or the end of study (148 weeks). For the
25 female rat combined lymphomas, the closest match of the Kaplan-Meier curves to the model
26 predicted values appears to be around the average life span of 96 weeks.
27 The AIC values for the time-to-tumor modeling are all higher than those for the quantal
28 modeling. However, since the LLFs on which the AIC are based are different for time-to-tumor
29 versus quantal models, the AIC values from time-to-tumor models cannot be compared to those
30 from quantal models.
31 Dose-Response Modeling using PBPK Dose Metrics (Option 3). Both time-to-tumor
32 and quantal dose-response modeling were conducted using the incidence of lympho-
33 immunoblastic neoplasms and the combined lymphomas for the female rat and the hepatocellular
34 carcinomas, the lympho-immunoblastic neoplasms, and the combined lymphomas for the male
35 rat and the three PBPK dose metrics (blood methanol AUC, peak blood concentration, and the
36 total metabolized per day) obtained from the EPA's PBPK model (Table E-5).
37 Only Approach 1 was used for the time-to-tumor dose-response modeling using PBPK
38 dose metrics, and results are given in Table E-6a; HEDs are provided in Table E-6b.
E-7 DRAFT - DO NOT CITE OR QUOTE
-------
1 The results of the multistage quantal modeling are given in Table E-7. Most model runs
2 gave an adequate fit to the data by the ^ Goodness of Fit p-value (e.g., p-values > 0.05),
3 although for male rat hepatocellular carcinomas, the calculations using the AUC or amount
4 metabolized dose metric were unable to converge for the model fit or the derivation of a BMDL.
5 A plot of the model fit for male rat combined lymphomas using total methanol metabolized per
6 day as the dose metric (the endpoint and dose metric used in the derivation of the oral CSF) is
7 shown in Figure E-10; HEDs are provided in Table E-8 where the BMDL was calculated.
E.2. INHALATION ORAL CANCER UNIT RISK POD
8 The NEDO (1987, 1987/2008b) study was conducted using a standard protocol with
9 exposure for 104 weeks, followed by sacrifice of all animals surviving to 104 weeks. As with
10 the Soffretti (2002a) study, pharmacokinetic dose metrics for use in the dose-response
11 assessment were determined for the inhalation exposures using the EPAs PBPK model.
E.2.1. Selection of the Data Modeled and HED
12 The individual animal data from the NEDO (1987/2008b) study were provided in a 2008
13 translation of the study from Japanese to English. Although the translation provided the number
14 of days on study and the neoplastic responses seen in each animal, the translation did not provide
15 results if a tissue was examined histopathologically with no neoplastic responses. This makes it
16 difficult to determine which of the individual animals were not examined, although the tables did
17 indicate that, for some of the animals, selected organs (specifically a few lungs in males and a
18 few adrenal glands in females) were not examined. Therefore, time-to-tumor analysis could not
19 be conducted with results from the inhalation data as was done with the oral data. However,
20 survival analysis of all the data from the NEDO (1987/2008b) study did not indicate that there
21 were any survival problems (Figures E-ll and E-12). This suggests that a time-to-tumor analysis
22 is not necessary.
23 The tumors with significantly increased incidence that were considered for dose-response
24 modeling were the female rat adrenal gland pheochromocytomas and the male rat lung tumors
25 (papillary adenomas and adenocarcinomas combined or papillary adenomas, adenocarcinomas
26 and adenomatosis combined); Table E-9 gives the incidence of these tumors.
E.2.1.1. Dose-Response Modeling
27 Quantal Dose-Response Modeling using Pharmacokinetic Dose Metrics. For the
28 selected endpoints from the NEDO (1987) study, only quantal dose-response modeling using
29 pharmacokinetic internal dose metrics estimated by the PBPK model (described in Section 3.4
30 and Appendix B) was conducted. Each of the dose metrics provided in Table E-10 was used in
31 the BMDS software with the incidence data in Table E-9 to estimate the BMDs and 95% lower
32 confidence limits (BMDLs) associated with a 10% extra risk (BMDLio).
E-8 DRAFT - DO NOT CITE OR QUOTE
-------
E.2.2. Results for the IUR POD
1 Quantal dose-response modeling was conducted using the incidence of adrenal gland
2 phoechromocytomas in female rats and the combined incidence of lung adenomas and
3 adenocarcinomas or lung adenomas, adenocarcinomas and adenomatosis in male rats using the
4 pharmacokinetics dose metrics derived from the EPA's PBPK model. The results of this
5 modeling are given in Table E-l 1. The multistage model gave an adequate fit to the data in all
6 instances as determined by the £ goodness-of-fit/?-value (e.g., ^-values > 0.05). A plot of the
7 model fit for female rat pheochromocytomas using total methanol metabolized per day as the
8 dose metric (the endpoint and dose metric used in the derivation of the IUR) is shown in
9 Figure E-13. HECs are provided in Table E-12.
E.3. ANALYSIS OF APAJA (1980) DRINKING WATER STUDY
10 The Apaja (1980) study was similar to the Soffritti et al. (2002a) study in that it was a life
11 span drinking water study. The primary differences are that Apaja (1980) used Eppley Swiss
12 Webster mice, did not stop exposure at 104 weeks and did not employ an untreated concurrent
13 control group. Methanol exposure groups of this study served as controls for malonaldehyde
14 exposed mice. As with the Soffretti (2002a) study, pharmacokinetic dose metrics for use in the
15 dose-response assessment were determined for the oral exposures using the EPA's PBPK mouse
16 model.
E.3.1. Selection of the Data Modeled and HED
17 Individual animal data from the Apaja (1980) study were not available. Therefore, time-
18 to-tumor analysis could not be conducted. The tumors with significantly increased incidence that
19 were considered for dose-response modeling were malignant lymphomas in male and female
20 mice; Table E-13 gives the incidence of these tumors.
E. 3.1.1. Dose-Response Modeling
21 Quantal Dose-Response Modeling using Pharmacokinetic Dose Metrics. A 1st degree
22 multisage model was used to evaluate the malignant lymphoma response from the Apaja (1980)
23 study versus pharmacokinetic dose metrics without background95 was conducted. Each of the
24 dose metrics provided in Table E-14 was used in the BMDS software with the incidence data in
25 Table E-13 to estimate the BMDs and 95% lower confidence limits (BMDLs) associated with a
26 10% extra risk (BMDLio).
95 The assumption of zero background dose is consistent with what was done for the derivation of the oral CSF and
EPA practice.
E-9 DRAFT - DO NOT CITE OR QUOTE
-------
E.3.2. Results for the IUR POD
1 Quantal dose-response modeling was conducted using the incidence of malignant
2 lymphoma in male and female Eppley Swiss Webster mice using the pharmacokinetics dose
3 metrics derived from the EPA's PBPK model. The results of this modeling are given in
4 Table E-15. The 1st degree multistage model gave an adequate fit to the data in all instances as
5 determined by the $ goodness-of-fit/>-value (e.g., ^-values > 0.05). A plot of the model fits for
6 male and female mice using total methanol metabolized per day as the dose metric is shown in
7 Figure E-14. BMDLio HECs associated with each dose metric are provided in Table E-16.
E.4. BACKGROUND DOSE ANALYSES
8 The primary purpose of this cancer analysis is for the determination of cancer risk
9 associated with increases in the levels of methanol or its metabolites (e.g., formate,
10 formaldehyde) over background. Thus, the PBPK model estimates of internal dose used in the
11 dose-response analyses described above do not describe or account for background levels of
12 methanol or its metabolites. However, background levels of methanol may have contributed to
13 the response levels reported for some of the tumors associated with methanol. If this
14 contribution is large, it could significantly impact cancer risk estimates.
15 Of the Soffritti et al. (2002a), NEDO (1985/2008b) and Apaja (1989) studies, only the
16 NEDO study reported methanol blood levels in test animals. Based on background levels of
17 methanol in the blood of F344 rats reported by NEDO (1987; 1985/2008b) PBPK model
18 estimates of AUC methanol (mg-h/L), Cmax methanol (mg/L) and total metabolites (mg/day) in
19 control animals were obtained. AUC methanol, Cmax methanol and methanol metabolite
20 background levels were estimated to be 97.16 mg-h/L, 4.05 mg/L and 7.79 mg/day for female
21 F344 rats and 96.02, 4.00 and 10.81 mg/day for male F344 rats, respectively.
22 Available bioassays do not allow for the quantification of the relationship between
23 background levels and background responses, but the "background dose" Multistage-cancer
24 model contained in version 2.1 of EPA's BMDS software can be used to estimate what
25 background dose would be necessary to explain the dose-response data evaluated in this
26 assessment if all of the background response were due to background levels of methanol or its
27 metabolites. The results of dose-response modeling of cancer endpoints evaluated in this
28 assessment using the Multistage-cancer "background dose" model are shown in Table E-17.
29 As can be seen from Table E-17, the Multistage-cancer-bgdose model predicts that AUC
30 methanol, Cmax methanol and methanol metabolite background doses necessary to explain the
31 dose-response data if all of the background response were due to background levels of methanol
32 or its metabolites are well above the actual background levels for these metrics predicted by the
33 PBPK model. In the case of lymphoma responses in Sprague-Dawley rats, the background doses
34 for all metrics predicted by the Multistage-cancer-bgdose model are 20-2,000 fold higher than
E-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 the actual background doses predicted in F344 rats by the PBPK model. In the case of female
2 pheochromocytoma and male lung responses in F344 rats, the background doses predicted by the
3 Multistage-cancer-bgdose model are 3-5 fold higher than the actual background doses predicted
4 by the PBPK model. The fact that background dose estimates from the Multistage-cancer-bgdose
5 model are several-fold and sometimes more than 1,000-fold higher than the PBPK model
6 predictions suggests that the contribution of background levels of methanol or its metabolites to
7 the background responses of these tumors is relatively small.
E-11 DRAFT - DO NOT CITE OR QUOTE
-------
Table E-l. Calculation of mg/kg-day doses
Dose (ppm)
0
500
5000
20000
Female Sprague-Dawley Rats
Body
weight
(kg)
0.33
0.33
0.33
0.34
Water
consump.
(g/day or
mL/day)
42.55
43.05
41.11
37.26
Dose
(mg/kg
-day)
0
66.0
624.1
2177
Average
time of
death
(wk)
98
96
94
98
Median
time of
death
(wk)
102
99
97
101
Male Sprague-Dawley Rats
Body
weight
(kg)
0.50
0.49
0.50
0.54
Water
consump.
(g/day or
mL/day)
52.57
52.06
52.58
48.32
Dose
(mg/kg -
day)
0
53.2
524
1780
Average
time of
death
(wk)
91
97
93
93
Median
time of
death
(wk)
91
98
93
100
Averaged
over all
dosed groups
(excluding
control)
Averaged
over all
groups
(including
control)
0.33
0.33
96
97
99
100
0.51
0.51
94
93
97
96
Source: Soffretti et al. (2002a).
Table E-2. Incidence for neoplasms considered for dose-response modeling
Dose
(ppm)
Dose
(mg/kg-
day)
Number of
animals
examined
Heptocellular
carcinomas
Histiocytic
sarcoma
Leukemia
monocytic
Leukemia
myloid
Lymphoma
lymphoblastic
Lymphoma
lympocytic
Lymphoma
lympho-
immunoblastic
All
lymphomas
combined
Female Sprague-Dawley rats
0
500
5000
20000
0
66.0
624.1
2177
100
100
100
100
Cochran Armitage Trend Test
1
2
2
3
0.19
3
3
3
3
0.5
0
1
1
1
0.3
0
1
0
0
0.8
9
17
19a
21a
0.04
9
19a
20a
22b
0.04
Male Sprague-Dawley rats
0
500
5000
20000
0
53.2
524
1780
100
100
100
99
Cochran Armitage Trend Test
0
2
2
3
0.10
2
4
1
1
0.9
1
0
0
0
0.8
8
4
6
1
0.98
1
3
1
1
0.7
16
24
28a
37b
0.0007
17
27
29a
38b
0.001
Tisher's Exact/>-value < 0.05; bFisher's Exact p-value < 0.01
Source: Soffretti et al. (2002a).
E-12
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-3. Results from multistage (1°) quantal modeling rat data using mg/kg-day
exposures and default HED derivation method
Female
Sprague-
Dawley rat
Male
Sprague-
Dawley rat
All organs
lympho-
immunoblastic
All organs - all
lymphomas
Hepatocellular
carcinoma
All organs
lympho-
immunoblastic
All organs - all
lymphomas
AIC
359.16
371.95
72.84
455.67
468.79
p-value
0.19
0.11
0.38
0.34
0.24
Scaled
residual at
observed
dose closest
toBMD
-0.07
-0.07
Animal values
BMD10
2179.51
2141.88
BMDL10
1058.99
1033.69
Human values
Human
equivalent
BMDIV
277.5
270.9
CSF"
3.6E-4
3.7E-4
Failed8
0.14
0.11
714.26
744.35
448.43
456.11
131.0
133.3
7.6E-4
7.5E-4
aModel failed to optimize.
Calculated as CSF = O.I/Human BMDL10.
Source: Soffretti et al. (2002a).
Table E-4. Results from time-to-tumor modeling data using mg/kg-day exposures and
default HED derivation method
AIC
Prediction
time (weeks)
Human values'"
MLE
(mg/kg-day)
LED10
(mg/kg-day)
CSFC
Approach 1 - Model fit to actual death times, dose estimates computed at 105 weeeks"
Female
Sprague-
Dawley rat
Male
Sprague-
Dawley rat
All organs lympho-immunoblastic
All organs all lymphomas
Hepatocellular carcinoma
All organs lympho-immunoblastic
All organs all lymphomas
831.42
900.46
105.45
1254.58
1309.86
105
105
105
105
105
729.0
679.1
4250.7
302.0
356.7
348.6
320.9
783.3
173.8
191.8
2.9E-4
3.1E-4
1.3E-04
5.8E-4
5.2E-4
Approach 2 - Truncating study at 105 Weeks3
Female
Sprague-
Dawley rat
Male
Sprague-
Dawley rat
All organs lympho-immunoblastic
All organs all lymphomas
Hepatocellular carcinoma
All organs lympho-immunoblastic
All organs - all lymphomas
631.81
664.43
1.16E+05d
914.42
935.29
105
105
105
105
105
370.7
431.9
1013.7
150.8
157.1
178.5
198.0
612.2
91.2
92.2
5.6E-4
5.1E-4
1.6E-4
1.1E-3
1.1E-3
Individual animal pathology data needed for the modeling reported in this table can be obtained from the Ramazzini Foundation web site
(http: //www. ramazzini. it/fondazione/study .asp).
bHuman values are computed by converting the animal doses to HED before modeling by multiplying by the body weight37' animal-to-
human extrapolation value (0.26 for females and 0.29 for males).
"Calculated as CSF = O.I/Human LED10.
dModel failed to optimize.
Source: Soffretti et al. (2002a).
E-13
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-5. PBPK model estimated dose-metrics for doses
Sex
Female
Sprague-
Dawley
Male
Sprague-
Dawley
Dose (mg/kg-day)
66
624.1
2177
53.2
524
1780
Body weight
(kg)
0.33
0.33
0.34
0.49
0.50
0.54
AUC
(mg-h/L)
66.84
9543.63
91262.27
55.76
7500.14
80420.25
Peak (mg/L)
5.96
500.59
4157.98
4.81
395.56
3629.08
Amount metabolized
(mg/day)
18.39
126.68
141.60
21.82
168.85
200.03
Source: Soffretti et al. (2002a).
Table E-6a. Results from time-to-tumor modeling of data using PBPK dose metrics
Approach 1 - Model fit
to actual death times,
dose estimates
computed at 105
weeeks
Female
Sprague-
Dawley
rat
Male
Sprague-
Dawley
rat
All organs
lympho-
immunoblastic
All organs all
lymphomas
Hepatocellular
carcinoma
All organs
lympho-
immunoblastic
All organs - all
lymphomas
Prediction
Time (wk)
105
105
105
105
105
AUC (mg-h/L)
AIC
832.43
901.43
110.47
1242.72
1297.74
MLE
152281
143138
undefined
55409
66211
LED10
64486
59574
119186
30059
33315
Peak (mg/L)
AIC
832.35
901.36
110.45
1244.61
1297.66
MLE
6763.4
6356.1
undefined
2460.2
2939.5
LED10
2895.3
2675.3
5385.9
1341.3
1487.2
Amount Metabolized
(mg/d)
AIC
829.24
898.24
109.60
1242.66
1295.95
MLE
160.45
145.53
undefined
141.63
143.14
LED10
90.7
82.2
300.4
77.3
82.7
Source: Soffretti et al. (2002a)
Table E-6b. HEDs from time-to-tumor modeling of data using PBPK dose metrics
Approach 1 - Model fit to actual death times,
dose estimates computed at 105 weeeks
Female
Sprague-
Dawley rat
Male
Sprague-
Dawley rat
All organs lympho-immunoblastic
All organs all lymphomas
Hepatocellular carcinoma
All organs lympho-immunoblastic
All organs - all lymphomas
AUC (mg-h/L)
LED10 (mg/kg-day)
647.0
618.3
965.4
594.0
657.1
Peak (mg/L)
LED10 (mg/kg-day)
676.2
645.4
1023.9
457.3
478.1
Amount
Metabolized (mg)
LED10 (mg/kg-day)
87.7
79.4
260.0
62.4
66.7
Source: Soffretti et al. (2002a)
E-14
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-7. Results of Multistage (1°) quantal modeling of data using PBPK dose
metrics
AUC (mg-h/L)
AIC
P-
value
Scaled
residual
at dose
nearest
toBMD
BMD10
BMDL10
Peak (mg/L)
AIC
P-
value
Scaled
residual
at dose
nearest to
BMD
BMD10
BMDL10
Amount Metabolized (mg/d)
AIC
P-
value
Scaled
residual
at dose
nearest to
BMD
BMD10
BMDL10
Female Sprague-Dawley rats
All organs lympho-immunoblastic
357.9
0.13
-0.132
112790
49771
359.9
0.13
-1.56
5034.3
2240.3
357.9
0.34
-0.913
139.8
77.2
All organs - all lymphomas
372.78
0.075
-0.128
111621
48789.6
372.17
0.077
-0.145
4982.2
2196.3
370.57
0.18
-0. 002
134.4
74.5
Male Sprague-Dawley rats
Hepatocellular carcinoma
73.21
0.37
Failed8
73.19
0.37
Failed8
72.55
0.36
Failed8
All organs lympho-immunoblastic
457.45
0.14
1.214
36865
21989
457.3
0.15
1.173
1639.0
981.7
455.46
0.36
-0.750
94.4
61.6
All organs - all lymphomas
468.09
0.15
0.878
37822
22195
467.98
0.15
0.838
1686.5
993.1
467.35
0.20
-0.901
101.7
63.9
aBMD computation failed. BMD is larger than three times maximum input doses.
Source: Soffretti et al. (2002a).
Table E-8. Application of human PBPK model to derive HEDs from results of
multistage (1°) quantal modeling of data using PBPK dose metrics
Female
Sprague-
Dawley rat
Male
Sprague-
Dawley rat
All organs lympho-
immunoblastic
All organs - all
lymphomas
Hepatocellular
carcinoma
All organs lympho-
immunoblastic
All organs - all
lymphomas
AUC (mg-h/L)
HEDBMDL10(mg/kg-
day)
560.96
555.20
N/A
395.99
397.25
Peak (mg/L)
RED BMDL10
(mg/kg-day)
584.37
578.18
N/A
405.56
407.22
Amount metabolized
(mg/d)a
RED BMDL10
(mg/kg-day)
74.54
71.92
N/A
49.65
51.50b
"Total metabolized methanol was selected as the preferred dose metric (see discussion in section 5.4.1.3). Before
applying human PBPK model to obtain these HED BMDL10 estimates, Table E-7 mg/d values were converted to
human mg/d by multiplying by either: (BWhuman)/4/ (BWrat)y4=(70 kg)y7 (0.33 kg )y<=55.6 for male rats or,
(BWhuman)yV(BWrat)y4= (70 kg)yV(0.26 kg)y'= 66.5 for female rats.
bThis value was used in the derivation of the methanol oral cancer slope factor.
Source: Soffretti et al. (2002a).
E-15
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-9. Incidence for neoplasms considered for dose-response modeling
Dose (ppm)
Examined
Adrenal gland
phoechromocytoma
Lung adenoma and
adenocarcinoma
Lung adenoma,
adenocarcinoma,
adenomatosis
Female F344 rats
0
10
100
1,000
50
51
49
51
2
3
2
7
Male F344 rats
0
10
100
1,000
52
50
52
52
Cochran Armitage Trend Test p_values
0.015
1
5
2
T
0.0259
5
6
7
11
0.0415
"Fisher's Exact ^-values < 0.05
Source: NEDO (1987, 1987/2008b).
Table E-10. PBPK dose metrics for doses
Dose (ppm)
AUC
(mg-h/L)
Peak (mg/L)
Amount
metabolized (mg)
Female F344 rats
0
10
100
1,000
0.00
3.70
37.53
434.29
0.00
0.19
1.93
22.69
0.00
0.30
2.94
28.96
Male F344 rats
0
10
100
1,000
0
3.70
37.53
433.61
0
0.19
1.93
22.68
0
0.41
4.11
40.19
Source: NEDO (1987, 1987/2008b).
E-16
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-ll. Benchmark results from multistage quantal dose-response modeling data
using PBPK dose-metrics
Model
AUC (mg-h/L)
AIC
P-
value
Scaled
residual
at dose
nearest
toBMD
BMD10
BMDL10
Peak (mg/L)
AIC
/)-value
Scaled
residual at
dose
nearest to
BMD
BMD10
BMDL10
Amount metabolized (mg/d)
AIC
/)-value
Scaled
residual
at dose
nearest
to BMD
BMD10
BMDL10
Female F344 rats
Adrenal glands - pheochromocytoma
Multistage (3°)
101.37
0.88
0.000
442.1
217.0
101.37
0.88
0.000
23.1
11.3
101.37
0.88
0.000
29.5
14.6
Male F344 rats
Lung - adenomas and adenocarcinomas
Multistage (3°)
107.99
0.16
0.000
454.9
230.7
107.99
0.16
0.000
23.8
12.1
107.99
0.16
0.001
42.2
21.6
Lung - adenomas, adenocarcinomas and adenomatosis
Multistage (1°)
168.61
0.89
-0.035
378.9
168.8
168.61
0.89
-0.035
19.8
8.83
168.59
0.90
-0.038
34.8
15.6
Source: NEDO (1987, 1987/2008b).
Table E-12. Application of human PBPK model to derive HECs from BMDLio
estimates in Table E-ll using multistage quantal modeling
Female F344
rat
Male F344 rat
Adrenal glands -
phoechromocytoma
Lung - adenomas and
adenocarcinomas
Lung - adenomas,
adenocarcinomas and
adenomatosis
AUC (mg-h/L)
HEC BMCL10
(mg/m3)
384.2262
401.96305
316.20336
Peak (mg/L)
HEC BMCL10
(mg/m3)
464.79844
485.15995
386.70859
Amount metabolized (mg/d)a
HEC BMCL10 (mg/m3)
81.85b
101.27
73.12
a Total metabolized methanol was selected as the preferred dose metric (see discussion in section 5.4.2.3). Before
applying human PBPK model to obtain these HEC BMCL10 estimates, Table E-7 mg/d values were converted to
human mg/d by multiplying by either: (BWhuman)/4/ (BWrat)y4=(70 kg)y
-------
Dose (ppm)
Examined
Malignant Lymphoma
Male Swiss Webster mice
Historical untreated controls 3
10
100
1,000
100
25
25
25
8
1
6d
4
aToth et al. (1977); bHinderer et al. (1979); >value = 0.06;
Source: Apaja(1980)
< o.05
Table E-14. PBPK dose metrics for doses in Apaja (1980)
Daily
Dose
(mg/kg-d)
Weekly
Avg. Dose
(mg/kg-d)
Body
Weight
(kg)
AUC
(mg-h/L)
Peak (mg/L)
Amount
metabolized (mg)
Female Swiss Webster mice
0
560.01
1000.00
2099.98
0
479.99
857.13
1800.00
0.040
0.040
0.040
0.040
0
484.86
3466.74
19503.87
0
88.40
383.10
1462.07
0
18.38
28.45
39.18
Male Swiss Webster mice
0
550.00
970.00
1800.00
0
471.42
831.39
1542.86
0.045
0.045
0.045
0.045
0
500.63
3405.26
14998.60
0
89.77
373.72
1162.69
0
20.29
31.15
41.86
Source: Apaja (1980).
Table E-15. Benchmark results from Multistage-cancer dose-response modeling data
for malignant lymphoma in Swiss Webster mice (Apaja, 1980) using PBPK dose-
metrics
Gender
Female mice3
Male mice3
AUC (mg-h/L)
AIC
288.91
122.09
P-
value
0.32
0.083
Scaled
residual
at dose
nearest
toBMD
1.354
-0.883
BMD10
5808.3
11164.2
BMDL10
2957.6
4343.04
Peak (mg/L)
AIC
288.42
121.58
/)-value
0.43
0.12
Scaled
residual at
dose
nearest to
BMD
1.090
-0.641
BMD10
428.5
798.3
BMDL10
225.1
338.9
Amount metabolized (mg/d)
AIC
288.07
120.97
7>-value
0.55
0.21
Scaled
residual
at dose
nearest
to BMD
-0.937
1.309
BMD10
22.7
35.8
BMDL10
10.1
18.3
3Multistage-cancer (1°) used for AUC and Peak metrics; Multistage-cancer (2°) used for Amount metabolized metric
Source: Apaja (1980).
1
E-18
DRAFT - DO NOT CITE OR QUOTE
-------
Table E-16. Application of human PBPK model to derive HEDs from BMDLio
estimates of Table E-15, Multistage (1°) modeling of malignant lymphoma in Swiss
mice (Apaja, 1980) using PBPK dose metrics
Female mice
Male mice
AUC (mg-h/L)
RED BMDL10 (mg/kg-day)
248
269
Peak (mg/L)
RED BMDL10 (mg/kg-
day)
281
305
Amount metabolized (mg/d)a
RED BMDL10
(mg/kg-day)
9.7
14.7
"Before applying human PBPK model Table 11 mg/day values were converted to human mg/day by multiplying by
either: (BWhuman)y7 (BWratf=(70 kgf/ (0.33 kg f=55.6 for male rats or, (BWhuman)yV(BWratf = (70 kgf/(0.26 kg f
= 66.5 for female rats.
Table E-17. Benchmark results for all tumor types using BMDS 2.1 multistage
"background dose" models and PBPK dose-metrics
Model
AUC (mg-h/L)
AIC
P-
value
Background
Dose
BMD10
BMDL10
Peak (mg/L)
AIC
p-
value
Background
Dose
BMD10
BMDL10
Amount metabolized (mg/d)
AIC
p-
value
Background
Dose
BMD10 BMDL10
Female Sprague-Dawley rats (Soffritti et al., 2002a)
All organs lympho-immunoblastic
Multistage (1°) 361.99 0.045 168476 112790 49771 359.93 0.14 7480.0 5034.3 2240.3 357.87 0.34
169.0
139.8 77.2
All organs - all lymphomas
Multistage (1°) 372.78 0.075 179247 111621 48790 372.72 0.077 7961.1 4982.2 2196.3 370.57 0.18
175.1
134.4 74.1
Male Sprague-Dawley rats (Soffritti et al., 2002a)
Hepatocellular carcinoma
Multistage (1°) 97.52 o.oo
Failed8
73.19 0.37
Failed"
72.55 0.36
Failed8
All organs lympho-immunoblastic
Multistage (1°) 455.i1 0.18 85497.6 36501 21901 454.97 0.20 3781.3 1625.7 979.1 453.98 0.30
190.2
97.8 63.2
All organs - all lymphomas
Multistage (1°) 468.09 0.15 96684.3 37822 22195 467.98 0.15 4285.6 1686.5 993.1 467.35 0.20
221.2
101.7 63.9
Female F344 rats (NEDO, 1985/2008b)
Adrenal glands - pheochromocytoma
Multistage (2°) ioi.46 0.84 553.1 450.9 214.8 103.45 0.56
28.9
23.6
11.2 101.47 0.83
36.7
30.1
14.4
Male F344 rats (NEDO, 1985/2008b)
Lung - adenomas and adenocarcinomas
Multistage (1°) 108.15 | 0.14 | 248.16 507.9 I 226.0 | 108.15 0.14 13.0 | 26.6 | 11.8 | 108.19 0.14
23.0
47.3 21.0
Lung - adenomas, adenocarcinomas amd adenomatosis
Multistage (1°) 168.61 0.89 430.38 378.9 168.8 168.61 0.89 22.5
19.8
8.8 168.59 0.90
39.3
34.8
15.6
Female Swiss Webster mice (Apaja, 1989)
Malignant lymphomas
Multistage (1°) 288.91 | 0.32 | 11933.3 5808 I 2958 | 288.42 0.43 863.1 | 428.5 | 225.1 | 289.00 0.37
16.8
9.1
Male Swiss Webster mice (Apaja, 1989)
Multistage (1°) 122.09 0.083 9573.9 11164 4343.0 121.58 0.12 646.9 798.3 338.9 121.30 0.20
28.6
38.1 17.6
aBMD computation failed. BMD is larger than three times maximum input doses.
Source: NEDO (1987, 1987/2008b).
E-19
DRAFT - DO NOT CITE OR QUOTE
-------
Female Rats Survival
60 80 100
Time on Study (weeks)
120
140
160
-Control •
-500 ppm •
-5000 ppm
20000 ppm
Figure E-l. Female rat survival.
Source: Soffretti et al. (2002a).
Male Rats Survival
30 45 60 75 90 105 120 135 150
Time on Study (weeks)
-Control 500 ppm 5000 ppm 20000 ppm
Figure E-2. Male rats survival.
Source: Soffretti et al. (2002a).
E-20
DRAFT - DO NOT CITE OR QUOTE
-------
14:2701/11/2007
Female Rats - Multistage Weibull Model (t= 105 weeks)
Time-to-Tumor Weibull
Observed (90%% Binomial limits)
1000 1500
Animal Dose (mg/kg-day)
Figure E-3. Female - Lymphomas lympho-immunoblastic - Multistage Weibull Model
-Approach 1.
Source: Soffretti et al. (2002a).
14:0201/11/2007
Female Rats - Multistage Weibull Model (1= 105 weeks)
Time-to-Tumor Weibull
Observed (90%% Binomial limits)
1000 1500
Animal Dose (mg/kg-day)
Figure E-4. Female - All lymphomas - Multistage Weibull Model - Approach 1.
Source: Soffretti et al. (2002a).
E-21
DRAFT - DO NOT CITE OR QUOTE
-------
15:1905/21/2008
Risk Assessment: 2 ; Chemical: Methanol; Sex: Male
Time-to-TumorWeibull
Observed (90%% Binomial limits)
0 200 400 600 800 1000 1200 1400 1600 1800
Dose (mg/kg/day)
Figure E-5. Male - Hepatocellular carcinoma - Multistage Weibull Model - Approach
1.
Source: Soffretti et al. (2002a).
16:0301/11/2007
Male Rats- Multistage Weibull Model (t= 105 weeks)
Time-to-Tumor Weibull
Observed (90%% Binomial limits)
200 400 600 800 1000 1200 1400 1600 1800
Animal Dose (mg/kg-day)
Figure E-6. Male - Lymphomas lympho-immunoblastic - Multistage Weibull Model
Approach 1.
Source: Soffretti et al. (2002a).
E-22
DRAFT - DO NOT CITE OR QUOTE
-------
16:0401/11/2007
Male Rats- Multistage Weibull Model (t= 105 weeks)
Time-to-Tumor Weibull
Observed (90%% Binomial limits)
200 400 600 800 1000 1200 1400 1600 1800
Animal Dose (mg/kg-day)
Figure E-7. Male - All lymphomas - Multistage Weibull Model - Approach 1.
Source: Soffretti et al. (2002a).
Female Rats All Lymphomas Combined
Kaplan Meier Plots and Multistage Weibull Model
Figure E-8. Female rats -All lymphomas time-to-tumor model fit and Kaplan Meier
curves.
Source: Soffretti et al. (2002a).
E-23
DRAFT - DO NOT CITE OR QUOTE
-------
Male Rats All Lymphomas Combined
Kaplan Meier Plots and Multistage Weibull Model
Figure E-9. Male rats -All lymphomas time-to-tumor model fit and Kaplan Meier
curves.
Source: Soffretti et al. (2002a).
Multistage Cancer Model with 0.95 Confidence Level
.3
V
£
0.1
08:55 09/03 2009
Multistage Cancer
Linear extrapolation
200
Figure E-10. Male rats-All lymphomas; dose = amount metabolized (mg/day); 1°
multistage model.
Source: Soffretti et al. (2002a).
E-24
DRAFT - DO NOT CITE OR QUOTE
-------
•S 0.7
0.5
I.
0.2
50
Female Rats Survival
60
70
Time on Study (weeks)
-Control •
- lOppm -
-lOOppm
1000 ppm
Figure E-ll. Female rat survival.
Source: NEDO (1987, 1987/2008b).
i
0.9
0.8 -
0.7 -
0.6 -
0.5 -
0.4 -
0.3 -
0.2 -
0.1 -
50
Male Rats Survival
65
95
Time on Study (weeks)
-Control •
- lOppm -
-100 ppm 1000 ppm
Figure E-12. Male rat survival.
Source: NEDO (1987, 1987/2008b).
E-25
DRAFT - DO NOT CITE OR QUOTE
-------
08:58 09/03 2009
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
15
dose
Figure E-13. Female rats- pheochromocytomas; dose = amount metabolized (mg/d); 3°
multistage model.
Source: NEDO (1987, 1987/2008b)
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
09:58 09/03 2009
15 20 25 30 35 40
dose
Multistage Cancer Model with 0.95 Confidence Level
Multistage Cancer
Linear extrapolation
10 15 20 25 30 35 40
09:56 09/03 2009
Figure E-14. Plots for female (top; p= 0.55) and male (bottom; p= 0.21) mice -
malignant lymphoma; dose=amount metabolized (mg/d); 2° multistage model.
Source: Apaja(1980)
E-26
DRAFT - DO NOT CITE OR QUOTE
------- |