DRAFT - DO NOT CITE OR QUOTE
EPA/635/R-11/001
www.epa.gov/iris
f/EPA
TOXICOLOGICAL REVIEW
OF
METHANOL (NON-CANCER)
(CAS No. 67-56-1)
In Support of Summary Information on the
Integrated Risk Information System (IRIS)
March 2011
NOTICE
This document is an External Peer Review Draft. This information is distributed solely for the
purpose of pre-dissemination peer review under applicable information quality guidelines. It has
not been formally disseminated by EPA. It does not represent and should not be construed to
represent any Agency determination or policy. It is being circulated for review of its technical
accuracy and science policy implications.
U.S. Environmental Protection Agency
Washington, DC
-------
DISCLAIMER
This document is a preliminary draft for review purposes only. This information is
distributed solely for the purpose of pre-dissemination peer review under applicable information
quality guidelines. It has not been formally disseminated by EPA. It does not represent and
should not be construed to represent any Agency determination or policy. Mention of trade names
or commercial products does not constitute endorsement or recommendation for use.
March 2011 « DRAFT - DO NOT CITE OR QUOTE
-------
CONTENTS TOXICOLOGICAL REVIEW OF METHANOL
(CAS NO. 67-56-1)
LIST OF TABLES VI
LIST OF FIGURES IX
LIST OF ABBREVIATIONS AND ACRONYMS XH
FOREWORD XVin
AUTHORS, CONTRIBUTORS, AND REVIEWERS XIX
1. INTRODUCTION 1-1
2. CHEMICAL AND PHYSICAL INFORMATION 2-1
3. TOXICOKINETICS 3-1
3.1. OVERVIEW 3-1
3.2. KEY STUDIES 3-9
3.3. HUMAN VARIABILITY IN METHANOL METABOLISM 3-15
3.4. PHYSIOLOGIC ALLY BASED TOXICOKINETIC MODELS 3-16
3.4.1. Model Requirements for EPA Purposes 3-16
3.4.2. MethanolPBPK Models 3-19
3.4.3. Selected Modeling Approach 3-23
3.4.4. Mouse Model Calibration and Sensitivity Analysis 3-30
3.4.5. Rat Model Calibration 3-35
3.4.6. Human Model Calibration 3-39
3.4.7. Monkey PK Data and Analysis 3-46
3.4.8. Summary and Conclusions 3-51
4. HAZARD IDENTIFICATION 4-1
4.1. STUDIES IN HUMANS - CASE REPORTS, OCCUPATIONAL AND
CONTROLLED STUDIES 4-1
4.1.1. Case Reports 4-1
4.1.2. Occupational Studies 4-10
4.1.3. Controlled Studies 4-12
4.2. ACUTE, SUBCHRONIC AND CHRONIC STUDIES IN ANIMALS - ORAL
AND INHALATION 4-14
4.2.1. Oral Studies 4-14
4.2.2. Inhalation Studies 4-17
4.3. REPRODUCTIVE AND DEVELOPMENTAL STUDIES - ORAL AND
INHALATION 4-26
4.3.1. Oral Studies 4-26
4.3.2. Inhalation Studies 4-28
4.3.3. Other Reproductive and Developmental Toxicity Studies 4-41
March 2011 i« DRAFT - DO NOT CITE OR QUOTE
-------
4.4. NEUROTOXICITY 4-46
4.4.1. Oral Studies 4-47
4.4.2. Inhalation Studies 4-49
4.4.3. Studies Employing In Vitro, S.C. and IP. Exposures 4-56
4.5. IMMUNOTOXICITY 4-61
4.6. MECHANISTIC DATA AND OTHER STUDIES IN SUPPORT OF THE MOA 4-66
4.6.1. Role of Methanol and Metabolites in the Developmental Toxicity of Methanol 4-66
4.6.2. Role of Folate Deficiency in the Developmental Toxicity of Methanol 4-71
4.6.3. Methanol-Induced Formation of Free Radicals, Lipid Peroxidation, and Protein
Modifications 4-72
4.6.4. Exogenous Formate Dehydrogenase as a Means of Detoxifying the Formic Acid that
Results from Methanol Exposure 4-75
4.7. SYNTHESIS OF MAJOR NONCANCER EFFECTS 4-76
4.7.1. Summary of Key Studies in Methanol Toxicity 4-76
4.8. NONCANCER MOA INFORMATION 4-84
4.9. EVALUATION OF CARCINOGENICITY 4-87
4.10. SUSCEPTIBLE POPULATIONS AND LIFE STAGES 4-87
4.10.1. Possible Childhood Suscepptibility 4-87
4.10.2. Possible Gender Differences 4-88
4.10.3. Genetic Susceptibility 4-88
5. DOSE-RESPONSE ASSESSMENTS AND CHARACTERIZATION 5-1
5.1. INHALATION REFERENCE CONCENTRATION (RFC) 5-1
5.1.1. Choice of Principal Study and Critical Effect(s) 5-2
5.1.2. Methods of Analysis for the POD—Application of PBPK and BMD Models 5-7
5.1.3. RfC Derivation -Including Application of Uncertainty Factors 5-15
5.1.4. Previous RfC Assessment 5-20
5.2. ORAL REFERENCE DOSE (RFD) 5-21
5.2.1. Choice of Principal Study and Critical Effect-with Rationale and Justification 5-21
5.2.2. RfD Derivation-Including Application of UFs 5-24
5.2.3. RfD Derivation-Application of UFs 5-24
5.2.4. Previous RfD Assessment 5-25
5.3. UNCERTANTIES IN THE INHALATION RFC AND ORAL RFD 5-25
5.3.1. Choice of Endpoint 5-27
5.3.2. Choice of Dose Metric 5-28
5.3.3. Choice of Model for BMDL Derivations 5-30
5.3.4. Choice of Animal-to-Human Extrapolation Method 5-30
5.3.5. Route-to-Route Extrapolation 5-31
5.3.6. Statistical Uncertainty at the POD 5-31
5.3.7. Choice of Bioassay 5-31
5.3.8. Choice of Species/Gender 5-32
5.3.9. Human Population Variability 5-32
5.4. CANCER ASSESSMENT 5-32
March 2011 iv DRAFT - DO NOT CITE OR QUOTE
-------
6. MAJOR CONCLUSIONS IN CHARACTERIZATION OF HAZARD AND DOSE
RESPONSE 6-1
6.1. HUMAN HAZARD POTENTIAL 6-1
6.2. DOSE RESPONSE 6-3
6.2.1. Noncancer/Inhalation 6-4
6.2.2. Noncancer/Oral 6-6
7. REFERENCES 7-1
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION A-l
APPENDIX B. DEVELOPMENT, CALIBRATION AND APPLICATION OF A
METHANOL PBPK MODEL B-l
APPENDIX C. RFC DERIVATION OPTIONS C-l
APPENDIX D. RFC DERIVATION-COMPARISON OF DOSE METRICS D-l
March 2011 v DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF TABLES
Table 2-1. Relevant physical and chemical properties of methanol 2-1
Table 3-1. Background blood methanol and formate levels in humans 3-2
Table 3-2. Human blood methanol and formate levels following methanol exposure 3-3
Table 3-3. Monkey blood methanol and formate levels following methanol exposure 3-4
Table 3-4. Mouse blood methanol and formate levels following methanol exposure 3-4
Table 3-5. Rat blood methanol and formate levels following methanol exposure 3-5
Table 3-6. Plasma formate concentrations in monkeys 3-13
Table 3-7. Serum folate concentrations in monkeys 3-13
Table 3-8. Routes of exposure optimized in models - optimized against blood
concentration data 3-22
Table 3-9. Key methanol kinetic studies for model validation 3-25
Table 3-10. Parameters used in the mouse, rat and human PBPK models 3-29
Table 3-11. Primate Kms reported in the literature 3-42
Table 3-12. Parameter estimate results obtained using acslXtreme to fit all human data
using either saturable or first-order metabolism 3-42
Table 3-13. Comparison of LLFs for Michaelis-Menten and first-order metabolism 3-44
Table 3-14. Monkey group exposure characteristics 3-51
Table 4-1. Mortality rate for subjects exposed to methanol-tainted whiskey in relation to
their level of acidosis3 4-2
Table 4-2. Reproductive and developmental toxicity in pregnant Sprague-Dawley rats
exposed to methanol via inhalation during gestation 4-30
Table 4-3. Reproductive parameters in Sprague-Dawley dams exposed to methanol during
pregnancy then allowed to deliver their pups 4-31
Table 4-4. Developmental effects in mice after methanol inhalation 4-34
Table 4-5. Benchmark doses at two added risk levels 4-34
Table 4-6. Reproductive parameters in monkeys exposed via inhalation to methanol during
prebreeding, breeding, and pregnancy 4-37
Table 4-7. 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 4-39
Table 4-8. Maternal and litter parameters when pregnant female C57BL/6J mice were
injected i.p. with methanol 4-42
March 2011 vi DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-9. 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 4-45
Table 4-10. Brain weights of rats exposed to methanol vapors during gestation and
lactation 4-54
Table 4-11. Effect of methanol on Wistar rat acetylcholinesterase activities 4-57
Table 4-12. Effect of methanol on neutrophil functions in in vitro and in vivo studies in
male Wistar rats 4-62
Table 4-13. Effect of intraperitoneally injected methanol on total and differential leukocyte
counts and neutrophil function tests in male Wistar rats 4-63
Table 4-14. Effect of methanol exposure on animal weight/organ weight ratios and on cell
counts in primary and secondary lymphoid organs of male Wistar rats 4-65
Table 4-15. The effect of methanol on serum cytokine levels in male Wistar rats 4-66
Table 4-16. 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 4-67
Table 4-17. Summary of ontogeny of relevant enzymes in CD-I mice and humans 4-68
Table 4-18. Dysmorphogenic effect of methanol and formate in neurulating CD-I mouse
embryos in culture (GD8) 4-69
Table 4-19. 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 4-74
Table 4-20. 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 4-75
Table 4-21. Summary of noncancer effects reported in studies of methanol toxicity in
experimental animals (oral) 4-77
Table 4-22. Summary of studies of methanol toxicity in experimental animals (inhalation
exposure) 4-78
Table 5-1. Summary of studies considered most appropriate for use in derivation of an
RfC 5-6
Table 5-2. The EPA PBPK model estimates of methanol blood levels (AUC)a in rat dams
following inhalation exposures and reported brain weights of 6 week old male
pups 5-11
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 5-13
Table 5-4. Summary of PODs for critical endpoints, application of UFs and conversion to
HEC values using BMD and PBPK modeling 5-16
March 2011 vii DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-5. Summary of uncertainties in methanol noncancer risk assessment 5-26
Table B-l. Parameters used in the mouse and human PBPK models B-6
Table B-2. Primate kms reported in the literature B-31
Table B-3. Parameter estimate results obtained using acslXtreme to fit all human data
using either saturable or first-order metabolism B-34
Table B-4. Comparison of LLF for Michaelis-Menten and first-order metabolism B-34
Table B-5. PBPK model predicted Cmax and 24-hour AUC for mice and humans exposed
toMeOH B-37
Table B-6. Mouse total MeOH metabolism/metabolites produced following inhalation
exposures B-98
Table B-7. Human total MeOH metabolism/metabolites produced from inhalation
exposures B-99
Table B-8. Human total MeOH metabolism/metabolites produced following oral
exposures B-99
Table B-9. Repeated daily oral dosing of humans with MeOH B-101
Table C-l. The EPA PBPK model estimates of methanol blood levels (AUC) in rat dams
following inhalation exposures and reported brain weights of 6-week old male
pups C-3
Table C-2. Comparison of BMDiso results for decreased brain weight in male rats at 6
weeks of age using modeled AUC of methanol as a dose metric C-5
Table C-3. Comparison of BMDos results for decreased brain weight in male rats at 6
weeks of age using modeled AUC of methanol as a dose metric C-10
Table C-4. EPA's PBPK model estimates of methanol blood levels (Cmax) in rats
following inhalation exposures C-l5
Table C-5. Comparison of BMDiso results for decreased brain weight in male rats at 8
weeks of age using modeled Cmax of methanol as a dose metric C-16
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 C-21
Table C-l. EPA's PBPK model estimates of methanol blood levels (Cmax) in mice
following inhalation exposures C-26
Table C-8. Comparison of BMD modeling results for cervical rib incidence in mice using
modeled Cmax of methanol as a common dose metric C-27
Table C-9. Comparison of BMD modeling results for cervical rib incidence in mice using
modeled Cmax of methanol as a common dose metric C-35
Table C-10. Comparison of BMD modeling results for VDR in female monkeys using
AUC blood methanol as the dose metric C-44
March 2011 viii DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF FIGURES
Figure 3-1. Methanol metabolism and key metabolic enzymes in primates and rodents 3-7
Figure 3-2. Folate-dependent formate metabolism. Tetrahydrofolate (THF)-mediated one
carbon metabolism is required for the synthesis of purines, thymidylate, and
methionine 3-8
Figure 3-3. Plot of fetal (amniotic) versus maternal methanol concentrations in GD20 rats. ..3-11
Figure 3-4. Schematic of the PBPK model used to describe the inhalation, oral, and i.v.
route pharmacokinetics of methanol 3-26
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 3-31
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) 3-32
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 3-34
Figure 3-8. NP rat i.v. route methanol blood kinetics 3-37
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 3-37
Figure 3-10. Model fits to datasets from oral exposures to 100 and 2,500 mg/kg methanol
in female Sprague-Dawley rats 3-39
Figure 3-11. Urinary methanol elimination concentration (upper panels) and cumulative
amount (lower panel) following inhalation exposures to methanol in human
volunteers 3-41
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 3-43
Figure 3-13. Inhalation exposures to methanol in human volunteers 3-45
Figure 3-14. Blood methanol concentration data from NP and pregnant monkeys 3-48
Figure 3-15. Chamber concentration profiles for monkey methanol exposures 3-50
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.5-14
Figure 5-2. PODs (in mg/m3) for selected endpoints with corresponding applied UFs 5-17
Figure B-l. Schematic of the PBPK model used to describe the inhalation, oral, and i.v.
route pharmacokinetics of MeOH B-3
March 2011 ix DRAFT - DO NOT CITE OR QUOTE
-------
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 B-8
Figure B-3. Simulation of inhalation exposures to MeOH in NP mice from Perkins et al.
(1995a) (8-hour exposures) and Dorman et al. (1995), (6-hour exposures) B-9
Figure B-4. Oral exposures to MeOH in pregnant mice on GD8 (Dorman et al., 1995) or
NP and GDIS B-ll
Figure B-5. Mouse intravenous route MeOH blood kinetics B-13
Figure B-6. Mouse model inhalation route sensitivity coefficients for metabolic
parameters B-17
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 B-17
Figure B-8. Mouse model sensitivity coefficients for oral exposures to MeOH B-18
Figure B-9. Mouse daily drinking water ingestion pattern (A) and resulting predicted blood
concentration fora 6 d/wk exposure (B) B-19
Figure B-10. NP rat i.v.-route methanol blood kinetics B-21
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 B-22
Figure B-12. Model fits to data sets from oral exposures to 100 (squares) or 2,500
(diamonds) mg/kg MeOH in female Sprague-Dawley rat B-24
Figure B-13. Simulated Sprague-Dawley rat inhalation exposures to 500, 1,000, or 2,000
ppm MeOH B-26
Figure B-14. Simulated rat oral exposures of Sprague-Dawley rats to 65.9, 624.2, or 2,177
mg MeOH/kg/day B-27
Figure B-15. Rat daily drinking water ingestion pattern (A) and resulting predicted blood
concentration fora 2-day exposure (B) B-28
Figure B-16. Urinary MeOH elimination concentration (upper panel) and cumulative
amount (lower panel), following inhalation exposures to MeOH in human
volunteers B-30
Figure B-17. Control (nonexposed) blood methanol concentrations B-32
Figure B-18. Data showing the visual quality of the fit using optimized first-order or
Michaelis-Menten kinetics to describe the metabolism of MeOH in humans B-33
Figure B-19. Inhalation exposures to MeOH in human volunteers B-35
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) B-3 8
March 2011 x DRAFT - DO NOT CITE OR QUOTE
-------
Figure B-21. Fit of the model to i.v. data using different metabolism and uptake parameter
optimizations B-43
Figure B-22. Fit of the model to oral data using different metabolism and uptake parameter
optimizations B-44
Figure B-23. Fit of the model to inhalation data using different metabolism and uptake
parameter optimizations B-44
Figure B-24. Simulated human oral exposures to 0.1 mg MeOH/kg/-day comparing the
first few days for four exposure scenarios B-100
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 C-8
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 C-13
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 C-l9
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 C-24
Figure C-5. Nested Logistic Model, 0.1 Extra Risk - Incidence of Cervical Rib in Mice
versus Cmax Methanol, GD 6-15 inhalational study C-33
Figure C-6. Nested Logistic Model, 0.05 Extra Risk - Incidence of Cervical Rib in Mice
versus Cmax Methanol, GD 6-15 inhalational study C-41
Figure C-l. 3rd Degree Polynomial Model, BMR of 1 Control Mean S.D. - VDR in
female monkeys using AUC blood methanol as the dose metric C-46
Figure D-l. Exposure-response data for methanol-induced CR plus SNR malformations in
mice at various concentration-time combinations D-2
Figure D-2. Internal dose-response relationships for methanol-induced CR plus SNR
malformations in mice at various concentration-time combinations for three
dose metrics D-3
March 2011 xi DRAFT - DO NOT CITE OR QUOTE
-------
LIST OF ABBREVIATIONS AND 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
ANOVA analysis of variance
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
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
March 2011
XII
DRAFT - DO NOT CITE OR QUOTE
-------
CHL
CI
Cls
CNS
con-A
CR
CSF
CT
CYP450
d, 5, A
D2
DA
DIPE
DMDC
DNA
DNT
DOPAC
DPC
DTK
EPS A
EKG
EO
EPA
ERF
EtOH
F
F0
FI
F2
F344
FAD
FAS
FD
FP
FR
FRACIN
Chinese hamster lung (cells)
confidence interval
clearance rate
central nervous system
carbon dioxide
concanavalin-A
crown-rump length
Cancer slope factor
computed tomography
cytochrome P450
delta, difference, change
dopamine receptor
dopamine
diisopropylether
dimethyl dicarbonate
deoxyribonucleic acid
developmental neurotoxicity test(ing)
dihydroxyphenyl acetic acid
days past conception
delayed-type hypersensitivity
European Food Safety Authority
electrocardiogram
Executive Order
U.S. Environmental Protection Agency
European Ramazzini Foundation
ethanol
fractional bioavailability
parental generation
first generation
second generation
Fisher 344 rat strain
folic acid deficient
folic acid sufficient
formate dehydrogenase
folate paire
folate reduced
fraction inhaled
March 2011
XIII
DRAFT - DO NOT CITE OR QUOTE
-------
FS folate sufficient
FSH follicular stimulating hormone
y-GT gamma glutamyl transferase
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
HCHO formaldehyde
HCOO formate
Hct hematocrit
HEC human equivalent concentration
FED 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
March 2011
DRAFT - DO NOT CITE OR QUOTE
-------
K1C
KAI
KAS
KBL
KIA
KLH
KLL
Km
Km2
KSI
L, dL, mL
LD50
LDH
LH
LLF
LMI
LOAEL
M, mM, jiM
MeOH
MLE
M-M
MN
MOA
4-MP
MRI
MTBE
MIX
N2O/O2
NAD+
NADH
NET
NCEA
ND
NEDO
NIEHS
NIOSH
first order clearance of methanol from the blood to the bladder for urinary
elimination
first order uptake from the intestine
first order methanol oral absorption rate from stomach
rate constant for urinary excretion from bladder
first order uptake from intestine
keyhole limpet hemocyanin
alternate first order rate constant
substrate concentration at half the enzyme maximum velocity (Vmax)
Michaelis-Menten rate constant for low affinity metabolic clearance of
methanol
first order transfer between stomach and intestine
liter, deciliter, milliliter
median lethal dose
lactate dehydrogenase
luteinizing hormone
(maximum) log likelihood function
leukocyte migration inhibition (assay)
lowest-observed-adverse-effect level
molar, millimolar, micromolar
methanol
maximum likelihood estimate
Michaelis-Menten
micronuclei
mode of action
4-methylpyrazale messenger RNA
magnetic resonance imaging
methyl tertiary butyl ether
methotrexate
nitrous oxide
nicotinamide adenine dinucleotide
reduced form of nicotinamide adenine dinucleotide
nitroblue tetrazolium (test)
National Center for Environmental Assessment
not determined
New Energy Development Organization (of Japan)
National Institute of Environmental Health Sciences
National Institute of Occupational Safety and Health
March 2011
xv DRAFT - DO NOT CITE OR QUOTE
-------
nmol nanomole
NOAEL no-observed-adverse-effect level
NOEL no-observed-effect level
NP nonpregnant
NR not reported
NRC National Research Council
NS not specified
NTP National Toxicology Program
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, ppm parts per billion, parts per million
PWG Pathology Working Group of the NTP of NIEHS
Q wave the initial deflection of the QRS complex
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
March 2011
xvi DRAFT - DO NOT CITE OR QUOTE
-------
S.D. standard deviation
S.E. standard error
SEM standard error of mean
SGPT serum glutamate pyruvate transaminase
SHE Syrian hamster embryo
SOD superoxide dismutase
SOP standard operating procedure(s)
t time
TV,, ty2 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)
UFA UF associated with interspecies (animal to human) extrapolation
UFD UF associated with deficiencies in the toxicity database
UFH 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-affmity/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
March 2011
xvii 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 and
reference concentration, 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.
March 2011 xviii 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
March 2011 xix DRAFT - DO NOT CITE OR QUOTE
-------
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
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
March 2011 xx DRAFT - DO NOT CITE OR QUOTE
-------
Angela Howard, Ph.D.
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
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
March 2011 xxi DRAFT - DO NOT CITE OR QUOTE
-------
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
ToxStrategies
Katy, TX
Cynthia VanLandingham, M.S.
ENVIRON International (Appendix E)
650 Poydras Street
New Orleans, LA
Lutz Weber, Ph.D., DABT
Oak Ridge Institute for Science and Education
Badger Road
Oak Ridge, TN
Errol Zeiger, Ph.D.
Alpha-Gamma Technologies, Inc.
3301 Benson Drive
Raleigh, NC
March 2011 xxii DRAFT - DO NOT CITE OR QUOTE
-------
REVIEWERS
This document has been provided for review to EPA scientists and interagency reviewers
from other 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
Ha 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
March 2011 xxiii DRAFT - DO NOT CITE OR QUOTE
-------
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
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
Connie Meacham, M.S.
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
Deborah Wales
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
March 2011 xxiv 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 magnitude)
9 of a daily exposure to the human population (including sensitive subgroups) that is likely to be
10 without an appreciable risk of deleterious effects during a lifetime. The inhalation RfC
11 (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
16 (>24 hours up to 30 days), and subchronic (>30 days up to 10% of lifetime) exposure durations,
17 all of 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 Development of these hazard identification and dose-response assessments for the
21 noncancer effects of methanol has followed the general guidelines for risk assessment as set forth
22 by the National Research Council (NRC) (1983). EPA Guidelines and Risk Assessment Forum
23 Technical Panel Reports that may have been used in the development of this assessment include
24 the following: Guidelines for the Health Risk Assessment of Chemical Mixtures (U.S. EPA,
25 1986a), Recommendations for and Documentation of Biological Values for Use in Risk
26 Assessment (U.S. EPA, 1988), Guidelines for Developmental Toxicity Risk Assessment (U.S.
27 EPA, 1991), Interim Policy for Particle Size and Limit Concentration Issues in Inhalation
28 Toxicity Studies (U.S. EPA, 1994a), Methods for Derivation of Inhalation Reference
29 Concentrations and Application of Inhalation Dosimetry (U.S. EPA, 1994b), Use of the
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 1 -1 DO NOT CITE OR QUOTE
-------
1 Benchmark Dose Approach in Health Risk Assessment (U.S. EPA, 1995), Guidelines for
1 Reproductive Toxicity Risk Assessment (U.S. EPA, 1996), Guidelines for Neurotoxicity Risk
3 Assessment (U.S. EPA, 1998), Science Policy Council Handbook: Risk Characterization (U.S.
4 EPA, 2000c), Benchmark Dose Technical Guidance Document (U.S. EPA, 2000a),
5 Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures (U.S.
6 EPA, 2000d), A Review of the Reference Dose and Reference Concentration Processes (U.S.
7 EPA, 2002), Science Policy Council Handbook: Peer Review (U.S. EPA, 2006b), and A
8 Framework for Assessing Health Risks of Environmental Exposures to Children (U.S. EPA,
9
10 The literature search strategy employed for this compound was based on the Chemical
11 Abstracts Service Registry Number (CASRN) and at least one common name. Any pertinent
12 scientific information submitted by the public to the IRIS Submission Desk was also considered
13 in the development of this document. The relevant literature was reviewed through October,
14 2010.
March 2011 1-2 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. Some relevant physical and chemical properties are listed in Table 2-1
4 below (HSDB. 2009: IPCS. 1997).
Table 2-1. Relevant physical and chemical properties of methanol
CASRN:
Empirical formula:
Molecular weight:
Vapor pressure:
Vapor Density:
Specific gravity:
Boiling point:
Melting point:
Water solubility:
Log octanol-water partition coefficient:
Conversion factor (in air):
67-56-1
CH3OH
32.04
160 mmHg at 30 °C
1.11
0.7866 g/mL (25 °C)
64.7 °C
-98 °C
Miscible
-0.82 to -0.68
1 ppm= 1.31 mg/m3;
1 mg/m3 = 0.763 ppm
5 Methanol is a clear, colorless liquid that has an alcoholic odor (IPCS, 1997). Endogenous
6 levels of methanol are present in the human body as a result of both metabolism1 and dietary
7 sources such as fruit, fruit juices, vegetables and alcoholic beverages,2 and can be measured in
8 exhaled breath and body fluids (CERHR. 2004: IPCS. 1997: Turner et al.. 2006). Dietary
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
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 (SiragusaetaL 1988). Increased levels of
methanol in blood and exhaled breath have also been observed after the consumption of ethanol (Fisher et al.. 2000).
March 2011
2-1
DRAFT-DO NOT CITE OR QUOTE
-------
1 exposure to methanol also occurs through the intake of some food additives. The artificial
2 sweetener aspartame and the beverage yeast inhibitor dimethyl dicarbonate (DMDC) release
3 methanol as they are metabolized (Stegink et al., 1989). In general, aspartame exposure does not
4 contribute significantly to the background body burden of methanol (Butchko et al., 2002). Oral,
5 dermal, or inhalation exposure to methanol in the environment, consumer products, or workplace
6 also occur.
7 Methanol is a high production volume chemical with many commercial uses and it is a
8 basic building block for hundreds of chemical products. Many of its derivatives are used in the
9 construction, housing or automotive industries. Consumer products that contain methanol
10 include varnishes, shellacs, paints, windshield washer fluid, antifreeze, adhesives, de-icers, and
11 Sterno heaters. In 2009, the Methanol Institute (2009b) estimated a global production capacity
12 for methanol of about 35 million metric tons per year (close to 12 billion gallons), a production
13 capacity in the United States (U.S.) of nearly 3.7 million metric tons (1.3 billion gallons), and a
14 total U.S. demand for methanol of over 8 million metric tons. Methanol is among the highest
15 production volume chemicals reported in the U.S. EPA's Toxic Release Inventory (TRI).3 It is
16 among the top chemicals on the 2008 TRI lists of chemicals with the largest total on-site and off-
17 site recycling (6th), energy recovery (2nd) and treatment (1 st) (U.S. EPA. 2009b). TRI also
18 reports that approximately 135,000,000 pounds of methanol was released or disposed of in the
19 United States in 2008, making methanol among the top five chemicals on the list entitled "TRI
20 On-site and Off-site Reported Disposed of or Otherwise Released in pounds for facilities in All
21 Industries for Hazardous Air Pollutant Chemicals U.S. 2008" (U.S. EPA. 2009d).
22 While production has switched to other regions of the world, demand for methanol is
23 growing steadily in almost all end uses. A large reason for the increase in demand is its use in the
24 production of biodiesel, a low-sulfur, high-lubricity fuel source. Global demand for biodiesel is
25 forecast to increase by 32% per year, rising from 30 million gallons in 2004, to 150 million
26 gallons by 2008, and to 350 million gallons by 2013 (Methanol Institute. 2009a). Power
27 generation and fuel cells could also be large end users of methanol in the near future (Methanol
28 Institute, 2009b).
3 The information in TRI does not indicate whether (or to what degree) the public has been exposed to toxic
chemicals. Therefore, no conclusions on the potential risks can be made based solely on this information (including
any ranking information). For more detailed information on this subject refer to The Toxics Release Inventory (TRI)
and Factors to Consider When Using TRI Data (U.S. EPA. 2009c).
March 2011 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 (CERHR, 2004;
3 IPCS, 1997; Turner et al., 2006). Table 3-1 summarizes background blood methanol levels in
4 healthy 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 & 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.4
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
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
4 In general, background levels among people who are on normal/non-restricted diets will be higher than those
reported.
March 2011 3-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 higher rate of increase of methanol concentrations in the body when exposed to the same
2 concentration in air.
Table 3-1. Background blood methanol and formate levels in humans
Description of human subjects
12 males on restricted diet (no
metnanol-containing or methanol-
producing foods) for 12 hr
22 adults on restricted diet (no
metnanol-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 etal. (1995)
Lee etal. (1992)
Lee etal. (1992)
Battermanetal.(1998)
Batterman and
Franzblau (1997)
Davoli etal. (1986)
Stegink etal. (1981)
Stegink et al. (1983)
Source: CERHR (2004).
March 2011
3-2 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; post
exposure samples
Males and females;
post exposure serum
levels
Males without
exercise; post
exposure blood
methanol and
plasma formate
Males with exercise;
post exposure blood
methanol and
plasma formate
Females; post
exposure samples
Exposure
Oral
Oral
Oral
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Exposure
duration or
method
1 dose in
juice
beverage
1 dose in
water
75 min
4hr
6hr
6hr
8hr
Methanol
exposure
concentration
0
3.4mg/kgbwa
10 mg/kg bwa
15mg/kgbwa
20 mg/kg bwa
0
3.4mg/kgbwa
5 mg/kg bwa
10 mg/kg bwa
o
0.6-0.87
mg/kg bwa
0
191 ppm
0
200 ppm
o
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.
t i oo^
(iyy^i
Batterman et
al. (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).
March 2011
3-3 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; post
exposure blood level
Exposure
route
Inhalation
Inhalation
Exposure
duration
2.5 hr/day,
7days/wk during
premating,
mating, and
gestation (348
days)
6hr
]Methanol
exposure
concentration
0
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
and Shen et
al (1999-
2QQ4)
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; post
exposure plasma methanol
and peak formate level
Mouse;CD-l;female; post
exposure blood methanol
level
Mouse;CD-l;female; mean
post exposure plasma
methanol level
Mouse;CD-l;female; plasma
level 1 hr post dosing
Mouse;CD-l;female; peak
plasma level
Exposure
route
Inhalation
Inhalation
Inhalation
Oral-
Gavage
Oral-
Gavage
Exposure
duration
6 hr on GD8
8hr
7 hr/day on
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
1,500 mg/kg
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.
ri995a)
Rogers et al.
(1993)
Dorman et al.
(1995)
4 -MP=4 -me thy Ipy razole
Source: CERHR (2004).
March 2011
3-4 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;
post exposure blood methanol
level on 3 days
Rat;Sprague-Dawley; female;
post exposure blood methanol
level
Rat;LongEvans;female; post
exposure plasma level on
GD7-GD12
Rat;LongEvans;female; 1 hr
post exposure blood level
Rat;Long-Evans;male and
female; 1 hr post exposure
blood level in pups
Rat/Fischer-344/male; post
exposure blood level
Rat;Long-Evans;male; post-
exposure serum level
Rat/Fischer-344/male; 25 min
post exposure blood level for
4-wk animals; -250 min post
exposure for 104-wk animals
Rat/Fischer-344/ female; 25
min post exposure blood level
for 4-wk animals; —250 min
post exposure for 104-wk
animals
Rat;Long-Evans;male; peak
blood formate level
Rat;Long-Evans;male; peak
blood methanol and formate
Exposure
route
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Inhalation
Oral-
gavage
Exposure
duration
7 hr/day
for 19 days
8hr
7 hr/day on
GD7-GD1
9
6 hr/day on
GD6-
PND21
6 hr/day on
PND1-
PND21
6hr
6hr
19.5 hr/day
for 4/104
19 hr/day
for 4/104
wk
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
0 ppm
10 ppm
100 ppm
1,000 ppm
Oppm
10 ppm
100 ppm
1,000 ppm
OFS
OFS
1,200 ppm-FS
1,200 ppm-FR
2,000 ppm-FS
2,000 ppm-FR
3,500mg/kgbw-FS
3,500 mg/kgbw-FP
3,500 mg/kgbw-FR
3,OOOmg/kgbw/day-FS
3,OOOmg/kgbw/dayFR
2,000 mg/kgbw/dayFS
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
4.01/3.78
1.56/3.32
3.84/3.32
53.59/12.08
13.39 /3.60
6.73/3.70
4.34/4.32
88.33/8.50
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
No data
No data
8.3
10.1
8.3
46
8.3
83
382
860
9.2
718
9 2
538
Reference
Nelson et al.
(19851
Pollack and
Brouwer
(1996);
Perkins et
al. (1995a)
Stanton et
al. (19951
Weiss et al.
(1996)
Horton et
al. (1992)
Cooper et
al. (1992)
NEDO
(2008b)
NEDO
(2008b)
Lee et al.
(1994)
Lee et al.
(1994)
FS = Folate sufficient; FR = Folate reduced; FP = Folate paire
Source: CERHR (2004).
March 2011
3-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 At doses that do not saturate metabolic pathways, a small percentage of methanol is
2 excreted directly in urine. Because of the high blood:air partition coefficient for methanol and
3 rapid metabolism in all species studied, the bulk of clearance occurs by metabolism, though
4 exhalation and urinary clearance become more significant when doses or exposures are
5 sufficiently high to saturate metabolism (subsequently in this document, "clearance" refers to
6 elimination by all routes, including metabolism, as indicated by the decline in methanol blood
7 concentrations.) Metabolic saturation and the corresponding clearance shift have not been
8 observed in humans and nonhuman primates because doses used were limited to the linear range,
9 but the enzymes involved in primate metabolism are also saturable.
10 The primary route of methanol elimination in mammals is through a series of oxidation
11 reactions that form formaldehyde, formate, and carbon dioxide (Figure 3-1). As noted in
12 Figure 3-1, methanol is converted to formaldehyde by alcohol dehydrogenase-1 (ADH1) in
13 primates and by catalase (CAT) and ADH1 in rodents. Although the first step of metabolism
14 occurs through different pathways in rodents and nonhuman primates, Kavet and Nauss (1990)
15 report that the reaction proceeds at similar rates (Vmax = 30 and 48 mg/h/kg in rats and nonhuman
16 primates, respectively). In addition to enzymatic metabolism, methanol can react with hydroxyl
17 radicals to spontaneously yield formaldehyde (Harris et al., 2003). Mannering et al. (1969) also
18 reported a similar rate of methanol metabolism in rats and monkeys, with 10 and 14% of a 1 g/kg
19 dose oxidized in 4 hours, respectively; the rate of oxidation by mice was about twice as fast, 25%
20 in 4 hours. In an HEI study by Pollack and Brouwer (1996), the metabolism of methanol was
21 2 times as fast in mice versus rats, with a Vmax for elimination of 117 and 60.7 mg/h/kg,
22 respectively. Despite the faster elimination rate of methanol in mice versus rats, mice
23 consistently exhibited higher blood methanol levels than rats when inhaling equivalent methanol
24 concentrations (See Tables 3-4 and 3-5). Possible explanations for the higher methanol
25 accumulation in mice include faster respiration (inhalation rate/body weight) and increased
26 fraction of absorption by the mouse (Perkins et al., 1995a). Because smaller species generally
27 have faster breathing rates than larger species, humans would be expected to absorb methanol via
28 inhalation more slowly than rats or mice inhaling equivalent concentrations. If humans eliminate
29 methanol at a comparable rate to rats and mice, then humans would also be expected to
30 accumulate less methanol than those smaller species. However, if humans eliminate methanol
31 more slowly than rats and mice, such that the ratio of absorption to elimination stays the same,
32 then humans would be expected to accumulate methanol to the same internal concentration but to
33 take longer to reach that concentration.
34 In all species, formaldehyde is rapidly converted to formate, with the half-life for
35 formaldehyde being ~1 minute. Formaldehyde is oxidized to formate by two metabolic pathways
March 2011 3-6 DRAFT - DO NOT CITE OR QUOTE
-------
1 (Teng et al., 2001). The first pathway (not shown in Figure 3-1) involves conversion of free
2 formaldehyde to formate by the so-called low-affinity pathway (affinity = l/Km = 0.002/(jM)
3 mitochondrial aldehyde dehydrogenase-2 (ALDH2). The second pathway (Figure 3-1) involves a
4 two-enzyme system that converts glutathione-conjugated formaldehyde
5 (^-hydroxymethylglutathione [FDVIGSH]) to the intermediate ^-formylglutathione, which is
6 subsequently metabolized to formate and glutathi one (GSH) by ^-formylglutathione hydrolase.5
7 The first enzyme in this pathway, formaldehyde dehydrogenase-3 (ADH3), is rate limiting, and
8 the affinity of HMGSH for ADH3 (affinity = l/Km = 0.15/nM) is about a 100-fold higher than
9 that of free formaldehyde for ALDH2. In addition to the requirement of GSH for ADH3 activity,
10 oxidation by ADH3 is nicotinamide adenine dinucleotide- (NAD+-)dependent. Under normal
11 physiological conditions NAD+ levels are about two orders of magnitude higher than NADH, and
12 intracellular GSH levels (mM range) are often high enough to rapidly scavenge formaldehyde
13 (Meister & Anderson, 1983; Svensson et al., 1999): thus, the oxidation of HMGSH is favorable.
14 In addition, genetic ablation of ADH3 results in increased formaldehyde toxicity (Deltour et al.,
15 1999). These data indicate that ADH3 is likely to be the predominant enzyme responsible for
16 formaldehyde oxidation at physiologically relevant concentrations, whereas ALDHs likely
17 contribute to formaldehyde elimination at higher concentrations (Dicker & Cedebaum, 1986).
Primates CH3OH (Methanol) Rodents
ihydro
(ADH1)
Alcohol dehydrogenase HCHO Catalase (CAT) andADHl
(Formaldehyde)
U + GSH)
HMGSH
(hydroxymethyl-GSH)
Formaldehyde dehydrogenase \, Formaldehyde dehydrogenase
(ADH3) (S-formyl glutathione) (ADH3)
U - GSH)
S-formylglutathione hydrolase HCOO (Formate) S-formylglutathione hydrolase
Folate dependent pathway CAT-peroxide and
(see Figure 3-2) CO2 (Carbon dioxide) Folate-dependent pathway
(see Figure 3-2)
Figure 3-1. Methanol metabolism and key metabolic enzymes in primates
and rodents.
Source: IPCS (1997).
5 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).
March 2011 3-7 DRAFT - DO NOT CITE OR QUOTE
-------
1 Rodents convert formate to carbon dioxide (CO2) through a folate-dependent enzyme
2 system and a CAT-peroxide system (Dikalova et al., 2001). Formate can undergo adenosine
3 triphosphate- (ATP-) dependent addition to tetrahydrofolate (THF), which can carry either one or
4 two one-carbon groups. Formate can conjugate with THF to form jV10-formyl-THF and its
5 isomer A^-formyl-TFIF, both of which can be converted to TV5, jV70-methenyl-TFIF and
6 subsequently to other derivatives that are ultimately incorporated into DNA and proteins via
7 biosynthetic pathways (Figure 3-2). There is also evidence that formate generates CO2" radicals,
8 and can be metabolized to CO2 via CAT and via the oxidation of jV10-formyl-THF (Dikalova et
9 al.. 2001).
Cytoplasm
Mitochondria
10-foraiylTHF
NV^^^ format
methenylTHF THF
methyleneTHF
iO-formylTHF
PURINES
methenylTHF
methyleneTHF
I MTHFR
5-methylTHF
homocysteme
AdoHyc
dTMP
Methionine
AdoMct
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).
10 Unlike rodents, formate metabolism in primates occurs solely through a folate-dependent
11 pathway. Black et al. (1985) reported that hepatic TFIF levels in monkeys are 60% of that in rats,
12 and that primates are far less efficient in clearing formate than are rats and dogs. Studies
13 involving [14C]formate suggest that -80% is exhaled as 14CO2, 2-7% is excreted in the urine, and
14 -10% undergoes metabolic incorporation (Hanzlik et al., 2005, and references therein). Mice
15 deficient in formyl-TFIF dehydrogenase exhibit no change in LD50 (via intraperitoneal [i.p.]) for
16 methanol or in oxidation of high doses of formate. Thus it has been suggested that rodents
March 2011
3-8 DRAFT - DO NOT CITE OR QUOTE
-------
1 efficiently clear formate via folate-dependent pathways, peroxidation by CAT, and by an
2 unknown third pathway; conversely, primates do not appear to exhibit such capacity and are more
3 sensitive to metabolic acidosis following methanol poisoning (R. J. Cook et al., 2001).
4 Blood methanol and formate levels measured in humans under various exposure
5 scenarios are reported in Table 3-2. As noted in Table 3-2, 75-minute to 6-hour exposures of
6 healthy humans to 200 parts per million (ppm) methanol vapors, the American Council of
7 Governmental Industrial Hygienists (ACGIH) threshold limit value (TLV) for occupational
8 exposure (2000), results in increased levels of blood methanol but not formate. A limited number
9 of monitoring studies indicate that levels of methanol in outdoor air are orders of magnitude
10 lower than the TLV (1997). Table 3-3 indicates that exposure of monkeys to 600 ppm methanol
11 vapors for 2.5 hours increased blood methanol but not blood formate levels. Normal dietary
12 exposure to aspartame, which releases 10% methanol during metabolism, is unlikely to
13 significantly increase blood methanol or formate levels (Butchko et al., 2002). Data in Table 3-2
14 suggest that exposure to high concentrations of aspartame is unlikely to increase blood formate
15 levels; no increase in blood formate levels were observed in adults ingesting "abusive doses"
16 (100-200 mg/kg) of aspartame (Stegink et al.. 1981). Kerns et al. (2002) studied the kinetics of
17 formate in 11 methanol-poisoned patients (mean initial methanol level of 57.2 mmol/L or 1.83
18 g/L) and determined an elimination half-life of 3.4 hours for formate. Kavet and Nauss (1990)
19 estimated that a methanol dose of 11 mM or 210 mg/kg is needed to saturate folate-dependent
20 metabolic pathways in humans. There are no data on blood methanol and formate levels
21 following methanol exposure of humans with reduced ADH activity or marginal folate tissue
22 levels, a possible concern regarding sensitive populations. As discussed in greater detail in
23 Section 3.2, a limited study in folate-deficient monkeys demonstrated no increase in blood
24 formate levels following exposure to 900 ppm methanol vapors for 2 hours. In conclusion,
25 limited available data suggest that typical occupational, environmental, and dietary exposures are
26 likely to increase baseline blood methanol but not formate levels in most humans.
3.2. KEY STUDIES
27 Some recent toxicokinetic and metabolism studies (Burbacher, Grant, et al., 1999;
28 Burbacher. Grant et al.. 2004: Dormanetal.. 1994: Medinsky et al.. 1997: Pollack & Brouwer,
29 1996) provide key information on interspecies differences, methanol metabolism during
30 gestation, metabolism in the nonhuman primate, and the impact of folate deficiency on the
31 accumulation of formate.
32 As part of an effort to develop a physiologically based toxicokinetic model for methanol
33 distribution in pregnancy, Pollack and Brouwer (1996) conducted a large study that compared
March 2011 3-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 toxicokinetic differences in pregnant and nonpregnant (NP) rats and mice. Methanol disposition6
2 was studied in Sprague-Dawley rats and CD-I mice that were exposed to 100-2,500 mg/kg of
3 body weight pesticide-grade methanol in saline by intravenous (i.v.) or oral routes. Exposures
4 were conducted in NP rats and mice, pregnant rats on gestation days (GD)7, GDI4, and GD20,
5 and pregnant mice on GD9 and GDIS. Disposition was also studied in pregnant rats and mice
6 exposed to 1,000-20,000 ppm methanol vapors for 8 hours. Three to five animals were
7 examined at each dose and exposure condition.
8 Based on the fit of various kinetic models to methanol measurements taken from all
9 routes of exposure, the authors concluded that high exposure conditions resulted in nonlinear
10 disposition of methanol in mice and rats.7 Both linear and nonlinear pathways were observed
11 with the relative contribution of each pathway dependent on concentration. At oral doses of
12 100-500 mg/kg of body weight, methanol was metabolized to formaldehyde and then formic acid
13 through the saturable nonlinear pathway. A parallel, linear route characteristic of passive-
14 diffusion accounted for an increased fraction of total elimination at higher concentrations.
15 Nearly 90% of methanol elimination occurred through the linear route at the highest oral dose of
16 2,500 mg/kg of body weight.
17 Oral exposure resulted in rapid and essentially complete absorption of methanol. No
18 significant change in blood area under the curve (AUC) methanol was seen between NP and
19 GD7, GD14 and GD20 rats exposed to single oral gavage doses of 100 and 2,500 mg/kg, nor
20 between NP and GD9 and GDIS mice at 2,500 mg/kg. The data as a whole suggested that the
21 distribution of orally and i.v. administered methanol was similar in rats versus mice and in
22 pregnant rodents versus NP rodents with the following exceptions:
23 • There was a statistically significant increase in the ratio of apparent volume of distribution
24 (Vd) to fractional bioavailability (F) by -20% (while F decreased but not significantly),
25 between NP and GD20 rats exposed to 100 mg/kg orally. However, this trend was not seen
26 in rats or mice exposed to 2,500 mg/kg, and the result in rats at 100 mg/kg could well be a
27 statistical artifact since both Vd and F were being estimated from the same data, making the
28 model effectively over-parameterized.
6 Methanol concentrations in whole blood and urine were determined by gas chromatography with flame ionization
detection (Pollack & Kawagoe. 1991).
7 A model incorporating parallel linear and nonlinear routes of methanol clearance was required to fit the data from
the highest exposure groups.
March 2011 3-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 • There were statistically significant decreases in the fraction of methanol absorbed by the fast
2 process (resulting in a slower rise to peak blood concentrations, though the peak is
3 unchanged) and in the Vmax for metabolic elimination between NP and GDIS mice. No
4 such differences were observed between NP and GD9 mice.
5 • The authors estimated a twofold higher Vmax for methanol elimination in mice versus rats
6 following oral administration of 2,500 mg/kg methanol, suggesting that similar oral doses
7 would result in lower methanol concentrations in the mouse versus rat.
8 Methanol penetration from maternal blood to the fetal compartment was examined in
9 GD20 rats by microdialysis.8 A plot of the amniotic concentration versus maternal blood
10 concentration (calculated from digitization of Figure 17 of Pollack and Brouwer (1996) report) is
11 shown in Figure 3-3. The ratio is slightly less than 1:1 (dashed line in plot) and appears to be
12 reduced with increasing methanol concentrations, possibly due to decreased blood flow to the
13 fetal compartment. Nevertheless, this is a very minor departure from linearity, consistent with a
14 substrate such as methanol that penetrates cellular membranes readily and distributes throughout
15 total body water.
_ 50001
~3>
,§ 4000 H
c
IS 3000-
•*->
c
c 2000H
o
o
•| 1000-
"E
< 0
1000 2000 3000 4000
Maternal blood concentration (mg/L)
5000
Figure 3-3. Plot of fetal (amniotic) versus maternal methanol concentrations
in GD20 rats. Note: 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).
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.
March 2011
3-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 Inhalation exposure resulted in less absorption in both rats and mice as concentrations of
2 methanol vapors increased, which was hypothesized to be due to decreased breathing rate and
3 decreased absorption efficiency from the upper respiratory tract.9 Based on blood methanol
4 concentrations measured following 8-hour inhalation exposures to concentrations ranging from
5 1,000-20,000 ppm, the study authors (Pollack & Brouwer, 1996) concluded that methanol
6 accumulation in the mouse occurred at a two- to threefold greater rate compared to the rat. They
7 speculated that faster respiration rate and more complete absorption in the nasal cavity of mice
8 may explain the higher methanol accumulation and greater sensitivity to certain developmental
9 toxicity endpoints (see Section 4.3.2).
10 The Pollack and Brouwer (1996) study was useful for comparing effects in pregnant and
11 NP rodents exposed to high doses, but the implication of these results for humans exposed to
12 ambient levels of methanol is not clear (2004).
13 Burbacher, Shen et al. (1999; 2004) examined toxicokinetics mMacaca fascicularis
14 monkeys prior to and during pregnancy. The study objectives were to assess the effects of
15 repeated methanol exposure on disposition kinetics, determine whether repeated methanol
16 exposures result in formate accumulation, and examine the effects of pregnancy on methanol
17 disposition and metabolism. Reproductive, developmental and neurological toxicity associated
18 with this study were also examined and are discussed in Sections 4.3.2 and 4.4.2. In a 2-cohort
19 design, 48 adult females (6 animals/dose/group/cohort) were exposed to 0, 200, 600, or
20 1,800 ppm methanol vapors (99.9% purity) for 2.5 hours/day, 7 days/week for 4 months prior to
21 breeding and during the entire breeding and gestation periods. Six-hour methanol clearance
22 studies were conducted prior to and during pregnancy. Burbacher, Shen et al. (1999; 2004)
23 reported that:
24
25 • At no point during pregnancy was there a significant change in endogenous methanol blood
26 levels, which ranged from 2.2-2.4 mg/L throughout.
27 • PK studies were performed initially (Study 1), after 90 days of pre-exposure and prior to
28 mating (Study 2), between GD66 and GD72 (Study 3), and again between GD126 and
29 GDI32 (Study 4). These studies were analyzed using classical PK (one-compartment)
30 models.
9 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 (~!/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.
March 2011 3-12 DRAFT - DO NOT CITE OR QUOTE
-------
1 • Disproportionate mean, dose-normalized, and net blood methanol dose-time profiles in the
2 600 and 1,800 ppm groups suggested saturation of the metabolism-dependent pathway. Data
3 from the 600 ppm group fit a linear model, while data from the 1,800 ppm group fit a
4 Michaelis-Menten model.
5 • Methanol elimination rates modestly increased between Study 1 and Study 2 (90 days prior to
6 mating). This change was attributed to enzyme induction from the subchronic exposure.
7 • Blood methanol levels were measured every 2 weeks throughout pregnancy, and while there
8 was measurement-to-measurement variation, there was no significant change or trend over
9 the course of pregnancy. There appears to be an upward trend in elimination half-life and
10 corresponding downward trend in blood methanol clearance between Studies 2, 3, and 4.
11 However, the changes are not statistically significant and the time-courses for blood methanol
12 concentration (elimination phase) appear fairly similar.
13 • Significant differences between pre-breeding and gestational blood plasma formate levels
14 were observed but were not dose dependent (Table 3-6).
15 • Significant differences in serum folate levels in periods prior to and during pregnancy were
16 not dose dependent (Table 3-7).
Table 3-6. Plasma formate concentrations in monkeys
Exposure Group
Control
200 ppm
600 ppm
1,800 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: Buibacher, Shen et al. (1999).
Table 3-7. Serum folate concentrations in monkeys
Exposure Group
Control
200 ppm
600 ppm
1,800 ppm
Mean serum folate level (ng/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: Buibacher. Shenetal. (1999).
17 An HEI review committee (Burbacher, Shen, et al., 1999) noted that this was a quality
18 study using a relevant species. Although the study can be used to predict effects in adequately
19 nourished individuals, the study may not be relevant to persons who are folate deficient.
March 2011
3-13 DRAFT - DO NOT CITE OR QUOTE
-------
1 A series of studies by Medinsky et al. (1997) and Dorman et al. (1994) examined
2 metabolism and pharmacokinetics of [14C]methanol and [14C]formate in normal and folate-
3 deficient cynomolgus, M. fascicularis monkeys that were exposed to environmentally relevant
4 concentrations of [14C]methanol through an endotracheal tube while anesthetized. In the first
5 stage of the study, 4 normal 12-year-old cynomolgus monkeys were each exposed to 10, 45, 200,
6 and 900 ppm [14C]methanol vapors (>98% purity) for 2 hours. Each exposure was separated by
7 at least 2 months. After the first stage of the study was completed, monkeys were given a folate-
8 deficient diet supplemented with 1% succinylsulfathiozole (an antibacterial sulfonamide used to
9 inhibit folic acid biosynthesis from intestinal bacteria) for 6-8 weeks in order to obtain folate
10 concentrations of <3 ng/mL serum and <120 ng/mL erythrocytes. Folate deficiency did not alter
11 hematocrit level, red blood cell count, mean corpuscular volume, or mean corpuscular
12 hemoglobin level. The folate-deficient monkeys were exposed to 900 ppm [14C]methanol for
13 2 hours. The results of the Medinsky et al. (1997) and Dorman et al. (1994) studies showed:
14 • Dose-dependent changes in toxicokinetics and metabolism did not occur as indicated by a
15 linear relationship between inhaled [14C]methanol concentration and end-of-exposure
16 blood [14C]methanol level, [14C]methanol AUC and total amounts of exhaled
17 [14C]methanol and [14C]carbon dioxide.
18 • Methanol concentration had no effect on elimination half-life (<1 hour) and percent
19 urinary [14C]methanol excretion (<0.01%) at all doses.
20 • Following exposure to 900 ppm methanol, urinary excretion or exhalation of
21 [14C]methanol did not differ significantly between monkeys in the folate sufficient and
22 deficient state. There was no significant [14C] formate accumulation at any dose.
23 • Peak blood [14C]formate levels were significantly higher in folate-deficient monkeys, but
24 did not exceed endogenous blood levels reported by the authors to be between 0.1 and
25 0.2 mmol/L (4.6-9.2 mg/L).
26 An HEI review committee (Medinsky et al., 1997) noted that absolute values in this study
27 cannot be extrapolated to humans because the use of an endotracheal tube in anesthetized
28 animals results in an exposure scenario that is not relevant to humans. However, the data in this
29 study suggest that a single exposure to an environmentally relevant concentration of methanol is
30 unlikely to result in a hazardous elevation in formate levels, even in individuals with moderate
31 folate deficiency.
March 2011 3-14 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 Li
3 (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
15 not 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 (2004). There is evidence that children under 5 years of age have reduced ADH
24 activity. A study by Pikkarainen and Raiha (1967) measured liver ADH activity using ethanol as
25 a substrate and found that 2-month-old fetal livers have -3-4% of adult ADH liver activity. ADH
26 activity in 4-5 month old fetuses is -10% of adult activity, and an infant's activity is -20% of
27 adult activity. ADH continues to increase in children with age and reaches a level that is within
28 adult ranges at 5 years of age. Adults were found to have great variation in ADH activity
29 (1,625-6,530/g liver wet weight or 2,030-5,430 mU/100 mg soluble protein). Smith et al. (1971)
30 also compared liver ADH activity in 56 fetuses (9-22 weeks gestation), 37 infants (premature to
31 <1 year old), and 129 adults (>20 years old) using ethanol as a substrate. ADH activity was 30%
32 of adult activity in fetuses and 50% of adult activity in infants. There is evidence that some
33 human infants are able to efficiently eliminate methanol at high exposure levels, however,
34 possibly via CAT (Tran et al., 2007).
March 2011 3-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 ADH3 exhibits little or no activity toward small alcohols, thus the previous discussion is
2 not relevant to the ontogeny of formaldehyde elimination (clearance). While such data on ADH3
3 activity does not exist, ADH3 mRNA is abundantly expressed in the mouse fetus (Ang et al.,
4 1996) and is detectible in human fetal tissues (third trimester), neonates and children (Estonius et
5 al.. 1996: Hines & McCarver. 2002).
6 As noted earlier in this section, folate-dependent reactions are important in the
7 metabolism of formate. Individuals who are commonly folate deficient include those who are
8 pregnant or lactating, have gastrointestinal (GI) disorders, have nutritionally inadequate diets, are
9 alcoholics, smoke, have psychiatric disorders, have pernicious anemia, or are taking folic acid
10 antagonist medications such as some antiepileptic drugs (2004; 1997). Groups which are known
11 to have increased incidence of folate deficiencies include Hispanic and African American
12 women, low-income elderly, and mentally ill elderly (2004). A polymorphism in methylene
13 tetrahydrofolate reductase reduces folate activity and is found in 21% of Hispanics in California
14 and 12% of Caucasians in the United States. Genetic variations in folic acid metabolic enzymes
15 and folate receptor activity are theoretical causes of folate deficiencies.
3.4. PHYSIOLOGICALLY BASED TOXICOKINETIC MODELS
16 In accordance with the needs of this human health risk assessment, particularly the
17 derivation of human health effect benchmarks from studies of the developmental effects of
18 methanol inhalation exposure in mice (J. M. Rogers, Mole, et al., 1993) and rats (NEDO, 1987)
19 mouse and rat models were developed to allow for the estimation of mouse and rat internal dose
20 metrics. A human model was developed to extrapolate those internal metrics to inhalation and
21 oral exposure concentrations that would result in the same internal dose in humans (human
22 equivalent concentrations [HECs] and human equivalent doses [HEDs]). The procedures used
23 for the development, calibration and use of these models are summarized in this section, with
24 further details provided in Appendix B, "Development, Calibration and Application of a
25 Methanol PBPK Model.
3.4.1. Model Requirements for EPA Purposes
3.4.1.1. MOA and Selection of a Dose Metric
26 Dose metrics closely associated with one or more key events that lead to the selected
27 critical effect are preferred for dose-response analyses compared to metrics not clearly correlated.
28 For instance, internal (e.g., blood, target tissue) measures of dose are preferred over external
29 measures of dose (e.g., atmospheric or drinking water concentrations), especially when, as with
March 2011 3-16 DRAFT - DO NOT CITE OR QUOTE
-------
1 methanol, blood methanol concentrations increase disproportionally with dose (J. M. Rogers,
2 Mole, et al., 1993). This is likely due to the saturable metabolism of methanol. In addition,
3 respiratory and GI absorption may vary between and within species. Mode of action (MOA)
4 considerations can also influence whether to model the parent compound with or without its
5 metabolites for selection of the most adequate dose metric.
6 As discussed in Section 4.3, developmental effects following methanol exposures have
7 been noted in both rats and mice (NEDO, 1987; Nelson et al., 1985; J. M. Rogers, Barbee, et al.,
8 1993; J. M. Rogers, Mole, et al., 1993), but are not as evident or clear in primate exposure
9 studies (L. S. Andrews et al., 1987: Burbacher, Grant, et al., 2004: Clary, 2003: Nelson et al.,
10 1985: J. M. Rogers, Barbee, et al., 1993: J. M. Rogers, Mole, et al., 1993). The report of the New
11 Energy Development Organization (NEDO, 1987) of Japan, which investigated developmental
12 effects of methanol in rats, indicated that there is a potential that developing rat brain weight is
13 reduced following maternal and neonatal exposures. These exposures included both in utero and
14 postnatal exposures. The methanol PBPK models developed for this assessment do not explicitly
15 describe these exposure routes. Mathematical modeling efforts have focused on the estimation of
16 human equivalent external exposures that would lead to an increase in internal blood levels of
17 methanol or its metabolites presumed to be associated with developmental effects as reported in
18 rats (NEDO, 1987) and mice (J. M. Rogers, Mole, et al., 1993).
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 (2004: Dorman et al.,
22 1995). The CERHR Expert Panel based their assessment of potential methanol toxicity on an
23 assessment of circulating blood levels (2004). While recent in vitro evidence indicates that
24 formaldehyde is more embryotoxic than methanol and formate (Harris et al., 2004: 2003), the
25 high reactivity of formaldehyde would limit its unbound and unaltered transport as free
26 formaldehyde from maternal to fetal blood (Thrasher & Kilburn, 2001), and the capacity for the
27 metabolism of methanol to formaldehyde is likely lower in the fetus and neonate versus adults
28 (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 to
30 the target tissue.
31 Given the reactivity of formaldehyde, models that predict levels of formaldehyde in the
32 blood are difficult to validate. However, production of formaldehyde or formate following
33 exposure to methanol can be estimated by summing the total amount of methanol cleared by
March 2011 3-17 DRAFT - DO NOT CITE OR QUOTE
-------
1 metabolic processes.10 This metric of formaldehyde or formate dose has limited value since it
2 ignores important processes that may differ between species, such as elimination (all routes) of
3 these two metabolites, but it can be roughly be equated to the total amount of metabolites
4 produced and may be the more relevant dose metric if formaldehyde is found to be the proximate
5 toxic moiety. Thus, both blood methanol and total metabolism metrics are considered to be
6 important components of the PBPK models. Dose metric selection and MOA issues are
7 discussed further in Sections 3.3, 4.6, and 4.8.
3.4.1.2. Criteria for the Development of Methanol 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 • Must simulate blood methanol concentrations and total methanol metabolism. Blood
15 methanol is the recommended dose metric for developmental effects, but total metabolism
16 may be a useful metric.
17 • Must be capable of simulating experimental blood methanol and total metabolism for the
18 inhalation route of exposure in mice and rats (a) and humans (b), and the oral route in rats (c)
19 and humans (d). These routes are important for determining dose metrics in the most
20 sensitive test species under the conditions of the toxicity study and in the relevant exposure
21 routes in humans.
22 • 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. Blood
24 methanol concentrations will depend strongly on ventilation rate, which varies significantly
25 between species.
26 • Must address the potential for saturable metabolism of methanol. Saturable metabolism has
27 the potential to bring nonlinearities into the exposure: tissue dose relationship.
28 • Model complexity should be consistent with modeling needs and limitations of the available
29 data. Model should adequately describe the biological mechanisms that determine the
30 internal dose metrics (blood methanol and total metabolism) to assure that it can be reliably
31 used to predict those metrics in exposure conditions and scenarios where data are lacking.
10 This assumption is more likely to be appropriate for formaldehyde than formate as formaldehyde is a direct
metabolite of methanol.
March 2011 3-18 DRAFT - DO NOT CITE OR QUOTE
-------
1 Compartments or processes should not be added that cannot be adequately characterized by
2 the available data.
3 Although the rat and mouse models are useful for the evaluation of the dose metrics
4 associated with methanol's developmental effects and the relevant toxicity studies, including
5 gestational exposures, no pregnancy-specific PBPK model exists for methanol, and inadequate
6 data exists for the development and validation of a fetal/gestational/conceptus compartment.
7 However, EPA determined that nonpregnancy models for the appropriate species and routes of
8 exposure could prove to be valuable because levels of methanol in NP, pregnant and fetal blood
9 are expected to be similar following the same oral or inhalation exposure. Pollack and Brouwer
10 (1996) determined that methanol distribution in rats and mice following repeated oral and i.v.
11 exposures up to day 20 of gestation is "virtually unaffected by pregnancy, with the possible
12 exception of the immediate perinatal period." The critical window for methanol induction of
13 cervical rib malformations in CD-I mice has been identified as occurring between GD6 and GD7
14 (7. M. Rogers, Barbee, et al., 1993; J. M. Rogers & Mole, 1997), a developmental period roughly
15 equivalent to week 3 of human development (Chernoff & Rogers, 2004). Methanol blood
16 kinetics measured during and after inhalation exposure in NP and pregnant mice on GD6-GD10
17 and GD6-GD15 (Dorman et al.. 1995: Perkins et al.. 1995a, 1996: J. M. Rogers. Mole, et al..
18 1993) are also similar. Further, the available data indicate that the maternal blood:fetal partition
19 coefficient is approximately 1 at dose levels most relevant to this assessment (Horton et al., 1992:
20 Ward etal., 1997). The same has been found in rat (Guerri & Sanchis, 1985: Zorzano & Herrera,
21 1989) and sheep (Brien et al., 1985: Cumming et al., 1984) studies of ethanol, a structurally
22 related chemical that also penetrates cellular membranes readily and distributes throughout total
23 body water. Consequently, fetal methanol concentrations are expected to be roughly equivalent
24 to that in the mother's blood. Thus, pharmacokinetics and blood dose metrics for NP mice and
25 humans are expected to provide reasonable approximations of pregnancy levels and fetal
26 exposure, particularly during early gestation, that improve upon default estimations from external
27 exposure concentrations.
3.4.2. Methanol PBPK Models
28 As has been discussed, methanol is well absorbed by both inhalation and oral routes and
29 is readily metabolized to formaldehyde, which is rapidly converted to formate in both rodents and
30 humans. As was discussed in Section 3.1, the enzymes responsible for metabolizing methanol
31 are different in rodents and humans. Several rat, mouse and human PBPK models which attempt
32 to account for these species differences have been published (Fisher et al., 2000: Horton et al.,
33 1992: Perkins etal., 1995a: Ward etal., 1997). In addition, a gestational model for a similar
March 2011 3-19 DRAFT - DO NOT CITE OR QUOTE
-------
1 water soluble compound, isopropanol, with the potential to be adapted to methanol
2 pharmacokinetics, was of interest (Clewell et al., 2001; Gentry et al., 2002; Gentry et al., 2003).
3 Three PK models (Bouchard etal.. 2001: Gentry etal.. 2003: Wardetal.. 1997) were identified
4 as potentially appropriate for use in animal-to-human extrapolation of methanol metabolic rates
5 and blood concentrations. An additional methanol PBPK model by Fisher et al. (2000) was
6 considered principally because it had an important feature - pulmonary compartmentalization
7 (see below for details) - worth adopting in the final model.
3.4.2.1. Ward et al. (1997)
8 The PBPK model of Ward et al. (1997) describes inhalation, oral and i.v. routes of
9 exposure and is parameterized for both NP and pregnant mice and rats (Table 3-8). The model
10 has not been parameterized for humans.
11 Respiratory uptake of methanol is described as a constant infusion into arterial blood at a
12 rate equal to the minute ventilation times the inhaled concentration and includes a parameter for
13 respiratory bioavailability, which for methanol is <100%. This simple approach is nonstandard
14 for volatile compounds but is expected to be appropriate for a compound like methanol, for
15 which there is little clearance from the blood via exhalation. Oral absorption is described as a
16 biphasic process, dependent on a rapid and a slow first-order rate constant. This is conceptually
17 similar to the isopropanol model discussed below (Clewell etal., 2001: Gentry et al., 2002),
18 which also employs slow and fast absorption processes but functionally separates them into
19 stomach and duodenal compartments.
20 Methanol elimination in the Ward et al. (1997) model is primarily via saturable hepatic
21 metabolism. The parameters describing this metabolism come from the literature, primarily
22 previous work by Ward and Pollack (1996) and Pollack et al. (1993). A first-order elimination of
23 methanol from the kidney compartment includes a lumped metabolic term that accounts for both
24 renal and pulmonary excretion.
25 The model adequately fits the experimental blood kinetics of methanol in rat and mice
26 and is therefore suitable for simulating blood dosimetry in the relevant test species and routes of
27 exposure (oral and i.v.). The Ward et al. (1997) model meets criteria 1, 2a, 2c, 3, 4, and 5. The
28 most significant limitation is the absence of parameters for the oral and inhalation routes in the
29 human. A modified version of this model that includes human parameters and a standard PBPK
30 lung compartment might be suitable for the purposes of this assessment.
3.4.2.2. Bouchard et al. (2001)
31 The Bouchard et al. (2001) model is not actually a PBPK model but is an elaborate
32 classical PK model, since the transfer rates are not determined from blood flows, ventilation,
March 2011 3-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 partition coefficients, and the like. The Bouchard et al. (2001) model uses a single compartment
2 for methanol: a central compartment represented by a volume of distribution where the
3 concentration is assumed to equal that in blood. The model was developed for inhalation and i.v.
4 kinetics only. Methanol is primarily eliminated via saturable metabolism. The model adequately
5 simulates blood kinetics in NP rats and humans following inhalation exposure and in NP rats
6 following i.v. exposure; there is no description for oral absorption. Because methanol distributes
7 with total body water (Horton etal., 1992; Ward etal., 1997), this simple model structure is
8 sufficient for predicting blood concentrations of methanol following inhalation and i.v. dosing.
9 The Bouchard et al. (2001) model has the advantage of simplicity, reflecting the
10 minimum number of compartments necessary for representing blood methanol pharmacokinetics.
11 Because volume of distribution can be easily and directly estimated for water-soluble
12 compounds like methanol or fit directly to experimental kinetics data, concern over the
13 scalability of this parameter is absent. The model has been parameterized for a required human
14 exposure route, inhalation (Table 3-8). The model meets criteria 1, 2b, 3, 4, and 5 described in
15 Section 3.4.1.2. However, the Bouchard model has specific and significant limitations. The
16 model has neither been parameterized for the mouse, a test species of concern (Table 3-8), nor for
17 the oral route in humans. As such, the model cannot be used to conduct the necessary
18 interspecies extrapolation.
3.4.2.3. Ward et al. (1997)
19 The PBPK model of Ward et al. (1997) describes inhalation, oral and i.v. routes of
20 exposure and is parameterized for both NP and pregnant mice and rats (Table 3-8). The model
21 has not been parameterized for humans.
22 Respiratory uptake of methanol is described as a constant infusion into arterial blood at a
23 rate equal to the minute ventilation times the inhaled concentration and includes a parameter for
24 respiratory bioavailability, which for methanol is <100%. This simple approach is nonstandard
25 for volatile compounds but is expected to be appropriate for a compound like methanol, for
26 which there is little clearance from the blood via exhalation. Oral absorption is described as a
27 biphasic process, dependent on a rapid and a slow first-order rate constant. This is conceptually
28 similar to the isopropanol model discussed below (Clewell et al., 2001; Gentry et al., 2002),
29 which also employs slow and fast absorption processes but functionally separates them into
30 stomach and duodenal compartments.
31 Methanol elimination in the Ward et al. (1997) model is primarily via saturable hepatic
32 metabolism. The parameters describing this metabolism come from the literature, primarily
33 previous work by Ward and Pollack (1996) and Pollack et al. (1993). A first-order elimination of
March 2011 3-21 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
J
4
5
6
7
methanol from the kidney compartment includes a lumped metabolic term that accounts for both
renal and pulmonary excretion.
The model adequately fits the experimental blood kinetics of methanol in rat and mice
and is therefore suitable for simulating blood dosimetry in the relevant test species and routes of
exposure (oral and i.v.). The Ward et al. (1997) model meets criteria 1, 2a, 2c, 3, 4, and 5. The
most significant limitation is the absence of parameters for the oral and inhalation routes in the
human. A modified version of this model that includes human parameters and a standard PBPK
lung compartment might be suitable for the purposes of this assessment.
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
-
9
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).
3.4.2.4. Gentry et al. and Clewell et al.
The rat and human models described in three papers by Gentry et al. (2002; 2003) and
Clewell et al. (2001) 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 rat and humans) and for several routes of exposure.
Therefore the isopropanol model was not considered further.
March 2011
3-22 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, methanol pharmacokinetics
3 and blood dose metrics for NP laboratory animals and humans are expected to improve upon
4 default extrapolations from external exposures as estimates of fetal exposure during early
5 gestation. The same level of confidence cannot be placed on the whole-body rate of metabolism,
6 in particular as a surrogate for formaldehyde dose. Because of formaldehyde's reactivity and the
7 limited fetal metabolic (ADH) activity (see Sections 3.3 and 4.10.1), fetal formaldehyde
8 concentration increases (from methanol) will probably not equal maternal increases in
9 formaldehyde concentration. But since there is no model that explicitly describes formaldehyde
10 concentration in the adult, let alone the fetus, the metabolism metric is the closest one can come
11 to predicting fetal formaldehyde dose. This metric is expected to be a better predictor of
12 formaldehyde dose than applied methanol dose or even methanol blood levels, which do not
13 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 removal
28 is described by two pathways, one to urinary elimination and one via metabolism to CC>2 (which
29 is exhaled). All of these metabolic and elimination steps are described as first-order processes,
30 but the explicit descriptions of formaldehyde and formate kinetics significantly distinguish the
31 model of Bouchard et al. (2001) from that of Ward et al. (1997), which only describes methanol.
32 There are two other important distinctions between the Ward et al. (1997) and Bouchard
33 et al. (2001) models. The former is currently capable of simulating blood data for all exposure
34 routes in mice but not humans, while the latter is capable of simulating human inhalation route
March 2011 3-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 blood pharmacokinetics but not those in mice. The Ward et al. (1997) model has more
2 compartments than is necessary to adequately represent methanol disposition but has been fit to
3 PK data in pregnant and NP mice for all routes of exposure (i.v., oral, and inhalation). The Ward
4 et al. (1997) model has also been fit to i.v. and oral route PK data in rats. Based primarily on the
5 extensive amount of fitting that has already been demonstrated for this model, it was determined
6 that a modified Ward et al. (1997) model, with the addition of a lung compartment as described
7 by Fisher et al. (2000), should be used for the purposes of this assessment. See Appendix B for a
8 more complete discussion of the selected modeling approach and modeling considerations.
3.4.3.1. Available PK Data
9 Although limited human data are available, several studies exist that contain PK and
10 metabolic data in mice, rats, and nonhuman primates for model parameterization. Table 3-9
11 contains references that were used to verify the model fits as reported in Ward et al. (1997).
3.4.3.2. Model Structure
12 A model was developed which includes compartments for alveolar air/blood methanol
13 exchange, liver, fat, bladder (human simulations) and the rest of the body (Figure 3-4). This
14 model is a revision of the model reported by Ward et al. (1997), reflecting significant
15 simplifications (removal of compartments for placenta, embryo/fetus, and extraembrionic fluid)
16 and three elaborations (addition of an intestine lumen compartment to the existing stomach
17 lumen compartment, use of a saturable rate of absorption from the stomach (but not intestine),
18 and addition of a bladder compartment which impacts simulations for human urinary excretion),
19 while maintaining the ability to describe methanol blood kinetics in mice, rats, and humans. A
20 fat compartment was included because it is the only tissue with a tissue:blood partitioning
21 coefficient appreciably different than 1, and the liver is included because it is the primary site of
22 metabolism. A bladder compartment was also added for use in simulating human urinary
23 excretion to capture the difference in kinetics between changes in blood methanol concentration
24 and urinary methanol concentration; the difference in model fit to human urinary data with vs.
25 without the bladder compartment is shown in Figure 3-11. The model code describes inhalation,
26 oral, and i.v. dose routes, and data exist (Table 3-9) that were used to fit parameters and evaluate
27 model predictions for all three of those routes in both mice and rats. In humans, inhalation
28 exposure data were available for model calibration and validation but not oral or i.v. data.
29 However, oral exposures were simulated in humans assuming a continuous, zero-order ingestion
30 rate, thereby obviating the need for oral uptake parameters.
March 2011 3-24 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-9. Key methanol kinetic studies for model validation
Reference
Batterman & Franzblau
(1997)
Batterman et al., (1998)
Burbacher, Shen et al.
(2004): Burbacher, Grant
et al. (2004)
Osterloh et al. (1996):
Chuwers et al. (1995):
D'Alessandro etal. (1994)
Medinsky et al. (1997):
Dorman et al., (1994)
Gonzalez-Quevedo et al.,
(2002)
Horton et al. (1992)
Perkins et al., (1995a,
1995b. 1996)
Pollack and Brouwer
(1996): Pollack et al.,
(1993)
Rogers and Mole, (1997):
Rogers etal. (1993);
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
(6hr)
1,000-20,000
(8hr)
1,000-20,000
(8hr)
1,000-15,000
(7 hr, 10 days)
78-231 (8hr)
Oral/derm
al/ff
Dermal
IP: 2
mg/kg-day,
2wk
Oral:
100-2,500
mg/kg
Oral:
2,500 mg/k
g
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,
GD14&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
figuresA
Figure 1
Figure 1,
Osterloh et
al. (1996)
Figure 7
Figures 2,
3, 6, 7, 8
aData obtained from the reported figure
1 PK data from intravenous exposures were used to test or further refine the parameters for
2 methanol metabolism in mice and rats. Monkey data were evaluated for insight into primate
3 kinetics. Data from Batterman et al., (19981 Osterloh et al. (19961 and Sedivec et al. (1981)
4 were used to estimate (fit) model parameters for humans subsequent to the addition of the
5 bladder compartment. The fact that optimized human parameters were similar to those predicted
6 in monkeys was important to the validation process (Bouchard et al., 2001)(see section 3.4.7 and
7 Appendix B). Blood levels of methanol have been reported following i.v., oral, and inhalation
8 exposure in rats and mice and inhalation exposure in nonhuman primates and humans.
9 The metabolism of methanol was represented in mice, rats, and humans by specifying
10 separate rate constants for the species-specific enzymes: two saturable processes for mice and
March 2011
3-25 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
J
4
5
6
7
11
Sprague-Dawley (SD) rats and one for F344 rats and humans. The requirement for two
saturable processes in the mouse and SD rat models may reflect saturation of CAT and ADH1.
Simulated methanol elimination by these metabolic processes is not linked in the PBPK model to
production of formaldehyde or formate. For the PBPK model, methanol metabolism is simply
another route of methanol elimination. Metabolism of formaldehyde (to formate) is not
explicitly simulated by the model, and this model tracks neither formate nor formaldehyde.
Since the metabolic conversion of formaldehyde to formate is rapid (<1 minute) in all species
(Kavet & Nauss, 1990), the rate of methanol metabolism may approximate a formate production
rate, though this has not been verified.
Inhalation
Exposure
••w.
Bladder
KBLJthuman
Urine |\
KLL or
Vma
only)
K — -
Venous Blood
r
^
Alveolar Air
Lung Blood
Body
Fat
Liver
^^ Exhaled Air
Arterial Blood
Vmax2
Km2
K AS or Vm AS, KM AS
Metabolites
Kfec
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, Km, Vmax2, and Km2, Michaelis-Menten rate constants for high
affinity/low capacity and low affinity/high capacity metabolism of MeOH;
KLL, alternate first-order rate constant; KBL, rate constant for urinary
excretion from bladder. Both metabolic pathways were used to describe
MeOH metabolism in the mouse and SD rat, while a single pathway
describes metabolism in the F344 rat and human.
11 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.
March 2011
3-26 DRAFT - DO NOT CITE OR QUOTE
-------
1 The primary purpose of this assessment is for the determination of noncancer risk
2 associated with exposures that increase the body burden of methanol or its metabolites (e.g.,
3 formate, formaldehyde) above prevailing, endogenous levels. Thus, the focus of model
4 development was on obtaining predictions of increased body burdens over background following
5 external exposures. To accomplish this, the PBPK models used in this assessment do not account
6 for background levels of methanol, formaldehyde or formate. In addition, background levels
7 were subtracted from the reported data before use in model fitting or validation (in many cases
8 the published data already have background subtracted by study authors). This approach for
9 dealing with endogenous background levels of methanol and its metabolites assumes that:
10 (1) endogenous levels do not contribute significantly to the adverse effects of methanol or its
11 metabolites; and (2) the exclusion of endogenous levels does not significantly alter PBPK model
12 predictions. There is uncertainty associated with these assumptions. Human data are not
13 available to evaluate whether there is a relationship between background levels of methanol or its
14 metabolites. To test the assumption that the exclusion of endogenous background levels does not
15 significantly alter PBPK model predictions, EPA performed the following alternative analysis
16 using models that incorporate background levels of methanol and its metabolites.
17 3.4.3.2.1. Alternative modeling approach - incorporation of background. If background
18 methanol levels are high enough compared to those which induce metabolic saturation, they may
19 have a significant impact on parameter estimation and hence internal dose predictions. To gauge
20 the impact of background levels on PBPK model predictions of exposure-induced changes in
21 internal doses, alternate (test) versions of the rat and human PBPK models were created which
22 incorporate a zero-order liver infusion term for methanol designed to approximate reported rat
23 and human background levels. Internal dose estimates for various exposure levels obtained from
24 the PBPK models that exclude background up front could then be compared with those from
25 models for which background levels were modeled, but then subtracted for benchmark dose
26 (BMD) modeling. For example, when background levels are included in the PBPK model and
27 the metric is blood AUC, BMD analysis used the PBPK-predicted difference, AUC(exposed rats)
28 - AUC(control rats), as the dose metric. After obtaining an internal dose point of departure
29 (POD) at a specific effect level for the rat with that metric, the human equivalent internal dose
30 was taken to be POD + AUC(human background). In short the level of effect (above
31 background) was correlated with the internal dose above background in the animal, then the
32 human background internal dose was added to the POD obtained with that metric to yield an
33 estimate of the dose when humans would have the same level of effect.
34 The two PBPK modeling approaches (i.e., including or excluding background levels in
35 the PBPK model) did not differ significantly (<1%) with respect to their internal dose point of
March 2011 3-27 DRAFT - DO NOT CITE OR QUOTE
-------
1 departure (POD, level above background) estimates from the principal rat noncancer studies.
2 HEC and HED estimates from the principal rat noncancer studies using the human PBPK model
3 with background included were only about 14% lower than those estimated using the human
4 PBPK model with background excluded. Because the more complex PBPK modeling required to
5 include background levels was estimated to have a minimal impact on dose extrapolations, the
6 use of simpler methanol models that do not incorporate background levels is considered adequate
7 for the purposes of this assessment.
3.4.3.3. Model Parameters
8 The EPA methanol model uses a consistent set of physiological parameters obtained
9 predominantly from the open literature (Table 3-10); the Ward et al. (1997) model employed a
10 number of data-set specific parameters.12 Parameters for blood flow, ventilation, and metabolic
11 capacity were scaled as a function of body weight raised to the 0.75 power, according to the
12 methods of Ramsey and Andersen (1984).
12 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.
March 2011 3-28 DRAFT - DO NOT CITE OR QUOTE
-------
Table 3-10. Parameters used in the mouse, rat and human PBPK models
Body weight (kg)
Mouse
0.03"
Rat
SD | F344
0.275*
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 et al. (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. (2000M
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 et al. (1997)
Calculated
Fitted
Oral absorption
VmASC
(mg/hr/kg075)
KMASC (mg/kg)
KSI (hr'1)
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/KMAS = the first-order constant
of Sultatos et al.
March 2011
3-29 DRAFT - DO NOT CITE OR QUOTE
-------
Mouse
Rat
SD | F344
Human
Source
Partition coefficients
LiverBlood
FatBlood
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 & 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 ) of 25.4 was used (average
from 1995a). The QPC used to fit the human data was obtained from U.S. EPA (2000b). This QPC was somewhat
higher than calculated from Brown et al. (1997) (-13 L/hr/kg°75)
eVmax, Km, and Vmax2, Km2 represent the two saturable metabolic elimination processes assumed to occur solely in
the liver. The Vmax used in the model = VmaxC (mg/kg° 75-hr) XBW°75. 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)
gHuman 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
5 inhalation studies of Rogers and Mole (1997) (Figure 3-5), Perkins et al. (1995a, 1995b), and
March 2011
3-30 DRAFT - DO NOT CITE OR QUOTE
-------
1 Rogers et al. (1993), as well as in NP and early gestation (GD8) mice of Dorman et al. (1995)
2 (Figure 3-6). Parameter values used in the calibrated model are given in Table 3-10.
3 The mouse model was also calibrated for the oral route by fitting all but one of the rate
4 constants for oral uptake of methanol to the oral-route blood methanol kinetics of Ward et al.
5 (1997; 1995). The best model fit to the mouse oral route blood methanol PK data was obtained
6 using a two-compartment GI tract model, as depicted in Figure 3-4. Because the oral data in rats
7 led to the conclusion that a saturable rate of uptake from the stomach lumen was necessary (see
8 section 3.4.5), the same equation was used for uptake in the mouse. But attempts to identify the
9 uptake saturation constant, KMASC, from the mouse data were unsuccessful; therefore KMASC
10 for the mouse was set equal to the value obtained for rats. Adjusting the other mouse oral uptake
11 parameters gave an adequate fit to those data. This calibration allows inhalation to oral dose-
12 route extrapolations in the mouse, which can then be extrapolated to identify human oral route
13 exposures equivalent to mouse inhalation exposures (if equivalent human exposures exist).
10,000
1,000
I
o
Oj
100
10
— 1000 ppm sim
— 2000 ppm sim
— 5000 ppm sim
— 7500 ppm sim
— 10000 ppm sirn
15000 pprn sim
A 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 pprn data (GD7)
n 15000 pprn data (GD6 & 10)
12 15
Time (hr)
IE:
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. (1993). 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. (1993)
March 2011
3-31 DRAFT - DO NOT CITE OR QUOTE
-------
8,000-
7,000-
J 6,000-
£ 5,000-
O 4,000-
w
g
3,000-
5 2,000 H
1,000 :
0-
+ 2,500 ppm IMP mouse
A 5,000 ppm NP mouse
o 10,000 ppm NP mouse
- 6 hr, 15,000 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. (I995a).
1 The parameterization of methanol metabolism (high-and low-affinity metabolic
2 pathways) was also verified by simulation of datasets describing the pharmacokinetics of
3 methanol following i.v. administration. The results of this calibration of the methanol PBPK
4 model are described in Appendix B and were generally consistent with both the available
5 inhalation and oral-route data. Up to 20 hours post exposure, blood methanol kinetics appears
6 similar for NP and pregnant mice. However, some data suggests that clearance in GDIS mice is
7 slower than in NP and earlier in gestation (GD10 and less), particularly beyond 20 hours post
8 exposure (see the 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 single
10 i.v. dose of 2,500 mg/kg, but were available for GDIS mice following administration of a broader
11 range of doses: 100, 500, and 2,500 mg/kg. The i.v. maternal PK data in GDIS mice appeared to
12 show an unexpected dose-dependent nonlinearity in initial blood concentrations. Before
March 2011
3-32 DRAFT - DO NOT CITE OR QUOTE
-------
1 discussing the nonlinearity, it is first noted that data values used here were obtained from a
2 computational "command file" provided by Ward et al. (1997). These values appear to be
3 consistent with the plots in their publication but are inconsistent with some of the values in their
4 Table 6 (Ward et al., 1997). In particular, the initial maternal blood concentration (i.e., the Cmax)
5 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
6 published table. The corresponding data point in their Figure 5 A is distinctly centered above
7 4,000 mg/L (digitizing yields 4,213 mg/L), and so must be 4,250 rather than 3,251 mg/L.
8 Therefore the data values listed in the command file were used in the subsequent analysis, rather
9 than those in the published table.
10 After i.v. dosing the ratio of the administered doses to the first concentrations measured
11 by Ward et al. (1997) (5-minute time points) were 0.588 L/kg, 0.585 L/kg, and 0.397 L/kg at
12 doses of 2,500, 500, and 100 mg/kg, respectively. The discrepancy between the first two values
13 and the third value suggests either a dose dependence in the Vd or some source of experimental
14 variability.13 It may be that Vd, which is not impacted by any other PBPK parameters and is only
15 determined by the biochemical partitioning properties of methanol, is 1.5-fold lower at 100
16 mg/kg than at the higher concentrations, while the Vd at 500 and 2,500 mg/kg are exactly as
17 predicted by the PBPK model without adjustment. However, it was found that the PBPK model,
18 obtained with measured partition coefficients and otherwise calibrated to inhalation data, could
19 adequately fit the data at the nominal dose of 100 mg/kg without other parameter adjustment
20 simply by simulating a dose of 200 mg/kg, as shown in Appendix B, Figure B-5. The fact that
21 the alternate dose (200 mg/kg) differs by a factor of 2 from the nominal dose suggests that the
22 data could also be the result of a simple dilution error in dose preparation. If the first two of the
23 dose/concentration values were not virtually identical (0.588 and 0.585 L/kg), but instead the
24 500 mg/kg value was more intermediate between those for 2,500 and 100 mg/kg, then a regular
25 dose dependence in Vd would seem more likely. However, based on these values, the U.S. EPA
26 has concluded that the apparent dose dependency is probably the result of a dosing error and
27 therefore, that dose-dependent parameter changes (e.g., in the partition coefficients) should not
28 be introduced in an attempt to otherwise better fit these data.
29 Further, the nominal "nonlinearity" between the maternal blood and conceptus shown in
30 Figure 8 of Ward et al. (1997) is the result of those data being plotted on a log-y/linear-x scale.
31 Replotting the data from Tables 5 and 6 (using the value of 4,250 mg/L from the command file as
32 the GDIS maternal Cmax for the 2,500 mg/kg) shows the results to be linear, especially in the
13 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.
March 2011 3-33 DRAFT - DO NOT CITE OR QUOTE
-------
1 low-dose region which is of the most concern (Figure 3-7). Therefore, the current model uses a
2 consistent set of parameters that are not varied by dose and fit the 2,500 and 500 mg/kg i.v. data
3 adequately, although they do not fit the 100 mg/kg i.v. data unless, as noted above, a presumed
4 i.v. dose of 200 mg/kg is employed. With that exception, both the single set of parameters used
5 herein and the assumption that maternal blood methanol is a good metric of fetal exposure are
6 well supported by the data.
10000
re
f 1000
0
« 100
Q. ,
0)
O
0
0 10
(
A
g, - '" " D
/
k
>
D 1000 2000 3000
- Y~
4 rat
n mouse
- y = x
4000 5000
Maternal ADC (ug/mL-day)
_ 3500
•c 3000
-
-I 2500
Ul
^ 2000
< 1500
Q. 1000
0)
o
o 500
O
0
i
B
/'
n/
U
0 1000 2000 3000
D
* rat
• mouse
- y = x
4000 5000
Maternal ADC (ug/mL-day)
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).
March 2011
3-34 DRAFT - DO NOT CITE OR QUOTE
-------
1 To summarize the mouse model calibration: using the single set of parameters listed for
2 the mouse in Table 3-10, the PBPK model has been shown to adequately fit or reproduce
3 methanol PK data from a variety of laboratories and publications, including both NP mice and
4 pregnant mice up to GD10. Two saturable metabolic pathways are thus described by the model
5 and supported by the data. Also, it is thereby demonstrated that a model based on NP mouse
6 physiology adequately describes (predicts) dosimetry in the pregnant mouse dam through GD10.
7 Finally, as illustrated in Figure 3-7b, methanol PK in the conceptus and dam of both mice
8 (including lower doses at GDIS) and rats (GDI4 and GD20) are virtually identical, except for the
9 very highest doses in mice. Thus the existing model appears to be adequate for predicting
10 internal methanol doses, including fetal exposures, at bioassay conditions.
11 An evaluation of the importance of selected parameters on mouse model estimates of
12 blood methanol AUC was performed by conducting a sensitivity analysis using the subroutines
13 within acslXtreme v2.3 (Aegis Technologies, Huntsville, Alabama). The analysis was conducted
14 by measuring the change in model output corresponding to a 1% change in a given model
15 parameter when all other parameters were held fixed. Sensitivity analyses were conducted for
16 the inhalation and oral routes. The inhalation route analysis was conducted under the exposure
17 conditions of Rogers and Mole (1997) and Rogers et al. (1993): 7-hour inhalation exposures at
18 the no-observed-effect level (NOEL) concentration of 1,000 ppm. The oral route sensitivity
19 analysis was conducted for an oral dose of 1,000 mg/kg.
20 The parameters with the largest sensitivity coefficients for the inhalation route at
21 1,000 ppm (absolute values >1) were pulmonary ventilation scaling coefficient (QPC) and
22 maximum velocity of the high-affinity/low-capacity pathway (VmaxC). The sensitivity coefficient
23 for QPC increases during the exposure period as metabolism begins to saturate. Following oral
24 exposure, mouse blood methanol AUC was sensitive to the rate constants for oral uptake. Blood
25 AUC was most sensitive to the maximum and saturation rate constants for uptake from the
26 stomach (VmASC and KMASC). The sensitivity coefficient for VmASC decreased during the
27 first hours after exposure from 1 to less than 0.1 at the end of exposure. Blood methanol AUC
28 was also modestly sensitive to first-order uptake from the intestine (KAI), and first-order transfer
29 between stomach and intestine (KSI), the rate constants for uptake from the intestine and transfer
30 rates between compartments, respectively. For a more complete description of this sensitivity
31 analysis for the mouse methanol PBPK model see Appendix B.
3.4.5. Rat Model Calibration
32 The rat model was calibrated to fit data from i.v., inhalation, and oral exposures in rats,
33 using data provided in the command file of Ward et al. (1997) and obtained from figures in
March 2011 3-35 DRAFT - DO NOT CITE OR QUOTE
-------
1 Horton et al. (1992) using Digitizlt. Holding other parameters constant, the rat PBPK model was
2 initially calibrated against the entire set of i.v.-route blood PK data (Figure 3-8) by fitting
3 Michaelis-Menten constants for one high-affmity/low-capacity and one low-affmity/high-
4 capacity enzyme to both the Ward et al. (1997) data for Sprague-Dawley (SD) rats and the Horton
5 et al. (1992) data for Fischer 344 (F344) rats, assuming that any difference between the two data
6 sets (100 mg/kg data) were from experimental variability and that a single set of parameters
7 could be fit to data for both strains of rat. However when the resulting parameters were then
8 used to simulate the F344 inhalation uptake data of Horton et al. (with the fractional absorption
9 for inhalation, FRACIN, adjusted to fit those data), it was found that the clearance rate predicted
10 (decline in blood concentrations) after the end of inhalation exposure was much more rapid than
11 shown by the data. More careful examination of the i.v. data then revealed that there too the
12 clearance for F344 rats was slower than for SD rats, and that the metabolic parameters obtained
13 from fitting the combined i.v. data best represented the SD rat data. It was concluded that the
14 combined data set indicated a true strain difference in metabolic parameters. The metabolic
15 parameters for SD rats were then obtained by fitting only the Ward et al. (1997) i.v. data (both
16 doses).
17 The 100 mg/kg i.v. data of Horton et al. (1992) were combined with their inhalation data
18 and a simultaneous optimization of the metabolic parameters and FRACIN for F344 rats was
19 attempted over that data set. For this data set, however, the optimization either converged with
20 the metabolic Vmax for the high affinity (low Km) pathway at zero, or with that Km value
21 increasing to be statistically indistinguishable from the high Km value. Therefore the Vmax for
22 the high affinity pathway was allowed to be zero, the Km for that pathway was not estimated, and
23 only a single Vmax and low affinity (high Km) were fit to those data, with a simultaneous
24 identification of FRACIN. Since there are no inhalation data for SD rats, this value of FRACIN
25 was assumed to apply for both strains. The optimized parameters for both strains of rats are
26 given in Table 3-10.
27 When the model was calibrated using the available inhalation and i.v. data for F344 rats
28 (Horton etal., 1992), a low fractional absorption of 20% was optimized to best fit the data, vs.
29 66.5% for the mouse. This lower fractional absorption is consistent with values presented by
30 Perkins et al. (1995a), who also found that the fractional absorption of methanol from inhalation
31 studies was lower in rats than in mice.
March 2011 3-36 DRAFT - DO NOT CITE OR QUOTE
-------
5000
- 4000
0)
2500 mg/kg IV in rats
16 24 32 40 48
Time (hr)
O)
O
o>
O
£
m
150
120 -
90 -
60 -
100 mg/kg IV in rats
*
A
Ward '97 data
Morton '92 data
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
z
O
o
o
a 2000 ppm data
+ 1200 ppm data
A 200 ppm data
— 2000 ppm simulation
— 1200 ppm simulation
— 200 ppm simulation
— 2000 ppm, 7-hr simulation
0 5 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 hour, a 7-hour simulated exposure is also shown for
comparison.
Source: Horton et al. (1992).
March 2011
3-37 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 2,500 mg/kg data with that model structure
6 using linear absorption and inter-compartment transfer rates. In particular the shorter-time data
7 for 2,500 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 2,500 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 absorption from the stomach, it was also found that
15 the 2,500 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 of Figure 3-10) was considerably
20 improved with an almost identical (excellent) fit to the 100 mg/kg data (saturable curve can be
21 distinguished from the linear curve just after the peak is reached in the lower panel of
22 Figure 3-10). For the purpose of scaling across individuals, strains, and species, the Km for
23 absorption from the stomach (KMAS) was assumed to scale in proportion to the stomach
24 (lumen) volume; i.e., with BW1. The Vmax (VmAS) was assumed to scale as BW°'75, with the
25 result that for low doses the effective linear rate constant (VmAS/KMAS) scales as BW"0'25,
26 which is a standard assumption for linear rates. Since the quantity on which the rate depends is
27 the total amount in the stomach (mg methanol), the resulting scaling constant for the Km,
28 KMASC, conveniently has units of mg/kg BW; i.e., the standard units for oral dosing.
March 2011 3-38 DRAFT - DO NOT CITE OR QUOTE
-------
2,700
2,100-
E 1,800-
I 1,500 -
a) 1,200-
•o 900-
J 600
CD
300-
0-
O
a
y -0-
— 2500 mg/kg linear uptake simulation
2500 mg/kg saturable uptake simulation
O 2500 mg/kg data
— 100 mg/kg linear uptake simulation
100 mg/kg saturable uptake simulation
n 100 mg/kg data
100
80-
60-
40
o
o
12
18
Time (hr)
24
30
36
— 100 mg/kg linear uptake simulation
100 mg/kg saturable uptake simulation
n 100 mg/kg data
4
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
1 The mouse model was scaled to humans by setting either a standard human body weight
2 (70 kg) or study-specific body weights and using human tissue compartment volumes and blood
3 flows, and then calibrated to fit the human inhalation exposure data available from the open
March 2011
3-39 DRAFT - DO NOT CITE OR QUOTE
-------
1 literature, which comprised data from four publications (Ernstgard et al., 2005): (Batterman et
2 al.. 1998: Osterloh et al.. 1996: Sedivec et al.. 1981).
3 Since the human data included time-course data for urinary elimination, a first-order rate
4 of loss of methanol from the blood (Kl) was used to provide an estimate of methanol elimination
5 to the bladder compartment in humans, and the rate of elimination from that compartment then
6 characterized by a second constant (KBL). Note that the total amount eliminated by this route
7 depends only on Kl, while KBL affects the rate at which the material cleared from the blood then
8 appears in the urine. Inhalation-route urinary methanol kinetic data described by Sedivec et al.
9 (1981) (Figure 3-11) was used in the model calibration to inform this rate constant. Without use
10 of the bladder compartment and rate constant, the fit of the model predictions to the data in
11 Figure 3-11 is quite poor (middle panel), and a statistical test on the improvement of fit obtained
12 by introducing the additional parameter (KBL) is significant (p < 0.0001). Conversion between
13 the PBPK-model-predicted rate of urinary excretion (mg/hours) or cumulative urinary excretion
14 (mg) and the urine methanol concentration data reported by the authors was achieved by
15 assuming 0.5 mL/hours/kg body weight total urinary output (Horton et al., 1992). The resulting
16 values of K1C and KBL, shown in Table 3-10, differ somewhat depending on whether first-order
17 or saturable liver metabolism is used. These are only calibrated against a small dataset and
18 should be considered an estimate. Urinary elimination is a minor route of methanol clearance
19 with little impact on blood methanol kinetics. However urine concentration is an indirect
20 indicator of the time-course in the blood and hence including this term in the model is useful in
21 overall model calibration.
22 Although the high doses used in the mouse studies clearly warrant the use of a second
23 metabolic pathway with a high Km, the human exposure data all represent lower concentrations
24 and may not require or allow for accurate calibration of a second metabolic pathway. Horton et
25 al. (1992) employed two sets of metabolic rate constants to describe human methanol disposition,
26 similar to the description used for rats and mice, but in vitro studies using monkey tissues with
27 nonmethanol substrates were used as justification for this approach. Although Bouchard et al.
28 (2001) described their metabolism using Michaelis-Menten metabolism, Starr and Festa (2003)
29 reduced that to an effective first-order equation and showed adequate fits. Perkins et al. (1995a)
30 estimated a Km of 320 + 1273 mg/L (mean + S.E.) by fitting a one-compartment model to data
31 from a single estimated oral dose. In addition to the extremely high standard error, the large
32 standard error for the associated Vmax (93 + 87 mg/kg/hours) indicates that the set of Michaelis-
33 Menten constants was not uniquely identifiable using this data. Other Michaelis-Menten
34 constants have been used to describe methanol metabolism in various models for primates
35 (Table 3-11).
March 2011 3-40 DRAFT - DO NOT CITE OR QUOTE
-------
Sedivecetal., 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
10
9
I 4
g 3
§ 2
U
1 -
Sedivec et a\., urine
concentration with
no bladder compartment
• 231 ppm
i 157 ppm
O 78 ppm
/ O
12 16
Time(h)
20
24
2.5 -
Sedivec et al.,
total urinary excretion
12 16
Time (h)
20
24
Figure 3-11. Urinary methanol elimination concentration (upper panels) and
cumulative amount (lower panel) following inhalation exposures to methanol
in human volunteers. Middle panel shows that without a bladder
compartment the shape of the urine time-course is quite discrepant from the
data. 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).
March 2011
3-41 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 + 8.7a'b
Reference
Jacobsenetal.. (1988)
Noker et al., (1980)
Makar et al., (1968)
Eells et al., (19831
Horton et al. (1992)
Fisher et al., (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
aThe 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 et al. Q995b).
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.
1 To estimate both 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 (Table 3-12). The metabolic (first-order or saturable) and urinary elimination constants
4 were numerically fit to the human datasets, while holding the value for FRACIN at 0.8655
5 (estimated from the results of Sedivec et al. [(1981)]) and holding the ventilation rate constant at
6 16.5 L/hours/kga75 and QPC at 24 L/hours/kg0'75 (values used by EPA [2000d] for modeling the
7 inhalation-route kinetics of vinyl chloride). Other human-specific physiological parameters were
8 used, as reported in Table 3-10. Final fitted parameters that have been used in the saturable
March 2011
3-42 DRAFT - DO NOT CITE OR QUOTE
-------
model are given in Table 3-10. The resulting fits of two different possible parameterizations,
first-order ["linear"] (dashed lines) or optimized Km/Vmax (solid lines), are shown in Figures 3-11
and 3-12.
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 rate, a
March 2011
3-43 DRAFT - DO NOT CITE OR QUOTE
-------
1 likelihood ratio test was performed.14 The hypothesis that one metabolic description is better
2 than another is calculated using the likelihood functions evaluated at the maximum likelihood
3 estimates. Since the parameters are optimized in the model using the maximum log likelihood
4 function (LLF), the resultant LLF is used for the statistical comparison of the models. The
5 equation states that two times the log of the likelihood ratio follows a chi square (%2) distribution
6 with r degrees of freedom:
7 - 2[log(/l(model l) / l(model 2))] = - 2[log l(model l) - log l(model 2)] = x\
8 The likelihood ratio test states that if the two times the difference between the maximum
9 LLFs of the two different descriptions of metabolism is greater than the $ distribution then the
10 model fit 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
19 dose candidate points of departure (PODs), for example the BMDLio for the inhalation-induced
20 brain-weight changes from NEDO (1987) with methanol blood AUC as the metric, is 90.9 mg-
21 hr/L, which corresponds to an average blood concentration of 3.8 mg/L. Therefore, the
22 Michaelis-Menten metabolism rate equation appears to be sufficiently supported by the existing
23 data with 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
X2r (99.95%
confidence)1"
12.22
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.
14 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 etal.. 1990).
March 2011 3-44 DRAFT - DO NOT CITE OR QUOTE
-------
1 While the use of Michaelis-Menten kinetics might allow predictions across a wide
2 exposure range (into the nonlinear region), extrapolation above 1,000 ppm is not suggested since
3 the highest human exposure data are for 800 ppm. Extrapolation to higher concentrations is
4 potentially misleading since the nonlinearity in the exposure-internal-dose relationship for
5 humans is uncertain above this point. However, the use of a BMDL should place the exposure
6 concentrations well within the linear range of the model.
7 The data from (Ernstgard et al., 2005) were used to assess the use of the first-order
8 metabolic rate constant to a dataset collected under conditions of light work. Historical measures
9 of QPC (52.6 L/hours/kga75) and QCC (26 L/hours/kg0'75) for individuals exposed under
10 conditions of 50 watts of work from that laboratory (52.6 L/hours/kg0'75) (Corley et al., 1994;
11 Ernstgard, 2005; Johanson et al., 1986) were used for the 2-hour exposure period (Figure. 3-13).
12 Otherwise, there were no changes in the model parameters (no fitting to these data). The results
13 are remarkably good, given the lack of parameter adjustment to data collected in a different
14 laboratory and using different human subjects than those to which the model was calibrated.
Ernstgard et al. 2005
4 6
Time (hr)
8
10
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/kg0 7S (Johanson et al., 1986), and a QCC of
26 L/hours/kg0 7S (Corlev et al., 1994) was used by the model.
Source: Ernstgard et al. (2005).
March 2011
3-45 DRAFT - DO NOT CITE OR QUOTE
-------
3.4.6.2. Oral Route
1 There were no methanol human data available for calibration or validation of the oral
2 route for the human model. In the absence of methanol data to estimate rate constants for oral
3 uptake, human oral absorption parameters reported values for ethanol (Sultatos et al., 2004) are
4 set in the code, except that saturable absorption from the stomach was retained with the KMASC
5 equal to the mouse value. The maximum rate of absorption form the stomach, VMASC, was
6 then set such that for a 70-kg person, VMAS/KM (the effective first-order rate constant at low
7 doses) matched the first-order absorption rate from Sultatos et al. (0.21 hr"1). Also, while
8 Sultatos et al. included a rate of metabolism for ethanol in the stomach, the corresponding fecal
9 elimination rate was set to zero here, effectively assuming 100% absorption of methanol for
10 humans. However, human oral dosimetry was described as zero-order uptake, in which
11 continuous infusion at a constant rate into the stomach equal to the daily dose/24 hours was
12 assumed and human internal doses were computed at steady state. Since absorption is 100% for
13 the human model, at steady state the net rate of absorption must equal the rate of infusion to the
14 stomach, irrespective of the other parameter values. (Changes in the absorption constants simply
15 cause the amount of methanol in each GI compartment at steady state to change until the net rate
16 of absorption from the stomach and intestine equals the rate of infusion.) Thus the human
17 absorption constants were set to what is considered a reasonable estimate, given the lack of
18 human oral PK data, but the simulations are conducted in a way that makes the result insensitive
19 to their values; having human values set does allow for simulations of non-constant infusion,
20 should such be desired. Since the AUC was computed for a continuous oral exposure, its value is
21 just 24 hours times the steady-state blood concentration at a given oral uptake rate.
3.4.7. Monkey PK Data and Analysis
22 In order to estimate internal doses (blood AUCs) for the monkey health-effects study of
23 Burbacher, Grant et al. (1999) and further elucidate the potential differences in methanol
24 pharmacokinetics between NP and pregnant individuals (2nd and 3rd trimester), a focused
25 reanalysis of the data of Burbacher, Shen et al. (1999) was performed. Individual blood
26 concentration measurements prior to and following exposure are shown in scatter plots in
27 Appendix B of Burbacher, Shen et al. (1999). More specifically, the monkeys in the study were
28 exposed for 2.5 hours/day, with the methanol concentration raised to approximately the target
29 concentration for the first 2 hours of each exposure and the last 30 minutes providing a chamber
30 "wash-out" period, when the exposure chamber concentration was allowed to drop to 0. Blood
31 samples were taken and analyzed for methanol concentration at 30 minutes, 1, 2, 3, 4, and 6
32 hours after removal from the chamber (or 1, 1.5, 2.5, 3.5, 4.5, and 6.5 hours after the end of
March 2011 3-46 DRAFT - DO NOT CITE OR QUOTE
-------
1 active exposure). These data were analyzed to compare the PK in NP versus pregnant animals,
2 and fitted with a simple PK model to estimate 24-hour blood AUC values for each exposure
3 level. Dr. Burbacher graciously provided the original data, which were used in this analysis.
4 Two cohorts of monkeys were examined, but the data (plots) did not indicate a systematic
5 difference between the two, so the data from the two cohorts were combined. The data from the
6 scatter plots of Burbacher, Shen et al. (1999) for the NP (pre-pregnancy), first pregnancy (2nd
7 trimester), and second pregnancy (3rd trimester) studies are compared in Figure 3-14, along with
8 model simulations (explained below). Since the pregnancy time points were from animals that
9 had been previously exposed for 87 days plus the duration of pregnancy to that time point, the
10 pre-exposed NP animals were used for comparison, rather than naive animals, with the
11 expectation that effects due to changes in enzyme expression (i.e., induction) from the subchronic
12 exposure would not be a distinguishing factor. Note that each exposure group included a pre-
13 exposure baseline or background measurement, also shown. To aid in distinguishing the data
14 visually, the NP data are plotted at times 5 minutes prior to the actual blood draws and the 3rd
15 trimester at 5 minutes after each blood draw.
16 Overall there appears to be no significant or systematic difference among the NP and
17 pregnant groups. The solid lines are model simulations calibrated to only the 2nd trimester data
18 (details below), but they just as adequately represent average concentrations for the NP and 3rd
19 trimester data. Likewise, a PK model calibrated to the NP PK data adequately predicted the
20 maternal methanol concentrations in the pregnant monkeys (results not shown). Since any
21 maternal:fetal methanol differences are expected to be similar in experimental animals and
22 humans (with the maternal:fetal ratio being close to one due to methanol's high aqueous
23 solubility and relatively limited metabolism by the fetus), the predicted levels for the 2nd
24 trimester maternal blood are used in place of measured or predicted fetal concentrations.
March 2011 3-47 DRAFT - DO NOT CITE OR QUOTE
-------
1800 ppm
64.0
-200
-100
0 100 200
Time (min)
300
400
600 ppm
24.0 i
x Pre-pregnancy
+ 2ndTrimester
- SrdTrimester
Simulation
-200
-100
0 100 200
Time (min)
300
400
200 ppm
10.0 i
-200
-100
0 100 200
Time (min)
300
400
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, Shen et al. (1999).
3.4.7.1. PK Model Analysis for Monkeys
1 To analyze and integrate the PK data of Burbacher, Shen et al. (1999), the one-
2 compartment model for Michaelis-Menten kinetics used by Burbacher, Grant et al. (1999) and
3 Burbacher, Shen et al. (1999) was extended by the addition of a chamber compartment to capture
March 2011
3-48 DRAFT - DO NOT CITE OR QUOTE
-------
1 the kinetics of concentration change in the exposure chamber, as shown in Figure 3-15. The data
2 in Figure 3-15 (Burbacher, Grant, et al., 1999; digitized from Figure 5 of Burbacher, Shen, et al.,
3 1999) show an exponential rise to and fall from the approximate target concentration during the
4 exposure period. The use of a single-compartment model for the chamber allows this dynamic to
5 be captured, so that the full concentration-time course is used in simulating the monkey internal
6 concentration rather than an approximate step function (i.e. rather than assuming an
7 instantaneous rise and fall). The pair of equations representing the time-course in the chamber
8 and monkey are as follows (bolded parameters are fit to data):
9 Chamber: dCch/dt = [(CCM'S - Cch)-Fch - RaJ/V^
10 Monkey: dC^/dt = [R^ - Vmax-Cn4/(Km + C^MV^-BW)
11 with R^ = Cch-Rc-(1000-BW)°74 -F and Cnet = C^ + Cbg.
12 d: delta, change
13 Cch: instantaneous chamber concentration (mg/L)
14 t: time (hour)
15 CCM: chamber in-flow methanol concentration (mg/L), which was set to the concentrations
16 corresponding to those reported in Table 2 of Burbacher, Shen et al. (1999), using the
17 "Breeding" column for the NP (87 days pre-exposed; values in Table 3-14)
18 S: exposure switch, set to 1 when exposure is on (first 2 hours) and 0 when off
19 Fch: chamber air-flow, 25,200 L/hours, as specified by Burbacher, Shen et al. (2004)
20 and Burbacher, Grant et al. (2004)
21 Rinh: net rate of methanol inhalation by the monkeys (mg/hr)
22 VCh (1,220 L): chamber volume, initially set to 1,380 L ("accessible volume" stated by
23 Burbacher, Shen et al. (2004) and Burbacher, Grant et al. (2004), but allowed to vary
24 below that value to account for volume taken by equipment, monkey, and to allow for
25 imperfect mixing
26 Cmk: instantaneous inhalation-induced monkey blood methanol concentration (mg/L); this is
27 added to the measured background/endogenous concentration before comparison to data
28 Vmax (39.3 mg/hr): fitted (nonscaled) Michaelis-Menten maximum elimination rate
29 Km (14.6 mg/L): fitted (nonscaled) Michaelis-Menten saturation constant
30 Vmk (0.75 L/kg): fitted volume of distribution for monkey
31 BW: monkey body weight (kg); for NP monkeys set to group average values in data of
32 Burbacher, Shen et al. (1999) and Burbacher, Grant et al. (1999)
33 Rc: allometric scaling factor for total monkey respiration (0.12 L/hours/g0'74 =
34 2 mL/minute/g0'74), as used by Burbacher, Shen et al. (1999) and Burbacher, Grant et al.
35 (1999)(note that scaling is to BW in g, not kg)
March 2011 3-49 DRAFT - DO NOT CITE OR QUOTE
-------
1 F: fractional absorption of inhaled methanol, set to 0.6 (60%), the (rounded) value measured
2 in humans by Sedivec et al. (1981): F and Vmk cannot be uniquely identified, given the
3 model structure, so F was set to the (approximate) human value to obtain a realistic
4 estimate of Vmk
5 Cnet: net blood concentration, equal to sum of the inhalation-induced concentration (Cmk) and
6 the background blood level (Cbg) (mg/L)
7 Cbg: background (endogenous) methanol concentration, set to the pre-exposure group-
8 specific mean from the data of Burbacher, Shen et al. (1999) and Burbacher, Grant et al.
9 (1999)
b
3-
"JT
0
1s
=
o
c
o
2000 -
1800
1600
1400
1200
1000
800
600
400
200
(
A
/ 1800 ppm
Chamber concentration profile
L
I
600 ppm
~m- - m - -m- -m- - -m- -m- - m - -m- -*•-•*- -m- -,
U
' 200 ppm
'.--+---«•---+---+---•»---+---«•---+---+---•»---+--
*W
I I I
) 0.5 1 1.5 2
Time (hr)
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: Buibacher, Shen et al. (1999).
10 The model was specifically fit to the 2nd trimester monkey data, assuming that the
11 parameters were the same for all the exposure groups and concentrations. While the discussion
12 above and data show little difference between the NP and two pregnancy groups, the 2nd
13 trimester group was presumed to be most representative of the average internal dosimetry over
14 the entire pregnancy. Further, the results of Mooney and Miller (2001) show that developmental
15 effects on the monkey brain stem following ethanol exposure are essentially identical for
16 monkeys exposed only during early pregnancy versus full-term, indicating that early pregnancy is
17 a primary window of vulnerability.
18 Model simulation results are the lines shown in Figures 3-14 and 3-15. The model
19 provides a good fit to the monkey blood and chamber air concentration data. While the chamber
March 2011
3-50 DRAFT - DO NOT CITE OR QUOTE
-------
1 volume was treated as a fitted parameter, which was not done by Burbacher, Shen et al. (1999),
2 the chamber concentration data support this estimate. The model does an adequate job of fitting
3 the data for all exposure groups without group-specific parameters. In particular, the data for all
4 exposure levels can be adequately fit using a single value for the volume of distribution (Vmk) as
5 well as each of the metabolic parameters. While one may be able to show statistically distinct
6 parameters for different groups or exposure levels (by fitting the model separately to each), as
7 was done by Burbacher, Shen et al. (1999), it is unlikely that such differences are biologically
8 significant, given the fairly large number of data points and the large variability evident in the
9 blood concentration data. Thus, the single set of parameters listed with the parameter
10 descriptions above will be used to estimate internal blood concentrations for the dose-response
11 analysis. The chamber concentrations for "pregnancy" exposures recorded by Burbacher, Shen et
12 al. (1999: Table 2) and average body weights for each exposure group at the 2nd trimester time
13 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, Shen et al. (1999) and Burbacher, Grant et al. (1999X 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
14 Mouse, rat, and human versions of a methanol PBPK model have been developed and
15 calibrated to data available in the open literature. The model simplifies the structure used by
16 Ward et al. (1997), while adding specific refinements such as a standard lung compartment
17 employed by Fisher et al. (2000) and a two-compartment GI tract.
18 Although the developmental endpoints of concern are effects which occur during in utero
19 and (to a lesser extent) lactational exposure, no pregnancy-specific PBPK model exists for
20 methanol and inadequate data exists for the development and validation of a
21 fetal/gestational/conceptus compartment. The fact that the unique physiology of pregnancy and
22 the fetus/conceptus are not represented in a methanol model would be important if methanol
23 pharmacokinetics differed significantly during pregnancy or if the observed partitioning of
March 2011 3-51 DRAFT - DO NOT CITE OR QUOTE
-------
1 methanol into the fetus/conceptus versus the mother showed a concentration ratio significantly
2 greater than or less than 1. Methanol pharmacokinetics during GD6-GD10 in the mouse are not
3 different from NP mice (Pollack & Brouwer, 1996), and the maternal blood:fetus/conceptus
4 partition coefficient is reported to be near 1 (Horton et al., 1992; Ward et al., 1997). At GDIS in
5 the mouse, maternal blood levels are only modestly different from those in NP animals (see
6 Figures B-4 and B-5 [Appendix B] for examples), and in general the PBPK model simulations
7 for the NP animal match the pregnancy data as well as the nonpregnancy data. Likewise,
8 maternal blood kinetics in monkeys differs little from those in NP animals (see Section 3.4.7 for
9 details). Further, in both mice and monkeys, to the extent that late-pregnancy blood levels differ
10 from NP for a given exposure, they are higher; i.e., the difference between model predictions and
11 actual concentrations is in the same direction. These data support the assumption that the ratio of
12 actual target-tissue methanol concentration to (predicted) NP maternal blood concentrations will
13 be about the same across species, and hence, that using NP maternal blood levels in place of fetal
14 concentrations will not lead to a systematic error when extrapolating risks.
15 The findings in the mouse (similar blood methanol kinetics between NP and pregnant
16 animals prior to GDIS and a maternal blood:fetal partition coefficient close to 1) are assumed to
17 be applicable to the rat. However, the critical gestational window for the reduced brain weight
18 effect observed in the NEDO (1987) rat study is broader than for the mouse cervical rib effect. In
19 addition, NEDO (1987) rats were exposed not only to methanol gestationally but also
20 lactationally and via inhalation after parturition. The additional routes of exposure presented to
21 the pups in this study present uncertainties (see additional discussion in Section 5.3.2) and
22 suggest that average blood levels in pups might be greater than those of the dam.
23 Methanol is transported directly from the maternal circulation to fetal circulation via the
24 placenta, but transfer via lactation involves distribution to the breast tissue, then milk, then
25 uptake from the pup's GI tract. Therefore blood or target-tissue levels in the breast-feeding infant
26 or pup are likely to differ more from maternal levels than do fetal levels. In addition, the health-
27 effects data indicate that most of the effects of concern are due to fetal exposure, with a relatively
28 small influence due to post birth exposures. Further, it would be extremely difficult to
29 distinguish the contribution of post birth exposure from pre birth exposure to a given effect in a
30 way that would allow the risk to be estimated from estimates of both exposure levels, even if one
31 had a lactation/child PBPK model that allowed for prediction of blood (or target-tissue) levels in
32 the offspring. Finally, one would still expect the target-tissue concentrations in the offspring to
33 be closely related to maternal blood levels (which depend on ambient exposure and determine the
34 amount delivered through breast milk), with the relationship between maternal levels and those
35 in the offspring being similar across species. Further, as discussed to a greater extent in Sections
March 2011 3-52 DRAFT - DO NOT CITE OR QUOTE
-------
1 5.1.2 and 5.3.2, it is likely that the difference in blood levels between rat pups and dams would
2 be similar to the difference between mothers and human offspring. Therefore, it is assumed that
3 the potential differences between pup and dam blood methanol levels do not have a significant
4 impact on this risk assessment and the estimation of HECs.
5 Therefore, the development of a lactation/child PBPK model appears not to be necessary,
6 given the minimal change that is likely to result in risk extrapolations, and use of (NP) maternal
7 blood levels as a measure of risk in the offspring is considered preferable over use of default
8 extrapolation methods. In particular, the existing human data allow for predictions of maternal
9 blood levels, which depend strongly on the rate of maternal methanol clearance. Since bottle-fed
10 infants do not receive methanol from their mothers, they are expected to have lower or, at most,
11 similar overall exposures for a given ambient concentration than the breast-fed infant, so that use
12 of maternal blood levels for risk estimation should also be adequately protective for that group.
13 The model fits to the mouse oral-route methanol kinetic data, using a consistent set of
14 parameters (Figure B-4 in Appendix B), are fairly good for doses of 1,500 mg/kg but
15 underpredict blood levels by 30% or more after a dose of 2,500 mg/kg. In particular, the oral
16 mouse model consistently underpredicts the amount of blood methanol reported in two studies
17 (Wardetal.. 1997: 1995). Ward et al. (1997) utilized a different Vmax for each oral absorption
18 dataset; the GDIS and the GD8 data from Dorman et al. (1995) were both fit using a Vmax of
19 -80 mg/kg/hours (body weights were not listed; the model assumed that GD8 and GDIS mice
20 were both 30 g; Ward et al. (1997) did not scale by body weight). Additionally, lower partition
21 coefficients for placenta (1.63 versus 3.28) and embryonic fluid (0.0037 versus 0.77) were used
22 for GD8 and GDIS. The current refined model adequately fits the oral PK data using a single set
23 of parameters that is not varied by dose or source of data.
24 The rat models were able to adequately predict the limited inhalation, oral and i.v.
25 datasets available. Low-dose exposures were emphasized in model optimization due to their
26 greater relevance to risk assessment. Based on a rat inhalation exposure to 500 ppm, the HEC
27 would be 281 ppm (by applying an AUC of 201.3 [Figure B-12] to Equation 1 of Appendix B).
28 The final mouse, rat, and human methanol PBPK models fit multiple datasets for
29 inhalation, oral, and i.v., from multiple research groups using consistent parameters that are
30 representative of each species but are not varied within species or by dose or source of data.
31 Also, a simple PK model calibrated to NP monkey data, which were shown to be essentially
32 indistinguishable from pregnant monkey PK data, was used to estimate blood methanol AUC
33 values (internal doses) in that species. In Section 5, the models and these results are used to
34 estimate chronic human exposure concentrations from internal dose metrics.
March 2011 3-53 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 terms
14 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
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 4-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 & Marsden, 1994). Bhatia and Marsden (1994) have discussed the various behavioral and motor
2 consequences of focal lesions of the basal ganglia from 240 case-study reports. Lesions in the
3 subcortical white matter adjacent to the basal ganglia often occur as well (Airas et al., 2008;
4 Bhatia & Marsden, 1994; Rubinstein et al., 1995). In the case reports of Patankar et al. (1999), it
5 was noted that the severity and extent of necrosis in the lenticular nuclei do not necessarily
6 correlate with clinical outcome.
7 In one of the earliest reviews of methanol overexposure, Bennett et al. (1953) described a
8 mass accidental poisoning when 323 persons, ranging in age from 10 to 78 years, in Atlanta,
9 Georgia, consumed "whisky" adulterated with as much as 35-40% methanol. In all, 41 people
10 died. Of the 323 individuals, 115 were determined to be acidotic with symptoms (visual
11 impairment, headache [affecting ~62%], dizziness [affecting ~30%], nausea, abdominal pain and
12 others) beginning around 24 hours post exposure. Visual impairment was mostly characterized
13 by blurred or indistinct vision; some who were not acidotic experienced transient visual
14 disturbances. The cardiovascular parameters were unremarkable. The importance of acidosis to
15 outcome is shown in Table 4-1. Among the key pathological features were cerebral edema, lung
16 congestion, gastritis, pancreatic necrosis, fatty liver, epicardial hemorrhages, and congestion of
17 abdominal viscera.
18 In another early investigation of methanol poisoning (involving 320 individuals), Benton
19 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 acidosisa
Subjects
All patients
Acidotic (CO2 <20 mEq)
Acidotic (CO2 <10 mEq)
Number
323
115
30
Percent deaths
6.2
19
50
"These data do not include those who died outside the hospital or who were moribund on arrival.
Source: Bennett etal. (1953).
20 Riegel and Wolf (1966), in a case report involving a 60-year-old woman who ingested
21 methanol, noted that nausea and dizziness occurred within 30 minutes of ingestion. She
22 subsequently passed out and remained unconscious for 3 days. Upon awakening she had
23 paralysis of the vocal cords and was clinically blind in one eye after 4 months. Some aspects of
24 Parkinson-like symptoms were evident. There was a pronounced hypokinesia with a mask-like
March 2011 4-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 face resembling a severe state of Parkinson's disease. The patient had difficulty walking and
2 could only make right turns with difficulty. There was no memory loss.
3 Treatment of a 13-year-old girl who ingested an unspecified amount of a windshield-
4 washer solution containing 60% methanol was described by Guggenheim et al. (1971). She
5 displayed profound acidosis; her vital signs, once she was treated for acidosis, were normal by 36
6 hours after hospital admission. During the ensuing 6 months after discharge from the hospital,
7 visual acuity (20/400, both eyes) worsened, and she experienced muscle tremors, arm pain, and
8 difficulty in walking. A regimen of levadopa treatment greatly improved her ability to function
9 normally.
10 Ley and Gali (1983) also noted symptoms that are Parkinson like following methanol
11 intoxication. In this case report respiratory support was needed; the woman was in a coma.
12 Once stabilized, she exhibited symptoms similar to those noted in other case study reports, such
13 as blurred vision, movement difficulty, and tremors. Computerized Axial Tomography scan
14 findings highlighted the central nervous system (CNS) as an important site for methanol
15 poisoning.
16 Rubinstein et al. (1995) presented evidence that a methanol blood level of 36 mg/dL
17 (360 mg/L) is associated with a suite of CNS and ocular deficits that led to a 36-year-old man
18 (who subsequently died) becoming comatose. CT scans at 1-2 days following ingestion were
19 normal. However, MRI scans at day 4 revealed lesions in the putamen and peripheral white
20 matter of the cerebral and cerebellar hemispheres. Bilateral cerebellar cortical lesions had been
21 reported in an earlier case of methanol poisoning by Chen et al. (1991).
22 Finkelstein and Vardi (2002) reported that long-term inhalation exposure of a woman
23 scientist to methanol without acute intoxication resulted in a suite of delayed neurotoxic
24 symptoms (e.g., hand tremor, dystonia, bradykinesia, and other decrements in body movement).
25 Despite treatment with levadopa, an increase in the frequency and severity of effects occurred.
26 Exposure to bromine fumes was concomitant with exposure to methanol.
27 Hantson et al. (1997) found, in four cases, that MRI and brain CT scans were important
28 tools in revealing specific brain lesions (e.g., in the putamina and white matter). The first subject
29 was a 57-year-old woman who complained of blurred vision, diplopia, and weakness 24 hours
30 after ingesting 250 mL of a methanolic antifreeze solution. Upon hospital admission she was
31 comatose and in severe metabolic acidosis. An MRI scan at 9 days indicated abnormal
32 hyperintense foci in the putamina (decreased in size by day 23) and subtle lesions (no change by
33 day 23) in the white matter. Upon her discharge, bilateral deficits in visual acuity and color
34 discrimination persisted.
March 2011 4-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 Similar deficits (metabolic acidosis, visual acuity, and color discrimination) were seen in
2 a man who ingested 300 mL of 75% methanol solution. His blood methanol level was
3 163 mg/dL (1,630 mg/L). An MRI administered 24 hours after hospital admission revealed
4 abnormal hyperintense foci in the putamina, with less intense lesions in the white matter. Like
5 the first subject, a subsequent MRI indicated the foci decreased in size over time, but visual
6 impairments persisted.
7 The third individual, a male, ingested an unspecified amount of a methanolic solution.
8 His blood methanol level was 1,290 mg/dL (12,900 mg/L), and he was in a coma upon hospital
9 admission. An MRI revealed lesions in the putamina and occipital subcortical white matter. A
10 follow-up CT scan was performed after 1 year and showed regression of the putaminal lesions
11 but no change in the occipital lesions. Upon his discharge, severe visual impairment remained
12 but no extrapyramidal signs were observed.
13 The last case was a man who became comatose 12 hours after ingesting 100 mL
14 methanol. His blood methanol level at that time was 60 mg/dL (600 mg/L). An MRI revealed
15 lesions in the putamina; at 3 weeks these lesions were observed to have decreased in size. Upon
16 his discharge, the neurological signs had improved but optic neuropathy (in visual evoked
17 potential) was observed.
18 In a separate publication, Hantson et al. (1997) reported a case of a 26-year-old woman
19 who had ingested 250-500 mL methanol during the 38th week of pregnancy. Her initial blood
20 methanol level was 230 mg/dL (2,300 mg/L) (formate was 33.6 mg/dL or 336 mg/L), yet only a
21 mild metabolic acidosis was indicated. No distress to the fetus was observed upon gynecologic
22 examination. Six days after therapy was initiated (methanol was not present in blood), she gave
23 birth. No further complications with either the mother or newborn were noted.
24 There have been several case reports involving infant or toddler exposures to methanol
25 (Brent et al.. 1991: De Brabander et al.. 2005: Kahn & Blum. 1979: Wu et al.. 1995). The report
26 by Wu et al. (1995) involved a 5-week-old infant with moderate metabolic acidosis and a serum
27 methanol level of 1,148 mg/dL (11,480 mg/L), a level that is ordinarily fatal. However, this
28 infant exhibited no toxic signs and survived without any apparent permanent problems. De
29 Brabander et al. (2005) reported the case of a 3-year-old boy who ingested an unknown amount
30 of pure methanol; at 3 hours after ingestion, the blood methanol level was almost 30 mg/dL (300
31 mg/L). Ethanol infusion as a therapeutic measure was not well tolerated; at 8 hours after
32 ingestion, fomepizole was administered, and blood methanol levels stabilized below 20 mg/dL
33 (200 mg/L), a level above which is considered to be toxic by the American Academy of Clinical
34 Toxicology (Barceloux et al., 2002). Neither metabolic acidosis nor visual impairment was
March 2011 4-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 observed in this individual. Hantson et al. (1997), in their review, touted the efficacy of
2 fomepizole over ethanol in the treatment of methanol poisoning
3 Bilateral putaminal lesions, suggestive of nonhemorrhagic necrosis in the brain of a man
4 who accidentally ingested methanol, were reported by Arora et al. (2005). Approximately
5 10 hours after MRI examination, he developed blurred vision and motor dysfunction. After
6 5 months, visual deficits persisted along with extrapyramidal symptoms. Persistent visual
7 dysfunction was also reported in another methanol poisoning case (Arora et al., 2007): the vision
8 problems developed -46 hours subsequent to the incident.
9 Vara-Castrodeza et al. (2007) applied diffusion-weighted MRI on a methanol-induced
10 comatose woman. Diffusion-weighted MRI provides an image contrast distinct from standard
11 imaging in that contrast is dependent on the molecular motion of water (Schaefer et al., 2000).
12 The neuroradiological findings were suggestive of bilateral putaminal hemorrhagic necrosis,
13 cerebral and intraventricular hemorrhage, diffuse cerebral edema, and cerebellar necrosis.
14 Diffusion-weighted MRI allows for differentiation of restricted diffusion which is indicative of
15 nonviable tissue. In this case, treatment for acidosis (blood methanol levels had risen to
16 1,000 mg/L) was unsuccessful and the patient died.
17 Emergency treatment was unable to save the life of a 38-year-old man who presented with
18 abdominal pain and convulsions after methanol intoxication (W. R. Henderson & Brubacher,
19 2002). A review of a head CT scan performed before the individual went into respiratory arrest
20 revealed bilateral globus pallidus ischemia.
21 Discrete lesions of the putamen, cerebral white matter, and corpus callosum were
22 observed upon MRI (8 days post ingestion) in a man exposed to methanol (blood level 370 mg/L)
23 complaining of vision loss (Keles et al., 2007). Standard treatments corrected the acidosis (pH
24 6.8), and at 1-month follow-up, his cognitive function improved but blindness and bilateral optic
25 atrophy were described as permanent. The follow-up MRI showed persistent putaminal lesions
26 with cortical involvement.
27 Fontenot and Pelak (2002) described a case of a woman who presented with persistent
28 blurred vision and a worsening mental status 36 hours after ingestion of an unspecified amount
29 of methanol. The initial CT scan revealed mild cerebral edema. The blood methanol level at this
30 time was 86 mg/dL (860 mg/L). A repeat CT scan 48 hours after presentation showed
31 hypodensities in the putamen and peripheral white matter. One month after discharge, cognitive
32 function improved, and the patient experienced only a mild lower-extremity tremor.
33 Putaminal necrosis and edema of the deep white matter (the corpus callosum was not
34 affected) was found upon MRI examination of a 50-year-old woman who apparently ingested an
35 unknown amount of what was believed to be pure laboratory methanol (Kuteifan et al., 1998).
March 2011 4-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 Her blood methanol level was 39.7 mM (127 mg/dL; 1,272 mg/L) upon hospital admission and
2 dropped to 102 mg/dL (1,020 mg/L) at 10 hours and to 71 mg/dL (710 mg/L) at 34 hours. The
3 woman, a chronic alcoholic, was in a vegetative state when found and did not improved over the
4 course of a year.
5 MRI and CT scans performed on a 51-year-old man with generalized seizures who had a
6 blood methanol level of 95 mM (304 mg/dL; 3,044 mg/L) revealed bilateral hemorrhagic
7 necrosis of the putamen and caudate nuclei (Gaul etal., 1995). In addition, there was extensive
8 subcortical necrosis and bilateral necrosis of the pontine tegmentum and optic nerve. The patient
9 died several hours after the scans were performed.
10 The relation of methanol overexposure to brain hemorrhage was a focus of the report by
11 Phang et al. (1988), which followed the treatment of 7 individuals, 5 of whom died within
12 72 hours after hospital admission. In two of the deceased individuals, CT scans and autopsy
13 revealed putaminal hemorrhagic necrosis. The investigators postulated that the association of
14 methanol with hemorrhagic necrosis may be complicated by the use of heparin during
15 hemodialysis treatment for acidosis
16 Treatment of two men who had drunk a solution containing 58% methanol and presented
17 with impaired vision, coma, and seizures was discussed in a case report by Bessell-Browne and
18 Bynevelt (2007). A CT scan on one individual revealed bilateral putaminal and cerebral lesions.
19 Blood methanol levels were 21 mg/L. This individual, despite standard treatments, never
20 regained consciousness. The second individual, upon MRI, showed scattered hemorrhage at the
21 grey-white interface of the cerebral hemispheres.
22 There have been two case reports (Adanir et al., 2005; Downie et al., 1992) that involved
23 percutaneous and inhalation exposure. Use of a methanol-containing emollient by a woman with
24 chronic pain led to vision loss, hyperventilation and finally, coma (Adanir et al., 2005).
25 Subsequent to standard treatment followed by hospital discharge, some visual impairment and
26 CNS decrements remained. The methanol blood threshold for ocular damage and acidosis
27 appeared to be -20 mg/L. Dutkiewicz et al. (1980) have determined the skin absorption rate to
28 be 0.192 mg/cm2/minute. In the case report of Aufderheide et al. (1993), two firefighters were
29 transiently exposed to methanol by inhalation and the percutaneous route. Both only complained
30 of a mild headache and had blood methanol levels of 23 and 16 mg/dL (230 and 160 mg/L),
31 respectively.
32 Bebarta et al. (2006) conducted a prospective observational study of seven men who had
33 purposefully inhaled a methanol-containing product. Four had a blood methanol level upon
34 hospital presentation of >24 mg/dL (240 mg/L); the mean formic acid level was 71 |ig/dL. One
35 individual had a blood methanol level of 86 mg/dL (860 mg/L) and a blood formic acid level of
March 2011 4-6 DRAFT - DO NOT CITE OR QUOTE
-------
1 250 |ig/mL upon hospital admission. This latter individual was treated with fomepizole. No
2 patient had an abnormal ophthalmologic examination. All seven stabilized quickly and acidosis
3 was normalized in 4 hours.
4 Numerous other case reports documenting putaminal necrosis/hemorrhage and/or
5 blindness have been reported (Blanco et al., 2006; Chen et al., 1991; Feany et al., 2001; Hsu et
6 al.. 1997:Pelletieretal., 1992).
7 Hovda et al. (2005) presented a combined prospective and retrospective case series study
8 of 51 individuals in Norway (39 males and 12 females, many of whom were alcoholics) who
9 were hospitalized after consuming tainted spirits containing 20% methanol and 80% ethanol. In
10 general, serum methanol concentrations were highest among those most severely affected. The
11 poor outcome was closely correlated with the degree of metabolic acidosis. It was noted by the
12 investigators that the concomitant consumption of ethanol prevented more serious sequelae in 2/5
13 individuals who presented with detectable ethanol levels and were not acidotic despite 2 having
14 the highest blood methanol levels. However, others with detectable levels of ethanol along with
15 severe metabolic acidosis (two of whom died) presumably had subtherapeutic levels of ethanol in
16 their system.
17 In a later report, Hovda et al. (2007) focused on formate kinetics in a 63-year-old male
18 who died 6 days after being admitted to the hospital with headache, vomiting, reduced vision,
19 and dizziness. The investigators speculated that the prolonged metabolic acidosis observed (T1/2
20 for formic acid was 77 hours before dialysis, compared to a typical normal range of 2.5-12 hours)
21 may have been related to retarded formate elimination.
22 Hovda and colleagues (Hunderi et al., 2006) found a strong correlation between blood
23 methanol concentration and the osmolal gap (R2 = 0.92) among 17 patients undergoing dialysis
24 after consuming methanol-contaminated spirits. They concluded that the osmolal gap could be
25 taken as a priori indication of methanol poisoning and be used to guide initiation and duration of
26 dialysis. As they indicated, many hours of dialysis could be safely dispensed with. The osmolal
27 gap pertains to the effect that methanol (and other alcohols) has on the depression of the freezing
28 point of blood in the presence of normal solutes. Braden et al. (1993) demonstrated in case
29 studies that the disappearance of the osmolal gap correlates with the correction of acidosis; they
30 cautioned that methanol and ethanol should not be assumed to be the main factors in causing
31 osmolal gap as glycerol and acetone and its metabolites can as well. A more detailed discussion
32 of the anion and osmolal gap has been provided by Henderson and Brubacher (2002).
33 Hassanian-Moghaddam et al. (2007) compiled data on the prognostic factor relating to
34 outcome in methanol-poisoning cases in Iran. They examined 25 patients, 12 of whom died; 3 of
35 the survivors were rendered blind. There was a significant difference in mean pH of the first
March 2011 4-7 DRAFT - DO NOT CITE OR QUOTE
-------
1 arterial blood gas measurements of those who subsequently died compared with survivors. It
2 was concluded that poor prognosis was associated with pH <7, coma upon admission, and
3 >24-hours delay from intake to admission.
4 The use of blood methanol levels as predictors of outcome is generally not recommended
5 (Barceloux et al., 2002). These investigators cited differences in sampling time, ingestion of
6 ethanol, and levels of toxic (e.g., formic acid) metabolites among the complicating factors. As an
7 illustration, the case report by Prabhakaran et al. (1993) cites two women who ingested a
8 methanol solution (photocopying diluent) at about the same time, were admitted to the hospital
9 about the same time (25-26 hours after ingestion) and had identical plasma methanol
10 concentrations (83 mg/dL; 830 mg/L) upon admission, but different outcomes. Patient #1 was in
11 metabolic acidosis and had an unstable conscious state even after treatment. Upon discharge at
12 day 6, there were no apparent sequelae. Patient #2 had severe metabolic acidosis, fixed and
13 dilated pupils, and no brain stem reflexes. This patient died at day 3 even though therapeutic
14 measures had been administered.
15 In a discussion of 3 fatal methanol-overexposure cases, Andresen et al. (2008) found
16 antemortem blood methanol levels of 540 and 740 mg/dL (5,400 and 7,400 mg/L) in two
17 individuals. At autopsy brain stem blood levels were 738 and 1,008 mg/dL (7,380 and
18 10,080 mg/L), respectively. These brain levels were much higher than blood levels postmortem.
19 Autopsy revealed brain and pulmonary edema in all three individuals; in the two who had the
20 longer survival times, there was hemorrhagic necrosis of the putamen and hemorrhages of the
21 tissue surrounding the optic nerve. In their study of 26 chronic users of methylated spirits, Meyer
22 et al. (2000) found that the best predictor of death or a poor outcome in chronic abusers was a pH
23 <7.0; there was no correlation between blood methanol levels and outcome. Mahieu et al. (1989)
24 considered a latency period before treatment exceeding 10 hours and a blood formate level >50
25 mg/dL (500 mg/L) as predictive of possible permanent sequelae. Liu et al. (1998) in their
26 examination of medical records of 50 patients treated for methanol poisoning over a 10-year
27 period found that: (1) deceased patients had a higher mean blood methanol level than survivors;
28 and (2) initial arterial pH levels <7.0 (i.e., severe metabolic acidosis). Coma or seizure was also
29 associated with higher mortality upon hospital admission.
30 Numerous cases of methanol poisoning have been documented in a variety of countries.
31 In Tunisia, 16 cases of methanol poisoning were discussed by Brahmi et al. (2007). Irreversible
32 blindness occurred in two individuals, with others reporting CNS symptoms, GI effects, visual
33 disturbances, and acidosis. Putaminal necrosis was also described in case reports from Iran
34 (Sefidbakht et al.. 2007). Of 634 forensic autopsies carried out in Turkey during 1992-2003, 18
35 appeared to be related to methanol poisoning (Azmak, 2006). Brain edema and focal necrosis of
March 2011 4-8 DRAFT - DO NOT CITE OR QUOTE
-------
1 the optic nerve were among various sequelae noted. Dethlefs and colleagues (Dethlefs & Naraqi,
2 1978; Naraqi et al., 1979) described permanent ocular damage in 8/24 males who ingested
3 methanol in Papua New Guinea.
4 In summary, most cases of accidental/intentional methanol poisoning reveal a common
5 set of symptoms, many of which are likely to be presented upon hospital admission. These
6 include:
7 • blurred vision and bilateral or unilateral blindness
8 • convulsions, tremors, and coma
9 • nausea, headache, and dizziness
10 • abdominal pain
11 • diminished motor skills
12 • acidosis
13 • dyspnea
14 • behavioral and/or emotional deficits
15 • speech impediments
16 Acute symptoms generally are nausea, dizziness, and headache. In the case reports cited
17 above, the onset of symptom sets as well as their severity varies depending upon how much
18 methanol was ingested, whether or not and when appropriate treatment was administered, and
19 individual variability. A longer time between exposure and treatment, with few exceptions,
20 results in more severe outcomes (e.g., convulsions, coma, blindness, and death). The diminution
21 of some acute and/or delayed symptoms may reflect concomitant ingestion of ethanol or how
22 quickly therapeutic measures (one of which includes ethanol infusion) were administered in the
23 hospital setting.
24 Those individuals who are in a metabolic acidotic state (e.g., pH <7.0) are typically the
25 individuals who manifest the more severe symptoms. Many case reports stress that, unlike blood
26 pH levels <7.0, blood levels of methanol are not particularly good predictors of health outcome.
27 According to a publication of the American Academy of Clinical Toxicology (Barceloux et al.,
28 2002), "the degree of acidosis at presentation most consistently correlates with severity and
29 outcome."
30 As the case reports demonstrate, those individuals who present with more severe
31 symptoms (e.g., coma, seizures, severe acidosis) generally exhibit higher mortality (even after
32 treatment) than those without such symptoms. In survivors of poisoning, persistence or
33 permanence of vision decrements and particularly blindness often have been observed
34 Correlation of symptomatology with blood levels of methanol has been shown to vary
35 appreciably between individuals. Blood methanol levels in the case reports involving ingestion
March 2011 4-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 ranged from values of 30 to over 1,000 mg/dL (300 to over 10,000 mg/L). The lowest value
2 (20 mg/dL; 200 mg/L) reported (Adanir et al., 2005) involved a case of percutaneous absorption
3 (with perhaps associated inhalation exposure) that led to vision and CNS deficits after hospital
4 discharge. In one case report (Rubinstein et al., 1995) involving ingestion, coma and subsequent
5 death were associated with an initial blood methanol level of 36 mg/dL (360 mg/L).
6 Upon MRI and CT scans, the more seriously affected individuals typically have focal
7 necrosis in both brain white matter and more commonly, in the putamen. Bilateral hemorrhagic
8 and nonhemorrhagic necrosis of the putamen is considered by many radiologists as the most
9 well-known sequelae of methanol overexposure.
4.1.2. Occupational Studies
10 Occupational health studies have been carried out to investigate the potential effects of
11 chronic exposure to lower levels of methanol than those seen in acute poisoning cases such as
12 those described above. For example, Frederick et al. (1984) conducted a health hazard
13 evaluation on behalf of the National Institute for Occupational Safety and Health (NIOSH) to
14 determine if vapor from duplicating fluid (which contains 99% methanol) used in mimeograph
15 duplicating machines caused adverse health effects in exposed persons. A group of 84 teacher's
16 aides were selected for study, 66 of whom responded with a completed medical questionnaire. A
17 group of 297 teachers (who were not exposed to methanol vapors to the same extent as the
18 teacher's aides) completed questionnaires as a control group. A 15-minute breathing zone sample
19 was taken from 21 duplicators, 15 of which were greater than the NIOSH-recommended short
20 term ceiling concentration of 800 ppm (1048 mg/m3). The highest breathing zone concentrations
21 were in the vicinity of duplicators for which no exhaust ventilation had been provided
22 (3,080 ppm [4,036 mg/m3] was the highest value recorded). Upon comparison of the self-
23 described symptoms of the 66 teacher's aides with those of 66 age-matched teachers chosen from
24 the 297 who responded, the number of symptoms potentially related to methanol were
25 significantly higher in the teacher's aides. These included blurred vision (22.7 versus 1.5%),
26 headache (34.8 versus 18.1%), dizziness (30.3 versus 1.5%), and nausea (18 versus 6%). By
27 contrast, symptoms that are not usually associated with methanol exposure (painful urination,
28 diarrhea, poor appetite, and jaundice) were similar in incidence among the groups.
29 To further investigate these disparities, NIOSH physicians (not involved in the study)
30 defined a hypothetical case of methanol toxicity by any of the following four symptom
31 aggregations: (1) visual changes; (2) one acute symptom (headache, dizziness, numbness,
32 giddiness, nausea or vomiting) combined with one chronic symptom (unusual fatigue, muscle
33 weakness, trouble sleeping, irritability, or poor memory); (3) two acute symptoms; or (4) three
March 2011 4-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 chronic symptoms. By these criteria, 45% of the teacher's aides were classified as being
2 adversely affected by methanol exposure compared to 24% of teachers (p < 0.025). Those
3 teacher's aides and teachers who spent a greater amount of time using the duplicators were
4 affected at a higher rate than those who used the machines for a lower percentage of their work
5 day.
6 Tanner (1992) reviewed the occupational and environmental causes of Parkinson!sm,
7 spotlighting the potential etiological significance of manganese, carbon monoxide, repeated head
8 trauma (such as suffered by boxers), and exposure to solvents. Among the latter, Tanner (1992)
9 discussed the effects of methanol and n-hexane on the nervous system. Acute methanol
10 intoxication resulted in inebriation, followed within hours by GI pain, delirium, and coma.
11 Tanner (1992) pinpointed the formation of formic acid, with consequent inhibition of cytochrome
12 oxidase, impaired mitochondrial function, and decreased ATP formation as relevant biochemical
13 and physiological changes for methanol exposure. Nervous system injury usually includes
14 blindness, Parkinson-like symptoms, dystonia, and cognitive impairment, with injury to
15 putaminal neurons most likely underlying the neurological responses.
16 Kawai et al. (1991) carried out a biomarker study in which 33 occupationally exposed
17 workers in a factory making methanol fuel were exposed to concentrations of methanol of up to
18 3,577 ppm (4,687 mg/m3), as measured by personal samplers of breathing zone air. Breathing
19 zone exposure samples were correlated with the concentrations of methanol in urine at the end of
20 the shift in 38 exposed individuals and 30 controls (r = 0.82). Eleven of 22 individuals who
21 experienced high exposure to methanol (geometric mean of 459 ppm [601 mg/m3]) complained
22 of dimmed vision during work while 32% of this group of workers experienced nasal irritation.
23 These incidences were statistically significant compared to those of persons who worked in low-
24 exposure conditions (geometric mean of 31 ppm [41 mg/m3]). One 38-year-old female worker
25 who had worked at the factory for only 4 months reported that her visual acuity had undergone a
26 gradual impairment. She also displayed a delayed light reflex.
27 Lorente et al. (2000) carried out a case control study of 100 mothers whose babies had
28 been born with cleft palates. Since all of the mothers had worked during the first trimester,
29 Lorente et al. (2000) examined the occupational information for each subject in comparison to
30 751 mothers whose babies were healthy. Industrial hygienists analyzed the work histories of all
31 subjects to determine what, if any, chemicals the affected mothers may have been exposed to
32 during pregnancy. Multivariate analysis was used to calculate odds ratios, with adjustments
33 made for center of recruitment, maternal age, urbanization, socioeconomic status, and country of
34 origin. Occupations with positive outcomes for cleft palate in the progeny were hairdressing
35 (OR = 5.1, with a 95% confidence interval [CI] of 1.0-26) and housekeeping (OR = 2.8, with a
March 2011 4-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 95% CI of 1.1-7.2). Odds ratios for cleft palate only and cleft lip with or without cleft palate
2 were calculated for 96 chemicals. There seemed to be no consistent pattern of association for any
3 chemical or group of chemicals with these impairments, and possible exposure to methanol was
4 negative for both outcomes.
4.1.3. Controlled Studies
5 Two controlled studies have evaluated humans for neurobehavioral function following
6 exposure to -200 ppm (262 mg/m3) methanol vapors in a controlled setting. The occupational
7 TLV established by the American Conference of Governmental Industrial Hygienists (2000) is
8 200 ppm (262 mg/m3). In a pilot study by Cook et al. (19911 12 healthy young men (22-32 years
9 of age) served as their own controls and were tested for neurobehavioral function following a
10 random acute exposure to air or 191 ppm (250 mg/m3) methanol vapors for 75 minutes. The
11 majority of results in a battery of neurobehavioral endpoints were negative. However, statistical
12 significance was obtained for results in the P-200 and N1-P2 component of event-related
13 potentials (brain wave patterns following light flashes and sounds), the Sternberg memory task,
14 and subjective evaluations of concentration and fatigue. As noted by the Cook et al. (1991),
15 effects were mild and within normal ranges. Cook et al. (1991) acknowledged limitations in their
16 study design, such as small sample size, exposure to only one concentration for a single duration
17 time, and difficulties in masking the methanol odor from experimental personnel and study
18 subjects.
19 In a randomized double-blind study, neurobehavioral testing was conducted on 15 men
20 and 11 women (healthy, aged 26-51 years) following exposure to 200 ppm (262 mg/m3) methanol
21 or water vapors for 4 hours (Chuwers et al., 1995): subjects served as their own controls in this
22 study. Exposure resulted in elevated blood and urine methanol levels (up to peak levels of
23 6.5 mg/L and 0.9 mg/L, respectively) but not formate concentrations. The majority of study
24 results were negative. No significant findings were noted for visual, neurophysiological, or
25 neurobehavioral tests except for slight effects (p < 0.05) on P-300 amplitude (brain waves
26 following exposure to sensory stimuli) and Symbol Digit testing (ability to process information
27 and psychomotor skills). Neurobehavioral performance was minimally affected by methanol
28 exposure at this level. Limitations noted by Chuwers et al. (1995) are that studies of alcohol's
29 affect on P-300 amplitude suggest that this endpoint may be biased by unknown factors and some
30 experimenters and subjects correctly guessed if methanol was used.
31 Although the slight changes in P-200 and P-300 amplitude noted in both the Chuwers et
32 al. (1995) and Cook et al. (1991) studies may be an indication of moderate alterations in
33 cognitive function, the results of these studies are generally consistent and suggest that the
March 2011 4-12 DRAFT - DO NOT CITE OR QUOTE
-------
1 exposure concentrations employed were below the threshold for substantial neurological effects.
2 This is consistent with the data from acute poisoning events which have pointed to a serum
3 methanol threshold of 200 mg/L for the instigation of acidosis, visual impairment, and CNS
4 deficits.
5 Mann et al. (2002) studied the effects of methanol exposure on human respiratory
6 epithelium as manifested by local irritation, ciliary function, and immunological factors. Twelve
7 healthy men (average age 26.8 years) were exposed to 20 and 200 ppm (26.2 and 262 mg/m3,
8 respectively) methanol for 4 hours at each concentration; exposures were separated by 1-week
9 intervals. The 20 ppm (26.2 mg/m3) concentration was considered to be the control exposure
10 since previous studies had demonstrated that subjects can detect methanol concentrations of
11 20 ppm (26.2 mg/m3) and greater. Following each single exposure, subclinical inflammation was
12 assessed by measuring concentrations of interleukins (IL-8, IL-lp, and IL-6) and prostaglandin
13 E2 in nasal secretions. Mucociliary clearance was evaluated by conducting a saccharin transport
14 time test and measuring ciliary beat frequency. Interleukin and prostaglandin data were evaluated
15 by a 1-tailed Wilcoxon test, and ciliary function data were assessed by a 2-tailed Wilcoxon test.
16 Exposure to 200 (262 mg/m3) versus 20 ppm (26.2 mg/m3) methanol resulted in a statistically-
17 significant increase in IL-lp (median of 21.4 versus 8.3 pg/mL) and IL-8 (median of 424 versus
18 356 pg/mL). There were no significant effects on IL-6 and prostaglandin E2 concentration,
19 ciliary function, or on the self-reported incidence of subjective symptoms of irritation. The
20 authors concluded that exposure to 200 ppm (262 mg/m3) methanol resulted in a subclinical
21 inflammatory response.
22 In summary, adult human subjects acutely exposed to 200 ppm (262 mg/m3) methanol
23 have experienced slight neurological (Chuwers et al., 1995) and immunological effects
24 (increased subclinical biomarkers for inflammation) with no self-reported symptoms of irritation
25 (Mann et al., 2002). These exposure levels were associated with peak methanol blood levels of
26 6.5 mg/L (Chuwers et al., 1995), which is approximately threefold higher than background
27 methanol blood levels reported for adult human subjects on methanol-restrictive diets
28 (Table 3-1). Nasal irritation effects have been reported by adult workers exposed to 459 ppm
29 (601 mg/m3) methanol (Kawai et al., 1991). Frank effects such as blurred vision, bilateral or
30 unilateral blindness, coma, convulsions/tremors, nausea, headache, abdominal pain, diminished
31 motor skills, acidosis, and dyspnea begin to occur as blood levels approach 200 mg methanol/L,
32 while 800 mg/L appears to be the threshold for lethality. Data for subchronic, chronic or in utero
33 human exposures are very limited and inconclusive.
March 2011 4-13 DRAFT - DO NOT CITE OR QUOTE
-------
4.2. ACUTE, SUBCHRONIC AND CHRONIC STUDIES IN ANIMALS - ORAL AND
INHALATION
1 A number of studies in animals have investigated the acute, subchronic, and chronic
2 toxicity of methanol. Most are via the inhalation route. Presented below are summaries of the
3 noncancer effects reported in these bioassays. Carcinogenic effects are not described or
4 discussed in this assessment.
4.2.1. Oral Studies
4.2.1.1. Acute Toxicity
5 Although there are few studies that have examined the short-term toxic effects of
6 methanol via the oral route, a number of median lethal dose (LD50) values have been published
7 for the compound. As listed in Lewis (1992), these include 5,628 mg/kg in rats, 7,300 mg/kg in
8 mice, and 7,000 mg/kg in monkeys.
4.2.1.2. Subchronic Toxicity
9 An oral repeat dose study was conducted by the EPA (1986c) in rats. Sprague-Dawley
10 rats (30/sex/dose) were gavaged with 0, 100, 500, or 2,500 mg/kg-day of methanol. Six weeks
11 after dosing, 10 rats/sex/dose group were subjected to interim sacrifice, while the remaining rats
12 continued on the dosing regimen until the final sacrifice (90 days). This study generated data on
13 weekly body weights and food consumption, clinical signs of toxicity, ophthalmologic
14 evaluations, mortality, blood and urine chemistry (from a comprehensive set of hematology,
15 serum chemistry, and urinalysis tests), and gross and microscopic evaluations for all test animals.
16 Complete histopathologic examinations of over 30 organ tissues were done on the control and
17 high-dose rats. Histopathologic examinations of livers, hearts, and kidneys and all gross lesions
18 seen at necropsy were done on low-dose and mid-dose rats. There were no differences between
19 dosed animals and controls in body weight gain, food consumption, or upon gross or microscopic
20 evaluations. Elevated levels (p < 0.05 in males) of serum alanine transaminase (ALT)15 and
21 serum alkaline phosphatase (SAP), and increased (but not statistically significant) liver weights
22 in both male and female rats suggest possible treatment-related effects in rats bolus dosed with
23 2,500 mg methanol/kg-day despite the absence of supportive histopathologic lesions in the liver.
24 Brain weights of high-dose group (2,500 mg/kg-day) males and females were significantly less
25 than those of the control group at terminal sacrifice. Based on these findings, 500 mg/kg-day of
26 methanol is considered an NOEL from this rat study.
15 Also known as serum glutamate pyruvate transaminase (SGPT)
March 2011 4-14 DRAFT - DO NOT CITE OR QUOTE
-------
4.2.1.3. Chronic Noncancer Toxicity
1 A report by Soffritti et al. (2002) summarized a European Ramazzini Foundation (ERF)
2 chronic duration experimental study of methanol16 in which the compound was provided to
3 100 Sprague-Dawley rats/sex/group ad libitum in drinking water at concentrations of 0, 500,
4 5,000, and 20,000 ppm (v/v). The animals were 8 weeks old at the beginning of the study. In
5 general, ERF does not randomly assign animals to treatment groups, but assigns all animals from
6 a given litter to the same treatment group (Bucher, 2002). All rats were exposed for up to
7 104 weeks, then maintained until they died naturally. Rats were housed in groups of 5 in
8 Makrolon cages (41 x 25 x 15 cm) in a room that was maintained at 23 ± 2°C and 50-60%
9 relative humidity. The in-life portion of the experiment ended at 153 weeks with the death of the
10 last animal. Mean daily drinking water, food consumption, and body weights were monitored
11 weekly for the first 13 weeks, every 2 weeks thereafter for 104 weeks, then every 8 weeks until
12 the end of the experiment. Clinical signs were monitored 3 times/day, and the occurrence of
13 gross changes was evaluated every 2 weeks. All rats were necropsied at death then underwent
14 histopathologic examination of organs and tissues.17
15 Soffritti et al. (2002) reported no substantial dose-related differences in survival, but no
16 data were provided. Using individual animal data available from the ERF website,18 Cruzan
17 (2009) reports that male rats treated with methanol generally survived better than controls, with
18 50% survival occurring at day 629, 686, 639 and 701 in the 0, 500, 5,000, and 20, 000 mg/L
19 groups, respectively. There were no significant differences in survival between female control
20 and treatment groups, with 50% survival occurring at day 717, 691, 678 and 708 in the 0, 500,
21 5,000, and 20,000 mg/L groups, respectively. Body weight and water and food consumption
22 were monitored in the study, but the data were not documented in the published report.
23 Soffritti et al. (2002) reported that water consumption in high-dose females was reduced
24 compared to controls between 8 and 56 weeks and that the mean body weight in high-dose males
25 tended to be higher than that of control males. Overall, there was no pattern of compound-
26 related clinical signs of toxicity, and the available data did not provide any indication that the
27 control group was not concurrent with the treated group (Cruzan, 2009). Soffritti et al. (2002)
16 Soffritti et al. (2002) report that methanol was obtained from J.T. Baker, Deventer, Holland, purity grade 99.8%.
17 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.
18 http://www.ramazzini.it/fondazione/foundation.asp.
March 2011 4-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 further reported that there were no compound-related signs of gross pathology or histopathologic
2 lesions indicative of noncancer toxicological effects in response to methanol.
3 Apaja (1980) performed dermal and drinking water chronic bioassays in which male and
4 female Eppley Swiss Webster mice (25/sex/dose group; 8 weeks old at study initiation) were
5 exposed 6 days per week until natural death to various concentrations of malonaldehyde and
6 methanol. The stated purpose of the study was to determine the carcinogen!city of
7 malonaldehyde, a product of oxidative lipid deterioration in rancid beef and other food products
8 in advanced stages of degradation. However, due to its instability, malonaldehyde was obtained
9 from the more stable malonaldehyde bis(dimethyacetal), which was hydrolyzed to
10 malonaldehyde and methanol in dilute aqueous solutions in the presence of a strong mineral acid.
11 In the drinking water portion of this study, mice were exposed to 3 different concentrations of
12 the malonaldehyde/methanol solution and three different control solutions of methanol alone,
13 0.222%, 0.444% and 0.889% methanol in drinking water (222, 444 and 889 ppm, assuming a
14 density of 1 g/ml), corresponding to the stoichiometric amount of methanol liberated by
15 hydrolysis of the acetal in the three test solutions. The methanol was described as Mallinckrodt
16 analytical grade. No unexposed control groups were included in these studies. However, the
17 author provided pathology data from historical records of untreated Swiss mice of the Eppley
18 colony used in two separate chronic studies, one involving 100 untreated males and 100
19 untreated females (Toth etal., 1977) and the other involving 100 untreated females
20 histopathological analyzed by Apaja (Apaja, 1980).
21 Mice in the Apaja (1980) study were housed five/plastic cage and fed Wayne Lab-Blox
22 pelleted diet. Water was available ad libitum throughout life. Liquid consumption per animal
23 was measured 3 times/week. The methanol dose in the dermal study (females only) was 21.3 mg
24 (532 mg/kg-day using an average weight of 0.04 kg as approximated from Figure 4 of the study),
25 three times/week. The methanol doses in the drinking water study were reported as 22.6, 40.8
26 and 84.5 mg/day (560, 1,000 and 2,100 mg/kg-day using an average weight of 0.04 kg as
27 approximated from Figures 14-16 of the study) for females, and 24.6, 43.5 and 82.7 mg/day (550,
28 970, and 1,800 mg/kg-day using an average weight of 0.045 kg as approximated from Figures
29 14-16 of the study) for males, 6 days/week. The animals were checked daily and body weights
30 were monitored weekly. The in-life portion of the experiment ended at 120 weeks with the death
31 of the last animal. Like the Soffritti et al. (2002) study, test animals were sacrificed and
32 necropsied when moribund.19
19 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
March 2011 4-16 DRAFT - DO NOT CITE OR QUOTE
-------
1 The authors reported that survival of the methanol exposed females of the drinking water
2 study was lower than untreated historical controls (p < 0.05), but no significant differences in
3 survival was noted for males. An increase in liver parenchymal cell necrosis was reported in the
4 male and female high-dose groups, with the incidence in females (8%) being significant
5 (p < 0.01) relative to untreated historical controls. Incidence of acute pancreatitis was higher in
6 high-dose males (p <0.001), but did not appear to be dose-related in females, increasing at the
7 mid- (p <0.0001) and low-doses (p <0.01) when compared to historical controls but not
8 appearing at all in the high-dose females. Significant increases relative to untreated historical
9 controls were noted in amyloidosis of the spleen, nephropathy and pneumonia, but the increases
10 did not appear to be dose related.
4.2.2. Inhalation Studies
4.2.2.1. Acute Toxicity
11 Lewis (1992) reported a 4-hour median lethal concentration (LCso) for methanol in rats of
12 64,000 ppm (83,867 mg/m3).
13 Japan's NEDO sponsored a series of toxicological tests on monkeys (M fascicularis),
14 rats, and mice, using inhalation exposure.20 These are unpublished studies; accordingly, they
15 were externally peer reviewed by EPA in 2009.21 A short-term exposure study evaluated
16 monkeys (sex unspecified) exposed to 3,000 ppm (3,931 mg/m3), 21 hours/day for 20 days (1
17 animal), 5,000 ppm (6,552 mg/m3) for 5 days (1 animal), 5,000 ppm (6,552 mg/m3) for 14 days
18 (2 animals), and 7,000 and 10,000 ppm (9,173 and 13,104 mg/m3, respectively) for up to 6 days
19 (1 animal at each exposure level) (NEDO, 1987). Most of the experimental findings were
20 discussed descriptively in the report, without specifying the extent of change for any of the
21 effects in comparison to seven concurrent controls. However, the available data indicate that
22 clinical signs of toxicity were apparent in animals exposed to 5,000 ppm (all exposure durations)
23 or higher concentrations of methanol. These included reduced movement, crouching, weak
24 knees, involuntary movements of hands, dyspnea, and vomiting. In the discussion section of the
or testes, prostate glands and tumors or other gross pathological lesions.
20 In their bioassays, NEDO (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 (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."
21 An external peer review (ERG. 2009) was conducted by EPA in 2009 to evaluate the accuracy of experimental
procedures, results, and interpretation and discussion of the findings presented in these study reports.
March 2011 4-17 DRAFT - DO NOT CITE OR QUOTE
-------
1 summary report, the authors stated that there was a sharp increase in the blood levels of methanol
2 and formic acid in monkey exposed to >3,000 ppm (3,931 mg/m3) methanol. They reported that
3 methanol and formic acid concentrations in the blood of monkeys exposed to 3,000 ppm or less
4 were 80 mg/L and 30 mg/L, respectively.22 In contrast, monkeys exposed to 5,000 ppm or higher
5 concentrations of methanol had blood methanol and formic acid concentrations of 5,250 mg/L
6 and 1,210 mg/L, respectively. Monkeys exposed to 7,000 ppm and 10,000 ppm became critically
7 ill and had to be sacrificed prematurely. Food intake was said to be little affected at 3,000 ppm,
8 but those exposed to 5,000 ppm or more showed a marked reduction. Clinically, the monkeys
9 exposed to 5,000 ppm or more exhibited reduced movement, weak knees, and involuntary
10 movement of upper extremities, eventually losing consciousness and dying.
11 There were no significant changes in growth, with the exception of animals exposed to
12 the highest concentration, where body weight was reduced by 13%. There were few compound-
13 related changes in hematological or clinical chemistry effects, although animals exposed to 7,000
14 and 10,000 ppm showed an increase in white blood cells. A marked change in blood pH values
15 at the 7,000 ppm and 10,000 ppm levels (values not reported) was attributed to acidosis due to
16 accumulation of formic acid. A range of histopathologic changes to the CNS was apparently
17 related to treatment. Severity of the effects was increased with exposure concentration. Lesions
18 included characteristic degeneration of the bilateral putamen, caudate nucleus, and claustrum,
19 with associated edema in the cerebral white matter. Necrosis of the basal ganglia was noted
20 following exposure to 5,000 ppm for 5 days (1 animal) and 14 days(l animal). Exposure to
21 3,000 ppm was considered to be close to the threshold for these necrotic effects, as the monkeys
22 exposed at this level experienced little more than minimal fibrosis of responsive stellate cells of
23 the thalamus, hypothalamus and basal ganglion. The authors reported that no clinical or
24 histopathological effects of the visual system were apparent, but that exposure to 3,000 ppm
25 (3,931 mg/m3) or more caused dose-dependent fatty degeneration of the liver, and exposure to
26 5,000 ppm (6,552 mg/m3) or more caused vacuolar degeneration of the kidneys, centered on the
27 proximal uniferous tubules.
4.2.2.2. Subchronic Toxicity
28 A number of experimental studies have examined the effects of subchronic exposure to
29 methanol via inhalation. For example, Sayers et al. (1944) employed a protocol in which 2 male
30 dogs were repeatedly exposed (8 times daily for 3 minutes/exposure) to 10,000 ppm
31 (13,104 mg/m3) methanol for 100 days. One of the dogs was observed for a further 5 days before
22 Note that Burbacher, Shen et al. (1999) and Burbacher, Grant et al. (2004) 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).
March 2011 4-18 DRAFT - DO NOT CITE OR QUOTE
-------
1 sacrifice; the other dog was observed for 41 days postexposure. There were no clinical signs of
2 toxicity, and both gained weight during the study period. Blood samples were drawn on a regular
3 basis to monitor hematological parameters, but few if any compound-related changes were
4 observed. Ophthalmoscopic examination showed no incipient anomalies at any point during the
5 study period. Median blood concentrations of methanol were 65 mg/L (range 0-280 mg/L) for
6 one dog, and 140 mg/L (70-320 mg/L) for the other.
7 White et al. (1983) exposed 4 male Sprague-Dawley rats/group, 6 hours/day, 5 days/week
8 to 0, 200, 2,000, or 10,000 ppm (0, 262, 2,621, and 13,104 mg/m3) methanol for periods of 1, 2,
9 4, and 6 weeks. Additional groups of 6-week-exposure animals were granted a 6-week
10 postexposure recovery period prior to sacrifice. The lungs were excised intact and lavaged
11 6 times with known volumes of physiological saline. The lavage supernatant was then assayed
12 for lactate dehydrogenase (LDH) and 7V-acetyl-/?-Z)-glucosamidase (/?-NAG) activities. Other
13 parameters monitored in relation to methanol exposure included absolute and relative lung
14 weights, lung DNA content, protein, acid RNase and acid protease, pulmonary surfactant,
15 number of free cells in lavage/unit lung weight, surface protein, LDH, and/?-NAG. As discussed
16 by the authors, none of the monitored parameters showed significant changes in response to
17 methanol exposure.
18 Andrews et al. (1987) carried out a study of methanol inhalation in 5 Sprague-Dawley
19 rats/sex/group and 3 M. fascicularis monkeys/sex/group, 6 hours/day, 5 days/week, to 0, 500,
20 2,000, or, 5,000 ppm (0, 660, 2,620, and 6,552 mg/m3) methanol for 4 weeks. Clinical signs
21 were monitored twice daily, and all animals were given a physical examination once a week.
22 Body weights were monitored weekly, and animals received an ophthalmoscopic examination
23 before the start of the experiment and at term. Animals were sacrificed at term by exsanguination
24 following i.v. barbiturate administration. A gross necropsy was performed, weights of the major
25 organs were recorded, and tissues and organs taken for histopathologic examination. As
26 described by the authors, all animals survived to term with no clinical signs of toxicity among the
27 monkeys and only a few signs of irritation to the eyes and nose among the rats. In the latter case,
28 instances of mucoid nasal discharges appeared to be dose related. There were no differences in
29 body weight gain among the groups of either rats or monkeys, and overall, absolute and relative
30 organ weights were similar to controls. The only exception to this was a decrease in the absolute
31 adrenal weight of female high-concentration monkeys and an increase in the relative spleen
32 weight of mid-concentration female rats. These changes were not considered by the authors to
33 have biological significance. For both rats and monkeys, there were no compound-related
34 changes in gross pathology, histopathology, or ophthalmoscopy. These data suggest a NOAEL of
March 2011 4-19 DRAFT - DO NOT CITE OR QUOTE
-------
1 5,000 ppm (6,600 mg/m3) for Sprague-Dawley rats and monkeys under the conditions of the
2 experiment.
3 Two studies by Poon et al. (1994; 1995) examined the effects of methanol on Sprague-
4 Dawley rats when inhaled for 4 weeks. The effects of methanol were evaluated in comparison to
5 those of toluene and toluene/methanol mixtures (Poon et al., 1994), and to gasoline and
6 gasoline/methanol mixtures (Poon et al., 1995). In the first case (Poon et al., 1994), 10 Sprague-
7 Dawley rats/sex/group were exposed via inhalation, 6 hours/day, 5 days/week to 0, 300, or
8 3,000 ppm (0, 393, 3,930 mg/m3) methanol for 4 weeks. Clinical signs were monitored daily,
9 and food consumption and body weight gain were monitored weekly. Blood was taken at term
10 for hematological and clinical chemistry determinations. Weights of the major organs were
11 recorded at necropsy, and histopathologic examinations were carried out. A 10,000 x g Hver
12 supernatant was prepared from each animal to measure aniline hydroxylase, aminoantipyrine N-
13 demethylase, and ethoxyresorufin-O-deethylase activities. For the most part, the responses to
14 methanol alone in this experiment were unremarkable. All animals survived to term, and there
15 were no clinical signs of toxicity among the groups. Body weight gain and food consumption did
16 not differ from controls, and there were no compound-related effects in hematological or clinical
17 chemistry parameters or in hepatic mixed function oxidase activities. However, the authors
18 described a reduction in the size of thyroid follicles that was more obvious in female than male
19 rats. The authors considered this effect to possibly have been compound related, although the
20 incidence of this feature for the 0, 300, and 3,000 ppm-receiving females was 0/6, 2/6, and 2/6,
21 respectively.
22 The second experimental report by Poon et al. (1995) involved the exposure of
23 15 Sprague-Dawley rats/sex/group, 6 hours/day, 5 days/week for 4 weeks to 0 or 2,500 ppm (0
24 and 3,276 mg/m3) to methanol as part of a study on the toxicological interactions of methanol
25 and gasoline. Many of the toxicological parameters examined were the same as those described
26 in Poon et al. (1994) study. However, in this study urinalysis featured the determination of
27 ascorbic and hippuric acids. Additionally, at term, the lungs and tracheae were excised and
28 aspirated with buffer to yield bronchoalveolar lavage fluid that was analyzed for ascorbic acid,
29 protein, and the activities of gamma-glutamyl transferase (y-GT), AP and LDH. Few if any of the
30 monitored parameters showed any differences between controls and those animals exposed to
31 methanol alone. However, two male rats had collapsed right eyes, and there was a reduction in
32 relative spleen weight in females exposed to methanol. Histopathologic changes in methanol-
33 receiving animals included mild panlobular vacuolation of the liver in females and some mild
34 changes to the upper respiratory tract, including mucous cell metaplasia. The incidence of the
35 latter effect, though higher, was not significantly different than controls in rats exposed to
March 2011 4-20 DRAFT - DO NOT CITE OR QUOTE
-------
1 2,500 ppm (3,267 mg/m3) methanol. However, there were also signs of an increased severity of
2 the effect in the presence of the solvent. No histopathologic changes were seen in the lungs or
3 lower respiratory tract of rats exposed to methanol alone.
4.2.2.3. Chronic Noncancer Toxicity
4 Information on the chronic noncancer toxicity of inhalation exposure to methanol has
5 come from NEDO (1987) which includes the results of experiments on 1) monkeys exposed for
6 up to 3 years, 2) rats and mice exposed for 12 months, 3) mice exposed for 18 months, and 4)
7 rats exposed for 2 years. These are unpublished studies; accordingly, they were externally peer
8 reviewed by EPA in 2009.23
9 In the monkeys, 8 animals (sex unspecified) were exposed to 10, 100, or 1,000 ppm (13,
10 131, and 1,310 mg/m3) methanol, 21 hours/day, for 7 months (2 animals), 19 months,
11 (3 animals), or 29 months (3 animals). There was no indication in the NEDO (1987) report that
12 this study employed a concurrent control group. One of the 3 animals receiving 100 ppm
13 methanol and scheduled for sacrifice at 29 months was terminated at 26 months. Clinical signs
14 were monitored twice daily, body weight changes and food consumption were monitored weekly,
15 and all animals were given a general examination under anesthetic once a month. Blood was
16 collected for hematological and clinical chemistry tests at term, and all animals were subject to a
17 histopathologic examination of the major organs and tissues.
18 While there were no clinical signs of toxicity in the low-concentration animals, there was
19 some evidence of nasal exudate in monkeys in the mid-concentration group. High-concentration
20 (1,000 ppm) animals also displayed this response and were observed to scratch themselves over
21 their whole body and crouch for long periods. Food and water intake, body temperature, and
22 body weight changes were the same among the groups. NEDO (1987) reported that there was no
23 abnormality in the retina of any monkey. When animals were examined with an
24 electrocardiogram, there were no abnormalities in the control or 10 ppm groups. However, in the
25 100 ppm group, one monkey showed a negative change in the T wave. All 3 monkeys exposed to
26 1,000 ppm (1,310 mg/m3) displayed this feature, as well as a positive change in the Q wave. This
27 effect was described as a slight myocardial disorder and suggests that 10 ppm (13.1 mg/m3) is a
28 NOAEL for chronic myocardial effects of methanol and mild respiratory irritation. There were
29 no compound-related effects on hematological parameters. However, 1 monkey in the 100 ppm
30 (131 mg/m3) group had greater than normal amounts of total protein, neutral lipids, total and free
31 cholesterol, and glucose, and displayed greater activities of ALT and aspartate transaminase
23 An external peer review (ERG. 2009) was conducted by EPA in 2009 to evaluate the accuracy of experimental
procedures, results, and interpretation and discussion of the findings presented in these study reports.
March 2011 4-21 DRAFT - DO NOT CITE OR QUOTE
-------
1 (AST). The authors expressed doubts that these effects were related to methanol exposure and
2 speculated that the animal suffered from liver disease.24
3 Histopathologically, no degeneration of the optical nerve, cerebral cortex, muscles, lungs,
4 trachea, tongue, alimentary canal, stomach, small intestine, large intestine, thyroid gland,
5 pancreas, spleen, heart, aorta, urinary bladder, ovary or uterus were reported (neuropathological
6 findings are discussed below in Section 4.4.2. Most of the internal organs showed no compound-
7 related histopathologic lesions. However, there were signs of incipient fibrosis and round cell
8 infiltration of the liver in monkeys exposed to 1,000 ppm (1,310 mg/m3) for 29 months. NEDO
9 (1987) indicated that this fibrosis occurred in 2/3 monkeys of the 1,000 ppm group to a "strictly
10 limited extent." They also qualitatively reported a dose-dependent increase in "fat granules" in
11 liver cells "centered mainly around the central veins" at all doses, but did not provide any
12 response data. The authors state that 1,000 ppm (1,310 mg/m3) represents a chronic lowest-
13 observed-adverse-effect level (LOAEL) for hepatic effects of inhaled methanol, suggesting that
14 the no effect level would be 100 ppm (131 mg/m3). However, this is a tenuous determination
15 given the lack of information on the pathological progression and significance of the appearance
16 of liver cell fat granules at exposures below 1,000 ppm and the lack detail (e.g., time of sacrifice)
17 for the control group.
18 Dose-dependent changes were observed in the kidney; NEDO (1987) described the
19 appearance of Sudan-positive granules in the renal tubular epithelium at 100 ppm (131 mg/m3)
20 and 1,000 (1,310 mg/m3) and hyalinization of the glomerulus and penetration of round cells into
21 the renal tubule stroma of monkeys exposed to methanol at 1,000 (1,310 mg/m3). The former
22 effect was more marked at the higher concentration and was thought by the authors to be
23 compound-related. This would indicate a no effect level at 10 ppm (13.1 mg/m3) for the chronic
24 renal effects of methanol. The authors observed atrophy of the tracheal epithelium in four
25 monkeys. However, the incidence of these effects was unrelated to dose and therefore, could not
26 be unequivocally ascribed to an effect of the solvent. No other histopathologic abnormalities
27 were related to the effects of methanol. Confidence in these determinations is considerably
28 weakened by uncertainty over whether a concurrent control group was used in the chronic
29 study.25
30 NEDO (1987) describes a 12-month inhalation study in which 20 F344 rats/sex/group
31 were exposed to 0, 10, 100, or 1,000 ppm (0, 13.1, 131, and 1,310 mg/m3) methanol,
32 approximately 20 hours/day, for a year. Clinical signs of toxicity were monitored daily; body
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.
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.
March 2011 4-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 weights and food consumption were recorded weekly for the first 13 weeks, then monthly.
2 Blood samples were drawn at term to measure hematological and clinical chemistry parameters.
3 Weights of the major organs were monitored at term, and a histopathologic examination was
4 carried out on all major organs and tissues. Survival was high among the groups; one high-
5 concentration female died on day 337 and one low-concentration male died on day 340. As
6 described by the authors, a number of procedural anomalies arose during this study. For
7 example, male controls in two cages lost weight because of an interruption to the water supply.
8 Another problem was that the brand of feed was changed during the study. Fluctuations in some
9 clinical chemistry and hematological parameters were recorded. The authors considered the
10 fluctuations to be minor and within the normal range. Likewise, a number of histopathologic
11 changes were observed, which, in every case, were considered to be unrelated to exposure level
12 or due to aging.
13 A companion experiment featured the exposure of 30 B6C3F1 mice/sex/group for 1 year
14 to the same concentrations as the F344 rats (NEDO, 1987). Broadly speaking, the same suite of
15 toxicological parameters was monitored as described above, with the addition of urinalysis.
16 10 mice/sex/group were sacrificed at 6 months to provide interim data on the parameters under
17 investigation. A slight atrophy in the external lacrimal gland was observed in both sexes and was
18 significant in the 1,000 ppm male group compared with controls. An apparently dose-related
19 increase in moderate fatty degeneration of hepatocytes was observed in males (1/20, 4/20, 6/20
20 and 8/20 in the 0, 10, 100, and 1,000 ppm dose groups, respectively) which was significantly
21 increased over controls at the 1,000 ppm dose. However the incidence of moderate to severe
22 fatty degeneration was observed in untreated animals maintained outside of the chamber. In
23 addition, there was a clear correlation between fatty degeneration and body weight (a change
24 which was not associated with treatment at 12 months); heavier animals tended to have more
25 severe cases of fatty degeneration. The possibility of renal deficits due to methanol exposure was
26 suggested by the appearance of protein in the urine. However, this effect was also seen in
27 controls and did not display a dose-response effect. Therefore, it is unlikely to be a consequence
28 of exposure to methanol. NEDO (1987) reported other histopathologic and biochemical (e.g.,
29 urinalysis and hematology) findings that do not appear to be related to treatment, including a
30 number of what were considered to be spontaneous tumors in both control and exposure groups.
March 2011 4-23 DRAFT - DO NOT CITE OR QUOTE
-------
1 NEDO (1987. 2008a)26 exposed 52 male and 53 female B6C3F1 mice/group for
2 18 months at the same concentrations of methanol (0, 10, 100 and 1,000 ppm) and with a similar
3 experimental protocol to that described in the 12-month studies.27 Animals were sacrificed at
4 the end of the 18-month exposure period. NEDO (2008a) reported that "there was no
5 microbiological contamination that may have influenced the result of the study" and that the
6 study included an assessment of general conditions, body weight change, food consumption rate,
7 laboratory tests (urinalysis, hematological, and plasma biochemistry) and pathological tests
8 (pathological autopsy,28 organ weight check and histopathology29). As stated in the summary
9 report (NEDO, 1987), a few animals showed clinical signs of toxicity, but the incidence of these
10 responses was not related to dose. Likewise, there were no compound-related changes in body
11 weight increase, food consumption,30 urinalysis, hematology, or clinical chemistry parameters.
12 High-concentration males had lower testis weights compared to control males. Significant
13 differences were detected for both absolute and relative testis weights. One animal in the high-
14 dose group had severely atrophied testis weights, approximately 25% of that of the others in the
15 dose group. Exclusion of this animal in the analysis still resulted in a significant difference in
16 absolute testis weight compared to controls but resulted in no difference in relative testis weight.
17 High-concentration females had higher absolute kidney and spleen weights compared to
18 controls, but there was no significant difference in these organ weights relative to body weight.
19 At necropsy, there were signs of swelling in spleen, preputial glands, and uterus in some animals.
20 Some animals developed nodes in the liver and lung although, according to the authors, none of
21 these changes were treatment-related. NEDO (2008a) reported that all nonneoplastic changes
22 were "nonspecific and naturally occurring changes that are often experienced by 18-month old
23 B6C3F1 mice" and that fatty degeneration of liver that was suspected to occur dose-dependently
24 in the 12-month NEDO (1987) study was not observed in this study.
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. 2008a). The translation was submitted to EPA by the Methanol
Institute and has been certified by NEDO as accurate and complete (Hashimoto & Nedo. 2008). An external peer
review (ERG. 2009) was conducted by EPA in 2009 to evaluate the accuracy of experimental procedures, results,
and interpretation and discussion of the findings presented in these study reports.
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.
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 (NEDO. 2008a) reports sporadic reductions in food consumption of the 1,000 ppm group, but no
associated weight loss or abnormal test results.
March 2011 4-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 Another study reported in NEDO (1987. 2008b)31 was a 24-month bioassay in which 52
2 F344 rats/sex/group were kept in whole body inhalation chambers containing 0, 10, 100, or
3 1,000 ppm (0, 13.1, 131, and 1,310 mg/m3) methanol vapor. Animals were maintained in the
4 exposure chambers for approximately 19.5 hours/day for a total of 733-736 days (males) and
5 740-743 days (females). Animals were monitored once a day for clinical signs of toxicity, body
6 weights were recorded once a week, and food consumption was measured weekly in a 24-animal
7 subset from each group. Urinalysis was carried out on the day prior to sacrifice for each animal,
8 the samples being monitored for pH, protein, glucose, ketones, bilirubin, occult blood, and
9 urobilinogen. Routine clinical chemistry and hematological measurements were carried out and
10 all animals were subject to necropsy at term, with a comprehensive histopathological
11 examination of tissues and organs.32
12 There was some fluctuation in survival rates among the groups in the rat study, though
13 apparently unrelated to exposure concentration.33 In all groups, at least 60% of the animals
14 survived to term. A number of toxicological responses were described by the authors, including
15 atrophy of the testis, cataract formation, exophthalmia, small eye ball, alopecia, and paralysis of
16 the hind leg. However, according to the authors, the incidence of these effects were unrelated to
17 dose and more likely represented effects of aging. NEDO (2008b) reported a mild, nonsignificant
18 (4%) body weight suppression among 1,000 ppm females between 51 and 72 weeks, but that
19 body weight gain was largely similar among the groups for the duration of the experiment. Food
20 consumption was significantly lower than controls in high-concentration male rats during the day
21 210-365 time interval, but no corresponding weight loss was observed. Among hematological
22 parameters, mid- and high-concentration females had a significantly (p > 0.05) higher
23 differential leukocyte count than controls, but dose dependency was not observed. Serum total
24 cholesterol, triglyceride, free fatty acid, and phospholipid concentrations were significantly
25 (p > 0.05) lower in high-concentration females compared to controls. Likewise, serum sodium
26 concentrations were significantly (p > 0.05) lower in mid- and high-concentration males
27 compared to controls. High-concentration females had significantly lower (p > 0.05) serum
31 This study is described in a summary report (NEDO. 1987) and a more detailed, 10-volume translation of the
original chronic rat study report (NEDO. 2008b). The translation was submitted to EPA by the Methanol Institute
and has been certified by NEDO as accurate and complete (Hashimoto & Nedo. 2008). An external peer review
(ERG. 2009) was conducted by EPA in 2009 to evaluate the accuracy of experimental procedures, results, and
interpretation and discussion of the findings presented in these study reports.
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.
March 2011 4-25 DRAFT - DO NOT CITE OR QUOTE
-------
1 concentrations of inorganic phosphorus but significantly (p > 0.05) higher concentrations of
2 potassium compared to controls. Glucose levels were elevated in the urine of high-concentration
3 male rats relative to controls, and female rats had lower pH values and higher bilirubin levels in
4 mid- and high-concentration groups relative to controls. In general, NEDO (1987, 2008b)
5 reported that these variations in urinary, hematology, and clinical chemistry parameters were not
6 related to chemical exposure.
7 NEDO (1987) reported that there was little change in absolute or relative weights of the
8 major organs or tissues. When the animals were examined grossly at necropsy, there were some
9 signs of swelling in the pituitary and thyroid, but these effects were judged to be unrelated to
10 treatment. The most predominant effect was the dose-dependent formation of nodes in the lung
11 of males (2/52, 4/52, 5/52, and 10/52 [p < 0.01] for control, low-, mid-, and high-concentration
12 groups, respectively). Histopathologic examination pointed to a possible association of these
13 nodes with the appearance of pulmonary adenoma (1/52, 5/52, 2/52, and 6/52 for control, low-,
14 mid- and high-concentration groups, respectively) and a single pulmonary adenocarcinoma in the
15 high-dose group (1/52).
4.3. REPRODUCTIVE AND DEVELOPMENTAL STUDIES - ORAL AND
INHALATION
16 Many studies have been conducted to investigate the reproductive and developmental
17 toxicity of methanol. The purpose of these studies was principally to determine if methanol has
18 a similar toxicology profile to another widely studied teratogen, ethanol.
4.3.1. Oral Studies
19 Three studies were identified that investigated the reproductive and developmental effects
20 of methanol in rodents via the oral route (Fu et al., 1996; J. M. Rogers, Mole, et al., 1993;
21 Sakanashi et al., 1996). Two of these studies also investigated the influence of folic acid-deficient
22 (FAD) diets on the effects of methanol exposures (Fu et al., 1996; Sakanashi etal., 1996).
23 Rogers et al. (1993) conducted a developmental toxicity study in which methanol in water
24 was administered to pregnant female CD-I mice via gavage on GD6-GD15. Eight test animals
25 received 4 g/kg-day methanol given in 2 daily doses of 2g/kg; 4 controls received distilled water.
26 By analogy to the protocol of an inhalation study of methanol that was described in the same
27 report, it is assumed that dams were sacrificed on GDI7, at which point implantation sites, live
28 and dead fetuses, resorptions/litter, and the incidences of external and skeletal anomalies and
29 malformations were determined. In the brief summary of the findings provided by the authors, it
30 appears that cleft palate (43.5% per litter versus 0% in controls) and exencephaly (29% per litter
March 2011 4-26 DRAFT - DO NOT CITE OR QUOTE
-------
1 versus 0% in controls) were the prominent external defects following maternal methanol
2 exposure by gavage. Likewise, an increase in totally resorbed litters and a decrease in the
3 number of live fetuses per litter were evident. However, it is possible that these effects may have
4 been caused or exacerbated by the high bolus dosing regimen employed. It is also possible that
5 effects were not observed due to the limited study size. The small number of animals in the
6 control group relative to the test group limits the power of this study to detect treatment-related
7 responses.
8 Sakanashi et al. (1996) tested the influence of dietary folic acid intake on various
9 reproductive and developmental effects observed in CD-I mice exposed to methanol. Starting
10 5 weeks prior to breeding and continuing for the remainder of the study, female CD-I mice were
11 fed folic acid free diets supplemented with 400 (low), 600 (marginal) or 1,200 (sufficient) nmol
12 folic acid/kg. After 5 weeks on their respective diets, females were bred with CD-I male mice.
13 On GD6-GD15, pregnant mice in each of the diet groups were given twice-daily gavage doses of
14 2.0 or 2.5 g/kg-day methanol (total dosage of 4.0 or 5.0 g/kg-day). On GDIS, mice were
15 weighed and killed, and the liver, kidneys and gravid uteri removed and weighed. Maternal liver
16 and plasma folate levels were measured, and implantation sites, live and dead fetuses, and
17 resorptions were counted. Fetuses were weighed individually and examined for cleft palate and
18 exencephaly. One third of the fetuses in each litter were examined for skeletal morphology.
19 They observed an approximate 50% reduction in liver and plasma folate levels in the mice fed
20 low versus sufficient folic acid diets in both the methanol exposed and unexposed groups.
21 Similar to Rogers et al. (1993), Sakanashi et al. (1996) observed that an oral dose of 4-5 g/kg-day
22 methanol during GD6-GD15 resulted in an increase in cleft palate in mice fed sufficient folic
23 acid diets, as well as an increase in resorptions and a decrease in live fetuses per litter. They did
24 not observe an increase in exencephaly in the FAS group at these doses, and the authors suggest
25 that this may be due to diet and the source of CD-I mice differing between the two studies.
26 In the case of the animals fed the folate deficient diet, there was a 50% reduction in
27 maternal liver folate concentration and a threefold increase in the percentage of litters affected by
28 cleft palate (86.2% versus 34.5% in mice fed sufficient folic acid) and a 10-fold increase in the
29 percentage of litters affected by exencephaly (34.5% versus 3.4% in mice fed sufficient folic
30 acid) at the 5 g/kg methanol dose. Sakanashi et al. (1996) speculate that the increased methanol
31 effect from the FAD diet could have been due to an increase in tissue formate levels (not
32 measured) or to a critical reduction in conceptus folate concentration following the methanol
33 exposure. Plasma and liver folate levels at GDIS within each dietary group were not
34 significantly different between exposed versus unexposed mice. However, these measurements
35 were taken 3 days after methanol exposure. Dorman et al. (1995) observed a transient decrease
March 2011 4-27 DRAFT - DO NOT CITE OR QUOTE
-------
1 in maternal red blood cells (RBCs) and conceptus folate levels within 2 hours following
2 inhalation exposure to 15,000 ppm methanol on GD8. Thus, it is possible that short-term
3 reductions in available folate during GD6-GD15 may have affected fetal development.
4 Fu et al. (1996) also tested the influence of dietary folic acid intake on reproductive and
5 developmental effects observed in CD-I mice exposed to methanol. This study was performed
6 by the same laboratory and used a similar study design and dosing regimen as Sakanashi et al.
7 (1996), but exposed the pregnant mice to only the higher 2.5 g/kg-day methanol (total dosage of
8 5.0 g/kg-day) on GD6-GD10. Like Sakanashi et al. (1996). Fu et al. (1996) measured maternal
9 liver and plasma folate levels on GDIS and observed similar, significant reductions in these
10 levels for the FAD versus FAS mice. However, Fu et al. (1996) also measured fetal liver folate
11 levels at GDIS. This measurement does not address the question of whether methanol exposure
12 caused short-term reductions in fetal liver folate because it was taken 8 days after the GD6-GD10
13 exposure period. However, it did provide evidence regarding the extent to which a maternal FAD
14 diet can impact fetal liver folate levels in this species and strain. Significantly, the maternal FAD
15 diet had a greater impact on fetal liver folate than maternal liver folate levels. Relative to the
16 FAS groups, fetal liver folate levels in the FAD groups were reduced 2.7-fold for mice not
17 exposed to methanol (1.86 ± 0.15 nmol/g in the FAD group versus 5.04 ± 0.22 nmol/g in the FAS
18 group) and 3.5-fold for mice exposed to methanol (1.69 ± 0.12 nmol/g in the FAD group versus
19 5.89 ± 0.39 nmol/g in the FAS group). Maternal folate levels in the FAD groups were only
20 reduced twofold both for mice not exposed (4.65 ± 0.37 versus 9.54 ± 0.50 nmol/g) and exposed
21 (4.55 ±0.19 versus 9.26 ± 0.42 nmol/g). Another key finding of the Fu et al. (1996) study is that
22 methanol exposure during GD6-GD10 appeared to have similar fetotoxic effects, including cleft
23 palate, exencephaly, resorptions, and decrease in live fetuses, as the same level of methanol
24 exposure administered during GD6-GD15 (J. M. Rogers, Mole, et al., 1993; Sakanashi et al.,
25 1996). This is consistent with the hypothesis made by Rogers et al. (1993) that the critical period
26 for methanol-induced cleft palate and exencephaly in CD-I mice is within GD6-GD10. As in the
27 studies of Sakanashi et al. (1996) and Rogers et al. (1993), Fu et al. (1996) reported a higher
28 incidence of cleft palate than exencephaly.
4.3.2. Inhalation Studies
29 Nelson et al. (1985) exposed 15 pregnant Sprague-Dawley rats/group to 0, 5,000, 10,000,
30 or 20,000 ppm (0, 6,552, 13,104, and 26,209 mg/m3) methanol (99.1% purity) for 7 hours/day.
31 Exposures were conducted on GDI-GDI9 in the two lower concentration groups and
32 GD7-GD15 in the highest concentration group, apparently on separate days. Two groups of 15
33 control rats were exposed to air only. Day 1 blood methanol levels measured 5 minutes after the
March 2011 4-28 DRAFT - DO NOT CITE OR QUOTE
-------
1 termination of exposure in NP rats that had received the same concentrations of methanol as
2 those animals in the main part of the experiment were 1.00 ± 0.21, 2.24 ± 0.20, and 8.65 ± 0.40
3 mg/mL for those exposed to 5,000, 10,000 and 20,000 ppm methanol, respectively. Evidence of
4 maternal toxicity included a slightly unsteady gait in the 20,000 ppm group during the first few
5 days of exposure. Maternal bodyweight gain and food intake were unaffected by methanol.
6 Dams were sacrificed on GD20, and 13-30 litters/group were evaluated. No effect was observed
7 on the number of corpora lutea or implantations or the percentage of dead or resorbed fetuses.
8 Statistical evaluations included analysis of variance (ANOVA) for body weight effect, Kruskal-
9 Wallis test for endpoints such as litter size and viability and Fisher's exact test for malformations.
10 Fetal body weight was significantly reduced at concentrations of 10,000 and 20,000 ppm by 7%
11 and 12-16%, respectively, compared to controls. An increased number of litters with skeletal
12 and visceral malformations were observed at > 10,000 ppm, with statistical significance obtained
13 at 20,000 ppm. Numbers of litters with visceral malformations were 0/15, 5/15, and 10/15 and
14 with skeletal malformations were 0/15, 2/15, and 14/15 at 0, 10,000, and 20,000 ppm,
15 respectively. Visceral malformations included exencephaly and encephaloceles. The most
16 frequently observed skeletal malformations were rudimentary and extra cervical ribs. The
17 developmental and maternal NOAELs for this study were identified as 5,000 ppm (6,552 mg/m3)
18 and 10,000 ppm (13,104 mg/m3), respectively.
19 NEDO (1987) sponsored a teratology study in Sprague-Dawley rats that included an
20 evaluation of postnatal effects in addition to standard prenatal endpoints. Thirty-six pregnant
21 females/group were exposed to 0, 200, 1,000, or 5,000 ppm (0, 262, 1,310, and 6,552 mg/m3)
22 methanol vapors (reagent grade) on GD7-GD17 for 22.7 hours/day. Statistical significance of
23 results was evaluated by t-test, Mann-Whitney U test, Fisher's exact test, and/or Armitage's ^
24 test.
25 Contrary to the Nelson et al. (1985) report of a 10,000 ppm NOAEL for this rat strain, in
26 the prenatal portion of the NEDO (1987) study, reduced body weight gain and food and water
27 intake during the first 7 days of exposure were reported for dams in the 5,000 ppm group.
28 However, it was not specified if these results were statistically significant. One dam in the
29 5,000 ppm group died on GDI9, and one dam was sacrificed on GDIS in moribund condition.
30 On GD20, 19-24 dams/group were sacrificed to evaluate the incidence of reproductive deficits
31 and such developmental parameters as fetal viability, weight, sex, and the occurrence of
32 malformations. As summarized in Table 4-2, adverse reproductive and fetal effects were limited
33 to the 5,000 ppm group and included an increase in late-term resorptions, decreased live fetuses,
34 reduced fetal weight, and increased frequency of litters with fetal malformations, variations, and
March 2011 4-29 DRAFT - DO NOT CITE OR QUOTE
-------
1 delayed ossifications. Malformations or variations included defects in ventricular septum,
2 thymus, vertebrae, and ribs.
3 Postnatal effects of methanol inhalation were evaluated in the remaining 12 dams/group
4 that were permitted to deliver and nurse their litters. Effects were only observed in the
5 5,000 ppm group, and included a 1-day prolongation of the gestation period and reduced post-
6 implantation survival, number of live pups/litter, and survival on PND4 (Table 4-3). When the
7 delay in parturition was considered, methanol treatment had no effect on attainment of
8 developmental milestones such as eyelid opening, auricle development, incisor eruption, testes
9 descent, or vaginal opening. There were no adverse body weight effects in offspring from
10 methanol treated groups. The weights of some organs (brain, thyroid, thymus, and testes) were
11 reduced in 8-week-old offspring exposed to 5,000 ppm methanol during prenatal development.
Table 4-2. 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)
March 2011
4-30 DRAFT - DO NOT CITE OR QUOTE
-------
Effect
Exposure concentration (ppm)
0
200
1,000
5,000
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, cp < 0.001, as calculated by the authors.
Values are means ± S.D. Values in parentheses are the numbers of litters.
Source: NEDO (1987).
Table 4-3. 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
M b
ap < 0.05, °p < 0.01, cp < 0.001, as calculated by the authors.
Values are means ± S.D. Values in parentheses are the numbers of litters.
Source: NEDO (1987).
March 2011
4-31 DRAFT - DO NOT CITE OR QUOTE
-------
1 NEDO (1987) contains an account of a two-generation reproductive study that evaluated
2 the effects of pre- and postnatal methanol (reagent grade) exposure (20 hours/day) on
3 reproductive and other organ systems of Sprague-Dawley rats. The F0 generation (30 males and
4 30 females per exposure group)34 was exposed to 0, 10, 100, and 1,000 ppm (0, 13.1, 131, and
5 1,310 mg/m3) from 8 weeks old to the end of mating (males) or to the end of lactation period
6 (females). The FI generation was exposed to the same concentrations from birth to the end of
7 mating (males) or to weaning of F2 pups 21 days after delivery (females). Males and females of
8 the F2 generation were exposed from birth to 21 days old (one animal/sex/litter was exposed to
9 8 weeks of age). NEDO (1987) noted reduced brain, pituitary, and thymus weights, and early
10 testicular descent in the offspring of F0 and FI rats exposed to 1,000 ppm methanol. The early
11 testicular descent is believed to be an indication of earlier fetal development as indicated by the
12 fact that it was correlated with increased pup body weight. However, no histopathologic effects
13 of methanol were observed. As discussed in the report, NEDO (1987) sought to confirm the
14 possible compound-related effect of methanol on the brain by carrying out an additional study in
15 which Sprague-Dawley rats were exposed to 0, 500, 1,000, and 2,000 ppm (0, 655, 1,310, and
16 2,620 mg/m3) methanol from the first day of gestation through the FI generation (see
17 Section 4.4.2).
18 Rogers et al. (1993) evaluated development toxicity in pregnant female CD-I mice
19 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,
20 9,894, 13,104, and 19,656 mg/m3) methanol vapors (> 99.9% purity) in a chamber for
21 7 hours/day on GD6-GD15 in a 3-block design experiment. The numbers of mice exposed at
22 each dose were 114, 40, 80, 79, 30, 30, and 44, respectively. During chamber exposures to air or
23 methanol, the mice had access to water but not food. In order to determine the effects of the
24 chamber exposure conditions, an additional 88 control mice were not handled and remained in
25 their cages; 30 control mice were not handled but were food deprived for 7 hours/day on
26 GD6-GD15. Effects in dams and litters were statistically analyzed using the General Linear
27 Models procedure and multiple t-tesi of least squares means for continuous variables and the
28 Fisher's exact test for dichotomous variables. An analysis of plasma methanol levels in 3
29 pregnant mice/block/treatment group on GD6, GD10, and GDIS revealed a dose-related increase
30 in plasma methanol concentration that did not seem to reach saturation levels, and methanol
31 plasma levels were not affected by gestation stage or number of previous exposure days. Across
32 all 3 days, the mean plasma methanol concentrations in pregnant mice were approximately 97,
34 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."
March 2011 4-32 DRAFT - DO NOT CITE OR QUOTE
-------
1 537, 1,650, 3,178, 4,204, and 7,330 ug/mL in the 1,000, 2,000, 5,000, 7,500, 10,000, and
2 15,000 ppm exposure groups, respectively.
3 The dams exposed to air or methanol in chambers gained significantly less weight than
4 control dams that remained in cages and were not handled. There were no methanol-related
5 reductions in maternal body weight gain or overt signs of toxicity. Dams were sacrificed on
6 GDI 7 for a comparison of developmental toxicity in methanol-treated groups versus the chamber
7 air-exposed control group. Fetuses in all exposure groups were weighed, assessed for viability,
8 and examined for external malformations. Fetuses in the control, 1,000, 2,000, 5,000, and
9 15,000 ppm groups were also examined for skeletal and visceral defects. Incidence of
10 developmental effects is listed in Table 4-4. A statistically significant increase in cervical
11 ribs/litter was observed at concentrations of 2,000, 5,000, and 15,000 ppm. At doses of
12 > 5,000 ppm the incidences of cleft palates/litter and exencephaly/litter were increased with
13 statistical significance achieved at all concentrations with the exception of exencephaly which
14 increased but not significantly at 7,500 ppm.35 A significant reduction in live pups/litter was
15 noted at > 7,500 ppm, with a significant increase in fully resorbed litters occurring at
16 > 10,000 ppm. Fetal weight was significantly reduced at > 10,000 ppm. Rogers et al. (1993a)
17 identified a developmental NOAEL and LOAEL of 1,000 ppm and 2,000 ppm, respectively.
18 They also provide BMD maximum likelihood estimates (benchmark concentration [BMC];
19 referred to by the authors as MLE) and estimates of the lower 95% confidence limit on the BMC
20 (benchmark concentration, 95% lower bound [BMCL]; referred to as benchmark dose [BMD] by
21 Rogers et al. (1993) for 5% and 1% added risk, by applying a log-logistic dose-response model to
22 the mean percent/litter data for cleft palate, exencephaly and resorption. The BMCos and
23 BMCLos values for added risk estimated by Rogers et al. (1993) are listed in Table 4-5. From
24 this analysis, the most sensitive indicator of developmental toxicity was an increase in the
25 proportion of fetuses per litter with cervical rib anomalies. The most sensitive BMCL and BMC
26 from this effect for 5% added risk were 305 ppm (400 mg/m3) and 824 ppm (1,080 mg/m3),
27 respectively.36
35 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.
36 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. (J. M. Rogers. Mole, et al.. 1993) 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.
March 2011 4-33 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-4. 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: *p < 0.05, bp < 0.01, °p < 0.001, as calculated by the authors.
Source: Rogers et al. (1993).
Table 4-5. Benchmark doses at two added risk levels
Endpoint
Cleft Palate (CP)
Exencephaly (EX)
CP and EX
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
BMCL01 (ppm)
1,798
784
1,816
2,949
1,915
58
Source: Rogers et al. (1993).
March 2011
4-34 DRAFT - DO NOT CITE OR QUOTE
-------
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/niL 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 GD11-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 incidence
24 of fetuses with 25 presacral vertebrae (26 is normal). However, an increased incidence of fetuses
25 with 27 presacral vertebrae was evident when dams were exposed on GD7. These results
26 indicate that gastrulation and early organogenesis is a period of increased embryonic sensitivity
27 to methanol.
28 Burbacher, Shen et al. (1999) and Burbacher, Grant et al. (1999) carried out toxicokinetic
29 and reproductive/developmental studies of methanol in M. fascicularis monkeys that were
30 published by the Health Effects Institute (HEI) in a two-part monograph. Some of the data were
31 subsequently published in the open scientific literature (Burbacher, Grant, et al., 2004;
32 Burbacher, Shen, et al., 2004). The experimental protocol featured exposure to 2 cohorts of 12
33 monkeys/group to low exposure levels (relative to the previously discussed rodent studies) of 0,
34 200, 600, or 1,800 ppm (0, 262, 786, and 2,359 mg/m3) methanol vapors (99.9% purity),
35 2.5 hours/day, 7 days/week, during a premating period and mating period (-180 days combined)
March 2011 4-35 DRAFT - DO NOT CITE OR QUOTE
-------
1 and throughout the entire gestation period (-168 days). The monkeys were 5.5-13 years old and
2 were a mixture of feral-born and colony-bred animals. The study included an evaluation of
3 maternal reproductive performance and tests to assess infant postnatal growth and newborn
4 health, reflexes, behavior, and development of visual, sensorimotor, cognitive, and social
5 behavioral function (see Section 4.4.2 for a review of the developmental neurotoxicity findings
6 from this study). Blood methanol levels, clearance, and the appearance of formate were also
7 examined and are discussed in Section 3.2.
8 With regard to reproductive parameters, there was a statistically significant decrease
9 (p = 0.03) in length of pregnancy in all treatment groups, as shown in Table 4-6. Maternal
10 menstrual cycles, conception rate, and live birth index were all unaffected by exposure. There
11 were also no signs of an effect on maternal weight gain or clinical toxicity among the dams. The
12 decrease in pregnancy length was largely due to complications of pregnancy requiring Cesarean
13 section (C-section) deliveries in the methanol exposure groups. The C-section deliveries were
14 performed in response to signs of difficulty in the pregnancy and thus may serve as supporting
15 evidence of reproductive dysfunction in the methanol-exposed females.
16 While pregnancy duration was virtually the same in all exposure groups, there were some
17 indications of increased pregnancy duress only in methanol-exposed monkeys. C-sections were
18 done in 2 monkeys from the 200 ppm group and 2 from the 600 ppm group due to vaginal
19 bleeding, presumed, but not verified, to be from placental detachment.37 A monkey in the
20 1,800 ppm group also received a C-section after experiencing nonproductive labor for 3 nights.
21 In addition, signs of prematurity were observed in 1 infant from the 1,800 ppm group that was
22 born after a 150-day gestation period. The authors speculated that the shortened gestation length
23 could be due to a direct effect of methanol on the fetal hypothalamus-pituitary-adrenal (HPA)
24 axis or an indirect effect of methanol on the maternal uterine environment. Other fetal
25 parameters such as crown-rump length and head circumference were unchanged among the
26 groups. Infant growth and tooth eruption were unaffected by prenatal methanol exposure.
37 Burbacher, Grant, et al. (2004) and Burbacher, Shen et al. (2004) 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.
March 2011 4-36 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-6. 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.;
"Live-born offspring only; bp < 0.05, as calculated by the authors.
Source: Burbacher, Grant et al. (2004).
1 In later life, 2 females out of the total of 9 offspring in the 1,800 ppm group experienced a
2 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, Shen et al. (1999) and Burbacher, Grant et al. (1999) studies
9 suggest that methanol exposure can cause reproductive effects, manifested as a shortened mean
10 gestational period due to pregnancy complications that precipitated delivery via a C-section, and
11 developmental neurobehavioral effects which may be related to the shortened gestational period
12 (see Section 4.4.2). The low exposure of 200 ppm may signify a LOAEL for reproductive
13 effects. However, the decrease in gestational length was marginally significant and largely the
14 result of human intervention (C-section) for reasons (presumably pregnancy complications) that
15 were not objectively confirmed with clinical procedures (e.g., placental ultrasound). Also, this
16 effect did not appear to be dose related, the greatest gestational period decrease having occurred
17 at the lowest (200 ppm) exposure level. Thus, a clear NOAEL or LOAEL cannot be determined
18 from 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
March 2011
4-37 DRAFT - DO NOT CITE OR QUOTE
-------
1 series, the most significant changes were in blood testosterone levels of animals exposed to
2 200 ppm methanol, the lowest concentration evaluated. At this exposure level, animals exposed
3 for 6 weeks had testosterone levels that were 32% of those seen in controls. However, higher
4 concentrations of methanol were associated with testosterone levels that were closer to those of
5 controls. However, the lack of a clear dose-response is not necessarily an indication that the
6 effect is not related to methanol. The higher concentrations of methanol could be causing other
7 effects (e.g., liver toxicity) which can influence the results. Male rats exposed to 10,000 ppm
8 methanol for 6 weeks displayed blood levels of LH that were about 3 times higher (mean ± S.D.)
9 than those exposed to air (311 ± 107% versus 100 ± 23%). In discussing their results, the authors
10 placed the greater emphasis on the fact that an exposure level equal to the ACGIH TLV
11 (200 ppm) had caused a significant depression in testosterone formation in male rats.
12 A follow-up study report by the same research group (Cameron et al., 1985) described the
13 exposure of 5 male Sprague-Dawley rats/group, 6 hours/day for either 1 day or 1 week, to
14 methanol, ethanol, n-propanol, or n-butanol at their respective TLVs. Groups of animals were
15 sacrificed immediately after exposure or after an 18-hour recovery period, and the levels of
16 testosterone, LH, and corticosterone measured in serum. As shown in Table 4-7, the data were
17 consistent with the ability of these aliphatic alcohols to cause a transient reduction in the
18 formation of testosterone. Except in the case of n-butanol, rapid recovery from these deficits can
19 be inferred from the 18-hour postexposure data.
March 2011 4-38 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-7. 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
Corticosterone
Control
Methanol
Ethanol
n-Propanol
n-Butanol
200
1,000
200
50
100 ± 20
115 ±18
111 ±32
112±21
143±lla
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 et al. (1985).
1 In a series of studies that are relevant to the reproductive toxicity of methanol in males,
2 Lee et al. (1991) exposed 8-week-old male Sprague-Dawley rats (9-10/group) to 0 or 200 ppm (0
3 and 262 mg/m3) methanol, 8 hours/day, 5 days/week, for 1, 2, 4, or 6 weeks to measure the
4 possible treatment effects on testosterone production. Study results were evaluated by one factor
5 ANOVA followed by Student's Mest. In the treated rats, there was no effect on serum
6 testosterone levels, gross structure of reproductive organs, or weight of testes and seminal
7 vesicles. Lee et al. (1991) also studied the in vitro effect of methanol on testosterone production
8 from isolated testes, but saw no effect on testosterone formation either with or without the
9 addition of human chorionic gonadotropin hormone.
March 2011
4-39 DRAFT - DO NOT CITE OR QUOTE
-------
1 In a third experiment from the same report, Lee et al. (1991) examined testicular
2 histopathology to determine if methanol exposure produced lesions indicative of changing
3 testosterone levels; the effects of age and folate status were also assessed. This is relevant to the
4 potential toxicity of methanol because folate is the coenzyme of tetrahydrofolate synthetase, an
5 enzyme that is rate limiting in the removal of formate. Folate deficiency would be expected to
6 cause potentially toxic levels of methanol, formaldehyde, and formate to be retained. The same
7 authors examined the relevance of folate levels, and by implication, the overall status of formate
8 formation and elimination in mediating the testicular functions of Long-Evans rats. Groups of
9 4-week-old male Long-Evans rats were given diets containing either adequate or reduced folate
10 levels plus 1% succinylsulfathiazole, an antibiotic that, among other activities,38 would tend to
11 reduce the folate body burden. At least 9 rats/dietary group/dose were exposed to 0, 50, 200, or
12 800 ppm (0, 66, 262, and 1,048 mg/m3) methanol vapors starting at 7 months of age while
13 8-12 rats/dietary group/dose were exposed to 0 or 800 ppm methanol vapors at 15 months of age.
14 The methanol exposures were conducted continuously for 20 hours/day for 13 weeks. Without
15 providing details, the study authors reported that visual toxicity and acidosis developed in rats
16 fed the low folate diet and exposed to methanol. No methanol-related testicular lesions or
17 changes in testes or body weight occurred in rats that were fed either the folate sufficient or
18 deficient diets and were 10 months old at the end of treatment. Likewise, no methanol-lesions
19 were observed in 18-month-old rats that were fed diets with adequate folate. However, the
20 incidence but not severity of age-related testicular lesions was increased in the 18-month-old rats
21 fed folate-deficient diets. Subcapsular vacuoles in germinal epithelium were noted in 3/12
22 control rats and 8/13 rats in the 800 ppm group. One rat in the 800 ppm group had atrophied
23 seminiferous tubules and another had Leydig cell hyperplasia. These effects, as well as the
24 transient decrease in testosterone levels observed by Cameron et al. (1984; 1985), could be the
25 result of chemically-related strain on the rat system as it attempts to maintain hormone
26 homeostasis.
27 Dorman et al. (1995) conducted a series of in vitro and in vivo studies of developmental
28 toxicity in ICR BR (CD-I) mice associated with methanol and formate exposure. The studies
29 used HPLC grade methanol and appropriate controls. PK and developmental toxicity parameters
30 were measured in mice exposed to a 6-hour methanol inhalation (10,000 or 15,000 ppm),
31 methanol gavage (1.5 g/kg) or sodium formate (750 mg/kg by gavage) on GD8. In the in vivo
38 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.
March 2011 4-40 DRAFT - DO NOT CITE OR QUOTE
-------
1 inhalation study, 12-14 dams/group were exposed to 10,000 ppm methanol for 6 hours on GD8,39
2 with and without the administration of fomepizole (4-methylpyrazole) to inhibit the metabolism
3 of methanol by ADH1. Dams were sacrificed on GD10, and folate levels in maternal RBC and
4 conceptus (decidual swelling) were measured, as well as fetal neural tube patency (an early
5 indicator of methanol-induced dysmorphogenic response). The effects observed included a
6 transient decrease in maternal RBC and conceptus folate levels within 2 hours following
7 exposure and a significant (p < 0.05) increase in the incidence of fetuses with open neural tubes
8 (9.65% in treated versus 0 in control). These responses were not observed following sodium
9 formate administration, despite peak formate levels in plasma and decidual swellings being
10 similar to those observed following the 6-hour methanol inhalation of 15,000 ppm. This suggests
11 that these methanol-induced effects are not related to the accumulation of formate. As this study
12 provides information relevant to the identification of the proximate teratogen associated with
13 developmental toxicity in rodents, it is discussed more extensively in Section 4.6.1.
4.3.3. Other Reproductive and Developmental Toxicity Studies
14 Additional information relevant to the possible effects of methanol on reproductive and
15 developmental parameters has been provided by experimental studies that have exposed
16 experimental animals to methanol during pregnancy via i.p. injections (J. M. Rogers et al., 2004).
17 Relevant to the developmental impacts of the chemical, a number of studies also have examined
18 the effects of methanol when included in whole-embryo culture (J. E. Andrews et al., 1995; J. E.
19 Andrews et al.. 1993: J. E. Andrews et al.. 1998: Hansen et al.. 2005: Harris et al.. 2003).
20 Pregnant female C57BL/6J mice received 2 i.p. injections of methanol on GD7 (J. M.
21 Rogers et al., 2004). The injections were given 4 hours apart to provide a total dosage of 0, 3.4,
22 and 4.9 g/kg. Animals were sacrificed on GDI7 and the litters were examined for live, dead, and
23 resorbed fetuses. Rogers et al. (2004) monitored fetal weight and examined the fetuses for
24 external abnormalities and skeletal malformations. Methanol-related deficits in maternal and
25 litter parameters observed by Rogers et al. (2004) are summarized in Table 4-8.
39 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.
March 2011 4-41 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-8. Maternal and litter parameters when pregnant female C57BL/6J
mice were injected i.p. with methanol
Parameter
No. pregnant at term
Wt gain GD7-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
O.liO.l
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
O.liO.l
0.70 ± 0.02a
Values are means ± SEM.
ap < 0.05, as calculated by the authors.
Source: Rogers et al. (2004).
1 Rogers et al. (2004) used a number of sophisticated imaging techniques, such as confocal
2 laser scanning and fluorescence microscopy, to examine the morphology of fetuses excised at
3 GD7, GD8, and GD9. They identified a number of external craniofacial abnormalities, the
4 incidence of which was, in all cases, significantly increased in the high-dose group compared to
5 controls. For some responses, such as microanophthalmia and malformed maxilla, the incidence
6 was also significantly increased in animals receiving the lower dose. Fifteen compound-related
7 skeletal malformations were tabulated in the report. In most cases, a dose-response effect was
8 evident, resulting in statistically significant incidences in affected fetuses and litters, when
9 compared to controls. Apparent effects of methanol on the embryonic forebrain included a
10 narrowing of the anterior neural plate, missing optical vesicles, and holoprosencephaly (failure of
11 the embryonic forebrain to divide). The authors noted that there was no sign of incipient cleft
12 palate or exencephaly, as had been observed in CD-I mice exposed to methanol via the oral and
13 inhalation routes (J. M. Rogers, Mole, et al., 1993).
14 In order to collect additional information on cell proliferation and histological changes in
15 methanol-treated fetuses, Degitz et al. (2004) used an identical experimental protocol to that of
16 Rogers et al. (2004) by administering 0, 3.4, or 4.9 g methanol/kg in distilled water i.p. (split
17 doses, 4 hours apart) to C57BL/6J mice on GD7. Embryos were collected at various times on
18 GD8 and GD10. Embryos from dams exposed to 4.9 g/kg and examined on GD8 exhibited
19 reductions in the anterior mesenchyme, the mesenchyme subjacent to the mesencephalon and the
20 base of the prosencephalon (embryonic forebrain), and in the forebrain epithelium. The optic pits
March 2011
4-42 DRAFT - DO NOT CITE OR QUOTE
-------
1 were often lacking; where present their epithelium was thin and there were fewer neural crest
2 cells in the mid- and hindbrain regions.
3 At GD9, there was extensive cell death in areas populated by the neural crest, including
4 the forming cranial ganglia. Dose-related abnormalities in the development of the cranial nerves
5 and ganglia were seen on GD7. In accordance with an arbitrary dichotomous scale devised by
6 the authors, scores for ganglia V, VIE, and IX were significantly (not otherwise specified)
7 reduced at all dose levels, and ganglia YE and X were reduced only at the highest dose. At the
8 highest dose (4.9 g/kg), the brain and face were poorly developed and the brachial arches were
9 reduced in size or virtually absent. Flow cytometry of the head regions of the embryos from the
10 highest dose at GD8 did not show an effect on the proportion of cells in S-phase.
11 Cell growth and development were compared in C57BL/6J and CD-I mouse embryos
12 cultured in methanol (Degitz, Rogers, et al., 2004). GD8 embryos, with 5-7 somites, were
13 cultured in 0, 1, 2, 3, 4, or 6 mg methanol/mL for 24 hours and evaluated for morphological
14 development. Cell death was increased in both strains in a developmental stage- and region-
15 specific manner at 4 and 6 mg/niL after 8 hours of exposure. The proportions of cranial region
16 cells in S-phase were significantly (p < 0.05) decreased at 6 mg/mL following 8- and 18-hour
17 exposures to methanol. After 24 hours of exposure, C57BL/6J embryos had significantly
18 (p < 0.05) decreased total protein at 4 and 6 mg/kg. Significant (p < 0.05) developmental effects
19 were seen at 3, 4, and 6 mg/kg, with eye dysmorphology being the most sensitive endpoint. CD-I
20 embryos had significantly decreased total protein at 3, 4, and 6 mg/kg, but developmental effects
21 were seen only at 6 mg/kg. It was concluded that the C57BL/6J embryos were more severely
22 affected by methanol in culture than the CD-I embryos.
23 Andrews et al. (1993) carried out a comparative study of the developmental toxicity of
24 methanol in whole Sprague-Dawley rat or CD-I mouse embryos. Nine-day rat embryos were
25 explanted and cultured in rat serum containing 0, 2, 4, 8, 12, or 16 mg/mL methanol for 24 hours
26 then transferred to rat serum alone for a further 24 hours. Eight-day mouse embryos were
27 cultured in 0, 2, 4, 6, or 8 mg/mL methanol in culture medium for 24 hours. At the end of the
28 culture period, embryos were examined for growth, development and dysmorphogenesis. For the
29 rats, doses of 8 mg/mL and above resulted in a concentration-related decrease in somite number,
30 head length, and developmental score. Some lethality was seen in embryos incubated at 12
31 mg/mL methanol. For the mouse embryos, incubation concentrations of 4 mg/mL methanol and
32 above resulted in a significant decrease in developmental score and crown-rump length. The
33 high concentration (8 mg/mL) was associated with embryo lethality. These data suggest that
34 mouse embryos are more sensitive than rat embryos to the developmental effects of methanol.
35 Using a similar experimental system to examine the developmental toxicity of formate and
March 2011 4-43 DRAFT - DO NOT CITE OR QUOTE
-------
1 formic acid in comparison to methanol, Andrews et al. (1995) showed that the formates are
2 embryotoxic at doses that are four times lower than equimolar doses of methanol. Andrews et al.
3 (1998) showed that exposure to combinations of methanol and formate was less embryotoxic
4 than would be expected based on simple toxicity additivity, suggesting that the embryotoxicity
5 observed following low-level exposure to methanol is mechanistically different from that
6 observed following exposure to formate.
7 A study by Hansen et al. (2005) determined the comparative toxicity of methanol and its
8 metabolites, formaldehyde and sodium formate, in GD8 mouse (CD-I) and GD10 rat (Sprague-
9 Dawley) conceptuses. Incubation of whole embryos was for 24 hours in chemical-containing
10 media (mouse: 4-12 mg/mL methanol, 1-6 |ig/mL formaldehyde, 0.5-4 mg/mL sodium formate;
11 rat: 8-20 mg/mL, 1-8 |ig/mL, 0.5-8 mg/mL). Subsequently, the visceral yolk sac (VYS) was
12 removed and frozen for future protein and DNA determination. The embryos were examined
13 morphologically to determine growth and developmental parameters such as viability, flexure
14 and rotation, crown-rump length, and neuropore closure. In other experiments, the chemicals
15 were injected directly into the amniotic space. For each response, Table 4-9 provides a
16 comparison of the concentrations or amounts of methanol, formaldehyde, and formate that
17 resulted in statistically significant changes in developmental abnormalities compared to controls.
18 For a first approximation, these concentrations or amounts may be taken as threshold-
19 dose ranges for the specific responses under the operative experimental conditions. The data
20 show consistently lower threshold values for the effects of formaldehyde compared to those of
21 formate and methanol. The mouse embryos were more sensitive towards methanol toxicity than
22 rat embryos, consistent with in vivo findings, whereas the difference in sensitivity disappeared
23 when formaldehyde was administered. Hansen et al. (2005) hypothesized that, while the MOA
24 for the initiation of the organogenic defects is unknown, the relatively low threshold levels of
25 formaldehyde for most measured effects suggest formaldehyde involvement in the embryotoxic
26 effects of methanol. By contrast, formate, the putative toxicant for the acute effects of methanol
27 poisoning (acidosis, neurological deficits), did not appear to reproduce the methanol-induced
28 teratogenicity in these whole embryo culture experiments.
March 2011 4-44 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-9. 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 fig)
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
aCR = crown-rump length,
bVYS = visceral yolk sac.
NR = not reported
Source: Hansen et al. (2005): Harris et al. (2004) (adapted).
1 Harris et al. (2003) provided biochemical evidence consistent with the concept that
2 formaldehyde might be the ultimate embryotoxicant of methanol by measuring the activities of
3 enzymes that are involved in methanol metabolism in mouse (CD-I) and rat (Sprague-Dawley)
4 whole embryos at different stages of development. Specific activities of the enzymes ADH1,
5 ADH3, and CAT, were determined in rat and mouse conceptuses during the organogenesis period
6 of 8-25 somites. Activities were measured in heads, hearts, trunks, and VYS from early- and
7 late-stage mouse and rat embryos. While CAT activities were similar between rat and mouse
8 embryos, mouse ADH1 activities in the VYS were significantly lower throughout organogenesis
9 when compared to the rat VYS or embryos of either species. ADH1 activities of heads, hearts,
10 and trunks from mouse embryos were significantly lower than those from rats at the 7-12 somite
11 stage. However, these interspecies differences were not evident in embryos of 20-22 somites.
March 2011
4-45 DRAFT - DO NOT CITE OR QUOTE
-------
1 ADH3 activities were lower in mouse versus rat VYS, irrespective of the stage of development.
2 However, while ADH3 activities in mouse embryos were markedly lower than those of rats in the
3 early stages of development, the levels of activity were similar to at the 14-16 somite stage and
4 beyond. A lower capacity to transform formaldehyde to formate might explain the increased
5 susceptibility of mouse versus rat embryos to the toxic effects of methanol. The hypothesis that
6 formaldehyde is the ultimate embryotoxicant of methanol is supported by the demonstration of
7 diminished ADH3 activity in mouse versus rat embryos and by the demonstration by Hansen et
8 al. (2005) that formaldehyde has a far greater embryotoxicity than either formate or methanol
9 itself.
10 That formate can induce similar developmental lesions in whole rat and mouse
11 conceptuses was demonstrated by Andrews et al. (1995), who evaluated the developmental
12 effects of sodium formate and formic acid in rodent whole embryo cultures in vitro. Day 9 rat
13 (Sprague-Dawley) embryos were cultured for 24 or 48 hours and day 8 mouse (CD-I) cultures
14 were incubated for 24 hours. As tabulated by the authors, embryos of either species showed
15 trends towards increasing lethality and incidence of abnormalities with exposure concentration.
16 Among the anomalies observed were open anterior and posterior neuropores, plus rotational
17 defects, tail anomalies, enlarged pericardium, and delayed heart development.
4.4. NEUROTOXICITY
18 A substantial body of information exists on the toxicological consequences to humans
19 who consume or are exposed to large amounts of methanol. As discussed in Section 4.1,
20 neurological consequences of acute methanol intoxication in humans include Parkinson-like
21 responses, visual impairment, confusion, headache, and numerous subjective symptoms. The
22 occurrence of these symptoms has been shown to be associated with necrosis of the putamen
23 when neuroimaging techniques have been applied (Salzman, 2006). Such profound changes
24 have been linked to tissue acidosis that arises when methanol is metabolized to formaldehyde and
25 formic acid through the actions of ADH1 and ADH3. However, the well-documented impact of
26 the substantial amounts of formate that are formed when humans and animals are exposed to
27 large amounts of methanol may obscure the potentially harmful effects that may arise when
28 humans and animals exposed to smaller amounts. Human acute exposure studies (Chuwers et al.,
29 1995: M. R. Cooketal., 1991) (See Section 4.1.3) at TLV levels of 200 ppm would indicate that
30 some measures of neurological function (e.g., sensory evoked potentials, memory testing and
31 psychomotor testing) were impaired in the absence of measureable formate production.
March 2011 4-46 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 & 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 of
12 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 not
16 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 FAD40 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.41 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 levels
40 Along with the FAD diet, 1% succinylsulphathiazole was also given to inhibit folic acid biosynthesis from
intestinal bacteria.
41 Assuming that Wistar rat drinking water consumption is 60 mL/kg-day (V. V. Rogers et al.. 2002X 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.
March 2011 4-47 DRAFT - DO NOT CITE OR QUOTE
-------
1 were determined at PND21 and neurobehavioral parameters (motor performance using the
2 spontaneous locomotor activity test and cognitive performance using the conditioned avoidance
3 response [CAR] test), and neurochemical parameters (dopaminergic and cholinergic receptor
4 binding and dopamine levels) were measured at PND45. The expression of growth-associated
5 protein (GAP 43), a neuro-specific protein in the hippocampus that is primarily localized in
6 growth cone membranes and is expressed during developmental regenerative neurite outgrowth,
7 was examined using immunohistochemistry and western blot analysis.
8 A loss in body weight gain was observed at PND7, PND14, and PND21 in animals
9 exposed to 2% (11,15 and 19% weight gain reduction) and 4% (17, 24 and 29% weight gain
10 reduction) methanol in the FAD group and only at 4% (9, 14 and 17% weight gain reduction)
11 methanol in the FAS group. No significant differences in food and water intake were observed
12 among the different treatment groups. Liver folate levels in the FAD group were decreased by
13 63% in rats prior to mating and 67% in pups on PND21.
14 Based on reports of Parkinson-like symptoms in survivors of severe methanol poisoning
15 (see Section 4.1), Aziz et al. (2002) hypothesized that methanol may cause a depletion in
16 dopamine levels and degeneration of the dopaminergic nigrostriatal pathway.42 Consistent with
17 this hypothesis, they found dopamine levels were significantly decreased (32% and 51%) in the
18 striatum of rats in the FAD group treated with 2% and 4% methanol, respectively. In the FAS
19 group, a significant decrease (32%) was observed in the 4% methanol-exposed group.
20 Methanol treatment at 2% and 4% was associated with significant increases in activity, in
21 the form of distance traveled in a spontaneous locomotor activity test, in the FAS group (13%
22 and 39%, respectively) and more notably, in the FAD group (33% and 66%, respectively) when
23 compared to their respective controls. Aziz et al. (2002) suggest that these alterations in
24 locomotor activity may be caused by a significant alteration in dopamine receptors and disruption
25 in neurotransmitter availability. Dopamine receptor (D2) binding in the hippocampus of the FAD
26 group was significantly increased (34%) at 1% methanol, but was significantly decreased at 2%
27 and 4% methanol exposure by 20% and 42%, respectively. In the FAS group, D2 binding was
28 significantly increased by 22% and 54% in the 2% and 4% methanol-exposed 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
42 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 (KimetaL 2003).
March 2011 4-48 DRAFT - DO NOT CITE OR QUOTE
-------
1 the methanol-exposed FAD pups may be due to alterations in the number of cholinergic
2 (muscarinic) receptor proteins in the hippocampal region of the brain. Muscarinic receptor
3 binding was significantly increased in the 2% (20%) and 4% (42%) methanol-exposed group in
4 FAD animals, while FAS group animals had a significant increase in cholinergic binding only in
5 the 4% methanol exposed group (21%). High concentrations of methanol may saturate the
6 body's ability to remove toxic metabolites, including formaldehyde and formate, and this may be
7 exacerbated in FAD pups having a low store of folate.
8 Immunohistochemistry showed an increase in the expression of GAP-43 protein in the
9 dentate granular and pyramidal cells of the hippocampus in 2% and 4% methanol-exposed
10 animals in the FAD group. The FAS group showed increased expression only in the 4%
11 methanol-exposed group. The Western blot analysis also confirmed a higher expression of
12 GAP-43 in the 2% and 4% methanol-exposed FAD group rats. Aziz et al. (2002) suggested that
13 up-regulation of GAP-43 in the hippocampal region may be associated with axonal growth or
14 protection of the nervous system from methanol toxicity.
15 The Aziz et al. (2002) study provides evidence that hepatic tetrahydrofolate is an
16 important contributing factor in methanol-induced developmental neurotoxicity in rodents. The
17 immature blood-brain barrier and inefficient drug-metabolizing enzyme system make the
18 developing brain a particularly sensitive target organ to the effects of methanol exposure.
4.4.2. Inhalation Studies
19 A review by Carson et al. (1981) has summarized a number of older reports of studies on
20 the toxicological consequences of methanol exposure. In one example relevant to neurotoxicity,
21 the review cites a research report of Chen-Tsi (1959) who exposed 10 albino rats/group (sex and
22 strain unstated) to 1.77 and 50 mg/m3 (1.44 and 40.7 ppm) methanol vapor, 12 hours/day, for
23 3 months. Deformation of dendrites, especially the dendrites of pyramidal cells, in the cerebral
24 cortex was included in the description of histopathological changes observed in adult animals
25 following exposure to 50 mg/m3 (40.7 ppm) methanol vapor. One out often animals exposed to
26 the lower methanol concentration also displayed this feature.
27 Information on the neurotoxicity of methanol inhalation exposure in adult monkeys
28 (M. fascicularis) has come from NEDO (1987) which describes the results of a number of
29 experiments. The study included an acute study, a chronic study monkeys, and a repeated
30 exposure experiment (of variable duration depending upon exposure level), followed by recovery
31 period (1-6 months), and an experiment looking at chronic formaldehyde exposure (1 or 5 ppm),
32 a combustion product of methanol. This last experiment was apparently only a pilot and included
33 only one monkey per exposure condition.
March 2011 4-49 DRAFT - DO NOT CITE OR QUOTE
-------
1 In the chronic experiment 8 monkeys were included per exposure level (control, 10, 100,
2 1,000 ppm or 13, 131, and 1,310 mg/m3, respectively, for 21 hours/day); however, animals were
3 serially sacrificed at 3 time points: 7 months, 19 months, or >26 months. This design reduced the
4 number of monkeys at each exposure level to 2 subjects at 7 months and 3 subjects at the
5 subsequent time points (see Section 4.2.2). One of the 3 animals receiving 100 ppm methanol
6 and scheduled for sacrifice at 29 months was terminated at 26 months.
7 Histopathologically, no overt degeneration of the retina, optical nerve, cerebral cortex, or
8 other potential target organs (liver and kidney) was reported in the chronic experiment.
9 Regarding the peripheral nervous system, 1/3 monkeys exposed to 100 ppm (131 mg/m3) and 2/3
10 exposed to 1,000 ppm (1,310 mg/m3) for 29 months showed slight but clear changes in the
11 peroneal nerves. There was limited evidence of CNS degeneration inside the nucleus of the
12 thalamus of the brain at exposure to 100 ppm (131 mg/m3) or 1,000 ppm (1,310 mg/m3) for
13 7 months or longer. Abnormal appearance of stellate cells (presumed astroglia) within the
14 cerebral white matter was also observed in a high proportion (7/8 in both mid- and high-exposure
15 groups) of monkeys exposed to 100 ppm and 1,000 ppm for 7 months or more. All monkeys that
16 had degeneration of the inside nucleus of the thalamus also had degeneration of the cerebral
17 white matter. According to NEDO (1987), the stellate cell response was transient and "not
18 characteristic of degeneration." The authors also noted that the stellate cell response was "nearly
19 absent in normal monkeys in the control group" and "in the groups exposed to a large quantity of
20 methanol or for a long time their presence tended to become permanent, so a relation to the long
21 term over which the methanol was inhaled is suspected." However, all control group responses
22 are reported in a single table in the section of the NEDO (1987) report that describes the acute
23 monkey study, with no indication as to when the control group was sacrificed.
24 In the recovery experiment, monkeys were exposed to 1,000, 2,000, 3,000, or 5,000 ppm
25 methanol, followed by recovery periods of various duration. Monkeys exposed to 3,000 ppm for
26 20 days followed by a 6-month recovery period experienced relatively severe fibrosis of
27 responsive stellate cells and lucidation of the medullary sheath. However, resolution of some of
28 the glial responses was noted in the longer duration at lower exposure levels, with no effects
29 observed on the cerebral white matter in monkeys exposed for 7 months to 1,000 ppm methanol
30 followed by a 6-month recovery period. In general, the results from the recovery experiment
31 corroborated results observed in the chronic experiment. NEDO (1987) interpreted the lack of
32 glial effects after a 6-month recovery as an indication of a transient effect. The authors failed to
33 recognize that glial responses to neural damage do not necessarily persist following resolution of
34 neurodegeneration (Aschner & Kimelberg, 1996).
35 The limited information available from the NEDO (1987) summary report suggests that
36 100 ppm (131 mg/m3) may be an effect level following continuous, chronic exposure to
37 methanol. However, the current report does not indicate a robust dose response for the
March 2011 4-50 DRAFT - DO NOT CITE OR QUOTE
-------
1 neurodegenerative changes in the thalamus and glial changes in the white matter. The number of
2 animals at each exposure level for each serial sacrifice also limits statistical power
3 (2-3 monkeys/time point/exposure level). Confidence in this study is also weakened by the lack
4 of documentation for a concurrent control group.
5 Weiss et al. (1996) exposed 4 cohorts of pregnant Long-Evans rats (10-12 dams/
6 treatment group/cohort) to 0 or 4,500 ppm (0 and 5,897 mg/m3) methanol vapor (high-
7 performance liquid chromatography [HPLC] grade), 6 hours/day, from GD6 to PND21. Pups
8 were exposed together with the dams during the postnatal period. Average blood methanol levels
9 in pups on PND7 and PND14 were about twice the level observed in dams. However, methanol
10 exposure had no effect on maternal gestational weight gain, litter size, or postnatal pup weight
11 gain up to PND1843. Neurobehavioral tests were conducted in neonatal and adult offspring; the
12 data generated from those tests were evaluated by repeated measures ANOVA. Three
13 neurobehavioral tests conducted in 13-26 neonates/group included a suckling test, conditioned
14 olfactory aversion test, and motor activity test. In contrast to earlier test results reported by
15 Infurna and Weiss (1986), methanol exposure had no effect on suckling and olfactory aversion
16 tests conducted on PND5 and PND10, respectively. Results of motor activity tests in the
17 methanol group were inconsistent, with decreased activity on PND18 and increased activity on
18 PND25. Tests that measured motor function, operant behavior, and cognitive function were
19 conducted in 8-13 adult offspring/group. Some small performance differences were observed
20 between control and treated adult rats in the fixed wheel running test only when findings were
21 evaluated separately by sex and cohort. The test requires the adult rats to run in a wheel and
22 rotate it a certain amount of times in order to receive a food reward. A stochastic spatial
23 discrimination test examined the rats' ability to learn patterns of sequential responses. Methanol
24 exposure had no effect on their ability to learn the first pattern of sequential responses, but
25 methanol-treated rats did not perform as well on the reversal test. The result indicated possible
26 subtle cognitive deficits as a result of methanol exposure. A morphological examination of
27 offspring brains conducted on PND1 and PND21 indicated that methanol exposure had no effect
28 on neuronal migration, numbers of apoptotic cells in the cortex or germinal zones, or
29 myelination. However, neural cell adhesion molecule (NCAM) 140 andNCAM 180 gene
30 expression in treated rats was reduced on PND4 but not 15 months after the last exposure.
31 NCAMs are glycoproteins required for neuron migration, axonal outgrowth, and establishing
32 mature neuronal function patterns.
43 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.
March 2011 4-51 DRAFT - DO NOT CITE OR QUOTE
-------
1 Stanton et al. (1995) exposed 6-7 pregnant female Long-Evans rats/group to 0 or
2 15,000 ppm (0 and 19,656 mg/m3) methanol vapors (> 99.9% purity) for 7 hours/day on
3 GD7-GD19. Mean serum methanol levels at the end of the 1st, 4th, 8th, and 12th days of
4 exposure were 3,836, 3,764, 3,563, and 3,169 |ig/mL, respectively. As calculated by authors,
5 dams received an estimated methanol dose of 6,100 mg/kg-day. A lower body weight on the first
6 2 days of exposure was the only maternal effect; there was no increase in postimplantation loss.
7 Dams were allowed to deliver and nurse litters. Parameters evaluated in pups included mortality,
8 growth, pubertal development, and neurobehavioral function. Examinations of pups revealed
9 that two pups from the same methanol-exposed litter were missing one eye; aberrant visually
10 evoked potentials were observed in those pups. A modest but significant reduction in body
11 weight gain on PND1, PND21, and PND35 was noted in pups from the methanol group. For
12 example, by PND35, male pups of dams exposed to methanol had a mean body weight of
13 129 grams versus 139 grams in controls (p < 0.01). However, postnatal mortality was unaffected
14 by exposure to methanol. The study authors did not consider a 1.7-day delay in vaginal opening
15 in the methanol group to be an adverse effect. Preputial separation was not affected by prenatal
16 methanol exposure. Neurobehavioral status was evaluated using 8 different tests on specific days
17 up to PND160. Tests included motor activity on PND13-PND21, PND30, and PND60, olfactory
18 learning and retention on PND18 and PND25, behavioral thermoregulation on PND20-21, T-
19 maze delayed alternation learning on PND23-PND24, acoustic startle reflex on PND24, reflex
20 modification audiometry on PND61-PND63, passive avoidance on PND73, and visual evoked
21 potentials on PND160. A single pup/sex/litter was examined in most tests, and some animals
22 were subjected to multiple tests. The statistical significance of neurobehavioral testing was
23 assessed by one-way ANOVA, using the litter as the statistical unit. Results of the
24 neurobehavioral testing indicated that methanol exposure had no effect on the sensory, motor, or
25 cognitive function of offspring under the conditions of the experiment. However, given the
26 comparatively small number of animals tested for each response, it is uncertain whether the
27 statistical design had sufficient power to detect small compound-related changes.
28 NEDO (1987) sponsored a teratology study that included an evaluation of postnatal
29 effects in addition to standard prenatal endpoints in Sprague-Dawley rats. Thirty-six pregnant
30 females/group were exposed to 0, 200, 1,000, or 5,000 ppm (0, 262, 1,310, and 6,552 mg/m3)
31 methanol vapors (reagent grade) on GD7-GD17 for 22.7 hours/day. Statistical significance of
32 results was evaluated by t-test, Mann-Whitney U test, Fisher's exact test, and/or Armitage's ^
33 test.
34 Postnatal effects of methanol inhalation were evaluated in the remaining 12 dams/group
35 that were permitted to deliver and nurse their litters. Effects were only observed in the
March 2011 4-52 DRAFT - DO NOT CITE OR QUOTE
-------
1 5,000 ppm. There were no adverse effects on offspring body weight from methanol exposure.
2 However, the weights of some organs (brain, thyroid, thymus, and testes) were reduced in
3 8-week-old offspring following prenatal-only exposure to 5,000 ppm methanol. An unspecified
4 number of offspring were subjected to neurobehavioral testing or necropsy, but results were
5 incompletely reported.
6 NEDO (1987) also contains an account of a two-generation reproductive study that
7 evaluated the effects of pre- and postnatal methanol (reagent grade) exposure (20 hours/day) on
8 reproductive and other organ systems of Sprague-Dawley rats and in particular the brain. The F0
9 generation (30 males and 30 females per exposure group)44 was exposed to 0, 10, 100, and
10 1,000 ppm (0, 13.1, 131, and 1,310 mg/m3) from 8 weeks old to the end of mating (males) or to
11 the end of lactation period (females). The FI generation was exposed to the same concentrations
12 from birth to the end of mating (males) or to weaning of F2 pups 21 days after delivery (females).
13 Males and females of the F2 generation were exposed from birth to 21 days old (1
14 animal/sex/litter was exposed to 8 weeks of age). NEDO (1987) noted reduced brain, pituitary,
15 and thymus weights, in the offspring of F0 and FI rats exposed to 1,000 ppm methanol. As
16 discussed in the report, NEDO (1987) sought to confirm the possible compound-related effect of
17 methanol on the brain by carrying out an additional study in which Sprague-Dawley rats were
18 exposed to 0, 500, 1,000, and 2,000 ppm (0, 655, 1,310, and 2,620 mg/m3) methanol from the
19 first day of gestation through the FI generation. Brain weights were measured in 10-14
20 offspring/sex/group at 3, 6, and 8 weeks of age. As illustrated in Table 4-10, brain weights were
21 significantly reduced in 3-week-old males and females exposed to > 1,000 ppm. At 6 and 8
22 weeks of age, brain weights were significantly reduced in males exposed to > 1,000 ppm and
23 females exposed to 2,000 ppm. Due to the toxicological significance of this postnatal effect and
24 the fact that it has not been measured or reported elsewhere in the peer-reviewed methanol
25 literature, the brain weight changes observed by NEDO (1987) following gestational and
26 postnatal exposures and following gestation-only exposure (in the teratology study discussed
27 above) are evaluated quantitatively and discussed in more detail in Section 5 of this review.
44 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."
March 2011 4-53 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-10. Brain weights of rats exposed to methanol vapors during
gestation and lactation
Offspring
age
3wka
6wka
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%)
^Exposed throughout gestation and FI generation.
bExposed 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, Shen et al. (1999) and Burbacher, Grant et al. (1999) carried out toxicokinetic,
2 reproductive, developmental and postnatal neurological and neurobehavioral studies of methanol
3 in M. fascicularis monkeys that were published by HEI in a two-part monograph. Some of the
4 data were subsequently published in the open scientific literature (Burbacher, Grant, et al., 2004;
5 Burbacher, Shen, et al., 2004). The experimental protocol featured exposure to 2 cohorts of 12
6 monkeys/group to low-exposure levels (relative to the previously discussed rodent studies) of 0,
7 200, 600, or 1,800 ppm (0, 262, 786, and 2,359 mg/m3) methanol vapors (99.9% purity),
8 2.5 hours/day, 7 days/week, during a premating period and mating period (-180 days combined)
9 and throughout the entire gestation period (-168 days). The monkeys were 5.5-13 years old and
10 were a mixture of feral-born and colony-bred animals. The outcome study included an evaluation
11 of maternal reproductive performance (discussed in Section 4.3.2) and tests to assess infant
12 postnatal growth and newborn health, neurological outcomes included reflexes, behavior, and
13 development of visual, sensorimotor, cognitive, and social behavioral function. Blood methanol
14 levels, elimination, and the appearance of formate were also examined and are discussed in
March 2011
4-54 DRAFT - DO NOT CITE OR QUOTE
-------
1 Section 3.2. The effects observed were in the absence of appreciable increases in maternal blood
2 formate levels.
3 Neurobehavioral function was assessed in 8-9 infants/group during the first 9 months of
4 life (Burbacher. Grant et al.. 1999: Burbacher. Grant et al.. 2004). Although results in 7/9 tests
5 were negative, 2 effects were possibly related to methanol exposure. The Visually Directed
6 Reaching (VDR) test is a measure of sensorimotor development and assessed the infants' ability
7 to grasp for a brightly colored object containing an applesauce-covered nipple. Beginning at 2
8 weeks after birth, infants were tested 5 times/day, 4 days/week. Performance on this test,
9 measured as age from birth at achievement of test criterion (successful object retrieval on 8/10
10 consecutive trials over 2 testing sessions), was reduced in all treated male infants. The times
11 (days after birth) to achieve the criteria for the VDR test were 23.7 ± 4.8 (n = 3), 32.4 ± 4.1 (n =
12 5), 42.7 ± 8.0 (n = 3), and 40.5 ± 12.5 (n = 2) days for males and 34.2 ± 1.8 (n = 5), 33.0 ± 2.9 (n
13 = 4), 27.6 ± 2.7 (n = 5), and 40.0 ± 4.0 (n = 7) days for females in the control to 1,800 ppm
14 groups, respectively. Statistical significance was obtained in the 1,800 ppm group when males
15 and females were evaluated together (p = 0.04) and in the 600 ppm (p = 0.007) for males only.
16 However, there was no significant difference between responses and/or variances among the dose
17 levels for males and females combined (p = 0.244), for males only (p = 0.321) and for males
18 only, excluding the high-dose group (p = 0.182). Yet there was a significant dose-response trend
19 for females only (p = 0.0265). The extent to which VDR delays were due to a direct effect of
20 methanol on neurological development or a secondary effect due to the methanol-induced
21 decrease in length of pregnancy and subsequent prematurity is not clear. Studies of reaching
22 behavior have shown that early motor development in pre-term human infants without major
23 developmental disorders differs from that of full-term infants (Fallang et al., 2003). Clinical
24 studies have indicated that the quality of reaching and grasping behavior in pre-term infants is
25 generally less than that in full-term infants (Fallang et al., 2003; Plantinga et al., 1997). For this
26 reason, measures of human infant development generally involve adjustment of a child's "test
27 age" if he or she had a gestational age of fewer than 38 weeks, often by subtracting weeks
28 premature from the age measured from birth (Wilson & Cradock, 2004). When this type of
29 adjustment is made to the Burbacher et al. (1999; 2004) VDR data, the dose-response trend for
30 males only becomes worse (p = 0.448) and the dose-response trend for the females only is
31 improved (p = 0.009), though the variance in the data could not be modeled adequately. Thus,
32 only the unadjusted VDR response for females only exhibited a dose response that could be
33 adequately modeled for the purposes of this assessment (see Appendix C).
34 At 190-210 days of age, the Fagan Test of infant intelligence was conducted. The
35 paradigm makes use of the infant's proclivity to direct more visual attention to novel stimuli
March 2011 4-55 DRAFT - DO NOT CITE OR QUOTE
-------
1 rather than familiar stimuli. The test measures the time infants spend looking at familiar versus
2 novel items. Deficits in the Fagan task can qualitatively predict deficits in intelligence quotient
3 (IQ) measurements assessed in children at later ages (Fagan & Singer, 1983). Control monkey
4 infants in the Burbacher et al. (1999; 2004) study spent more than 62% ± 4% (mean for both
5 cohorts) of their time looking at novel versus familiar monkey faces, while none of the treated
6 monkeys displayed a preference for the novel faces (59% ± 2%, 54% ± 2% and 59% ± 2% in
7 200, 600 and 1,800 ppm groups, respectively). Unlike the VDR results discussed previously,
8 results of this test did not appear to be gender specific and were neither statistically significant
9 (ANOVA p = 0.38) nor related to exposure concentration. The findings indicated a cohort effect
10 which appeared to reduce the statistical power of this analysis. The authors' exploratory analysis
11 of differences in outcomes between the 2 cohorts indicated an effect of exposure in the second
12 cohort and not the first cohort due to higher mean performance in controls of cohort 2 (70% +
13 5% versus 55% ± 4% for cohort 1). In addition, this latter finding could reflect the inherent
14 constraints of this endpoint. If the control group performs at the 60% level and the most
15 impaired subjects perform at approximately the 50% chance level (worse than chance
16 performance would not be expected), the range over which a concentration-response relationship
17 can be expressed is limited. Because of the longer latency between assessment and birth, these
18 results would not be confounded with the postulated methanol-induced decrease in gestation
19 length of the exposed groups of this study. Negative results were obtained for the remaining
20 seven tests that evaluated early reflexes, gross motor development, spatial and concept learning
21 and memory, and social behavior. Infant growth and tooth eruption were unaffected by methanol
22 exposure.
4.4.3. Studies Employing In Vitro, S.C. and I.P. Exposures
23 There is some experimental evidence that the presence of methanol can affect the activity
24 of acetylcholinesterase (Tsakiris et al., 2006). Although these experiments were carried out on
25 erythrocyte membranes in vitro, the apparent compound-related changes may have implications
26 for possible impacts of methanol and/or its metabolites on acetylcholinesterase at other centers,
27 such as the brain. Tsakiris et al. (2006) prepared erythrocyte ghosts from blood samples of
28 healthy human volunteers by repeated freezing-thawing. The ghosts were incubated for 1 hour at
29 37°C in 0, 0.07, 0.14, 0.6 or 0.8 mmol/L methanol and the specific activities of
30 acetylcholinesterase monitored. Respective values (in change of optical density units/minute-mg
31 protein) were 3.11 ± 0.15, 2.90 ± 0.10, 2.41 ± 0.10 (p < 0.05), 2.05 ±Q.ll(p< 0.01), and 1.81 ±
32 0.09 (p < 0.001). More recently, Simintzi et al. (2007) carried out an in vitro experiment to
33 investigate the effects of aspartame metabolites, including methanol, on 1) a pure preparation of
March 2011 4-56 DRAFT - DO NOT CITE OR QUOTE
-------
1 acetylcholinesterase, and 2) the same activity in homogenates of frontal cortex prepared from the
2 brains of (both sexes of) Wistar rats. The activities were measured after incubations with 0, 0.14,
3 0.60, or 0.8 mmoles/L (0, 4.5, 19.2, and 25.6 mg/L) methanol, and with methanol mixed with the
4 other components of aspartame metabolism, phenylalanine and aspartic acid. After incubation at
5 37°C for 1 hour, the activity of acetylcholinesterase was measured spectrophotometrically. As
6 shown in Table 4-11, the activities of the acetylcholinesterase preparations were reduced dose
7 dependency after incubation in methanol. Similar results were also obtained with the other
8 aspartame metabolites, aspartic acid, and phenylalanine, both individually or as a mixture with
9 methanol. While the implications of this result to the acute neurotoxicity of methanol are
10 uncertain, the authors speculated that methanol may bring about these changes through either
11 interactions with the lipids of rat frontal cortex or perturbation of proteinaceous components.
Table 4-11. 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.
ap<0.01.
V< 0.001.
Source: Simintzi et al. (2007).
12 In another experiment of relevance to neurotoxicity, the impact of repeat methanol
13 exposure on amino acid and neurotransmitter expression in the retina, optic nerve, and brain was
14 examined by Gonzalez-Quevedo et al. (2002). The goal of the study was to determine whether a
15 sustained increase in formate levels, at concentrations below those known to produce toxic
16 effects from acute exposures, can induce biochemical changes in the retina, optical nerve, or
17 certain regions of the brain. Male Sprague-Dawley rats (5-7/group; 100-150 g) were divided into
18 6 groups and treated for 4 weeks according to the following plan. Four groups of animals
19 received tap water ad libitum as drinking water for 1 week. During the second week, groups 1
20 and 2 (control and methanol respectively) received saline subcutaneously, (s.c.) and groups 3 and
March 2011
4-57 DRAFT - DO NOT CITE OR QUOTE
-------
1 4 (methotrexate45 [MTX] and methotrexate-methanol [MTX-methanol], respectively) received
2 MTX s.c. (0.2 mg/kg-day). During the 3rd week, MTX was reduced to 0.1 mg/kg and 20%
3 methanol (2g/kg-day) was given i.p. to groups 2 (methanol) and 4 (MTX-methanol). Groups 1
4 (control) and 3 (MTX) received equivalent volumes of saline administered i.p. The treatment
5 was continued until the end of the fourth week. Groups 5, (taurine46 [Tau]) and 6, (Tau-MTX-
6 methanol) received 2% Tau in their drinking water ad libitum during the first 4 weeks, after
7 which they were treated in the same manner as groups 1 and 4, respectively. Weights were
8 documented weekly on all animals. Blood for formate and amino acid determinations and biopsy
9 samples of retina, optic nerve, hippocampus, and posterior cortex of each animal were collected
10 at the end of the experiment. Formate levels were not affected by Tau alone or MTX alone.
11 While methanol alone increased blood formate levels, MTX-methanol, and Tau-MTX-methanol
12 produced a threefold increase in blood formate levels as compared to controls and a twofold
13 increase as compared to methanol alone. The amino acids aspartate, glutamate, asparagine,
14 serine, histidine, glutamine, threonine, glycine, arginine, alanine, hypotaurine,
15 gamma-aminobutyric acid (which is also a neurotransmitter), and tyrosine were measured in
16 blood, brain, and retinal regions.
17 None of the amino acids measured were altered in the blood of methanol-, MTX-, or
18 MTX-methanol-treated animals. Tau was increased in the blood of animals treated with taurine
19 in the drinking water (Tau and Tau-MTX-methanol) and histidine was increased in the Tau group
20 but not in the Tau-MTX-methanol group.
21 The levels of aspartate, Tau, glutamine, and glutamate were found to be altered by
22 treatment in various areas of the brain. Aspartate was increased in the optic nerve of animals
23 treated with MTX-methanol and Tau-MTX-methanol, indicating a possible relation to formate
24 accumulation. The authors note that L-aspartate is a major excitatory amino acid in the brain and
25 that increased levels of excitatory amino acids can trigger neuronal cell damage and death (Albin
26 & Greenamvre, 1992). Aspartate, glutamine and Tau were found to be increased with respect to
27 controls in the hippocampus of the three groups receiving methanol. Glutamate was significantly
28 increased in the hippocampus in the methanol and the Tau-MTX-methanol groups with respect to
29 controls, but no statistically significant difference was found in the MTX-methanol group when
30 compared to controls, methanol alone, or the Tau-MTX-methanol groups. The authors suggest
31 that increased levels of aspartate and glutamine in the hippocampus could provide an explanation
45 Methotrexate depletes folate stores (resulting in an increase in the formate levels of methanol exposed animals) by
interfering with tetrahydrofolate(THF) regeneration (DormanetaL 1994).
46 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-Quevado et al. 2002).
March 2011 4-58 DRAFT - DO NOT CITE OR QUOTE
-------
1 for some of the CNS symptoms observed in methanol poisonings on the basis of their observed
2 impact on cerebral arteries (Huang et al., 1994). The fact that these increases resulted primarily
3 from methanol without MTX is significant in that it indicates methanol can cause excitotoxic
4 effects without formate mediation. The treatments used did not produce any significant changes
5 in amino acid levels in the posterior cortex.
6 The neurotransmitters serotonin (5-HT) and dopamine (DA) and their respective
7 metabolites, 5-hydroxyindolacetic acid (5-HIAA) and dihydroxyphenylacetic acid (DOPAC),
8 were measured in the brain regions described. The levels of these monoamines were not affected
9 by formate accumulation, as the only increases were observed for 5-HT and 5-HIAA following
10 methanol-only exposure. 5-HT was increased in the retina and hippocampus of methanol-only
11 treated animals, and the metabolite 5-HIAA was increased in the hippocampus of methanol-only
12 treated animals; DA and DOPAC levels were not altered by the treatments in any of the areas
13 measured. The posterior cortex did not show any changes in monoamine levels for any treatment
14 group.
15 Rajamani et al. (2006) examined several oxidative stress parameters in male Wistar rats
16 following methotrexate-induced folate deficiency. Animals (6/group) were divided into 3e
17 groups: saline controls, methotrexate (MTX) controls, and MTX-methanol treated animals.
18 Animals in the MTX-only group were treated with 0.2 mg/kg-day MTX s.c. injection for 7 days
19 and following confirmation of folate deficiency, received either saline for MTX control and
20 saline controls or a single dose of 3 g/kg methanol (20% w/v in saline) i.p. on day 8. On the 9th
21 day, all animals were sacrificed and blood and tissue samples were collected. The optic nerve,
22 retina, and brain were collected and the brain was dissected into the following regions: cerebral
23 cortex, cerebellum, mid-brain, pons medulla, hippocampus and hypothalamus. Each region was
24 homogenized, then centrifuged at 300 x g for 15 minutes and the supernatant was examined for
25 indicators of oxidative stress including the free radical scavengers superoxide dismutase (SOD),
26 CAT, glutathione peroxidase, and reduced GSH levels. The levels of protein thiols, protein
27 carbonyls, and amount of lipid peroxidation were also measured. Compared to controls the
28 levels of SOD, CAT, GSH peroxidase, oxidized GSH, protein carbonyls and lipid peroxidation
29 were elevated in all of the brain regions where it was measured, with greater increases observed
30 in the MTX-methanol treated animals than in the MTX alone group. The level of GSH and
31 protein thiols was decreased in all regions of the brain, with a greater decrease observed in the
32 MTX-methanol-treated animals than MTX-treated animals. In addition, expression of HSP70, a
33 biomarker of cellular stress, was increased in the hippocampus. Overall, these results suggest that
34 methanol treatment of folate-deficient rats results in increased oxidative stress in the brain, retina
35 and optic nerve.
March 2011 4-59 DRAFT - DO NOT CITE OR QUOTE
-------
1 To determine the effects of methanol intoxication on the HPA axis, a combination of
2 oxidative stress, immune and neurobehavioral parameters were observed (Parthasarathy et al.,
3 2006a). Adult male Wistar albino rats (6 animals/group) were treated with either 0 or 2.37g/kg-
4 day methanol i.p. for 1, 15 or 30 days. Oxidative stress parameters examined included SOD,
5 CAT, GSH peroxidase, GSH, and ascorbic acid (Vitamin C). Plasma corticosterone levels were
6 measured, and lipid peroxidation was measured in the hypothalamus and the adrenal gland. An
7 assay for DNA fragmentation was conducted in tissue from the hypothalamus, the adrenal gland
8 and the spleen. Immune function tests conducted included the footpad thickness test for delayed
9 type hypersensitivity (DTH), a leukocyte migration inhibition assay, the hemagglutination assay
10 (measuring antibody titer), the neutrophil adherence test, phagocytosis index, and a nitroblue
11 tetrazolium (NET) reduction and adherence assay used to measure the killing ability of
12 polymorphonuclear leukocytes (PMNs). The open field behavior test was used to measure
13 general locomotor and explorative activity during methanol treatment in the 30-day treatment
14 group, with tests conducted on days 1, 4, 8, 12, 16, 20, 24, and 28. All enzymatic (SOD, CAT,
15 and GSH peroxidase) and nonenzymatic antioxidants (GSH and Vitamin C) were significantly
16 increased in the 1-day methanol-exposed group as compared to controls. However, with
17 increasing time of treatment, all of the measured parameters were significantly decreased when
18 compared with control animals. Lipid peroxidation was significantly increased in both the
19 hypothalamus and the adrenal gland at 1, 15, and 30 days, with the 30-day treated animals also
20 significantly increased when compared to the 15-day methanol-treated animals.
21 Leukocyte migration and antibody titer were both significantly increased over controls for
22 all time points, while footpad thickness was significantly decreased in 15- and 30-day treated
23 animals. Neutrophil adherence was significantly decreased after 1 and 30 days of exposure. A
24 significant decrease in the NET reduction and adherence was found when comparing PMNs from
25 the 30-day treated animals with cells from the 15-day methanol-treated group.
26 The open field behavior tests showed a significant decrease in ambulation from the 4th
27 day on and significant decreases in rearing and grooming from the 20th day on. A significant
28 increase was observed in immobilization from the 8th day on and in fecal bolus from the 24th
29 day on in methanol-exposed animals.
30 While corticosterone levels were significantly increased following 1 or 15 days of
31 methanol treatment, they were significantly decreased after 30 days of treatment, as compared to
32 controls. Following 30 days of methanol treatment, DNA from the hypothalamus, the adrenal
33 gland, and the spleen showed significant fragmentation. The authors conclude that exposure to
34 methanol-induced oxidative stress, disturbs HPA-axis function, altering corticosterone levels and
35 producing effects in several nonspecific and specific immune responses.
March 2011 4-60 DRAFT - DO NOT CITE OR QUOTE
-------
4.5. IMMUNOTOXICITY
1 Parthasarathy et al. (2005) provided data on the impact of methanol on neutrophil
2 function in an experiment in which 6 male Wistar rats/group were given a single i.p. exposure of
3 2,370 mg/kg methanol mixed 1:1 in saline. Another group of 6 animals provided blood samples
4 that were incubated with methanol in vitro at a methanol concentration equal to that observed in
5 the in vivo-treated animals 30 and 60 minutes postexposure. Total and differential leukocyte
6 counts were measured from these groups in comparison to in vivo and in vitro controls.
7 Neutrophil adhesion was determined by comparing the neutrophil index in the untreated blood
8 samples to those that had been passed down a nylon fiber column. The cells' phagocytic ability
9 was evaluated by their ability to take up heat-killed Candida albicam. In another experiment,
10 neutrophils were assessed for their killing potential by measuring their ability to take up then
11 convert NET to formazan crystals.47 One hundred neutrophils/slides were counted for their total
12 and relative percent formazan-positive cells.
13 The blood methanol concentrations 30 and 60 minutes after dosing were 2,356 ± 162 and
14 2,233 ±146 mg/L, respectively. The mean of these values was taken as the target concentration
15 for the in vitro methanol incubation. In the in vitro studies, there were no differences in total and
16 differential leukocyte counts, suggesting that no lysis of the cells had occurred at this methanol
17 concentration. This finding contrasts with the marked difference in total leukocytes observed as
18 a result of methanol incubation in vivo, in which, at 60 minutes after exposure, 16,000 ± 1,516
19 cells/mm3 were observed versus 23,350 ± 941 in controls (p < 0.001). Some differences in
20 neutrophil function were observed in blood samples treated with methanol in vitro and in vivo.
21 These differences are illustrated for the 60-minute postexposure samples in Table 4-12.
47 Absence of NET reduction indicates a defect in some of the metabolic pathways involved in intracellular microbial
killing.
March 2011 4-61 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-12. 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
Values are means ± S.D. for six animals.
><0.01.
V< 0.001.
Source: Parthasarathy et al. (2005).
1 Parthasarathy et al. (2005) observed differences in the neutrophil functions of cells
2 exposed to methanol in vitro versus in vivo, most notably in the phagocytic index that was
3 reduced in vitro but significantly increased in vivo. However, functions such as adherence and
4 NET reduction showed consistency in the in vitro and in vivo responses. The authors noted that,
5 by and large, the in vivo effects of methanol on neutrophil function were more marked than those
6 in cells exposed in vitro.
7 Another study by Parthasarathy et al. (2005) also exposed 6 male Wistar rats/group i.p. to
8 methanol at approximately 1/4 the LD50 (2.4 g/kg). The goal was to further monitor possible
9 methanol-induced alterations in the activity of isolated neutrophils and other immunological
10 parameters. The exposure protocol featured daily injections of methanol for up to 30 days in the
11 presence or absence of sheep RBCs. Blood samples were assessed for total and differential
12 leukocytes, and isolated neutrophils were monitored for changes in phagocytic and avidity
13 indices, NET reduction, and adherence. In the latter test, blood samples were incubated on a
14 nylon fiber column, then eluted from the column and rechecked for total and differential
15 leukocytes. Phagocytosis was monitored by incubating isolated buffy coats from the blood
16 samples with heat-killed C. albicam. NET reduction capacity examined the conversion of the
17 dye to formazan crystals within the cytoplasm. The relative percentage of formazan-positive
18 cells in each blood specimen gave a measure of methanol's capacity to bring about cell death.
19 As tabulated by the authors, there was a dose-dependent reduction in lymphoid organ weights
20 (spleen, thymus, and lymph node) in rats exposed to methanol for 15 and 30 days via i.p.
21 injection, irrespective of the presence of sheep RBCs. Methanol also appeared to result in a
22 reduction in the total or differential neutrophil count. These and potentially related changes to
23 neutrophil function are shown in Table 4-13.
March 2011
4-62 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-13. 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 ± 83 ^
121 ± W*-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 ±
2553*b
16±3a
79 ±5
18,633 ±2057
8±3
89 ±4
16,675 ± 1908
23±4a
78.5 ± 4a
14,067 ± 930*-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 ± 1.0^
29.5 ± 5.0^
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).
a/> < 0.05 from respective control.
bp < 0.05 between 15-and 30-day treatment groups.
Source: Parthasarathy et al. (2005).
1 The study provided data that showed altered neutrophil functions following repeated daily
2 exposures of rats to methanol for periods up to 30 days. This finding is indicative of a possible
3 effect of methanol on the immunocompetence of an exposed host.
4 Parthasarathy et al. (2006) reported on additional immune system indicators as part of a
5 study to determine the effects of methanol intoxication on the HPA axis. As described in
6 Section 4.4.3, immune function tests conducted included the footpad thickness test for DTK, a
7 leukocyte migration inhibition assay, the hemagglutination assay (measuring antibody titer), the
8 neutrophil adherence test, phagocytosis index, and a NET reduction and adherence assay used to
9 measure the killing ability of PMNs.
10 Leukocyte migration and antibody titer were both significantly increased over controls for
11 all time points, while footpad thickness was significantly deceased in 15- and 30-day treated
12 animals. Neutrophil adherence was significantly decreased after 1 and 30 days of exposure. A
March 2011
4-63 DRAFT - DO NOT CITE OR QUOTE
-------
1 significant decrease in the NET reduction and adherence was found when comparing PMNs from
2 the 30-day treated animals with cells from the 15-day methanol-treated group.
3 Parthasarathy et al. (2007) reported the effects of methanol on a number of specific
4 immune functions. As before, 6 male Wistar rats/group were treated with 2,370 mg/kg methanol
5 in a 1:1 mixture in saline administered intraperitoneally for 15 or 30 days. Animals
6 scheduled/designated for termination on day 15 were immunized intraperitoneally with 5 x 109
7 sheep RBCs on the 10th day. Animals scheduled for day 30 termination were immunized on the
8 25th day. Controls were animals that were not exposed to methanol but immunized with sheep
9 RBCs as described above. Blood samples were obtained from all animals at sacrifice and
10 lymphoid organs including the adrenals, spleen, thymus, lymph nodes, and bone marrow were
11 removed. Cell suspensions were counted and adjusted to 1 x 108 cells/mL. Cell-mediated
12 immune responses were assessed using a footpad thickness assay and a leucocyte migration
13 inhibition (LMI) test, while humoral immune responses were determined by a hemagglutination
14 assay, and by monitoring cell counts in spleen, thymus, lymph nodes, femoral bone marrow, and
15 in splenic lymphocyte subsets. Plasma levels of corticosterone were measured along with levels
16 of such cytokines as TNF-a, IFN-y, IL-2, and IL-4. DNA damage in splenocytes and thymocytes
17 was also monitored using the Comet assay.
18 Table 4-14 shows decreases in the animal weight/organ weight ratios for spleen, thymus,
19 lymph nodes and adrenal gland as a result of methanol exposure. However, the splenocyte,
20 thymocyte, lymph node, and bone marrow cell counts were time-dependently lower in methanol-
21 treated animals.
March 2011 4-64 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-14. 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
o.os±o.or
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 ±0.06^
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 4-15.
March 2011
4-65 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-15. 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 ± 68.8^
78.8±4.4a
585 ± 45a
950±59.6a
1
2
3
4
Values are means ± six animals.
ap < 0.05 versus control groups.
bp < 0.05 versus 15-day treated group.
Source: Parthasarathy et al. (2007).
Drawing on the results of DNA single strand breakage in this experiment, the authors
speculated that methanol-induced apoptosis could suppress specific immune functions such as
those examined in this research report. Methanol appeared to suppress both humoral and cell-
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
March 2011
4-66 DRAFT - DO NOT CITE OR QUOTE
-------
1
2
o
J
4
5
6
7
8
9
10
11
12
48 .
to 10,000 ppm methanol for 6 hours on GD8, with and without the administration of
fomepizole (4-methylpyrazole) to inhibit the metabolism of methanol by ADH1. Dams were
sacrificed on GD10, and fetuses were examined for neural tube patency. As shown in Table 4-16,
the incidence of fetuses with open neural tubes was significantly increased in the 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 4-16. 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
Values are means ± S.D.
ap < 0.05, as calculated by the authors.
Source: Dorman etal. (1995) (adapted).
13 The data in Table 4-16 suggest that the formate metabolite is not responsible for the
14 observed increase in open neural tubes in CD-I mice following methanol exposure. Formate
15 administered by gavage (750 mg/kg) did not increase this effect despite the fact that this formate
16 dose produced the same toxicokinetic profile as a 6-hour exposure to 10,000 ppm methanol
17 vapors (1.05 mM formate in maternal blood and 2.0 mmol formate/kg in decidual swellings).
18 However, the data are consistent with the hypotheses that the formaldehyde metabolite of
48 Dorman et al. (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.
March 2011
4-67 DRAFT - DO NOT CITE OR QUOTE
-------
1 methanol may play a role. Both CAT and ADH1 activity are immature at days past conception
2 (DPC)8 (Table 4-17). If fetal ADH1 is more mature than fetal CAT, it is conceivable that the
3 decrease in the open neural tube response observed for methanol combined with fomepizole
4 (Table 4-16) may be due to fomepizole having a greater effect on the metabolism of fetal
5 methanol to formaldehyde than is observed in adult rats. Unfortunately, the toxicity studies were
6 carried out during a period of development where ADH1 expression and activity are just starting
7 to develop (Table 4-17); therefore, it is uncertain whether any ADH1 was present in the fetus to
8 be inhibited by fomepizole.
Table 4-17. 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
+
-
3
N/A
+
+
""Activity of CAT and ADH1 are expressed as nmol/minute/mg and pmol/minute/mg, respectively.
Source: Harris et al. (2003).
9 Dorman et al. (1995) provide additional support for their hypothesis that methanol's
10 developmental effects in CD-I mice are not caused by formate in an in vitro study involving the
11 incubation of GD8 whole CD-I mouse embryos with increasing concentrations of methanol or
12 formate. Developmental anomalies were observed on GD9, including cephalic dysraphism,
13 asymmetry and hypoplasia of the prosencephalon, reductions of brachial arches I and n, scoliosis,
March 2011
4-68 DRAFT - DO NOT CITE OR QUOTE
-------
1 vesicles on the walls of the mesencephalon, and hydropericardium (Table 4-18). The
2 concentrations of methanol used for embryo incubation (0-375 mM) were chosen to be broadly
3 equivalent to the peak methanol levels in plasma that have been observed (approximately
4 100 mM) after a single 6-hour inhalation exposure to 10,000 ppm (13,104 mg/m3). As discussed
5 above, these exposure conditions induced an increased incidence of open neural tubes on GD10
6 embryos when pregnant female CD-I mice were exposed on GD8. (Table 4-16). Embryonic
7 lesions such as cephalic dysraphism, prosencephalic lesions, and brachial arch hypoplasia were
8 observed with 250 mM (8,000 mg/L) methanol and 40 mM (1,840 mg/L) formate. The study
9 authors noted that a formate concentration of 40 mM (1,840 mg/L) greatly exceeds blood formate
10 levels in mice inhaling 15,000 ppm methanol (0.75 mM = 35 mg/L), a teratogenic dose.
Table 4-18. Dysmorphogenic effect of methanol and formate in neurulating
CD-I mouse embryos in culture (GD8)
Treat-ment
Vehicle
Methanol
Formate
Concen-tration
(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. abnor-
mal
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
Mode-rate
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
Hypo-plasia
2
1
2
3
T
9a
2
4
1
2
3
Asym-metry
0
0
2
1
1
1
0
2
2
1
5a
Total
2
1
4
4
8
10a
2
6
o
3
o
3
8
Bra-chial
arch- hypo-
plasia
0
0
1
1
6a
8a
1
0
0
1
13a
zp < 0.05, as calculated by the authors.
Source: Dormanetal. (1995) (adapted).
11 As discussed in Section 4.3.3, a series of studies by Harris et al. (2004; 2003) also
12 provide evidence as to the moieties that may be responsible for methanol-induced developmental
13 toxicity. Harris et al. (2004) have shown that among methanol and its metabolites, viability of
14 cultured rodent embryos is most affected by formate. In contrast, teratogenic endpoints (of
15 interest to this risk assessment) in cultured rodent embryos are more sensitive to methanol and
16 formaldehyde than formate. Data from these studies indicate that developmental toxicity may be
17 more related to formaldehyde than methanol, as formaldehyde-induced teratogenicity occurs at
March 2011
4-69 DRAFT - DO NOT CITE OR QUOTE
-------
1 several orders of magnitude lower than methanol (Table 4-10) (Hansen et al., 2005; Harris et al.,
2 2004). It should also be noted that CAT, ADH1, and ADH3 activities are present in both the rat
3 embryo and VYS at stages as early as 6-12 somites (Harris et al., 2003): thus, it is presumable
4 that in these ex vivo studies methanol is metabolized to formaldehyde and formaldehyde is
5 subsequently metabolized to S-formylglutathione.
6 Studies involving GSH also lend support that formaldehyde may be a key proximal
7 teratogen. Inhibition of GSH synthesis with butathione sulfoximine (BSO) has little effect on
8 developmental toxicity endpoints, yet treatment with BSO and methanol or formaldehyde
9 increases developmental toxicity (Harris et al., 2004). Among the enzymes involved in methanol
10 metabolism, only ADH3-mediated metabolism of formaldehyde is GSH dependent. This
11 hypothesis that ADH3-mediated metabolism of formaldehyde is important for the amelioration of
12 methanol's developmental toxicity is also supported by the diminished ADH3 activity in the
13 mouse versus rat embryos, which is consistent with the greater sensitivity of the mouse to
14 methanol developmental toxicity (Harris et al., 2003) (Section 4.3.3). Similarly reasonable
15 explanations for this greater mouse sensitivity are not readily apparent for the two MO As
16 described below that attribute methanol toxicity to methanol metabolism per se, either through
17 the depletion of folate (Section 4.6.2) or the generation of reactive oxidant species (Section
18 4.6.3). Mouse livers actually have considerably higher hepatic tetrahydrofolate and total folate
19 than rat or monkey liver. Harris et al. (2003) and Johlin et al. (1987) have shown that CAT
20 activity in the embryo and VYS of rats and mice appear similar..
21 Without positive identification of the actual moiety responsible for methanol-induced
22 teratogenicity, MOA remains unclear. If the moiety is methanol, then it is possible that
23 generation of NADH during methanol oxidation creates an imbalance in other enzymatic
24 reactions. Studies have shown that ethanol intake leads to a > 100-fold increase in cellular
25 NADH, presumably due to ADH1 -mediated reduction of the cofactor NAD+ to NADH
26 (Cronholm, 1987; M. E. Smith & Newman, 1959). This is of potential importance because, for
27 example, ethanol intake has been shown to increase the in vivo and in vitro enzymatic reduction
28 of other endogenous compounds (e.g., serotonin) in humans (Davis etal., 1967; Svensson et al.,
29 1999). In rodents, CAT-mediated methanol metabolism may obviate this effect; in humans,
30 however, methanol is primarily metabolized by ADH1.
31 If the teratogenic moiety of methanol is formaldehyde, then reactivity with protein
32 sulfhydryls and nonprotein sulfhydryls (e.g., GSH) or DNA protein cross-links may be involved.
33 Metabolic roles ascribed to ADH3, particularly regulation of S-nitrosothiol biology (Foster &
34 Stamler, 2004), could also be involved in the MOA. Recently, Staab et al. (2008) have shown
35 that formaldehyde alters other ADH3-mediated reactions through cofactor recycling and that
March 2011 4-70 DRAFT - DO NOT CITE OR QUOTE
-------
1 formaldehyde alters levels of cellular S-nitrosothiol, which plays a key role in cellular signaling
2 and many cellular functions and pathways (Hess et al., 2005).
3 Studies such as those by Harris et al. (2004; 2003) and Dorman et al. (1995) suggest that
4 formate is not the metabolite responsible for methanol's teratogenic effects. The former
5 researchers suggest that formaldehyde is the proximate teratogen, and provide evidence in
6 support of that hypothesis. However, questions remain. Researchers in this area have not yet
7 reported using a sufficient array of enzyme inhibitors to conclusively identify formaldehyde as
8 the proximate teratogen. Studies involving other inhibitors or toxicity studies carried out in
9 genetically engineered mice, while not devoid of confounders, might further inform regarding the
10 methanol MOA for developmental toxicity. Even if formaldehyde is ultimately identified as the
11 proximate teratogen, methanol would likely play a prominent role, at least in terms of transport to
12 the target tissue. The high reactivity of formaldehyde would limit its unbound and unaltered
13 transport as free formaldehyde from maternal to fetal blood (Thrasher & Kilburn, 2001), and, as
14 has been discussed, the capacity for the metabolism of methanol to formaldehyde is likely lower
15 in the fetus and neonate versus adults (Section 3.3).
4.6.2. Role of Folate Deficiency in the Developmental Toxicity of Methanol
16 As discussed in Sections 3.1 and 4.1, humans and other primates are susceptible to the
17 effects of methanol exposure associated with formate accumulation because they have lower
18 levels of hepatic tetrahydrofolate-dependent enzymes that help in formate oxidation.
19 Tetrahydrofolate-dependent enzymes and critical pathways that depend on folate, such as purine
20 and pyrimidine synthesis, may also play a role in the developmental toxicity of methanol.
21 Studies of rats and mice fed folate-deficient diets have identified adverse effects on reproductive
22 performance, implantation, fetal growth and developmental defects, and the inhibition of folate
23 cellular transport has been associated with several developmental abnormalities, ranging from
24 neural tube defects to neurocristopathies such as cleft-lip and cleft-palate, cardiacseptal defects,
25 and eye defects (Antony, 2007). Folate deficiency has been shown to exacerbate some aspects of
26 the developmental toxicity of methanol in mice (see discussion of (Fuet al., 1996), and
27 (Sakanashi etal., 1996), in Section 4.3.1) and rats (see discussion of (Aziz et al., 2002), in
28 Section 4.4.1).
29 The studies in mice focused on the influence of FAD on the reproductive and skeletal
30 malformation effects of methanol. Sakanashi et al. (1996) showed that dams exposed to
31 5 g/kg-day methanol on GD6-GD15 experienced a threefold increase in the percentage of litters
32 affected by cleft palate and a 10-fold increase in the percentage of litters affected by exencephaly
33 when fed a FAD (resulting in a 50% decrease in liver folate) versus a FAS diet. They speculated
March 2011 4-71 DRAFT - DO NOT CITE OR QUOTE
-------
1 that the increased methanol effect from FAD diet could have been due to an increase in tissue
2 formate or a critical reduction in conceptus folate concentration immediately following the
3 methanol exposure. The latter appears more likely, given the high levels of formate needed to
4 cause embryotoxicity (Section 4.3.3) and the decrease in conceptus folate that is observed within
5 2 hours of GD8 methanol exposure (Dorman et al., 1995). Fu et al. (1996) confirmed the
6 findings of Sakanashi et al. (1996) and also determined that the maternal FAD diet had a much
7 greater impact on fetal liver folate than maternal liver folate levels.
8 The rat study of Aziz et al. (2002) focused on the influence of FAD on the developmental
9 neurotoxicity of methanol. Experiments by Aziz et al. (2002) involving Wistar rat dams and
10 pups exposed to methanol during lactation provide evidence that methanol exposure during this
11 postnatal period affects the developing brain. These effects (increased spontaneous locomotor
12 activity, decreased conditioned avoidance response, disturbances in dopaminergic and cholinergic
13 receptors and increased expression of GAP-43 in the hippocampal region) were more
14 pronounced in FAD as compared to FAS rats. This suggests that folic acid may play a role in
15 methanol-induced neurotoxicity. These results do not implicate any particular proximate
16 teratogen, as folate deficiency can increase levels of both methanol, formaldehyde and formate
17 (Medinsky et al., 1997). Further, folic acid is used in a number of critical pathways such as
18 purine and pyrimidine synthesis. Thus, alterations in available folic acid, particularly to the
19 conceptus, could have significant impacts on the developing fetus apart from the influence it is
20 presumed to have on formate removal.
4.6.3. Methanol-induced Formation of Free Radicals, Lipid Peroxidation, and Protein
Modifications
21 Oxidative stress in mother and offspring has been suggested to be part of the teratogenic
22 mechanism of a related alcohol, ethanol. Certain reproductive and developmental effects (e.g.,
23 resorptions and malformation rates) observed in Sprague-Dawley rats following ethanol exposure
24 were reported to be ameliorated by antioxidant (Vitamin E) treatment (Wentzel & Eriksson,
25 2006; Wentzel et al., 2006). A number of studies have examined markers of oxidative stress
26 associated with methanol exposure.
27 Skrzydlewska et al. (2005) provided inferential evidence for the effects of methanol on
28 free radical formation, lipid peroxidation, and protein modifications, by studying the protective
29 effects of N-acetyl cysteine and the Vitamin E derivative, U83836E, in the liver of male Wistar
30 rats exposed to the compound via gavage. Forty-two rats/group received a single oral gavage
31 dose of either saline or 50% methanol. This provided a dose of approximately 6,000 mg/kg, as
32 calculated by the authors. Other groups of rats received the same concentration of methanol, but
March 2011 4-72 DRAFT - DO NOT CITE OR QUOTE
-------
1 were also injected intraperitoneally with either N-acetylcysteine or U-83836E. N-acetylcysteine
2 and U-83836E controls were also included in the study design. Animals in each group were
3 sacrificed after 6, 14, and 24 hours or after 2, 5, or 7 days. Livers were rapidly excised for
4 electron spin resonance (ESR) analysis, and 10,000 x g supernatants were used to measure GSH,
5 malondialdehyde, a range of protein parameters, including free amino and sulfhydryl groups,
6 protein carbonyls, tryptophan, tyrosine, and bityrosine, and the activity of cathepsin B.
7 Skrzydlewska et al. (2005) provided data that showed an increase in an ESR signal at
8 g = 2.003 in livers harvested 6 and 12 hours after methanol exposure. The signal, thought to be
9 indicative of free radical formation, was opposed by N-acetylcysteine and U83836E. Other
10 compound-related changes included: 1) a significant decrease in GSH levels that was most
11 evident in rats sacrificed 12 and 24 hours after exposure; 2) increased concentrations in the lipid
12 peroxidation product, malondialdehyde (by a maximum of 44% in the livers of animals sacrificed
13 2 days after exposure); 3) increased specific concentrations of protein carbonyl groups and
14 bityrosine; but 4) reductions in the specific level of tryptophan. Given the ability of N-
15 acetylcysteine and U83836E to oppose these changes, at least in part, the authors speculated that
16 a number of potentially harmful changes may have occurred as a result of methanol exposure.
17 These include free radical formation, lipid peroxidation, and disturbances in protein structure.
18 However, it is unclear whether or not the metabolites of methanol, formaldehyde, and/or formate,
19 were involved in any of these changes.
20 Rajamani et al. (Rajamani et al., 2006) examined several oxidative stress parameters in
21 male Wistar rats following methotrexate-induced folate deficiency. Compared to controls, the
22 levels of free radical scavengers SOD, CAT, GSH peroxidase, oxidized GSH, protein carbonyls,
23 and lipid peroxidation were elevated in several regions of the brain, with greater increases
24 observed in the MTX-methanol-treated animals than in the MTX-alone group. The level of GSH
25 and protein thiols was decreased in all regions of the brain, with a greater decrease observed in
26 the MTX-methanol-treated animals than MTX-treated animals.
27 Dudka (2006) measured the total antioxidant status (TAS) in the brain of male Wistar rats
28 exposed to a single oral gavage dose of methanol at 3 g/kg. The animals were kept in a nitrous
29 oxide atmosphere (N2O/O2) throughout the experiment to reduce intrinsic folate levels, and
30 various levels of ethanol and/or fomepizole (as ADH antidotes) were administered i.p. after
31 4 hours. Animals were sacrificed after 16 hours, the brains homogenized, and the TAS
32 determined spectrophotometrically. As illustrated graphically by the author, methanol
33 administration reduced TAS in brain irrespective of the presence of ADH antidotes. The author
34 speculated that, while most methanol is metabolized in the liver, some may also reach the brain.
March 2011 4-73 DRAFT - DO NOT CITE OR QUOTE
-------
1 Metabolism to formate might then alter the NADH/NAD+ ratio resulting in an increase in
2 xanthine oxidase activity and the formation of the superoxide anion.
3 Parthasarathy et al. (2006) investigated the extent of methanol-induced oxidative stress in
4 rat lymphoid organs. Six male Wistar rats/group received 2,370 mg/kg methanol (mixed 1:1 with
5 saline) injected i.p. for 1, 15 or 30 days. A control group received a daily i.p. injection of saline
6 for 30 days. At term, lymphoid organs such as the spleen, thymus, lymph nodes, and bone
7 marrow were excised, perfused with saline, then homogenized to obtain supernatants in which
8 such indices of lipid peroxidation as malondialdehyde, and the activities of CAT, SOD, and GSH
9 peroxidase were measured. Parthasarathy et al. (2006) also measured the concentrations of GSH
10 and ascorbic acid (nonenzymatic antioxidants) and the serum concentrations of a number of
11 indicators of liver and kidney function, such as ALT, AST, blood urea nitrogen (BUN), and
12 creatinine.
13 Table 4-19 shows the time-dependent changes in serum liver and kidney function
14 indicators that resulted from methanol administration. Treatment with methanol for increasing
15 durations resulted in increased serum ALT and AST activities and the concentrations of BUN and
16 creatinine.
Table 4-19. 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.
ap < 0.05 versus controls.
Source: Parthasarathy et al. (2006) (adapted).
March 2011
4-74 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-20. 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.
ap < 0.05, versus controls.
Source: Parthasarathy et al. (2006) (adapted).
1 Table 4-20 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. (2006) concluded that exposure
7 to methanol may cause oxidative stress by altering the oxidant/antioxidant balance in lymphoid
8 organs in the rat.
4.6.4. Exogenous Formate Dehydrogenase as a Means of Detoxifying the Formic Acid that
Results from Methanol Exposure
9 In companion reports, Muthuvel et al. (2006; 2006) used 6 male Wistar rats/group to test
10 the ability of exogenously-administered formate dehydrogenase (FD) to reduce the serum levels
11 of formate that were formed when 3 g/kg methanol was administered i.p. to rats in saline. In the
12 first experiment, purified FD (from Candida boitinif) was administered by i.v. conjugated to the
13 N-hydroxysuccinimidyl ester of monomethoxy polyethylene glycol propionic acid (PEG-FD)
March 2011
4-75 DRAFT - DO NOT CITE OR QUOTE
-------
1 (Muthuvel, Rajamani, Senthilvelan, et al., 2006). In the second, rats were administered FD-
2 loaded erythrocytes (Muthuvel, Rajamani, Manikandan, et al., 2006). In the former case, some
3 groups of rats were made folate deficient by means of a folate-depleted diet; in the latter, folate
4 deficiency was brought about by i.p. administration of methotrexate. In some groups, the rats
5 received an infusion of an equimolar mixture of carbonate and bicarbonate (each at 0.33 mol/L)
6 to correct the formate-induced acidosis. As illustrated by the authors, methanol-exposed rats
7 receiving a folate-deficient diet showed significantly higher levels of serum formate than those
8 receiving a folate-sufficient diet. However, administration of native or PEG-FD reduced serum
9 formate in methanol-receiving folate-deficient rats to levels seen in animals receiving methanol
10 and the folate-sufficient diet.
11 In the second report, Muthuvel et al. (2006) carried out some preliminary experiments to
12 show that hematological parameters of normal, reconstituted but unloaded, and reconstituted and
13 FD-loaded erythrocytes, were similar. In addition, they showed that formate levels of serum were
14 reduced in vitro in the presence of FD-loaded erythrocytes. Expressing blood formate
15 concentration in mmol/L at the 1-hour time point after carbonate/bicarbonate and enzyme-loaded
16 erythrocyte infusion via the tail vein, the concentration was reduced from 10.63 ±1.3 (mean ±
17 S.D.) in methanol and methotrexate-receiving controls to 5.83 ± 0.97 (n = 6). This difference
18 was statistically significant at the/? < 0.05 level. However, FD-loaded erythrocytes were less
19 efficient at removing formate in the absence of carbonate/bicarbonate. Effective elimination of
20 formate appears to require an optimum pH for the FD activity in the enzyme-loaded erythrocytes.
4.7. SYNTHESIS OF MAJOR NONCANCER EFFECTS
4.7.1. Summary of Key Studies in Methanol Toxicity
21 A substantial body of information exists on the toxicological consequences to humans
22 who consume or are acutely exposed to large amounts of methanol. Neurological and
23 immunological effects have been noted in adult human subjects acutely exposed to as low as
24 200 ppm (262 mg/m3) methanol (Chuwers et al., 1995; Mann et al., 2002). Nasal irritation
25 effects have been reported by adult workers exposed to 459 ppm (601 mg/m3) methanol. Frank
26 effects such as blurred vision and bilateral or unilateral blindness, coma, convulsions/tremors,
27 nausea, headache, abdominal pain, diminished motor skills, acidosis, and dyspnea begin to occur
28 as blood levels approach 200 mg methanol/L, and 800 mg/L appears to be the threshold for
29 lethality. Data for subchronic, chronic or in utero human exposures are very limited.
30 Determinations regarding longer term effects of methanol are based primarily on animal studies.
March 2011 4-76 DRAFT - DO NOT CITE OR QUOTE
-------
1 An end-point-by-end-point survey of the primary noncancer effects of methanol in
2 experimental animals is given in the following paragraphs. Tabular summaries of the principal
3 toxicological studies that have examined the noncancer effects of methanol when experimental
4 animals were exposed to methanol via the oral or inhalation routes are provided in Tables 4-21
5 and 4-22. Most studies focused on developmental and reproductive effects. A large number of
6 the available studies were performed by routes of exposure (e.g., i.p.) that are less relevant to the
7 assessment. The data are summarized separate sections that address oral exposure (Section
8 4.7.1.1) and inhalation exposure (Section 4.7.1.2).
Table 4-21. Summary of noncancer effects reported in studies of methanol
toxicity in experimental animals (oral)
Species, strain,
number/sex
Rat
Sprague-Dawley
30/sex/group
Rat
Sprague-Dawley
100/sex/group
Mouse
Swiss
Dose/duration
0, 100, 500, and
2,500 mg/kg-day for
13 wk
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 2 101 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)
500
ND
ND
LOAEL
(mg/kg-day)
2,500
ND
1,800-2,100
Effect
Reduction of brain
weights, increase in the
serum activity of ALT
and AP. Increased liver
weights
No noncancer effects
were reported
Increased incidence of
liver parenchymal cell
necrosis
Reference
U.S. EPA
(1986b)
Soffritti et al.
(2002)
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-GD17 or
GD17-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.
(1993)
NA = Not applicable; ND = Not determined; M= male, F=female.
March 2011
4-77 DRAFT - DO NOT CITE OR QUOTE
-------
Table 4-22. Summary of studies of methanol toxicity in experimental animals
(inhalation exposure)
Species, strain,
number/sex
Monkey
M. fascicularis,
1 or 2 animals/group
Dog (2)
Rat
Sprague-Dawley
5 males/ group
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, 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
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
4wk
0 or 2,500 ppm, 6
hr/day, 5 days/wk for 4
wk
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)
ND
NA
NA
NA
5,000
5,000
NA
NA
ND
ND
NA
NA
100
100
LOAEL
(ppm)
ND
NA
200
200
NA
NA
300
2,500
ND
ND
NA
NA
1,000
1,000
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
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.
Reference
NEDO
(19871
Sayers et al.
(19441
Cameron et
al. (19841
Cameron et
al. (1985)
Andrews et
al. (19871
Poon et al.
(19941
Poon et al.
(19951
NEDO
0987)
March 2011
4-78 DRAFT - DO NOT CITE OR QUOTE
-------
Species, strain,
number/sex
Rat
Sprague-Dawley
15/pregnant
females/group
Rat
Sprague-Dawley
36/pregnant
females/group
Rat
Sprague-Dawley
FI and F2 generations
of a two-generation
study
Rat
Sprague-Dawley
Follow-up study of
brain weights inFi
generation of
10-14/sex/group inFi
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, 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,
onGD7-GD17
0, 10, 100, or
1000 ppm, 20 hr/day;
FI- birth to end of
mating (M) or weaning
(F);F2-birthto8wks
0, 500, 1,000, and
2,000 ppm; GDO
through FI generation
0, 1,000, 2,000, 5,000,
7,500, 10,000, or
15,000 ppm, 7 hr/day
onGD6-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)
5,000
1,000
100
500
1,000
NA
NA
NA
ND
LOAEL
(ppm)
10,000
5,000
1,000
1,000
2,000
10,000
15,000
4,500
NDa
Effect
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.
Reduced weight of brain, pituitary,
and thymus at 8, 16 and 24 wk
postnatal in FI 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)
Increased incidence of extra cervi-
cal 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
Nelson et al.
(1985)
NEDO
(1987)
Rogers et al.
(1993)
Rogers and
Mole (1997)
Stanton et al.
(1995)
Weiss et al.
(1996)
Burbacher et
al. (1999;
2004; 1999;
2004)
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.
March 2011
4-79 DRAFT - DO NOT CITE OR QUOTE
-------
4.7.1.1. Oral
1 There have been very few subchronic, chronic, or in utero experimental studies of oral
2 methanol toxicity. In one such experiment, an EPA-sponsored 90-day gavage study in Sprague-
3 Dawley rats suggested a possible effect of the compound on the liver (U.S. EPA, 1986b). In the
4 absence of gross or histopathologic evidence of toxicity, fluctuations on some clinical chemistry
5 markers of liver biochemistry and increases in liver weights at the highest administered dose
6 (2,500 mg/kg-day) justify the selection of the mid-dose level (500 mg/kg-day) as a NOAEL for
7 this effect under the operative experimental conditions. That the bolus effect may have been
8 important in the induction of those few effects that were apparent in the subchronic study is
9 suggested by the outcome of lifetime drinking water study of methanol that was carried out in
10 Sprague-Dawley rats by Soffritti et al. (2002). According to the authors, no noncancer
11 toxicological effects of methanol were observed at drinking water concentrations of up to
12 20,000 ppm (v/v). Based on default assumptions on drinking water consumption and body
13 weight gain assumptions, the high concentration was equivalent to a dose of 1,780 mg/kg-day in
14 males and 2,177 mg/kg-day in females. In the stated absence of any changes to parameters
15 reflective of liver toxicity in the Soffritti et al. (2002) study, the slight impacts to the liver
16 observed in the subchronic study (U.S. EPA, 1986b) at 2,500 mg/kg-day suggest the latter dose to
17 be a minimal LOAEL. Logically, the true but unknown threshold would at the high end of the
18 range from 500 (the default NOAEL) to 2,500 mg/kg-day for liver toxicity via oral gavage.
19 Two studies have pointed to the likelihood that oral exposure to methanol is associated
20 with developmental neurotoxicity or developmental deficits. When Infurna and Weiss (1986) ex-
21 posed pregnant Long-Evans rats to 2% methanol in drinking water (providing a dose of approxi-
22 mately 2,500 mg/kg-day), they observed no reproductive or developmental sequelae other than
23 from 2 tests within a battery of fetal behavioral tests (deficits in suckling ability and homing be-
24 havior). In the oral section of the Rogers et al. (1993) study, such teratological effects as cleft pa-
25 late and exencephaly and skeletal malformations were observed in fetuses of pregnant female
26 mice exposed to daily gavage doses of 4,000 mg/kg methanol during GD6-GD15. Likewise, an
27 increase in totally resorbed litters and a decrease in the number of live fetuses/litter appear likely
28 to have been an effect of the compound. Similar skeletal malformations were observed by Rogers
29 and Mole (1997), Rogers et al. (1993), and Nelson et al. (1985) following inhalation exposure.
4.7.1.2. Inhalation
30 Some clinical signs, gross pathology, and histopathological effects of methanol have been
31 seen in experimental animals including adult nonhuman primates exposed to methanol vapor.
32 Results from an unpublished study (NEDO, 1987) of M. fasciculoris monkeys, chronically
March 2011 4-80 DRAFT - DO NOT CITE OR QUOTE
-------
1 exposed to concentrations as low as 10 ppm for up to 29 months, resulted in histopathological
2 effects in the liver, kidney, brain and peripheral nervous system. These results were generally
3 reported as subtle and do not support a robust dose response over the range of exposure levels
4 used. Confidence in the methanol-induced findings of effects in adult nonhuman primates is
5 limited because this study utilized a small number (2-3) of animals/dose level/time of sacrifice
6 and inadequately reporting of results (i.e., lack of clear documentation of a concurrent control
7 group). In addition, the monkeys used in this study were all wild-caught. All of these concerns
8 limit the study's utility in derivation of an RfC.
9 A number of studies have examined the potential toxicity of methanol to the male
10 reproductive system (Cameron et al., 1984; Cameron et al., 1985; E. Lee et al., 1991). The data
11 from Cameron et al. (1984; 1985) showed a transient but not necessarily dose-related decrease in
12 serum testosterone levels of male Sprague-Dawley rats. Lee et al. (1994) reported the appearance
13 of testicular lesions in 18-month-old male Long-Evans rats that were exposed to methanol for
14 13 weeks and maintained on folate-deficient diets. Taken together, the Lee et al. (1994) and
15 Cameron et al. (1984; 1985) study results could indicate chemically-related strain on the rat
16 system as it attempts to maintain hormone homeostasis. However, the available data are
17 insufficient to definitively characterize methanol as a toxicant to the male reproductive system.
18 When Sprague-Dawley rats were exposed to methanol, 6 hours/day for 4 weeks, there
19 were some signs of irritation to the eyes and nose. Mild changes to the upper respiratory tract
20 were also described in Sprague-Dawley rats that were exposed for 4 weeks to up to 300 ppm
21 methanol (Poon et al., 1995). Other possible effects of methanol in rats included a reduction in
22 size of thyroid follicles (Poon et al., 1994), panlobular vacuolation of the liver, and a decrease in
23 spleen weight (Poon et al., 1995). NEDO (1987) reported dose-related increases in moderate
24 fatty degeneration in hepatocytes of male mice exposed via inhalation for 12 months, but this
25 finding was not observed in the NEDO (1987) 18-month mouse inhalation study. Nodes reported
26 in the liver of mice from the 18-month study may have been precancerous, but the 18-month
27 study duration was not of sufficient duration to make a determination.
28 One of the most definitive and quantifiable toxicological impacts of methanol when
29 administered to experimental animals via inhalation is related to the induction of developmental
30 abnormalities in fetuses exposed to the compound in utero. Developmental effects have been
31 demonstrated in a number of species, including monkeys, but particularly rats and mice. Most
32 developmental teratological effects appear to be more severe in the latter species. For example,
33 in the study of Rogers et al. (1993) in which pregnant female CD-I mice were exposed to
34 methanol vapors on GD6-GD15 at a range of concentrations, reproductive and fetal effects
35 included an increase in the number of resorbed litters, a reduction in the number of live pups, and
March 2011 4-81 DRAFT - DO NOT CITE OR QUOTE
-------
1 increased incidence of exencephaly, cleft palate, and the number of cervical ribs. While the
2 biological significance of the cervical rib effect has been the subject of much debate (See
3 discussion of Chernoff and Rogers (2004) in Section 5), it appears to be the most sensitive
4 indicator of developmental toxicity from this study, with a NOAEL of 1,000 ppm (1,310 mg/m3).
5 In rats, however, the most sensitive developmental effect, as reported in the NEDO (1987) two-
6 generation inhalation studies, was a postnatal reduction in brain weight at 3, 6 and 8 weeks
7 postnatally, which was significantly lower than controls when pups and their dams were exposed
8 to 1,000 ppm (1,310 mg/m3) during gestation and throughout lactation. The NOAEL reported in
9 this study was 500 ppm (655 mg/m3).
10 Rogers and Mole (1997) addressed the question of which period of gestation was most
11 critical for the adverse developmental effects of methanol in CD-I rats. Such malformations and
12 anomalies as cleft palate, exencephaly, and a range of skeletal defects, appeared to be induced
13 with a greater incidence when the dams were exposed on or around GD6. These findings were
14 taken to indicate that methanol is most toxic to embryos during gastrulation and in the early
15 stages of organogenesis. However, NEDO (1987) gestation-only and two-generation studies
16 showed that significant reductions in brain weight were observed at a lower exposure levels
17 when pups and their dams were exposed during lactation as well as gestation, indicating that
18 exposure during the later stages of organogenesis, including postnatal development, can
19 significantly contribute to the severity of the effects in this late-developing organ system.
20 In comparing the toxicity (NOAELs and LOAELs) for the onset of developmental effects
21 in mice and rats exposed in utero, there is suggestive evidence from the above studies that mice
22 may be more susceptible to methanol than rats. Supporting evidence for this proposition has
23 come from in vitro studies in which rat and mouse embryos were exposed to methanol in culture
24 (J. E. Andrews et al., 1993). Further evidence for species-by-species variations in the
25 susceptibility of experimental animals to methanol during organogenesis has come from
26 experiments on monkeys (Burbacher, Grant et al., 1999; Burbacher, Grant, et al., 2004;
27 Burbacher, Shen, et al., 1999; Burbacher, Shen, et al., 2004). In these studies, exposure of
28 monkeys to methanol during premating, mating, and throughout gestation resulted in a shorter
29 period of gestation in dams exposed to as low as 200 ppm (263 mg/m3). The shortened gestation
30 period was largely the result of C-sections performed in the methanol-exposure groups "in
31 response to signs of possible difficulty in the maintenance of pregnancy," including vaginal
32 bleeding. Though statistically significant, the finding of a shortened gestation length may be of
33 limited biological significance. Gestational age, birth weight and infant size observations in all
34 exposure groups were within normal ranges for M. fascicularis monkeys, and vaginal bleeding
35 1-4 days prior to delivery of a healthy infant does not necessarily imply a risk to the fetus (as
March 2011 4-82 DRAFT - DO NOT CITE OR QUOTE
-------
1 cited in CERHR, (2004)). An ultrasound examination could have substantiated fetal or placental
2 problems arising from presumptive pregnancy duress (see Section 4.3.2). As discussed in
3 Section 4.4.2, there is also evidence from this study that methanol caused neurobehavioral effects
4 in exposed monkey infants that may be related to the gestational exposure. However, the data are
5 not conclusive, and a dose-response trend is not robust. There is insufficient evidence to
6 determine if the primate fetus is more or less sensitive than rodents to methanol teratogenesis.
7 Several other uncertainties contributed to decreased confidence in the use of this primate in
8 quantitative estimates of risk. These included: a mixture of wild- and colony-derived monkey
9 mothers used in the study; the use of a cohort design necessitated by the complexity of this study
10 also seemingly resulted in limitations in power to detect effects (e.g., Fagan test results for
11 controls); and no apparent adjustment in statistical analysis for results from the neurobehavioral
12 battery of tests employed leading to concern about inflation of type 1 error. Because of the
13 uncertainties associated with these results, including the fact that the decrease in gestational
14 length was not exacerbated with increasing methanol exposure, EPA was not able to identify a
15 definitive NOAEL or LOAEL from this study. This study does support the weight of evidence
16 for developmental neurotoxicity in the hazard characterization of low-level methanol exposure.
17 Weiss et al. (1996) and Stanton et al. (1995) evaluated the developmental and
18 developmental neurotoxicological effects of methanol exposure on pregnant female Long-Evans
19 rats and their progeny. In the former study, exposure of dams to 15,000 ppm (19,656 mg/m3),
20 7 hours/day on GD7-GD19 resulted in reduced weight gain in pups, but produced little other
21 evidence of adverse developmental effects. The authors subjected the pups to a number of
22 neurobehavioral tests that gave little if any indication of compound-related changes. This study,
23 while using high exposure levels, was limited in its power to detect effects due to the small
24 number of animals used. In the Weiss et al. (1996) study, exposure of pregnant female Long-
25 Evans rats to 0 or 4,500 (0 and 5,897 mg/m3) methanol from GD6 to PND21 likewise provided
26 fluctuating and inconsistent results in a number of neurobehavioral tests that did not necessarily
27 indicate any compound-related impacts. The finding of this study indicated subtle cognitive
28 defects not on the learning of an operant task but in the reversal learning. This study also
29 reported exposure-related changes in neurodevelopmental markers of NCAMs on PND4.
30 NCAMs are a family of glycoproteins that is needed for migration, axonal outgrowth, and
31 establishment of the pattern for mature neuronal function.
32 Taking all of these findings into consideration reinforces the conclusion that the most
33 appropriate endpoints for use in the derivation of an RfC for methanol are associated with
34 developmental neurotoxicity and developmental toxicity. Among an array of findings indicating
35 developmental neurotoxicity and developmental malformations and anomalies that have been
March 2011 4-83 DRAFT - DO NOT CITE OR QUOTE
-------
1 observed in the fetuses and pups of exposed dams, an increase in the incidence of cervical ribs of
2 gestationally exposed mice (J. M. Rogers, Mole, et al., 1993) and a decrease in the brain weights
3 of gestationally and lactationally exposed rats (NEDO, 1987) appear to be the most robust and
4 most sensitive effects.
4.8. NONCANCER MOA INFORMATION
5 A review by Jacobsen and McMartin (1986) has provided a comprehensive summary of
6 the mechanism by which methanol brings about its acute toxic effects. Overwhelmingly, the
7 evidence points to methanol poisoning being a consequence of formate accumulation. This
8 compound is formed from formaldehyde under the action of ADH3. Formaldehyde itself is
9 formed from methanol under the action of ADH1. Evidence for the involvement of formate
10 comes from the delay in the onset of harmful symptoms, detection of formate in the blood
11 stream, and the profound acidosis that develops 12-24 hours after exposure to methanol.
12 Treatments for methanol poisoning include the i.v. administration of buffer to correct the
13 acidosis, hemodialysis to remove methanol from the blood stream, and i.v. administration of
14 either ethanol or fomepizole to inhibit the activity of ADH1. Therapies to increase endogenous
15 levels of folate may enhance the activity of THF synthetase, an enzyme that catalyzes the
16 oxidation of formate to CO2. Jacobsen and McMartin (1986) have drawn attention to the
17 accumulation of lactate in advanced stages of severe methanol poisoning, a possible consequence
18 of formate inhibition of mitochondrial respiration and tissue hypoxia. The additional decrease in
19 blood pH is likely to enhance the nonionic diffusion of formic acid across cell membranes, with
20 resulting CNS-depression, hypotension, and further lactate production.
21 Jacobsen and McMartin (1986) summarized a body of evidence that also points to the
22 formate-related acidosis as the etiologically important factor in ocular damage. The hypothesis
23 suggests that ocular toxicity is due to the inhibition of cytochrome oxidase in the optic nerve by
24 formate. This would cause inhibition of ATP formation and consequent disruption of optic nerve
25 function.
26 While it is well established that the toxic consequences of acute methanol poisoning arise
27 from the action of formate, there is less certainty on how the toxicological impacts of longer-term
28 exposure to lower levels of methanol are brought about. For example, since developmental
29 effects in experimental animals appear to be significant adverse effects associated with in utero
30 methanol exposure, it is important to determine potential MO As for how these specific effects
31 are brought about.
32 As described in Section 4.6.1, data from experiments carried out by Dorman et al. (1995),
33 formate is not the probable proximate teratogen in pregnant CD-I mice exposed to high
March 2011 4-84 DRAFT - DO NOT CITE OR QUOTE
-------
1 concentrations of methanol vapor. This conclusion is based on the fact that there appeared to be
2 little, if any, accumulation of formate in the blood of methanol-exposed mice, and exencephaly
3 did not occur until formate levels were grossly elevated. Another line of argument is based on
4 the observation that treatment of pregnant mice with a high oral dose of formate did not induce
5 neural tube closure defects at media concentrations comparable to those observed in uterine
6 decidual swelling after maternal exposure to methanol. Lastly, methanol- but not formate-
7 induced neural tube closure defects in mouse embryos in vitro at media concentrations
8 comparable to the levels of methanol detected in blood after a teratogenic exposure.
9 Harris et. al (Hansen et al.. 2005: Harris et al.. 2004: Harris et al.. 2003) carried out a
10 series of physiological and biochemical experiments on mouse and rat embryos exposed to
11 methanol, formaldehyde and formate, concluding that the etiologically important substance for
12 embryo dysmorphogenesis and embryolethality was likely to be formaldehyde rather than the
13 parent compound or formate. Specific activities for enzymes involved in methanol metabolism
14 were determined in rat and mouse embryos during the organogenesis period of 8-25 somites
15 (Harris et al., 2003). The experiment was based on the concept that differences in the
16 metabolism of methanol to formaldehyde and formic acid by the enzymes ADH1, ADH3, and
17 CAT may contribute to hypothesized differences in species sensitivity that were apparent in
18 toxicological studies. A key finding was that the activity of ADH3 (converting formaldehyde to
19 formate) was lower in mouse VYS than that of rats throughout organogenesis, consistent with the
20 greater sensitivity of the mouse to the developmental effects of methanol exposure. Another
21 study (Harris et al., 2004) which showed that the inhibition of GSH synthesis increases the
22 developmental toxicity of methanol also lends support to this hypothesis because ADH3-
23 mediated metabolism of formaldehyde is the only enzyme involved in methanol metabolism that
24 is GSH-dependent. These findings provide inferential evidence for the proposition that
25 formaldehyde may be the ultimate teratogen through diminished ADH3 activity. This concept is
26 further supported by the demonstration that the LOAELs for the embryotoxic effects of
27 formaldehyde in rat and mouse embryos were much lower than those for formate and methanol
28 (Hansen et al., 2005). Taking findings from both sets of experiments together, Harris et. al.
29 (Hansen et al.. 2005: Harris et al.. 2004: Harris et al.. 2003) concluded that the demonstrable
30 lower capacity of mouse embryos to transform formaldehyde to formate (by ADH3) could
31 explain the increased susceptibility of mouse versus rat embryos to the toxic effects of methanol.
32 While studies such as those by Harris et al. (2004: 2003) and Dorman et al. (1995: 1996)
33 strongly suggest that formate is not the metabolite responsible for methanol's teratogenic effects,
34 there are still questions regarding the relative involvement of methanol versus formaldehyde. In
35 vitro evidence suggests that formaldehyde is the more embryotoxic moiety, but methanol would
March 2011 4-85 DRAFT - DO NOT CITE OR QUOTE
-------
1 likely play a prominent role, at least in terms of transport to the target tissue. The high reactivity
2 of formaldehyde would limit its unbound and unaltered transport as free formaldehyde from
3 maternal to fetal blood (Thrasher & Kilburn, 2001), and the capacity for the metabolism of
4 methanol to formaldehyde is likely lower in the fetus and neonate versus adults (see discussion in
5 Section 3.3)
6 In humans, metabolism of methanol occurs primarily through ADH1, whereas in rodents
7 methanol metabolism involves primarily CAT, as well as ADH1. There are no known studies
8 that compare enzyme activities of human ADH1 and rodent CAT. Assuming that relative
9 expression and activity of ADH1 is comparable across species, rodents are expected to clear
10 methanol more rapidly than humans due to involvement of CAT. In fact, even among rodents the
11 metabolism of methanol may be quite different, as one study has demonstrated that the rate of
12 methanol oxidation in mice is twice the rate in rats, as well as nonhuman primates (Mannering et
13 al., 1969). Despite a faster rate of methanol metabolism, mice have consistently shown higher
14 blood methanol levels than rats following exposure to equivalent concentrations (Tables 3-4 and
15 3-5). A faster respiration rate and increased fraction of absorption by mice is thought to be the
16 reason for the higher blood methanol levels compared to rats (Perkins et al., 1995a). Using the
17 exposure conditions of Horton et al. (1992) for rats, when the respiration rate scaling coefficient
18 (QPC) was increased from the rat value of 16.4 to the mouse value of 25.4 while holding all
19 other parameters constant, peak blood concentrations were predicted by the PBPK model to
20 increase by 1.4-fold at 200 ppm and 1.8-fold at 2,000 ppm (where metabolism is becoming
21 saturated). Because smaller species generally have faster breathing rates than larger species (in
22 the PBPK model, the respiration rate/BW is 3 times slower in humans versus rats and almost 10
23 times slower versus mice), humans would be expected to accumulate less methanol than rats or
24 mice inhaling equivalent concentrations and given the same metabolism rate. However, Horton
25 et al. (1992) measured a blood concentration in rats exposed to 200 ppm methanol of about 3.7
26 mg/L after 6 hours of exposure while Sedevic et al. (1981) measured around 5.5 mg/L in human
27 volunteers after 6 hours of exposure to 231 ppm. Correcting for the higher exposure, human
28 blood concentrations would be around 4.8 mg/L if exposed at 200 ppm. Simulations with the
29 mouse model predict a blood level of 5.7 mg/L after 6 hours of exposure to 200 ppm, only 20%
30 higher than this interpolated human value. Thus the slower inhalation rate in humans is offset by
31 the slower metabolic rate, leading to equivalent blood concentrations. (If the same rate of
32 metabolism/BW as mice is used, human blood concentrations are predicted to decrease by
33 approximately fivefold.). These differences are considered in Section 5 for the characterization
34 of human and rodent PBPK models used for the derivation of human equivalent concentrations
35 (HECs).
March 2011 4-86 DRAFT - DO NOT CITE OR QUOTE
-------
4.9. EVALUATION OF CARCINOGENICITY
1 Carcinogenicity will be addressed in a separate document.
4.10. SUSCEPTIBLE POPULATIONS AND LIFE STAGES
4.10.1. Possible Childhood Suscepptibility
2 Studies in animals have identified the fetus as being more sensitive than adults to the
3 toxic effects of methanol; the greatest susceptibility occurs during gastrulation and early
4 organogenesis (CERHR, 2004). Table 4-17 summarizes some of the data regarding the relative
5 ontogeny of CAT, ADH1, and ADH3 in humans and mice. Human fetuses have limited ability to
6 metabolize methanol as ADH1 activity in 2-month-old and 4-5 month-old fetuses is 3-4% and
7 10% of adult activity, respectively (Pikkarainen & Raiha, 1967). ADH1 activity in 9-22 week old
8 fetal livers was found to be 30% of adult activity (M. Smith et al., 1971). Likewise, ADH1
9 activity is -20-50% of adult activity during infancy (Pikkarainen & Raiha, 1967; M. Smith et al.,
10 1971). Activity continues to increase until reaching adult levels at 5 years of age (Pikkarainen &
11 Raiha, 1967). However, no difference between blood methanol levels in 1-year-old infants and
12 adults was observed following ingesting the same doses of aspartame, which releases 10%
13 methanol by weight during metabolism (Stegink et al., 1983). Given that the exposure was
14 aspartame as opposed to methanol, it is difficult to draw any conclusions from this study vis-a-vis
15 ontogeny data and potential influences of age differences in aspartame disposition. With regard
16 to inhalation exposure, increased breathing rates relative to adults may result in higher blood
17 methanol levels in children compared to adults (CERHR, 2004). It is also possible that
18 metabolic variations resulting in increased methanol blood levels in pregnant women could
19 increase the fetus' risk from exposure to methanol. In all, unresolved issues regarding the
20 identification of the toxic moiety increase the uncertainty with regards to the extent and
21 pathologic basis for early life susceptibility to methanol exposure.
22 The prevalence of folic acid deficiency has decreased since the United States and Canada
23 introduced a mandatory folic acid food fortification program in November 1998. However,
24 folate deficiency is still a concern among pregnant and lactating women, and factors such as
25 smoking, a poor quality diet, alcohol intake, and folic antagonist medications can enhance
26 deficiency (CERHR, 2004). Folate deficiency could affect a pregnant woman's ability to clear
27 formate, which has also been demonstrated to produce developmental toxicity in rodent in in
28 vitro studies at high-doses (Dorman et al., 1995). It is not known if folate-deficient humans have
29 higher levels of blood formate than individuals with adequate folate levels. A limited study in
30 folate-deficient monkeys demonstrated no formate accumulation following an endotracheal
31 exposure of anesthetized monkeys to 900 ppm methanol for 2 hours (Dorman et al., 1994). The
March 2011 4-87 DRAFT - DO NOT CITE OR QUOTE
-------
1 situation is obscured by the fact that folic acid deficiency during pregnancy by itself is thought to
2 contribute to the development of severe congenital malformations (Pitkin, 2007).
4.10.2. Possible Gender Differences
3 There is limited information on potential differences in susceptibility to the toxic effects
4 of methanol according to gender. However, one study reported a higher background blood
5 methanol level in human females versus males (Batterman & Franzblau, 1997). In rodents,
6 fetuses exposed in utero were found to be the most sensitive subpopulation. One study suggested
7 a possible increased sensitivity of male versus female rat fetuses and pups. When rats were
8 exposed to methanol pre- and postnatally, 6- and 8-week-old male progeny had significantly
9 lower brain weights at 1,000 ppm, compared to those in females that demonstrated the same
10 effect only at 2,000 ppm (NEDO, 1987). In general, there is little evidence for substantial
11 disparity in the level or degree of toxic response to methanol in male versus female experimental
12 animals or humans. However, it is possible that the compound-related deficits in fetal brain
13 weight that were evident in the pups of FI generation Sprague-Dawley rats exposed to methanol
14 in the NEDO (1987) study may reflect a threshold neurotoxicological response to methanol. It is
15 currently unknown whether higher levels of exposure would result in brain sequelae comparable
16 to those observed in acutely exposed humans.
4.10.3. Genetic Susceptibility
17 Polymorphisms in enzymes involved in methanol metabolism may affect the sensitivity of
18 some individuals to methanol. For example, as discussed in Chapter 3, data summarized in
19 reviews by Agarwal (2001). Burnell et al. (1989). Bosron and Li (1986). and Pietruszko (1980)
20 discuss genetic polymorphisms for ADH. Class IADH, the primary ADH in human liver, is a
21 dimer composed of randomly associated polypeptide units encoded by three genetic loci
22 (ADH1 A, ADH1B, and ADH1C). Polymorphisms are observed at the ADH1B (ADH1B*2,
23 ADH1B*3) and ADH1C (ADH1C*2) loci. The ADH1B*2 phenotype is estimated to occur in
24 -15% of Caucasians of European descent, 85% of Asians, and less that 5% of African Americans.
25 Fifteen percent of African Americans have the ADH1B*3 phenotype, while it is found in less
26 than 5% of Caucasian Europeans and Asians. The only reported polymorphisms in ADH3 occur
27 in the promoter region, one of which reduces the transcript!onal activity in vitro nearly twofold
28 (Hedberg et al., 2001). While polymorphisms in ADH3 are described in more than one report
29 (Cichoz-Lach et al., 2007; Hedberg et al., 2001), the functional consequence(s) for these
30 polymorphisms remains unclear.
March 2011 4-88 DRAFT - DO NOT CITE OR QUOTE
-------
5. DOSE-RESPONSE ASSESSMENTS AND CHARACTERIZATION
5.1. INHALATION REFERENCE CONCENTRATION (RfC)49
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.50 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
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
49 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.
50 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).
March 2011 5-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 human variability, animal-to-human differences, and limitations in the database influence
2 derivationoftheRfC.
5.1.1. Choice of Principal Study and Critical Effect(s)
5.1.1.1. Key Inhalation Studies
3 While a substantial body of information exists on the toxicological consequences to
4 humans exposed to large amounts of methanol, no human studies exist that would allow for
5 quantification of subchronic, chronic, or in utero effects of methanol exposure. Table 4-22
6 summarizes available experimental animal inhalation studies of methanol. Several of these
7 studies, including the monkey chronic (NEDO, 1987) and developmental (Burbacher, Grant, et
8 al., 1999: Burbacher, Grant et al., 2004; Burbacher, Shen, et al., 1999; Burbacher, Shen, et al.,
9 2004) studies, the male rat reproductive studies (Cameron et al., 1984; Cameron et al., 1985; R
10 Lee et al., 1991), and the 4-week rat studies (Poon et al., 1994), are lacking in key attributes (e.g.,
11 documented dose response, documented controls, and duration of exposure) necessary for their
12 direct use in the quantification of a chronic RfC. These studies will be considered in this chapter
13 for their contributions to the overall RfC uncertainty. Several inhalation reproductive or
14 developmental studies were adequately documented and are of appropriate size and design for
15 quantification and derivation of an RfC. These studies are considered for use in the derivation of
16 an RfC and are summarized below.
5.1.1.2. Selection of Critical Effect(s)
17 Developmental effects have been assessed in a number of toxicological studies of
18 monkeys, rats, and mice. The findings of Rogers and Mole (1997) indicate that methanol is toxic
19 to mouse embryos in the early stages of organogenesis, on or around GD7. In the study of
20 Rogers et al. (1993), in which pregnant female CD-I mice were exposed to methanol vapors
21 (1,000, 2,000, and 5,000 ppm) on GD6-GD15, reproductive and fetal effects included an increase
22 in the number of resorbed litters, a reduction in the number of live pups, and increased incidences
23 of exencephaly, cleft palate, and the number of cervical ribs. They reported a NOAEL for
24 cervical rib malformations at 1,000 ppm (1,310 mg/m3) and a LOAEL of 2,000 ppm (2,620
25 mg/m3, 49.6% per litter versus 28.0% per litter in the control group). Increased incidence of
26 cervical ribs was also observed in the rat organogenesis study (NEDO, 1987) in the 5,000 ppm
27 dose group (65.2% per litter versus 0% in the control group), indicating that the endpoint is
28 significant across species.
29 The biological significance of the cervical rib endpoint within the regulatory arena has
30 been the subject of much debate (Chernoff & Rogers, 2004). Previous studies have classified
March 2011 5-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 this endpoint as either a malformation (birth defect of major importance) or a variation
2 (morphological alternation of minor significance). There is evidence that incidence of
3 supernumerary ribs (including cervical ribs) is not just the addition of extraneous, single ribs but
4 rather is related to a general alteration in the development and architecture of the axial skeleton
5 as a whole. In CD-I mice exposed during gestation to various types of stress, food and water
6 deprivation, and the herbicide dinoseb, supernumerary ribs were consistently associated with
7 increases in length of the 13th rib (Branch et al., 1996). This relationship was present in all fetal
8 ages examined in the study. The authors concluded that these findings are consistent with
9 supernumerary ribs being one manifestation of a basic alteration in the differentiation of the
10 thoraco-lumbar border of the axial skeleton. The biological significance of this endpoint is
11 further strengthened by the association of supernumerary ribs with adverse health effects in
12 humans. The most common effect produced by the presence of cervical ribs is thoracic outlet
13 disease (Fernandez Noda et al., 1996; M. S. Henderson, 1914; Nguyen et al., 1997). Thoracic
14 outlet disease is characterized by numbness and/or pain in the shoulder, arm, or hands. Vascular
15 effects associated with this syndrome include cerebral and distal embolism (Beam et al., 1993;
16 Connell et al., 1980; Short, 1975), while neurological symptoms include extreme pain, migraine,
17 and symptoms similar to Parkinson's (Evans, 1999; Fernandez Noda et al., 1996; Saxton et al.,
18 1999). Schumacher et al. (1992) observed 242 rib anomalies in 218 children with tumors (21.8%)
19 and 11 (5.5%) in children without malignancy, a statistically significant (p < 0.001) difference
20 that indicates a strong association between the presence of cervical ribs and childhood cancers.
21 A number of rat studies have confirmed the toxicity of methanol to embryos during
22 organogenesis (NEDO, 1987: Nelson et al., 1985: Weiss et al., 1996). NEDO (1987) reported
23 reduced brain, pituitary, and thymus weights in FI and F2 generation Sprague-Dawley rats at
24 1,000 ppm methanol. In a follow-up study of the FI generation brain weight effects, NEDO
25 (1987) reported decreased brain, cerebellum, and cerebrum weights in FI males exposed at
26 1,000 ppm methanol from GDO through the FI generation. The exposure levels used in these
27 studies are difficult to interpret because dams were exposed prior to gestation, and dams and
28 pups were exposed during gestation and lactation. However, it is clear that postnatal exposure
29 increases the severity of brain weight reduction. In another experiment in which NEDO (1987)
30 exposed rats only during organogenesis (GD7-GD17), the observed decreases in brain weights in
31 offspring at 8 weeks of age were less severe than in the studies for which exposure was continued
32 postnatally. This finding is not unexpected, given that the brain undergoes tremendous growth
33 beginning early in gestation and continuing in the postnatal period. Rats are considered altricial
34 (i.e., born at relatively underdeveloped stages), and many of their neurogenic events occur
35 postnatally (Clancy et al., 2007). Brain effects from postnatal exposure are also relevant to
March 2011 5-3 DRAFT - DO NOT CITE OR QUOTE
-------
1 humans given that, in humans, gross measures of brain growth increase for at least 2-3 years after
2 birth, with the growth rate peaking approximately 4 months after birth (Rice & Barone, 2000).
3 A change in brain weight is considered to be a biologically significant effect (U.S. EPA,
4 1998). This is true regardless of changes in body weight because brain weight is generally
5 protected during malnutrition or weight loss, unlike many other organs or tissues (U.S. EPA,
6 1998). Thus, change in absolute brain weight is an appropriate measure of effects on this critical
7 organ system. Decreases in brain weight have been associated with simultaneous deficits in
8 neurobehavioral and cognitive parameters in animals exposed during gestation to various
9 solvents, including toluene and ethanol (Coleman et al., 1999; Gibson et al., 2000; Hass et al.,
10 1995). NEDO (1987) reports that brain, cerebellum, and cerebrum weights decrease in a dose-
11 dependant manner in male rats exposed to methanol throughout gestation and the F i generation.
12 Developmental neurobehavioral effects associated with methanol inhalation exposure
13 have been investigated in monkeys. Burbacher et al. (1999; 2004; 1999; 2004) exposedM
14 fascicularis monkeys to 0, 262, 786, and 2,359 mg/m3 methanol, 2.5 hours/day, 7 days/week
15 during premating/mating and throughout gestation (approximately 168 days). In these studies,
16 exposure of monkeys to up to 1,800 ppm (2,359 mg/m3) methanol during premating, mating, and
17 throughout gestation resulted in no changes in reproductive parameters other than a shorter
18 period of gestation in all exposure groups that did not appear to be dose related. The shortened
19 gestation period was largely the result of C-sections performed in the methanol exposure groups
20 "in response to signs of possible difficulty in the maintenance of pregnancy," including vaginal
21 bleeding. As discussed in Section 4.7.1.2, though statistically significant, the shortened gestation
22 finding may be of limited biological significance given questions concerning its relation to the
23 methanol exposure. Developmental parameters, such as fetal crown-rump length and head
24 circumference, were unaffected, but there appeared to be neurotoxicological deficits in methanol-
25 exposed pups. VDR was significantly reduced in the 786 mg/m3 group for males and the 2,359
26 mg/m3 group for both sexes. However, a dose-response trend for this endpoint was only
27 exhibited for females. In fact, this is the only effect reported in the Burbacher et al. (1999; 2004;
28 1999; 2004) studies for which a significant dose-response trend is evident. As discussed in
29 Section 4.4.2, confidence may have been increased by statistical analyses to adjust for multiple
30 testing (CERHR, 2004). Yet it is worth noting that the dose-response trend for VDR in females
31 remained significant with (p = 0.009) and without (p = 0.0265) an adjustment for the shortened
32 gestational periods, and it is a measure of functional deficits in sensorimotor development that is
33 consistent with early developmental CNS effects (brain weight changes discussed above) that
34 have been observed in rats.
March 2011 5-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 Another test, the Fagan test of infant intelligence, indicated small but not significant
2 deficits of performance (time spent looking at novel faces versus familiar faces) in treated
3 monkeys. Although not statistically significant and not quantifiable, the results of this test are
4 also important when considered in conjunction with the brain weight changes noted in the NEDO
5 (1987) rat study. As discussed in Section 4.7.1.2, the monkey data are not conclusive, and there
6 is insufficient evidence to determine if the primate fetus is more or less sensitive than rodents to
7 methanol teratogenesis. Taken together, however, the NEDO (1987) rat study and the Burbacher
8 et al. (1999; 2004; 1999; 2004) monkey study suggest that prenatal exposure to methanol can
9 result in adverse effects on developmental neurology pathology and function, which can be
10 exacerbated by continued postnatal exposure.
11 A number of studies described in Section 4.3.2 and summarized in Section 4.7.1.2 have
12 examined the potential toxicity of methanol to the male reproductive system (Cameron et al.,
13 1984; Cameron et al., 1985; E. Lee et al., 1991). Some of the observed effects, including a
14 transient decrease in testosterone levels, could be the result of chemically related strain on the rat
15 system as it attempts to maintain hormone homeostasis. However, the data are insufficient to
16 definitively characterize methanol as a toxicant to the male reproductive system.
17 The studies considered for use in the derivation of an RfC are summarized in Table 5-1.
18 As discussed in Sections 5.1.3.1 and 5.3, there is uncertainty associated with the selection of an
19 effect endpoint from the methanol database for use in the derivation of an RfC. Taking into
20 account the limitations of the studies available for quantification purposes, decreased brain
21 weight at 6 weeks in male Sprague-Dawley rats exposed throughout gestation and the postnatal
22 period (NEDO, 1987) was chosen as the critical effect for the purposes of this dose-response
23 assessment as it can be reliably quantified and represents both a sensitive organ system and a key
24 period of development. RfC derivations utilizing alternative endpoints (e.g., cervical rib effects
25 in mice and delayed sensorimotor development in monkeys) and alternative methods (e.g., use of
26 different BMRs) are summarized in Appendix C and in Section 5.1.3.1.
March 2011 5-5 DRAFT - DO NOT CITE OR QUOTE
-------
Table 5-1. Summary of studies considered most appropriate for use in
derivation of an RfC
REFERENCE
NEDO (19871
Two-generation
study
NEDO (1987)
Follow-up study
of FI generation
NEDO (1987)
Teratology study
Nelson et al.
(1985)
Rogers et al.
(1993)
Burbacher et al.
(1999: 2004:
1999; 2004)
Species
(strain)
Rat
Sprague-
Dawley
Rat
Sprague-
Dawley
Rat
Sprague-
Dawley
Mouse
CD-I
M.
fascicularis
Sex
M,F
M,F
F
F
Number/
dose group
Not specified -
F! andF2
generation
10-147 sex/
group- F!
generation
10-12/sex/
group
15 pregnant
dams/group
30-114
pregnant dams/
group
12 pregnant
monkeys/grou
P
Exposure
Duration
F! -Birth to end
of mating (M)
or weaning (F);
F 2 -birth to 8
wk
GDO through
F! generation
GD7-GD17
GDl-GD19or
GD7-GD15
GD6-GD15
2.5 hr/day,
7 days/wk,
during
premating,
mating and
gestation
Critical Effect
Reduced weight of
brain, pituitary, and
thymus at 8, 16, and
24 wk postnatal in F!
and at 8 wkinF2
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
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)
100
500
1,000
5,000
1,000
LOAEL
(ppm)
1,000
1,000
5,000
10,000
2,000
b
aAnimals 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.
March 2011
5-6 DRAFT - DO NOT CITE OR QUOTE
-------
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 several
5 measures of dose for methanol, including the following:
6 • Cmax - The peak concentration of methanol in the blood during the exposure period;
7 • AUC - Area under the curve, which represents the cumulative product of concentration and
8 time for methanol in the blood; and
9 • Total metabolism - The production of metabolites of methanol, namely formaldehyde and
10 formate.
11 As described in Section 3.4.3.2, the focus of model development is on obtaining accurate
12 predictions of increased body burdens over endogenous background levels of methanol and its
13 metabolites. The PBPK models do not describe or account for background levels of methanol,
14 formaldehyde or formate.
15 Although there remains uncertainty surrounding the identification of the proximate
16 teratogen of importance (methanol, formaldehyde, or formate), the dose metric chosen for
17 derivation of an RfC was based on blood methanol levels. This decision was primarily based on
18 evidence that the toxic moiety is not likely to be the formate metabolite of methanol (CERHR,
19 2004) and evidence that levels of the formaldehyde metabolite following methanol maternal
20 and/or neonate exposure would be much lower in the fetus and neonate than in adults. While
21 recent in vitro evidence indicates that formaldehyde is more embryotoxic than methanol and
22 formate, the high reactivity of formaldehyde would limit its unbound and unaltered transport as
23 free formaldehyde from maternal to fetal blood (Thrasher & Kilburn, 2001), and the capacity for
24 the metabolism of methanol to formaldehyde is likely lower in the fetus and neonate versus
25 adults (see discussion in Section 3.3). Thus, even if formaldehyde is identified as the proximate
26 teratogen, methanol would likely play a prominent role, at least in terms of transport to the target
27 tissue. Further discussions of methanol metabolism, dose metric selection, and MOA issues are
28 covered in Sections 3.3, 4.6, and 4.8.
29 A BMDL was then derived in terms of the internal dose metric utilized. Finally, the
30 BMDL values were converted to FtECs via the use of a PBPK model parameterized for humans.
31 The next section describes the rationale for and application of the benchmark modeling
32 methodology for the RfC derivation.
March 2011 5-7 DRAFT - DO NOT CITE OR QUOTE
-------
5.1.2.1. Application of the BMD/BMDL Approach
1 Several developments over the past few years impact the derivation of the RfC: (1) EPA
2 has developed draft BMD assessment methods (U.S. EPA, 1995, 2000a) and supporting software
3 (Appendix C) to improve upon the previous NOAEL/LOAEL approach; (2) MOA studies have
4 been carried out that can give more insight into methanol toxicity; and (3) EPA has refined PBPK
5 models for methanol on the basis of the work of Ward et al. (1997) (see Section 3.4. for
6 description of the EPA model). The EPA PBPK model provides estimates of HECs from rodent
7 exposures that are supported by pharmacokinetic information available for rodents and humans.
8 The following sections describe how the BMD/BMDL approach, along with the EPA PBPK
9 model, is used to obtain a POD for use in the derivation of an RfC for methanol in accordance
10 with current draft BMD technical guidance (U.S. EPA. 2000a).
11 The BMD approach attempts to fit models to the dose-response data for a given endpoint.
12 It has the advantage of taking more of the dose-response data into account when determining the
13 POD, as well as estimating the dose for which an effect may have a specific probability of
14 occurring. The BMD approach also accounts, in part, for the quality of the study (e.g., study
15 size) by estimating a BMDL, the 95% lower bound confidence limit on the BMD. The BMDL is
16 closer to the BMD (higher) for large studies and further away from the BMD (lower) for small
17 studies. Because the BMDL approach will account, in part, for a study's power, dose spacing,
18 and the steepness of the dose-response curve, it is generally preferred over the NOAEL approach.
19 When possible, all experimental data points are included in this assessment to ensure
20 adequate fit of a BMD model and derivation of a BMDL. A summary of the POD values
21 determined by BMD analysis for the critical endpoint (as well as other considered endpoints) (see
22 Appendix C for modeling results), application of UFs, and conversion to HECs using the BMD
23 and PBPK approach, is included in Section 5.1.3.1.
March 2011 5-8 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.51 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 (U.S. EPA. 2000a).52
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), described in Section 4.4.2 and shown in Table 4-10,
12 indicate that there is not an obvious cumulative effect of ongoing exposure on brain-weight
13 decrements in rats exposed postnatally; i.e., the dose response in terms of percent of control is
14 about the same at 3 weeks postnatal as at 8 weeks postnatal in rats exposed throughout gestation
15 and the FI generation. However, there does appear to be a greater brain-weight effect in rats
16 exposed postnatally versus rats exposed only during organogenesis (GD7-GD17). In male rats
17 exposed during organogenesis only, there is no statistically significant decrease in brain weight at
18 8 weeks after birth at the 1,000 ppm exposure level. Conversely, in male rats exposed to the
19 same level of methanol throughout gestation and the FI generation, there was an approximately
20 5% decrease in brain weights (statistically significant at the/? < 0.01 level). The fact that male
21 rats exposed to 5,000 ppm methanol only during organogenesis experienced a decrease in brain
22 weight of 10% at 8 weeks postnatal indicates that postnatal exposure is not necessary for the
23 observation of persistent postnatal effects. However, the fact that this decrease was less than the
24 13% decrease observed in male rats exposed to 2,000 ppm methanol throughout gestation and the
25 8 week postnatal period indicates that both exposure concentration and duration are important
26 components of the ability of methanol to cause this effect. The extent to which the observation
27 of the increased effect is due to a cumulative effect in rats exposed postnatally versus recovery in
28 rats for which exposure was discontinued at birth is not clear.
51USEPA's BMDS 2.1.1 (U.S. EPA. 2009a') 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).
52Akaike's Information Criterion (AIC) (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.
March 2011 5-9 DRAFT - DO NOT CITE OR QUOTE
-------
1 The fact that brain weight is susceptible to both the level and duration of exposure
2 suggests that a dose metric that incorporates a time component would be the most appropriate
3 metric to use. For these reasons, and because it is more typically used in internal-dose-based
4 assessments and better reflects total exposure within a given day, daily AUC (measured for
5 22 hours exposure/day) was chosen as the most appropriate dose metric for modeling the effects
6 of methanol exposure on brain weights in rats exposed throughout gestation and continuing into
7 theFi generation.
8 Application of the EPA methanol PBPK model (described in Section 3.4) to the NEDO
9 (1987) study, in which developing rats were exposed during gestation and the postnatal period,
10 presents complications that need to be discussed. The neonatal rats in this study were exposed to
11 methanol gestationally before parturition as well as lactationally and inhalationally after
12 parturition. The PBPK model developed by EPA only estimates internal dose metrics for
13 methanol exposure in NP adult mice and rats. Experimental data indicate that inhalation-route
14 blood methanol kinetics in NP mice and pregnant mice on GD6-GD10 are similar (Dorman et al.,
15 1995: Perkins et al.. 1995b: J. M. Rogers. Barbee. et al.. 1993: J. M. Rogers. Mole, et al.. 1993).
16 In addition, experimental data indicate that the maternal blood:fetal partition coefficient for mice
17 is approximately 1 (see Sections 3.4.1.2 and 3.4.4). Assuming that these findings apply for rats,
18 the data indicate that PBPK estimates of PK and blood dose metrics for NP rats are better
19 predictors of fetal exposure during gestation than would be obtained from default extrapolations
20 from external exposure concentrations. However, as is discussed to a greater extent in Section
21 5.3, the additional routes of exposure presented to the pups in this study (lactation and inhalation)
22 present uncertainties that suggest the average blood levels in pups in the NEDO (1987) report
23 might be greater than those of the dam. The assumption made in this assessment is that, if such
24 differences exist between human mothers and their offspring, they are not expected to be
25 significantly greater than that which has been postulated for rats. Thus, the PBPK model-
26 estimated adult blood methanol level is considered to be an appropriate dose metric for the
27 purpose of this analysis and HEC derivation.
5.1.2.2. HMD Approach Applied to Brain Weight Data in Rats
28 The NEDO (1987) study reported decreases in brain weights in developing rats exposed
29 during gestation only (GD7-GD17) or during gestation and the postnatal period, up to 8 weeks
30 (see Section 4.4.2). Because of the biological significance of decreases in brain weight as an
31 endpoint in the developing rat and because this endpoint was not evaluated in other peer-
32 reviewed studies, BMD analysis was performed using these data. For the purposes of deriving an
33 RfC for methanol from developmental endpoints using the BMD method and rat data, decreases
34 in brain weight at 6 weeks of age in the more sensitive gender, males, exposed throughout
March 2011 5-10 DRAFT - DO NOT CITE OR QUOTE
-------
1 gestation and continuing into the FI generation (both through lactation and inhalation routes)
2 were utilized. Decreases in brain weight at 6 weeks (gestational and postnatal exposure), rather
3 than those seen at 3 and 8 weeks, were chosen as the basis for the RfC derivation because they
4 resulted in lower estimated BMDs and BMDLs. Decreased brain weights in male rats at 8 weeks
5 age after gestation-only exposure were not utilized because they were less severe at the same
6 dose level (1,000 ppm) compared to gestation and postnatal exposure.
7 The first step in the current BMD analysis is to convert the inhalation doses, given as ppm
8 values from the studies, to an internal dose metric using the EPAPBPK model (see Section 3.4).
9 For decreased brain weight in male rats, AUC of methanol in blood (hr x mg/L) is chosen as the
10 appropriate internal dose metric for the reasons discussed in Section 5.1.2.1. Predicted AUC
11 values for methanol in the blood of rats are summarized in Table 5-2. These AUC values are
12 then used as the dose metric for the BMD analysis of response data shown in Table 5-2 for
13 decreased brain weight at 6 weeks in male rats following gestational and postnatal exposure.53
14 The full details of this analysis are reported in Appendix C. More details concerning the PBPK
15 modeling were presented in Section 3.4.
16
Table 5-2. The EPA PBPK model estimates of methanol blood levels (AUC)a
in rat dams following inhalation exposures and reported brain weights of 6
week old male pups.
Exposure level (ppm)
0
500
1,000
2,000
Methanol in blood AUC (hr
x mg/L)A in Rats
0
79.1
226.5
966.0
Mean male rat (Ft generation) brain weight
at 6 weeks8
1.78 ±0.07
1.74 ±0.09
1.69±0.06C
1.52±0.07d
aAUC values were obtained by simulating 22 hr/day exposures for 5 days and calculated for the last 24 hours of that
period.
bExposed throughout gestation and FI generation. Values are means ± S.D.
cp < 0.01, dp < 0.001, as calculated by the authors.
Source: NEDO (1987).
17 The current draft BMD technical guidance (U.S. EPA, 2000a) suggests that, in the
18 absence of knowledge as to what level of response to consider adverse, a change in the mean
All BMD assessments in this review were performed using BMDS version 2.1.1 (U.S. EPA. 2009a).
March 2011
5-11 DRAFT - DO NOT CITE OR QUOTE
-------
1 equal to one S.D. from the control mean can be used as a BMR for continuous endpoints.
2 However, it has been suggested that other BMRs, such as 5% change relative to estimated control
3 mean, are also appropriate when performing BMD analyses on fetal weight change as a
4 developmental endpoint (Kavlock et al., 1995). Therefore, both a one S.D. change from the
5 control mean and a 5% change relative to estimated control mean were considered (see Appendix
6 C for RfC derivations using alternative BMRs). For this endpoint, a one S.D. change from the
7 control mean returned the lowest BMDL estimates and was considered the most suitable BMR
8 for use in the RfC derivation. All models were fit using restrictions and option settings suggested
9 in the draft EPA BMD Technical Guidance Document (U.S. EPA. 2000a).
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 is provided
13 in Table 5-3. BMDL values in Table 5-3 represent the 95% lower-bound confidence limit on the
14 AUC estimated to result in a mean that is one S.D. from the control mean. There is a 2.5-fold
15 range of BMDL estimates from adequately fitting models, indicating considerable model
16 dependence. In addition, the fit of the Hill and more complex Exponential models is better than
17 the other models in the dose region of interest as indicated by a lower scaled residual at the dose
18 group closest to the BMD (0.09 versus -0.67 or -0.77) and visual inspection. In accordance with
19 draft EPA BMD Technical Guidance (2000a), the BMDL from the Hill model (bolded), is
20 selected as the most appropriate basis for an RfC derivation because it results in the lowest
21 BMDL from among a broad range of BMDLs and provides a superior fit in the low dose region
22 nearest the BMD. The Hill model dose-response curve for decreased brain weight in male rats is
23 presented in Figure 5-1, with response plotted against the chosen internal dose metric of AUC of
24 methanol in rats. The BMDLiso was determined to be 90.9 hr x mg/L using the 95% lower
25 confidence limit of the dose-response curve expressed in terms of the AUC for methanol in
26 blood.
March 2011 5-12 DRAFT - DO NOT CITE OR QUOTE
-------
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
277.75
277.75
277.75
277.75
170.43
260.42
260.42
171.95
171.95
BMDL1SD
(AUC,
hr x mg/L)A
224.85
224.85
224.85
224.85
90.86
208.68
208.68
96.85
96.85
/rvalue
0.5387
0.5387
0.5387
0.5387
0.836
0.613
0.613
0.82
0.82
AICC
-203.84
-203.84
-203.84
-203.84
-203.04
-204.10
-204.10
-203.03
-203.03
Scaled residual0
-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. usingBMDS 2.1.1 (U.S. EPA, 2009, 200772) and model options and restrictions suggested by EPABMD
technical guidance (U.S. EPA. 2000a).
bln accordance with draft EPA BMD Technical Guidance guidance (2000a), 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).
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).
March 2011
5-13 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
Hill
BMDL BMD
200
400 600
dose
800
1000
09:05 08/24 2009
1
2
3
4
5
6
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.
Once the BMDLiso was obtained in units of hr x mg/L, it was used to derive a chronic
RfC. The first step is to calculate the HEC using the PBPK model described in Appendix B. An
algebraic equation is provided (Equation 1 of Appendix B) that describes the relationship
between predicted methanol AUC and the human equivalent inhalation exposure concentration
(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
12
HEC (mg/m3) =1.31 x 139 ppm = 182 mg/m3
March 2011
5-14 DRAFT - DO NOT CITE OR QUOTE
-------
5.1.3. RfC Derivation - Including Application of Uncertainty Factors
5.1.3.1. Comparison between Endpoints and BMDL Modeling Approaches
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 applied54 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 (J. M. Rogers, Mole, et al., 1993) because they generally involve a larger number of
14 subjects. Reference values estimated for cervical rib incidence in mice using Cmax as the dose
15 metric were 13.6 and 10.4 mg/m3 using BMDLio and BMDLos PODs, respectively (see
16 Appendix D for discussion of choice of Cmax as the appropriate dose metric for incidence of
17 cervical rib in mice). The reference value estimated for alterations in sensorimotor development
18 and performance as measured by the VDR test in female monkeys using AUC as the dose metric
19 was 1.7 mg/m3 using the BMDLso as the POD. As discussed in Section 4.4.2, confidence in this
20 endpoint is reduced by a marginal dose-response trend in one sex (females) and a limited sample
21 size. Although the VDR test demonstrates that prenatal and continuing postnatal exposure to
22 methanol can result in neurotoxicity, the use of such statistically borderline results is not
23 warranted in the derivation of the RfC, given the availability of better dose-response data in other
24 species. Decreases in brain weight at 6 weeks of age in male rats exposed during gestation and
25 throughout the F i 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
54 The rationale for the selection of these UFs is discussed later in Section 5.1.3.
March 2011 5-15 DRAFT - DO NOT CITE OR QUOTE
-------
1 standard recommended by EPA's draft technical guidance (2000a) and yields a lower BMDL than
2 5% relative deviance for this data set. Thus, the RfC is:
3 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. (1993)
BMDL10 mouse
cervical rib Cmax
94.3 mg/L
1360
10
3
3
1
1
100
13.6
BMDLos mouse
cervical rib Cmax
44.7 mg/L
1036
10
3
3
1
1
100
10.4
Burbacher et al.
(1999: 19991
BMDL1SD
female monkey
VDRa AUC
81.7hrxmg/L
165
10
3
3
1
1
100
1.7
NEDO (1987)
BMDL05
rat brain wt.b
AUC
123.8 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 x Cmax). 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 5.1.3.2.1 to 5.1.3.2.4.
fThis endpoint (bolded) was used for the derivation of the RfC.
March 2011
5-16 DRAFT - DO NOT CITE OR QUOTE
-------
10000
Rogers et al., 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; BWDI_(HEC)05 weight; BMDL(HEC)1SD
BMDL(HEC)10 reaching; BMDL(HEC)1SD
• POD I
1 Interspecies
Extrapolation
•
Inter-individual Variation
1 Database
^ RfC
Figure 5-2. PODs (in mg/m ) for selected endpoints with corresponding
applied UFs (chosen RfC value is circled)
5.1.3.2. Application 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 UFn- 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
March 2011
5-17 DRAFT - DO NOT CITE OR QUOTE
-------
1 of the body burden of methanol or its metabolites, sensitization of an individual to methanol
2 effects, or augmentation of underlying conditions or changes in processes that share common
3 features with methanol effects. Additionally, inherent differences in an individual's genetic
4 make-up, diet, gender, age, or disease state may affect the pharmacokinetics and
5 pharmacodynamics of methanol, influencing susceptibility intrinsically. Co-exposure to a
6 pollutant that alters metabolism or other clearance processes, or that adds to background levels of
7 metabolites may also affect the pharmacokinetics and pharmacodynamics of methanol,
8 influencing susceptibility extrinsically (see Section 4.10). The determination of the UF for
9 human variation is supported by several types of information, including information concerning
10 background levels of methanol in humans, variation in pharmacokinetics revealed through
11 human studies and from PBPK modeling, variation of methanol metabolism in human tissues,
12 and information on physiologic factors (including gender and age), or acquired factors (including
13 diet and environment) that may affect methanol exposure and toxicity.
14 In using the AUC of methanol in blood as the dose metric for derivation of health
15 benchmarks for methanol, the assumption is made that concentrations of methanol in blood over
16 time are related to its toxicity, either through the actions of the parent or it subsequent
17 metabolism. However, the formation of methanol's metabolites has been shown in humans to be
18 carried out by enzymes that are inducible, highly variable in activity, polymorphic, and to also be
19 involved in the metabolism of other drugs and environmental pollutants. Hence, differences in
20 the metabolism of methanol that are specific for target tissue, gender, age, route of
21 administration, and prior exposure to other environmental chemicals may give a different pattern
22 of methanol toxicity if metabolism is required for that toxicity. Eighty-five percent of Asians
23 carry an atypical phenotype of ADH that may affect their ability to metabolize methanol
24 (Agarwal. 2001: Bosron & Li. 1986: Pietruszko. 1980). Also, polymorphisms in ADH3
25 occurring in the promoter region reduce the transcriptional activity in vitro nearly twofold,
26 although no studies have reported differences in ADH3 enzyme activity in humans (Hedberg et
27 al.. 2001).
28 Although data on the specific potential for increased susceptibility to methanol are
29 lacking, there is information on PK and pharmacodynamic factors suggesting that children may
30 have differential susceptibility to methanol toxicity (see Section 4.10.1). Thus, there is
31 uncertainty in children's responses to methanol that should be taken into consideration for
32 derivation of the UF for human variation that is not available from either measured human data
33 or PBPK modeling analyses. The enzyme primarily responsible for metabolism of methanol in
34 humans, ADH, has been reported to be reduced in activity in newborns. Differences in
35 pharmacokinetics include potentially greater pollutant intake due to greater ventilation rates,
March 2011 5-18 DRAFT - DO NOT CITE OR QUOTE
-------
1 activity, and greater intake of liquids in children. In terms of differences in susceptibility to
2 methanol due to pharmacodynamic considerations, the substantial anatomical, physiologic, and
3 biochemical changes that occur during infancy, childhood, and puberty suggest that there are
4 developmental periods in which the endocrine, reproductive, immune, audiovisual, nervous, and
5 other organ systems may be especially sensitive.
6 There are some limited data from short-term exposure studies in humans and animal
7 experiments that suggest differential susceptibility to methanol on the basis of gender. Gender
8 can provide not only different potential targets for methanol toxicity but also differences in
9 methanol pharmacokinetics and pharmacodynamics. NEDO (1987) reported that in rats exposed
10 to methanol pre- and postnatally, 6- and 8-week-old male progeny had significantly lower brain
11 weights at 1,000 ppm, whereas females only showed decreases at 2,000 ppm. In general, gender-
12 related differences in distribution and clearance of methanol may result from the greater muscle
13 mass, larger body size, decreased body fat, and increased volumes of distribution in males
14 compared to females.
15 5.1.3.2.2. Animal-to-human extrapolation UFA. A factor of 3 was applied to account for
16 uncertainties in extrapolating from rodents to humans. Application of a full UF of 10 would
17 depend on two areas of uncertainty: toxicokinetic and toxicodynamic uncertainty. In this
18 assessment, the toxicokinetic component is largely addressed by the determination of a HEC
19 through the use of PBPK modeling. Given the chosen dose metric (AUC for methanol blood),
20 uncertainties in the PBPK modeling of methanol are not expected to be greater for one species
21 than another. The analysis of parameter uncertainty for the PBPK modeling performed for
22 human, mouse, and rat data gave similar results as to how well the model fit the available data.
23 Thus, the human and rodent PBPK model performed similarly using this dose metric for
24 comparisons between species. As discussed in Section 5.3 below, uncertainty does exist
25 regarding the relation of maternal blood levels estimated by the model to fetal and neonatal blood
26 levels that would be obtained under the (gestational, postnatal and lactational) exposure scenario
27 employed in the critical study. However, at environmentally relevant exposure levels, it is
28 assumed that the ratio of the difference in blood concentrations between a human infant and
29 mother would be similar to and not significantly greater than the difference between a rat dam
30 and its fetus. Key parameters and factors which determine the ratio of fetal or neonatal human
31 versus mother methanol blood levels either do not change significantly with age (partition
32 coefficients, relative blood flows) or scale in a way that is common across species
33 (allometrically). For this reason and because EPA has confidence in the ability of the PBPK
34 model to accurately predict adult blood levels of methanol, the PK uncertainty is reduced and a
35 value of 1 was applied. Rodent-to-human pharmacodynamic uncertainty is covered by a factor of
March 2011 5-19 DRAFT - DO NOT CITE OR QUOTE
-------
1 3, as is the practice for deriving RfCs (U.S. EPA, 1994b). Therefore, a factor of 3 is used for
2 interspecies uncertainty.
3 5.1.3.2.3. Database UFD. A database UF of 3 was applied to account for deficiencies in the
4 toxicity database. The database for methanol toxicity is quite extensive: there are chronic and
5 developmental toxicity studies in rats, mice, and monkeys, a two-generation reproductive toxicity
6 study in rats, and neurotoxicity and immunotoxicity studies. However, there is uncertainty
7 regarding which test species is most relevant to humans. In addition, limitations of the
8 developmental toxicity database employed in this assessment include gaps in testing and
9 imperfect study design, reporting, and analyses. Developmental studies were conducted at levels
10 inducing maternal toxicity, a full developmental neurotoxicity test (DNT) in rodents has not been
11 performed and is warranted given the critical effect of decreased brain weight, there are no
12 chronic oral studies in mice, and chronic and developmental studies in monkeys were generally
13 inadequate for quantification purposes, for reasons discussed in Section 5.1.1.1. Problems of
14 interpretation of developmental and reproductive studies also arise given the dose spacing
15 between lowest and next highest level. For these reasons, an UF of 3 was applied to account for
16 deficiencies in the database.
17 5.1.3.2.4. Extrapolation from subchronic to chronic and LOAEL-to-NOAEL extrapolation
18 UFs. A UF was not necessary to account for extrapolation from less than chronic results because
19 developmental toxicity (cervical rib and decreased brain weight) was used as the critical effect.
20 The developmental period is recognized as a susceptible lifestage where exposure during certain
21 time windows is more relevant to the induction of developmental effects than lifetime exposure
22 (U.S. EPA, 1991).
23 A UF for LOAEL-to-NOAEL extrapolation was not applied because BMD analysis was
24 used to determine the POD, and this factor was addressed as one of the considerations in
25 selecting the BMR. In this case, a BMR of one S.D. from the control mean in the critical effect
26 was selected based on the assumption that it represents a minimum biologically significant
27 change.
5.1.4. Previous RfC Assessment
28 The health effects data for methanol were assessed for the IRIS database in 1991 and
29 were determined to be inadequate for derivation of an RfC.
March 2011 5-20 DRAFT - DO NOT CITE OR QUOTE
-------
5.2. ORAL REFERENCE DOSE (RfD)
1 In general, the RfD is an estimate of a daily exposure to 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 BMDL, with uncertainty/variability factors applied to reflect limitations of the data used. The
5 RfD is expressed in terms of mg/kg-day of exposure to an agent and is derived by a similar
6 methodology as is the RfC. Ideally, studies with the greatest duration of exposure and conducted
7 via the oral route of exposure give the most confidence for derivation of an RfD. For methanol,
8 the oral database is currently more limited than the inhalation database. With the development of
9 PBPK models for methanol, the inhalation database has been used to help bridge data gaps in the
10 oral database to derive an RfD.
5.2.1. Choice of Principal Study and Critical Effect-with Rationale and Justification
11 No studies have been reported in which humans have been exposed subchronically or
12 chronically to methanol by the oral route of exposure and thus, would be suitable for derivation
13 of an oral RfD. Data exist regarding effects from oral exposure in experimental animals, but they
14 are more limited than data from the inhalation route of exposure (see Sections 4.2, 4.3, and 4.4).
15 Only 2 oral studies of 90-days duration or longer in animals have been reported (Soffritti
16 et al.. 2002: U.S. EPA. 1986b) for methanol. EPA (1986b) reported that there were no
17 differences in body weight gain, food consumption, or gross or microscopic evaluations in
18 Sprague-Dawley rats gavaged with 100, 500, or 2,500 mg/kg-day versus control animals. Liver
19 weights in both male and female rats were increased, although not significantly, at the 2,500
20 mg/kg-day dose level, suggesting a treatment-related response despite the absence of
21 histopathologic lesions in the liver. Brain weights of high-dose group males and females were
22 significantly less than control animals at terminal (90 days) sacrifice. The data were not reported
23 in adequate detail for dose-response modeling and BMD estimation. Based primarily on the
24 qualitative findings presented in this study, the 500 mg/kg-day dose was deemed to be a
25 NOAEL.55
26 The only lifetime oral study available was conducted by Soffritti et al. (2002) in Sprague-
27 Dawley rats exposed to 0, 500, 5,000, 20,000 ppm (v/v) methanol, provided ad libitum in
28 drinking water. Based on default, time-weighted average body weight estimates for Sprague-
29 Dawley rats (U.S. EPA. 19881 average daily doses of 0, 46.6, 466, and 1,872 mg/kg-day for
55 U.S. EPA (1986b) did not report details required for a BMD analysis such as standard deviations for mean
responses.
March 2011 5-21 DRAFT - DO NOT CITE OR QUOTE
-------
1 males and 0, 52.9, 529, 2,101 mg/kg-day for females were reported by the study authors. All rats
2 were exposed for up to 104 weeks, and then maintained until natural death. The authors report
3 no substantial changes in survival nor was there any pattern of compound-related clinical signs of
4 toxicity. The authors did not report noncancer lesions, and there were no reported compound-
5 related signs of gross pathology or histopathologic lesions indicative of noncancer toxicological
6 effects in response to methanol.
7 Five oral studies investigated the reproductive and developmental effects of methanol in
8 rodents (Aziz et al., 2002; Fu et al., 1996; Infurna & Weiss, 1986; J. M. Rogers, Mole, et al.,
9 1993; Sakanashi etal., 1996), including three studies that investigated the influence of FAD diets
10 on the effects of methanol exposures (Aziz et al., 2002; Fu et al., 1996; Sakanashi etal., 1996).
11 Infurna and Weiss (1986) exposed pregnant Long-Evans rats to 2,500 mg/kg-day in drinking
12 water on either GDIS-GDI? or GD17-GD19. Litter size, pup birth weight, pup postnatal weight
13 gain, postnatal mortality, and day of eye opening were not different in treated animals versus
14 controls. Mean latency for nipple attachment and homing behavior (ability to detect home
15 nesting material) were different in both methanol treated groups. These differences were
16 significantly different from controls. Rogers et al. (1993) exposed pregnant CD-I mice via
17 gavage to 4 g/kg-day methanol, given in 2 equal daily doses. Incidence of cleft palate and
18 exencephaly was increased following maternal exposure to methanol. Also, an increase in totally
19 resorbed litters and a decrease in the number of live fetuses per litter were observed.
20 Aziz et al. (2002). Fu et al. (1996). and Sakanashi et al. (1996) investigated the role of
21 folic acid in methanol-induced developmental neurotoxicity. Like Rogers et al. (1993), the
22 former 2 studies observed that an oral gavage dose of 4-5 g/kg-day during GD6-GD15 or
23 GD6-GD10 resulted in an increase in cleft palate in mice fed sufficient folic acid diets, as well as
24 an increase in resorptions and a decrease in live fetuses per litter. Fu et al. (1996) also observed
25 an increase in exencephaly in the FAS group. Both studies found that an approximately 50%
26 reduction in maternal liver folate concentration resulted in an increase in the percentage of litters
27 affected by cleft palate (as much as threefold) and an increase in the percentage of litters affected
28 by exencephaly (as much as 10-fold). Aziz et al. (2002) exposed rat dams throughout their
29 lactation period to 0, 1, 2, or 4% v/v methanol via the drinking water, equivalent to
30 approximately 480, 960 and 1,920 mg/kg-day.56 Pups were exposed to methanol via lactation
31 from PND1-PND21. Methanol treatment at 2% and 4% was associated with significant
32 increases in activity (measured as distance traveled in a spontaneous locomotor activity test) in
56 Assuming that Wistar rat drinking water consumption is 60 mL/kg-day (V. V. Rogers et al.. 2002X 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.
March 2011 5-22 DRAFT - DO NOT CITE OR QUOTE
-------
1 the FAS group (13 and 39%, respectively) and most notably, in the FAD group (33 and 66%,
2 respectively) when compared to their respective controls. At PND45, the CAR in FAD rats
3 exposed to 2% and 4% methanol was significantly decreased by 48% and 52%, respectively,
4 relative to nonexposed controls. In the FAS group, the CAR was only significantly decreased in
5 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
6 Given the oral database limitations, including the limited reporting of noncancer findings
7 in the subchronic (U.S. EPA. 1986b) and chronic studies (Soffritti et al.. 2002) of rats and the
8 high-dose levels used in the two rodent developmental studies, EPA has derived an RfD by using
9 relevant inhalation data and route-to-route extrapolation with the aid of the EPAPBPK model
10 (see Sections 3.4 and 5.1). Several other factors support use of route-to-route extrapolation for
11 methanol. The limited data for oral administration indicate similar effects as reported via
12 inhalation exposure (e.g., the brain and fetal skeletal system are targets of toxicity). Methanol
13 has been shown to be rapidly and well-absorbed by both the oral and inhalation routes of
14 exposure (CERHR, 2004; Kavet & Nauss, 1990). Once absorbed, methanol distributes rapidly to
15 all organs and tissues according to water content, regardless of route of exposure.
16 As with the species-to-species extrapolation used in the development of the RfC, the dose
17 metric used for species-to-species and route-to-route extrapolation of inhalation data to oral data
18 is the AUC of methanol in blood. Simulations for human oral methanol exposure were
19 conducted using the model parameters as previously described for human inhalation exposures,
20 with human oral kinetic/absorption parameters from Sultatos et al. (2004, 090530) (i.e., KAS =
21 0.2, KSI = 3.17, and KAI = 3.28). Human oral exposures were assumed to occur during six
22 drinking episodes during the day, at times 0, 3, 5, 8, 11, and 15 hours from the first ingestion of
23 the day. For example, if first ingestion occurred at 7 am, these would be at 7 am, 10 am, 12
24 noon, 3 pm, 6 pm, and 10 pm. Each ingestion event was treated as occurring over 3 minutes,
25 during which the corresponding fraction of the daily dose was infused into the stomach lumen
26 compartment. The fraction of the total ingested methanol simulated at each of these times was
27 25%, 10%, 25%, 10%, 25%, and 5%, respectively. Six days of exposure were simulated to allow
28 for any accumulation (visual inspection of plots showed this to be finished by the 2nd or 3rd
29 day), and the results for the last 24 hours were used. Dividing the exposure into more and
30 smaller episodes would decrease the estimated peak concentration but have little effect on AUC.
31 This dose metric was used for dose-response modeling to derive the POD, expressed as a BMDL.
32 The BMDL was then back-calculated using the EPAPBPK model to obtain an equivalent oral
33 drinking water dose in terms of mg/kg-day.
March 2011 5-23 DRAFT - DO NOT CITE OR QUOTE
-------
5.2.2. RfD Derivation-Including Application of UFs
5.2.2.1. Consideration of Inhalation Data
1 Inhalation studies considered for derivation of the RfC are used to supplement the oral
2 database using the route-to-route extrapolation, as previously described. BMD approaches were
3 applied to the existing inhalation database, and the EPAPBPK model was used for species-to-
4 species extrapolations. The rationale and approach for determining the RfC is described above
5 (Section 5.1), and the data used to support the derivation of the RfC were extrapolated using the
6 EPA PBPK model to provide an oral equivalent POD.
5.2.2.2. Selection of Critical Effect(s) from Inhalation Data
7 Methanol-induced effects on the brain in rats (weight decrease) and fetal axial skeletal
8 system in mice (cervical ribs and cleft palate) were consistently observed at lower levels, than
9 other targets, in the oral and inhalation databases. Analysis of inhalation developmental toxicity
10 studies shows lower BMDLs for decreased male brain weight in rats exposed throughout
11 gestation and the FI generation (NEDO, 1987) than BMDLs associated with the fetal axial
12 skeletal system in mice (see Section 5.1.3.1). Therefore, the BMDL for decreases in brain weight
13 in male rats is chosen to serve as the basis for the route-to-route extrapolation and calculation of
14 the RfD.
5.2.2.3. Selection of the POD
15 The BMDL chosen for the RfC is used to determine the POD for the RfD. This value is
16 based on a developmental toxicity dataset that includes in utero and postnatal exposures and is
17 below the range of estimates for other developmental datasets consisting of exposure only
18 throughout organogenesis. The neonatal brain is the target organ chosen for derivation of the
19 RfC. The BMDL for the RfC (AUC of 90.9 hr x mg/L methanol in blood) is converted using the
20 EPA model to a human equivalent oral exposure of 3 8.6 mg/kg-day.57
5.2.3. RfD Derivation-Application of UFs
21 In an approach consistent with the RfC derivation, UFs are applied to the oral POD of
22 38.6 mg/kg-day to address interspecies extrapolation, intraspecies variability, and database
23 uncertainties for the RfD. Because the same dataset, endpoint, and PBPK model used to derive
24 the RfC were also used to calculate the oral POD, the total UF of 100 is applied to the BMDL of
25 38.6 mg/kg-day to yield an RfD of 0.4 mg/kg-day for methanol.
57 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.
March 2011 5-24 DRAFT - DO NOT CITE OR QUOTE
-------
1 RfD = 38.6 mg/kg-day + 100 = 0.4 mg/kg-day (rounded to one significant figure)
5.2.4. Previous RfD Assessment
2 The previous IRIS assessment for methanol included an RfD of 0.5 mg/kg-day that was
3 derived from a EPA (1986b) subchronic oral study in which Sprague-Dawley rats (30/sex/dose)
4 were gavaged daily with 0, 100, 500, or 2,500 mg/kg-day of methanol. There were no
5 differences between dosed animals and controls in body weight gain, food consumption, gross or
6 microscopic evaluations. Elevated levels of SGPT, serum alkaline phosphatase (SAP), and
7 increased but not statistically significant liver weights in both male and female rats suggest
8 possible treatment-related effects in rats dosed with 2,500 mg methanol/kg-day, despite the
9 absence of supportive histopathologic lesions in the liver. Brain weights of both high-dose group
10 males and females were significantly less than those of the control group. Based on these
11 findings, 500 mg/kg-day of methanol was considered a NOAEL in this rat study. Application of
12 a 1,000-fold UF (interspecies extrapolation, susceptible human subpopulations, and subchronic
13 to chronic extrapolation) yielded an RfD of 0.5 mg/kg-day.
5.3. UNCERTANTIES IN THE INHALATION RFC AND ORAL RFD
14 The following is a more extensive discussion of the uncertainties associated with the RfC
15 and RfD for methanol beyond that which is addressed quantitatively in Sections 5.1.2, 5.1.3, and
16 5.2.2. A summary of these uncertainties is presented in Table 5-5.
March 2011 5-25 DRAFT - DO NOT CITE OR QUOTE
-------
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
APBPKmodelwas
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, 2006b) 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). Burbacher, Grant et al.
(2004) and Burbacher, Shen et al. (2004)
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.
March 2011
5-26 DRAFT - DO NOT CITE OR QUOTE
-------
5.3.1. Choice of Endpoint
1 The impact of endpoint selection (on brain weight decrease in male rats) the derivation of
2 the RfC and RfD was discussed in Sections 5.1.3.1 and 5.2.2.2. Potential RfC values considered
3 ranged from 1.7 to 13.6 mg/m3, depending on whether neurobehavioral function in male
4 monkeys, brain weight decrease in male rats, or cervical ribs incidence in mice was chosen as the
5 critical effect for derivation of the POD, with the former endpoint representing the lower end of
6 the RfC range. The use of other endpoints, particularly pre-term births identified in the
7 Burbacher et al. (1999; 2004; 1999; 2004) monkey study, would potentially result in lower
8 reference values, but significant uncertainties associated with those studies preclude their use as
9 the basis for an RfC.
10 Burbacher et al. (1999; 2004; 1999; 2004) exposed M. fascicularis monkeys to 0, 262,
11 786, and 2,359 mg/m3 methanol 2.5 hours/day, 7 days/week during premating/mating and
12 throughout gestation (approximately 168 days). They observed a slight but statistically
13 significant gestation period shortening in all exposure groups that was largely due to C-sections
14 performed in the methanol exposure groups "in response to signs of possible difficulty in the
15 maintenance of pregnancy," including vaginal bleeding. As discussed in Sections 4.3.2 and
16 5.1.1.2, there are questions concerning this effect and its relationship to methanol exposure. An
17 ultrasound was not done to confirm the existence of real fetal or placental problems.
18 Neurobehavioral function was assessed in infants during the first 9 months of life. Two tests out
19 of nine, returned positive results possibly related to methanol exposure. VDR performance was
20 reduced in all treated male infants, and was significantly reduced in the 2,359 mg/m3 group for
21 both sexes and the 786 mg/m3 group for males. However, an overall dose-response trend for this
22 endpoint was only observed in females. As discussed in Section 4.4.2, confidence in this
23 endpoint may have been increased by statistical analyses to adjust for multiple testing (CERHR,
24 2004), but it is a measure of functional deficits in sensorimotor development that is consistent
25 with early developmental CNS effects (brain weight changes discussed above) that have been
26 observed in rats. The Pagan test of infant intelligence indicated small but not significant deficits
27 of performance (time spent looking a novel faces versus familiar faces) in treated infants.
28 Although these results indicate that prenatal and continuing postnatal exposure to methanol can
29 result in neurotoxicity to the offspring, especially when considered in conjunction with the gross
30 morphological effects noted in NEDO (1987), the use of such statistically borderline results is
31 not warranted in the derivation of the RfC, given the availability of better dose-response data in
32 other species.
33 NEDO (1987) also examined the chronic neurotoxicity of methanol inM fascicularis
34 monkeys exposed to 13.1, 131, or 1,310 mg/m3 for up to 29 months. Multiple effects were noted
March 2011 5-27 DRAFT - DO NOT CITE OR QUOTE
-------
1 at 131 mg/ m3, including slight myocardial effects (negative changes in the T wave on an EKG),
2 degeneration of the inside nucleus of the thalamus, and abnormal pathology within the cerebral
3 white tissue in the brain. The results support the identification of 13.1 mg/m3 as the NOAEL for
4 neurotoxic effects in monkeys exposed chronically to inhaled methanol. However, as discussed
5 in Section 4.2.2.3, there exists significant uncertainty in the interpretation of these results and
6 their utility in deriving an RfC for methanol. These uncertainties include lack of appropriate
7 control group data, limited nature of the reporting of the neurotoxic effects observed, and use of
8 wild-caught monkeys in the study. Thus, while the NEDO (1987) study suggests that monkeys
9 may be a more sensitive species to the neurotoxic effects of chronic methanol exposure than
10 rodents, the substantial deficits in the reporting of data preclude the quantification of data from
11 this study for the derivation of an RfC.
12 The increased incidence of cervical ribs was identified as a biologically significant,
13 potential co-critical effect based on the findings of Rogers et al. (1993). Mice were exposed to
14 1,000, 2,000, or 5,000 ppm, and incidence of cervical ribs was statistically increased at
15 2,000 ppm. However, given that the reference values for the increased incidence of cervical ribs
16 are estimated to be approximately five times higher than the reference values calculated using
17 decreases in brain weight in male rats (NEDO, 1987) decreased brain weight was chosen as the
18 basis for the derivation of the RfC.
5.3.2. Choice of Dose Metric
19 A recent review of the reproductive and developmental toxicity of methanol by a panel of
20 experts concluded that methanol, not its metabolite formate, is likely to be the proximate
21 teratogen and that blood methanol level is a useful biomarker of exposure (CERHR, 2004;
22 Dorman et al., 1995). The CERHR Expert Panel based their assessment of potential methanol
23 toxicity on an assessment of circulating blood levels (CERHR, 2004). In contrast to the
24 conclusions of the NTP-CERHR panel, in vitro data from Harris et al. (2004: 2003) suggest that
25 the etiologically important substance for embryo dysmorphogenesis and embryolethality was
26 likely to be formaldehyde rather than the parent compound or formate. Although there remains
27 uncertainty surrounding the identification of the proximate teratogen of importance (methanol,
28 formaldehyde, or formate), the dose metric chosen for derivation of an RfC was based on blood
29 methanol levels. This decision was primarily based on evidence that the toxic moiety is not
30 likely to be the formate metabolite of methanol (CERHR, 2004), and evidence that levels of the
31 formaldehyde metabolite following methanol maternal and/or neonate exposure would be lower
32 in the fetus and neonate than in adults. While recent in vitro evidence indicates that
33 formaldehyde is more embryotoxic than methanol and formate, the high reactivity of
March 2011 5-28 DRAFT - DO NOT CITE OR QUOTE
-------
1 formaldehyde would limit its unbound and unaltered transport as free formaldehyde from
2 maternal to fetal blood (Thrasher & Kilburn, 2001) (see discussion in Section 3.3). Thus, even if
3 formaldehyde is ultimately identified as the proximate teratogen, methanol would likely play a
4 prominent role, at least in terms of transport to the target tissue. Further discussions of methanol
5 metabolism, dose metric selection, and MOAissues are in Sections 3.3, 4.6, and 4.8.
6 There exists some concern in using the FI generation NEDO (1987) rat study as the basis
7 from which to derive the RfC. This concern mainly arises from issues related to the low
8 confidence that the PBPK model is accurately predicting dose metrics for neonates exposed
9 through multiple and simultaneous routes. The PBPK model was structured to predict internal
10 dose metrics for adult NP animals and was optimized using adult metabolic and physiological
11 parameters. Young animals have very different metabolic and physiological profiles than adults
12 (enzyme activities, respiration rates, etc.). This fact, coupled with multiple routes of exposure,
13 make it likely that the PBPK did not accurately predict the internal dose metrics for the offspring.
14 Stern et al. (1996) reported that when rat pups and dams were exposed together during lactation
15 to 4,500 ppm methanol in air, methanol blood levels in pups from GD6-PND21 were
16 approximately 2.25 times greater than those of dams. This discrepancy persisted until PND48,
17 when postnatal exposure continued to PND52. It is logical to assume that similar differences in
18 blood methanol levels would also be observed in the NEDO (1987) FI study, as the exposure
19 scenario is similar to that of Stern et al. (1996). Differences between pup and dam blood
20 methanol levels might be expected to be slightly greater than twofold in the NEDO (1987) FI
21 study as the exposure was continuous (versus 6 hours/day in the Stern et al. (1996) paper) and
22 lasted for a longer duration (-64 days versus 37). Under a similar scenario, human newborns
23 may experience higher blood levels than their mothers as a result of breast feeding. As has been
24 discussed in Chapter 3, children have a limited capacity to metabolize methanol via ADH;
25 however, there is some evidence that human infants are able to efficiently eliminate methanol at
26 high-exposure levels, possibly via CAT (Tran et al., 2007). At environmentally relevant exposure
27 levels, it is assumed that the ratio of the difference in blood concentrations between infant and
28 mother would not be significantly greater than the twofold difference that has been observed in
29 rats.58 For this reason and because EPA has confidence in the ability of the PBPK model to
30 accurately predict adult blood levels of methanol, the maternal blood methanol levels for the
31 estimation of HECs from the NEDO (1987) study were used as the dose metric.
58 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).
March 2011 5-29 DRAFT - DO NOT CITE OR QUOTE
-------
5.3.3. Choice of Model for BMDL Derivations
1 The Hill model adequately fit the dataset for the selected endpoint (goodness-of-fit/?-
2 value = 0.84). Data points were well predicted near the BMD (scaled residual = 0.09) (see
3 Figure 5-1). There is a 2.5-fold range of BMDL estimates from adequately fitting models,
4 indicating considerable model dependence. The BMDL from the Hill model was selected, in
5 accordance with EPA BMD Technical Guidance (2000a), because it results in the lowest BMDL
6 from among a broad range of BMDLs and provides a superior fit in the low dose region nearest
7 the BMD.
5.3.4. Choice of Animal-to-Human Extrapolation Method
8 APBPK model developed by the EPA, adapted from Ward et al. (1997), was used to
9 extrapolate animal-to-human concentrations. An AUC blood level of methanol (90.9 hr x mg/L)
10 associated with a one S.D. change from the control mean for brain weights in rats was estimated
11 using the rat PBPK model. Then the human PBPK model was used to convert back to a human
12 equivalent exposure concentration or a BMCLHEc/iso of 182 mg/m3. If no PBPK models were
13 available, a BMCLnEc/iso of 424 mg/m3 would have been derived by adjusting the 556.5 mg/m3
14 BMCLiso for external exposure concentration for duration and the animal-to-human standard
15 adjustment factor for systemic effects (the ratio of animal and human blood:air partition
16 coefficients). This value is approximately twofold higher than the value derived using the PBPK
17 model. However, as discussed above, use of PBPK-estimated maternal blood methanol levels for
18 the estimation of HECs allows for the use of data-derived extrapolations rather than standard
19 methods for extrapolations from external exposure levels.
20 As discussed in Section 3.4, the PBPK models do not describe or account for background
21 levels of methanol, formaldehyde or formate, and background levels were subtracted from the
22 reported data before use in model fitting or validation (if not already subtracted by study authors),
23 as described below. This approach was taken because the relationship between background doses
24 and background responses is not known, because the primary purpose of this assessment is for
25 the determination of noncancer risk associated with increases in the levels of methanol or its
26 metabolites (e.g., formate, formaldehyde) over background, and because the subtraction of
27 background levels is not expected to have a significant impact on PBPK model parameter
28 estimates (see further discussion in Section 3.4.3.2).
March 2011 5-30 DRAFT - DO NOT CITE OR QUOTE
-------
5.3.5. Route-to-Route Extrapolation
1 To estimate an oral dose POD for decrease in brain weight in rats, a route-to-route
2 extrapolation was performed on the inhalation exposure POD used to derive the RfC. One way
3 to characterize the uncertainty associated with this approach is to compare risk levels (BMDL
4 values) using the dose metric, AUC methanol, for developmental decreases in brain weight
5 derived from 1) an existing oral subchronic study and 2) from a model estimating this metric
6 from an existing inhalation subchronic study. There are currently no oral developmental studies
7 investigating decreases in brain weight available to compare to the risk values estimated using
8 the second procedure. However, the fact that the oral BMDL of 38.6 mg/kg-day estimated in this
9 assessment from the NEDO (1987) inhalation study of neonate rats via a PBPK model is lower
10 than the NOAEL of 500 mg/kg-day identified in EPA (1986b) methanol study of adult rats is
11 consistent with other studies which suggest that fetal/neonatal organisms are a sensitive
12 subpopulation.
5.3.6. Statistical Uncertainty at the POD
13 There is uncertainty in the selection of the BMR level. For decreased brain weight in rats,
14 no established standard exists, so a BMR of one S.D. change from the control mean was used.
15 Parameter uncertainty can be assessed through CIs. Each description of parameter uncertainty
16 assumes that the underlying model and associated assumptions are valid. For the Hill model
17 applied to the data for decreased brain weight in rats, there is a degree of uncertainty at the one
18 S.D. level (the POD for derivation of the RfC), with the 95% one-sided lower confidence limit
19 (BMDL) being -50% below the maximum likelihood estimate of the BMD.
5.3.7. Choice of Bioassay
20 The NEDO (1987) study was used for development of the RfC and RfD because it
21 resulted in the lowest BMDL. It was also a well-designed study, conducted in a relevant species
22 with an adequate number of animals per dose group, and with examination of appropriate
23 developmental toxicological endpoints. Developmental (Burbacher, Grant, et al., 1999;
24 Burbacher, Grant, et al., 2004; Burbacher, Shen, et al., 1999; Burbacher, Shen, et al., 2004) and
25 chronic studies (NEDO, 1987) of methanol have been performed in monkeys. As discussed
26 above in Section 5.3.1 and other sections of this assessment, while the monkey may be a
27 sensitive species for use in the determination of human risk, reporting deficits and study
28 uncertainties preclude their use in the derivation of an RfC.
March 2011 5-31 DRAFT - DO NOT CITE OR QUOTE
-------
5.3.8. Choice of Species/Gender
1 The RfC and RfD were based on decreased brain weight at 6 weeks postbirth in male rats
2 (the gender most sensitive to this effect) (NEDO, 1987). This decrease in brain weight also
3 occurs in female rats; however, if the decreased brain weight in female rats had been used, higher
4 RfC and RfD values would have been derived (approximately 66% higher than the male derived
5 values).
5.3.9. Human Population Variability
6 The extent of interindividual variation of methanol metabolism in humans has not been
7 well characterized. As discussed in Section 4.10, there are a number of issues that may lead to
8 sensitive human subpopulations. Potentially sensitive subpopulations would include individuals
9 with polymorphisms in the enzymes involved in the metabolism of methanol and individuals
10 with significant folate deficiencies. Sensitive lifestages would include children and neonates, as
11 they have increased respiration rates compared to adults, which may increase their methanol
12 blood levels compared to adults. Also, children have been shown to have decreased ADH
13 activity relative to adults, thus decreasing their ability to metabolize and eliminate methanol. As
14 demonstrated by these examples, there exists considerable uncertainty pertaining to human
15 population variability in methanol metabolism, which provides justification for the 10-fold
16 intraspecies UF used to derive the RfC and RfD.
5.4. CANCER ASSESSMENT
17 A cancer dose-response estimation is not addressed in this document. However, the Agency is
18 currently reviewing the literature and will develop a cancer assessment for methanol at a later
19 date.
March 2011 5-32 DRAFT - DO NOT CITE OR QUOTE
-------
6. MAJOR CONCLUSIONS IN 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 potential
7 for human exposure exists today in the form of the artificial sweetener, aspartame, which is a
8 methyl ester of the dipeptide aspartyl-phenylalanine. Methanol is the major anti-freeze
9 constituent of windshield washer fluid. Its use as a fuel additive for internal combustion engines
10 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 CC>2, 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 can result from relatively low doses due to metabolic acidosis that appears to affect
22 predominantly the nervous system, with potentially lasting effects such as blindness, Parkinson-
23 like symptoms, and cognitive impairment. These effects can be observed in humans when blood
24 methanol levels exceed 200 mg/L. The species differences in toxicity from acute exposures
25 appear to be the result of a limited ability of humans to metabolize formic acid.
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 6-1 DRAFT - DO NOT CITE OR QUOTE
-------
1 Despite the existence of many case reports on acute human exposures, the knowledge
2 base for long-term, low-level exposure of humans to methanol is limited. The current TLV for
3 methanol is 200 ppm (262 mg/m3) (ACGIH, 2000). Controlled experiments with human
4 volunteers indicate that only minor neurobehavioral changes occur following 4-hour exposure to
5 this concentration. A limited study on self-reported health effects in 66 persons exposed to
6 methanol at levels that came close to or exceeded the NIOSH short-term ceiling of 800 ppm
7 (1048 mg/m3), in comparison with an age-matched group of 66 less or not exposed persons,
8 suggested a statistically significant increase in the incidence of CNS-related symptoms, such as
9 dizziness, nausea, headache, and blurred vision (Frederick et al., 1984). Impaired vision and
10 nasal irritation were observed in a study of 33 methanol-exposed workers (Kawai et al., 1991).
11 A number of reproductive, developmental, subchronic and chronic exposure duration
12 studies have been conducted in mice, rats, and monkeys. This summary will focus primarily on
13 reproductive and developmental toxicity as the main endpoints of concern. Sections 4.7, 5.1.1
14 and 5.2.1 contain more extensive summaries that consider the dose-related effects that have been
15 observed in other organ systems following subchronic or chronic exposure.
16 Although there is no evidence in humans, methanol has shown to be a reproductive and
17 developmental toxicant in several animal studies. No studies have been reported in which
18 humans have been exposed subchronically or chronically to methanol by the oral route of
19 exposure, and thus would be suitable for derivation of an oral RfD. Data exist regarding effects
20 from oral exposure in experimental animals, but they are more limited than data from the
21 inhalation route of exposure (see Sections 4.2, 4.3, and 4.4). Two oral studies in rats (Soffritti et
22 al., 2002; U.S. EPA, 1986b), one oral study in mice (Apaja, 1980) and several inhalation studies
23 in monkeys, rats and mice (NEDO, 1987, 2008a, 2008b) of 90-days duration or longer have been
24 reported. While some noncancer effects of methanol exposure were noted in these studies,
25 principally in the liver and brain, they were either not quantifiable due to study limitations or
26 occurred at high doses relative to reproductive/developmental effects. As discussed below, the
27 results of inhalation reproductive/developmental toxicity studies in rats (NEDO, 1987), mice (I
28 M. Rogers, Mole, et al., 1993), and monkeys (Burbacher, Grant, et al., 1999; Burbacher, Grant, et
29 al., 2004: Burbacher, Shen, et al., 1999: Burbacher, Shen, et al., 2004) are the principal
30 considerations for both the RfD and RfC values derived in this assessment.
31 A larger number of studies have used the inhalation route to assess the potential of
32 reproductive or developmental toxicity of methanol in mice, rats, and monkeys, with
33 concentrations ranging from 200 to 20,000 ppm (blood levels reaching as high as 8.65 mg/mL).
34 To sum up the findings, rat dams survived even the highest doses without gross signs of toxicity,
35 but their offspring were severely affected (Nelson etal., 1985). Two more inhalation studies,
March 2011 6-2 DRAFT - DO NOT CITE OR QUOTE
-------
1 Rogers et al. (1993; 1993) and Rogers and Mole (1997), confirmed that methanol causes
2 exencephaly and cleft palate in mice, the most sensitive days being GD6 and GD7 (i.e., early
3 organogenesis). These severe malformations were observed at exposure concentrations of
4 5,000 ppm or above. Nelson et al. (1985) and Rogers et al. (1993) also observed an increased
5 occurrence of ossification disturbances and skeletal anomalies at methanol concentrations
6 > 2,000 ppm, of which cervical ribs in mouse fetuses is considered the critical effect for toxicity
7 value derivation in this review. A study conducted in pregnant cynomolgus monkeys that were
8 exposed to 200-600 ppm methanol for 2.5 hours/day throughout premating, mating, and gestation
9 showed no signs of maternal or fetal toxicity. The potential compound-related effects noted were
10 a shortening of the gestation period by less than 5% and developmental neurotoxicity,
11 particularly delayed sensorimotor development monkeys (Burbacher, Grant, et al., 1999;
12 Burbacher. Grant et al.. 2004: Burbacher. Shen. et al.. 1999: Burbacher. Shen. et al.. 2004).
13 While all of the above studies were conducted with exposure durations of 7 hours/day or
14 less, NEDO (1987) conducted a series of developmental/reproductive studies in rats that used
15 exposure times of 20 hours/day or more at concentrations between 500 and 5,000 ppm. Atwo-
16 generation study by these researchers that exposed the dams throughout pregnancy and the pups
17 through 8 weeks of age, demonstrated dose-dependent reductions in brain weights that forms the
18 basis for the RfC derived in this review.
6.2. DOSE RESPONSE
19 As described in Chapter 3, background levels of methanol and its metabolites are
20 produced through endogenous metabolic processes. Potential risks resulting from these
21 endogenous levels are not determined in this IRIS assessment. This assessment focuses on the
22 determination of noncancer risk associated with exogenous methanol exposures that increase the
23 body burden of methanol or its metabolites (e.g., formate, formaldehyde) above endogenous
24 background levels. Average background blood levels in healthy adults following restriction of
25 methanol-producing foods from the diet are reported in Section 3.1 (Table 3-1). The mouse, rat
26 and human PBPK models developed for this assessment predict increased blood levels of
27 methanol and its metabolites over background following oral or inhalation exposure to methanol
28 (see further discussion in Section 3.4.3.2). Consequently, this assessment provides estimates of
29 noncancer risk from oral and inhalation exposures above sources of methanol that contribute to
30 background blood levels.
March 2011 6-3 DRAFT - DO NOT CITE OR QUOTE
-------
6.2.1. Noncancer/Inhalation
1 Clearly defined toxic endpoints at moderate exposure levels have been observed only in
2 reproductive and developmental toxicity studies. Three endpoints from developmental toxicity
3 studies were considered for derivation of the RfC: formation of cervical ribs in CD-I mice
4 exposed to methanol during organogenesis (J. M. Rogers, Mole, et al., 1993), deficits in
5 sensorimotor development as measured by VDR tests administered to monkeys exposed to
6 methanol monkeys (Burbacher, Grant, et al., 1999; Burbacher, Grant, et al., 2004; Burbacher,
7 Shen, et al., 1999; Burbacher, Shen, et al., 2004), and reduced brain weights in rats exposed to
8 methanol from early gestation through 8 weeks of postnatal life (NEDO, 1987). For the purpose
9 of comparability and to better illustrate methodological uncertainty, reference values were
10 derived for all of these endpoints using a BMD modeling approach which evaluated several
11 models and various measures of risk. In the present review, mostly because of a paucity of
12 adequate long-term or developmental oral studies and the existence of several inhalation studies
13 that examined sensitive subpopulations (pregnant mothers, developing fetuses and neonates) in
14 various species, it was decided to use the critical effect from an inhalation study to derive an
15 RfD. Thus, the criteria and rationales on which the RfC assessment is based also form the basis
16 for the RfD derivation.
17 The Rogers et al. (1993) inhalation study is a multidose developmental study that was
18 considered for use in the derivation of a reference value. The exposure concentrations in this
19 study were 0, 1,000, 2,000, and 5,000 ppm administered for 7 hours/day on GD7-GD17. The
20 BMD evaluation, based on the nested log-logistic model of BMDS version 2.1.1 (U.S. EPA,
21 2009a), produced BMD/BMDL values in terms of internal peak blood methanol (Cmax). PBPK
22 modeling was used to convert the internal animal dose metrics to HECs, and a UF of 100 was
23 applied to yield RfCs of 10.4 mg/m3 and 13.6 mg/m3 for 5 and 10% extra risk, respectively.
24 Reproductive and developmental neurobehavioral effects observed in monkeys following
25 methanol inhalation exposure monkeys (Burbacher, Grant, et al., 1999; Burbacher, Grant, et al.,
26 2004; Burbacher, Shen, etal., 1999; Burbacher, Shen, et al., 2004) were also considered for use
27 in the derivation of a reference value. M. fascicularis monkeys were exposed to 0, 262, 786, and
28 2,359 mg/m3 methanol 2.5 hours/day, 7 days/week during premating/mating and throughout
29 gestation (approximately 168 days). Delayed sensorimotor development as measured by a VDR
30 test was the only effect in this study that exhibited a dose-response and is a measure of a
31 functional deficit that is consistent with early developmental CNS effects (e.g., brain weight
32 changes) that have been observed in rats (NEDO, 1987). Though there is uncertainty associated
33 with this effect and its relation to methanol exposure, a BMD analysis was performed for
34 comparative purposes. BMD/BMDL values for the VDR endpoint were estimated using AUCs
March 2011 6-4 DRAFT - DO NOT CITE OR QUOTE
-------
1 derived from a monkey PBPK model of blood methanol data reported in the Burbacher et al.
2 (1999) monkeys study. A human methanol PBPK modeling was then used to convert the internal
3 AUC BMDL to an HEC, and a UF of 100 was applied to yield a reference value estimate of 1.7
4 mg/m3.
5 Reduced brain weight was evaluated based on the results of a two-generation study by
6 NEDO (1987) in which fetal rats and their dams were exposed from the first day of gestation
7 until 8 weeks of age, and brain weights were determined at 3, 6, and 8 weeks of age. To obtain
8 reference value estimates from these studies, a rat PBPK model was used to predict PODs in
9 terms of internal doses, which were converted to HEC values via a human PBPK model and
10 divided by UFs (see Table 5-4). BMD modeling was executed using two different BMRs, one
11 S.D. (as is usual with continuous data) and 5% relative (to control response) risk. The resulting
12 reference value estimates were 2.4 and 1.8 mg/m3 (5% relative risk and 1 S.D., respectively) for
13 reduced brain weight at 6 weeks of age following gestational and postnatal exposure.
14 Despite the variety of approaches, different critical effects, and different data sources, all
15 reference value estimates fell within a narrow range. The reference value associated with the
16 BMD estimate of the dose corresponding to a one S.D. decrease in brain weight in male rats at 6
17 weeks post-birth observed in the NEDO (1987) developmental toxicity study is considered most
18 suitable for derivation of the methanol chronic RfC due to the relevance of the exposure
19 scenario/study design and endpoint (see Sections 5.1.2.2 and 5.3) to the potential for
20 developmental effects in neonatal humans, the relative robustness of the dose
21 response data and because it resulted in one of the lowest reference values of the BMD
22 derivations (see Table 5-4). Thus, the proposed chronic RfC for exposure to methanol is
23 2 mg/m3, an evaluation that includes a UFH of 10 for intraspecies variability, a UFA of 3 to
24 address the pharmacodynamic component of interspecies variability, and a UFD of 3 for database
25 uncertainty.
26 The confidence in this RfC is medium to high. Confidence in the NEDO (1987)
27 developmental studies is medium to high. While there are issues with the lack of reporting
28 detail, the critical effect (brain weight reduction) has been reproduced in an oral study of adult
29 rats (U.S. EPA, 1986b), and the exposure regimen involving pre- and postnatal exposures
30 addresses a potentially sensitive human subpopulation. Confidence in the database is medium.
31 Despite the fact that skeletal and brain effects have been demonstrated and corroborated in
32 multiple animal studies in rats, mice, and monkeys, some study results were not quantifiable,
33 there is uncertainty regarding which is the most relevant test species, and there is limited data
34 regarding reproductive or developmental toxicity of methanol in humans. There is also
35 uncertainty regarding the potential active agent—the parent compound, methanol, formaldehyde,
March 2011 6-5 DRAFT - DO NOT CITE OR QUOTE
-------
1 or formic acid. There are deficiencies in our knowledge of the metabolic pathways of methanol
2 in the human fetus during early organogenesis, when the critical effects can be induced in
3 animals. Thus, the medium-to-high confidence in the critical study and the medium confidence
4 in the database together warrant an overall confidence descriptor of medium to high.
6.2.2. Noncancer/Oral
5 There is a paucity of scientific data regarding the outcomes of chronic oral exposure to
6 methanol. No data exist for long-term methanol exposure of humans. A subchronic (90-day)
7 oral study in Sprague-Dawley rats reported brain and liver weight changes, with some evidence
8 for minor liver damage at 2,500 mg/kg-day that was not supported by histopathologic findings
9 (U.S. EPA, 1986b). Liver necrosis was reported in Eppley Swiss Webster mice that consumed
10 approximately 2000 mg/kg-day (Apaja, 1980). In the only other study that administered
11 methanol chronically to animals by the oral route, Soffritti et al. (2002) reported that, overall,
12 there was no pattern of compound-related clinical signs of toxicity in Sprague-Dawley rats
13 exposed to up to approximately 2,000 mg/kg-day. The authors further reported that there were no
14 compound-related signs of gross pathology nor histopathologic lesions indicative of noncancer
15 toxicological effects in response to methanol; however, they did not provide any detailed data to
16 illustrate these findings.
17 As discussed above and in Section 5.1.1, reproductive and developmental effects are
18 considered the most sensitive and quantifiable effects reported in studies of methanol. Oral
19 reproductive and developmental studies employed single doses that were too high to be of use.
20 In the absence of suitable reproductive or developmental data from oral exposure studies, it was
21 decided to conduct a route-to-route extrapolation and to use the critical effect from the inhalation
22 study (brain weight) to derive an RfD. Thus, the POD (in terms of AUC methanol in blood) used
23 for the derivation of the RfC was also used for the derivation of the RfD. This POD was
24 converted to an HED via a human PBPK model and divided by a UF of 100 to obtain an RfD
25 value of 0.4 mg/kg-day. As for the RfC, the 100-fold UF includes a UFH of 10 for intraspecies
26 variability, a UFA of 3 to address pharmacodynamic uncertainty, and a UFD of 3 for database
27 uncertainty.
28 The confidence in the RfD is medium to high. Despite the relatively high confidence in
29 the critical studies, all limitations to confidence as presented for the RfC also apply to the RfD.
30 Confidence in the RfD is slightly lower than for the RfC due to the lack of adequate oral studies
31 for the RfD derivation, necessitating a route-to-route extrapolation.
March 2011 6-6 DRAFT - DO NOT CITE OR QUOTE
-------
7. REFERENCES
ACGffl. (American Conference of Governmental Industrial Hygienists). (2000). 2000 TLVs®
andBEIs®: based on the documentations of the threshold limit values for chemical
substances and physical agents and biological exposure indices. Cincinnati, OH: Author.
Adanir, J., Ozkalkanti, M. Y, & Aksun, M. (2005). Percutaneous methanol intoxication: Case
report. European Journal ofAnaesthesiology, 22, 560-561.
Agarwal, D. P. (2001). Genetic polymorphisms of alcohol metabolizing enzymes. Pathologic et
Biologic, 49, 703-709.
Airas, L., Paavilainen, T., Marttila, R. J., & Rinne, J. (2008). Methanol intoxication-induced
nigrostriatal dysfunction detected using 6-[18F]fluoro-L-dopa PET. NeuroToxicology, 29,
671-674. doi: 10.1016/j.neuro.2008.03.010
Akaike, H. (1973). Information theory and an extension of the maximum likelihood principle. In
B. N. Petrov & F. Csaki (Eds.), 2nd International Symposium on Information Theory (pp.
267-281). Budapest, Hungary: Akademiai Kiado.
Albin, R. L., & Greenamyre, J. T. (1992). Alternative excitotoxic hypotheses. Neurology, ¥2,
733-738.
Allen, B. C., Kavlock, R. J., Kimmel, C. A., & Faustman, E. M. (1994). Dose-response
assessment for developmental toxicity II: Comparison of generic benchmark dose
estimates with no observed adverse effect levels. Toxicological Sciences, 23(4), 487-495.
doi: 10.1006/faat. 1994.1133
Andresen, H., Schmoldt, H., Matschke, J., Flachskampfc, F. A., & Turka Ee. (2008). Fatal
methanol intoxication with different survival times -morphological findings and
postmortem methanol distribution. Forensic Science International, 779, 206-210.
Andrews, J. E., Ebron-McCoy, M., Kavlock, R. J., & Rogers, J. M. (1995). Developmental
toxicity of formate and formic acid in whole embryo culture: a comparative study with
mouse and rat embryos. Teratology, 57, 243-251.
Andrews, J. E., Ebron-McCoy, M., Logsdon, T. R., Mole, L. M., Kavlock, R. J., & Rogers, J. M.
(1993). Developmental toxicity of methanol in whole embryo culture: a comparative
study with mouse and rat embryos. Toxicology, 57, 205-215.
Andrews, J. E., Ebron-McCoy, M., Schmid, J. E., & Svensgaard, D. (1998). Effects of
combinations of methanol and formic acid on rat embryos in culture. Birth Defects
Research, Part A: Clinical and Molecular Teratology, 58, 54-61.
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 7-1 DRAFT - DO NOT CITE OR QUOTE
-------
Andrews, L. S., Clary, J. I, Terrill, J. B., & Bolte, H. F. (1987). Subchronic inhalation toxicity of
methanol. Journal of'Toxicology and Environmental Health, 20, 117-124.
Ang, H. L., Deltour, L., Hayamizu, T. R, Zgombic-Knight, M., & Duester, G. (1996). Retinoic
acid synthesis in mouse embryos during gastrulation and craniofacial development linked
to class IV alcohol dehydrogenase gene expression. Journal of Biological Chemistry, 271,
9526-9534.
Antony, A. C. (2007). In utero physiology: role of folic acid in nutrient delivery and fetal
development. American Journal of Clinical Nutrition, 85 (suppl.), 598S- 603S.
Apaja, M. (1980). Evaluation of toxicity and carcinogenicity ofmalonaldehyde: An experimental
study in Swiss mice. Ada Universitatis Ouluensis, Series D, Medica 55 (Vol. 8). Finland:
Anat Pathol Microbiol.
Arora, V, Nijjar, I. B. S., Multani, A. S., Singh, J. P., Abrol, R., Chopra, R., & Attri, R. (2007).
MRI findings in methanol intoxication: a report of two cases. British Journal of
Radiology, 80, 243-246.
Arora, V, Nijjar, I. S., Thukral, H., & Roopa. (2005). Bilateral putaminal necrosis caused by
methanol intoxication- a case report. Neuroradiology, 15(3), 341-342.
Aschner, M., & Kimelberg, H. K. (Eds.). (1996). The role ofglia in neurotoxicity. Boca Raton,
FL: CRC Press.
Aufderheide, T. P., White, S. M., Brady, W. J., & Stueven, H. A. (1993). Inhalational and
percutaneous methanol toxicity in two firefighters. Annals of Emergency Medicine, 22,
1916-1918.
Aziz, M. H., Agrawal, A. K., Adhami, V M., Ali, M. M., Baig, M. A., & Seth, P. K. (2002).
Methanol-induced neurotoxicity in pups exposed during lactation through mother; role of
folic acid. Neurotoxicology and Teratology, 24, 519-527.
Azmak, D. (2006). Methanol related deaths in Edirne. Legal Medicine, 8, 39-42.
Barceloux, D. G, Bond, G. R., Krenzelok, E. P., Cooper, H., & Vale, J. A. (2002). American
academy of clinical toxicology practice guidelines on the treatment of methanol
poisoning. Clinical Toxicology, 40(4), 415-446.
Batterman, S. A., & Franzblau, A. (1997). Time-resolved cutaneous absorption and permeation
rates of methanol in human volunteers. International Archives of Occupational and
Environmental Health, 70, 341-351.
Batterman, S. A., Franzblau, A., D'Arcy, J. B., Sargent, N. E., Gross, K. B., & Schreck, R. M.
(1998). Breath, urine, and blood measurements as biological exposure indices of short-
term inhalation exposure to methanol. International Archives of Occupational and
Environmental Health, 71, 325-335.
Beam, P., Patel, J., & O'Flynn, W. R. (1993). Cervical ribs: a cause of distal and cerebral
embolism. Postgraduate Medical Journal, 69, 65-68.
March 2011 7-2 DRAFT - DO NOT CITE OR QUOTE
-------
Bebarta, V. S., Heard, K., & Dart, R. C. (2006). Inhalational abuse of methanol products:
elevated methanol and formate levels without vision loss. American Journal of
Emergency Medicine, 24, 725-728.
Bennett, I. L., Jr., Gary, F. H., Mitchell, G. L., Jr., & Cooper, M. N. (1953). Acute methyl alcohol
poisoning: A review based on experiences in an outbreak of 323 cases. Medicine, 32, 431-
463.
Benton, C. D., Jr., & Calhoun, R P., Jr. (1952). The ocular effects of methyl alcohol poisoning:
Report of a catastrophe involving three hundred and twenty persons. Transactions of the
American Academy of Ophthalmology and Otolaryngology, 56, 875-885.
Bessell-Browne, R. J., & Bynevelt, M. (2007). Two cases of methanol poisoning: CT and MRI
features. Australasian Radiology, 57(2), 175-178.
Bhatia, K. P., & Marsden, C. D. (1994). The behavioural and motor consequences of focal lesions
of the basal ganglia in man. Brain, 117, 859-876.
Black, K. A., Eells, J. T., Noker, P. E., Hawtrey, C. A., & Tephly, T. R. (1985). Role of hepatic
tetrahydrofolate in the species difference in methanol toxicity. Proceedings of the
National Academy of Sciences, 82, 3854-3858.
Blanco, M., Casado, R., Vazquez, F., & Pumar Jm. (2006). CT and MR imaging findings in
methanol intoxication. American Journal of Neuroradiology, 27(2), 452-454.
Bosron, W. F., & Li, T. K. (1986). Genetic polymorphism of human liver alcohol and aldehyde
dehydrogenases, and their relationship to alcohol metabolism and alcoholism.
Hepatology, 6, 502-510.
Bouchard, M., Brunei, R. C., Droz, P. O., & Carrier, G. (2001). Abiologically based dynamic
model for predicting the disposition of methanol and its metabolites in animals and
humans. Toxicological Sciences, 64, 169-184.
Braden, G. L., Strayhorn, C. H., Germain, M. J., Mulhern, J. G, & Skutches, C. L. (1993).
Increased osmolal gap in alcoholic acidosis. Archives of Internal Medicine, 753(20),
2377-2380.
Brahmi, N., Blel, Y, Abidi, N., Kouraichi, N., Thabet, H., Hedhili, A., & Amamou, M. (2007).
Methanol poisoning in Tunisia: Report of 16 cases. Clinical Toxicology, 45, 717-720.
Branch, S., Rogers, J. M., Brownie, C. F., & Chernoff, N. (1996). Supernumerary lumbar rib:
Manifestation of basic alteration in embryonic development of ribs. Journal of Applied
Toxicology, 16(2), 115-119.
Brent, J., Lucas, M., Kulig, K., & Rumack, B. H. (1991). Methanol poisoning in a six-week-old
infant. Journal of Paediatrics and Child Health, 118, 644-646.
Brien, J. F., Clarke, D. W., Richardson, B., & Patrick, J. (1985). Disposition of ethanol in
maternal blood, fetal blood, and amniotic fluid of third-trimester pregnant ewes.
American Journal of Obstetrics and Gynecology, 152, 583-590.
March 2011 7-3 DRAFT - DO NOT CITE OR QUOTE
-------
Brown, R. P., Delp, M. D., Lindstedt, S. L., Rhomberg, L. R., & Bellies, R. P. (1997).
Physiological parameter values for physiologically based pharmacokinetic models.
Toxicology and Industrial Health, 73(4), 407-484. doi: 10.1177/074823379701300401
Bucher, J. R. (2002). The National Toxicology Program rodent bioassay: Designs, interpretations,
and scientific contributions. Annals of the New York Academy of Sciences, 982, 198-207.
Burbacher, T. M., Grant, K., Shen, D., Damian, D., Ellis, S., & Liberate, N. (1999). Reproductive
and offspring developmental effects following maternal inhalation exposure to methanol
in nonhuman primates Part II: developmental effects in infants exposedprenatally to
methanol. Cambridge, MA: Health Effects Institute.
Burbacher, T. M., Grant, K. S., Shen, D. D., Sheppard, L., Damian, D., Ellis, S., & Liberate, N.
(2004). Chronic maternal methanol inhalation in nonhuman primates (Macaca
fascicularis): reproductive performance and birth outcome. Neurotoxicology and
Teratology, 26, 639-650.
Burbacher, T. M., Shen, D., Grant, K., Sheppard, L., Damian, D., Ellis, S., & Liberate, N. (1999).
Reproductive and offspring developmental effects following maternal inhalation exposure
to methanol in nonhuman primates Part I: methanol disposition and reproductive toxicity
in adult females. Cambridge, MA: Health Effects Institute.
Burbacher, T. M., Shen, D. D., Lalovic, B., Grant, K. S., Sheppard, L., Damian, D., . . . Liberate,
N. (2004). Chronic maternal methanol inhalation in nonhuman primates (Macaca
fascicularis): exposure and toxicokinetics prior to and during pregnancy. Neurotoxicology
and Teratology, 26,201 -221.
Burnell, J. C., Li, T. K., & Bosron, W. R (1989). Purification and steady-state kinetic
characterization of human liver b3b3 alcohol dehydrogenase. Biochemistry, 28, 6810-
6815.
Burwell, R. D., Whealin, J., & Gallagher, M. (1992). Effects of aging on the diurnal pattern of
water intake in rats. Behavioral and Neural Biology, 55(3), 196-203.
Butchko, H. H., Stargel, W. W., Comer, C. P., Mayhew, D. A., Benninger, C., Blackburn, G L., . .
. Trefz, F. K. (2002). Aspartame: Review of safety. Regulatory Toxicology and
Pharmacology, 35, S1-S93.
Cameron, A. M., Nilsen, O. G, Haug, E., & Eik-Nes, K. B. (1984). Circulating concentrations of
testosterone, luteinizing hormone and follicle stimulating hormone in male rats after
inhalation of methanol. Archives of Toxicology, 7, 441-443.
Cameron, A. M., Zahlsen, K., Haug, E., Nilsen, O. G, & Eik-Nes, K. B. (1985). Circulating
steroids in male rats following inhalation of n-alcohols. Archives of Toxicology, 8, 422-
424.
Carson, B. L., McCann, J. L., Ellis, H. V, HI, Herndon, B. L., & Baker, L. H. (1981). Methanol
health effects [final task 7 report]. Ann Arbor, MI: U.S. EPA.
Caspi, R., Foerster, H., Fulcher, C. A., Hopkinson, R., Ingraham, J., Kaipa, P., ... Karp, P. D.
(2006). MetaCyc: a multiorganism database of metabolic pathways and enzymes. Nucleic
Acids Research, 34, D511-D516.
March 2011 7-4 DRAFT - DO NOT CITE OR QUOTE
-------
CERHR. (NTP Center for the Evaluation of Risks to Human Reproduction). (2004). NTP-
CERHR expert panel report on the reproductive and developmental toxicity of methanol.
Reproductive Toxicology, 18, 303-390. doi: 10.1016/j.reprotox.2003.10.013
Chen-Tsi, C. (1959). [Materials on the hygienic standardization of the maximally permissible
concentration of methanol vapors in the atmosphere]. Gigiena i Sanitariia, 24, 7-12.
Chen, J. C., Schneiderman, J. R, & Wortzman, G. (1991). Methanol poisoning: Bilateral
putaminal and cerebellar cortical lesions on CT and MR. Journal of Computer Assisted
Tomography, 15, 522-524.
Chernoff, N., & Rogers, J. M. (2004). Supernumerary ribs in developmental toxicity bioassays
and in human populations: Incidence and biological significance. Journal of Toxicology
and Environmental Health, Part B: Critical Reviews, 7, 437-449.
Chuwers, P., Osterloh, J., Kelly, T., D'Alessandro, A., Quinlan, P., & Becker, C. (1995).
Neurobehavioral effects of low-level methanol vapor exposure in healthy human
volunteers. Environmental Research, 77(2), 141-150.
Cichoz-Lach, H., Partycka, J., Nesina, I, Wojcierowski, J., Slomka, M., & Celinski, K. (2007).
Genetic polymorphism of alcohol dehydrogenase 3 in digestive tract alcohol damage.
HepatoGastroenterol, 54(16), 1222-1227.
Clancy, B., Finlay, B. L., Darlington, R. B., & Anand, K. J. (2007). Extrapolating brain
development from experimental species to humans. NeuroToxicology, 28(5), 931-937.
doi: 10.1016/j.neuro.2007.01.014
Clarke, D. W., Steenaart, N. A. E., & Brien, J. R (1986). Disposition of ethanol and activity of
hepatic and placental alcohol dehydrogenase and aldehyde dehydrogenases in the third-
trimester pregnant guinea pig for single and short-term oral ethanol administration.
Alcoholism: Clinical and Experimental Research, 10(3), 330-336.
Clary, J. J. (2003). Methanol, is it a developmental risk to humans? Regulatory Toxicology and
Pharmacology, 37, 83-91. doi: 10.1016/s0273-2300(02)00031-4
Clewell, H. J., IE, Gentry, P. R., Gearhart, J. M., Covington, T. R., Banton, M. I, & Andersen, M.
E. (2001). Development of a physiologically based pharmacokinetic model of isopropanol
and its metabolite acetone. Toxicological Sciences, 63(2), 160-172.
Coleman, C. N., Mason, T., Hooker, E. P., & Robinson, S. E. (1999). Developmental effects of
intermittent prenatal exposure to 1,1,1-trichloroethane in the rat. Neurotoxicology and
Teratology, 21(6), 699-708. doi: 10.1016/s0892-0362(99)00035-5
Collins, A. S., Sumner, S. C. J., Borghoff, S. J., & Medinsky, M. A. (1999). A physiological
model for tert-amyl methyl ether and tert-amyl alcohol: hypothesis testing of model
structures. Toxicological Sciences, 49, 15-28.
Connell, J. L., Doyle, J. C., & Gurry, J. R (1980). The vascular complications of cervical ribs.
Australian and New Zealand Journal of Surgery, 50(2), 125-130. doi: lO.llll/j.1445-
2197.1980.tb06648.x
March 2011 7-5 DRAFT - DO NOT CITE OR QUOTE
-------
Cook, M. R., Bergman, F. I, Cohen, H. D., Gerkovich, M. M., Graham, C., Harris, R. K., &
Siemann, L. G (1991). Effects of methanol vapor on human neurobehavioral measures.
(Report No. Research Report Number 42). Boston, MA: Health Effects Institute.
Cook, R. I, Champion, K. M., & Giometti, C. S. (2001). Methanol toxicity and formate
oxidation in NEUT2 mice. Archives of Biochemistry and Biophysics, 393(2), 192-198.
doi: 10.1006/abbi.2001.2485
Cooper, R. L., Mole, M. L., Rehnberg, G. L., Goldman, J. M., McElroy, W. K., Hein, I, & Stoker,
T. E. (1992). Effect of inhaled methanol on pituitary and testicular hormones in chamber
acclimated and non-acclimated rats. Toxicology, 71, 69-81.
Corley, R. A., Bormett, G. A., & Ghanayem, B. I. (1994). Physiologically-based pharmacolinetics
of 2-butoxyethanol and its major metabolite 2-butoxyacetic acid, in rats and humans.
Toxicology and Applied Pharmacology, 129(1), 61-79. doi: 10.1006/taap. 1994.1229
Cronholm, T. (1987). Effect of ethanol on the redox state of the coenzyme bound to alcohol
dehydrogenase studied in isolated hepatocytes. BiochemicalJournal, 248(2), 567-572.
Cruzan, G. (2009). Assessment of the cancer potential of methanol. Critical Reviews in
Toxicology, 39(4), 347-363. doi: 10.1080/10408440802475199
Cumming, M. E., Ong, B. Y, Wade, J. G, & Sitar, D. S. (1984). Maternal and fetal ethanol
pharmacokinetics and cardiovascular responses in near-term pregnant sheep. Canadian
Journal of Physiology and Pharmacology, 62(12), 1435-1439.
d'Alessandro, A., Osterloh, J. D., Chuwers, P., Quinlan, P. J., Kelly, T. J., & Becker, C. E. (1994).
Formate in serum and urine after controlled methanol exposure at the threshold limit
value. Environmental Health Perspectives, 102, 178-181.
Davis, V. E., Brown, H., Huff, J. A., & Cashaw, J. L. (1967). The alteration of serotonin
metabolism to 5-hydroxytryptophol by ethanol ingestion in man. Journal of Laboratory
and Clinical Medicine, 69(1), 132-140.
Davoli, E., Cappellini, L., Airoldi, L., & Fanelli, R. (1986). Serum methanol concentrations in
rats and in men after a single dose of aspartame. Food and Chemical Toxicology, 24(3),
187-189.
De Brabander, N., Wojciechowski, M., De Decker, K., De Weerdt, A., & Jorens, P. G. (2005).
Fomepizole as a therapeutic strategy in paediatric methanol poisoning: a case report and
review of the literature. European Journal of Pediatrics, 164(3), 158-161. doi:
10.1007/s00431-004-1588-5
Degitz, S. J., Rogers, J. M., Zucker, R. M., & Hunter, E. S., IE. (2004). Developmental toxicity
of methanol: pathogenesis in CD-I and C57BL/6J mice exposed in whole embryo culture.
Birth Defects Research, Part A: Clinical and Molecular Teratology, 70, 179-184. doi:
10.1002/bdra.20009
Degitz, S. J., Zucker, R. M., Kawanishi, C. Y, Massenburg, G. S., & Rogers, J. M. (2004).
Pathogenesis of methanol-induced craniofacial defects in C57BL/6J mice. Birth Defects
Research, Part A: Clinical and Molecular Teratology, 70, 172-178. doi:
10.1002/bdra.20010
March 2011 7-6 DRAFT - DO NOT CITE OR QUOTE
-------
Deltour, L., Foglio, M. H., & Duester, G. (1999). Metabolic deficiencies in alcohol
dehydrogenase Adhl, Adh3, and Adh4 null mutant mice. Overlapping roles of Adhl and
Adh4 in ethanol clearance and metabolism of retinol to retinoic acid. Journal of
Biological Chemistry, 2 74(24), 16796-16801.
Dethlefs, R., & Naraqi, S. (1978). Ocular manifestations and complications of acute methyl
alcohol intoxication. MedicalJournal of Australia, 2(10), 483-485.
Devore, J. L. (1995). Probability and statistics for engineering and the sciences (4 ed.). Belmont,
CA: Duxbury Press.
Dicker, E., & Cedebaum, A. I. (1986). Inhibition of the low-Km mitachondrial aldehyde
dehydrogenase by diethyl maleate and phorone in vivo and in vitro: implications for
formaldehyde metabolism. BiochemicalJournal, 240(3), 821-827.
Dikalova, A. E., Kadiiska, M. B., & Mason, R. P. (2001). An in vivo ESR spin-trapping study:
Free radical generation in rats from formate intoxication-role of the Fenton reaction.
Proceedings of the National Academy of Sciences, 98(24), 13549-13553. doi:
10.1073/pnas.251091098
Dorman, D. C., Bolon, B., Struve, M. F., LaPerle, K. M. D., Wong, B. A., Elswick, B., & Welsch,
F. (1995). Role of formate in methanol-induced exencephaly in CD-I mice. Teratology,
52(1), 30-40.
Dorman, D. C., Moss, O. R., Farris, G. M., Janszen, D., Bond, J. A., & Medinsky, M. A. (1994).
Pharmacokinetics of inhaled [14C]methanol and methanol-derived [14C]formate in
normal and folate-deficient cynomolgus monkeys. Toxicology and Applied
Pharmacology, 725(2), 229-238. doi: 10.1006/taap. 1994.1202
Dorman, D. C., & Welsch, F. (1996). Developmental toxicity of methanol in rodents, CUT
Activities, pp. 1-7.
Downie, A., Khattab, T. M., Malik, M. L, & Samara, I. N. (1992). A case of percutaneous
industrial methanol toxicity. Occupational Medicine, 42(1), 47-49.
Dudka, J. (2006). The total antioxidant status in the brain after ethanol or 4-methylpyrazole
administration to rats intoxicated with methanol. Experimental and Toxicologic
Pathology, 57, 445-448.
Dutkiewicz, B., Konczalik, J., & Karwacki, W. (1980). Skin absorption and per os administration
of methanol in men. International Archives of Occupational and Environmental Health,
¥7(1), 81-88.
Eells, J. T., Black, K. A., Tedford, C. E., & Tephly, T. R. (1983). Methanol toxicity in the
monkey: effects of nitrous oxide and methionine. Journal of Pharmacology and
Experimental Therapeutics, 227(2), 349-353.
ERG. (Eastern Research Group Inc.). (2009). External letter peer review of reports documenting
methanol studies in monkeys, rats and mice performed by the New Energy Development
Organization (NEDO). Lexington, MA: Author.
Ernstgard, L. (2005). [E-mail correspondance from Lena Ernstgard to Torka Poet].
March 2011 7-7 DRAFT - DO NOT CITE OR QUOTE
-------
Ernstgard, L., Shibata, E., & Johanson, G. (2005). Uptake and disposition of inhaled methanol
vapor in humans. Toxicological Sciences, 88, 30-38.
Estonius, M., Svensson, S., & Hoog, J. O. (1996). Alcohol dehydrogenase in human tissues:
Localization of transcripts coding for five classes of the enzyme. FEES Letters, 397(2-3),
338-342. doi: 10.1016/s0014-5793(96)01204-5
Evans, A. L. (1999). Pseudoseizures as a complication of painful cervical ribs. Developmental
Medicine and Child Neurology, 41, 840-842. doi: 10.1017/s0012162299001668
Fagan, J. E, & Singer, L. T. (1983). Infant recognition memory as a measure of intelligence. In L.
P. Lipsitt (Ed.), Advances in infancy research (Vol. 2, pp. 31-78). New York, NY: Ablex.
Fallang, B., Saugstad, O. D., Gragaard, J., & Hadders-Algra, M. (2003). Kinematic quality of
reaching movements in preterm infants. Pediatric Research, 53, 836-842.
Feany, M. B., Anthony, D. C., Frosch, M. P., Zane, W., & De Girolami, U. (2001). August 2000:
two cases with necrosis and hemorrhage in the putamen and white matter. Brain
Pathology, 11, 125.
Fernandez Noda, E. L, Nufiez-Arguelles, J., Perez Fernandez, J., Castillo, J., Perez Izquierdo, M.,
& Rivera Luna, H. (1996). Neck and brain transitory vascular compression causing
neurological complications, results of surgical treatment on 1300 patients. Journal of
Cardiovascular Surgery, 37, 155-166.
Finkelstein, Y, & Vardi, J. (2002). Progressive parkinsonism in a young experimental physicist
following long-term exposure to methanol. NeuroToxicology, 23, 521-525.
Fiserova-Bergerova, V, & Diaz, M. L. (1986). Determination and prediction of tissue-gas
partition coefficients. International Archives of Occupational and Environmental Health,
58, 75-87. doi: 10.1007/bf00378543
Fisher, J. W., Dorman, D. C., Medinsky, M. A., Welsch, F., & Conolly, R. B. (2000). Analysis of
respiratory exchange of methanol in the lung of the monkey using a physiological model.
Toxicological Sciences, 53, 185-193.
Fontenot, A. P., & Pelak, V. S. (2002). Development of neurologic symptoms in a 26-year-old
woman following recovery from methanol intoxication. Chest, 122, 1436-1439.
Foster, M. W., & Stamler, J. (2004). New insights into protein s-nitrosylation-mitochondria as a
model system. Journal of Biological Chemistry, 279(24), 25891-25897.
Frederick, L. J., Schulte, P. A., & Apol, A. (1984). Investigation and control of occupational
hazards associated with the use of spirit duplicators. American Industrial Hygiene
Association Journal, 45, 51-55.
Fu, S. S., Sakanashi, T. M., Rogers, J. M., & Hong. (1996). Influence of dietary folic acid on the
developmental toxicity of methanol and the frequency of chromosomal breakage in the
CD-I mouse. Reproductive Toxicology, 10, 455-463.
Gannon, K. S., Smith, J. C., Henderson, R., & Hendrick, P. (1992). A system for studying the
microstructure of ingestive behavior in mice. Physiology and Behavior, 51, 515-521.
March 2011 7-8 DRAFT - DO NOT CITE OR QUOTE
-------
Gaul, H. P., Wallace, C. I, Auer, R. N., & Chen Fong, T. (1995). MRI findings in methanol
intoxication. American Journal of Neuroradiology, 16, 1783-1786.
Gentry, P. R., Covington, T. R., Andersen, M. E., & Clewell, H. I, IE. (2002). Application of a
physiologically based pharmacokinetic model for isopropanol in the derivation of a
reference dose and reference concentration. Regulatory Toxicology and Pharmacology,
51-68.
Gentry, P. R., Covington, T. R., & Clewell, H. I, IE. (2003). Evaluation of the potential impact
of pharmacokinetic differences on tissue dosimetry in offspring during pregnancy and
\actation.RegulatoryToxicologyandPharmacology, 35(1), 1-16. doi: 10.1016/s0273-
2300(03)00047-3
Gibson, M. A. S., Butters, N. S., Reynolds, J. N., & Brien, J. R (2000). Effects of chronic
prenatal ethanol exposure on locomotor activity, and hippocampal weight, neurons, and
nitric oxide synthase activity of the young postnatal guinea pig. Neurotoxicology and
Teratology, 22(2), 183-182.
Gonzalez-Quevado, A., Obregon Furbina, M., Rousso, T., & Lima, L. (2002). Effect of chronic
methanol administration on amino-acids and monoamines in retina, optic nerve, and brain
of the rat. Toxicology and Applied Pharmacology, 185, 77-84.
Guerri, C., & Sanchis, R. (1985). Acetaldehyde and alcohol levels in pregnant rats and their
fetuses. Alcohol, 2, 267-270.
Guggenheim, M. A., Couch, J. R., & Weinberg, W. (1971). Motor dysfunction as a permanent
complication of methanol ingestion. Archives of Neurology, 24, 550-554.
Hansen, J. M., Contreras, K. M., & Harris, C. (2005). Methanol, formaldehyde, and sodium
formate exposure in rat and mouse conceptuses: A potential role of the visceral yolk sac
in embryotoxicity. Birth Defects Research, Part A: Clinical and Molecular Teratology,
73, 72-82.
Hantson, P., Duprez, T., & Mahieu, P. (1997). Neurotoxicity to the basal ganglia shown by
magnetic resonance imaging (MRI) following poisoning by methanol and other
sub stances . Clinical Toxicology, 35, 151-161.
Hantson, P., Lambermont, J. Y, & Mahieu, P. (1997). Methanol poisoning during late pregnancy.
Clinical Toxicology, 35(2), 187-191.
Hanzlik, R. P., Fowler, S. C., & Eells, J. T. (2005). Absorption and elimination of formate
following oral administration of calcium formate in female human subjects. Drug
Metabolism and Disposition, 33(2), 282-286. doi: 10. 1124/dmd. 104.001289
Harris, C., Dixon, M., & Hansen, J. M. (2004). Glutathione depletion modulates methanol,
formaldehyde and formate toxicity in cultured rat conceptuses. Cell biology and
toxicology, 20, 133-145.
Harris, C., Wang, S. W., Lauchu, J. J., & Hansen, J. M. (2003). Methanol metabolism and
embryotoxicity in rat and mouse conceptuses : Comparisons of alcohol dehydrogenase
(ADH1), formaldehyde dehydrogenase (ADH3), and catalase. Reproductive Toxicology,
77(3), 349-357. doi: 10.1016/s0890-6238(03)00013-3
March 2011 7-9 DRAFT - DO NOT CITE OR QUOTE
-------
Hashimoto, M., & Nedo. (2008). Report letter from Masahiro Hashimoto, Director General of
the Policy Planning and Coordination Department of NEDO, to John Lynn, CEO of
Methanol Institute, certifying the English translations of the 1985 NEDO (New Energy
and Industrial Technology Development Organization) originalJapanese methanol test
reports are accurate and complete. Kawasaki City, Japan: New Energy and Industrial
Technology Development Organization.
Hass, U., Lund, S. P., Simonsen, L., & Fries, A. S. (1995). Effects of prenatal exposure to xylene
on postnatal development and behavior in rats. Neurotoxicology and Teratology, 17(3),
341-349.
Hassanian-Moghaddam, H., Pajoumand, A., Dadgar, S. M., & Shadnia, S. H. (2007). Prognostic
factors in methanol poisoning. Human and Experimental Toxicology, 26(7), 583-586.
Hedberg, J. J., Backlund, M., Stromberg, P., Lonn, S., Dahl, M. L., Ingelman-Sundberg, M., &
Hoog, J. O. (2001). Functional polymorphism in the alcohol dehydrogenase 3 (ADH3)
promoter. Pharmacogenetics, 11(9), 815-824.
F£EI. (Health Effects Institute). (1987). Automotive methanol vapors and human health: An
evaluation of existing scientific information and issues for future research. Boston, MA:
Author.
Henderson, M. S. (1914). Cervical rib: Report of thirty-one cases. Journal of Bone and Joint
Surgery: American Volume, 11, 408-430.
Henderson, W. R., & Brubacher, J. (2002). Methanol and ethylene glycol poisoning: A case study
and review of current literature. Canadian Journal of Emergency Medicine, 4(1), 34-40.
Hines, R. N., & McCarver, D. G. (2002). The ontogeny of human drug-metabolizing enzymes:
Phase I. Oxidative enzymes. Journal of Pharmacology and Experimental Therapeutics,
300(2), 355-360. doi: 10.1124/jpet.300.2.355
Horton, V. L., Higuchi, M. A., & Rickert, D. E. (1992). Physiologically based pharmacokinetic
model for methanol in rats, monkeys, and humans. Toxicology and Applied
Pharmacology, 117(1), 26-36.
Hovda, K. E., Hunderi, O. H., Tafjord, A. B., Dunlop, O., Rudberg, N., & Jacobsen, D. (2005).
Methanol outbreak in Norway 2002-2004: epidemiology, clinical features and prognostic
signs. Journal of Internal Medicine, 258, 181-190.
Hovda, K. E., Mundal, H., Urdal, P., McMartin, K., & Jacobsen, D. (2007). Extremely slow
formate elimination in severe methanol poisoning: A fatal case report. Clinical
Toxicology, ¥5,516-521.
HSDB. (Hazardous Substances Data Bank). (2009, December 22, 2009). Methanol: Human
health effects, from http://toxnet.nlm.nih.gOv/cgi-bin/sis/search/f7./temp/~XKgvsM:l
Hsu, H. H., Chen, C. Y, Chen, F. H., Lee, C. C., Chou, T. Y, & Zimmerman, R. A. (1997). Optic
atrophy and cerebral infarcts caused by methanol intoxication: MRI. Neuroradiology, 39,
192-194.
March 2011 7-10 DRAFT - DO NOT CITE OR QUOTE
-------
Huang, Q. R, Gebrewold, A., Zhang, A., Altura, B. T., & Altura, B. M. (1994). Role of excitatory
amino acids in regulation of rat pial microvasculature. American Journal of Physiology,
266, R158-R163.
Hunderi, O. H., Hovda, K. E., & Jacobsen, D. (2006). Use of the osmolal gap to guide the start
and duration of dialysis in methanol poisoning. Scandinavian Journal of Urology and
Nephrology, 40, 70-74.
Infurna, R., & Weiss, B. (1986). Neonatal behavioral toxicity in rats following prenatal exposure
to methanol. Teratology, 33, 259-265.
IPCS. (International Programme on Chemical Safety). (1997). Methanol. Geneva, Switzerland:
World Health Organization.
Jacobsen, D., & McMartin, K. E. (1986). Methanol and ethylene glycol poisonings. Mechanisms
of toxicity, clinical course, diagnosis and treatment. Medical Toxicology, 1, 309-334.
Jacobsen, D., Webb, R., Collins, T. D., & McMartin, K. E. (1988). Methanol and formate kinetics
in late diagnosed methanol intoxication. Medical Toxicology, 3, 418-423.
Johanson, G, Kronborg, H., Naslund, P. H., & Nordqvist, M. B. (1986). Toxicokinetics of
inhaled 2-butoxyethanol (ethylene glycol monobutyl ether) in man. Scandinavian Journal
of Work, Environment and Health, 12, 594-602.
Johlin, R C., Fortman, C. S., Nghiem, D. D., & Tephly, T. R. (1987). Studies on the role of folic
acid and folate-dependent enzymes in human methanol poisoning. Molecular
Pharmacology, 31, 557-561.
Kahn, A., & Blum, D. (1979). Methyl alcohol poisoning in an 8-month-old boy: An unusual route
of intoxication. Journal of Paediatrics and Child Health, 94, 841-843.
Kavet, R., & Nauss, K. M. (1990). The toxicity of inhaled methanol vapors. Critical Reviews in
Toxicology, 21, 21-50.
Kavlock, R. J., Allen, B. C., Faustman, E. M., & Kimmel, C. A. (1995). Dose-response
assessments for developmental toxicity. IV. Benchmark doses for fetal weight changes.
Toxicological Sciences, 26,211 -222.
Kawai, T., Yasugi, T., Mizunuma, K., Horiguchi, S., Hirase, Y, Uchida, Y, & Ikeda, M. (1991).
Methanol in urine as a biological indicator of occupational exposure to methanol vapor.
International Archives of Occupational and Environmental Health, 63, 311-318.
Keles, G. T., Orguc, S., Toprak, B., Ozaslan, S., & Sakarya, M. (2007). Methanol poisoning with
necrosis corpus callosum. Clinical Toxicology, 45(3), 307-308.
Kerns, W., U, Tomaszewski, C., McMartin, K., Ford, M., & Brent, J. (2002). Formate kinetics in
methanol poisoning. Clinical Toxicology, 40, 137-143.
Keys, D. A., Schultz, I. R., Mahle, D. A., & Fisher, J. W. (2004). A quantitative description of
suicide inhibition of dichloroacetic acid in rats and mice. Toxicological Sciences, 82(2),
381-393. doi: 10.1093/toxsci/kfh276
March 2011 7-11 DRAFT - DO NOT CITE OR QUOTE
-------
Kim, S. W., Jang, Y. I, Chang, J. W., & Hwang, O. (2003). Degeneration of the nigrostriatal
pathway and induction of motor deficit by tetrahydrobiopterin: An in vivo model relevant
to Parkinson's disease. Neurobiology of Disease, 13(2), 167-176.
Kraut, J. A., & Kurtz, I. (2008). Toxic alcohol ingestions: Clinical features, diagnosis, and
management. Clinical Journal of the American Society of Nephrology, 3(1), 208-225.
Kuteifan, K., Oesterle, H., Tajahmady, T., Gutbub, A. M., & Laplatte, G. (1998). Necrosis and
haemorrhage of the putamen in methanol poisoning shown on MRI. Neuroradiology,
40(3), 158-160.
Lee, E., Brady, A. N., Brabec, M. J., & Fabel, T. (1991). Effects of methanol vapors on
testosterone production and testis morphology in rats. Toxicology and Industrial Health,
7, 261-275.
Lee, E. W., Garner, C. D., & Terzo, T. S. (1994). Animal model for the study of methanol
toxicity: Comparison of folate-reduced rat responses with published monkey data.
Journal of Toxicology and Environmental Health, 41, 71-82.
Lee, E. W., Terzo, T. S., D'Arcy, J. B., Gross, K. B., & Schreck, R. M. (1992). Lack of blood
formate accumulation in humans following exposure to methanol vapor at the current
permissible exposure limit of 200 ppm. American Industrial Hygiene Association
Journal, 53, 99-104.
Lewis, R. J., Sr. (1992). Sax's dangerous properties of industrial materials: v///(8th ed.). New
York, NY: Van Nostrand Reinhold.
Ley, C. O., & Gali, F. G. (1983). Parkinsonian syndrome after methanol intoxication. European
Neurology, 22, 405-409.
Liu, J. J., Daya, M. R., Carrasquillo, O., & Kales, S. N. (1998). Prognostic factors in patients
with methanol poisoning. Clinical Toxicology, 36, 175-181.
Lorente, C., Cordier, S., Bergeret, A., De Walle, H. E. K., Goujard, J., Ayme, S., . . . Bianchi, F.
(2000). Maternal occupational risk factors for oral clefts. Scandinavian Journal of Work,
Environment and Health, 26, 137-145.
Mahieu, P., Hassoun, A., & Lauwerys, R. (1989). Predictors of methanol intoxication with
unfavorable outcome. Human Toxicology, 5(2), 135-137.
Makar, A. B., Tephly, T. R., & Mannering, G. J. (1968). Methanol metabolism in the monkey.
Molecular Pharmacology, 4, 471-483.
Mann, W. J., Muttray, A., Schaefer, D., Klimek, L., Faas, M., & Konietzko, J. (2002). Exposure
to 200 ppm of methanol increases the concentrations of interleukin-lbeta and interleukin-
8 in nasal secretions of healthy volunteers. Annals of Otology, Rhinology, and
Laryngology, 111, 633-638.
Mannering, G. J., Van Harken, D. R., Makar, A. B., Tephly, T. R., Watkins, W. D., & Goodman, J.
I. (1969). Role of the intracellular distribution of hepatic catalase in the peroxidative
oxidation of methanol. Annals of the New York Academy of Sciences, 168, 265-280.
March 2011 7-12 DRAFT - DO NOT CITE OR QUOTE
-------
Medinsky, M. A., Dorman, D. C., Bond, J. A., Moss, O. R., Janszen, D. B., & Everitt, J. I.
(1997). Pharmacokinetics of methanol and formate in female cynomolgus monkeys
exposed to methanol vapors. Boston, MA: Health Effects Institute.
Meister, A., & Anderson, M. E. (1983). Glutathione. Annual Review of Biochemistry, 52, 711-
760.
Methanol Institute. (Methanol Institute). (2009a, December 18, 2009). Biodiesel: A growing
market for methanol, from http://www.methanol.org/pdfFrame.cfm?pdf=Biodiesel2.pdf
Methanol Institute. (Methanol Institute). (2009b, December 18, 2009). Frequently asked
questions, from http://www.methanol.org/pdfFrame.cfm?pdf=faqs.pdf
Meyer, R. J., Beard, M. E., & Ardagh, M. W. (2000). Methanol poisoning. New Zealand Medical
Journal, 773(1102), 11-13.
Montserrat, C. A., Field, M. S., Perry, C., Ghandour, H., Chiang, E., Selhub, J., . . . Stover, P. J.
(2006). Regulation of folate-mediated one-carbon metabolism by 10-
formyltetrahydrofolate dehydrogenase. Journal of Biological Chemistry, 257(27), 18335-
18342.
Mooney, S. M., & Miller, M. W. (2001). Episodic exposure to ethanol during development
differentially affects brainstem nuclei in the macaque. Journal ofNeurocytology, 30, 973-
982. doi: 10.1023/a:1021832522701
Muthuvel, A., Rajamani, R., Manikandan, S., & Sheeladevi, R. (2006). Detoxification of formate
by formate dehydrogenase-loaded erythrocytes and carbicarb in folate-deficient methanol-
intoxicated rats. Clinica ChimicaActa, 367, 162-169. doi: 10.1016/j.cca.2005.12.007
Muthuvel, A., Rajamani, R., Senthilvelan, M., Manikandan, S., & Sheeladevi, R. (2006).
Modification of allergenicity and immunogenicity of formate dehydrogenase by
conjugation with linear mono methoxy poly ethylene glycol: Improvement in
detoxification of formate in methanol poisoning. Clinica Chimica Acta, 374, 122-128.
doi: 10.1016/j.cca.2006.06.003
Naraqi, S., Dethlefs, R. R, Slobodniuk, R. A., & Sairere, J. S. (1979). An outbreak of acute
methyl alcohol intoxication. Australian and New Zealand Journal of Medicine, 9(1), 65-
68.
NEDO. (New Energy Development Organization). (1987). Toxicological research of methanol as
a fuel for power station: summary report on tests with monkeys, rats and mice. Tokyo,
Japan: Author.
NEDO. (New Energy Development Organization). (2008a). Test report: 18-month inhalation
carcinogenicity study on methanol in B6C3F1 mice (test no. 4A-223). Tokyo, Japan
(Original work published 1985): Mitsubishi Kasei Institute of Toxicology and
Environmental Sciences.
NEDO. (New Energy Development Organization). (2008b). Test report: 24-month inhalation
carcinogenicity study on methanol in Fischer rats (Test No. 5A-268). Toyko, Japan
(Original work published 1985): Mitsubishi Kasei Institute of Toxicology and
Environmental Sciences.
March 2011 7-13 DRAFT - DO NOT CITE OR QUOTE
-------
Nelson, B. K., Brightwell, W. S., MacKenzie, D. R., Khan, A., Burg, J. R., Weigel, W. W., &
Goad, P. T. (1985). Teratological assessment of methanol and ethanol at high inhalation
levels in rats. Toxicological Sciences, 5, 727-736.
Nguyen, T., Baumgartner, E, & Nelems, B. (1997). Bilateral rudimentary first ribs as a cause of
thoracic outlet syndrome. Journal of the National Medical Association (JNMA), 89, 69-
73.
Noker, P. E., Eells, J. E, & Eephly, E R. (1980). Methanol toxicity: treatment with folic acid and
5-formyl tetrahydrofolic acid. Alcoholism: Clinical and Experimental Research, 4, 378-
383.
NRC. (National Research Council). (1983). Risk assessment in the federal government:
Managing the process. Washington, DC: National Academy Press.
Osterloh, J. D., D'Alessandro, A., Chuwers, P., Mogadeddi, H., & Kelly, E. J. (1996). Serum
concentrations of methanol after inhalation at 200 ppm. Journal of Occupational and
Environmental Medicine, 38, 571-576.
Parthasarathy, N. J., Kumar, R. S., & Devi, R. S. (2005). Effect of methanol intoxication on rat
neutrophil functions. Journal of Immunotoxicology, 2, 115-211. doi:
10.1080/15476910500187425
Parthasarathy, N. J., Kumar, R. S., Karthikeyan, P., & Sheela Devi, R. (2005). In vitro and in vivo
study of neutrophil functions after acute methanol intoxication in albino rats.
Toxicological and Environmental Chemistry, 87, 559-566. doi:
10.1080/02772240500382308
Parthasarathy, N. J., Kumar, R. S., Manikandan, S., & Devi, R. S. (2006). Methanol-induced
oxidative stress in rat lymphoid organs. Journal of Occupational Health, 48, 20-27.
Parthasarathy, N. J., Kumar, R. S., Manikandan, S., Narayanan, G. S., Kumar, R. V, & Devi, R.
S. (2006). Effect of methanol-induced oxidative stress on the neuroimmune system of
experimental rats. Chemico-BiologicalInteractions, 161, 14-25.
Parthasarathy, N. J., Srikumar, R., Manikandan, S., Narayanan, G. S., & Devi, R. S. (2007).
Effect of methanol intoxication on specific immune functions of albino rats. Cell biology
and toxicology, 23(3), 177-187.
Pastino, G. M., & Conolly, R. B. (2000). Application of a physiologically based pharmacokinetic
model to estimate the bioavailability of ethanol in male rats: distinction between gastric
and hepatic pathways of metabolic clearance. Toxicological Sciences, 55, 256-265.
Patankar, T., Bichile, L., Karnad, D., Prasad, S., & Rathod, K. (1999). Methanol poisoning:
computed tomography scan findings in four patients. Australasian Radiology, 43(4), 526-
528.
Pelletier, J., Habib, M. H., Khalil, R., Salamon, G, Bartoli, D., & Jean, P. (1992). Putaminal
necrosis after menhanol intoxication [letter]. Journal of Neurology Neurosurgery and
Psychiatry, 55, 234-235.
March 2011 7-14 DRAFT - DO NOT CITE OR QUOTE
-------
Peng, M. T., Chen, Y. T., Hung, S. H., & Yaung, C. L. (1990). Circadian rhythms of feeding and
drinking behavior of rats aged from 3 to 21 months. Proceedings of the National Science
Council, Republic of China. PartB, Life Sciences (Taipei), 14, 98-104.
Perkins, R. A., Ward, K. W., & Pollack, G. M. (1995a). Comparative toxicokinetics of inhaled
methanol in the female CD-I mouse and Sprague-Dawley rat. Toxicological Sciences, 28,
245-254.
Perkins, R. A., Ward, K. W., & Pollack, G. M. (1995b). A pharmacokinetic model of inhaled
methanol in humans and comparison to methanol disposition in mice and rats.
Environmental Health Perspectives, 103, 726-733.
Perkins, R. A., Ward, K. W., & Pollack, G. M. (1996). Methanol inhalation: site and other factors
influencing absorption, and an inhalation toxicokinetic model for the rat. Pharmaceutical
Research, 13, 749-755.
Phang, P. T., Passerini, L., Mielke, B., Berendt, R., & King, E. G. (1988). Brain hemorrhage
associated with methanol poisoning. Critical Care Medicine, 16, 137-140.
Pietruszko, R. (1980). Alcohol and aldehyde dehydrogenase isozymes from mammalian liver—
their structural and functional differences. Isozymes: Current Topics in Biological
Medical Research, 4, 107-130.
Pikkarainen, P. H., & Raiha, N. C. R. (1967). Development of alcohol dehydrogenase activity in
the human liver. Pediatric Research, 7(3), 165-168.
Pitkin, R. M. (2007). Folate and neural tube defects. American Journal of Clinical Nutrition, 85,
285S-288S.
Plantinga, Y, Perdock, J., & de Groot, L. (1997). Hand function in low-risk preterm infants: Its
relation to muscle power regulation. Developmental Medicine and Child Neurology, 39,
6-11.
Poet, T. S., Teeguarden, J. G, & Hinderliter, P. M. (2006). Final Report: Development,
calibration and application of a methanol PBPK model. Richland, WA: Battelle Pacific
Northwest National Laboratory.
Pollack, G. M., & Brouwer, K. L. R. (1996). Maternal-fetal pharmacokinetics of methanol.
(Report No. 74). Boston, MA: Health Effects Institute.
Pollack, G. M., Brouwer, K. L. R., & Kawagoe, J. L. (1993). Toxicokinetics of intravenous
methanol in the female rat. Toxicological Sciences, 21, 105-110.
Pollack, G. M., & Kawagoe, J. L. (1991). Determination of methanol in whole blood by capillary
gas chromatography with direct on-column injection. Journal ofChromatographyA, 570,
406-411.
Poon, R., Chu, L, Bjarnason, S., Potvin, M., Vincent, R., Miller, R. B., & Valli, V. E. (1994).
Inhalation toxicity study of methanol, toluene, and methanol/toluene mixtures in rats:
effects of 28-day exposure. Toxicology and Industrial Health, 10, 231-245.
March 2011 7-15 DRAFT - DO NOT CITE OR QUOTE
-------
Poon, R., Chu, I., Bjarnason, S., Vincent, R., Potvin, M., Miller, R. B., & Valli, V. E. (1995).
Short-term inhalation toxicity of methanol, gasoline, and methanol/gasoline in the rat.
Toxicology and Industrial Health, 77, 343-361.
Prabhakaran, V, Ettler, H., & Mills, A. (1993). Methanol poisoning: Two cases with similar
plasma concentrations but different outcomes. Canadian Medical Association Journal,
148, 981-984.
Rajamani, R., Muthuvel, A., Senthilvelen, M., & Sheeladevi, R. (2006). Oxidative stress induced
by methotrexate alone and in the presence of methanol in discrete regions of the rodent
brain, retina and optic nerve. Toxicology Letters, 165, 265-273.
Ramsey, J. C., & Andersen, M. E. (1984). A physiologically based description of the inhalation
pharmacokinetics of styrene in rats and humans. Toxicology and Applied Pharmacology,
73(1), 159-175. doi: 10.1016/0041-008x(84)90064-4
Rice, D., & Barone, S., Jr. (2000). Critical periods of vulnerability for the developing nervous
system: evidence from humans and animal models. Environmental Health Perspectives,
705,511-533.
Riegel, H., & Wolf, G. (1966). Severe neurological deficiencies as a consequence of methyl
alcohol poisoning. Fortschritte der Neurologic, Psychiatric, 34, 346-351.
Rogers, J. M., Barbee, B. D., & Mole, M. L. (1995). Exposure concentration and time (CxT)
relationships in the developmental toxicity of methanol in mice. Toxicologist, 15, 164.
Rogers, J. M., Barbee, B. D., & Rehnberg, B. R (1993). Critical periods of sensitivity for the
developmental toxicity of inhaled methanol. Teratology, 47, 395.
Rogers, J. M., Brannen, K. C., Barbee, B. D., Zucker, R. M., & Degitz, S. J. (2004). Methanol
exposure during gastrulation causes holoprosencephaly, facial dysgenesis, and cervical
vertebral malformations in C57BL/6J mice. Birth Defects Research, Part B:
Developmental and Reproductive Toxicology, 77(2), 80-88. doi: 10.1002/bdrb.20003
Rogers, J. M., & Mole, M. L. (1997). Critical periods of sensitivity to the developmental toxicity
of inhaled methanol in the CD-I mouse. Teratology, 55, 364-372.
Rogers, J. M., Mole, M. L., Chernoff, N., Barbee, B. D., Turner, C. L, Logsdon, T. R., &
Kavlock, R. J. (1993). The developmental toxicity of inhaled methanol in the CD-I
mouse, with quantitative dose-response modeling for estimation of benchmark doses.
Teratology, 47, 175-188.
Rogers, V. V, Wickstrom, M., Liber, K., & MacKinnon, M. D. (2002). Acute and subchronic
mammalian toxicity of naphthalenic acids from oil sands tailings. Toxicological Sciences,
66, 347-355.
Rubinstein, D., Escott, E., & Kelly, J. P. (1995). Methanol intoxication with putaminal and white
matter necrosis: MR and CT findings. American Journal of Neuroradiology, 16(1), 1492-
1494.
Sakanashi, T. M., Rogers, J. M., Fu, S. S., Connelly, L. E., & Keen, C. L. (1996). Influence of
maternal folate status on the developmental toxicity of methanol in the CD-I mouse.
Teratology, 54, 198-206.
March 2011 7-16 DRAFT - DO NOT CITE OR QUOTE
-------
Salzman, M. (2006). Methanol neurotoxicity. Clinical Toxicology, 44, 89-90.
Saxton, E. H., Miller, T. Q., & Collins, J. D. (1999). Migraine complicated by brachial
plexopathy as displayed by MRI and MRA: Aberrant subclavian artery and cervical ribs.
Journal of the National Medical Association (JNMA), 91, 333-341.
Sayers, R. R., Yant, W. P., Schrenk, H. H., Chornyak, J., Pearce, S. J., Patty, F. A., & Linn, J. G.
(1944). Methanol poisoning n Exposure of dogs for brief periods eight times daily to high
concentrations of methanol vapor in air. Journal of Industrial Hygiene and Toxicology,
26, 255-259.
Schaefer, P. W., Grant, P. E., & Gonzalez, R. G. (2000). Diffusion-weighted MR imaging of the
brain. Radiology, 277(2), 331-345.
Schumacher, R., Mai, A., & Gutjahr, P. (1992). Association of rib abnormalities and malignancy
in childhood. European Journal of Pediatrics, 757(6), 432-434.
Sedivec, V, Mraz, M., & Flek, J. (1981). Biological monitoring of persons exposed to methanol
vapours. International Archives of Occupational and Environmental Health, 48, 257-271.
Sefidbakht, S., Rasekhi, A. R., Kamali, K., Borhani, H. A., Salooti, A., Meshksar, A., ...
Nabavizadeh, S. A. (2007). Methanol poisoning: acute MR and CT findings in nine
patients. Neuroradiology, 49, 427-435.
Short, D. W. (1975). The subclavian artery in 16 patients with complete cervical ribs. Journal of
Cardiovascular Surgery, 16, 135-141.
Simintzi, I, Schulpis, K. H., Angelogianni, P., Liapi, C., & Tsakiris, S. (2007). The effect of
aspartame metabolites on the suckling rat frontal cortex acetylcholinesterase An in vitro
study. Food and Chemical Toxicology, 45, 2397-2401.
Siragusa, R. J., Cerda, J. J., Baig, M. M., Burgin, C. W., & Robbins, F. L. (1988). Methanol
production from the degradation of pectin by human colonic bacteria. American Journal
of Clinical Nutrition, 47, 848-851.
Skrzydlewska, E., Elas, M., & Ostrowska, J. (2005). Protective effects of N-acetylcysteine and
vitamin E derivative U83836E on proteins modifications induced by methanol
intoxication. Toxicology Mechanisms and Methods, 75,263-270.
Smith, M., Hopkinson, D. A., & Harris, H. (1971). Developmental changes and polymorphism in
human alcohol dehydrogenase. Annals of Human Genetics, 34(3), 251-271.
Smith, M. E., & Newman, H. W. (1959). The rate of ethanol metabolism in fed and fasting
animals. Journal of Biological Chemistry, 234, 1544-1549.
Soffritti, M., Belpoggi, F, Cevolani, D., Guarino, M., Padovani, M., & Maltoni, C. (2002).
Results of long-term experimental studies on the carcinogenicity of methyl alcohol and
ethyl alcohol in rats. In M. A. Mehlman (Ed.), Carcinogenesis bioassays and protecting
public health: commemorating the lifework of Cesare Maltoni and colleagues (pp. 46-
69). Bologna, Italy: Ann. N. Y. Acad. Sci.
March 2011 7-17 DRAFT - DO NOT CITE OR QUOTE
-------
Spiteri, N. J. (1982). Orcadian patterning of feeding, drinking and activity during diurnal food
access in rats. Physiology and Behavior, 25(1), 139-147. doi: 10.1016/0031-
9384(82)90115-9
Staab, C. A., Lander, J., Brandt, M., Lengqvist, J., Morgenstern, R., Grafstrom, R. C., & Hoog, J.
O. (2008). Reduction of S-nitrosoglutathione by alcohol dehydrogenase 3 is facilitated by
substrate alcohols via direct cofactor recycling and leads to GSH-controlled formation of
glutathione transferase inhibitors. BiochemicalJournal, 413(3), 493-504. doi:
10.1042/bj20071666
Staats, D. A., Fisher, J. W., & Connolly, R. B. (1991). Gastrointestinal absorption of xenobiotics
in physiologically based pharmacokinetic models: a two-compartment description. Drug
Metabolism and Disposition, 19(1), 144-148.
Stanton, M. E., Crofton, K. M., Gray, L. E., Gordon, C. J., Boyes, W. K., Mole, M. L., . . .
Bushnell, P. J. (1995). Assessment of offspring development and behavior following
gestational exposure to inhaled methanol in the rat. Toxicological Sciences, 28, 100-110.
Starr, T. B., & Festa, J. L. (2003). A proposed inhalation reference concentration for methanol.
Regulatory Toxicology and Pharmacology, 38, 224-231.
Stegink, L. D., Brummel, M. C., Filer, L. J., Jr., & Baker, G L. (1983). Blood methanol
concentrations in one-year-old infants administered graded doses of aspartame. Journal of
Nutrition, 113, 1600-1606.
Stegink, L. D., Brummel, M. C., McMartin, K., Martin-Amat, G, Filer, L. J., Jr., Baker, G. L., &
Tephly, T. R. (1981). Blood methanol concentrations in normal adult subjects
administered abuse doses of aspartame. Journal of Toxicology and Environmental Health,
7,281-290.
Stegink, L. D., Filer, L. J., Bell, E. F., Ziegler, E. E., & Tephly, T. R. (1989). Effect of repeated
ingestion of aspartame-sweetened beverage on plasma amino acid, blood methanol, and
blood formate concentrations in normal adults. Metabolism, 38(4), 357-363.
Steiner, E. C., Rey, T. D., & McCroskey, P. S. (1990). Reference guide for SimuSolv. Midland,
MI: Dow Chemical Company.
Stern, S., Reuhl, K., Soderholm, S., Cox, C., Sharma, A., Balys, M., . . . Weiss, B. (1996).
Perinatal methanol exposure in the rat I Blood methanol concentration and neural cell
adhesion molecules. Toxicological Sciences, 34, 36-46.
Sultatos, L. G, Pastino, G. M., Rosenfeld, C. A., & Flynn, E. J. (2004). Incorporation of the
genetic control of alcohol dehydrogenase into a physiologically based pharmacokinetic
model for ethanol in humans. Toxicological Sciences, 78, 20-31.
Svensson, S., Some, M., Lundsjo, A., Helander, A., Cronholm, T., & Hoog, J. O. (1999).
Activities of human alcohol dehydrogenases in the metabolic pathways of ethanol and
serotonin. European Journal of Biochemistry, 262(2), 324-329.
Tanner, C. M. (1992). Occupational and environmental causes of parkinsonism. Occupational
and Environmental Medicine, 7, 503-513.
March 2011 7-18 DRAFT - DO NOT CITE OR QUOTE
-------
Teeguarden, J. G, Deisinger, P. I, Poet, T. S., English, J. C., Faber, W. D., Barton, H. A., ...
Clewell, H. J., IE. (2005). Derivation of a human equivalent concentration for n-butanol
using a physiologically based pharmacokinetic model for n-butyl acetate and metabolites
n-butanol and n-butyric acid. Toxicological Sciences, 55(1), 429-446.
Teng, S., Beard, K., Pourahmad, J., Moridani, M., Easson, E., Poon, R., & O'Brien, P. J. (2001).
The formaldehyde metabolic detoxification enzyme systems and molecular cytotoxic
mechanism in isolated rat hepatocytes. Chemico-Biological Interactions, 130-132, 285-
296. doi: 10.1016/s0009-2797(00)00272-6
Tephly, T. R., & McMartin, K. E. (1984). Methanol metabolism and toxicity. Food Science and
Technology Bulletin: Functional Foods, 12, 111-140.
Thrasher, J. D., & Kilburn, K. H. (2001). Embryo toxicity and teratogenicity of formaldehyde.
Archives of Environmental Health, 56(4), 300-311. doi: 10.1080/00039890109604460
Toth, B. A., Wallcave, L., Patil, K., Schmeltz, I, & and Hoffmann, D. (1977). Induction of
tumors in mice with the herbicide succinic acid 2,2-dimethylhydrazide. Cancer Research,
37, 3497-3500.
Tran, M. N., Wu, A. H., & Hill, D. W. (2007). Alcohol dehydrogenase and catalase content in
perinatal infant and adult livers: potential influence on neonatal alcohol metabolism.
Toxicology Letters, 169(3), 245-252. doi: 10.1016/j.toxlet.2007.01.012
Tsakiris, T., Giannoulia-Karantana, A., Simintzi, I, & Schulpis, K. H. (2006). The effect of
aspartame metabolites on human erythrocyte membrane acetylcholinesterase activity.
Pharmacological Research, 53, 1-5. doi: 10.1016/j.phrs.2005.07.006
Turner, C., Spanel, P., & Smith, D. (2006). A longitudinal study of methanol in the exhaled breath
of 30 healthy volunteers using selected ion flow tube mass spectrometry, SIFT-MS.
Physiological Measurement, 27, 637-648. doi: 10.1088/0967-3334/27/7/007
U.S. EPA. (U.S. Environmental Protection Agency). (1986a). Guidelines for the health risk
assessment of chemical mixtures. (Report No. EPA/63 O/R-98/002). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum. Retrieved from
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=22567.
U.S. EPA. (U.S. Environmental Protection Agency). (1986b). Rat oralsubchronic toxicity study
with methanol. (Report No. TRLNo. 032-005). Washington, DC: Author.
U.S. EPA. (U.S. Environmental Protection Agency). (1988). Recommendations for and
documentation of biological values for use in risk assessment. (Report No. EPA/600/6-
87/008). Cincinnati, OH: U.S. Environmental Protection Agency, Office of Research and
Development. Retrieved from
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=34855.
U.S. EPA. (U.S. Environmental Protection Agency). (1991). Guidelines for developmental
toxicity risk assessment. (Report No. EPA/600/FR-91/001). Washington, DC: U.S.
Environmental Protection Agency, Risk Assessment Forum. Retrieved from
http://www.epa.gov/iris/backgrd.html.
March 2011 7-19 DRAFT - DO NOT CITE OR QUOTE
-------
U.S. EPA. (U.S. Environmental Protection Agency). (1994a). Interim policy for particle size and
limit concentration issues in inhalation toxicity studies. Washington, DC: U.S.
Environmental Protection Agency, Office of Pesticide Products, Health Effects Division.
Retrieved from http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=l86068.
U.S. EPA. (U.S. Environmental Protection Agency). (1994b). Methods for derivation of
inhalation reference concentrations and application of inhalation dosimetry. (Report No.
EPA/600/8-90/066F). Washington, DC: U.S. Environmental Protection Agency, Office of
Research and Development. Retrieved from
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=71993.
U.S. EPA. (U.S. Environmental Protection Agency). (1995). The use of the benchmark dose
approach in health risk assessment. (Report No. EPA/63O/R-94/007). Washington, DC:
U.S. Environmental Protection Agency, Risk Assessment Forum.
U.S. EPA. (U.S. Environmental Protection Agency). (1996). Guidelines for reproductive toxicity
risk assessment. (Report No. EPA/63O/R-96/009). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum. Retrieved from
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=283 8.
U.S. EPA. (U.S. Environmental Protection Agency). (1998). Guidelines for neurotoxicity risk
assessment. (Report No. EPA/630/R-95/001F). Washington, DC: U.S. Environmental
Protection Agency, Risk Assessment Forum. Retrieved from
http ://cfpub. epa. gov/ncea/cfm/recordi splay, cfm? deid= 12479.
U.S. EPA. (U.S. Environmental Protection Agency). (2000a). Benchmark dose technical
guidance document [external review draft]. (Report No. EPA/63O/R-00/001).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
Retrieved from http://www.epa.gov/raf/publications/benchmark-dose-doc-draft.htm.
U.S. EPA. (U.S. Environmental Protection Agency). (2000b, July 13, 2010). Integrated risk
information system: Vinyl chloride, from http://www.epa.gov/ncea/iris/subst/1001.htm
U.S. EPA. (U.S. Environmental Protection Agency). (2000c). Science policy council handbook:
Risk characterization. (Report No. EPA 100-B-00-002). Washington, D.C.: U.S.
Environmental Protection Agency, Office of Research and Development, Office of
Science Policy. Retrieved from http://www.epa.gov/osa/spc/pdfs/rchandbk.pdf.
U.S. EPA. (U.S. Environmental Protection Agency). (2000d). Supplementary guidance for
conducting health risk assessment of chemical mixtures. (Report No. EPA/63 O/R-00/002).
Washington, DC: U.S. Environmental Protection Agency, Risk Assessment Forum.
Retrieved from http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=20533.
U.S. EPA. (U.S. Environmental Protection Agency). (2002). A review of the reference dose and
reference concentration processes. (Report No. EPA/630/P-02/0002F). Washington, DC:
U.S. Environmental Protection Agency, Risk Assessment Forum. Retrieved from
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=55365.
March 2011 7-20 DRAFT - DO NOT CITE OR QUOTE
-------
U.S. EPA. (U.S. Environmental Protection Agency). (2006a). A framework for assessing health
risk of environmental exposures to children. (Report No. EPA/600/R-05/093F).
Washington, DC: U.S. Environmental Protection Agency, Office of Research and
Development. Retrieved from
http://cfpub.epa.gov/ncea/cfm/recordisplay. cfm?deid=l 58363.
U.S. EPA. (U.S. Environmental Protection Agency). (2006b). U.S. Environmental Protection
Agency peer review handbook. (Report No. EPA/1 OO/B-06/002). Washington, DC: U.S.
Environmental Protection Agency, Science Policy Council. Retrieved from
http://www. epa.gov/peerrevi ew/pdfs/peer_revi ew_handbook_2006.pdf.
U.S. EPA. (U.S. Environmental Protection Agency). (2009a, December 19, 2009). Benchmark
dose software (BMDS), from http://www.epa.gov/NCEA/bmds
U.S. EPA. (U.S. Environmental Protection Agency). (2009b, December 18, 2009). Section E -
Quantities of TRI chemicals in waste, 2008, from
http://www. epa.gov/tri/tridata/tri08/nati onal_analysis/index.htm
U.S. EPA. (U.S. Environmental Protection Agency). (2009c, December 18, 2009). The Toxics
Release Inventory (TRI) and factors to consider when using TRI data, from
http://www.epa.gov/tri/triprogram/FactorsToConPDF.pdf
U.S. EPA. (U.S. Environmental Protection Agency). (2009d, December 18, 2009). TRI on-site
and off-site reported disposed of or otherwise released in pounds for facilities in all
industries for hazardous air pollutant chemicals U.S. 2008, from
http://www. epa.gov/triexplorer/
Vara-Castrodeza, A., Perez-Castrillon, J. L., & Duenas-Laita, A. (2007). Magnetic resonance
imaging in methanol poisoning. Clinical Toxicology, 45, 429-430.
Ward, K. W., Blumenthal, G. M., Welsch, R, & Pollack, G. M. (1997). Development of a
physiologically based pharmacokinetic model to describe the disposition of methanol in
pregnant rats and mice. Toxicology and Applied Pharmacology, 145, 311-322.
Ward, K. W., Perkins, R. A., Kawagoe, J. L., & Pollack, G. M. (1995). Comparative
toxicokinetics of methanol in the female mouse and rat. Toxicological Sciences, 26, 258-
264.
Ward, K. W., & Pollack, G. M. (1996). Comparative toxicokinetics of methanol in pregnant and
nonpregnant rodents. Drug Metabolism and Disposition, 24, 1062-1070.
Weiss, B., Stern, S., Soderholm, S. C., Cox, C., Sharma, A., Inglis, G. B., . . . Gelein, R. (1996).
Developmental neurotoxicity of methanol exposure by inhalation in rats. (Report No. HEI
Research Report Number 73). Boston, MA: Health Effects Institute.
Wentzel, P., & Eriksson, U. J. (2006). Ethanol-induced fetal dysmorphogenesis in the mouse is
diminished by high antioxidative capacity of the mother. Toxicological Sciences, 2(2),
416-422. doi: 10.1093/toxsci/kfl024
Wentzel, P., Rydberg, U., & Eriksson, U. J. (2006). Antioxidative treatment diminishes ethanol-
induced congenital malformations in the rat. Alcoholism: Clinical and Experimental
Research, 30(10), 1752-1760. doi: 10.1111/j.l530-0277.2006.00208.x
March 2011 7-21 DRAFT - DO NOT CITE OR QUOTE
-------
White, L. R., Marthinsen, A. B. L., Richards, R. J., Eik-Nes, K. B., & Nilsen, O. G. (1983).
Biochemical and cytological studies of rat lung after inhalation of methanol vapour.
Toxicology Letters, 77(1-2), 1-5.
Wilson, S. L., & Cradock, M. M. (2004). Review: Accounting for prematurity in developmental
assessment and the use of age-adjusted scores. Journal of Pediatric Psychology, 29, 641-
649. doi: 10.1093/jpepsy/jsh067
Wu, A. H. B., Kelly, T., McKay, C., & Ostheimer, D. (1995). Definitive identification of an
exceptionally high methanol concentration in an intoxication of a surviving infant:
Methanol metabolism by first-order elimination kinetics. Journal of Forensic Sciences,
¥0,315-320.
Yuan, J. H. (1993). Modeling blood/plasma concentrations in dosed feed and dosed drinking
water toxicology studies. Toxicology and Applied Pharmacology, 119, 131-141.
Zorzano, A., & Herrera, E. (1989). Disposition of ethanol and acetaldehyde in late pregnant rats
and their fetuses. Pediatric Research, 25(1), 102-106.
March 2011 7-22 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX A. SUMMARY OF EXTERNAL PEER REVIEW AND PUBLIC
COMMENTS AND DISPOSITION
[page intentionally left blank]
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 A-1 DRAFT - DO NOT CITE OR QUOTE
-------
APPENDIX B. DEVELOPMENT, CALIBRATION AND APPLICATION OF A
METHANOL PBPK MODEL
B.I. SUMMARY
1 This appendix describes the development, calibration, and approach for application of
2 mouse, rat, and human PBPK models to extrapolate mouse and rat methanol inhalation-route
3 internal dose metrics to human inhalation exposure concentrations that result in the same internal
4 dose (HEC). The human oral methanol dose(s) yielding internal dose(s) equivalent to the mouse
5 or rat internal dose at the (HED) is also presented.
6 A PBPK model was developed to describe the blood kinetics of methanol (MeOH) in
7 mice and humans. The model includes compartments for lung/blood MeOH exchange, liver, fat,
8 and the rest of the body. To describe blood MeOH kinetics, the model employs two saturable
9 descriptions of MeOH metabolism in mice and SD rats, one saturable metabolic pathway in F344
10 rats and humans, and a first-order description of renal clearance (from blood) in humans. Renal
11 clearance is a minor pathway and does not appreciably affect MeOH blood kinetics, but methanol
12 concentrations in urine are an important indicator of the corresponding blood levels.
13 This model is a revision of the model reported by Ward et al. (1997), reflecting significant
14 simplifications (removal of compartments for placentae, embryo/fetus, and extraembrionic fluid)
15 and two elaborations (addition of an intestine lumen compartment to the existing stomach lumen
16 compartment and addition of a bladder compartment which impacts simulations for human
17 urinary excretion.), while maintaining the ability to describe MeOH blood kinetics. The model
18 reported here uses a single consistent set of parameters; the Ward et al. model employed a
19 number of data-set specific parameters. Other biokinetic MeOH models that were considered as
20 starting points for the current model also used varied parameters by dataset to achieve model fits
21 to the data. For example, the model of Bouchard et al. (2001) used different respiratory rates and
22 fractional inhalation absorbed for different human exposures.
23 The mouse model was calibrated against inhalation-route blood MeOH kinetic data and
24 verified using intravenous-route blood MeOH kinetic data. The rat model was calibrated against
25 low-dose intravenous data and validated with inhalation-route data. The human model was
26 calibrated against inhalation-route MeOH kinetic data. The models accurately described the
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 B-1 DRAFT—DO NOT CITE OR QUOTE
-------
1 inhalation route pharmacokinetics of MeOH. Mouse model simulations of oral- and i.v.-route
2 kinetics compare well to some but not all the experimental data.
3 The MeOH HECs predicted by the human model (based on 1,000 ppm inhalation
4 exposure in mice) were >1,000 ppm using either blood AUC or Cmax as the dose metrics. The
5 MeOH HED derived by cross-route extrapolation of this inhalation-route HEC was 110 mg/kg-
6 day, based on MeOH blood AUC following zero order uptake of MeOH (a constant rate of
7 delivery). Because of the lack of human data from high-dose exposures, it was not possible to
8 calibrate the human model for inhalation exposures above 1,000 ppm or oral exposures above
9 110 mg/kg-day. However, application of the human PBPK model to the internal experimental
10 animal doses estimated via the BMD approach resulted in RfC and RfD PODs that are below
11 1,000 ppm or 110 mg/kg-day, respectively (see Sections 5.1.3.1 and 5.2.2).
B.2. MODEL DEVELOPMENT
B.2.1. Model Structure
12 This model is a revision of the model reported by Ward et al. (1997), reflecting significant
13 simplifications and two elaborations, while maintaining the ability to describe MeOH blood
14 kinetics in mice, rats, and humans (Figure B-l). The kidney, pregnancy and the fetal
15 compartment have been removed. The kidney was lumped with the body compartment because
16 the blood:tissue partition coefficients for these tissues were similar. The elaborate time-
17 dependent descriptions of pregnancy were removed because analysis of the available
18 pharmacokinetic data indicates that blood MeOH kinetics in NP and pregnant mice are not
19 different enough to warrant separate descriptions. Because the maternal blood:fetal blood
20 partition coefficients were near 1, there was no need to explicitly model fetal kinetics; they will
21 be equivalent to maternal blood kinetics. Further supporting data exist for ethanol, which is quite
22 similar to MeOH in its partitioning and transport properties. In rats (Guerri & Sanchis, 1985;
23 Zorzano & Herrera, 1989), sheep (Brien et al., 1985; Cumming et al., 1984), and guinea pigs
24 (Clarke et al., 1986), fetal and maternal blood concentrations of ethanol are virtually
25 superimposable; maternal to fetal blood ratios are very close to 1, including during late gestation.
26 Also, fetal brain concentrations in guinea pigs (Clarke et al., 1986) were also very similar to the
27 mothers'.
28 In addition to the absolute maternal-fetal concentration similarity noted above, it is
29 common practice to use blood concentrations as an appropriate metric for risk extrapolation via
30 PBPK modeling for effects in various tissues, based on the reasonable expectation that any
31 tissue:blood differences will be similar in both the test species and humans. For example, even if
32 the brain:blood ratio was around 1.2 in the mouse or rat, the similar biochemical make-up of
March 2011 B-2 DRAFT—DO NOT CITE OR QUOTE
-------
1 brain tissue and blood in rats and humans leads to the expectation that the brain:blood levels in
2 humans (which depend on the biochemical make-up) will also be close to 1.2, and so the relative
3 "error" that might occur by using blood instead of brain concentration in evaluating the dose-
4 response in rats will be cancelled out by using blood instead of brain concentration in the human.
5 The fact that measured fetal blood levels are virtually identical to maternal levels for methanol
6 (and ethanol) tells us that the rate of metabolism in the fetus is not sufficient to significantly
7 reduce the fetal concentration versus maternal, and use of a PBPK model to predict maternal
8 levels will give a better estimate of fetal exposure than use of the applied dose or exposure,
9 because there are animal-human differences in adult PK of MeOH for which the model accounts,
10 based on PK data from humans as well as rodents.
Inhalal
Expos
RlaHHpr ^ ^
KBL| (human only)
Urine IV — *
KLLor
\/m3Y
tion
ure
-w
Venous Blood
^—- -
r
4
^^^_^
Alveolar Air
Lung Blood
Body
Fat
Liver
^E:
>j
•
<~
xha
Arterial Blood
,
Vmax2
Km2
Metabolites
^KASorVmAS.KMAS
Oral
Exposure
Feces
Figure B-l. Schematic of the PBPK model used to describe the inhalation,
oral, and i.v. route pharmacokinetics of MeOH. KAS, first-order oral
absorption rate from stomach; KAI, first-order uptake from the intestine;
KSI, first-order transfer between stomach and intestine; Vmax, Km, Vmax2,
and Km2, Michaelis-Menten rate constants for high affinity/low capacity and
low affinity/high capacity metabolism of MeOH; KLL, alternate first-order
rate constant; KBL, rate constant for urinary excretion from bladder. Both
metabolic pathways were used to describe MeOH metabolism in the mouse
and SD rat, while a single pathway describes metabolism in the F344 rat and
human.
March 2011
B-3
DRAFT—DO NOT CITE OR QUOTE
-------
1 A lung compartment was added to describe delivery of MeOH to blood as a function of
2 ventilation, partitioning, and blood flow rather than the less standard approach used by Ward et
3 al. (1997). A term was added to the gas uptake equations to describe the fractional respiratory
4 bioavailability of MeOH. A fat compartment was included because it is the only tissue with a
5 tissue:blood partitioning coefficient appreciably different than unity, and the liver is included
6 because it is the primary site of metabolism. A bladder compartment was added to better
7 describe the kinetics of human urinary data, where the drop in excretion rate is slower than the
8 predicted decline in blood methanol and hence rate of metabolite production. Also, to best
9 describe the observed rat dosimetry after oral exposure while maintaining metabolic parameters
10 fit to data from inhalation and IV exposure, a small rate of elimination from the intestine (lumen)
11 compartment to feces. (The mouse data could be adequately fit with this rate set to zero,
12 corresponding to 100% absorption; humans were assumed to have zero fecal elimination, like the
13 mouse.) The final models thus include compartments for fat, liver, the rest of the body, bladder
14 (only used for humans), and lung. The mouse and rat models describe inhalation, oral, and
15 intravenous route dosing and the human model describes inhalation and oral route dosing and the
16 rat model includes a non-zero rate of fecal elimination. Although there is an endogenous
17 background level of both MeOH and formate (See Section 3.3), the model does not explicitly
18 describe or account for background levels of MeOH or formate. In this analysis, when non-zero
19 background levels have been measured (in blood), that background was simply subtracted from
20 the concentrations measured during exposure. However, a zero-order rate of infusion could be
21 added to the liver, blood, or stomach compartments to mimic background levels if that was
22 considered necessary.
23 MeOH is well absorbed by the inhalation and oral routes, and is readily metabolized to
24 formaldehyde, which is rapidly converted to formate in both rodents and humans. Although the
25 enzymes responsible for metabolizing formaldehyde are different in rodents (CAT) and humans
26 (ALD) the metabolite, formate, is the same, and the metabolic rates are similar (Clary, 2003).
27 Most of the published rodent kinetic models for MeOH describe the metabolism of MeOH to
28 formaldehyde as a saturable process but differ in the handling of formate metabolism and
29 excretion (Bouchard et al.. 2001: Fisher et al.. 2000: Horton et al.. 1992: Wardetal.. 1997).
30 Ward et al. (1997) used one saturable and one first-order pathway for mice, and Horton et al.
31 (1992) applied two saturable pathways of metabolism to describe MeOH elimination in rats.
32 Bouchard et al. (2001) employed one metabolic pathway and a second pathway described as
33 urinary elimination in rats and humans. The need for two saturable metabolic pathways in the
34 mouse model was confirmed through simulation and optimization. High exposure (>2,000 ppm
35 MeOH) and low exposure (1,000 ppm MeOH) blood data could not be adequately fit either
36 visually or by more formal optimization without the second saturable metabolic pathway. The
March 2011 B-4 DRAFT—DO NOT CITE OR QUOTE
-------
1 optimization approach and results are found below and in the Additional Materials at the end of
2 this appendix.
3 While the PPK model explicitly describes the concentration of methanol, it only describes
4 the rate of metabolism or conversion of MeOH to its metabolites. Distribution and metabolism
5 of formaldehyde is not considered by the model, and this model tracks neither formate nor
6 formaldehyde. (The data that would be needed to parameterize or validate a specific description
7 of either of these metabolites is not available). Since the metabolic conversion of formaldehyde
8 to formate is rapid (< 1 minute) in all species (Kavet & Nauss, 1990), the MeOH metabolism rate
9 should approximate a formate production rate, though this has not been verified. Thus the rate of
10 MeOH metabolism predicted by the model can be used as a dose metric for either or both of
11 these metabolites, but scaling of that metabolic rate metric to humans requires that the rate be
12 normalized to BW°'75, (i.e., scaled rate = mg/kg°'75-time ); to account for the general expectation
13 metabolic elimination of the metabolites scales as BW°'75, hence is slower in humans.
14 The model was initially coded in acslXtreme vl.4 and updated in acslXtreme v 2.3. Most
15 procedures used to generate this report, except those for the optimization, may be run by
16 executing the corresponding .m files. The model code (acslXtreme .csl file) and supporting .m
17 files are available electronically and as text in the Additional Materials at the end of this
18 appendix. A key identifying .m files associated with figures and tables in this report is also
19 provided in the Additional Materials.
B.2.2. Model Parameters
20 Physiological parameters such as tissue volumes, blood flows, and ventilation rates were
21 obtained from the open literature (Table B-l). Parameters for blood flow, ventilation, and
22 metabolic capacity were scaled as a function of body weight raised to the 0.75 power, according
23 to the methods of Ramsey and Andersen (1984).
March 2011 B-5 DRAFT—DO NOT CITE OR QUOTE
-------
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 et al. (1997)
Calculated'
Flows (L/hr/kg0'75)
Alveolar
ventillation''
Cardiac output
25.4
25.4
16.4
16.4
16.5
24.0
Brown et al. (1997): Perkins et al. (1995a); U.S. EPA
(2000b)
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
o o
3 NA
4 22.3
65 100
NA
NA
22.7
5.2
72.1
1s*
, saturable
order
,T. 33.1
NA
,T. 23.7
NA
NA
NA
0.037 0.0342
3
95.7 NA
Brown et al. (1997)
Calculated
Fitted
Oral absorption
VmASC
(mg/hr/kg075)
KMASC (mg/kg)
KSI (hr"1)
KAI (hr"1)
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/KM — the first -order
constant of Sultatos et al.
Partition coefficients
LiverBlood
Fat:Blood
1.06
0.083
1.06
0.083
0.583*
0.142
Fiserova-Bergerova & Diaz (1986): Ward et al. (1997)
March 2011
B-6
DRAFT—DO NOT CITE OR QUOTE
-------
Blood:Air
Body:Blood
Lung:Blood
Bladder time-constant
(KBL, hf Y
Inhalation fractional
availability (%)
Mouse
1350'
0.66
1
Rat
SD | F344
1350
0.66
1
NA
0.665
0.20
Human
1626
0.805
1.07
0.564 0.612
0.866*
Source
Fiserova-Bergerova & Diaz (1986): Horton et al.
(19921
Rodent: estimated; human: (Fiserova-Bergerova &
Diaz, 1986) (human "body" assumed = muscle)
Fitted (human)
Rodent: fitted; human
(Ernstgardetal.,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/B W°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 (2000b). This QPC
was somewhat higher than calculated from Brown et al. (1997) (~13 L/hr/kg°75)
eVmax, Km, and Vmax2, Km2 represent the two saturable metabolic processes assumed to occur solely in the liver.
The Vmax used in the model = VmaxC (mg/kg° 75-hr) XBW°75. 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 (Teeguardenetal.. 2005); The Kl used in the model= K1C / BW°25
fKLLC - alternate human first-order metabolism rate (used only when VmaxC = Vmax2C = 0)
gHuman 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 elimination 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
1 Mouse model partition coefficients were used as reported (liver, fat, blood:air) or
2 estimated (lung, body). The mouse body compartment partition coefficient was set
3 approximately equal to the measured value for muscle (Ward et al., 1997). The mouse lung
4 partition coefficient was assumed to be 1.0, similar to the liver partition coefficient. This
5 parameter has no numerically significant impact on modeled blood dose metrics.
6 Human partition coefficients were reported by Horton et al. (1992), but were in fact
7 measured in rat tissues. The reported rat fat partition coefficient was considerably closer to unity
8 than reported for MeOH or ethanol by other researchers (Pastino & Conolly, 2000; Ward et al.,
9 1997) and assumed to be in error. Human partition coefficients were obtained from Fiserova-
10 Bergerova and Diaz (1986).
March 2011
B-7
DRAFT—DO NOT CITE OR QUOTE
-------
B.2.3. Mouse Model Calibration
B.2.3.1. Inhalation-Route Calibration
1 For purposes of conducting interspecies extrapolations of MeOH dosimetry, the
2 inhalation route was the most important route requiring calibration for the mouse model. The
3 critical endpoint and NOEL, which are the basis for the HEC estimation, are from inhalation-
4 route studies. The ability to predict blood MeOH concentrations from inhalation exposures was
5 therefore a priority. Pharmacokinetic data from other routes, i.v. and oral, were used to verify
6 elimination terms derived by fitting to the inhalation data or to estimate a MeOH oral uptake rate
7 constants. Holding other parameters constant, the mouse PBPK model was calibrated against
8 inhalation-route blood pharmacokinetic data (Figure B-2) by fitting five parameters: Michaelis-
9 Menten constants for one high affinity/low capacity and one low-affinity high-capacity enzyme
10 and the inhalation fractional availability term.
11
12
13
14
15
10,000
1,000
I
O
0)
£
•a
O
O
100
10
— 1000 ppm sirn
— 2000 ppm sim
— 5000 ppm sim
— 7500 ppm sim
— 10000 ppm sim
15000 ppm sirn
i 1000 ppm data (GD6 &. 10)
• 2000 ppm data (GD6 & 10)
• 5000 ppm data (GD6 &. 10)
* 7500 ppm data (GD6 & 10)
A 10000 ppm data (GD6 & 10)
T 10000 pprn data (GD7)
p 15000 pprn data (GD6 & 10)
12 15
Time (hr)
18
21
24
27
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/kg0'75)
reported in Brown et al. (1997). Where ventilation rates were reported for individual exposure
concentrations by Perkins et al. (1995a), they were used directly in the model and a notation was
made in the figure legend. Reported ventilation rates varied from 592 to 857 L/kg x 8 hr,
March 2011
B-8
DRAFT—DO NOT CITE OR QUOTE
-------
1 depending on exposure concentration (Perkins et al., 1995a). Adjusting these values to 2/3 total
2 (for alveolar ventilation) and allometrically scaling by BW°'75, values used in the model for these
3 exposures ranged from 20.5 to 29.7 (L/hr/kg0'75) (See Table B-l). A fractional availability of
4 73% of alveolar ventilation was visually optimized to best describe the inhalation-route blood
5 MeOH pharmacokinetic data. This percentage of uptake for inhalation exposures is similar to
6 values reported for other alcohols in rodents (Teeguarden et al., 2005), but considerably lower
7 than the value reported by Perkins et al. (1995a) of 126% of alveolar ventilation (85% of total
8 ventilation).
9 The calibrated model predicted blood MeOH concentration time-course agreed well with
10 measured values in adult mice in the inhalation studies of Rogers et al. (1997; 1993) (Figure B-
11 2), and Perkins et al. (1995a), as well as in NP and early gestation (GD8) mice of Dorman et al.
12 (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-
w
* 3,000 -I
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
15 20
Time (hr)
35
Figure B-3. Simulation of inhalation exposures to MeOH in NP mice from
Perkins et al. (1995a) (8-hour exposures) and Dorman et al. (1995), (6-hour
exposures). Data points represent measured blood MeOH concentrations
and lines represent PBPK model simulations. Note: data was obtained using
Digitizlt (Sharlt! Inc. Greensburg, PA) to digitize data from Figure 2 of
Perkins et al. (1995a) and Figure B-2 from Dorman et al. (1995). Default
ventilation rates (Table B-l) 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/hr/kgO.75), respectively. The cardiac output for these
simulations was set equal to the alveolar ventilation rate.
March 2011
B-9
DRAFT—DO NOT CITE OR QUOTE
-------
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-route
9 blood MeOH kinetics for each GD. For instance, the fraction of the total oral dose assigned to
10 the fast absorption process varied from 54 to 71%, depending on the data set. An alternative
11 approach with uptake attributed to stomach and intestine, which allows for greater flexibility in
12 fitting the data (Staats et al., 1991), was compared to a simpler one utilizing a single rate of
13 uptake. In both the current model and the model of Ward et al. (1997), orally ingested MeOH
14 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
March 2011 B-10 DRAFT—DO NOT CITE OR QUOTE
-------
1 Materials, Figure B-20). The two-compartment GI tract approach (with parameters that better fit
2 the low dose data) was retained in the model and used for all final mouse oral route simulations.
3,500-
3,000 -
^^
2,500-
2,000-
o<
E
•"_"•
O
Qj
1,500-
•o
J 1,000-
00
500-
0 -J
• Oral 1500 mg/kg GD 8 mouse
• Oral 2500 mg/kg NP mouse
n Oral 2500 mg/kg GD 18 mouse
— 1500 mg/kg simuluation
— 2500 mg/kg simuluation
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).
3 Using the two-compartment oral absorption model and adjusting only the absorption
4 parameters resulted in a good fit to the lower oral dose (1,500 mg/kg) (Dorman et al., 1995), but
5 consistently under-prediction of the 2,500 mg/kg oral dosing blood levels (Ward et al., 1997).
6 When the metabolic constants (VmaxC values) were decreased, the data from the higher dose
7 were fit, but the fit of the data for the 1,500 mg/kg dose was lost (see Additional Materials,
8 Figure B-19). Also, when using the lower metabolic rate constants required to fit the data of
9 Ward et al. (1997), the inhalation data of Rogers et al. (1993) could no longer be fit by the model
10 (see Additional Materials, Figure B-20). The two-compartment GI tract approach (with
11 parameters that better fit the low dose data) was retained in the model and used for all final
12 mouse oral route simulations.
B.2.3.3. Intravenous Route Simulation
13 The parameterization of MeOH metabolism (high-and-low affinity metabolic pathways)
14 was verified by simulation of data sets describing the intravenous-route pharmacokinetics of
15 MeOH. MeOH blood kinetics data in NP mice are only available for a single i.v. dose of
16 2,500 mg/kg (Ward et al., 1997). MeOH blood kinetics are also reported in GDIS mice
March 2011
B-11
DRAFT—DO NOT CITE OR QUOTE
-------
1 following administration of a broader range of doses: 100, 500, and 2,500 mg/kg. Because
2 MeOH kinetics appear similar for NP and pregnant mice after administration of 2,500 mg/kg
3 prior to 20 hours, the model is expected to fit data for both pregnant and NP mice using the same
4 set of parameters, and hence, data for both life stages were used to verify overall clearance
5 (including metabolism) of MeOH.
6 Initial blood concentrations of MeOH following i.v. administration were not proportional
7 to administered dose in the data from Ward et al. (1997), but were approximately 1.5-fold lower
8 in the 100 mg/kg dose group than expected if a dose-independent volume of distribution (Vo) is
9 assumed (Figure B-5 A). Initial blood concentrations were, however, proportional to
10 administered dose between 2,500 and 500 mg/kg. To account for this unexpected
11 nonproportionality, Ward et al. (1997) used higher partition coefficients for placenta and
12 embryonic fluid and a lower Vmax for the metabolism of MeOH for the 100 and 500 mg/kg doses
13 than for the 2,500 mg/kg dose. These adjustments to partition coefficients effectively make the
14 volume of distribution (VD) dose-dependent. However, the PBPK model obtained here, with
15 measured partition coefficients and otherwise calibrated to inhalation data as described above,
16 was capable of simulating both the 500 and 2,500 mg/kg data without adjustment or varying
17 parameters between those 2 doses.
18 From a physico-chemical (mechanistic) basis, VD should only depend on the tissue:blood
19 partitioning, which for small organic (non-polar) molecules such as methanol is not expected to
20 have any concentration- or dose-dependence. Hence the VD should be adequately predicted by
21 the PBPK model with the independently measured partition coefficients. If there were some
22 dose-dependence one would expect the value at 500 mg/kg to be intermediate between the values
23 at 100 and 2,500 mg/kg doses, but that was not the case. Further, no biochemical mechanism has
24 been suggested (by Ward et al. or others) which could explain such dose-dependence. Thus the
25 apparent change in VD at 100 mg/kg is highly unlikely, based on mechanistic considerations and
26 past experience with similar organic compounds. Finally, the data at the nominal dose of 100
27 mg/kg could also be adequately fit without other parameter adjustment simply by simulating a
28 dose of 200 mg/kg (dotted line, Figure B-5B). The fact that this alternate simulation differs in
29 dose by a factor of 2 suggests another possibility: a dilution error occurred in the preparation of
30 the dosing solution by Ward et al. (1997) (i.e., one serial dilution step was skipped).
March 2011 B-12 DRAFT—DO NOT CITE OR QUOTE
-------
10,000
~ 1,000
o
QJ
_o
m
100
10
- IV 2500 simulation
— IV 500 simulation
- IV 100 simulation
• IV 2500 NP mouse
o IV 2500 GD 8 mouse
IV 2500 GD 18 mouse
A IV 500 GD 18 mouse
• IV 100 GD 18 mouse
10
15
20
Time (hr)
1,000
100
I
O
Oj
10
-— IV 200 simulation
— IV 100 simulation
• IV 200 GD IS mouse
B
\
0.5
1,5
2 2.5
Time (hr)
3,5
4.5
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
GDIS 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 ug/ml) and so are not
shown.
Source: Adapted from Ward et al. (1997).
1 Given these considerations and observations regarding dosimetry and distribution, the
2 ability of the model to fit the high- (2,500 mg/kg) and mid-dose (500 mg/kg) intravenous-route
March 2011
B-13
DRAFT—DO NOT CITE OR QUOTE
-------
1 pharmacokinetic data without adjustment was considered sufficient to validate the parameters
2 calibrated from the inhalation studies for the metabolic elimination of MeOH. Metabolic
3 constants reasonably predict blood MeOH kinetics following a 2,500 mg/kg dose in NP animals
4 until 12 hours postexposure, but under predict blood MeOH in GD9 and GDIS mice at 8 hours
5 of exposure and beyond, and under-predict levels in both NP and pregnant mice at 15 hours and
6 beyond. At this high-dose, where blood kinetics of MeOH were reported in NP, GD9, and GDIS
7 mice, the data for the GDIS mice was inconsistent with the GD9 and NP animals. The GD9 data
8 at 12 hours appears inconsistent with the NP data, but then the 2 are nearly identical again at 15
9 hours, so it is not clear if that difference at 12 hours is real or just due to experimental variability.
10 Blood levels of MeOH were -500 mg/L in GDIS mice at 24 hours, but were nondetectable after
11 18 hours in the other groups (detection limit 2 mg/L). Blood concentrations were accurately
12 predicted following administration of 500 mg/kg MeOH (Figure B-5 A). The model predictions
13 did not match the 100 mg/kg data unless one assumed an error in dose preparation, as described
14 above (Figure B-5B). While it is very unusual to suggest such an error in published, peer-
15 reviewed experimental data, since no other adequate explanation (mechanism) is available, such
16 dose-dependence in VD has not been observed for similar organic compounds, and the error
17 suggested is a fairly simple one, this seems a far more likely explanation than the alternative.
18 The calibration of the MeOH PBPK model is consistent with both the available inhalation and
19 oral-route data.
B.2.3.4. Total methanol metabolism
20 Quantifying production of formaldehyde following MeOH exposure for use as an
21 alternative dose metric is of particular interest because formaldehyde is also undergoing toxicity
22 assessment. However, it is important to understand that because the model was developed to
23 describe blood MeOH kinetics, metabolism of MeOH to neither formaldehyde nor formate is
24 specifically described; the model tracks neither of these metabolites. While the metabolism of
25 MeOH described by the model may be presumed to equate with formaldehyde production, this
26 metabolic flux simply leaves the computational model system without specific attribution. Since
27 the metabolic conversion of formaldehyde to formate is rapid in all species (< 1 minute) (Kavet
28 & Nauss, 1990), the MeOH clearance rate may also approximate a formate production as well as
29 a formaldehyde production rate, though this has not been verified.
30 Thus, production of formaldehyde or formate following exposure to MeOH can only be
31 estimated by summing the total amount of MeOH eliminated by metabolic processes. If used,
32 this metric of formaldehyde or formate dose should be scaled by BW075 to adjust for expected
33 species differences in the clearance of these two metabolites (this is scaling reflects the generally
34 accepted assumption that metabolic elimination [of formaldehyde or formate] scales as BW075;
March 2011 B-14 DRAFT—DO NOT CITE OR QUOTE
-------
1 if the metabolic rate is scaled this way, then equal scaled rates in animals and humans is expected
2 to result in equal body burdens or concentrations of the toxic metabolites). The total rate of
3 MeOH metabolism is assumed to equal the total amount of metabolites produced. Values of total
4 MeOH metabolism as a function of exposure in mice and humans are presented in the Additional
5 Materials (Tables B-6, B-7, and B-8).
B.2.3.5. Formal optimization of mouse model parameters
6 Formal optimization of five parameters (inhalation fractional availability and the Vmax
7 and Kms for high and low affinity MeOH metabolism) was attempted using optimization
8 routines in acslXtreme v2.01.1.2. Under the best circumstances, formal optimizations offer the
9 benefit of repeatability and confirmation that global optima have not been significantly missed by
10 user-guided visual optimization. Incorporating judgments regarding the value of specific data
11 sets, while possible when visually fitting, is more difficult when using optimization routines.
12 This is an important distinction between these approaches for this modeling exercise.
13 The mouse inhalation route NOEL was less than 1,000 ppm MeOH. The model is
14 calibrated against inhalation-route data because of the importance of this exposure route in the
15 assessment. Unfortunately, the vast majority of the MeOH data came from much higher
16 exposure concentrations. As expected, various attempts at formal optimization lead to improved
17 fits for some but never all data sets. This is to be expected when there is significant variability in
18 the underlying data. Various data-weighting schemes were included to improve overall
19 optimization while maintaining a good fit to the lowest concentration (1,000 ppm) data. In the
20 end, formal optimization provided no significant improvement over the fractional availability and
21 metabolic parameter values obtained by visual optimization, so these were retained in the final
22 version of the model.
23 Further details on the approach and results from the formal optimization are found in the
24 Additional Materials in outline format with supporting figures. More complete documentation
25 was not developed because the products of the optimizations were not used in the final model.
26 The documentation is intended only to demonstrate that appropriate optimizations were
27 conducted and what the results of those optimizations were.
B.2.4. Mouse Model Sensitivity Analysis
28 An evaluation of the importance of selected parameters on mouse model estimates of
29 blood MeOH AUC was performed by conducting a sensitivity analysis using the subroutines
30 within acslXtreme. Files for reproducing the sensitivity analysis are available in the model as
31 described in the additional materials. The analysis was conducted by measuring the change in
32 model output corresponding to a 1% change in a given model parameter when all other
33 parameters were held fixed. A normalized sensitivity coefficient of 1 indicates that there is a
March 2011 B-15 DRAFT—DO NOT CITE OR QUOTE
-------
1 one-to-one relationship between the fractional change in the parameter and model output; values
2 close to zero indicate a small effect on model output. A positive value for the normalized
3 sensitivity coefficient indicates that the output and the corresponding model parameter are
4 directly related while a negative value indicates they are inversely related.
5 Sensitivity analyses were conducted for the inhalation and oral routes. The
6 inhalation-route analysis was conducted under the exposure conditions of Rogers and Mole
7 (1997) and Rogers et al. (1993), 7-hour inhalation exposures at the NOEL concentration of
8 1,000 ppm. The oral route sensitivity analysis was conducted for an oral dose of 1,000 mg/kg.
9 Parameters with sensitivity coefficients less than 0.1 are not reported. The parameters
10 with the largest sensitivity coefficients for the inhalation route at 1,000 ppm were pulmonary
11 ventilation, VmaxC, and partitioning to the body compartment (Figures B-6 [metabolism] and B-7
12 [flows and partition coefficients]). MeOH AUC was also sensitive to KM2 and VmaxC. The
13 sensitivity coefficient for pulmonary ventilation increases from 1 to -1.75 during the exposure
14 period as metabolism begins to saturate. The sensitivity coefficient is 1 for concentrations 100
15 ppm MeOH or less or when hepatic elimination is nonlimiting.
16 Oral-route mouse blood MeOH AUC was sensitive to the rate constants for uptake.
17 Blood AUC was most sensitive to the first-order rate constant for uptake from the stomach, KAS,
18 during the first hour after exposure, becoming less important over time (Figure B-8). Blood
19 MeOH AUC was also modestly sensitive to KAI, and KSI, the rate constants for uptake from the
20 intestine and transfer rates between compartments, respectively.
March 2011 B-16 DRAFT—DO NOT CITE OR QUOTE
-------
0)
u
1
0)
o
u
>
.*;
5
in
c
0)
1 "
.U "
"
.U "
y| -
Km
r~ /
-. VmiY^P
\
%
X
\
V
V
*
^ .^
* 'X
* -1
^_. _.,— -»—- VmaxC - - -
6 8 10
Time (hr)
12
14
16
Figure B-6. Mouse model inhalation route sensitivity coefficients for
metabolic parameters. Sensitivity coefficients calculated for an exposure of
1,000 ppm MeOH are reported for blood MeOH AUC. Note: Km, Vmax
refer to the high-affinity, low-capacity pathway and Km2, Vmax2 refer to the
low-affinity, high-capacity pathway.
0)
0
fc
0)
o
o
'in
0)
t
.0
.0
-i =; -
1
OC
o -
-u.o
- 1
(
x" " '^%-
^, - ^* *
QPC "
|
/QCC
_?b£
1
) 2 4 6 8 10 12 14 1
Time (hr)
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.
March 2011
B-17
DRAFT—DO NOT CITE OR QUOTE
-------
6 8 10
Time (hr)
12
14
16
Figure B-8. Mouse model sensitivity coefficients for oral exposures to MeOH.
The sensitivity of blood MeOH AUC to oral absorption rate constants (KAS:
stomach; KAI: intestine; KSI: transfer between compartments) is reported.
B.2.5. Mouse drinking water ingestion pattern
1 To simulate exposures of mice via drinking water under bioassay conditions, an ingestion
2 pattern first used by Keys et al. (2004), based on data from Yuan (1993) was used. The pattern
3 specifies a fraction of percent of total daily ingestion consumed in each half-hour interval. The
4 first interval was shifted to correspond to the beginning of the active (dark) period, for
5 consistency with patterns used for humans and rats. A Table function was used in acslXtreme to
6 interpolate an instantaneous rate between the measured (30-min) values, with normalization so
7 that the 24-hour integral equals 100%. The daily pattern is shown in Figure B-9A and the
8 resulting blood concentration for a mouse exposed for 6 days per week (2100 mg/kg) is shown in
9 Figure B-9B.
March 2011
B-18
DRAFT—DO NOT CITE OR QUOTE
-------
A
8 10 12 14 16 18 20 22 24
Time (h)
^ 1,500-!
en
E 1,200-
C
o
'-I—1
ID
-P
C
0)
u
c
o
u
TJ
O
_0
CD
900-
600-
300-
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.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
March 2011
B-19
DRAFT—DO NOT CITE OR QUOTE
-------
1 (Figure B-10) by fitting Michaelis-Menten constants for one high affinity/low capacity and one
2 low-affinity, high-capacity enzyme, using the optimization routines in acslXtreme v2.3. Also
3 shown for comparison in Figure B-10A are the 100 mg/kg IV data of Horton et al. (1992),
4 obtained using Fischer 344 (F344) rats (data extracted from figures using Digitizlt), with a model
5 simulation (heavy red line) which differs from that for the SD rat only due to the predicted effect
6 of know body weight differences. While the fit to the Ward et al. (1997) data for SD rats is
7 excellent, especially for the lower dose, the rate of clearance (disappearance from the blood,
8 mostly due to metabolism) is over-predicted for the F344 rat when parameters fit to SD rat data
9 are used. The 100 mg/kg IV data, with an alternate simulation for the F344 rat obtained with
10 distinct parameters (see below) is expanded in Figure B-10B, emphasizing the difference in
11 clearance between the two strains.
12 We then attempted to fit the model to the inhalation data of Horton et al. (1992) by
13 adjusting only the inhalation fractional uptake (FRACIN). The results, shown in Figure B-l 1 A,
14 are clearly poor. While the model does match the uptake portion of the inhalation data for the
15 1200 and 2000 ppm exposures, it under-predicts the peak concentration reached at 200 ppm.
16 Further, the post-exposure clearance predicted by the model is much more rapid than indicated by
17 the data, as occurred with the IV kinetics (Figure B-10). (Since the peak concentration for the
18 2000 ppm inhalation exposure actually occurred at 7 hr, we also simulated a 7-hr exposure,
19 shown by the thin black line. The result indicates that the data are more consistent with and
20 better predicted by the longer exposure duration, but clearance is still over-predicted post-
21 exposure.) Therefore we concluded that the data show a clear strain difference in metabolism,
22 and should support at least a partially independent set of parameters for SD and F344 rats.
23 We then combined the 100 mg/kg IV and inhalation data of Horton et al. (1992) (for F344
24 rats) and attempted to simultaneously identify the four metabolic parameters (Vmax and Km for
25 two pathways) and FRACIN. However when this was done the resulting values for the two Km's
26 were ~ 90 ± 50 mg/L and 70 ± 40 mg/L (Km and Km2, respectively), which are clearly
27 indistinguishable from a statistical standpoint. If instead the Km's were fixed at the more distinct
28 values identified from the SD rat IV data (6.3 and 65 mg/L), the optimization routine tended to
29 set the Vmax associated with the lower Km to zero. Thus the F344 rat data of Horton et al.
30 (1992) appear to be most consistent with a single metabolic pathway, even though the observed
31 concentrations spanned almost 2 orders of magnitude. Therefore those data (including the 100
32 mg/kg IV data) were simultaneously fit by adjusting a single Vmax and Km, along with the
33 inhalation fraction, FRACIN, with the second metabolic pathway set to zero (Figures B-l OB and
34 B-11B).
March 2011 B-20 DRAFT—DO NOT CITE OR QUOTE
-------
10,000
1,000
o 100
(U
—
o
_Q 10
CD
— 2500 mg/kg SD sim
— 100 mg/kg SD rat sim
— 100 mg/kg F344 rat sim
• 2500 mg/kg SD rat IV
n 100 mg/kg SD rat IV
+ 100 mg/kg F344 rat IV
10 15 20 25
Time (hr)
30
35
40
45
100
I
O
£ 10
O
_o
CD
B
100 mg/kg SD rat sim
100 mg/kg F344 rat sim
100 mg/kg SD rat IV
100 mg/kg F344 rat IV
3 4
Time (hr)
Figure B-10. NP rat i.v.-route methanol blood kinetics. MeOH was infused
into: female Sprague-Dawley rats (275 g) at target doses of 100 (open squares
and thin black line) or 2,500 (filled squares and heavy blue line) mg/kg; or
(filled diamonds and heavy red lines) male F-344 rats (220 g) at target doses
of 100 mg/kg. Data points represent measured blood concentrations and
lines represent PBPK model simulations with (A) metabolic parameters fit to
the Sprague-Dawley rat data or (B) metabolic parameters fit to F-344 data
(see text for further details).
Source: Ward et al. (1997: squares'): Horton et al. (1992: diamonds).
March 2011
B-21
DRAFT—DO NOT CITE OR QUOTE
-------
100
01
E
—
I
o
v
10
0.1
\ \
— 1200 ppm simulation
— 2000 ppm simulation ,_,
— 200 ppm simulation P 2000 ppm data
* 1200 ppm data A 200 ppm data
— 2000 ppm, 7-hr simulation
8
Time (hr)
10
A
12
14
16
100
10
x
o
01
0.1
— 2000 ppm simulation
— 200 ppm simulation
* 1200 ppm data
— 2000 ppm, 7-hr simulation
~v
— 1200 ppm simulation
n 2000 ppm data
A 200 ppm data
8
Time (hr)
10
12
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 obtained 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)
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
March 2011
B-22
DRAFT—DO NOT CITE OR QUOTE
-------
1 simultaneously fit the inhalation data for all exposure levels, although a wide range of metabolic
2 saturation (Km) values were tested. We could obtain a better fit of the high-concentration data by
3 constraining FRACIN to a higher value, for example, but then the fits to the lower concentration
4 data were compromised (not shown). Examining Figure 2 of Horton et al. (1992), the
5 experimental variability (indicated by the error bars) on the 2000 ppm data was much larger than
6 the 200 or 1200 ppm data, and as indicated by the simulations in Figure B-ll here, there is at
7 least the appearance that the exposure was actually for 7 hr instead of 6 hr. (To be clear, the 2000
8 ppm data were used in the optimization with the duration of inhalation set to 6 hr, but the routine
9 selected parameters which only poorly fit those data.) Since our greatest concern is in predicting
10 dosimetry at lower exposure levels, near to the points of departure, we decided to retain the fits
11 shown here. The corresponding parameters are listed in Table B-l. The fractional absorption
12 (20%) was lower than that estimated for mice (66.5%), but Perkins et al. (1995a) also found
13 lower fractional absorption of inhaled methanol in rats vs. mice.
14 Finally, first-order oral absorption parameters were first fit to the lower dose (100 mg/kg)
15 oral absorption data reported by Ward et al. (1997), using the optimization routines in acslXtreme
16 v2.3 (Figure B-12, heavy/solid lines). (Since the animals used were SD rats, the SD-specific
17 metabolic parameters were used.) While the fit to the low-concentration data was quite good
18 (Figure B-12, lower panel), the fit to the 2500 mg/kg data (Figure B-12, upper panel) exhibited a
19 much faster and higher peak than shown by the data. Even when the model was fit to both the
20 high- and low-concentration data simultaneously, the fit to the high-concentration data could not
21 be significantly improved without completely degrading the low-concentration fit (not shown).
22 Also note that the 2500 mg/kg linear simulation completely over-estimates all the data points;
23 i.e., the area-under-the-curve for this dose is higher than indicated by the data, indicating that the
24 assumption of 100% absorption is not valid. Therefore, an alternative model using a saturable
25 (Michaelis-Menten) equation for absorption from the stomach and fecal elimination (linear term)
26 from the intestine was considered (thin lines) and found to significantly improve the high-
27 concentration simulation, with a nearly identical fit the low-concentration data. While methanol
28 absorption is not known to be regulated by transporters or other processes that would give rise to
29 rate saturation, it is clear form the discrepancy between the linear model and the 2500 mg/kg data
30 that uptake is slower than predicted by such a model and its use would lead to an over-prediction
31 of internal concentrations. Therefore parameters for the saturable uptake model are reported in
32 Table B-l and the KMASC applied to mice and humans. Note that since the saturation constant
33 corresponds to a fairly large dose (620 mg/kg), the model is still effectively linear at low- to
34 moderate dose rates.
March 2011 B-23 DRAFT—DO NOT CITE OR QUOTE
-------
2,700-
^ 2,400-
^ 2,100
E 1,800-
31 1,500-
« 1,200 -
^ 900-
600 -
300
0
2500 mg/kg linear uptake simulation
2500 mg/kg saturable uptake simulation
2500 rng/kg data
100 mg/kg linear uptake simulation
100 rng/kg saturable uptake simulation
100 rng/kg data
12
18
Time (hr)
24
30
36
Qi
E
'••—•'
O
Qj
100
80-1
60
40
0
— 100 mg/kg linear uptake simulation
100 rng/kg saturable uptake simulation
n 100 mg/kg data
Time (hr)
Figure B-12. Model fits to data sets from oral exposures to 100 (squares) or
2,500 (diamonds) mg/kg MeOH in female Sprague-Dawley rat (Expanded
scale in lower panel). Symbols are concentrations obtained from the
command file. The thick lines represent PBPK model fits using a linear (first-
order) equation for absorption from the stomach compartment with no fecal
elimination, while the thin lines use a Michaelis-Menton equation with a
small fraction eliminated in the feces. All other GI rates, including
absorption from the intestine, are first order.
Source: Ward et al.Q997).
March 2011
B-24
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 ppm MeOH are 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 time
24 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 continuous 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.
March 2011 B-25 DRAFT—DO NOT CITE OR QUOTE
-------
CD
12 24 36 48 60
Time (hr)
72
84
96
Exposure
concentration
(ppm)
500
1,000
2000
*~rnax
(mg/L)
3.6
10.6
48.5
AUC
(hrmg/L)
79
227
968
AMET
(mgEq)
17.6
34.8
67.2
1
2
3
4
Figure B-13. Simulated Sprague-Dawley rat inhalation exposures to 500,
1,000, or 2,000 ppm MeOH. Rat BW was set to 0.33 kg. Simulations are
shown for both continuous (thin, dashed/dotted lines in plot) and 22
hours/day exposures (thick, solid lines in plot). Cmax, AUC, and amount
metabolized (AMET) are determined from the 22 hour/day simulations, run
for a total of 5 days (120 hours), with the AUC and AMET calculated for the
last 24 hours of the simulation.
Figure B-14 depicts simulations run to mimic a single oral exposure, treated as a
continuous infusion for 12 hours (assuming 12-hour period when rats are awake and active).
Total AUC and AMET and AUC24 and AMET24 for the first 24 hours after start of exposure
were calculated.
March 2011
B-26
DRAFT—DO NOT CITE OR QUOTE
-------
I
o
0)
o
_0
CD
10,000-=
1,000-
100.
10-
1-
0.1-
0.01,
0.001
40
Time (hr)
60
80
Exposure Body Cmax AUC AUC24 AMET AMET24
dose weight (kg) (mg/L) (hr-mg/L) (hrmg/L) (mgEq) (mgEq)
(mg/kg-day)
66.0
624.1
177
0.33
0.33
0.34
9.3
631.6
2,832.4
104.8
9,525
72,617
104.8
8,817
45,138
21.2
155.6
347.4
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.)
1 To simulate ingestion of methanol in drinking water by rats under bioassay conditions, an
2 ingestion pattern based on the observations of Spiteri (1982) and Peng et al. (1990). While mice
3 ingest water in frequent, small bouts (Gannon et al., 1992) that are reasonably described as a
4 continuous delivery to the stomach, rats exhibit clear periods of ingestion alternating with
5 periods where no ingestion occurs (Peng et al., 1990; Spiteri, 1982). Based on those data a
6 reasonable representation of rat water ingestion can be described as serious of pulses. During the
7 dark/active period of each day (first 12 hr) each bout of drinking was assumed to last 45 min
8 followed by 45 min without ingestion (total of 8 bouts). During the light/inactive period (next 12
9 hr) drinking bouts were assumed to last only 30 min followed by 2.5 hr (150 min) without
10 drinking (4 bouts). An equal amount was assumed to be consumed in each bout within the dark
March 2011
B-27
DRAPT—DO NOT CITE OR QUOTE
-------
1 period, likewise within each light-period bout, with the respective amounts adjusted such that
2 80% of the total ingestion occurs during the dark and 20% during the light (Burwell et al., 1992).
3 The resulting absorption pattern is shown in Figure B-15A and a simulated blood concentration
4 time-curve (for 50 mg/kg/day dosing) is shown in Figure B-15B.
^ 0.14-1
-------
B.2.8. Human Model Calibration
B.2.8.1. Inhalation Route
1 The mouse model was scaled to human body weight (70 kg or study-specific average),
2 using human tissue compartment volumes and blood flows, and calibrated to fit the human
3 inhalation-exposure data available from the open literature, which comprised data from four
4 publications (Batterman et al., 1998; Ernstgard et al., 2005; Osterloh et al., 1996; Sedivec et al.,
5 1981).
6 A first-order rate of loss of MeOH from the blood, K1C, and a first-order bladder
7 compartment time constant, KBL, were used to provide an estimate of urinary MeOH
8 elimination. The inhalation-route urinary MeOH kinetic data described by Sedivec et al. (1981)
9 (Figure B-16) were used to inform these parameters. The urine MeOH concentration data
10 reported by the authors were converted to amount in urine by assuming 0.5 mL/hr/kg total
11 urinary output (Horton et al., 1992). Sedivec et al. (1981) measured a fractional uptake of 57.7%,
12 based on total amount inhaled. Since the PBPK model uses alveolar rather than total ventilation
13 and this is typically assumed to be 2/3 of total ventilation the fractional uptake of Sedivec et al.
14 (1981) was corrected by dividing by 2/3 to obtain a value for FRACIN of 0.8655. The resulting
15 values of K1C and KBL, shown in Table B-l, differ somewhat depending on whether first-order
16 or saturable liver metabolism is used. These are only calibrated against a small data set and
17 should be considered an estimate. Urine is a minor route of MeOH clearance with little impact
18 on blood MeOH kinetics.
19 Although the high-doses used in the mouse studies warrant the use of a second metabolic
20 pathway with a high Km, the human exposure data all represent lower concentrations and may
21 not require or allow for accurate calibration of a second metabolic pathway. Horton et al. (1992)
22 employed two sets of metabolic rate constants to describe human MeOH disposition, similar to
23 the description used for rats and mice, but in vitro studies using monkey tissues with non-MeOH
24 substrates were used as justification for this approach. Although Bouchard et al. (2001)
25 described their metabolism using Michaelis-Menten metabolism, Starr and Festa (2003) reduced
26 that to an effective first-order equation and showed adequate fits. Perkins et al. (1995a)
27 estimated a Km of 320 + 1273 mg/L (mean + S.E.) by fitting a one-compartment model to data
28 from a single oral poisoning to an estimated dose. In addition to the extremely high standard
29 error, the large standard error for the associated Vmax (93 + 87 mg/kg/hr) indicates that the set
30 of Michaelis-Menten constants was not uniquely identifiable using this data. Other Michaelis-
31 Menten constants that have been used to describe MeOH metabolism in various models for
32 primates are given in Table B-2. Because the Km calculated by Perkins et al. (1995a) from the
33 high-dose oral exposure is 320 mg/L, while the highest observed concentration in the data sets
March 2011 B-29 DRAFT—DO NOT CITE OR QUOTE
-------
1 considered here is 14 mg/L (Batterman et al., 1998), forcing the model to use this higher Km
2 would simply result in fits that are effectively indistinguishable from the linear model. A simple,
3 linear model is preferred over the use of a Km value that high.
10 -
9 -
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
24
3 i
2.5 -
Sedivec et al.,
total urinary excretion
20
24
Figure B-16. Urinary MeOH elimination concentration (upper panel) and
cumulative amount (lower panel), following inhalation exposures to MeOH in
human volunteers. Data points in lower panel represent estimated total
urinary MeOH elimination from humans exposed to 78 (diamonds), 157
(triangles), and 231 (circles) ppm MeOH for 8 hours, and lines represent
PBPK model simulations. Solid lines are model results with the saturable
equation for hepatic metabolism while dashed lines show results for liner
metabolism. Data digitized from Sedivec et al. (1981) and provided for
modeling by the EPA.
Source: Sedivec et al. (1981).
March 2011
B-30
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
Perkinds et al. (1995a)
Perkinds et al. Q995a)
Perkinds et al. Q995a)
Perkinds et al. (1995a)
Horton et al. (1992)
Fisher et al. (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-dependent
12 factor obtained by multiplying the slope of this regression (0.093 mg/L-hr) by the measurement
13 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 0.8655
16 (estimated from the results of Sedivec et al. as described above) and holding the ventilation rate
17 constant at 16.5 L/hr/kg0'75 and QPC at 24 L/hr/kg0'75 (values used by EPA [2000d] for modeling
18 the inhalation-route kinetics vinyl chloride). Other human-specific physiological parameters
19 were set as reported in Table B-l.
March 2011
B-31
DRAFT—DO NOT CITE OR QUOTE
-------
1.8
1.6
J 1.4
ro
£.1.2
I 1
2 0.8
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-16 and B-18). 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 y2 distribution with r degrees of freedom:
March 2011
B-32
DRAFT—DO NOT CITE OR QUOTE
-------
1 -2[log(/l(modell)//l(model2))] =-2[log/l(modell)-Iog/l(model2)] = X2r
2 The likelihood ratio test states that if twice the difference between the maximum LLF of
3 the two different descriptions of metabolism is greater than the %2 distribution, then the model fit
4 has been improved (Collins et al., 1999; Devore, 1995; 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 Michaelis-Menten kinetics to describe the metabolism of
MeOH in humans. The rate constants used for each simulation are given in
Table B-3.
Source: Batterman et al. (1998: top): Osterloh et al. (1996: bottom).
March 2011
B-33
DRAFT—DO NOT CITE OR QUOTE
-------
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
v r
v max^-"
First Order
KLLC
Optimized value
23.8
33.2
95.7
S.D.
8.8
10.1
5.4
Correlation matrix
-0.994
NA
LLF
-24.1
-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 BMDL10 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 (log*,) for M-M
-24.1
LLF (log*,) for 1st
order
-31.0
LLF
1st versus M-Ma
34.1
X2r (99% confidence) b
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.
March 2011
B-34
DRAFT—DO NOT CITE OR QUOTE
-------
1 While the use of Michaelis-Menten kinetics might allow predictions across a wide
2 exposure range (into the nonlinear region), extrapolation above 1,000 ppm is not suggested since
3 the highest human exposure data are for 800 ppm. Extrapolations to higher concentrations are
4 potentially misleading since the nonlinearity in the exposure-internal-dose relationship for
5 humans is uncertain above this point. The use of a BMD or internally applied UFs should place
6 the exposure concentrations within the range of the model.
7 The data from Ernstgard et al. (2005) was used to assess the use of the first-order
8 metabolic rate constant to a dataset collected under conditions of light work. Historical measures
9 of QPC (52.6 L/hr/kgO.75) and QCC (26 L/hr/kgO.75) for individuals exposed under conditions
10 of 50 w of work from that laboratory (52.6 L/hr/kgO.7) (Corlev et al.. 1994: Ernstgard. 2005:
11 Johanson et al., 1986) were used for the 2-hour exposure period (Figure B-19). Otherwise, there
12 were no changes in the model parameters (no fitting to these data). The results are remarkably
13 good, given the lack of parameter adjustment to data collected in a different laboratory, using
14 different human subjects than those to which the model was calibrated.
Ernstgard et al. (2005)
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 (Johanson et al., 1986), and a QCC of 26
L/hr/kgO.75 (Corlev et al., 1994) were used by the model.
Source: Ernstgard etal..(2005)
March 2011
B-35
DRAFT—DO NOT CITE OR QUOTE
-------
B.2.8.2. Oral Route
1 There were no human data available for calibration or validation of the oral route for the
2 human model. In the absence of data to estimate rate constants for oral uptake, the 'humanset.m'
3 file which sets parameters for human simulations applies the KMAS for the mouse with the other
4 absorption parameters set to match those identified for ethanol in humans by Sultatos et al.
5 (2004): VmASC was set such that for a 70-kg person, VMAS/KMAS matches the first-order
6 uptake constant of Sultatos et al. (2004) (0.21 hr-1). While Sultatos et al. (2004) include a term
7 for ethanol metabolism in the stomach, no such term is included here and the rate of fecal
8 elimination is set to zero, corresponding to 100% absorption. However zero-order ingestion, a
9 continuous infusion at a constant rate into the stomach lumen equal to the daily dose/24 hours,
10 was assumed for all human simulations. Since absorption was assumed to be 100% of
11 administered MeOH, at steady state the rate of uptake from the stomach and intestine
12 compartments (combined) must equal the rate of infusion to the stomach. Since Cmax is driven
13 by the oral absorption rate, which was assumed rather than fitted and verified, Cmax was not
14 used as a dose metric for human oral route simulations. AUC, which is less dependent on rate of
15 uptake, was used as the dose metric and for estimation of HEDs. Since the AUC was computed
16 for a continuous oral exposure, its value is just 24-hours times the steady-state blood
17 concentration at a given oral uptake rate.
B.2.8.3. Inhalation Route HECs and Oral Route HEDs
18 The atmospheric MeOH concentration resulting in a human daily average blood MeOH
19 AUC (hrxmg/L) or Cmax (mg/L) equal to that occurring in experimental animals following
20 exposure at the POD concentration is termed the HEC. Similarly, the oral dose (rate) resulting in
21 human daily average blood MeOH AUC (hrxmg/L) equivalent to that occurring in an
22 experimental animal at the POD concentration is termed the HED.
23 To determine the HEC for specific exposures in mice, the mouse PBPK model is first
24 used to determine the daily blood MeOH 24-hours AUC and Cmax associated with 7 hour/day
25 inhalation exposures. Mice were exposed each day for 10 days, so the full 10-day exposure was
26 simulated and an average 24-hours AUC calculated over that time, so no other duration
27 adjustment was needed. The human AUC was determined for the last 24 hours of a continuous
28 1,000-hour exposure, to assure steady state was achieved. The human Cmax was determined at
29 steady state and so is equivalent to the steady state blood MeOH concentration. Results are given
30 in Table B-5 and for inhalation shown in Figure B-20.
31 For example, for a 1,000 ppm exposure this resulted in model-predicted peak blood of
32 133 mg/L and an AUC of 770 (hrxmg/L). The human model can then be used to determine the
33 human MeOH exposure concentration leading to the same daily average AUC or Cmax under
March 2011 B-36 DRAFT—DO NOT CITE OR QUOTE
-------
1 continuous exposure conditions. Based on AUC, the HEC of the 1,000-ppm exposure is
2 684 ppm, while based on peak (human steady-state) concentration, the HEC is predicted to be
3 1110 ppm. The parameters used in the human model for these simulations are listed in Table B-l
4 for saturable kinetics.
5 The HED was calculated by using the human model to find the oral dose (mg/kg-day) that
6 gave a blood MeOH AUC equivalent to the mouse AUC following an exposure at the POD.
7 Zero-order absorption was assumed. For example, the human oral exposure equivalent to a
8 1,000-ppm inhalation exposure in mice (i.e., with an AUC of 770 mg-hr/L) is 165 mg/kg-day.
9 Since a 200 mg/kg-day oral exposure gave a human AUC of 1,284 mg-hr/L, which falls between
10 the values predicted for inhalation exposures at 800 ppm (1,090 mg-hr/L) and 1,000 ppm (2,090
11 mg-hr/L), this oral exposure rate was taken to be the upper end for the model to accurately
12 estimate an HED.
Table B-5. PBPK model predicted Cr
humans exposed to MeOH
and 24-hour AUC for mice and
Exposure
concentration
(ppm)
1
10
50
100
250
500
1,000
2,000
5,000
10,000
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/kg-day (oral), but are
included for comparison
13 Again, since the available human exposure data is to, at most, 800 ppm, the model could
14 not be calibrated for higher exposures that approximate most of the mouse and rat exposure
15 concentrations. The AUC in humans for similar exposure levels is ~3 times greater than in the
March 2011
B-37
DRAFT—DO NOT CITE OR QUOTE
-------
mouse, primarily because human exposure estimates are expected to result from 24-hour
exposures and the mice were exposed for 7 hours.
1000000
100000 -
10000 -
1000 -
100 -
10 -
1 -
0.1
•Mouse
Human
Human-linear
1 10 100 1000 10000
Concentration (ppm)
10000
10 100 1000
Concentration (ppm)
10000
10000
Human
Human-linear
10 100 1000
Concentration (ppm)
10000
J
4
5
6
7
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°7S to reflect cross-species
scaling (Human simulation results above 1,000 ppm are not considered
reliable but are shown for comparison).
While the PBPK computational code can be used in the future to derive HECs or HEDs
for other exposures, an alternative approach was developed that allows non-PBPK model users to
estimate MeOH HECs and HEDs from benchmark doses in the form of AUCs. This approach
uses algebraic equations describing the relationship between predicted MeOH 24-hour AUC or
scaled daily metabolic rate in the liver (MET, mg/kg°'75-day) (constant 24-hour exposure) and the
March 2011
B-38
DRAFT—DO NOT CITE OR QUOTE
-------
1 inhalation exposure level (i.e., an HEC in ppm) (Equations 1, Ib, 3 or 3b) or oral exposure rate
2 (i.e., an HED in mg/kg-day) (Equations 2, 2b, 4 or 4b). To use the equations to derive an HEC or
3 HED, the target human AUC is applied to the appropriate equation. Since these relationships are
4 for continuous exposures, blood concentration is constant, and hence extrapolation for a Cmax is
5 obtained by simply using AUC = 24*Cmax.
6
7 HEC(ppm<}ooo) = 0.02525 xAUC + 1290x U Equation 1
\w (rng/kg-day<20o) = 0.3448 xMET + 4286xMET Equation 4
v B B KMUJ 860.0-MET
11
12 Once the HEC or HED is calculated from the appropriate equation above (depending on which internal metric is
13 being used), the RfC or RfD is then just calculated by dividing with the extrapolation uncertainty factor (UF).
14
15 RfC (ppm) = HEC (ppm)/UF Equations
16 RfD (mg/kg-day) = HED (mg/kg-day)/UF Equation 6
B.2.9. Conclusions and Discussion
17 Mouse, rat, and human MeOH PBPK models have been developed and calibrated to data
18 in the open literature. The EPA chose to develop its own model because none of the existing
19 models satisfactorily fulfilled all of the criteria specified in Section 3.4.1 of Chapter 3. Further,
20 none of the existing models had been calibrated or tested against the larger collection of data
21 considered for each species here. As a result, while each model may fit the subset of the data to
22 which it had been calibrated better than the final model described here, without adjustment of
23 parameters from those published, each model either had features which made it incompatible
24 with risk extrapolation (e.g., parameters which vary with dose in an unpredictable way) or had an
25 inadequate fit to other data considered critical for establishing overall model soundness. The
26 EPA model simplifies the structure used by Ward et al. (1997) in some aspects while adding
27 specific refinements (e.g., a standard lung compartment and a two-compartment GI tract).
28 Although the developmental endpoints of concern are effects which occur during in utero
29 and (to a lesser extent) lactational exposure, it is not necessary for a MeOH PBPK model to
30 specifically describe pregnancy (i.e., specify a fetal/gestational/conceptus compartment) and
31 lactation in order for it to provide better cross-species extrapolation of risk than default methods.
32 Representation of the unique physiology of pregnancy and the fetus/conceptus would be
March 2011 B-39 DRAFT—DO NOT CITE OR QUOTE
-------
1 necessary if MeOH pharmacokinetics differed significantly during pregnancy or if the observed
2 partitioning of MeOH into the fetus/conceptus versus the mother showed a concentration ratio
3 significantly greater than or less than 1. MeOH pharmacokinetics GD6-GD10 in the mouse, are
4 not different from NP mice (Pollack & Brouwer, 1996), and the maternal blood:fetus/conceptus
5 partition coefficient is reported to be near 1 (Horton et al., 1992; Ward et al., 1997). At GDIS in
6 the mouse, maternal blood levels are only modestly different from those in NP animals (see
7 Figures B-4 and B-5 for examples), and in general the PBPK model simulations for the NP
8 animal match the pregnancy data as well as the NP data. Likewise maternal blood kinetics in
9 monkeys differs little from those in NP animals (see Section 3.4.7). Further, in both mice and
10 monkeys, to the extent that late-pregnancy blood levels differ from NP for a given exposure, they
11 are higher; i.e., the difference between model predictions and actual concentrations is in the same
12 direction. These data support the assumption that the ratio of actual target-tissue methanol
13 concentration to (predicted) NP maternal blood concentrations will be about the same across
14 species, and hence that using NP maternal blood levels in place of fetal concentrations will not
15 lead to a systematic error when extrapolating risks. Thus, a full representation of pregnancy and
16 the fetal/conceptus compartment appears to be unnecessary.
17 While lactational exposure is less direct than fetal exposure and blood or target-tissue
18 levels in the breast-feeding infant or pup are likely to differ more from maternal levels, the
19 health-effects data indicate that most of the effects of concern are due to fetal exposure, with only
20 a small influence due to postbirth exposures. Separating out the contribution of postbirth
21 exposure from pre-birth exposure to a given endpoint in a way that would allow the risk to be
22 estimated from estimates of both exposure levels would be extremely difficult, even if one had a
23 lactation/child PBPK model that allowed for prediction of blood (or target-tissue) levels in the
24 offspring. And one would still expect the target-tissue concentrations in the offspring to be
25 closely related to maternal blood levels (which depend on ambient exposure and determine the
26 amount delivered through breast milk), with the relationship between maternal levels and those
27 in the offspring being similar across species.
28 Therefore, the development of a lactation/child PBPK model appears not to be supported,
29 given the minimal change that is likely to result in risk extrapolations and use of (NP) maternal
30 blood levels as a measure of risk in the offspring is still considered preferable over use of default
31 extrapolation methods. In particular, the existing human data allow for accurate predictions of
32 maternal blood levels, which depend strongly on the rate of maternal methanol clearance. Failing
33 to use the existing data (via PBPK modeling) for human methanol clearance (versus that in other
34 species) would be to ignore this very important determinant of exposure to breast-fed infants.
35 And since bottle-fed infants do not receive methanol from their mothers, they are expected to
36 have lower or, at most, similar overall exposures for a given ambient concentration than the
March 2011 B-40 DRAFT—DO NOT CITE OR QUOTE
-------
1 breast-fed infant, so that use of maternal blood levels for risk estimation should also be
2 adequately protective for that group.
3 During model development, several inconsistencies between experimental blood MeOH
4 kinetic data embedded in the Ward et al. (1995) model and the published figures first reporting
5 these data were discovered. Therefore, data were digitized from the published literature when a
6 figure was available, and the digitized data was compared to the provided data. When the
7 digitized data and the data embedded in the computational files (i.e., provided to Battelle under
8 contract from the EPA) were within 3% of each other, the provided data was used; when the
9 difference was greater than 3%, the digitized data was used. Often, using the published figures as
10 a data source resulted in substantial improvements of the fit to the data in the cases where the
11 published figures were different from the embedded data.
12 The final MeOH PBPK model fits well inhalation-route blood kinetic data from separate
13 laboratories in rodents and humans. Intravenous-route blood MeOH kinetic data in NP mice
14 were only available for a single i.v. dose of 2,500 mg/kg, but were available for GDIS mice
15 following administration of a broader range of doses: 100, 500, and 2,500 mg/kg. Up to 20 hours
16 postexposure, blood MeOH kinetics appear similar for NP and pregnant mice after administration
17 of 2,500 mg/kg. The intravenous pharmacokinetic data in GDIS mice showed an unexpected
18 dose-dependent nonlinearity in initial blood concentrations, suggesting either a dose dependence
19 on the volume distribution, which is unlikely, or some source of experimental variability. To
20 account for this nonlinearity, Ward et al. (1997) used dose-specific partition coefficients for
21 placenta and embryonic fluid and Vmax for the metabolism of MeOH. The current model uses a
22 consistent set of parameters that are not varied by dose and therefore does not fit these 100 mg/kg
23 dose intravenous data. The model does fit the 500 and 2,500 mg/kg doses, and if a presumed i.v.
24 dose of 200 mg/kg (twice the reported 100 mg/kg) is employed, is able to predict initial blood
25 concentrations for the lowest dose data, as expected. The i.v. data from the Ward et al. (1995)
26 model does match the corresponding published figures.
27 The model fits to the mouse oral-route MeOH kinetic data using a consistent set of
28 parameters (Figure B-4) are reasonably good but not as good as fits to the inhalation data. The
29 model consistently underpredicts the amount of blood MeOH reported in two studies (Ward et
30 al., 1997; 1995). Ward et al. (1997) utilized a different Vmax for each oral absorption data set. In
31 the report by Ward et al. (1997) the GD18 and the GD8 data from Dorman et al. (1995) were both
32 fit using a Vmax of-80 mg/kg/hr (body weights were not listed, the model assumed that GD8
33 and GDIS mice were both 30 g; Ward et al. (1997) did not scaled by body weight), but lower
34 partition coefficients for placenta (1.63 versus 3.28) and embryonic fluid (0.0037 versus 0.77).
35 The current model adequately fits the oral pharmacokinetic data using a single set of parameters
36 that is not varied by dose or source of data.
March 2011 B-41 DRAFT—DO NOT CITE OR QUOTE
-------
1 The fits of the rat model to the limited dataset readily available were quite good. The low-
2 dose exposures of all routes were emphasized in model optimization since they were the doses
3 most relevant to risk assessment. Based on a rat inhalation exposure to 500 ppm, the human HEC
4 would be 300 ppm (by applying an AUC of 226 [Figure B-12] to Equation 1).
5 The mouse, rat, and human models fit multiple datasets from multiple research groups
6 using consistent parameters that are representative of each species, but are not varied within
7 species. Using the model, it will be possible to ascertain chronic human exposure concentrations
8 that are likely to be without an appreciable risk of deleterious effects.
B.3. ADDITIONAL MATERIAL
9 • Results from Optimizations
10 • acslXtreme Program (.csl) File (Electronic Attachment)
11 • acslXtreme procedure (.cmd) file
12 • Key to .m files for reproducing the results in this report
13 • Code for .m files
14 • Personal communication from Lena Ernstgard regarding human exposures reported in the
15 Ernstgard and Johanson, 2005 SOT poster
16 • Personal communication from Dr. Rogers regarding mouse exposures to MeOH
17 • Data and simulations for MeOH Metabolism/Total Metabolites Produced
18 • 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
19 The approach and results are presented below in outline format with supporting figures.
20 More complete documentation was not developed because the products of the optimizations were
21 not used in the final model. The documentation here is intended only to demonstrate that
22 appropriate optimizations were conducted and what the results of those optimizations were.
23 1. The Vmax for the low affinity pathway was set to 0 and the remaining VmaxC, Km, and
24 fractional availability were optimized using inhalation data only.
25 a. The optimizer was unable to find a value for Km that was greater than 0.
26 b. The resulting metabolic parameters essentially represented a zero order loss
27 process.
28 2. The Vmax for the low affinity pathway was set to 0 and the remaining VmaxC and Km
29 were optimized using all (oral, intravenous and inhalation) data.
30 a. The optimized single Km, 135 mg/L, was equal to the average of the 2 original
31 Kms.
March 2011 B-42 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
b. Fits to the MeOH blood levels following inhalation exposures > 2,000 ppm are
slightly improved, but the model fits to the 1,000 ppm exposure concentration
overpredict reported values by 20%.
Parameters for both metabolic pathways were optimized using all (oral, intravenous
and inhalation) data.
a. The fit to the high-dose intravenous data from Ward et al. (1997) (2,500 mg/kg)
was improved (Figure B-21).
b. The fit to the high-dose oral data, also from Ward et al. (1997), (2,500 mg/kg) was
improved (Figure B-22).
c. The fit to the mid-dose i.v. data (500 mg/kg) dose was not as good as using the
visually fit parameters (Figure B-21)
d. The fit to the low-dose oral data (1,500 mg/kg) was not as good when the visually
fit parameters were used (Figure B-22). The low-dose data was from Dorman et
al. (1995).
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 metabolism and
uptake parameter optimizations. Solid blue lines - visually optimized;
dashed red lines - metabolic parameters (Km, Km2, VmaxC, VmaX2C)
optimized using all inhalation data sets; dash/dot green lines - metabolic
parameters optimized using all data sets (inhalation, oral, and intravenous).
Source: Ward et al. (1997): Dorman et al. (1995).
Figure B-23. Fit of the model to inhalation data using different metabolism
and uptake parameter optimizations. Dotted black lines - model optimized
fractional inhalation, solid blue lines - visually optimized; dashed red lines -
metabolic parameters (Km, Km2, VmaxC, Vmax2C) optimized using all
inhalation data sets; dash/dot green lines - metabolic parameters optimized
using all data sets (inhalation, oral, and intravenous).
Source: Rogers et al. (1997).
March 2011
B-44
DRAFT—DO NOT CITE OR QUOTE
-------
B.3.1.2. 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 for
10 some, but never all, data sets. This is to be expected when there is clearly significant variability
11 in the underlying data. Various data weighting schemes were included to improve overall
12 optimization while maintaining a good fit to the 1,000 ppm data. In the end, optimization offered
13 no significant improvement over the fractional uptake and metabolic parameter values obtained
14 by visual optimization, so these were retained in the final version of the model.
B.3.1.3. acslXtreme 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
Based on MeOH Model by Ward et al (1997) with these revisions:
TS Poet, P Hinderliterand J Teeguarden (2006),
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 stomach 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
7) Code for incorporating endogenous/background MeOH, where the "source"
is a continuous infusion term in the liver was added. A set of equations
in the initial section sets the infusion rate based on other model parameters
such that when the endogenous background blood level is set by CVBBG = X mg/L
(orVCVBBG can be adjusted in a parameter estimation when INCBG = 1, otherwise
set INCBG = 0), or DCVBBG = X mg/L (the later to be used when dosing and
data analysis are for radiolabled methanol), the appropriate infusion rate
March 2011 B-45 DRAFT—DO NOT CITE OR QUOTE
-------
initial concentration in all tissues is set. Alternately one can set the
and endogenous urine concentration RURO = Y mg/L and the initial calculations
will set the other parameters appropriately. Finally, setting RINCBG > 0
with INCBG = 1 gives a linear increase in the zero-order infusion with time.
Code added by Paul Schlosser, U.S. EPA, Dec. 2009
Version is final version used for simulations
***** MODEL UN ITS***
Concentration, mg/L
Mass of Chemical, mg
Volume, L
Flow, L/hr
Body Weight Kg
=72 Character Line=
1
2
3
4
5
6
7
8
9
10
11
12
13
14 INITIAL
15 ! Initialize some Variables before start
16 Integer IDS, MULTE
17 REAL DRT(6), DRP(6) Istore drink water times, percents in array!
18
19 CONSTANT BW = 0.030 ! Body weight (kg)
20 CONSTANT QPC=15. ! Alveolar ventilation (L/hr/kg**0.75)
21
22 ! Blood Flows (fraction of cardiac output)
23 CONSTANT QCC = 15.0 ! Cardiac output (L/hr/kg**0.75)
24 CONSTANT QFC = 0.05 ! Fat
25 CONSTANT QLC = 0.25 ! Liver
26 ! Blood flow to rest of body Calculated by Flow Balance!
27 QRC = 1.0-(QFC + QLC)
28 QC = QCC*BW**0.75
29 OP = QPC*BW**0.75
30
31 ! Tissue Volumes for mice (fraction of body weight)
32 CONSTANT VAC = 0.0123 ! Arterial blood
33 CONSTANT VFC = 0.07 ! Fat
34 CONSTANT VLC = 0.055 ! Liver
35 CONSTANT VLuC = 0.0073 ! Lung tissue
36 CONSTANT VVBC = 0.0368 ! Venous blood
37 VRC = 0.91 - (VAC+VFC+VLC+VLuC+VVBC)
38
39 ! Partition Coefficients (Mouse values from Ward et al. (1997) used as default)
40 CONSTANT PB = 1350 ! MeOH Blood:Air; Use Morton value!
41 CONSTANT PF = 0.08 ! MeOH Fat:Blood
42 CONSTANT PL = 1.1 ! MeOH Liver:Blood
43 CONSTANT PLU = 1.0 ! MeOH Lung:Blood, compartment for dosing only
44 CONSTANT PR = 0.8 ! MeOH Rest of body:Blood
45
46 ! Hepatic Metabolism of MeOH
47 CONSTANT KM = 45.0 ! mg/L
48 CONSTANT VMAXC = 15.0 ! mg/hr/BW**0.75
49 VMAX = VMAXC*BW**0.75 ! mg/hr
50 CONSTANT VMAX2C = 15.0! 2nd saturable pathway
51 VMAX2 = VMAX2C*BW**0.75
52 CONSTANT KM2= 45.0
53 CONSTANT KLLC = 0.0! First-order metabolism
54 ! Set VMAXC = VMAX2C = 0, when KLLC > 0
55 KLL = KLLC/BW**0.25
March 2011 B-46 DRAFT—DO NOT CITE OR QUOTE
-------
1
2 ! MeOH Clearance from Blood!
3 CONSTANT K1C = 0.01 ! First-order clearance, BW**0.25/hr
4 K1 = K1C/BW**0.25 ! Scaled blood elimination, hr-1
5 ! This lumped term was used in the WARD model an accounted for
6 ! renal elimination and "additional" non-hepatic metabolism of
7 ! MeOH associated only with high dose i.v. data.
8 ! A 1st-order term should not be used to represent two processes
9 ! with different dose-dependencies.
10 ! This has not been used for mouse data (set=0), but was used to
11 ! approximate human urinary data!
12
13 ! Bladder compartment added by Paul Schlosser, October 2008
14 CONSTANT KBL=0.0 ! Bladder constant, 1/hr
15
16 ! Fractional Absorption of MeOH
17 CONSTANT FRACin = 0.85 ! Inhalation, value from Perkins et al (1996)
18 CONSTANT KFEC = 0.0 ! Fecal elimination constant, 1/hr
19 ! KFEC determines oral bioavailability
20
21 ! Molecular Weight of MeOH
22 CONSTANT MWMe = 32.0 ! mol wt, g/mol!
23
24 ! Closed Chamber Parameters
25 CONSTANT VChC = 100.0 ! Volume of closed chamber (L)
26 CONSTANT Rats = 0.0 ! Number of rats in chamber
27 CONSTANT kLoss = 0.0 ! Chamber loss rate /hr
28 ! Set RATS = 0.0 and KLOSS = 0.0 for open chamber
29
30 ! Blood Flows (L/hr)
31 QF = QFC*QC ! Fat
32 QL = QLC*QC ! Liver
33 OR = QRC*QC ! Rest of Body
34
35 ! Tissue Volumes (mL)
36 VAB = VAC*BW! Arterial blood volume
37 VF = VFC*BW ! Fat
38 VL = VLC*BW ! Liver
39 VLu = VLuC*BW ! Lung
40 VR = VRC*BW ! Rest of the body
41 VVB = WBC*BW ! Venous blood
42 VBL = VAB + WB ! Total blood
43
44 i Timing commands !
45 CONSTANT TCHNG = 6.0 ! End of exposure!
46 CONSTANT TSTOP = 24.0 ! End of experiment/simulation!
47 CONSTANT POINTS = 1000.0 ! No. points for simulation output!
48 CONSTANT REST = 100000.0 ! End of work period for human exercise
49 CONSTANT WORK = 100000.0 ! Start of work period for human exercise
50 SCHEDULE DS1 .AT.REST ! Change from work to rest conditions
51 SCHEDULE DS2.AT.WORK ! Change from rest to work conditions
52 ! Human Rest/Work (changes in blood-flow fractions to fat/liver not currently usedO
53 CONSTANT QPCHR=15.0, QCCHR=15.0, QLCHR=0.25, QFCHR=0.05 ! Rest
54 CONSTANT QPCHW=52.0, QCCHW=26.0, QLCHW=0.16, QFCHW=0.06 ! Work
55
March 2011 B-47 DRAFT—DO NOT CITE OR QUOTE
-------
1 ! Simulation Control !
2 ! Exposure Conditions Based on User Defined Initial Amounts of
3 ! Chemical (mg)
4 CONSTANT CONCppm = 0.0 ! Air Concentration in ppm
5 VCh = VChC-(Rats*BW) ! Volume of Occupied Chamber
6 CONCmg = CONCppm*MWMe/24451 ! Convert ppm to mg/Liter!
7 ACHO = CONCmg*VCH ! Init Amt in Chamber, mg!
8
9 ! Background levels, added by Paul M. Schlosser, U.S. EPA, 12/8/09
10 ! CVBbg (CVBBG) is the constant to be set to the background blood
11 ! concentration when dosing is with *non*-radio-labeled methanol, so
12 ! exogenous and endogenous methanol are indistinguishable.
13 ! dCVBbg (DCVBBG) is the constant to be set to the background blood
14 ! concentration when dosing is with "radio-labeled* methanol.
15 CONSTANT CVBbg = 1.6 ! Value from Rogers et al (1993) for CD-1 mice
16 constant vCVBbg = 0.0 ! Value for use as an adjustable variable
17 constant RINCBG=0.113 ! Relative increase in background appearance of
18 ! methanol per hour (multiplies time T and time-zero appearance)
19 ! Paul M. Schlosser, U.S. EPA, 12/2009
20 constant INCBG=0.0 ! Set to 1.0 when fitting vCVBbg and RINCBG
21 constant dCVBbg = 0.0 ! "Cold" (not-radiolabeled) background, for 14C data
22 constant RURO = 0.0 ! Initial urine concentration; used to set CVBBG when > 0
23
24 IF(K1C.EQ.O.O)THEN
25 CVBG = CVBBG + incbg*vCVBbg
26 ELSE
27 CVBG = CVBBG + incbg*vCVBbg + RURO*BWA0.25*0.5e-3/(K1C*WBC)
28 ENDIF
29
30 ! Following are calculations of initial conditions given an endogenouse
31 ! background blood concentration, CVBBG or VCVBBG, or urine concentration, RURO.
32 CVLbg = ((QF+QR)*QP/(QP+QC*PB) + QL + K1*VVB)*cvbg/QL
33 RAObg = (VMAX/(KM + CVLbg) + VMAX2/(KM2 + CVLbg) + KLL)*CVLbg + ...
34 (QC*QP/(QP+QC*PB) + K1*VVB)*cvbg
35 CABO=QC*cvbg/(QC+QP/PB)
36 AABO=CABO*VAB
37 AFO=VF*CABO*PF
38 ALO=VL*CVLbg*PL
39 ALuO=VLu*CABO*PLu
40 ARO=VR*CABO*PR
41 AVBO=WB*cvbg
42 ABLO=K1*cvbg*WB/KBL
43
44 ! Following are calculations of initial conditions given an endogenouse
45 ! background blood concentration, DCVBBG, set > 0 (with CVBBG, etc. = 0)
46 ! when dosing is with radio-labelled MeOH. Currently does not allow one
47 ! to use the equivalent of fitted background (VCVBBG), time-dependent
48 ! (RINCBG), or urine concntration (RURO) to set the background.
49 dCVLbg = ((QF+QR)*QP/(QP+QC*PB) + QL + K1*WB)*dCVBbg/QL
50 d RAObg = (VMAX/(KM + dCVLbg) + VMAX2/(KM2 + dCVLbg) + KLL)*dCVLbg + (QC*QP/(QP+QC*PB) +
51 K1*WB)*dCVBbg
52 dCABO=QC*dCVBbg/(QC+QP/PB)
53 dAABO=dCABO*VAB
54 dAFO=VF*dCABO*PF
55 dALO=VL*dCVLbg*PL
March 2011 B-48 DRAFT—DO NOT CITE OR QUOTE
-------
1 dALuO=VLu*dCABO*PLu
2 dARO=VR*dCABO*PR
3 dAVBO=WB*dCVBbg
4
5 ! Oral dosing
6 CONSTANT KAS = 0.1 ! 1st order oral abs, hr-1
7 CONSTANT KMASC = 550 ! Saturable oral abs Kmasc [=] mg/kg
8 KMAS = KMASC*BW
9 CONSTANT VASC = 1740 ! Saturable oral ab VmaxC, mg/hr/kgA0.75
10 VAS = VASC*BW**0.75 ! Saturable oral ab Vmax, mg/hr
11 CONSTANT KAI = 0.1 ! 1st order oral abs from intestine, hr-1
12 CONSTANT KSI = 0.5 ! 1st order transfer stom to intes hr-1
13 CONSTANT DOSE = 0.0 ! Oral dose in mg/kg BW
14 CONSTANT ODS = 0.0 ! Switch for zero order oral uptake
15 ! (Set to 1 for zero order, set to 0 for first order)
16 ODOSE = DOSE*BW*(1.0-ODS) ! Convert mg/kg to mg total (oral)
17 RAOZ = DOSE*BW*ODS/24.0 ! mg/hr for zero order dosing
18
19 ! Daily dose for steady drinking water by "sipping" (by rats)
20 CONSTANT DWDOSE = 0 ! mg/kg/d by periodic sipping
21 CONSTANT PER1 =1.5 ! Period between sipping episodes (hr) during dark
22 ! "Between" means from the start of 1 to the start of the next episode
23 CONSTANT DUR1 = 0.75 ! Duration of sipping episodes during dark (hr)
24 CONSTANT PER2 = 3.0 ! Period during light (hr) between sipping episodes
25 CONSTANT DUR2 = 0.5 ! Duration of sipping episodes during light (hr)
26 CONSTANT FNIGHT = 0.8 ! Fraction of drinking during night
27 constant days = 7.0 ! days/week of oral exposure
28 constant metd = 7.0 ! number of days at end over which AUCBF and AMETF
29 ! are averaged
30 tmetf = metd*24.0
31 dayon=24.0*days
32 ! Night sipping rate (mg/h) during episodes
33 DWRNIGHT = DWDOSE*BW*FNIGHT*PER1/(12.0*DUR1)
34 ! Day sipping rate (mg/h) during episodes
35 DWRDAY= DWDOSE*BW*(1-FNIGHT)*PER2/(12.0*DUR2)
36 IDOSE=0
37 ! Above assumes 12-hr each for day/night
38
39 ! Drinking Table from Deborah Keys for mice, as used in
40 ! A quantitative description of suicide inhibition of dichloroacetic acid in rats and mice.
41 ! Keys, D. A., Schultz, I. R., Mahle, D. A., Fisher, J. W. (2004).
42 ! Based on data of Yuan, J. (1993). Modeling blood/plasma concentrations in dosed feed and dosed
43 ! drinking water toxicology studies. .
44 constant rdrink = 1.0! Default for use of sipping w/ DWDOSE abopve
45 ! set rdrink = 0.0 to use pattern below
46 table mdrinkp, 1,49/0., .5, 1, 1.5, 2., 2.5, 3., 3.5, &
47 4., 4.5, 5., 5.5, 6., 6.5, 7., 7.5, &
48 8., 8.5, 9., 9.5, 10., 10.5, 11, 11.5, &
49 12., 12.5, 13., 13.5, 14., 14.5, 15., 15.5, &
50 16., 16.5, 17., 17.5, 18., 18.5, 19., 19.5, &
51 20., 20.5, 21,21.5, 22., 22.5, 23., 23.5, 24.0, &
52 0.12 , 0.9, 1.6, 1.8, 1.9, 2.9, 4.0, 4.5, 4.9, 4.9, &
53 4.8,4.4,4.0,5.0,5.9,5.3,4.5,3.9, &
54 3.2, 3.0, 2.7, 2.5, 2.3, 2.3, 2.3, 1.9, &
55 1.4, 1.4, 1.3, 1.3, 1.3, 1.1,0.8,0.8, &
March 2011 B-49 DRAFT—DO NOT CITE OR QUOTE
-------
1 0.8,0.6,0.5,0.7,0.8,0.6,0.4,0.2, &
2 0.05, 0.08, 0.14, 0.07, 0.06, 0.08, 0.12 /
3
4 ! Larger bolus dosing
5 CONSTANT DRDOSE=0.0 ! Total dose by drinking water in boluses, mg/kg day
6 ! Times for multiple oral drinks/day *after* 0
7 ! Must be ascending, 0 <= times < 24 hr
8 ! CONSTANT DRT=0, 2,4,6,8,10 ! Rat values
9 Constant DRT = 0.0, 3.0, 5.0, 8.0, 11.0, 15.0 ! Human values
10 ! DRTIME(1) assumed = 0 and not used
11 ! Fraction consumed by drinking at those times
12 CONSTANT DRP = 0.25, 0.1, 0.25, 0.1, 0.25, 0.05
13
14 ITotal oral bolus dose; initial value given at t=0 via initial condition
15 TODOSE = DRP(1)*DRDOSE*BW*(1.0-ODS) + ODOSE
16
17 ! IV dosing
18 CONSTANT IVDOSE = 0.0 ! IV dose, mg/kg
19 CONSTANT TINF = 0.025 ! Length of exposure (hrs), default = 1.5 min (bolus)
20 ! 1.5 min reported by Ward & Pollack, (1996)
21 TIV = IVDOSE*BW ! Expected amt infused, mg
22 IV1 = TIV/TINF ! Rate of infusion, mg/hg
23
24 ! For I.V. Runs, control step size if necessary by changing MaxT, not POINTS or CINT
25 MAXT =1.0 ! Maximum Step Size, Hours
26 !IF (IVDOSE.GE.1.0E-4) MAXT = 1.0E-4
27
28 ! Liver infusion
29 CONSTANT LIVRO = 0.0 ! Zero-order liver total, mg/kg/day
30 RLIVO = LIVRO*BW/TCHNG ! Rate in mg/hr
31
32 ! Dose Scheduling
33 CONSTANT MULTE=0 ! Default is *no* repeated dosing/inhalation
34 CIZONE = 1.0 ! Start with inhalation on
35 IVZONE = 1.0 ! Start with IV on
36 schedule ON.AT.24.0
37 SCHEDULE OFF.AT.TCHNG ! Turn off exposure at TCHNG
38 DAY = 0;
39 NEWDAY = 0; IDS = 2 ! First dose given as initial condition
40 IF (MULTE) SCHEDULE ORALDOSE.AT.DRT(2)
41 ALGORITHM IALG = 2 ! Gear algorithm
42 END ! END OF INITIAL
43
44 DYNAMIC
45
46 DERIVATIVE
4? ,„„„„„„„„„.„„„ MeOH .........................
48 IVR = IVZONE1V1 ! IV dosing; IVR = ate of infusion, mg/hg
49 ! Oral Dosing
50 OWING = ((DWRNIGHT*PULSE(0.0,PER1,DUR1)*PULSE(0.0,24.0,12.0) + &
51 DWRDAY*PULSE(0.0,PER2,DUR2)*(1-PULSE(0.0,24.0,12.0)) )*rdrink + &
52 (1-rdrink)*mdrinkp(mod(T,24.0))*0.02*DWDOSE*BW)*PULSE(0.0,168,dayon)
53 RAS = KAS*STOM + VAS*STOM/(KMAS+STOM)
54 RSTOM = OWING+ RAOZ-RAS-KSI*STOM ! Change in stomach (mg/hr)
55 RINT = KSI*STOM-RFEC-KAI*AINTEST ! Change in intestines (mg/hr)
March 2011 B-50 DRAFT—DO NOT CITE OR QUOTE
-------
1 RLZ = RLIVO*CIZONE ! Zero-order to liver
2 RAO = RAS + KAI*AINTEST + RLZ + RAObg*(1.0+incbg*Rincbg*T)
3 ! Oral absorption (mg/hr); last term is endogenous background rate
4 RFEC = KFEC*AINTEST
5 FEC = INTEG(RFEC, 0.0)
6 STOM = INTEG(RSTOM, TODOSE) ! Amt in stomach (mg)
7 AINTEST = INTEG(RINT, 0.0) ! Amt in intestines (mg)
8 OralDoseCheck = INTEG(RAO, 0.0)
9
10 ! Arterial Blood
11 RAAB = QC*(CVLU - CAB)
12 AAB = INTEG(RAAB, AABO) ! Amount, mg
13 CAB = AAB/VAB ! Concentration, mg/L
14 AAUCB = INTEG(CAB, 0.0) ! AUC, hr*mg/L
15
16 dRAAB = QC*(dCVLU - dCAB) ! non-radio-labelled background equations
17 dAAB = INTEG(dRAAB, dAABO) ! Amount, mg
18 dCAB = dAAB/VAB ! Concentration, mg/L
19
20 ! Fat
21 RF = QF*(CAB - CVF)
22 AF = INTEG(RF, AFO) ! Amount, mg
23 CF = AF/VF ! Concentration, mg/L
24 CVF = CF/PF ! AUC, hr*mg/L
25
26 dRF = QF*(dCAB - dCVF) ! non-radio-labelled background equations
27 dAF = INTEG(dRF, dAFO) ! Amount, mg
28 dCF = dAF/VF ! Concentration, mg/L
29 dCVF = dCF/PF! AUC, hr*mg/L
30
31 ! Liver
32 RAL = QL*(CAB - CVL) + RAO - RMETL - RMETL2 - RMETL3
33 AL = INTEG(RAL, ALO) ! Amount, mg
34 CL = AL/VL ! Concentration, mg/L
35 CVL = CL/PL ! Concentration, mg/L
36 AUCL = INTEG(CL, 0.0)! AUC, hr*mg/L
37
38 ! non-radio-labelled background equations ...
39 dRAL = QL*(dCAB - dCVL) + dRAObg*(1.0+incbg*Rincbg*T) - dRMETL - dRMETL2 - dRMETLS
40 dAL = INTEG(dRAL, dALO) ! Amount, mg
41 dCL = dAL/VL ! Concentration, mg/L
42 dCVL = dCL/PL ! Concentration, mg/L
43
44 tCVL = CVL + dCVL
45 ! tCVL = total of labelled and non-radio-labelled liver venous bloood, used
46 ! in metabolic saturation terms to account for mutual inhibition of both forms.
47
48 ! Liver Metabolism
49 RMETL = VMAX*CVL/(KM + tCVL)
50 METL = INTEG(RMETL, 0.0)
51 RMETL2 = VMAX2*CVL/(KM2 + tCVL)
52 METL2 = INTEG(RMETL2, 0.0)
53 RMETL3 = KLL*CVL
54 METL3 = INTEG(RMETL3, 0.0)
55
March 2011 B-51 DRAFT—DO NOT CITE OR QUOTE
-------
1 dRMETL = VMAX*dCVL/(KM + tCVL) ! non-radio-labelled background equations
2 dRMETL2 = VMAX2*dCVL/(KM2 + tCVL)
3 dRMETLS = KLL*dCVL
4
5
6 ! Total Amount Metabolized (Formate and Formaldehyde)
7 ! Does not include K1C for human MeOH excretion estimate
8 AMET = METL + METL2 + METL3
9 AMET24 = AMET*24.0/TSTOP
10 ! Total amount metabolized in last tmetf hr of exposure, averaged per day
11 AMETF = INTEG((RMETL+RMETL2+RMETL3)*PULSE(TSTOP-
12 tmetf,TSTOP,tmetf),0.0)*24.0/tmetf
13 ! (tmetf = 24.0*metd)
14 ! Chamber concentration (mg/L)
15 RACh = (Rats*QP*CLEx) - (FRACinh*Rats*QP*CCh) - (kl_oss*ACh)
16 ACh = INTEG(RACh, AChO)
17
18 ! The following calculation yields an air concentration equal to the
19 ! closed chamber value if a closed chamber run is in place and a
20 ! specified constant air concentration if an open chamber run is in place
21 CCh = ACh*Cizone/VCh
22 CCPPM = CCh*24451/MWMe
23 CLoss = INTEG(kLoss*ACh, 0.0)
24
25 ! Lungs
26 RALu = QP*(FRACinh*CCh - CLEx) + QC*(CVB - CVLu)
27 ALu = INTEG(RALu, ALuO)
28 CLu = ALu/VLu ! Concentration, mg/L
29 CVLu = CLu/PLu ! Exiting Concentration, mg/L
30
31 dRALu = QC*(dCVB - dCVLu) - QP*dCLEx ! non-radio-labelled background eqns
32 dALu = INTEG(dRALu, dALuO)
33 dCLu = dALu/VLu ! Concentration, mg/L
34 dCVLu = dCLu/PLu ! Exiting Concentration, mg/L
35
36 ! Amount Inhaled
37 Rlnh = FRACinh*QP*CCh
38 Alnh = INTEG(Rlnh, 0.0) ! mg per rat
39 AlnhC = Alnh*Rats ! mg for a group of rats
40
41 ! Amount Exhaled
42 CLEx = CVLu/PB ! Concentration, mg/L
43 dCLEx = dCVLu/PB ! non-radio-labelled background equations
44 ! changed from CVB/PB to CVLu/PB by Paul Schlosser, U.S. EPA, 12/8/09
45 ! This makes it a standard venous equilibrium gas exchange model!
46 RAEx = QP*CLEx
47 AEx = INTEG(RAEx, 0.0)*PULSE(0,TCHNG,TSTOP) ! Amount, mg per rat
48 AExC = AEx*Rats ! Amount, mg, for a group of rats
49 AxF = INTEG(RAEx*PULSE(TCHNG,24,24), 0.0) ! Amount exhaled post-exposure
50
51 ! Rest of Body
52 RAR = QR*(CAB - CVR)
53 AR = INTEG(RAR, ARO) ! Amount, mg
54 CR = AR/VR ! Concentration, mg/L
55 CVR = CR/PR ! Exiting Venous Concentration, mg/L
March 2011 B-52 DRAFT—DO NOT CITE OR QUOTE
-------
1 AUCR = INTEG(CR, 0.0) ! AUC, hr*mg/L
2
3 dRAR = QR*(dCAB - dCVR) ! non-radio-labelled background equations
4 dAR = INTEG(dRAR, dARO) ! Amount, mg
5 dCR = dAR/VR ! Concentration, mg/L
6 dCVR = dCR/PR ! Exiting Venous Concentration, mg/L
7
8 ! Venous Blood (mg)
9 RURB = K1*CVB*WB ! Lumped Clearance from Blood
10 RAVB = QF*CVF + QL*CVL + QR*CVR + IVR - QC*CVB - RURB
11 AVB = INTEG(RAVB, AVBO) ! Amount, mg
12 CVB = AVB/WB ! Concentration, mg/L
13
14 dRURB = K1*dCVB*WB ! non-radio-labelled background equations
15 dRAVB = QF*dCVF + QL*dCVL + QR*dCVR - QC*dCVB - dRURB
16 dAVB = INTEG(dRAVB, dAVBO)! Amount, mg
17 dCVB = dAVB/VVB ! Concentration, mg/L
18
19 AUCB = INTEG(CVB, 0.0) ! AUC, hr*mg/L (total over entire exposure)
20 AUCBB = AUCB*24.0/TSTOP ! Average over exposure, hr*mg/(L*day)
21 AUCBF = INTEG(CVB*PULSE(TSTOP-tmetf,TSTOP,tmetf),0)*24.0/tmetf
22 ! AUCBF = Last tmetf AUC averaged/day (tmetf = 24.0*metd)
23 ! For "steady state" AUC in blood over a day, set exposures to
24 ! several weeks to reach "periodicity", then use AUCBF w/ metd = 7
25
26 ! Bladder compartment, added by PS, U.S. EPA, 10/2008
27 RBL = KBL*ABL ! Rate of clearance from bladder (mg/hr)
28 ABL = INTEG((RURB-RBL),ABLO) ! Amount in bladder (mg)
29 RUR= RBL/(BW*0.5e-3)! Urine concentration = rate/[BW*(0.5e-3 L/h/kg BW)]
30 URB = INTEG(RBL, 0.0) ! Amount cleared to urine, mg
31 URBF = INTEG(RURB*PULSE(TSTOP-tmetf,TSTOP,tmetf),0)*24.0/tmetf
32 ! Amount cleared to urine in last tmetf averaged/day (tmetf = 24.0*metd)
33
34 I************************* Magg Balance **************************
35 Tbody = AAB + AF + AL + ALU + AR + AVB + ABL + STOM + AINTEST
36 MetabORCIrd = URB + METL + METL2 + METL3 + AEX + FEC
37 TMass = Tbody + MetabORCIrd
38 TDose = AinH + INTEG(IVR+DWING+RAOZ+RLZ,0.0) + TODOSE
39 Mass Bal=100*(TDose-TMass)/(TMass+1 e-12)
40 Icompare to TIV, ODOSE, or AINHC
41 ! Check Blood Flows
42 QTOT = QF + QL + QR
43 QRECOV= 100.0*QTOT/QC
44 END ! End of Derivative
45 TERMT(T.GE.TStop)
46
47 | Exposure Control
48 DISCRETE ORALDOSE ! Stom is amount in stomach
49 IDOSE = DRP(IDS)*DRDOSE*BW
50 STOM = STOM + IDOSE ! Drinking percent
51 TODOSE = TODOSE + IDOSE
52 IF(IDS.EQ.1)THEN
53 STOM = STOM + ODOSE
54 TODOSE = TODOSE + ODOSE
55 ENDIF
March 2011 B-53 DRAFT—DO NOT CITE OR QUOTE
-------
1 IDS = IDS+1
2 IF(IDS.EQ.7)THEN ! For 6 doses
3 IDS = 1
4 N EWDAY = N E WDAY + 24
5 SCHEDULE ORALDOSE.AT.NEWDAY ! Go to start of the next day
6 ELSE
7 SCHEDULE ORALDOSE.AT.(NEWDAY+DRT(IDS)) ! Go to next drink time
8 ENDIF
9 END ! OF DISCRETE ORALDOSE
10
11 DISCRETE OFF ! Turn IN HAL exposure off
12 CIZONE = 0.0
13 IVZONE = 0.0
14 DAY=DAY+1
15 IF (MULTE) SCHEDULE ON.AT.(DAY*24.0)
16 END ! OF DISCRETE OFF
17
18 DISCRETE ON
19 CIZONE=1.0
20 SCHEDULE OFF.AT.(T+TCHNG)
21 END ! OF DISCRETE ON
22
23 DISCRETE DS1 ! Human at rest
24 ! Equations scheduled for change during simulation repeated here
25 QC = QCCHR*BW**0.75
26 OP = QPCHR*BW**0.75
27 OF = QFC*QC ! QFCHR*QC ! Equations for alternate flow fractions
28 QL = QLC*QC ! QLCHR*QC ! But QFC and QLC taken to be 'at rest' values
29 QRC = 1.0-(QFC + QLC)
30 OR = QRC*QC
31 FRACINH = FRACIN
32 END ! OF DISCRETE DS1
33
34 DISCRETE DS2 ! Human at work (50W)
35 ! Equations scheduled for change during simulation repeated here
36 QC = QCCHW*BW**0.75
37 OP = QPCHW*BW**0.75
38 OF = QFC*QC ! QFCHW*QC ! Equations for alternate flow fractions
39 QL = QLC*QC ! QLCHW*QC ! But don't seem to work (fit data) well
40 QRC = 1.0 - (QFC + QLC)
41 QR = QRC*QC
42 FRACINH=FRACINW
43 END ! OF DISCRETE DS2
44
45 END ! End of Dynamic
46 END ! End of Program
B.3.2. acslXtreme procedure (.cmd) file
47 ! File MEOHCBMMfinal.CMD - FOR PBPK MODEL FOR METHANOL
48 ! taken from .cmd file from Ward et al. (1997), Edited by KWW - 06/02/96
49 ! Developed for this (CBMM) model - 4/15/15
50 ! Final with Digitized Data - 5/25/05
51 ! Final Version has fast and slow rates of oral absorption
52 ! Version 4 is final version used for simulations
March 2011 B-54 DRAFT—DO NOT CITE OR QUOTE
-------
1 ! Final Version 1.10.06
2 ! Beyond this comment, this file is left "as is" for archival purposes. But most if not all of
3 ! the functions and data sets defined here are replicated and/or replaced in the .m files below.
4 ! Only use these when there is no corresponding .m file.
5 ! - Edited by Paul Schlosser (U.S. EPA), October 2008
7 PREPARE T,CVB,MetB
8
9 ! Procedural blocks for general mouse/rat data
10 PROCED CDMICE ! Anatomic/physiologic data for mice
11 SET BW=0.03, TSTOP=1.5
12 SET IVDose=0, DOSE=0, CONCppm=0
13 SET PL=1.06, PF=0.083, PR=0.66, PB=1350
14 SET QPC=25.4,QCC=25.4,fracin=0.73
15 SET QLC = 0.25,QFC=0.05
16 SET KM=12,VmaxC=14.3,KLC=0.0,KAS=2
17 SET Vmax2c=19,km2=210,KAI=0.22,KSI=l.l
18 SET VAC = 0.0123,VFC = 0.07,VLC = 0.055
19 SET VLuC = 0.0073, WBC = 0.0368
20 ! Volumes from Brown et al. (1997)
21 IMouse QPC avg from Brown 29, 24 used in Corley et al. (1994) and others
22 !AVG of measured vent rates by Perkins et al. (1995a) 25.4 L/hr/kgA0.75
23 IBlood volume 4.9% total. As per Brown 25:75 split art:ven
24 IMetab originally from Ward et al. (1997)- KldC for mice =0
25 END
26
27 PROCED HUMAN
28 SET BW=70
29 SET IVDose=0, DOSE=0, CONCppm=0
30 SET PL=1.06, PR=0.66, fracin=0.75
31 SET VFC=0.214, VLC=0.026,VLUC=0.008
32 SETVAC=0.0198,WBC=0.0593
33 SET QPC=18.5, QCC=18.5, QLC=0.227, QFC=0.052
34 SET KM=12,VmaxC=l 1,KLC=0.044,KAS=2.0
35 SETKAI=0.22,KSI=1.1
36 SETPB = 1626,PF=0.14
37 SET Vmax2c=0
38 ! Volumes from Brown et al. (1997)
39 !QPC from Brown et al. (1997). upper end 13.4 L/hr/kgA0.75
40 INeed higher for data, 15 L/hr/kgA0.75 used in several published human models
41 IBlood volume 7.9% total. As per Brown 25:75 split art:ven
42 IFrac absorbed from Ernstgard SOT poster + personal communication
43 IHuman Partition Coef. equal to mice. Horton et al. (1992) used rat
44 Except Human Partition Coef blood and fat - from Fiserova-Bergerova & Diaz, (1986)
45 ! - but rat values are inconsistent with expected fat partitioning for an alcohol like this
46 ! - for example Pastino & Conolly (2000) EtOH model, fat PC =0. 1
47 END
48
49 PROCED SDRAT ! Anatomic/physiologic data for rats
50 SET BW=0.3, TSTOP=1.5
5 1 SET IVDose=0, DOSE=0, CONCppm=0
52 SETPL=1.6,PF=0.1,PR=1.3
53 SET KM=45,VmaxC=15,KLC=0. 1,KAS=5
54 SET VAC = 0.0185, VFC=0.07, VLC= 0.034, VLuC=0.005, WBC=0.0555
55 ! Volumes from Brown, et. Al. (1997)
56 !PC from Horton et al. (1992). PF reduced to 0. 1 from Horton's 1 . 1
March 2011 B-55 DRAFT—DO NOT CITE OR QUOTE
-------
1 IBlood volume 7.4% total. As per Brown 25:75 split art:ven
2 IMetab originally from Ward et al. (1997) - KldC for mice =0
3 !Rat model not calibrated
4 END
5
6 PROCED PREG
7 !For GD 18 mice, BW increased as estimated from Rogers et al. (1993)
8 ! Increased VFC as per Corley CRT development review
9 IThis just to give a WAG as to how data might change from B W and different volume of distribution
10 !Not invoked for any PROCs below as the default
11 ILiverto 140%ofNP
12 SET BW = 0.055, VFC=0.08,VLC=0.11,WBC=0.05
13 END
14
15 PROCED CLEARIT
16 SET IVDose=0, DOSE=0, CONCppm=0
17 END
18
19 PROCED SHOWIT
20 display Vmaxc,km,klc,pb,pf,pr,pl,kas,fracin
21 END
22
23 [Procedural blocks for all non-pregnant mouse data
24 !IV
25 PROCED MWARDIV25
26 IWard et al. (19971
27 IFigure 2, data from Ward model cmd
28 IData was checked via digitizit - within +1-5% of cmd file
29 CLEARIT
30 CDMICE
31 SET TSTOP=24.0
32 SET IVDOSE=2500., tchng=0.025
33 END
34
35 PROCED PMWARDIV25
36 PLOT /D=MWARDIV25, CVB
37 END
38
39 DATA MWARDIV25(T,CVB)
40 0.08 4481.8
41 0.25 4132.2
42 0.5 3888
43 1.00 3164.8
44 2.0 2303.5
45 4.00 1921.5
46 6 1883.8
47 8 1620
48 12 838
49 18 454.7
50 24 NaN
51 END
52
53 PROCED MWARD95IV25
54 ! Ward etal. (1995)
55 IFigure 2
56 IData via digitizit
March 2011 B-56 DRAFT—DO NOT CITE OR QUOTE
-------
1 CLEARIT
2 CDMICE
3 SET TSTOP=24.0
4 SET IVDOSE=2500., tchng=0.025
5 END
6
7 PROCED PMWARD95IV25
8 PLOT /D=MWARD95IV25, CVB
9 END
10
11 DATA MWARD95IV25(T,CVB)
12 0.53 3299.60
13 1.06 3244.54
14 1.54 3190.71
15 3.07 2803.13
16 4.07 2544.36
17 5.02 2237.77
18 6.02 2063.59
19 7.02 1873.10
20 8.02 1521.92
21 9.03 1670.30
22 10.03 1423.12
23 END
24
25 ! Procs for pegnant IV below: MWARDGD9IV25, MWARDGD18IV25, MWARDGD18IV5, MWARDGD18IV1
26 ! Oral
27 PROCED MWARDPO25
28 IWard et al. (1995)
29 IFigure 2, data from Ward model cmd
30 IData was checked via digitizit - within +1-5% of cmd file
31 CLEARIT
32 CDMICE
33 SET TSTOP=24, DOSE=2500
34 END
35
36 PROCED PMWARDPO25
37 PLOT /D=MWARDPO25, CVB
38 END
39
40 DATAMWARDPO25(T,CVB)
41 0.504 2370
42 0.96 2645
43 1.44 2705
44 1.992 2719
45 2.208 2781
46 3 2704
47 4.008 2370
48 4.992 2617
49 6 2516
50 7.008 2635
51 7.992 2213
52 9 2370
53 10.008 2028
54 10.992 1916
55 12 1347
56 13.008 1467
March 2011 B-57 DRAFT—DO NOT CITE OR QUOTE
-------
13.992
15
16.008
16.992
18
19.01
21
24
END
1354
1175
864.3
745.2
422.4
428
243
136
1
2
3
4
5
6
7
8
9
10
11 IProcs for pregnant Oral below: MDORGD8PO15, MWARDGD18PO25
12 ! Inhalation
13 ! QPC set to measured as in Perkins et al. (1995a) for each concentration
14
15 PROCED MPERKIN25
16 IPerkins et al. (1995a)
17 !Fig. 2 data in Ward cmd file
18 CLEARIT
19 CDMICE
20 SET TSTOP=24, CONCppm=2500, vchc=5000
21 SET QPC = 29., QCC=29.
22 SET TCHNG=8
23 END
24
25 PROCED PMPERKIN25
26 PLOT /D=MPERKIN25, CVB
27 END
28
29 IThis data from Digitizlt
30 DATAMPERKIN25(T,CVB)
31 2.0 414.0
32 4.0 453.0
33 6.0 586.0
34 8.25 694.0
35 12 282.0
36 16 0.6
37 END
38
39 IThis data from cmd file
40 !DATAMPERKIN25(T,CVB)
41 11.99 386.49
42 14.01 617.57
43 16.00 816.22
44 18.26 970.27
45 112.00 393.24
46 116.0 13.51
47 END
48
49 PROCED MPERKIN50
50 IPerkins et al. (1995a)
51 !Fig. 2, data in Ward cmd file
52 IData in command file higher than appears in figure
53 CLEARIT
54 CDMICE
55 SET TSTOP=24, CONCppm=5000, vchc=5000
56 SET TCHNG=8, qpc=24.,qcc=24.
March 2011 B-58 DRAFT—DO NOT CITE OR QUOTE
-------
1 END
2
3 PROCED PMPERKIN50
4 PLOT /D=MPERKIN50, CVB
5 END
6
7 Ithis from Digitizit, Fig 2, Perkins et al. (1995a)
8 DATAMPERKIN50(T,CVB)
9 1 644.00
10 2 877.00
11 3 1340.00
12 4 1450.00
13 6 2040.00
14 8.25 2290.0
15 12 1410.0
16 16 583.0
17 20 271.0
18 24 9.7
19 END
20
21 IThis data from cmd file
22 !DATAMPERKIN50(T,CVB)
23 !1.0 906.76
24 12.0 1202.7
25 !3.0 1828.38
26 !4.0 1986.49
27 !6.0 2800
28 !8.3 3125.68
29 112.0 1914.86
30 116.0 806.76
31 120.0 367.57
32 124.0 10.81
33 !END
34
35 PROCED MPERKIN100
36 IPerkins et al. (1995a)
37 !Fig. 2 data in Ward cmd file
38 INote, Table 6 in Ward paper - max value of 3260 +/-151
39 CLEARIT
40 CDMICE
41 SET TCHNG=8, CONCppm= 1,0000, tstop=36,vchc=5000
42 SETQPC=21,qcc=21
43 END
44
45 PROCED PMPERKIN100
46 PLOT /D=MPERKIN100, CVB
47 END
48
49 Ithis from Digitizit, Fig 2 Perkins et al
50 DATAMPERKIN100(T,CVB)
51 2.0 2080.0
52 4.0 2530.0
53 6.0 3350.0
54 8.25 3350.0
55 12 2370.0
56 16 1830.0
March 2011 B-59 DRAFT—DO NOT CITE OR QUOTE
-------
20 1080.0
24 591.0
28 44.6
END
IDATA MPERKIN100(T,CVB)
IThis from original cmd file
12.0 2809.46
!4.0 3405.4
16.0 4528.38
!8.3 4524.32
112.0 3212.16
116.0 2456.76
120.0 1439.19
124.0 798.65
128.0 55.4
!END
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 ! Procs for Preg mouse Inhalation date below: MDOR8IN10,MDOR8IN15
20 !and:MROGGD7IN10, MROGGD6IN1, MROGGD6IN2, MROGGD6IN5, MROGGD6IN10
21
22 ! pregnant mice
23 !IV
24 PROCED MWARDGD9IV25
25 ! Ward etal. (1996)
26 !Not used in the manuscript, only in cmd file
27 CLEARIT
28 CDMICE
29 SET TSTOP=24
30 SET IVDOSE=2500., TINF=0.025
31 END
32
33 PROCED PMWARDGD9IV25
34 PLOT /D=MWARDGD9IV25, CVB
35 END
36
37 DATA MWARDGD9IV25(T,CVB)
38 0.0833 4606.2
39 0.25 4079.5
40 0.5 3489.3
41 1 2939.6
42 2 3447.6
43 4 2605.0
44 6 2690.5
45 8 2574.9
46 12 1506.1
47 18 498.6
48 24. NaN
49 END
50
51 PROCED PROCED MWARDGD18IV25
52 ! Ward etal. (1996)
53 INote, Table 6 in Ward paper - max value of 3521+/- 492
54 CLEARIT
55 CDMICE
56 SET TSTOP=24
March 2011 B-60 DRAFT—DO NOT CITE OR QUOTE
-------
0.25
0.5
1.0
2.0
4.0
6.0
8.0
12
18
24.0
END
3445.1
2936.8
2470.5
2528.1
2292.3
2269.4
2057.0
1805.9
1482.2
496.1
1 SET IVDOSE=2500., TINF=0.025
2 END
o
6
4 PROCEDPMWARDGD18IV25
5 PLOT /D=MWARDGD18IV25, CVB
6 END
7
8 DATAMWARDGD18IV25(T,CVB)
9 0.0833 4250.0
10
11
12
13
14
15
16
17
18
19
20
21
22 PROCEDMWARDGD18IV5
23 ! Ward etal. (1996)
24 INote, Table 6 in Ward paper - max value of 868.8 +/- 53.9
25 CLEARIT
26 CDMICE
27 SET TSTOP=6
28 SET IVDOSE=500., TINF=0.025
29 END
30
31 PROCED PMWARDGD18IV5
32 PLOT /D=MWARDGD18IV5, CVB
33 END
34
35 DATAMWARDGD18IV5(T,CVB)
36 0.25 854.7
37 0.5 720.2
38 1.0 624.1
39 2.0 453.2
40 3.0 307.6
41 4.0 217.7
42 4.5 202.6
43 END
44
45 PROCED MWARDGD18IV1
46 IWard et al. (19961
47 IWard Proc GD8, but must be 18 as per PBPK manuscript
48 INote, Table 6 in Ward paper - max value of 252 +/-12.9
49 liable matches file
50 CLEARIT
51 CDMICE
52 SET TSTOP=4
53 SETIVDOSE=100.
54
55 PROCED PMWARDGD 18IV1
56 PLOT /D=MWARDGD18IV1, CVB
March 2011 B-61 DRAFT—DO NOT CITE OR QUOTE
-------
1 END
2
3 DATAMWARDGD18IV1(T,CVB)
4 0.25 252
5 0.52 242.2
6 1.0 222.7
7 2 176.4
8 3 134.2
9 3.5 94.41
10 END
11
12 ! Oral
13
14 PROCED MDORGD8PO15
15 IWard et al. (1997), cmd file
16 INote, Table 6 in Ward paper - max value of 1610 +/- 704
17 ! Table and file match w/in round off
18 IData must be from Dorman
19 IDorman Teratology, 1995, Fig. 1
20 Iwithin error for Digitiz data the same
21 CLEARIT
22 CDMICE
23 SET TSTOP=24,DOSE= 1500
24 END
25
26 PROCED PMDORGD8PO15
27 PLOT /D=MDORGD8PO15, CVB
28 END
29
30 DATAMDORGD8PO15(T,CVB)
31 1 1609.6
32 2 1331.2
33 4 1241.6
34 8 707.2
35 16 160.0
36 24 38.4
37 END
38
3 9 PROCED MWARDGD18PO25
40 IWard et al. (1996)
41 INote, Table 6 in Ward paper - max value of 3205 +/- 291
42 CLEARIT
43 CDMICE
44 SET TSTOP=24, DOSE=2500
45 END
46
47 PROCED PMWARDGD18PO25
48 PLOT /D=MWARDGD18PO25, CVB
49 END
50
51 Ifrom cmd file, replaced with digitized
52 IDATA MWARDGD18PO25(T,CVB)
53 10.25 2770.
54 10.5 3299.
55 II 3336.
56 12 3502.
March 2011 B-62 DRAFT—DO NOT CITE OR QUOTE
-------
1 !4 3217.
2 !6 2999.
3 !10 2036.
4 !12 1832.
5 !15 949.1
6 !18 403.5
7 !21 40.47
8 !24. 16.03
9 !END
10
11 IDigitizit data
12 DATAMWARDGD18PO25(T,CVB)
13 0.5 2024
14 1 2554
15 2 3193
16 4 3002
17 6 2933
18 10 1976
19 12 1922
20 15 1339
21 18 1033
22 21 832
23 24 580
24 END
25
26 ! Inhalation
27
28 PROCED MDOR8IN10
29 ! Ward et al. (1997)
30 INote, Table 6 in Ward paper - max value of 2080 +/- 800
31 !Fig 7? Table 6 attributes to Dorman
32 IDigitizit of Dorman Fig 2 matches cmd file
33 lactual exposure ppm 9900
34 CLEARIT
35 CDMICE
36 SET TCHNG=6, CONCppm=9900, tstop=36
37 END
38
39 PROCED PMDOR8IN10
40 PLOT /D=MDOR8IN10, CVB
41 END
42
43 DATA MDOR8IN10(T,CVB)
44 1 771.2
45 2 1017.6
46 4 1788.8
47 6 2076.8
48 8 2281.6
49 16 1152.0
50 24 268.8
51 END
52
53 PROCED MDOR8IN15
54 ! Ward et al. (1997)
55 INote, Table 6 in Ward paper - max value of 7136 +/-736
56 !Fig 7? Table 6 attributes to Dorman
March 2011 B-63 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 IDigitizit of Dorman Fig 2 matches cmd file
2 CLEARIT
3 CDMICE
4 SET TCHNG=6, CONCppm= 15000, tstop=36
5 SET vchc=5000000000
6 END
7
8 PROCED PMDOR8IN15
9 PLOT /D=MDOR8IN15, CVB
10 END
11
12 DATAMDOR8IN15(T,CVB)
13
14
15
16
17
18
19
20
21
22 IFiles above provided in cmd file from Ward et al. (1997) PBPK model
23 IFiles below added for this evaluation,
24 ! sources described in proc files and in notebook
25
26 PROCED MROGGD7IN10
27 ! Rogers et al. (1997)
28 ! Actual Values kindly provided by Rogers
29 CLEARIT
30 CDMICE
31 SET TCHNG=7, CONCppm= 1,0000, tstop=36
32 SET vchc=500000000,bw=0.032
33 END
34
35 PROCED PMROGGD7IN10
36 PLOT /D=MROGGD7IN10, CVB
37 END
38
39 DATA MROGGD7IN10(T,CVB)
40
41
42
43
44
45
46
47
48
49
50
51
52 PROCED MROGGD6IN1
53 CLEARIT
54 IRogers et al. (1993)
55 IRogers data from GD 6 and 10
56 !InTable2
March 2011 B-64 DRAFT—DO NOT CITE OR QUOTE
1
4
6
7
7.5
8
9
12
16
26
END
930
2800
3360
3990
3980
4120
3270
2630
1690
60
-------
1 CDMICE
2 SET TCHNG=7, CONCppm= 1,000, tstop=36,vchc=500000000,bw=0.032
3 END
4
5 PROCED PMROGGD6IN1
6 PLOT /D=MROGGD6IN1, CVB
7 END
8
9 DATAMROGGD6IN1(T,CVB)
10 7 63
11 7 131
12 END
13
14
15 PROCED MROGGD6IN2
16 ! Rogers et al. (1993)
17 IRogers data from GD 6 and 10
18 !InTable2
19 CLEARIT
20 CDMICE
21 SET TCHNG=7, CONCppm=2000, tstop=36, vchc=500000000,bw=0.032
22 END
23
24 PROCED PMROGGD6IN2
25 PLOT /D=MROGGD6IN2, CVB
26 END
27
28 DATAMROGGD6IN2(T,CVB)
29 7 487
30 7 641
31 END
32
33 PROCED MROGGD6IN5
34 ! Rogers et al. (1993)
35 IRogers data from GD 6 and 10
36 !InTable2
37 CLEARIT
38 CDMICE
39 SET TCHNG=7, CONCppm=5000, tstop=36,vchc=500000000,bw=0.032
40 END
41
42 PROCED PMROGGD6IN5
43 PLOT /D=MROGGD6IN5, CVB
44 END
45
46 DATAMROGGD6IN5(T,CVB)
47 7 2126
48 7 1593
49 END
50
51 PROCED MROGGD6IN10
52 ! Rogers et al. (1993)
53 IRogers data from GD 6, 10, 15
54 !InTable2
55 CLEARIT
56 CDMICE
March 2011 B-65 DRAFT—DO NOT CITE OR QUOTE
-------
1 SET TCHNG=7, CONCppm= 1,0000, tstop=36,vchc=500000000,bw=0.032
2 END
o
6
4 PROCED PMROGGD6IN10
5 PLOT /D=MROGGD6IN10, CVB
6 END
7
8 DATAMROGGD6IN10(T,CVB)
9 7 4653
10 7 4304
11 7 3655
12 END
13
14 Human inhalation dta
15
16 PROCED HJOHIN1
17 lErnstgard et al. (2005) SOT poster 200 ppm human
18 IDigitized from Fig 2
19 I Also personal communication - Ernstgard
20 !QPC from Johanson et al. (1986) Scand J. Work Env. 86 =52.6
21 !If Assume value = alveolar, similar to Astrand '83 value of 56 L/hr/krA0.75
22 IFracin - 50% of total (from poster) -76%
23 !QCC from Corley et al. (1994)
24 CLEARIT
25 HUMAN
26 SET TCHNG=2, CONCppm=100, tstop=16
27 SET QPC=52.6,qcc=26,vchc=500000000
28 END
29
30 PROCED PHJOHIN1
31 PLOT /D=HJOHIN1, CVB
32 END
33
34 DATAHJOHIN1(T,CVB)
35 0.20 0.87
36 0.46 1.50
37 0.97 2.31
38 1.46 3.24
39 1.91 3.65
40 2.17 3.52
41 2.50 2.55
42 2.91 2.23
43 3.51 1.59
44 4.01 1.72
45 5.02 0.41
46 6.00 0.50
47 9.24 0.12
48 END
49
50 PROCED HJOHIN2
51 lErnstgard et al. (2005) SOT poster 200 ppm human
52 IDigitized from Fig 2
53 I Also personal communication - Ernstgard
54 IQPC from Johanson et al. (1986) 86 =52.6
55 llf Assume value = alveolar, similar to Astrand '83 value of 56 L/hr/krA0.75
56 IFracin - 50% of total (from poster) -75%
March 2011 B-66 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
!QCC from Corley et al. (1994)
CLEARIT
HUMAN
SET TCHNG=2, CONCppm=200, tstop=16
SETQPC=52.6,qcc=26,vchc=500000000
END
PROCED PHJOHIN2
PLOT /D=HJOHIN2, CVB
END
DATA HJOHIN2(T,CVB)
0.22 1.63
0.49 2.92
0.92 4.76
1.47 6.30
1.90 7.65
2.16 6.20
2.47 5.49
2.91 4.96
3.50 3.64
4.00 3.43
4.99 1.94
5.97 1.03
8.90 0.21
END
PROCED HOSTERIN2
! Osterloh et al. (1996)
! Digitized data provided by EPA
! Subtracted background from exposure blood levels
CLEARIT
HUMAN
SET TCHNG=4, CONCppm=200, tstop=16
SET vchc=500000000, BW=78.2
END
PROCED PHOSTERIN2
PLOT /D=HOSTERIN2, CVB
END
DATA HOSTERIN2(T,CVB)
0.05 0.54
0.25 1.39
0.50 1.82
0.75 2.28
1.00 2.42
1.50 2.94
2.00 3.37
2.50 3.90
3.00 4.21
3.50 4.61
4.00 4.82
5.00 2.99
6.00 2.30
7.00 1.40
March 2011
B-67
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
7.95 1.07
END
PROCED HBATIN82
IBatterman et al., (1998) Int Arch Occ Health
[Digitized Data
CLEARIT
HUMAN
SET TCHNG=2, CONCppm=800,
SET vchc=500000000
END
PROCED PHBATIN82
PLOT /D=HBATIN82, CVB
END
DATA HBATIN82(T,CVB)
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
PROCED HBATIN81
IBatterman et al. (1998)
[Digitized Data
CLEARIT
HUMAN
SET TCHNG=1, CONCppm=800,
SET vchc=500000000
END
PROCED PHBATIN81
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. (1998)
tstop=16
tstop=16
March 2011
B-68
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
[Digitized Data
Ibody weight not provided
CLEARIT
HUMAN
SET TCHNG=0.5, CONCppm=800, tstop=16
SET vchc=500000000
END
PROCED PHBATIN830
PLOT /D=HBATIN830, CVB
END
DATA HBATIN830(T,CVB)
0.579 4.608
0.857 4.685
1.137 4.870
1.650 3.452
2.650 2.082
3.662 0.910
4.693 0.316
5.713 0.320
6.643 0.292
7.696 0.547
END
PROCED HSEDIN231
ISedivecetal. (1981)
IDigitized Data
INote, urine volumes not given, these are estimates
lurine production of 0.75 mg/hr, this for info purposes only! ! !
CLEARIT
HUMAN
SET TCHNG=8, CONCppm=231, tstop=24
SET vchc=500000000
END
PROCED PHSEDIN23 1
PLOT /D=HSEDIN231, Metb
END
DATA HSEDIN23 l(T,Metb)
0.043 0.0042
2.174 0.33
4.478 0.87
6.478 1.46
8.522 2.15
10.348 2.63
12.130 2.91
14.044 3.07
18.870 3.32
23.696 3.52
END
PROCED HSEDIN157
ISedivecetal. (198TJ
IDigitized Data
March 2011 B-69 DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
INote, urine volumes not given, these are estimates
lurine production of 0.75 mg/hr, this for info
HUMAN
SET TCHNG=8, CONCppm=157, tstop=24
SET vchc=500000000
END
PROCED PHSEDIN157
PLOT /D=HSEDIN157, Metb
END
DATA HSEDIN157(T,Metb)
0.126 0.0038
2.204 0.228
4.242 0.576
6.196 0.975
8.326 1.47
10.163 1.81
12.094 2.00
14.016 2.12
18.8966 2.34
23.776 2.53
END
PROCED HSEDIN78
ISedivecetal. (1981)
[Digitized Data
purposes only! !! CLEARIT
INote, urine volumes not given, these are estimates
lurine production of 0.75 mg/hr, this for info purposes only 1 1 1
CLEARIT
HUMAN
SET TCHNG=8, CONCppm=78, tstop=24
SET vchc=500000000
END
PROCED PHSEDIN78
PLOT /D=HSEDIN78, Metb
END
DATA HSEDIN78(T,Metb)
0.03 0.013
2.06 0.189
3.96 0.397
6.09 0.652
8.09 0.820
10.11 0.933
11.93 1.02
13.92 1.09
18.89 1.27
END
IAUC, Cmax estimation procedures
Proced mousin
ITo determine AUC for 7 hr exposure in mice
CLEARIT
CDMICE
March 2011 B-70 DRAFT—DO NOT CITE OR QUOTE
-------
1 SET TCHNG=7, tstop=24
2 SET vchc=50000000000
3 SET CONCppm=l
4 start /nc
5 d concppm,AUCB,amet,cvb
6 SET CONCppm=5
7 start /nc
8 d concppm,AUCB,amet,cvb
9 SET CONCppm=10
10 start /nc
11 d concppm,AUCB,amet,cvb
12 SET CONCppm=25
13 start/nc
14 d concppm,AUCB,amet,cvb
15 SET CONCppm=50
16 start /nc
17 d concppm,AUCB,amet,cvb
18 SET CONCppm=75
19 start/nc
20 d concppm,AUCB,amet,cvb
21 SET CONCppm=100
22 start /nc
23 d concppm,AUCB,amet,cvb
24 SET CONCppm=175
25 start /nc
26 d concppm,AUCB,amet,cvb
27 SET CONCppm=208.3
28 start /nc
29 d concppm,AUCB,amet,cvb
30 SET CONCppm=250
31 start /nc
32 d concppm,AUCB,amet,cvb
33 SET CONCppm=325
34 start /nc
35 d concppm,AUCB,amet,cvb
36 SET CONCppm=500
37 start /nc
38 d concppm,AUCB,amet,cvb
39 SET CONCppm=750
40 start /nc
41 d concppm,AUCB,amet,cvb
42 SET CONCppm= 1,000
43 start /nc
44 d concppm,AUCB,amet,cvb
45 SET CONCppm=2000
46 start /nc
47 d concppm,AUCB,amet,cvb
48 SET CONCppm=2500
49 start /nc
50 d concppm,AUCB,amet,cvb
51 SET CONCppm=5000
52 start /nc
53 d concppm,AUCB,amet,cvb
54 SET CONCppm= 1,0000
55 start /nc
56 d concppm,AUCB,amet,cvb
March 2011
B-71
DRAFT—DO NOT CITE OR QUOTE
-------
1 SET CONCppm=50000
2 start /nc
3 d concppm,AUCB,amet,cvb
4 END
5
6 Proced mousinC
7 !To determine 7 hr Cmax, note - not at SS
8 CLEARIT
9 CDMICE
10 SET TCHNG=7, tstop=7,VCHC=50000000000
11 SETCONCppm=l
12 start /nc
13 dconcppm,cvb
14 SET CONCppm=10
15 start/nc
16 d concppm,CVB
17 SET CONCppm=50
18 start/nc
19 dconcppm,CVB
20 SET CONCppm=100
21 start/nc
22 d concppm,CVB
23 SET CONCppm=250
24 start /nc
25 d concppm,CVB
26 SET CONCppm=500
27 start /nc
28 d concppm,CVB
29 SET CONCppm= 1,000
30 start /nc
31 d concppm,CVB
32 SET CONCppm=2000
33 start/nc
34 d concppm,CVB
35 SET CONCppm=2500
36 start /nc
37 d concppm,CVB
38 SET CONCppm=5000
39 start/nc
40 d concppm,CVB
41 SET CONCppm= 1,0000
42 start /nc
43 d concppm,CVB
44 SET CONCppm=50000
45 start /nc
46 d concppm,CVB
47 END
48
49 Proced humin
50 ! To determine 24 hr AUC,Cmax at SS for human
51 CLEARIT
52 human
53 SET TCHNG=360, tstop=l,000
54 SET vchc=5000000000, points=48
55 SET Concppm=l
56 Start /nc
March 2011
B-72
DRAFT—DO NOT CITE OR QUOTE
-------
1 d concppm,aucBb,cvb
2 SET CONCppm=10
3 start /nc
4 d concppm,aucBb,cvb
5 SET CONCppm=50
6 start /nc
7 d concppm,aucBb,cvb
8 SET CONCppm=100
9 start /nc
10 d concppm,aucBb,cvb
11 SET CONCppm=250
12 start /nc
13 d concppm,aucBb,cvb
14 SET CONCppm=500
15 start/nc
16 d concppm,aucBb,cvb
17 SET CONCppm=625
18 start/nc
19 d concppm,aucBb,cvb
20 SET CONCppm=750
21 start/nc
22 d concppm,aucBb,cvb
23 SET CONCppm=875
24 start /nc
25 d concppm,aucBb,cvb
26 SET CONCppm= 1,000
27 start /nc
28 d concppm,aucBb,cvb
29 SET CONCppm=2000
30 start /nc
31 d concppm,aucBb,cvb
32 SET CONCppm=2500
33 start/nc
34 d concppm,aucBb,cvb
35 SET CONCppm=5000
36 start /nc
37 d concppm,aucBb,cvb
3 8 SET CONCppm= 1,0000
39 start/nc
40 d concppm,aucBb,cvb
41 SET CONCppm=50000
42 start /nc
43 d concppm,aucBb,cvb
44 END
45
46 Proced humor
47 ! To determine 24 hr AUC
48 ! Oral exposure
49 CLEARIT
50 human
51 SET TCHNG= 1,000, tstop= 1,000
52 SET vchc=5000000000, points=48
53 SETdose=0.1
54 SET ODS=1
55 start /nc
56 d dose,aucBb
March 2011
B-73
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
SETdose=l
Start /nc
d dose,aucBb
SET dose=5
start /nc
d dose,aucBb
SETdose=10
start /nc
d dose,aucBb
SET dose=50
start /nc
d dose,aucBb
SETdose=100
start /nc
d dose,aucBb
SET dose=250
start /nc
d dose,aucBb
SET dose=350
start /nc
d dose,aucBb
SET dose=500
start /nc
d dose,aucBb
SET dose=750
start /nc
d dose,aucBb
SET dose= 1,000
start /nc
d dose,aucBb
SET dose=2500
start /nc
d dose,aucBb
SET dose=5000
start /nc
d dose,aucBb
S ODS=0
END
[Procedural blocks for all non-pregnant rat data
!Not calibrated!!!!!!
!Procs from Ward CMD file
PROCED WARDIV25
CLEARIT
SDRAT
SET TSTOP=48.
SET IVDOSE=2500.
END
PROCED PWARDIV25
PLOT /D=WARDIV25, CVB
END
DATA WARDIV25(T,CVB)
0.072 4849
March 2011
B-74
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
0.168 3926
0.24 2965
0.504 2836
1.008 3248
1.992 2589
3 2619
4.008 2514
7.008 2315
19.992 1495
22.992 1272
24 1214
25.992 982
28.008 957
30 860
37.992 238
39 200
40.008 150
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
March 2011
B-75
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
o
6
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
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
1.992 1611
3 1869
4.008 1896
7.008 2181
24 1365
25.992 1081
28.008 921
30 958.4
31.008 969.8
45 42.9
46.008 27.1
46.992 16.4
48 23.9
49.008 41.9
49.992 13.1
52.008 2.3
52.992 1
END
PROCED WARDPO1
CLEARIT
cdmice
SETBW=0.3
SETDOSE=100,tstop=8
END
PROCED PWARDPO1
PLOT /D=WARDPO1, CVB
END
DATA WARDPO1(T,CVB)
0.072 85.5
0.168 95.6
0.24 95.5
0.504 91.1
0.744 86.6
1.008 80.6
1.488 71.3
1.992 61.1
3 45.1
4.008 27.4
4.992 16.4
6 8.9
7.008 4.2
END
March 2011
B-76
DRAFT—DO NOT CITE OR QUOTE
-------
B.3.3. Procedural .m files for reproducing the results in Appendix B and Chapter 3
B.3.3.1. Key to ACSL Extreme v2.5.0.6 .m files
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
4 6, 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. (1995a) (8 hr exposures) and GD 8 mice from Dorman et
7 al. (1995) (6 hr exposures)
8 Ward_mouse_GD18.m Figure B-4 - Oral exposures to MeOH in pregnant and non-
9 pregnant 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 Ward et
11 al. (1997)
12 Apaja-mouse-drink.m Calculates internal doses for mice in Apaja (1980)
13
14 SDratold.m Sets parameters for Sprague-Dawley (SD) rat simulations with parameters fit
15 when background is subtracted
16 SDrat.m Sets parameters for Sprague-Dawley (SD) rat simulations with parameters fit
17 when background is included
18 F344ratold.m Sets parameters for F344 rat simulations with parameters fit when background is
19 subtracted
20 F344rat.m Sets parameters for F344 rat simulations with parameters fit when background is
21 included
22 Ward-rat-iv.m Figure B-10 - Simulations rat IV exposures from Ward et al. (1997) and Horton et
23 al. (1992)
24 Horton-rat-inhal.m Figure B-l 1 - Simulations rat inhalation exposures from Horton et al.
25 (1992)
26 Ward-rat-oral.m Figure B-12 - Simulations rat oral exposures from Ward et al. (1997)
27 Nedo-rat-inhal-devpmt-rat.m Figure B-l3 - Simulations rat inhalation (bioassay) exposures
28 (200, 500, 1,000, 2000, & 5000 ppm)
29 Nedo-rat-inhal-cancer.m Simulations for NEDO F344 rat cancer inhalation study
30 rat-infu-sims.m Figure B-14 - Simulations rat "oral" exposures (bioassay doses, but using
31 liver infusion; for illustration only)
32
33 humanset.m Sets human MeOH PBPK parameters with endogenous/background included
March 2011 B-77 DRAFT—DO NOT CITE OR QUOTE
-------
1 humanold.m Sets human MeOH PBPK parameters when endogenous/background levels are
2 subtracted (included)
3 Sedivec_human_inh.m Figure B-16 - Simulation of human urinary MeOH elimination
4 following Inhalation exposures from Sedivec et al. (1981)
5 Batterman_human_inh.m Figure B-17 (upper panel) - Simulations of human inhalation
6 exposure data of Batterman et al. 1998 (1998)
7 Osterloh_human_inh.m Figure B-17 (lower panel) - Simulations of human inhalation
8 exposure data of Osterloh et al. (1996)
9 Ernstgard_human_inh.m Figure B-18 - Simulations of human inhalation exposures to
10 MeOH from Ernstgard et al. (2005)
11
12 mouse_inh_sim.m Produces data for Table B-5, mouse inhalation exposures
13 human_inh_sim.m Produces data for Table B-5, human inhalation expsoures
14 human_oral_sim.m Produces data for Table B-5, human oral exposures
15 human_drink_compare.m Figure B-24 and Table B-9 (altenate drinking pattern comparison)
16
17 A set of separate human data files is then provided; mouse and rat data are included in the
18 corresponding .m files.
19
20 Note: many the rat and human .m files include a "switch" parameter, "inclbg" (case-
21 sensitive), such that setting the value to zero (default) yields simulations and plots (with data) for
22 the analysis with background subtracted. When inclbg = 1 the results are for the analysis with
23 background included. Other rat and human files simply include a line which, when un-
24 commented ("%" at beginning is removed) the results include background. Brief comments near
25 the top of each file also explain these switches.
26
27 Found in the Sensitivity Analysis Files Folder
28 Fig_B-6 Sensitivity of the mouse model to metabolic parameters (e.g., Km and Vmax) for
29 the inhalation route
30 Fig_B-7 Sensitivity of the mouse model to flow parameters (e.g., blood flow to liver) and
31 to the rest-of-body partition coefficient for the inhalation route
32 Fig_B-8 Sensitivity analysis of the rat model to oral absorption parameters for a bolus oral
33 exposure (1,000 mg/kg)
B.3.3.2. Code for .m files
34 % File CDmice.m
35 % Sets parameters for mouse simulations, MeOH PBPK model
36 CONCPPM=10; WESITG=0; WEDITG=0; CINT-0.1;
March 2011 B-78 DRAFT—DO NOT CITE OR QUOTE
-------
1 start @nocallback
2 BW=0.03; TSTOP=24; TCHNG=7; REST=20000; WORK=20000;
3 IVDOSE=0; DOSE=0; DRDOSE=0; RATS=0; KLOSS=0; LIVRO=0;
4 PL=1.06; PF=0.083; PR=0.66; PLU=1; PB=1350;
5 QPC=25.4; QCC=25.4; FRACIN=0.73; KFEC=0;
6 QLC=0.25; QFC=0.05;
7 KM=12; VMAXC=14.3; K1C=0.0; KAS=0.0; KLLC=0;
8 VMAX2C=19; KM2=210; KAI=0.5; KSI=5.0;
9 VAC=0.0123; VFC=0.07; VLC=0.055; VLUC=0.0073; WBC=0.0368;
10 CONCPPM=0.0; IVDOSE=0.0; DOSE=0.0; DWDOSE=0; MULTE=0; RDRINK=1;
11 DCVBBG=0; CVBBG=1.6; RURO=0; INCBG=0;
12 % Volumes from Brown et al. (1997)
13 % Mouse QPC avg from Brown 29; 24 used in Corley et al. (1994) and others
14 % AVG of measured vent rates by Perkins et al. (1995a) 25.4 L/hr/kgA0.75
15 % Blood volume 4.9% total. As per Brown 25:75 split artven
16 % Metab originally from Ward et al - KldC for mice = 0
17
18 % use mouselNH_fit-params.m % File contents copied below
19 % Updated parameters as obtained by Paul Schlosser, U.S. EPA
20 % August 11, 2009 [this file updated]
21
22 % Values generated through parameter estimation script 'mouselNH_fit.m'
23 VMAX2C = 3.222500e+00; KM2 = 660; VMAXC = 19; KM = 5.2; FRACIN = 6.650939e-01;
24
25 % Values generated through parameter estimation script 'mouseor_fit.m'
26 VASC = 1.8332466+03; KSI = 2.2; KAI = 0.33; KMASC = 620;
27
28
29 % File Rogers_mouse_inhal.m (Figure B-2)
30 % Produces MeOH PBPK figures for Rogers' mouse inhalation exposures
31 % Variables in the plot command are case sensitive
32 use CDmice
33 % set mouse parameters
34 % DATA BLOCKS
35 % These data blocks taken directly from MeOH CBMMv3.cmd
36 % Data for are T (hours), CV (mg/L)
37 % semicolons (";") creates a new line in a data file
38
39 % Rogers et al. (1997),
40 D7IN10 = [1, 930; 4, 2800; 6, 3360; 7, 3990; 7.5, 3980;
41 8, 4120; 9, 3270; 12, 2630; 16, 1690; 26, 60];
42
43 % Rogers et al., (1993)
44 D6IN1 = [7, 63; 7, 131]; D6IN2 = [7, 487; 7, 641];
45 D6IN5 = [7, 2126; 7, 1593]; D6IN7p5 = [7, 2801; 7, 3455];
46 D6IN10 = [7, 4653; 7, 4304]; D6IN15 = [7, 7720; 7, 7394];
47
48 % RUN MODEL
49 RATS=0.0; KLOSS=0.0; % -> open chamber
50 TCHNG=7; CONCPPM=10000; TSTOP=27.0; MULTE=0; BW=0.032;
51 CINT=TSTOP/1000; cs=[]; prepare ©clear T CVB
52 forCONCPPM=[1,2, 5, 7.5, 10, 15]*1000
53 start @nocallback
54 cs=[cs,_cvb];
55 % Since TSTOP & CINT not changing, assume _t also the same.
March 2011 B-79 DRAFT—DO NOT CITE OR QUOTE
-------
1 end
2
3 % PLOT COM MAN DS
4 % The rogers.aps file will retain changes made using the plot
5 % editor as long as the editor is called by clicking the
6 % words EDIT PLOT PROPERTIES not the little icon in the
7 % properties dialogue box
8 plot(_t,cs(:,1), _t,cs(:,2), _t,cs(:,3), _t,cs(:,4), _t,cs(:,5), _t,cs(:,6), ...
9 D6IN1(:,1),D6IN1(:,2),D6IN2(:,1),D6IN2(:,2),D6IN5(:,1),D6IN5(:,2), ...
10 D6IN7p5(:,1),D6IN7p5(:,2),D6IN10(:,1),D6IN10(:,2), ...
11 D7IN10(:,1),D7IN10(:,2),D6IN15(:,1),D6IN15(:,2), 'rogers.aps')
12
13 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
14 % Cannot save data with different # of rows to the same table.
15 cs=[_t,cs];
16 save cs @file='Rogersplotdata.csv' @format=ASCII @separator=ascii
17
18
19 % File: PerkinsDorm-mouse-inh.m (Figure B-3)
20 % Produces MeOH PBPK simulations Perkins etal. (1995a) inhalation exposures,
21 % and Ward et al. (1997) (pregnant) and Dorman et al. (1995) for comparison)
22 % Includes all nonpregnant and "early" GD ( open chamber
55 MULTE=0; TSTOP=24; CONCPPM=2500; QPC = 29; QCC=29; TCHNG=8;
March 2011 B-80 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
26 %FileWardGD18.m
27 % Creates Figure B-4, including Ward et al. (1997) NP and GD 18 mouse data
28 % and Dorman et al. (1995) GD 8 mouse data.
29 CDmice
30 TSTOP=25; DOSE=1500; CONCPPM=0; MULTE=0;
31 prepare ©clear T CVB
32 start @nocallback
33 T1=_t;P1=_cvb;
34 DOSE=2500; start @nocallback
35 D15=[1, 1609.6; 2, 1331.2; 4, 1241.6;
36 8, 707.2; 16, 160; 24, 38.4];
37 D25a=[0.5, 2370; 0.96, 2645; 1.44, 2705; 2, 2719;
38 2.2, 2781; 3, 2704; 4, 2370; 5, 2617; 6, 2516;
39 7, 2635; 8, 2213; 9, 2370; 10, 2028; 11, 1916;
40 12, 1347; 13, 1467; 14, 1354; 15, 1175; 16, 864.3;
41 17, 745.2; 18, 422.4; 19, 428; 21, 243; 24, 136];
42 D25b=[0.5, 2024; 1, 2554; 2, 3193; 4, 3002; 6, 2933;
43 10, 1976; 12, 1922; 15, 1339; 18, 1033; 21, 832; 24, 580];
44
45 plot(D15(:,1),D15(:,2),D25a(:,1),D25a(:,2),D25b(:,1),D25b(:,2),...
46 T1,P1,_t,_cvb,"wardgd18plot.aps")
47 % File Ward-mouse-iv.m
48 % M File for reproducing MeOH PBPK Figure B-5 For WARD iv mouse exposures
49 % (also Ward Pregnant Includes all nonpregnant and Pregnant)
50
51
52 % DATA BLOCKS
53 %Taken directly from MeOH CBMMv3.cmd, values are [T (hours), CV (mg/L)]
54 %Wardetal. (1995)
55 NPIV25=[0.08, 4481.8; 0.25, 4132.2; 0.5, 3888; 1, 3164.8; 2, 2303.5;
March 2011 B-81 DRAFT—DO NOT CITE OR QUOTE
-------
1 4, 1921.5; 6, 1883.8; 8, 1620; 12, 838; 18, 454.7; 24, 1.41];
2 %PROCED MWARDGD8IV25
3 GD8IV25=[0.0833, 4606.2; 0.25, 4079.5; 0.5, 3489.3; 1, 2939.6; 2, 3447.6;
4 4, 2605.0; 6, 2690.5; 8, 2574.9; 12, 1506.1; 18, 498.6; 24, 0.554];
5 %!Ward etal. (1996)
6 GD18IV25=[0.0833, 4250.0; 0.25, 3445.1; 0.5, 2936.8; 1, 2470.5; 2, 2528.1;
7 4, 2292.3; 6, 2269.4; 8, 2057.0; 12, 1805.9; 18, 1482.2; 24.0, 496.1];
8 % Ward etal. (1996)
9 GD18IV5=[0.25, 854.7; 0.5, 720.2; 1, 624.1;
10 2, 453.2; 3, 307.6; 4, 217.7; 4.5, 202.6];
11 % Ward etal. (1996)
12 GD18IV1=[0.25, 252; 0.52, 242.2; 1, 222.7; 2, 176.4; 3, 134.2; 3.5, 94.41];
13
14 % RUN MODEL
15 useCDMICE
16 TSTOP=24.0; IVDOSE=2500; TCHNG=0.025; start @nocallback
17 CVs25 = _cvb; Ts25 = _t; TSTOP=6; IVDOSE=500; start @nocallback
18 CVs5 = _cvb; Ts5 = _t; TSTOP=4; IVDOSE=100; start @nocallback
19 CVs1 = _cvb; Ts1 = _t; IVDOSE=200; start @nocallback
20
21 % PLOT COMMANDS
22 % The .aps file will retain changes made using the plot
23 % editor as long as the editor is called by clicking the
24 % words EDIT PLOT PROPERTIES not the little icon in the
25 % properties dialogue box
26 plot(Ts25, CVs25, Ts5, CVs5, Ts1, CVs1, NPIV25(:,1), NPIV25(:,2),...
27 GD8IV25(:,1), GD8IV25(:,2), GD18IV25(:,1), GD18IV25(:,2), ...
28 GD18IV5(:,1), GD18IV5(:,2), GD18IV1(:,1), GD18IV1(:,2), 'iv.aps')
29 plot(_t, _cvb, Ts1, CVs1, GD18IV1(:,1), GD18IV1(:,2), 'ivb.aps')
30
31 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
32 % Cant save data with different # of rows to the same table.
33 mytablel = [Ts25, CVs25, Ts5, CVs5, _t, _cvb, Ts1, CVs1];
34 save mytablel @file='WardlV.csv' @format=ASCII @separator=comma
35
36
37 % File Apaja-mouse-drink.m
38 % Calculates internal doses for mice in Apaja (1980)
39 use CDmice
40 DWDOSE=1; start @nocallback
41 DWDS=[0.045, 550; 0.045, 970; 0.045, 1800;
42 0.040, 560; 0.040, 1000; 0.040. 2100];
43 % Above are BWs and doses for males, then females, from Apaja (1980)
44 ODS=1; TSTOP=24*3*7; MULTE=1; DAYS=6.0; simres=[]; LIVRO=0;
45 PER1=1.5; DUR1=0.75; PER2=3.0; DUR2=0.5; FNIGHT=0.8; CINT=0.01;
46 CVBBG=0; DCVBBG=0; INCBG=0;
47 prepare @clear T CVB STOM
48 for RDRINK =0 %[1, 0] % 0 -> mouse drinking pattern
49 forij=1:length(DWDS)
50 BW=DWDS(ij,1); DWDOSE=DWDS(ij,2); start @nocallback
51 simres=[simres;[TDOSE*(24/TSTOP)/BW,BW,AUCBF,max(_cvb),AMETF/(BWA0.75)]];
52 end
53 plot(_t,_cvb)
54 end
55 simres=[simres(:,1)*7/6, simres];
March 2011 B-82 DRAFT—DO NOT CITE OR QUOTE
-------
1 simres/100 % Print values to screen (/100)
2 TDOSE*(24/TSTOP)/BW % Check that final total dose/day is correct
3 save simres @file='Apaja_mouse_drink_sims.csv' @format=ascii @separator=comma
4
5
6 % File SDratold.m
7 % Sets parameters for rat simulations, MeOH PBPK model,
8 % parameters fit to data with backgrouind subtracted
9 CONCPPM=10; WESITG=0; WEDITG=0; TSTOP=24; TCHNG=6; MULTE=0;
10 DCVBBG=0; CVBBG=0; RURO=0; INCBG=0; REST=20000; WORK=20000;
11 start @nocallback
12 BW=0.275; TSTOP=24; FRACIN=0.2; WESITG=0; WEDITG=0;
13 IVDOSE=0; DOSE=0; CONCPPM=0; DRDOSE=0; DWDOSE=0; ODS=0; LIVRO=0;
14 QCC=16.4; QPC=16.4; QFC=0.07; QLC=0.25;
15 PL=1.06; PF=0.083; PR=0.66; PB=1350;
16 VAC=0.0185; VFC=0.07; VLC=0.037; VLUC=0.005; WBC=0.0443;
17
18 VMAXC = 5.0; KM = 6.3; VMAX2C = 8.4; KM2 = 65; KLLC=0.0; K1C=0.0;
19
20 % Below are linear absorption params fit to 100 mg/kg
21 % oral data, w/ no fecal elimination
22 KAS=10.9; KSI=6.8; KAI=0.039; KFEC=0.0; VASC=0;
23
24 % Below are for saturable uptake model
25 % Values generated through parameter estimation script 'ratoral_fit.m'
26
27 KSI = 7.4; KAI = 0.051; VASC = 5570; KMASC = 620; KAS=0.0;
28 KFEC = 0.029;
29
30
31 %FileSDrat.m
32 % Sets parameters for rat simulations, MeOH PBPK model,
33 % parameters fit to data with background included
34 CONCPPM=10; WESITG=0; WEDITG=0; TSTOP=24; TCHNG=6; MULTE=0;
35 DCVBBG=0; CVBBG=3; RURO=0; INCBG=0; REST=20000; WORK=20000;
36 start @nocallback
37 BW=0.275; TSTOP=24; FRACIN=0.2; WESITG=0; WEDITG=0;
38 IVDOSE=0; DOSE=0; CONCPPM=0; DRDOSE=0; DWDOSE=0; ODS=0; LIVRO=0;
39 QCC=16.4; QPC=16.4; QFC=0.07; QLC=0.25;
40 PL=1.06; PF=0.083; PR=0.66; PB=1350;
41 VAC=0.0185; VFC=0.07; VLC=0.037; VLUC=0.005; WBC=0.0443;
42
43 VMAXC = 9.9; KM = 2.8; VMAX2C = 9.1; KM2 = 60; KLLC=0.0; K1C=0.0;
44
45 % Below are linear absorption params fit to 100 mg/kg
46 % oral data, w/ no fecal elimination
47 KAS=10.9; KSI=6.8; KAI=0.039; KFEC=0.0; VASC=0;
48
49 % Below are for saturable uptake model
50 % Values generated through parameter estimation script 'ratoral_fit.m'
51
52 KSI = 7.2; KAI = 0.05; VASC = 5500; KMASC = 620; KAS=0.0;
53 KFEC = 2.875611e-02;
54
55
March 2011 B-83 DRAFT—DO NOT CITE OR QUOTE
-------
1 % File F344ratold.m: parameters specific to F344 rat, fitted to data with background subtracted
2 % Created by Paul Schlosser, U.S. EPA, Aug. 2009
3 use SDratold
4 VMAXC = 22.3; KM = 100; VMAX2C=0;
5 %use rat_fit-params
6
7
8 % File F344rat.m: parameters specific to F344 rat, fitted to data with background included
9 % Created by Paul Schlosser, U.S. EPA, Aug. 2009; revised Dec. 2009
10 use SDrat
11 VMAXC = 21.5; KM = 92.5; VMAX2C=0;
12
13
14 % File: Ward-rat-iv.m
15 % Creates Figure B-9; rat MeOH PBPK model, to simulate
16 % Ward '97 rat iv 2500 & 100 mg/kg (BW=275, SD)
17 % and Morton '92 iv 100 mg/kg (BW=100, F344)
18 rwi25 =[0.072, 4849; 0.168, 3926; 0.24, 2965; 0.5, 2836; 1, 3248;
19 2, 2589; 3, 2619; 4, 2514; 7, 2315; 20, 1495; 23, 1272; 24, 1214;
20 26, 982; 28, 957; 30, 860; 38, 238; 39, 200; 40, 150; 41, 167; 43, 77];
21
22 % ward '97 rat iv 100 mg/kg (BW 275, SD)
23 rwi1=[0.072, 141.7; 0.168, 121.8; 0.24, 111.6; 0.5, 99.7; 0.744, 97.4;
24 1, 86.3; 1.488, 80.3; 2, 58; 3, 44.4; 4, 22.8; 5, 10.9; 6, 3.8];
25
26 % horton '92 rat iv 100 mg/kg (BW 220, F344)
27 rtii1=[0.24, 91.13; 0.52, 80.14; 0.90, 70.29; 1.38, 61.22;
28 1.86, 60.63; 2.79, 39.40; 4.30, 26.05; 5.81, 12.51];
29
30 DCVBBG=0; CVBBG=0; inclbg=0; use SDratold
31 % Line above set simulations for data without background.
32 % Uncomment the following line to show results with background.
33 %inclbg=1; CVBBG=3; use SDrat
34
35 %use rat_fit-params
36 prepare ©clear T CVB
37
38 TSTOP=48; IVDOSE=2500; TCHNG=0.016; BW=0.275; CINT=0.1; start ©nocallback
39 Twi25 = _t; Cwi25 = _cvb;
40 TSTOP=24; CINT=0.05; IVDOSE=100; start ©nocallback
41 Twi1 = _t; Cwi1 = _cvb;
42 BW=0.22; CVBBG=0; DCVBBG=3*inclbg; start ©nocallback
43
44 plot(Twi25, Cwi25, Twi1, Cwi1, _t, _cvb, rwi25(:,1), rwi25(:,2)+DCVBBG, ...
45 rwi1(:,1), rwi1(:,2)+DCVBBG, rhi1(:,1), rhi1(:,2)+CVBBG, 'rwi2500.aps')
46
47 use F344rat
48 CVBBG=0; DCVBBG=3*inclbg;
49 IVDOSE=100; BW=0.22; TCHNG=0.016; CINT=0.05; start ©nocallback
50 plot(Twi1, Cwi1, _t, _cvb, ...
51 rwi1(:,1), rwi1(:,2)+DCVBBG, rhi1(:,1), rhi1(:,2)+CVBBG, 'rwMOO.aps')
52 n=min([length(_t),length(Twi25),length(Twi1)])
53 res=[Twi25(1:n), Cwi25(1:n), Twi1(1:n), Cwi1(1:n), _t(1:n), _cvb(1:n)];
54 save res @file='Ward-rat-iv-sim.csv' @format=ascii @separator=comma
55
March 2011 B-84 DRAFT—DO NOT CITE OR QUOTE
-------
1
2 % File: Horton-rat-inhal.m
3 % MeOH PBPK model rat simulations for Morton '92 rat inhalation data
4 % Creates Figure B-10
5 hi20=[0.46, 18.70; 1, 23.76; 3, 59.73; 6, 80.12
6 7, 83.25; 9, 53.49; 12, 16.54; 15, 0.91];
7 hi12=[0.46, 4.89; 1, 8.02; 3, 20.57; 6, 26.63;
8 7, 16.12; 8, 9.28; 9, 5.23; 10.5, 2.93; 12, 0.98];
9 hi2=[0.48, 1.2; 3, 3.1; 6, 3.7; 6.47, 2.7;
10 7, 2.0; 8, 1.6; 9, 1.2];
11
12 use F344ratold
13 %To see fits with SDrat parameters, uncomment the next line
14 %use SDratold
15
16 % F344ratold is for simulations without background.
17 % Uncomment next lines, depending on strain, for simulations with background.
18 %use F344rat
19 %use SDrat
20 %CVBBG=0; DCVBBG=3;
21
22 prepare @clear T CVB
23 TSTOP=16; CONCPPM=2000; TCHNG=6; BW=0.22; CINT=0.1; start ©nocallback
24 120 = _t; c20 = _cvb;
25 CONCPPM=1200; TSTOP=13; start ©nocallback
26 t12=_t;c12 = _cvb;
27 CONCPPM=200; TSTOP=10; start ©nocallback
28 12 = _t; c2 = _cvb;
29 TSTOP=16; CONCPPM=2000; TCHNG=7; start ©nocallback
30
31 plot(hi20(:,1), hi20(:,2), t20, c20, hi12(:,1), hi12(:,2), t12, c12, ...
32 hi2(:,1), hi2(:,2),t2, c2, _t, _cvb, 'hi2000.aps')
33
34
35 % File: Ward-rat-oral.m
36 % MeOH PBPK model rat simulations for Ward '97 rat oral data
37 % Creates Figre B-11
38 use SDRatold
39 % Below resets to linear absorption fits
40 KAS=10.9; KSI=6.8; KAI=0.039; KFEC=0.0; VASC=0;
41 %SDRatold is for parameters fit when backcground is excluded.
42 inclbg=0; %Set =1 for simulations with background included, 0 for excluded.
43 ifinclbg==1
44 use SDRat
45 % Then linear absorption parameters...
46 KAS=10.9; KSI=6.8; KAI=0.039; KFEC=0.0; VASC=0;
47 end
48 BW=0.3; ODS=0; prepare ©clear T CVB
49 DOSE=100; TSTOP=10; start ©nocallback
50 t1=_t;c1=_cvb;
51 DOSE=2500; TSTOP=36; start ©nocallback
52 t2=_t;c2=_cvb;
53
54 %Now simulate with saturable uptake parameters
55 use SDratold
March 2011 B-85 DRAFT—DO NOT CITE OR QUOTE
-------
1 ifinclbg==1
2 use SDRat
3 end
4
5 DOSE=100; TSTOP=10; start @nocallback
6 t1A=_t;c1A=_cvb;
7 DOSE=2500; TSTOP=36; start @nocallback
8 t2A=_t;c2A=_cvb;
9
10 d1=[0.072, 85.5; 0.168, 95.6; 0.24, 95.5; 0.504, 91.1;
11 0.744, 86.6; 1.008, 80.6; 1.488, 71.3; 1.992, 61.1;
12 3, 45.1; 4.008, 27.4; 4.992, 16.4; 6, 8.9; 7.008, 4.2];
13
14 d25=[0.072, 862.7; 0.168, 1243; 0.24, 1356; 0.504, 1621;
15 1.008, 1641; 1.992, 1611; 3, 1869; 4.008, 1896; 7.008, 2181;
16 24, 1365; 25.992, 1081; 28.008, 921; 30, 958.4; 31.008, 969.8];
17
18 plot(t2,c2, t2A,c2A, d25(:,1),d25(:,2)+CVBBG, ...
19 t1,c1, t1A,c1A, d1(:,1),d1(:,2)+CVBBG,"wardratoralplot.aps")
20 plot(t2,c2, t2A,c2A, d25(:,1),d25(:,2)+CVBBG, ...
21 t1,c1, t1A,c1A, d1(:,1),d1(:,2)+CVBBG,"wardratoralplotb.aps")
22
23
24 % File: Nedo-rat-inhal-devpmt.m
25 % MeOH PBPK model rat simulations for rat inhalation exposures
26 % 200, 500, 1000, 2000, and 5000 ppm
27 % Internal doses for NEDO developmental inhalation exposures, Sprague-Dawley rats
28 % Creates Figure B-13 ('simres' is tabulated results)
29 use SDRatold
30 % SDRatold is for simulations with no background;
31 % uncomment the following to include background
32 % use SDRat
33
34 prepare ©clear T CVB
35 simres=[]; ts=[]; cs=[]; ts2=[]; cs2=[]; TSTOP=24*2*7; CINT=1;
36 for CONCPPM=[0,200, 500, 1000, 2000, 5000]
37 TCHNG=22; MULTE=1; start @nocallback
38 res=[CONCPPM,BW,max(_cvb),AUCBF,AMETF]
39 ts=[ts,_t]; cs=[cs,_cvb];
40 TCHNG=TSTOP; MULTE=0; %CONCPPM=22*cp/24;
41 start @nocallback
42 simres=[simres;[res,CONCPPM,max(_cvb),AUCBF,AMETF/(BWA0.75)]]
43 ts2=[ts2,_t]; cs2=[cs2,_cvb];
44 end
45
46 simres
47 plot(ts(:,2),cs(:,2),ts(:,3),cs(:,3),ts(:,4),cs(:,4), ...
48 ts2(:,2),cs2(:,2),ts2(:,3),cs2(:,3),ts2(:,4),cs2(:,4), ...
49 [2424],[0,95],'fig13.aps')
50 save simres @file='Nedo_devpomt_rat_inhal_sims_old.csv' @format=ascii @separator=comma
51 cs=[ts(:,1),cs,cs2]; save cs @file='Fig13_sims.csv' @format=ascii @separator=comma
52
53
54 % File Nedo-rat-inhal-cancer.m
55 % Simulations for NEDO F344 rat cancer inhalation study
March 2011 B-86 DRAFT—DO NOT CITE OR QUOTE
-------
1 use F344ratold
2 % F344ratold is for results without background
3 % Uncomment the following to include background
4 % use F344rat
5 bg=0; TCHNG=22; TSTOP=5*7*24; MULTE=1;
6 res=[]; CONCPPM=200; prepare ©clear T CVB
7 start @nocallback
8 TCHNG=19.5; ODS=1; cppm=[0,10,100,1000]; DCVBBG=0; INCBG=0;
9 bwm=[422.1, 418.3, 417.7, 410.0]/1000;
10 bwf=[268.7, 270.6, 267.0, 264.9]/1000;
11 CVBBG=bg*3.31
12 for iJ=1:length(cppm)
13 CONCPPM=cppm(iJ); BW=bwm(iJ); start @nocallback
14 res=[res;[CONCPPM,TCHNG,BW,AUCBF,max(_cvb),AMETF/(BWA0.75)]]
15 end
16 CVBBG=bg*4.54;
17 for iJ=1:length(cppm)
18 CONCPPM=cppm(iJ); BW=bwf(iJ); start @nocallback
19 res=[res;[CONCPPM,TCHNG,BW,AUCBF,max(_cvb),AMETF/(BWA0.75)]]
20 end
21
22 save res @file='Nedo_rat_cancer_sims_new.csv' @format=ascii @separator=comma
23
24
25 % File: rat-infu-sims.m
26 % MeOH PBPK model rat simulations for zero-order liver infusions
27 % Creates Figre B-14
28 use SDRatold
29 % SDRatold is for simulations with no background;
30 % uncomment the following to include background
31 %useSDRat
32
33 Iv0=[0.33, 65.9; 0.33, 624.1; 0.34, 2177;
34 0.49, 53.2; 0.50, 524;
35 0.54 1780];
36 % Above are BWs and doses from Soffritti et al. (2002)
37 prepare ©clear T CVB
38 TCHNG=12; MULTE=0; simres=[]; ts=[]; cs=[];
39 fori=1:3
40 BW=lvO(i,1); LIVRO=lvO(i,2); TSTOP=24; start @nocallback
41 res=[LIVRO,BW,max(_cvb),0,AUCB,0,AMET];
42 TSTOP=84; start @nocallback
43 res(4)=AUCB; res(6)= AM ET; simres=[simres;res];
44 ts=[ts,_t]; cs=[cs,_cvb];
45 end
46 simres/100
47 plot(ts(:,1),cs(:,1),ts(:,2),cs(:,2),ts(:,3),cs(:,3),'fig14b.aps')
48 save simres @file='rat_liver-infusion_sims.csv' @format=ascii @separator=comma
49
50
51 % File: humanset.m
52 % Sets parameters for human simulations *with background levels included*.
53 % Expects the user to define metabf = "linear" to use 1st-order metabolism parameters;
54 % otherwise metabf set to "non-linear" and Michaelis-Menten parameters used.
55 WESITG=0; WEDITG=0;
March 2011 B-87 DRAFT—DO NOT CITE OR QUOTE
-------
1 BW = 70; IVDOSE=0; DOSE=0; CONCPPM=0; LIVRO=0;
2 RURO=0; RINCBG=0; CVBBG=0; DCVBBG=0; REST=3000; WORK=3000;
3 VCVBBG = 0.987; INCBG = 0; TSTOP = 24; FRACIN = 0.8655; FRACINW=1;
4 RATS=0; KLOSS=0; % constant exposure/no chamber losses
5 PB = 1626; PL = 0.583; PF = 0.142; PR = 0.805; PLU=1.07; %1.0;
6 VFC = 0.214; VLC = 0.026; VLUC = 0.008; VAC = 0.0198; WBC = 0.0593;
7 QPC = 24.0; QCC = 16.5; QLC = 0.227; QFC = 0.052;
8 QPCHW=52.6; QCCHW=26; QPCHR=QPC; QCCHR=QCC;
9 % Below are old values, replaced by file calls further down!
10 % Values for Michaelis-Menten liver metabolism
11 KLLC = 0.0; KBL=0.612; KM = 23.7; VMAXC = 33.1; K1C = 0.0231;
12 %linear liver metabolism optimum
13 KLLC = 60.7; VMAXC = 0; VMAX2C = 0; K1C = 0.0397;
14
15 % Mouse oral uptake KMASC; others set to match ethanol values
16 % for humans from Sultatos et al. (2004), with VASC set so that
17 % VASC/KMAS = 0.21/h, the Sultatos et al. (2004) 1st-order constant,
18 % and KFEC = 0 corresponding to assumed 100% absorption.
19 KSI = 3.17; KAI = 3.28; KMASC = 620; KFEC=0;VASC = 0.21*KMASC;
20 % From file human_fit_nonlin_bld.m
21 K1C = 3.154532e-02; KBL = 6.538270e-01; VMAXC = 1.352751e+02;
22 RINCBG = 1.190381e-01; VCVBBG = 9.848708e-01; KM = 1.219040e+02; KLLC=0;
23 % Below sets 'no bladder values; uncomment next lines to use
24 %values from human_fit_nonlin_nbl.m
25 % K1C = 2.800062e-02; KBL = 1 .OOOOOOe+03; VMAXC = 1.0859076+05;
26 % RINCBG = 1.2280906-01; VCVBBG = 9.851535e-01; KM = 1.088033e+05;
27
28 exist metabf; % check if metabf defined
29 if-ans % If not...
30 metabf = "non-linear"
31 end
32 if metabf=="linear"
33 % Below are optimal values for 1st-order liver metabolism
34 VMAXC=0.0; VMAX2C=0.0;
35 % From file human_fit_linear_bld.m
36 K1C = 3.073116e-02; KBL = 6.782894e-01; KLLC = 7.200736e+01;
37 RINCBG = 1.251028e-01; VCVBBG = 9.844745e-01;
38 else metabf="non-linear";
39 end
40 %use human_fit-params
41 disp(['Simulation for ',ctot(metabf),' human kinetics']);
42
43
44 % File: humanold.m
45 % Sets parameters for human simulations. Parameters are as set or optimized
46 % for the human model when background leves are subtracted (not included).
47 WESITG=0; WEDITG=0;
48 BW = 70; IVDOSE=0; DOSE=0; CONCPPM=0; LIVRO=0;
49 RURO=0; RINCBG=0; CVBBG=0; DCVBBG=0; REST=3000; WORK=3000;
50 VCVBBG = 0.987; INCBG = 0; TSTOP = 24; FRACIN = 0.8655; FRACINW=1;
51 RATS=0; KLOSS=0; % constant exposure/no chamber losses
52 PB = 1626; PL = 0.583; PF = 0.142; PR = 0.805; PLU=1.07; %1.0;
53 VFC = 0.214; VLC = 0.026; VLUC = 0.008; VAC = 0.0198; WBC = 0.0593;
54 QPC = 24.0; QCC = 16.5; QLC = 0.227; QFC = 0.052;
55 QPCHW=52.6; QCCHW=26; QPCHR=QPC; QCCHR=QCC;
March 2011 B-88 DRAFT—DO NOT CITE OR QUOTE
-------
1 % Below are old values, replaced by file calls further down!
2 % Values for Michaelis-Menten liver metabolism
3 KLLC = 0.0; KBL=0.612; KM = 23.7; VMAXC = 33.1; K1C = 0.0342;
4
5 % Rat oral uptake KMASC; others set to match ethanol values
6 % for humans from Sultatos et al. (2004), with VASC set so that
7 % VASC/KMAS = 0.21/h, the Sultatos et al. 1st-order constant,
8 % and KFEC = 0 corresponding to assumed 100% absorption.
9 KSI = 3.17; KAI = 3.28; KMASC = 620; KFEC=0;VASC = 0.21*KMASC;
10
11
12 % File: Sedivec_human_inh.m
13 % Creates MeOH PBPK Figure B-16
14 % For human inhalation exposures, w/ data of Sedivec et al
15
16 % DATA BLOCKS
17 % These data blocks taken directly from MeOH CBMMv3.cmd
18 % Data are T (hours), CV (mg/L), cumulative urinary clearance (mg)
19 % Rounded to 3-4 sig figs
20 % Sedivec et al. (1981), Int Arch Occ Health, urine
21 load @file=Sedv231 .csv @format=ascii;
22 load @file=Sedv157.csv@format=ascii;
23 load @file=Sedv78.csv@format=ascii;
24
25 % RUN MODEL
26 use humanold
27 % Humanold is without background
28 inclbg = 0; % Set to 1 to run with background included, 0 for excluded
29 ifinclbg==1
30 metabf="non-linear"; use humanset
31 end
32 TCHNG=8; TSTOP=24; CONCPPM=231; CVBBG=0; DCVBBG=0; INCBG=0;
33 RURO=inclbg*Sedv231(1,2);
34 prepare @clear T RUR URB
35 start @nocallback
36 ur1 = _urb; t1 = _t; cu1=_rur;, % Save time series for urine MeOHc
37 CONCPPM=157; RURO=inclbg*Sedv157(1,2);
38 start @nocallback
39 ur2 = _urb; t2 = _t; cu2= _rur;
40 CONCPPM=78; RURO=inclbg*Sedv78(1,2);
41 start @nocallback
42
43 % PLOT COM MAN DS
44 pj=5-inclbg*3; pk=pj+1;
45 plot(t1,ur1,t2,ur2,_t,_urb,Sedv231(:,1),Sedv231(:,pk),...
46 Sedv157(:,1),Sedv157(:,pk),Sedv78(:,1),Sedv78(:,pk),'sedivic.aps')
47 plot(t1,cu1,t2,cu2,_t,_rur,Sedv231(:,1),Sedv231(:,pj),...
48 Sedv157(:,1),Sedv157(:,pj),Sedv78(:,1),Sedv78(:,pj),'sedivic2.aps')
49
50 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
51 % Cant save data with different # of rows to the same table.
52 mytablel = [t1 ,ur1 ,cu1 ,t2,ur2,cu2,_t,_urb,_rur];
53 eval(['save mytablel @file=Sedv_fit_KLLC.',num2str(round(KLLC)), ...
54 '.csv @format=ascii @separator=comma']);
55
March 2011 B-89 DRAFT—DO NOT CITE OR QUOTE
-------
1
2 % File: Batterman_human_inh.m
3 % Creates MeOH PBPK Figure B-17 (upper panel)
4 % For human inhalation exposures of Batterman et al. (1998)
5
6 %These data blocks taken directly from MeOH CBMMv3.cmd
7 %Data are T (hours), CV (mg/L)
8 % Batterman et al. (1998)
9 load @file=Batt81 .csv @format=ascii;
10 load @file=Batt82.csv@format=ascii;
11 load @file=Batt830.csv@format=ascii;
12
13 use humanold
14 % Humanold is without background
15 inclbg = 0; % Set to 1 to run with background included, 0 for excluded
16 ifinclbg==1
17 metabf="non-linear"; use humanset
18 end
19 CVBBG=inclbg*1.77; DCVBBG=0; INCBG=0; RURO=0;
20 prepare @clear T CVB
21 TCHNG=2; CONCPPM=800; TSTOP=8.2; start @nocallback
22 t2=_t; c2=_cvb; TCHNG=1; start @nocallback
23 t1=_t; c1=_cvb; TCHNG=0.5; start @nocallback
24 t30=_t; c30=_cvb;
25
26 % PLOT COM MAN DS
27 pj=4-2*inclbg;
28 plot(t2,c2, t1,c1, t30,c30, Batt82(:,1),Batt82(:,pj), ...
29 Batt81(:,1),Batt81(:,pj),Batt830(:,1),Batt830(:,pj),'batterman.aps')
30
31 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
32 % Cant save data with different # of rows to the same table.
33 le=1 :min([length(t1),length(t2),length(t30)]);
34 mytablel = [t2(le),c2(le),t1(le),c1(le),t30(le),c30(le)];
35 eval(['save mytablel @file=Batter_fit_KLLC.',num2str(round(KLLC)),'.csv '...
36 '@format=ascii @separator=comma']);
37
38
39 % File: Osterloh_human_inh.m
40 % Creates Fig B-17 (lower panel)
41 % Data from Osterloh et al. (1996)
42 % Digitized data provided by EPA (datl)
43 % Subtracted background from exposure blood levels
44 % by Paul Schlosser, U.S. EPA
45 use humanold
46 % Humanold is without background
47 inclbg = 0; % Set to 1 to run with background included, 0 for excluded
48 ifinclbg==1
49 metabf="non-linear"; use humanset
50 end
51 BW=78.2;
52 load @file=Osterloh.csv @format=ascii; dat=Osterloh;
53 load @file=Osterloh_con.csv @format=ascii; datc=Osterloh_con;
54
55 prepare ©clear T CVB
March 2011 B-90 DRAFT—DO NOT CITE OR QUOTE
-------
1 TSTOP=8.2; INCBG=inclbg; CVBBG=0; DCVBBG=0; RURO=0; CONCPPM=0;
2 start @nocallback
3 pj=4-2*inclbg;
4 plot(_t,_cvb,datc(:,1),datc(:,pj), 'osterloh_con.aps')
5 TCHNG=4; CONCPPM=200; TSTOP=16; start @nocallback
6 plot(_t,_cvb,dat(:,1),dat(:,pj), 'osterloh.aps')
7 mytable 1 =[_t,_cvb];
8 eval(['save mytablel @file=Oster_fit_KLLC.',num2str(round(KLLC)),'.CSV '...
9 '@format=ascii @separator=comma']);
10
11
12 % File: Ernstgard_human_inh.m
13 % Creates MeOH PBPK Figure B-18, w/ data of Ernstgard et al (2005; 2005)
14 % For human inhalation exposures w/ exercise
15 o/0 DATA BLOCKS
16 %These data blocks taken directly from MeOH CBMMv3.cmd
17 %Data are T (hours), CV (mg/L)
18 % Ernstgard et al. (2005) SOT poster, 100 ppm & 200 ppm human
19 load @file=Ernst_con.csv @format=ascii; erne = Ernst_con
20 load @file=Ernst100.csv @format=ascii; ernl = ErnstlOO
21 load @file=Ernst200.csv @format=ascii; ern2 = Ernst200
22
23 % RUN MODEL
24 use humanold
25 % Humanold is without background
26 inclbg = 0; % Set to 1 to run with background included, 0 for excluded
27 ifinclbg==1
28 metabf="non-linear"; use humanset
29 end
30 QPCHR=QPC; QCCHR=QCC; REST=2.0; WORK=0.0;
31 TCHNG=2.0; CONCPPM=0; TSTOP=10.0; QPCHW=52.6; QCCHW=26.0;
32 VCVBBG = 0.505; RINCBG = 0.128; INCBG=inclbg; RURO=0; CVBBG=0; DCVBBG=0;
33 FRACINW=FRACIN;
34 %CVBBG=0.665;INCBG=0;
35 prepare T CVB OP QC
36 start @nocallback
37 cvc = _cvb; tc = _t; CONCPPM=100; start @nocallback
38 cv1 = _cvb; t1 = _t; CONCPPM=200; start @nocallback
39
40 % PLOT COM MAN DS
41 pj=3-inclbg;
42 plot(t1, cv1, _t, _cvb,ern1(:,1),ern1(:,pj),ern2(:,1),ern2(:,pj),...
43 tc,cvc,ernc(:,1),ernc(:,pj),'ernstgard.aps')
44 % WRITE OUT DATA TO A TEXT FILE FOR IMPORTING INTO EXCEL
45 % Cant save data with different # of rows to the same table.
46 mytablel = [t1, cv1, _t, _cvb];
47 eval(['save mytablel @file=Ernst_nofit_KLLC.',num2str(round(KLLC)),'.csv'...
48 '@format=ascii @separator=comma']);
49
50
51 % File: mouse_inh_sim.m
52 % Runs simulations for Table B-5, mouse internal-dose calculations
53 % from inhalation exposure, over the concentration range specified
54 % in the 'for' statement below.
55 % Results saved to file 'MouselnhalSims.csv'.
March 2011 B-91 DRAFT—DO NOT CITE OR QUOTE
-------
1 use CDMice
2 BW=0.03; TCHNG=7; % 7 hr/day exposures
3 TSTOP=240; MULTE=1; % Run for 10 days; multi-day exposure 'on'
4 RATS=0.0; KLOSS=0.0; % -> open chamber
5 prepare ©clear T CVB
6 CONCPPM=10000; CINT=0.02; start @nocallback
7 % plot(_t,_cvb) % uncomment to see/check that periodicity reached by TSTOP
8 inhres=Q; CINT=0.2;
9 forCONCPPM=[1, 10,50, 100,250,500, 1000,2000,5000, 10000]
10 start @nocallback
11 inhres=[inhres;[CONCPPM, AUCBB, max(_cvb), AMET24/(BWA0.75)]];
12 end
13 save inhres @file=MouselnhalSims.csv @format=ASCII @separator=comma
14
15
16 % File: human_inh_sim.m
17 % Runs simulations for Table B-5, human internal-dose calculations
18 % from inhalation exposure, over the concentration range specified
19 % in the 'for' statement below.
20 % Results saved to file 'HumanlnhSims_KLLC.#.csv', where # is
21 % value of KLLC used (0 if non-linear/Michaelis-Menten kinetics).
22 % If metab="linear", 1st order kinetics used; otherwise non-linear.
23 use humanold
24 % humanold is for simulation without background
25 % uncomment the following line to include background
26 %metabf="non-linear"; use humanset
27 WESITG=0; WEDITG=0; MULTE=0; CINT=1.0; RATS=0.0; KLOSS=0.0;
28 CONCPPM=0; TSTOP=1000; TCHNG=1000; DOSE=0; DWDOSE=0; ODS=1;
29 CVBBG=2; DCVBBG=0; INCBG=0;
30 prepare @clear T CVB STOM
31 start @nocallback
32 simres=[CONCPPM,AUCBF,max(_cvb),AMETF/(BWA0.75)];
33 forCONCPPM=[1, 10,50, 100,250,500, 1000,2000,5000, 10000]
34 start @nocallback
35 simres=[simres;[CONCPPM,AUCBF-simres(1,2),max(_cvb)-simres(1,3),(AMETF/(BWA0.75))-
36 simres(1,4)]]
37 end
38 disp(['Simulation for ',ctot(metabf),' human kinetics']);
39 eval(['save simres @file=Human_new_lnhSims_KLLC.',num2str(round(KLLC)), ...
40 '.csv @format=ascii @separator=comma']);
41
42
43 % File: human_oral_sim.m
44 % Runs simulations for Table B-5, human internal-dose calculations from
45 % oral exposure, over the exposure range specified in the 'for' statement below.
46 % Results saved to file 'Hum_DW_Sims_KLLC.#.csv', where # is value of KLLC
47 % used (0 if non-linear/Michaelis-Menten kinetics).
48 % If metab="linear", 1st order kinetics used; otherwise non-linear.
49 use humanold
50 % humanold is for simulation without background
51 % uncomment the following line to include background
52 %metabf="non-linear"; 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=0; ODS=1; DRDOSE=0;
55 CVBBG=0; DCVBBG=0; INCBG=0;
March 2011 B-92 DRAFT—DO NOT CITE OR QUOTE
-------
1 prepare @clear T CVB STOM
2 start @nocallback
3 simres=[DOSE,AUCBF,max(_cvb),AMETF/(BWA0.75)];
4 forDOSE=[0.1, 1, 10,30,41.66,70,90, 110, 130, 160,200:100:800]
5 % [403.4,496.4,563.5,730.5,40,65.8]
6 start @nocallback
7 simres=[simres;[DOSE,AUCBF-simres(1,2),max(_cvb)-simres(1,3), ...
8 (AMETF/(BWA0.75))-simres(1,4)]]
9 end
10 disp(['Simulation for ',ctot(metabf),' human kinetics']);
11 eval(['save simres @file=Hum_DW_Sims_old_KLLC.',num2str(round(KLLC)),'.csv'...
12 '@format=ascii @separator=comma']);
13
14
15 % file human_drink_compare.m
16 % creates Figure B-24 and Table B-9
17 % Created by Paul Schlosser, U.S. EPA, 8/26/09
18 use humanold
19 % humanold is for simulation without background
20 % uncomment the following line to include background
21 %metabf="non-linear"; use humanset
22 WESITG=0; WEDITG=0; MULTE=0; CINT=0.1; TSTOP=48; DOSE=0.1; ODS=1; DRDOSE=0;
23 prepare @clear T CVB
24 start @nocallback
25 T1=_t; C1=_cvb; DOSE=0; LIVRO=0.1; TCHNG=12; MULTE=1; start @nocallback
26 T2=_t; C2=_cvb; LIVRO=0; ODS=0; DOSE=0.1; start @nocallback
27 T3=_t; C3=_cvb; DOSE=0; DRDOSE=0.1; start @nocallback
28 plot(T1 ,C1 ,T2,C2,T3,C3,_t,_cvb,'humoralsim.aps')
29
30 tbl=[];metd = 1.0;
31 fordse=[0.1, 1.0, 10, 100, 250, 500]
32 row=[]; LIVRO=0; DOSE=0; DRDOSE=dse; start @nocallback
33 row=[dse,max(_cvb),AUCBF,AMETF];
34 LIVRO=dse; TCHNG=12; MULTE=1; DOSE=0; DRDOSE=0; start @nocallback
35 row=[row,max(_cvb),AUCBF,AMETF];
36 LIVRO=0; DOSE=dse; DRDOSE=0; start @nocallback
37 tbl=[tbl;[row,max(_cvb),AUCBF,AMETF]];
38 end
39 tbl
40
41
B.3.3.3. Human data files
42 The following are data (.csv) files called and used by the human simulation and plotting
43 .m files above. The file itself includes only the lines of numbers (and "NaN" entries), not the
44 title. The format is that the first number in each row is time (h), the next one or two numbers are
45 data with background included, and the last one or two entries (mostly separated by an "NaN"
46 entry) are data with background subtracted.
47
48 File Sedv231.csv
49 0,0.971429,NaN,NaN,NaN,NaN
50 2,4.32857,0.185499965,NaN,3.338093417,0.11683327
March 2011 B-93 DRAFT—DO NOT CITE OR QUOTE
-------
1 4,6.78571,0.574499765,NaN,5.776185833,0.435833043
2 6,8.4,1.105999615,NaN,7.37142825,0.895999536
3 8,9.62857,1.736999565,NaN,8.580950667,1.454332798
4 10,7.64286,2.341499615,NaN,6.576193083,1.98483283
5 12,4.32857,2.760499665,NaN,3.2428555,2.32849953
6 14,2.41429,2.996499765,NaN,1.309527917,2.48783295
7 19,1.48571,3.337749765,NaN,0.333328958,2.631582926
8 24,1.2,3.57274939,NaN,0,2.66074921
9
10 FileSedv157.csv
11 0,0.887072,NaN,NaN,NaN,NaN
12 2,3.09137,0.13924547,NaN,2.1852415,0.076483453
13 4,4.86647,0.41776987,NaN,3.941285,0.29091188
14 6,5.84042,0.79251102,NaN,4.8961785,0.600223103
15 8,6.67141,1.23042507,NaN,5.708112,0.97137327
16 10,5.34209,1.65089757,NaN,4.3597355,1.323747933
17 12,2.73962,1.93375742,NaN,1.738209,1.53717599
18 14,1.7966,2.09252512,NaN,0.7761325,1.625177943
19 19,1.29882,2.36337437,NaN,0.23071125,1.713276771
20 24,1.11575,2.574649245,NaN,0,1.733464005
21
22 File Sedv78.csv
23 0,0.786793,NaN,NaN,NaN,NaN
24 2,1.6869,0.086579255,NaN,0.881013917,0.030835487
25 4,2.4725,0.232158255,NaN,1.647520833,0.119334203
26 6,3.12944,0.428226155,NaN,2.28536775,0.256985304
27 8,3.41439,0.657260205,NaN,2.551224667,0.426266038
28 10,2.39752,0.860677055,NaN,1.515261583,0.568593057
29 12,1.60951,1.000923105,NaN,0.7081585,0.64641276
30 14,1.35082,1.104534655,NaN,0.430375417,0.686261447
31 19,1.06171,1.31563103,NaN,0.093532708,0.732103408
32 24,1.01591,1.49742278,NaN,0,0.74028752
33
34 File BattSlcsv
35 1,8.27,NaN,6.5
36 1.25,7.97,NaN,6.2
37 1.5,7.17,NaN,5.4
38 2,6.07,NaN,4.3
39 3,4.57,NaN,2.8
40 4,3.27,NaN,1.5
41 5,2.71,NaN,0.94
42 6,2.49,NaN,0.72
43 7,2.29,NaN,0.52
44 8,2,NaN,0.23
45
46 File Batt82.csv
47 2,15.37,NaN,13.6
48 2.25,15.17,NaN,13.4
49 2.5,13.77,NaN,12
50 3,11.37,NaN,9.6
51 4,8.17,NaN,6.4
52 5,5.87,NaN,4.1
53 6,4.37,NaN,2.6
54 7,3.57,NaN,1.8
55 8,3.17,NaN,1.4
March 2011 B-94 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
File Batt830.csv
0.5,6.37,NaN,4.6
0.75,6.47,NaN,4.7
1,6.67,NaN,4.9
1.5,5.27,NaN,3.5
2.5,3.97,NaN,2.2
3.5,2.77,NaN,1
4.5,2.29,NaN,0.52
5.5,2.28,NaN,0.51
6.5,2.24,NaN,0.47
7.5,2.45,NaN,0.68
File Osterloh.csv
0,1.2269,NaN,NaN
0.25,2.4329,NaN, 1.18275
0.5,2.7998,NaN, 1.5264
0.75,3.2444,NaN, 1.94775
1,3.393,NaN,2.0731
1.5,4.1073,NaN,2.7409
2,4.5307,NaN,3.1178
2.5,4.9542,NaN,3.4948
3,5.5037,NaN,3.9978
3.5,5.733,NaN,4.1806
4,6.0789,NaN,4.48
5,4.4815,NaN,2.7896
6,3.7279,NaN, 1.943
7,2.9842,NaN, 1.1063
8,2.6574,NaN,0.6865
File Osterloh con.csv
0,0.8778,NaN,NaN
0.25,0.9391, NaN,0.03805
0.5,0.9762,NaN,0.0519
0.75,0.9261, NaN,-0.02145
1,0.9778,NaN,0.007
1.5,1. 149,NaN,0.1317
2,1.1456,NaN,0.0818
2.5,1. 055,NaN,-0.0553
3,1.2358,NaN,0.079
3.5,1. 1549,NaN,-0.0484
4,1.268,NaN,0.0182
5,1.4844,NaN,0.1416
6,1.4389,NaN,0.0031
7,1.5971, NaN,0.0683
8,1.6003,NaN,-0.0215
File Enrst con.csv
0,0.67857516,0.18347516
0.25,0.29961 5652,-0.206359348
0.5,0.386661924,-0.130188076
1 . 1 829,0.52082301 6,-0.0257331 34
1 .4636,0.637390944,0.078624344
1.8365,0.672340176,0.097352426
2.4886,0.694447776,0.091093676
March 2011
B-95
DRAFT—DO NOT CITE OR QUOTE
-------
1 2.9175,0.297052452,-0.324958798
2 3.5442,0.717500556,0.068227856
3 4.0342,0.264528648,-0.406059052
4 5.0684,1.134052596,0.418477196
5 5.9994,1.099269972,0.343196072
6 9.1639,0.836064576,-0.057665074
7 13.0139,0.932450508,-0.128754142
8
9 File ErnstlOO.csv
10 0,NaN,NaN
11 0.2284,1.544299164,0.87
12 0.5108,2.214591984,1.5
13 1.0422,3.209667876,2.31
14 1.5104,4.01014242,3.24
15 1.9779,4.554252108,3.65
16 2.2262,4.139750628,3.52
17 2.5357,3.278457756,2.55
18 2.939,2.766900708,2.23
19 3.5604,2.436331212,1.59
20 4.0581,2.436331212,1.72
21 5.0523,1.970793216,0.41
22 6.0469,1.717036416,0.5
23 9.2474,0.958335624,0.12
24
25 File Enst200.csv
26 0,NaN,NaN
27 0.1992,2.056176612,1.63
28 0.5139,3.493798596,2.92
29 0.9524,5.396263308,4.76
30 1.5454,7.109137728,6.3
31 1.982,8.334667728,7.65
32 2.2298,7.109137728,6.2
33 2.509,6.259789368,5.49
34 2.9747,5.45378472,4.96
35 3.565,4.853393568,3.64
36 4.0619,4.228472592,3.43
37 5.0554,3.142319796,1.94
38 6.0487,2.238195852,1.03
39 8.9379,1.223002044,0.21
March 2011 B-96 DRAFT—DO NOT CITE OR QUOTE
-------
B.3.4. Personal Communication from Lena Ernstgard Regarding Human Exposures
Reported in the Ernstgard and Johanson, 2005 SOT Poster
1 From: Lena Ernstgard [Lena.Ernstgard@imm.ki.se]
2 Sent: Wednesday, March 23, 2005 12:39 AM
3 To: Poet, Torka S
4 Subject: RE: Human MeOH poster
5 Hi,
6 We measured the ventilation rate and they ought to be similar to those reported by Dr. Johanson at the same
7 workload.
8 Sincerly,
9 Lena Ernstgard
10
11 At 18:41 2005-03-22, you wrote:
12
13 Thank you very much. Your net uptake is what we thought. Did you measure ventilation rates?
14
15 Thanks again,
16 Torka
17
18 Torka Poet, PhD
19 Center for Biological Monitoring and Modeling
20 Pacific Northwest National Laboratories
21 902 Battelle Blvd.
22 P.O. Box 999, MSIN P7-59
23 Richland, WA 99352
24 ph: (509)376-7740
25 fax: (509)376-9449
26 e-mail: Torka.poet@pnl.gov
27 (Express Mail Delivery: 790 Sixth Street, Zip Code 99354)
28
29 From: Lena Ernstgard [mailto:Lena.Ernstgard@imm.ki.se1
30 Sent: Sunday, March 20, 2005 11:21 PM
31 To: Poet, Torka S
32 Subject: Re: Human MeOH poster
33 Hi,
34 The manuscript has not been submitted yet, but it will be soon I hope. I will save your mail and send you a
35 copy as soon as possible.
36 When I say % of net uptake, i mean the relative uptake. It is calculated as: cone in exposure chamber -
37 (minus) exhaled cone / (divided by) cone in exposure chamber. I hope you understand how we have done.
38 Sincerly,
39 Lena Ernstgard
March 2011 B-97 DRAFT—DO NOT CITE OR QUOTE
-------
B.3.5. Total MeOH Metabolism/Metabolites Produced
Table B-6. Mouse total MeOH metabolism/metabolites produced following
inhalation exposures"
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
"Total over a 36-hour period during which mice were exposed for 7 hours to MeOH according to the conditions of
the dose-response study.
March 2011
B-98
DRAFT—DO NOT CITE OR QUOTE
-------
Table B-7. Human total MeOH metabolism/metabolites produced from
inhalation exposures"
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 metabolism/metabolites produced following
oral exposures"
Exposure concentration (mg/kg.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.3.6. 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 regardless of the oral uptake rate constants (assumed equal to the mouse),
March 2011
B-99
DRAFT—DO NOT CITE OR QUOTE
-------
1 hence avoids the difficulty that independent values of these constants are not available for
2 humans due to a lack of human oral PK data.. A more likely drinking scenario was tested by
3 using additional code within the model to simulate a 6-times/day drinking schedule, over the
4 course of 15 hours (see code below). The schedule is still an approximation, as it assumes 6
5 episodes of drinking, each considered to be a bolus. Specifically, it was assumed that humans
6 drank at 0, 3, 5, 8, 11, and 15 hours from the first ingestion of each day, with the respective
7 fractions of daily consumption being 25, 10, 25, 10, 25, and 5% at those times. The predicted
8 blood concentrations resulting from simulations of six daily boluses, once/day boluses, 12 h/d
9 infusion (zero order), or constant (zero order) are shown in Figure B-24 for a total dose of 0.1
10 mg/kg. Table B-9 shows PBPK model predicted Cmax, AUC, and Amet (for the last 24 hours of
11 repeated exposures) for humans exposed to MeOH via six daily boluses, 12 hour/day infusion, or
12 a single daily gavage.
0.07-
0.06-
0.05-
C
"
-------
Table B-9. Repeated daily oral dosing of humans with MeOH>
DOSE
(MG/KG)
0.1
1
10
100
250
500
SIX DAILY BOLUSES
Cmax
(mg/L)
0.0197
0.198
2.07
30.7
133
583
AUC (mg-
h/L)
0.0472
0.474
4.97
79.8
448
2.424
Amet
(mg)
1.98
19.8
198
1,970
4,810
7,290
12 HR/D AY INFUSION
Cmax
(mg/L)
0.0138
0.138
1.45
26.5
162
649
AUC (mg-
h/L)
0.0472
0.474
4.94
82.1
500
2,530
Amet
(mg)
1.98
19.8
198
1,970
4,790
7,270
SINGLE-DAILY BOLUS
Cmax
(mg/L)
0.0657
0.667
7.67
126
359
878
AUC (mg-
h/L)
0.05472
0.481
5.75
169
878
3,350
Amet
(mg)
1.98
19.8
198
1,930
4,640
7,410
*AUC in blood and Amet (amount metabolized) computed from 24-48 hr
March 2011
B-101
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 2.1.1 (U.S. EPA, 2009a), were fit to datasets for decreased brain weight in male
4 rats exposed throughout gestation and the postnatal period to 6 weeks and male rats exposed
5 during gestation on days 7-17 only (NEDO, 1987). Although there remains uncertainty
6 surrounding the identification of the proximate teratogen of importance (methanol,
7 formaldehyde, or formate), the dose metrics chosen for the derivation of RfCs were based on
8 blood methanol levels. This decision was primarily based on evidence that the toxic moiety is
9 not likely to be the formate metabolite of methanol (CERHR, 2004), and evidence that levels of
10 the formaldehyde metabolite following methanol maternal and/or neonate exposure would be
11 much lower in the fetus and neonate than in adults. While recent in vitro evidence indicates that
12 formaldehyde is more embryotoxic than methanol and formate, the high reactivity of
13 formaldehyde would significantly limit its transport from maternal to fetal blood, and the
14 capacity for the metabolism of methanol to formaldehyde is lower in the fetus and neonate versus
15 adults.
C.I.I. Decreased Brain Weight in Male Rats Exposed throughout Gestation and into the
Postnatal Period
16 The results of NEDO (1987), shown in Table 4-10, 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
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 C-1 DRAFT—DO NOT CITE OR QUOTE
-------
1 significant at the/? < 0.01 level). The extent to which this observation is due to recovery in rats
2 for which exposure was discontinued at birth versus a cumulative effect in rats exposed
3 postnatally is not clear. The fact that male rats exposed to 5,000 ppm methanol only during
4 organogenesis experienced a decrease of brain weight of 10% at 8 weeks postnatal indicates that
5 postnatal exposure is not necessary for the observation of persistent postnatal effects. However,
6 the fact that this decrease was less than the 13% decrease observed in male rats exposed to
7 2,000 ppm methanol throughout gestation, and the 8 week postnatal period indicates that the
8 absence of postnatal exposure allows for some measure of recovery.
9 It appears that once methanol exposure is discontinued, continuous biological processes
10 that are disrupted by exposure, manifesting as decreased brain weight, undergo some recovery
11 and brain weights begin to return to normal values. This indicates that brain weight is
12 susceptible to both the level and duration of exposure. Therefore, a dose metric that incorporates
13 a time component would be the most appropriate metric to use. For these reasons and because it
14 is more typically used in internal-dose-based assessments and better reflects total exposure within
15 a given day, daily AUC (measured for 22-hour exposure/day) was chosen as the most appropriate
16 dose metric for modeling the effects of methanol exposure on brain weights in rats exposed
17 throughout gestation and continuing into the FI generation.
18 Application of the EPA methanol PBPK model (described in Section 3.4) to the NEDO
19 (1987) study in which developing rats were exposed during gestation and the postnatal period
20 presents complications that need to be discussed. The neonatal rats in this study were exposed to
21 methanol gestationally before parturition, as well as lactationally and inhalationally after
22 parturition. The PBPK model developed by the EPA only estimates internal dose metrics for
23 methanol exposure in NP adult mice and rats. Experimental data indicate that inhalation-route
24 blood methanol kinetics in NP mice and pregnant mice on GD6-GD10 are similar (Dorman et al.,
25 1995: Perkins et al.. 1995a. 1995b: J. M. Rogers. Barbee. et al.. 1993: J. M. Rogers. Mole, et al..
26 1993). In addition, experimental data indicate that the maternal blood:fetal partition coefficient
27 for mice is approximately 1 (see Section 3.4.1.2). Assuming that these findings apply for rats,
28 they indicate that pharmacokinetic and blood dose metrics for NP rats are appropriate surrogates
29 for fetal exposure during early gestation. However, as is discussed to a greater extent in Section
30 5.3, the additional routes of exposure presented to the pups in this study (lactation and inhalation)
31 present uncertainties that make it reasonable to assume that average blood levels in pups in the
32 NEDO report are also greater than those of the dam. However, it is also reasonable to assume
33 that any differences seen between the pups and dams would also be seen in mothers and human
34 offspring. Therefore, the presumed differences between pup and dam blood methanol levels are
March 2011 C-2 DRAFT—DO NOT CITE OR QUOTE
-------
1 deemed relatively inconsequential, and the PBPK model-estimated adult blood methanol levels
2 are assumed to be appropriate dose metrics for the purpose of this analysis.
3 The first step in the current analysis is to convert the inhalation doses, given as ppm
4 values from the studies, to an internal dose surrogate or dose metric using the EPA PBPK model
5 (see Section 3.4). Predicted AUC values for methanol in the blood of rats and humans are
6 summarized in Table C-l. These AUC values are then used as the dose metric for the BMD
7 analysis of response data shown in Table C-l for decreased brain weight at 6 weeks in male rats
8 following gestational and postnatal exposure.59 Decreases in brain weight at 6 weeks (gestational
9 and postnatal exposure), rather than those seen at 3 and 8 weeks, were chosen as the basis for the
10 RfC derivation because they resulted in lower estimated BMDs and BMDLs. The details of this
11 analysis are reported below. More details concerning the PBPK modeling were presented in
12 Section 3.4.
13
14
15
16
17
18
Table C-l. The EPA PBPK model estimates of methanol blood levels (AUC) in
rat dams following inhalation exposures and reported brain weights of
6-week old male pups.
Exposure level
(ppm)
50
1,00
2,00
Methanol in blood AUC
(hr x mg/L)A in Rats
(
79.
226.
966.
Mean male rat (Ft generation) brain weight at
6 weeks8
1.78 ±0.07
1.74 ±0.09
1.69±0.06C
1.52±0.07d
aAUC values were obtained by simulating 22 hr/day exposures for 5 days and calculated for the last
24 hours of that period.
bExposed throughout gestation and Fj generation. Values are means ± S.D.
cp < 0.01, dp < 0.001, as calculated by the authors.
Source: NEDO (1987).
The current BMD technical guidance (2000a) suggests that in the absence of knowledge
as to what level of response to consider adverse, a change in the mean equal to 1 control S.D.
from the control mean can be used as a BMR for continuous endpoints. However, it has been
suggested that other BMRs, such as 5% change relative to estimated control mean, are also
appropriate when performing BMD analyses on fetal weight change as a developmental endpoint
All BMD assessments in this review were performed using BMDS version 2.1.1 (U.S. EPA. 2009a).
March 2011
C-3
DRAFT—DO NOT CITE OR QUOTE
-------
1 (Kavlock et al., 1995). Therefore, in this assessment, both a 1 control mean S.D. change and a
2 5% change relative to estimated control mean were considered. All models were fit using
3 restrictions and option settings suggested in the EPA BMD Technical Guidance Document
4 (2000a).
C. 1.1.1. BMD Approach with a BMR of 1 Control Mean S.D. - Gestation and into the
Postnatal Period
5 A summary of the results most relevant to the development of a POD using the BMD
6 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 6 weeks in male
7 rats exposed to methanol throughout gestation and continuing into the FI generation, with a
8 BMR of 1 control mean S.D (NEDO, 1987), is provided in Table C-2. The 6 week male brain
9 weight responses were chosen because they resulted in lower BMD and BMDL estimates than
10 male responses at 3 and 8 weeks and female responses at any time point (data not shown). Model
11 fit and was determined by statistics (AIC and $ residuals of individual dose groups) and visual
12 inspection, as recommended by EPA (U.S. EPA, 2000a). There is a 2.5-fold range of BMDL
13 estimates from adequately fitting models, indicating considerable model dependence. In
14 addition, the fit of the Hill and more complex Exponential models is better than the other models
15 in the dose region of interest as indicated by a lower scaled residual at the dose group closest to
16 the BMD (0.09 versus -0.67 or -0.77) and visual inspection. In accordance with EPA BMD
17 Technical Guidance (2000a), the BMDL from the Hill model (bolded), is selected as the most
18 appropriate basis for an RfC derivation because it results in the lowest BMDL from among a
19 broad range of BMDLs and provides a superior fit in the low dose region nearest the BMD. The
20 detailed results of the Hill model run, including text and plot (Figure C-l) are shown after Table
21 C-2. The BMDLiso was determined to be 90.9 hr x mg/L using the 95% lower confidence limit
22 of the dose-response curve expressed in terms of the AUC for methanol in blood.
March 2011 C-4 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
277.75
277.75
277.75
277.75
170.43
260.42
260.42
171.95
171.95
BMDL1SD
(AUC,
hr x mg/L)A
224.85
224.85
224.85
224.85
90.86
208.68
208.68
96.85
96.85
/rvalue
0.5387
0.5387
0.5387
0.5387
0.836
0.613
0.613
0.82
0.82
AICC
-203.84
-203.84
-203.84
-203.84
-203.04
-204.10
-204.10
-203.03
-203.03
Scaled residual0
-0.77
-0.77
-0.77
-0.77
0.09
-0.67
-0.67
0.09
0.09
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
aThe BMDL is the 95% lower confidence limit on the AUC estimated to decrease brain weight by 1 control mean
S.D. using BMDS 2.1.1 (U.S. EPA. 2009a) and model options and restrictions suggested by EPA BMD technical
guidance (U.S. EPA. 2000a).
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 (U.S. EPA. 2000a). 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).
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hil_m-6wk-brw_Hil-
Restrict.(d)
Gnuplot Plotting File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hil_m-6wk-
brw_Hil-Restrict.pit
Sat Dec 26 23:15:13 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
Power parameter restricted to be greater than 1
A constant variance model is fit
March 2011
C-5
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
Default Initial Parameter Values
alpha = 0.00539333
rho = 0 Specified
intercept = 1.78
v = -0.26
n = 1.08342
k = 400.5
Asymptotic Correlation Matrix of Parameter Estimates
the user,
alpha
intercept
v
k
( *** 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 )
alpha
1
6.7e-009
-3.6e-008
1.5e-008
intercept
6.7e-009
1
0.54
-0.64
v
-3.6e-008
0.54
1
-0.99
1.5e-008
-0.64
-0.99
1
Interval
Variable
Limit
alpha
0.00692039
intercept
1.81447
v
0.0679677
n
k
3876.35
Estimate
0.0049574
1.77822
-0.601684
1
1286.01
Parameter Estimates
Std. Err.
0.00100155
0.0184934
0.341665
NA
1321.63
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
0.0029944
1.74197
-1.27134
-1304.34
NA - Indicates that this parameter has hit a bound
implied by some ineguality 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
79.1
226.5
966
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
0.0876
-0.165
0.0887
-0.00679
March 2011
C-6
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
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)}s = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
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
Al
A2
A3
fitted
R
Log (likelihood)
105.539862
106.570724
105.539862
105.518430
77.428662
# Param's
5
8
5
4
2
AIC
-201.079724
-197.141449
-201.079724
-203.036861
-150.857324
Test 1:
Test
Test
Test
Explanation of Tests
Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Are Variances Homogeneous? (Al vs A2)
Are variances adeguately modeled? (A2 vs. A3)
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
58.2841
2.06173
2.06173
0.042863
p-value
<.0001
0.5597
0.5597
0.836
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.
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
March 2011
C-7
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
Benchmark Dose Computation
Specified effect = 1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 170.432
BMDL = 90.8618
Hill Model with 0.95 Confidence Level
c
o
Q.
(/)
0)
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).
11 Once the BMDLiso was obtained in units of hr x mg/L, it was used to derive a chronic
12 inhalation reference value. The first step is to calculate the HEC using the PBPK model
13 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
March 2011
DRAFT—DO NOT CITE OR QUOTE
-------
1 describes the relationship between predicted methanol AUC and the human equivalent inhalation
2 exposure concentration (HEC) in ppm.
3 BMDLHEC (ppm)= 0.02525*BMDLisD+(1290*BMDLisD)/(765.5 + BMDLiSD)
4 BMDLHEC (ppm)= 0.02525*90.9+(1290*90.9)/(765.5+ 90.9) = 139 ppm
5 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm
6 was converted using the conversion factor specific to methanol of 1 ppm = 1.31 mg/m3:
7 HEC (mg/m3) =1.31 x 139 ppm = 182 mg/m3
8 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
9 associated with animal to human differences, 10 for consideration of human variability, and 3 for
10 database deficiencies) to obtain the chronic inhalation reference value:
11 RfC (mg/m3) = 182 mg/m3 -100 = 1.8 mg/m3
C.I.1.2. BMD Approach with a BMR of 0.05 Change Relative to Estimated Control Mean
- Gestation and into the Postnatal Period (NEDO, 1987)
12 A summary of the results most relevant to the development of a POD using the BMD
13 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 6 weeks in male
14 rats exposed to methanol throughout gestation and continuing into the FI generation, with a
15 BMR of 0.05 change relative to estimated control mean, is provided in Table C-3. The 6 week
16 male brain weight responses were chosen because they resulted in lower BMD and BMDL
17 estimates than male responses at 3 and 8 weeks and female responses at any time point (data not
18 shown). Model fit was determined by statistics (AIC and ^ residuals of individual dose groups)
19 and visual inspection, as recommended by the EPA BMD Technical Guidance (U.S. EPA, 2000a).
20 There is a 2.4-fold range of BMDL estimates from adequately fitting models, indicating
21 considerable model dependence. In addition, the fit of the Hill and more complex Exponential
22 models is better than the other models in the dose region of interest as indicated by a lower
23 scaled residual at the dose group closest to the BMD (0.09 versus -0.67 or -0.77) and visual
24 inspection. In accordance with EPA BMD Technical Guidance (U.S. EPA, 2000a), the BMDL
25 from the Hill model (bolded), is selected as the most appropriate basis for an RfC derivation
26 because it results in the lowest BMDL from among a broad range of BMDLs and provides a
27 superior fit in the low dose region nearest the BMD. Output from the hill model, including text
28 and plot (Figure C-2), is shown after Table C-3. The BMDL05 was determined to be 123.9 hr x
March 2011 C-9 DRAFT—DO NOT CITE OR QUOTE
-------
1 mg/L, using the 95% lower confidence limit of the dose-response curve expressed in terms of the
2 AUC for methanol in blood.
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
BMDos (AUC,
hr x mg/L)A
343.82
343.82
343.82
343.82
222.98
BMDLos (AUC,
hr x mg/L)A
297.35
297.35
297.35
297.35
123.77
/rvalue
0.5387
0.5387
0.5387
0.5387
0.836
AICC
-203.84
-203.84
-203.84
-203.84
-203.04
Scaled Residual"
-0.77
-0.77
-0.77
-0.77
-0.09
Exponential 2 325.20 277.72 0.613 -204.10 -0.67
Exponentials 325.20 277.72 0.613 -204.10 -0.67
Exponential 4 223.74 129.86 0.82 -203.03 0.09
Exponentials 223.74 129.86 0.82 -203.03 0.09
aThe BMDL is the 95% lower confidence limit on the AUC estimated to decrease brain weight by 5% using BMDS
2.1.1 (U.S. EPA 2009a) and model options and restrictions suggested by EPABMD Technical Guidance (U.S.
EPA. 2000a).
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 (U.S. EPA. 2000a). 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)
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hil_m-6wk-brw_Hil-
ConstantVariance-BMROS-Restrict.(d)
Gnuplot Plotting File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-Fl\hil_m-6wk-
brw_Hil-ConstantVariance-BMR05-Restrict.pit
Sat Dec 26 23:05:11 2009
BMDS Model Run
The form of the response function is:
Y[dose] = intercept + v*dose/xn/(k^n + dose^n)
March 2011
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
63
64
65
66
67
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
Default Initial Parameter Values
alpha = 0.00539333
rho = 0 Specified
intercept = 1.78
v = -0.26
n = 1.08342
k = 400.5
Asymptotic Correlation Matrix of Parameter Estimates
the user,
alpha
intercept
v
k
( *** 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 )
alpha
1
6.7e-009
-3.6e-008
1.5e-008
intercept
6.7e-009
1
0.54
-0.64
v
-3.6e-008
0.54
1
-0.99
1.5e-008
-0.64
-0.99
1
Interval
Variable
Limit
alpha
0.00692039
intercept
1.81447
v
0.0679677
n
k
3876.35
Estimate
0.0049574
1.77822
-0.601684
1
1286.01
Parameter Estimates
Std. Err.
0.00100155
0.0184934
0.341665
NA
1321.63
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
0.0029944
1.74197
-1.27134
-1304.34
NA - Indicates that this parameter has hit a bound
implied by some ineguality constraint and thus
has no standard error.
Table of Data and Estimated Values of Interest
March 2011
C-ll
DRAFT—DO NOT CITE OR QUOTE
-------
1
9
^
3
4
5
6
7
Dose
0
79.1
226.5
966
N
12
12
11
14
Obs Mean
1.78
1.74
1.69
1.52
Est Mean
Obs Std Dev Est Std Dev Scaled Res.
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
1.78
1.74
1.69
1.52
0.07
0.09
0.06
0.07
0.0704
0.0704
0.0704
0.0704
0.0876
-0.165
0.0887
-0.00679
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Signal
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
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
Al
A2
A3
fitted
R
Log (likelihood)
105.539862
106.570724
105.539862
105.518430
77.428662
# Param's
5
8
5
4
2
AIC
-201.079724
-197.141449
-201.079724
-203.036861
-150.857324
Test 1:
Test
Test
Test
Explanation of Tests
Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Are Variances Homogeneous? (Al vs A2)
Are variances adeguately modeled? (A2 vs. A3)
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
58.2841
2.06173
2.06173
0.042863
p-value
<.0001
0.5597
0.5597
0.836
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.
to be appropriate here
The modeled variance appears
March 2011
C-12
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
The p-value for Test 4 is greater than .1. The model chosen seems
to adequately describe the data
Benchmark Dose Computation
Specified effect = 0.05
Risk Type = Relative risk
Confidence level = 0.95
BMD = 222.984
BMDL = 123.773
Hill Model with 0.95 Confidence Level
c
o
Q.
(/)
0)
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).
16 Once the BMDL0s was obtained in units of hr x mg/L, it was used to derive a chronic
17 inhalation reference value. The first step is to calculate the HEC using the PBPK model
March 2011
C-13
DRAFT—DO NOT CITE OR QUOTE
-------
1 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
2 describes the relationship between predicted methanol AUC and the human equivalent inhalation
3 exposure concentration (HEC) in ppm.
4 BMDLHEC (ppm)= 0.02525*BMDL05+(1290*BMDLo5)/(765.5 + BMDL05)
5 BMDLHEC (ppm)= 0.02525*123.77+(1290*123.77)7(765.5+ 123.77) = 183 ppm
6 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
7 converted using the conversion factor specific to methanol of 1 ppm = 1.31 mg/m3:
HEC (mg/m3) = 1.31 x 183 ppm = 240 mg/m3
9 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
10 associated with animal to human differences, 10 for consideration of human variability, and 3 for
11 database deficiencies) to obtain the chronic inhalation reference value:
12 RfC (mg/m3) = 240 mg/m3 - 100 = 2.4 mg/m3
C.1.2. Decreased Brain Weight in Male Rats Exposed During Gestation Only (GD7-GD17)
13 Cmax, as calculated by the EPAs PBPK model, was selected as the dose metric for this
14 exposure scenario, in concordance with the choice of this dose metric for the increased incidence
15 of cervical rib in mice in the Rogers et al. (J. M. Rogers, Mole, et al., 1993). Exposures occurred
16 only during the major period of organogenesis in both studies. As there is evidence that Cmax is a
17 better predictor of response than AUC for incidence of cervical rib (see Appendix D), it was
18 assumed appropriate to consider Cmax the better predictor for decreased brain weight as well.
19 The first step in the current analysis is to convert the inhalation doses, given as ppm
20 values from the studies, to an internal dose surrogate or dose metric using the EPA PBPK model
21 (see Section 3.4). Predicted Cmax values for methanol in the blood of rats are summarized in
22 Table C-4.
March 2011 C-14 DRAFT—DO NOT CITE OR QUOTE
-------
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
1 The current BMD technical guidance (U.S. EPA, 2000a) 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, are
5 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's BMD Technical Guidance
9 Document (U.S. EPA. 2000a).
C. 1.2.1. BMD Approach with a BMR ofl Control Mean S.D. (GD7-GD17)
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) (NEDO. 1987) for decreased brain weight at 8
12 weeks in male rats exposed to methanol during gestation from days 7-17, with a BMR of 1
13 control mean S.D, is provided in Table C-5. Male brain weight responses were chosen because
14 they resulted in lower BMD and BMDL estimates than female responses (data not shown).
15 Model fit was determined by statistics (AIC and $ residuals of individual dose groups) and
16 visual inspection, as recommended by EPA (2000b). The polynomial and power models reduced
17 to linear form and returned identical modeling results. In contrast, the more complex Hill and
18 Exponential models, which estimate a response "plateau" or asymptote, returned similar,
19 markedly nonlinear results. This is because these models approximated the response "plateau" to
20 be near the maximum drop in brain weight observed in the study (approximately 10% at the high
21 dose), resulting in a distinctly "L" shaped dose-response curve (see figure C-3).60 In this case, the
22 only PBPK model estimated Cmax dose that is associated with a significant response over
60 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 fourfold higher.
March 2011 C-15 DRAFT—DO NOT CITE OR QUOTE
-------
1 controls, the high-dose, is 60-fold higher than the mid-dose Cmax estimate. Thus, there are many
2 plausible curve shapes and, consequently, a wide range of BMDL estimates. Per EPA (2000a)
3 guidance and to err on the side of public health protection, the lowest BMDLiso of 10.26 mg
4 methanol/L in blood estimated from adequate and plausible models was chosen for use in the
5 RfC derivation (details of the Hill model results follow Table C-5). However, it should be noted
6 that there is a great deal of uncertainty and model dependence associated with these dose-
7 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
Hillb
Exponential 2
Exponential 3
Exponential 4b
BMD1SD (Cmax,
mg/L)A
207.18
207.18
207.18
207.18
43.08
199.98
199.98
39.53
BMDL1SD
(Cmax,mg/L)A
135.22
135.22
135.22
135.22
10.26
127.55
127.55
10.26
/rvalue
0.7881
0.7881
0.7881
0.7881
0.9602
0.9494
0.9494
Not reported
AICC
-173.12
-173.12
-173.12
-173.12
-171.59
-173.13
-173.13
-171.59
Scaled residual0
-0.43
-0.43
-0.43
-0.43
-0.10
-0.42
-0.42
0.10
aThe BMDL is the 95% lower confidence limit on the Cmax estimated to decrease brain weight by 1 control mean
S.D. using BMDS 2.1.1 (U.S. EPA. 2009a) and model options and restrictions suggested by EPA BMD technical
guidance (U.S. EPA. 2000a).
bPer EPA (2000a) 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).
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)
9
10
11
12
13
14
15
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-only\hilm-8wk-brwHil-
Restrict.(d)
Gnuplot Plotting File: C:\Usepa\BMDS21\Data\Methanol\NEDO\Gest-only\hilm-8wk-
brwHil-Restrict.pit
Tue Aug 25 12:40:30 2009
March 2011
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
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
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
Power parameter restricted to be greater than 1
The variance is to be modeled as Var(i) = exp(lalpha + rho * In(mean(i)
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
rho
intercept
v
n
k
-4.68678
0
2
-0.19
0.861776
303.331
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -n
have been estimated at a boundary point, or have been specified by the user,
and do not appear in the correlation matrix )
lalpha
rho
intercept
v
k
lalpha
1
-1
-0.083
0.6
-0.18
rho
-1
1
0.096
-0.6
0.18
intercept
-0.083
0.096
1
0.19
-0.55
V
0.6
-0.6
0.19
1
-0.73
k
-0.18
0.18
-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 ineguality constraint and thus
has no standard error.
March 2011
C-17
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
^
J
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
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Seal
0 11 2 2.01 0.047 0.0608 -0
1.2 11 2.01 2 0.075 0.0614 0
10.6 12 1.99 1.99 0.072 0.0662 0.
630.5 10 1.81 1.81 0.161 0.154 -0.
Model Descriptions for likelihoods calculated
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Signal
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 AIC
Al 83.205960 5 -156.411920
A2 92.060485 8 -168.120970
A3 90.797178 6 -169.594356
fitted 90.795933 5 -171.591867
R 70.761857 2 -137.523714
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.)
Tests of Interest
Test -2*log (Likelihood Ratio) Test df p-value
Test 1 42.5973 6 <.0001
Test 2 17.7091 3 0.000505
Test 3 2.52661 2 0.2827
Test 4 0.00248896 1 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. A non-homogeneous variance
model appears to be appropriate
The p-value for Test 3 is greater than .1. The modeled variance appears
ed Res
.371
.295
0954
0338
March 2011
C-18
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems
to adequately describe the data
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
Hill Model with 0.95 Confidence Level
c
o
Q.
o:
c
(0
2.05
2
1.95
1.9
1.85
1.8
1.75
1.7
Hill
EMDL BMD
100
200
300
dose
400
500
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).
March 2011
C-19
DRAFT—DO NOT CITE OR QUOTE
-------
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
9 BMDLHEC (ppm)= 0.02525*BMDLisD*24+(1290*BMDLisD*24)/(765.5 + BMDLiSD*24)
10 BMDLHEC (ppm)= 0.02525*10.3*24+(1290*10.3*24)/(765.5+ 10.3*24) = 321 ppm
11 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
12 converted using the conversion factor specific to methanol of 1 ppm = 1.31 mg/m3:
13 HEC (mg/m3) =1.31 x 321 ppm = 421 mg/m3
14 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
15 associated with animal to human differences, 10 for consideration of human variability, and 3 for
16 database deficiencies) to obtain the chronic inhalation reference value:
17 RfC (mg/m3) = 421 mg/m3 - 100 = 4.2 mg/m3
C.I.2.2. BMD Approach with a BMR of 0.05 Change Relative to Control Mean
(GD7-GD17)
18 A summary of the results most relevant to the development of a POD using the BMD
19 approach (BMD, BMDL, and model fit statistics) for decreased brain weight at 8 weeks in male
20 rats exposed to methanol during gestation from days 7 to 17, with a BMR of 0.05 change relative
21 to estimated control mean, is provided in Table C-6. Model fit was determined by statistics (AIC
22 and % residuals of individual dose groups) and visual inspection, as recommended by EPA
23 (2000a). Modeling considerations and uncertainties for this dataset were discussed in C. 1.2.1
24 and, as was done for the BMR of 1 S.D., the lowest BMDL0s of 21.07 mg methanol/L in blood
25 estimated from the BMDS exponential 4 model was chosen for use in the RfC derivation
26 (NEDO, 1987). Results from the exponential 4 model, including text and plot (see Figure C-4),
27 are shown after Table C-6.
March 2011 C-20 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
Hillb
Exponential 2
Exponential 3
Exponential 4b
BMDos (Cmax,
mg/L)A
328.84
328.84
328.84
328.84
92.30
320.62
320.62
76.36
BMDLos (Cmax,
mg/L)A
226.08
226.08
226.08
226.08
Not reported
215.13
215.13
21.07
/rvalue
0.7881
0.7881
0.7881
0.9446
0.9602
0.9494
0.9494
Not reported
AICC
-173.12
-173.12
-173.12
-173.12
-171.59
-173.13
-173.13
-171.59
Scaled residual0
0.02
0.02
0.02
0.02
0.10
0.02
0.02
0.10
aThe BMDL is the 95% lower confidence limit on the Cmax estimated to decrease brain weight by 5% using BMDS
2.1.1 (U.S. EPA. 2009a) and model options and restrictions suggested by EPA BMD Technical Guidance (2000a).
bPer EPA (2000a) 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
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
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 Run
The form of the response function by Model:
Model 2: Y[dose] = a * expfsign * b * dose}
Model 3: Y[dose] = a * expfsign * (b * dose)Ad}
Model 4: Y[dose] = a * [c-(c-l) * exp{-b * dose}]
Model 5: Y[dose] = a * [c-(c-l) * exp{-(b * dosej^
Note: Y[dose] is the median response for exposure = dose;
sign = +1 for increasing trend in data;
sign = -1 for decreasing trend.
Model 2 is nested within Models 3 and 4.
Model 3 is nested within Model 5.
March 2011
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
Model 4 is nested within Model 5.
Dependent variable = Mean
Independent variable = Dose
Data are assumed to be distributed: normally
Variance Model: exp(lnalpha +rho *ln(Y[dose]))
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
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
Parameter Estimates by Model
Model 2 Model 3 Model 4
Inalpha NC
rho NC
a NC
b NC
c NC
d NC
NC = No Convergence
NC
NC
NC
NC
NC
NC
7.03418
-18.1386
2.00677
0.00941775
0.902498
Table of Stats From Input Data
Dose N Obs Mean Obs Std Dev
0
1.2
10.6
630.5
11
11
12
10
2
2.01
1.99
1.81
0.047
0.075
0.072
0.161
Estimated Values of Interest
Model
4
Dose
0
1.2
10.6
630.5
Est Mean
2.007
2.005
1.988
1.812
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:
March 2011 C-22 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
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Signal
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma (1)^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + log (mean (i)
Model R: Yij = Mu + e(i)
Var{e(ij)} = Sigma/x2
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
March 2011
C-23
DRAFT—DO NOT CITE OR QUOTE
-------
Risk Type = Relative deviation
Confidence Level = 0.950000
4
5
6
7
8
9
10
11
Model
2
3
4
5
BMD and BMDL by Model
BMD
BMDL
0
0
76.3561
0
0
0
21.0664
0
Not computed
Not computed
Not computed
o
Q.
a:
c
ro
2.05
1.95
1.9
1.85
1.8
1.75
1.7
BMDL
14:1508/252009
Exponential Model 4 with 0.95 Confidence Level
Exponential
BMD
100
200
300
dose
400
500
600
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).
12 Once the BMDLos was obtained in units of mg/L, it was used to derive a chronic
13 inhalation reference value. The first step is to calculate the HEC using the PBPK model
14 described in Appendix B. An algebraic equation is provided (Equation 1 of Appendix B) that
March 2011
C-24
DRAFT—DO NOT CITE OR QUOTE
-------
1 describes the relationship between predicted methanol AUC and the human equivalent inhalation
2 exposure concentration (HEC) in ppm. This equation can also be used to estimate model
3 predictions for HECs from Cmax values because Cmax values, and AUC values were estimated at
4 steady-state for constant 24-hour exposures (i.e., AUC = 24 x Cmax).
5 BMDLHEC (ppm)= 0.02525*BMDL05*24+(1290*BMDLo5*24)/(765.5 + BMDL05*24)
6 BMDLHEC (ppm)= 0.02525*21.1*24+(1290*21.1*24)7(765.5+21.1*24) = 526 ppm
7 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
8 converted using the conversion factor specific to methanol of 1 ppm = 1.31 mg/m3:
9 HEC (mg/m3) =1.31 x 526 ppm = 690 mg/m3
10 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
11 associated with animal to human differences, 10 for consideration of human variability, and 3 for
12 database deficiencies) to obtain the chronic inhalation reference value:
13 RfC (mg/m3) = 690 mg/m3 - 100 = 6.9 mg/m3
14 C.2. RFC DERIVATIONS USING ROGERS ET AL.
14
15 For the purposes of deriving an RfC for methanol from developmental endpoints using
16 the BMD method and mouse data, cervical rib incidence data were evaluated from Rogers et al.
17 (1993). In this paper, Rogers et al. (1993) also utilized a BMD methodology, examining the
18 dosimetric threshold for cervical ribs and other developmental impacts by applying a log-logistic
19 maximum likelihood model to the dose-response data. Using air exposure concentrations (ppm)
20 as their dose metric, a value for the lower 95% confidence limit on the benchmark dose for 5%
21 additional risk in mice was 305 ppm (400 mg/m3), using the log-logistic model. Although the
22 teratology portion of the NEDO study (1987) also reported increases in cervical rib incidence in
23 Sprague-Dawley rats, the Rogers et al. (1993) study was chosen for dose-response modeling
24 because effects were seen at lower doses, it was peer-reviewed and published in the open
25 literature, and data on individual animals were available for a more statistically robust analysis
26 utilizing nested models available in BMDS 2.1.1 (U.S. EPA. 2009a).
27 The first step in the current BMD analysis is to convert the inhalation doses, given as ppm
28 values from the studies, to an internal dose surrogate or dose metric using the EPAs PBPK model
29 (see Section 3.4). For cervical rib malformations, Cmax of methanol in blood (mg/L) is chosen as
March 2011 C-25 DRAFT—DO NOT CITE OR QUOTE
-------
1 the appropriate internal dose metric (see Appendix D for further explanation). Predicted Cr
2 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
500
1,000
Methanol in blood Cmax (mg/L)A in mice
0.0216
0.218
1.14
2.46
7.83
26.4
134
4
5
6
7
8
9
10
11
12
13
14
15
16
17
aRounded to three significant figures.
These Cmax values are then used as the dose metric for the BMD analysis of cervical rib
incidence. A 10% BMR level is the value typically calculated for comparisons across chemicals
and endpoints for dichotomous responses because this level is near the low end of the observable
range for many types of toxicity studies. However, reproductive and developmental studies
having a nested design often have a greater sensitivity, and a 5% BMR is typically appropriate for
determination of a POD (Allen et al.. 1994: U.S. EPA. 2000a). Rogers et al. (1993) utilized a 5%
added risk for the BMR in the original study. This assessment utilizes both a 10% and 5% extra
risk level as a BMR for the determination of a POD.61 The nested suite of models available in
BMDS 2.1.1 (U.S. EPA, 2009a) was used to model the cervical rib data. In general, data from
developmental toxicity studies are best modeled using nested models, as these models account
for any intralitter correlation (i.e., the tendency of littermates to respond similarly to one another
relative to other litters in a dose group). All models were fit using restrictions and option settings
suggested in the EPA's BMD Technical Guidance Document (2000a).
C.2.1. BMD Approach with a BMR of 0.10 Extra Risk
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
61 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. 2000a') does not preclude the use of a BMR that is below observable response data and EPA
has deemed that Rogers et al. (1993) is adequate for the consideration of a 5% BMR.
March 2011
C-26
DRAFT—DO NOT CITE OR QUOTE
-------
1 exposed to methanol during gestation from days 6 to 15, with a BMR of 0.10 extra risk, is
2 provided in Table C-8. Model fit was determined by statistics (AIC and $ residuals of individual
3 dose groups) and visual inspection, as recommended by U.S. EPA (U.S. EPA, 2000a). The best
4 model fit to these data (from visual inspection and comparison of AIC values) was obtained using
5 the Nested Logistic (NLogistic) model. The textual and graphic (see Figure C-5) output from
6 this model follows Table C-8. The BMDLio was determined to be 94.3 mg/L using the 95%
7 lower confidence limit of the dose-response curve expressed in terms of the Cmax for methanol in
8 blood 0. M. Rogers. Mole, et al.. 1993).
Table C-8. Comparison of BMD modeling results for cervical rib incidence in
mice using modeled Cm,,\ of methanol as a common dose metric
Model
NLogisticb
NCTR
Rai and Van Ryzin
BMD10
(Cmax,mg/L)A
141.492
207.945
221.509
BMDL10 (Cmax,
mg/L)A
94.264
103.972
110.754
/rvalue
0.293
0.241
0.163
AICC
1046.84
1048.92
1051.65
Scaled residual0
0.649
0.662
0.661
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
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. 2000a).
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 et al. (1993).
NLogistic Model.
(Version: 2.13; Date: 02/20/2007)
Input Data File: U: \Methanol\BMDS\CervicalRib\Cmax\NLog_Cmax_10_def ault. (d)
Wed Nov 07 15:45:40 2007
BMD Method for RfC: Incidence of Cervical Rib in Mice versus Cmax Methanol, GD 6-15
inhalational study (Rogers,et al., 1993)
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.
March 2011
C-27
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
4.9
T-^
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 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.
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
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.
0000
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
.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
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
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.
,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
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
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
March 2011
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
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.
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.
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,0000
,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.
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.
7.
7.
7.
7.
8.
8.
8.
2.
3.
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
.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
.363
.363
.363
.363
.383
.383
.383
.703
.631
.562
.562
.506
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
.786
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
.902
.902
.902
.902
2.902
0.494
0.494
0.859
0.859
,150
,150
,150
,150
1.425
1.425
1.732
1.732
1.732
1.732
1.732
2.099
2.099
2.543
2.543
2.543
2.543
2.543
2.543
3.068
3.068
3.068
1.406
1.892
2.250
2.250
2.530
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
2
2
2
0
2
0
8
2
3
2
1
3
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
-0.2530
-0.2530
-0.2530
-1.1847
-0.4373
-1.2562
2.0195
0.8346
1.1101
-0.1967
-0.9842
0.3091
March 2011
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
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.
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
,0000
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.
5.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
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
.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
.743
.743
.743
.743
.743
.681
.681
.681
.681
.681
.681
.681
.681
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
5
5
5
5
5
6
6
6
6
6
6
6
6
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.
3.
3.
3.
3.
3.
4.
4.
4.
4.
4.
4.
4.
4.
,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
,714
,714
,714
,714
,714
,086
,086
,086
,086
,086
,086
,086
,086
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
3
5
5
4
4
6
2
4
5
6
5
4
5
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
-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
March 2011
C-30
DRAFT—DO NOT CITE OR QUOTE
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
10
i y
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
2005.0000 6.
2005.0000 6.
2005.0000 6.
2005.0000 6.
2005.0000 7.
2005.0000 7.
2005.0000 7.
2005.0000 7.
2005.0000 7.
2005.0000 8.
Combine litters
0000 0.681
0000 0.681
0000 0.681
0000 0.681
0000 0.623
0000 0.623
0000 0.623
0000 0.623
0000 0.623
0000 0.576
with adjacent
6
6
6
6
7
7
7
7
7
8
levels of the
within dose groups until the expected count
the fit of the
Grouped Data
Dose Lit
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
134.0000
526.0000
526.0000
526.0000
526.0000
526.0000
March 20 11
4.086
4.086
4.086
4.086
4.361
4.361
4.361
4.361
4.361
4.606
3
6
0
0
7
5
5
7
6
0
-0.5641
0.9945
-2.1227
-2.1227
1.1486
0.2781
0.2781
1.1486
0.7133
-1.7419
litter-specific covariate
exceeds 3
.0, to help
improve
X^2 statistic to chi-square.
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
6.0000
7.0000
7.0000
7.0000
8.0000
8.0000
8.0000
2.0000
3.0000
4.0000
5.0000
5.0000
Expected Observed
0.270
1.675
2.401
3.722
3.977
3.977
3.977
1.326
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
C-31
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
5
5
4
2
2
0
8
2
3
3
8
4
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
0.3044
-0.0284
-0.3578
-1.0166
-0.4373
-1.2562
2.0195
0.8346
1.1101
-0.8351
1.3670
-0.4926
DRAFT— DO
NOT CITI
-------
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
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
2005
2005
2005
2005
2005
2005
2005
Chi
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
-square =
To calculate
at
the mean
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
2
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
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
105.13 DF
the BMD and BMDL
litter specific
5.
5.
5.
5.
5.
5.
5.
5.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
2.
3.
5.
5.
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.
= 98
,592
,592
,592
,592
,592
,592
,592
,592
,105
,105
,105
,105
,105
,105
,105
,105
,105
,105
,105
,105
,496
,496
,496
,985
,985
,777
,577
,118
,118
,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
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
7
5
5
7
6
0
P-value = 0 .
, the litter
covariate of
specific
all the
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
-2
1
0
0
1
0
-1
2930
.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
.1227
.1486
.2781
.2781
.1486
.7133
.7419
covariate is fixed
data: 5
.379518
March 2011
C-32
DRAFT—DO NOT CITE OR QUOTE
-------
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.1
Extra risk
0.95
141.492
94.264
0.8
0.7
0.6
0.5
0.4
0.3
0.2
Nested Logistic Model with 0.95 Confidence Level
Nested Logistic
BMDL BMD
500
13:20 11/132007
1000
dose
1500
2000
Figure C-5. Nested Logistic Model, 0.1 Extra Risk - Incidence of Cervical
Rib in Mice versus C^x 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).
BMDLHEC (ppm) = 0.0224*BMDLio*24+(1334*BMDLio*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:
March 2011
C-33
DRAFT—DO NOT CITE OR QUOTE
-------
1 HEC (mg/m3) =1.31 x 1038 ppm = 1360 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
6 RfC (mg/m3) =1360 mg/m3 -100 = 13.6 mg/m3
C.2.2. BMD Approach with a BMR of 0.05 Extra Risk
7 A summary of the results most relevant to the development of a POD using the BMD
8 approach (BMD, BMDL, and model fit statistics) for increased incidence of cervical rib in mice
9 exposed to methanol during gestation from days 6 to 15, with a BMR of 0.05 extra risk, is
10 provided in Table C-9. Model fit was determined by statistics (AIC and ^ residuals of individual
11 dose groups) and visual inspection, as recommended by U.S. EPA (2000a). The best model fit to
12 these data (from visual inspection and comparison of AIC values) was obtained using the
13 NLogistic model. The text and graphic (see Figure C-6) output from this model follow Table C-
14 6. The BMDLos was determined to be 44.7 mg/L using the 95% lower confidence limit of the
15 dose-response curve expressed in terms of the Cmax for methanol in blood (J. M. Rogers, Mole, et
16 al.. 1993).
March 2011 C-34 DRAFT—DO NOT CITE OR QUOTE
-------
Table C-9. Comparison of BMD modeling results for cervical rib incidence in
mice using modeled Cmax of methanol as a common dose metric
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
Model
NLogisticb
NCTR
Rai and Van
Ryzin
BMD 05
(Cmax, mg/L)A
67.022
101.235
107.838
BMDL05
(Cmax, mg/L)A
44.651
50.618
53.919
/rvalue
0.293
0.241
0.163
AICC
1046.84
1048.92
1051.65
Scaled residual0
0.649
0.662
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 5% extra risk
(dichotomous endpoints) estimated by the model using the likelihood profile method (U.S. EPA.
2QOOa).
Vlodel 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_def ault. (d)
Wed Nov 07 15:45:40 2007
BMD Method for RfC: Incidence of Cervical Rib in Mice versus Cmax Methanol, GD6-GD15
inhalational study (J. M. Rogers, Mole, et al., 1993)
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
March 2011
C-35
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
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
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
Dose
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
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
. -Spec.
Cov.
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
Est
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
. Prob.
.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
Litter
Size
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
Expected
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.
,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
Observed
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
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
March 2011
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
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.
134.
134.
134.
134.
134.
134.
134.
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
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.
7.
7.
7.
7.
8.
8.
8.
2.
3.
4.
4.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
6.
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
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
.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
.363
.363
.363
.363
.383
.383
.383
.703
.631
.562
.562
.506
.506
.506
.506
.466
.466
.466
.466
.466
.466
.466
.466
.466
.786
.786
.786
.786
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
2.312
.902
.902
.902
.902
2.902
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
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.
,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
3.068
1.406
1.892
2.250
2.250
2.530
2.530
2.530
2.530
2.796
2.796
2.796
2.796
2.796
2.796
2.796
2.796
2.796
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
2
2
2
0
2
0
8
2
3
2
1
3
5
3
1
3
3
3
5
6
5
2
0
2
0.7656
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
-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
March 2011
C-37
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
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.
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
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.
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
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
0000
with
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
.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
.743
.743
.743
.743
.743
.681
.681
.681
.681
.681
.681
.681
.681
.681
.681
.681
.681
.623
.623
.623
.623
.623
.576
adjacent '.
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
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
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.
3.
3.
3.
3.
3.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
4.
,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
,714
,714
,714
,714
,714
,086
,086
,086
,086
,086
,086
,086
,086
,086
,086
,086
,086
,361
,361
,361
,361
,361
4.606
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
3
5
5
4
4
6
2
4
5
6
5
4
5
3
6
0
0
7
5
5
7
6
0
-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
-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
March 2011
C-38
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
within dose groups until the
the fit of the XA2 statistic
Grouped Data
Mean
Dose Lit. -Spec. Cov.
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.
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
1.
2.
3.
4.
5.
5.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
7.
7.
7.
7.
7.
7.
7.
8.
8.
8.
1.
2.
3.
3.
4.
5.
5.
5.
6.
7.
7.
7.
8.
8.
8.
2.
3.
4.
5.
5.
6.
6.
6.
6.
6.
6.
6.
6.
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
expected count exceeds 3.0, to help improve
to chi-square.
Scaled
Expected Observed Residual
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.
4.
5.
5.
5.
3.
3.
3.
1.
1.
4.
5.
5.
5.
5.
5.
5.
5.
5.
5.
5.
3.
3.
3.
3.
,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
,199
,086
,086
,086
,068
,068
,068
,406
,892
,500
,060
,060
,592
,592
,592
,592
,592
,592
,592
,592
,105
,105
,105
,105
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
5
5
4
2
2
0
8
2
3
3
8
4
6
8
11
2
2
9
5
6
0
4
5
1
-0
1
-0
-0
0
-0
-1
-0
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
.5586
.0237
.2443
.8049
.0098
.8614
.2970
.2458
.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
March 2011
C-39
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
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
2005
2005
2005
2005
2005
2005
2005
Chi
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
-square =
To calculate
at
the mean
7
7
7
7
7
7
7
7
8
8
8
9
9
1
2
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
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
.0000
105.13 DF
the BMD and BMDL
litter specific
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
3.
2.
3.
5.
5.
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.
= 98
,105
,105
,105
,105
,105
,105
,105
,105
,496
,496
,496
,985
,985
,777
,577
,118
,118
,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
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
7
5
5
7
6
0
P-value = 0 .
, the litter
covariate of
specific
all the
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
-2
1
0
0
1
0
-1
2930
.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
.1227
.1486
.2781
.2781
.1486
.7133
.7419
covariate is fixed
data: 5
.379518
March 2011
C-40
DRAFT—DO NOT CITE OR QUOTE
-------
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
0.05
Extra risk
0.95
67.0227
44.6514
Nested Logistic Model with 0.95 Confidence Level
0.8
0.7
0.6
0.5
0.4
0.3
0.2
11:42 11/162007
2000
Figure C-6. Nested Logistic Model, 0.05 Extra Risk - Incidence of Cervical
Rib in Mice versus Cm-^ Methanol, GD 6-15 inhalational study.
Source: Rogers et al. (1993).
1 Once the BMDL0s 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).
BMDLHEC (ppm) = 0.0224*BMDL05*24+(1334*BMDLo5*24)/(794+ BMDL05*24)
BMDLHEC (ppm) = 0.0224*44.7*24 + ((1334*44.7*24)7(794+ 44.7*24)) = 791 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:
March 2011
C-41
DRAFT—DO NOT CITE OR QUOTE
-------
HEC (mg/m3) = 1.31 x 791 ppm = 1036 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) = 1036 mg/m3 - 100 = 10.4 mg/m3
C.3. RfC-DERIVATIONS USING BURBACHER ET AL.
6 The BMD approach was utilized in the derivation of potential chronic inhalation
7 reference values from effects seen in monkeys due to prenatal methanol exposure. Deficits in
8 VDR were evaluated from Burbacher et al. (1999; 1999). In the application of the BMD
9 approach, continuous models in the EPA's BMDS 2.1.1 (U.S. EPA. 2009a) were fit to the dataset
10 for increased latency in VDR in neonatal monkeys. As the EPA's PBPK model was not
11 parameterized for monkeys, external concentration (ppm) was used as the dose metric.
12 The VDR test, which assesses time (from birth) it takes for an infant to grasp for a
13 brightly colored object containing an applesauce-covered nipple, is a measure of sensorimotor
14 development. Beginning at 2 weeks after birth, infants were tested 5 times/day, 4 days/week.
15 Performance on that test, measured as age from birth at achievement of test criterion (successful
16 object retrieval on 8/10 consecutive trials over 2 testing sessions), was reduced in all treated male
17 infants. The times (days after birth) to achieve the criteria for the VDR test were 23.7 ±4.8
18 (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
19 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
20 in the control to 1800 ppm groups, respectively. As discussed in Section 4.3.2, this type of
21 response data is sometimes adjusted to account for premature births by subtracting time (days)
22 premature from the time (days from birth) needed to meet the test criteria (Wilson & Cradock,
23 2004). When this type of adjustment is applied, the times (days after birth or, if shorter, days
24 after control mean gestation length) to achieve the criteria for VDR test were 22.0 ± 9.54 (n = 3),
25 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
26 (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
27 control to 1800 ppm groups, respectively. When these data were modeled within BMDS 2.1.1
28 (U.S. EPA, 2009a), there was no significant difference between unadjusted responses and/or
29 variances among the dose levels for males and females combined (p = 0.244), for males only (p =
30 0.321) and for males only with the high-dose group excluded (p = 0.182), or for adjusted
31 responses of males and females combined (p = 0.12), males only (p = 0.448) and males only with
March 2011 C-42 DRAFT—DO NOT CITE OR QUOTE
-------
1 the high-dose group excluded (p = 0.586).62 The only data that offered a significant dose-
2 response trend was that for unadjusted (p = 0.0265) and adjusted (p = 0.009) female responses,
3 but the model fits for the adjusted female response data were unacceptable. Only the unadjusted
4 female VDR response data offered both a dose-response trend and acceptable model fits. The
5 modeling results for this data set are presented in Table C-10.
6 The current BMD technical guidance (U.S. EPA, 2000a) suggests that in the absence of
7 knowledge as to what level of response to consider adverse, a change in the mean equal to
8 1 control S.D. from the control mean can be used as a BMR for continuous endpoints. A
9 summary of the results most relevant to the development of a POD using the BMD approach
10 (BMD, BMDL, and model fit statistics) for increased latency of VDR in female neonatal
11 monkeys exposed to methanol with a BMR of 1 control mean S.D. is provided in Table C-10.
12 Model fit was determined by statistics (AIC and £ residuals of individual dose groups) and
13 visual inspection, as recommended by EPA (2000a). The 3rd degree polynomial model returned
14 a lower AIC than the other models.63 The text and graphic (see Figure C-7) output from this
15 model follows Table C-10. The BMDLiso was determined to be 81.7 hrxmg/L, using the 95%
16 lower confidence limit of the dose-response curve expressed in terms of the ppm of external
17 methanol concentration.
62 BMDS (U.S. EPA. 2009a') 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.
63 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.
March 2011 C-43 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 residual0
0.5380
0.0994
0.0199
0.0000
0.0000
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
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. 2000a).
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: Burbacher et al. (1999).
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/x2 + ...
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
rho
beta_0
beta_l
beta_2
beta 3
4.07254
0
34.2
0
0
0
March 2011
C-44
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
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
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter (s) -beta 1 -beta 2
have been estimated at a boundary point, or have been
the user,
and do not appear in the correlation matrix )
lalpha rho beta 0 beta 3
lalpha 1 -1 -0.0076 0.018
rho -1 1 0.0076 -0.018
beta 0 -0.0076 0.0076 1 -0.37
beta_3 0.018 -0.018 -0.37 1
Parameter Estimates
95.0% Wald ci
Variable Estimate Std. Err. Lower Conf. Limit
Limit
lalpha -13.5062 9.81148 -32.7363
5.72395
rho 4.90831 2.77841 -0.537284
10.3539
beta 0 31.5013 1.49057 28.5798
34.4228
beta 1 8.36431e-025 NA
beta 2 0 NA
beta 3 3.19775e-006 1.53534e-006 1.88544e-007
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 Est Std Dev Sea
0 5 34.2 31.5 4.09 5.55
6.73 4 33 31.5 5.83 5.55
28.28 5 27.6 31.6 5.94 5.58
138.1 7 40 39.9 10.7 9.93 0
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 (1)^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 AIC
Al -51.042924 5 112.085848
A2 -47.867444 8 111.734888
A3 -49.286738 6 110.573475
fitted -50.589469 4 109.178938
R -55.013527 2 114.027055
specified by
Upper Conf.
6.20695e
led Res.
1.09
0.54
-1.59
.0199
March 2011
C-45
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
Test 1:
Test 2:
Test 3:
Test 4:
(Note:
Explanation of Tests
Do responses and/or variances differ among Dose levels?
(A2 versus R)
Are Variances Homogeneous? (Al vs A2)
Are variances adeguately modeled? (A2 versus A3)
Does the Model for the Mean Fit? (A3 versus fitted)
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.
to be appropriate here
The p-value for Test 4 is greater than .1.
to adeguately describe the data
Specified effect =
Risk Type
Confidence level =
BMD =
BMDL =
The modeled variance appears
The model chosen seems
Estimated S.D.s from the control mean
0.95
120.176
81.6513
Polynomial Model with 0.95 Confidence Level
i.
w
&
50
45
40
35
30
25
20
Polynomial
BMDL
BMD
20
40
60
80
100
120
140
15:3012/122008
dose
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.
Source: Burbacher et al. (1999: 1999)
March 2011
C-46
DRAFT—DO NOT CITE OR QUOTE
-------
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.
BMDLHEC (ppm) = 0.0224*BMDLisD+(1334*BMDLisD)/(794+ BMDLiSD)
6 BMDLHEC (ppm) = 0.0224*81.7+(l334*81.7)7(794+ 81.7) = 126.3 ppm
7 Next, because RfCs are typically expressed in units of mg/m3, the HEC value in ppm was
8 converted using the conversion factor specific to methanol of 1 ppm = 1.31 mg/m3:
9 HEC (mg/m3) = 1.31 x 126.3 ppm = 165 mg/m3
10 Finally, this HEC value was divided by a composite 100-fold UF (3 for uncertainty
11 associated with animal to human differences, 10 for consideration of human variability, and 3 for
12 database deficiencies) to obtain the chronic inhalation reference value:
13 RfC (mg/m3) = 165 mg/m3 - 100 = 1.7 mg/m3
March 2011 C-47 DRAFT—DO NOT CITE OR QUOTE
-------
APPENDIX D. RfC DERIVATION - COMPARISON OF DOSE METRICS
D.I. METHODS
D.I.I. 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, with
15 the responses from 5-hour exposures exceeding those from 7-hour exposures, and the responses
16 from 2-hour exposures exceeding those from 3-hour exposures. It was noted that the study was
17 done with a block-design, where the dams/litters for some concentration-time combination were
18 divided between multiple blocks and the average CR + SNR incidence in controls varied from
19 30-52% among the 8 blocks. Therefore block-control response (percent) was subtracted from
20 each exposed litter's response (percent) before calculating an average response among litters in a
21 given concentration-time combination. The resulting data are presented in Figure D-l.
Note. Hyperlinks to the reference citations throughout this document will take you to the NCEA HERO database (Health and
Environmental Research Online) at http://epa.gov/hero. HERO is a database of scientific literature used by U.S. EPA in the
process of developing science assessments such as the Integrated Science Assessments (ISAs) and the Integrated Risk
Information System (IRIS).
March 2011 D-l DRAFT—DO NOT CITE OR QUOTE
-------
105
90
5? 75
j/j 60
+
O 45
30
15
C
A: Raw data starting with _
grand-mean of controls^ A-^^—^^A
/ •" //*^T . o
x:'/^/*--
<*'"/''''
, / , rr /
—==^~- ~-~~ &
s ~~~~~
• 7-hr data
— A^ 5-hr data
— •— 3-hr data
O 2-hr data
X 1-hrdata
-.
) 5000 10000 15000
Concentration (ppm)
75
60
5? 45
W 30
+
0 15-
o e
-15
c
• 7-hr data
— A — 5-hr data
— • — 3-hr data
0 2-hr data
X 1-hrdata
A — ~^^
/ /• , /, ' '
o
^«
>^=^|K— — m' B: Data minus bkgd in
O each block, then summarized
) 5000 10000
Concentration (ppm)
15000
1
2
3
4
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
March 2011
D-2
DRAFT—DO NOT CITE OR QUOTE
-------
1 exposure, except for the response to 3 hours of 15,000 ppm methanol. Therefore these
2 background-corrected response measures will be used to evaluate the 3 dose metrics, with the
3 exception that the 3-hour 15,000 ppm data point will be dropped as an outlier. In particular, the
4 dose-response relationship based on these data will be plotted against each of the dose metrics to
5 determine which provides the most consistent overall dose-response relationship.
D.2. RESULTS
D.2.1. Dose Metric Comparisons
6 The average incidence of CR plus SNR from the concentration-time developmental
7 bioassay of Rogers et al. (1995), with block-specific control values subtracted from each litter
8 average before calculating overall average responses, is plotted in Figure D-2 against three dose
9 metrics: AUC, Cmax, and total amount metabolized of methanol (The volume units for Cmax and
10 AUC were adjusted to put all three data sets on approximately the same scale for comparison).
75
60
45
E
oTsO
O
0
-15
(
4
* Q,
• ci A n A
* *nQ A
* D A
A €
> 4 A[Eh
El •
A
n Cmax (mg/cL)
• AUC (hr-mg/mL)
A Amount metabolized
) 10 20 30 40 50 60
Dose metric
Figure D-2. Internal dose-response relationships for methanol-induced CR
plus SNR malformations in mice at various concentration-time combinations
for three dose metrics. 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. The set of response values plotted
for each metric is the same, only the metric associated with those responses
changes.
Source: Rogers et al. (1995).
March 2011
D-3
DRAFT—DO NOT CITE OR QUOTE
-------
1 While none of the metrics results are in complete alignment of the dose-response data, the
2 scatter for the Cmax dose-response (i.e., the range of response values associated with a given
3 small range of the dose metric - scatter in the y-direction) is quite a bit less than either of the
4 other two metrics. Thus, Cmax appears to be a better predictor of response than AUC or amount
5 metabolized. Looking at the exposure-response data in Figure D-l, one can see that 2- and
6 3-hour exposures at 5,000 ppm elicit no increase over control, while 5- and 7-hour exposures at
7 this level do.
8 If AUC or amount metabolized were true measures of risk, then one would expect a
9 graded response, where the 2- and 3-hour exposures were intermediate between controls and 5-7-
10 hour exposures. But the lack of response at those shorter times indicates that the concentration
11 (Cmax) has not risen high enough in such a short exposure to cause a response, while it has at the
12 longer durations. From Figure D-2, it appears that a Cmax of 11 mg/cL (1,100 mg/L) is a
13 NOAEL, with a linear increase in CR + SNR from that level to 38 mg/cL, after which the
14 response begins to plateau. Note that while the plot is of response above background, the plateau
15 is effectively at 100% total incidence: the highest points in Figure D-l are from the 7-hour
16 exposures at 15,000 ppm, where actual incidence was 98% (30% in controls); and the next
17 highest points are from the 5-hour 15,000 and 10,000 ppm exposures, where the incidences were
18 94% and 93%, respectively (32% in controls; both from the same block).
March 2011 D-4 DRAFT—DO NOT CITE OR QUOTE
------- |