United States
Environmental ProtecBon
Agency
Health Effects Document
for Perfluorooctane
Sulfonate (PFOS)
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
-------
Health Effects Document
for
Perfluorooctane Sulfonate (PFOS)
U.S. Environmental Protection Agency
Office of Water (43 04T)
Health and Ecological Criteria Division
Washington, DC 20460
EPA Document Number: 822R14002
Date: February 2014
Perfluorooctane sulfonate (PFOS) - February 2014 ii
Draft - Do Not Cite or Quote
-------
ACKNOWLEDGMENT
This document was prepared under the U.S. EPA Contract No. DW-8992342701, Work
Assignment No. 2011-001 with Oak Ridge National Laboratory. The Lead U.S. EPA Scientist is
Joyce Morrissey Donohue, Ph.D., Health and Ecological Criteria Division, Office of Science and
Technology, Office of Water.
The Oak Ridge National Laboratory is managed and operated by UT-Battelle, LLC., for the U.S.
Department of Energy under Contract No. DE-AC05-OOOR22725.
Perfluorooctane sulfonate (PFOS) - February 2014 iii
Draft - Do Not Cite or Quote
-------
Authors, Contributors, and Reviewers
Chemical manager
Joyce Morrissey Donohue, Ph.D.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency, Washington D.C.
Authors (EPA)
Amal Mahfouz, Ph.D.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency, Washington D.C.
Joyce Morrissey Donohue, Ph.D.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency, Washington D.C.
Tina Moore Duke, M.S.
Office of Water, Office of Science and Technology
Health and Ecological Criteria Division
U.S. Environmental Protection Agency, Washington D.C.
Authors (Oak Ridge National Laboratory)
Dana Glass-Mattie, D.V.M.
Environmental Sciences Division
Oak Ridge National Laboratory, Oak Ridge, TN
Carol S. Wood, Ph.D., D.A.B.T.
Environmental Sciences Division
Oak Ridge National Laboratory, Oak Ridge, TN
Peer Reviewers
Internal
Christopher Lau, Ph.D.
Perfluorooctane sulfonate (PFOS) - February 2014 iv
Draft - Do Not Cite or Quote
-------
National Health and Environmental Effects Research Laboratory, Office of Research and
Development
Reproductive Toxicology Division
U.S. Environmental Protection Agency, Research Triangle Park, NC
Greg Miller, Ph.D.
Office of Children's Health Protection, Office of the Administrator
U.S. Environmental Protection Agency, Washington, DC
John Wambaugh, Ph.D.
National Center for Computational Toxicology, Office of Research and Development
Systems Models for Chemical Toxicity and Exposure
U.S. Environmental Protection Agency, Research Triangle Park, NC
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency, Research Triangle Park, NC
External
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
-------
TABLE OF CONTENTS
ACKNOWLEDGMENT iii
Authors, Contributors, and Reviewers iv
TABLE OF CONTENTS vi
LIST OF TABLES viii
LIST OF FIGURES x
ABBREVIATIONS AND ACRONYMS xi
1.0 EXECUTIVE SUMMARY 1-
2.0 IDENTITY: CHEMICAL AND PHYSICAL PROPERTIES 2-
3.0 TOXICOKINETICS 3-
3.1 Absorption 3-
3.1.1 Oral Exposure 3-
3.1.2 Inhalation Exposure 3-2
3.1.3 Dermal Exposure 3-2
3.2 Distribution 3-2
3.2.1 Oral Exposure 3-5
3.2.2 Inhalation and Dermal Exposure 3-16
3.2.3 Other Routes of Exposure 3-17
3.3 Metabolism 3-17
3.4 Excretion 3-18
3.4.1 Oral Exposure 3-18
3.4.2 Inhalation Exposure 3-20
3.4.3 Dermal Exposure 3-20
3.4.4 Other Exposure Routes 3-20
3.5 Pharmacokinetic Considerations 3-20
3.5.1 Physiologically based models 3-20
3.5.2 Half-life data 3-27
3.5.3 Volume of Distribution Data 3-31
4.0 HAZARD IDENTIFICATION 4-
4.1 Human Effects 4-
4.1.1 Short-Term Studies and Case Reports 4-
4.1.2 Long-Term and Epidemiological Studies 4-
4.1.2.1 Noncancer Systemic Toxicity Studies 4-
4.1.2.2 Reproductive Hormones and Reproductive/Developmental Studies 4-5
4.1.2.3 Thyroid Effect Studies 4-9
4.1.2.4 Immunotoxicity 4-11
4.1.2.5 Carcinogenicity Studies 4-13
4.2 Animal Studies 4-14
4.2.1 Acute Toxicity 4-14
4.2.2 Short-Term Studies 4-16
4.2.3 Subchronic Studies 4-20
4.2.4 Neurotoxicity 4-24
4.2.5 Developmental/Reproductive Toxicity 4-26
4.2.6 Specialized Developmental Studies 4-37
4.2.7 Chronic Toxicity 4-41
4.2.8 Carcinogenicity 4-42
4.3 Other Key Data 4-44
4.3.1 Mutagenicity and Genotoxicity 4-44
4.3.2 Immunotoxicity 4-45
4.3.3 Physiological or Mechanistic Studies 4-49
Perfluorooctane sulfonate (PFOS) - February 2014 vi
Draft - Do Not Cite or Quote
-------
4.3.3.1 Noncancer Effects 4-49
4.3.4 Structure-Activity Relationship 4-61
4.4 Hazard Characterization 4-61
4.4.1 Synthesis and Evaluation of Major Noncancer Effects 4-62
4.4.2 Synthesis and Evaluation of Carcinogenic Effects 4-68
4.4.3 Mode of Action and Implications in Cancer Assessment 4-69
4.4.4 Weight of Evidence Evaluation for Carcinogenicity 4-70
4.4.5 Potentially Sensitive Populations 4-70
5.0 DOSE-RESPONSE ASSESSMENT 5-1
5.1 Dose-Response for Noncancer Effects 5-1
5.1.1 RfD Determination 5-1
5.1.1.1 Benchmark Dose Approach 5-7
5.1.1.2 Pharmacokinetic Model Approach 5-11
5.1.1.3 RfD Quantitation 5-21
5.1.2 RfC Determination 5-27
5.2 Dose-Response for Cancer Effects 5-27
6.0 REFERENCES 6-1
APPENDIX A: Summary of Data 1
APPENDIX B 1
Perfluorooctane sulfonate (PFOS) - February 2014 vii
Draft - Do Not Cite or Quote
-------
LIST OF TABLES
TABLE 2-1. Chemical and Physical Properties of PFOS 2-2
TABLE 3-1. Percent (%) Binding of PFOS in Rat, Monkey and Human Plasma3 3-3
TABLE 3-2. Average PFOS Level ((ig/mL or ppm) in Serum of Monkeys3 3-6
TABLE 3-3. PFOS Levels in the Serum and Liver of Rats3 3-7
TABLE 3-4. Mean (± SD) daily PFOS Consumption and Tissue Residue Levels in Rats Treated for 28
Days3 3-8
TABLE 3-5. Concentrations of PFOS in Male Rats' Whole Blood ((ig/mL) and Various Tissues
(Hg/g) After 28 Days3 3-8
TABLE 3-6. Levels of PFOS in serum and bile of rats treated for 5 days3 3-9
TABLE 3-7. PFOS Concentrations (Mean ± S.D.) in Samples From Pregnant Dams and Fetuses (GD 21
only) in (ig/mL (ppm) for Serum and Urine and (ig/g for Liver and Feces3 3-10
TABLE 3-8. Mean PFOS (± Standard Error) Concentrations in Serum, Liver and Brain Tissue in Dams
and Offspring3 3-11
TABLE 3-9. PFOS contents in serum, hippocampus and cortex of offspring (n=6)3 3-12
TABLE 3-10. Mean PFOS content in serum and lungs of rat offspring (n=6)3 3-12
TABLE 3-11. Levels of PFOS (Means ± SE) in Mouse Serum Following Treatment for 10 Days3 3-13
TABLE 3-12. Mean Concentration of PFOS (±SD) in Various Tissues of Mice3 3-14
TABLE 3-13. Ratios (means ± S.D.) between the concentrations of 35S-labeled PFOS in various organs
and blood of mouse dams, fetuses and pups versus the average concentration in maternal blood3 3-15
TABLE 3-14. Percent Distribution (%) of PFOS in Mice After a 50 mg/kg Subcutaneous Injection3 ..3-17
TABLE 3-15. Estimation of Toxicokinetic Parameters for PFOS3 3-18
TABLE 3-16. Mean % (± SE) of 14C-K+PFOS in rats after a single dose of 4.2 mg/kg3 3-19
TABLE 3-17. PFOS pharmacokinetic data summary for monkeys3 3-28
TABLE 3-18. PFOS pharmacokinetic data summary for Rats3 3-29
TABLE 3-19. PFOS pharmacokinetic data summary for mice3 3-30
TABLE 3-20. Summary of Half-life Data 3-31
TABLE 4-1. Association of Serum PFOS with Serum Lipids and Uric Acid 4-4
TABLE 4-2. Association of serum PFOS with reproductive and developmental outcomes 4-9
TABLE 4-3. Association of serum PFOS with the prevalence of thyroid disease and thyroid hormone
levels in studies of general and worker populations 4-11
TABLE 4-4. Mean (± SD) Values for Select Parameters in Rats Treated for 4 Weeks3 4-17
TABLE 4-5. Mean (± SD) Values for Select Parameters in Rats Treated for 28 Days3 4-18
TABLE 4-6. Mean (± SD) Values for Select Parameters in Monkeys Treated for 182 Days3 4-22
TABLE 4-7. Mean (± SD) Values for Select Parameters in Rats Treated for 14 Weeks3 4-24
TABLE 4-8. Fertility and Litter Observations in Dams Administered 0 to 2.0 mg PFOS/kg/Day3 4-31
TABLE 4-9. Effects Observed in the Mice Administered PFOS from GD 0 to GD 17/183 4-36
TABLE 4-10. Incidence of nonneoplastic liver lesions in rats_(number affected/total number) 4-42
TABLE 4-11. Tumor Incidence (%)3 4-43
TABLE 4-12. Genotoxicity of PFOS/« Vitro 4-44
TABLE 4-13. Genotoxicity of PFOS In Vivo 4-44
TABLE 4-14. Thyroid hormone levels in PFOS treated rats 4-51
Perfluorooctane sulfonate (PFOS) - February 2014 viii
Draft - Do Not Cite or Quote
-------
TABLE 4-15. Summary of PFAA Transactivation of Mouse and Human PPARa, (3/5 and ya 4-53
TABLE 5-1. NOAEL/LOAEL and Effects for Longer-term Duration Studies of PFOS 5-3
TABLE 5-2. NOAEL/LOAEL Data for Short-term Oral Studies of PFOS 5-5
TABLE 5-3. Benchmark Dose Modeling for a 5% Increased Risk of Developmental Toxicity in Rats.. 5-7
TABLE 5-4. Benchmark Dose Modeling for a 10% Increased Incidence of Liver Lesions in Rats 5-8
TABLE 5-5. Benchmark Dose Modeling fora 10% Increase in Liver Weight 5-10
TABLE 5-6. Description of prior distributions used 5-13
TABLE 5-7. Pharmacokinetic parameters used in the Andersen et al. (2006) model 5-15
TABLE 5-8. Predicted final serum concentration and time integrated serum concentration (AUC) for
different treatments of rat 5-16
TABLE 5-9. Predicted final serum concentration and time integrated serum concentration (AUC) for the
mouse 5-17
TABLE 5-10. Predicted final serum concentration and time integrated serum concentration (AUC) for
the monkey 5-17
TABLE 5-11. Average Serum concentrations Derived from the AUC and the duration of Dosing 5-19
TABLE 5-12. Human Equivalent Doses Derived from the Modeled Animal Average Serum Values ..5-21
TABLE 5-13. RfD Point of Departure Options from the PFOS Animal Studies 5-22
TABLE 5-14. The Impact of Quantification Approach on the RfD outcome for the PODs from the
available NOAELs 5-23
TABLE 5-15. The Impact of Quantification Approach on the RfD Outcome for the BMDLs from liver
and developmental endpoints 5-25
TABLE 5-16. The Impact of Quantification Approach on the RfD Outcomes for the HEDs from the
Pharmacokinetic Model Average Serum Values 5-26
TABLE A.I. PFOS Toxicokinetic Information 2
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version) 6
TABLE A.3. Summary of Animal Studies with Exposure to PFOS 14
Perfluorooctane sulfonate (PFOS) - February 2014 ix
Draft - Do Not Cite or Quote
-------
LIST OF FIGURES
Figure 2- 1. Chemical Structure of PFOS 2-1
Figure 3-1. Distribution of radiolabeled PFOS in dams and in fetuses/pups in the liver, lung, kidney and
brain 3-16
Figure 3-2. PFOS Contents in Urine, Feces and Overall Excretion in Male Rats Treated for 28 Days .3-19
Figure 3 -3. Schematic for a physiologically-motivated renal resorption pharmacokinetic model 3-21
Figure 3-4. Structure of model for PFOS in rats and monkeys 3-22
Figure 3-5. Structure of the PFOS PBPK model in monkeys and humans 3-23
Figure 3-6. Structure of the PBPK Model for PFOS in the Adult Sprague-Dawley Rat 3-25
Figure 3-7. Predicted Daily Average Concentration of PFOS in Maternal (black line) and Fetal (gray
line) Plasma at External Doses to the Dam 3-26
Figure 4-1. Functional categories of genes modified by PFOS in wild type and null mice 4-57
Figure 4-2. Function distribution and category enrichment analysis of the differential proteins 4-59
Figure 5-1. BMDS graphic output from selected model runs; data from Thomford, 2002 5-9
Figure 5- 2. BMDS graphic output from liver weight model runs; data from Seacat et al., 2002, 2003 5-11
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
-------
ABBREVIATIONS AND ACRONYMS
Ach
ACoA
ACOX1
ADAF
AIC
ALP
ALT
ANOVA
AP-1
Asp
AST
AUC
AWWARF
BGS
BMD
BMD
BMDS
BMI
BQL
BrdU
BUN
bw
°C
c
CaMKII
CAR
CAS
CCL
CCL3
CD
CFSE
CI
CL
CoA
CREB
CSF
CSM
Cte
cws
CYP4A22
Cytc
d
DA
DAUDA
acetylcholine
Acetyl CoA
peroxisomal acyl-coenzyme A oxidase
Age-Dependent Agjustment Factor
Akaike's Information Criterion
alkaline phosphatase
alanine transaminase
analysis of variance
activation protein-1
aspartate
aspartate aninotransferase
area under the curve
American Water Works Association Research Foundation
brain growth spurt
benchmark dose
benchmark dose - Lower 95th percentile confidence bound
benchmark dose software
body mass index
below quantifiable limit
bromodeoxyuridine
blood urea nitrogen
body weight
Celsius
Carbon
calcium/calmodulin-dependent protein kinase II
constitutive androstane receptor
Chemical Abstracts Service
Contaminant Candidate List
Contaminant Candidate List 3
circular dichroism
6-carboxyfluorescein succinimidyl ester
confidence interval
clearance
coenzyme A
cAMP response element-binding protein
Cancer Slope Factor or cerebrospinal fluid
cholestyramine
acyl CoA thioesterase
community water system
cytochrome P-450 4A22
cytochrome c
day
dansylamide or dopamine
1 l-(5-dimethylaminoapthalenesulphonyl)-undecanoic acid
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
XI
-------
DIO1 type 1 deiodinase
dL deciliter
DMEM Dulbecco's Minimal Essential Medium
DMSO dimethyl sulfoxide
DNA Deoxyribonucleic acid
DNBC Danish National Birth Cohort
DP dansyl-L-proline
DPPC dipalmitoylphosphatidylcholine
DWI drinking water intake
EAA excitatory amino acid
ECso half maximal effective concentration
ECF Electro-Chemical Fluorination
ED equilibrium dialysis
EPS A European Food Safety Authority
FOB functional observational battery
FT3 free triiodothyronine
FT4 free thyroxin
g gram
GABA gamma-aminobutyric acid
GAP-43 growth-associated protein-43
GD gestation day
GFAP glial fibrillary acidic protein
GGT gamma-glutamyl transpeptidase
GJIC gap junction intercellular communication
GLP good laboratory practice
Glu glutamate
Gly glycine
GS glutamine synthetase
GSH glutathione
GSI gonad-somatic index
FIDL high density lipoprotein
FLED human equivalent dose
FtL-60 human promyelocytic leukemia cell line
FDVIG-CoA 3-hydroxy-3-methylglutaryl coenzyme A
HOMA homeostatic model assessment
FIPT hypothalamic-pituitary-thyroid
HPLC/
ESMSMS High Performance Liquid Chromatography - electrospray tandem mass
spectrometry
FIRL health reference level
HSA Human Serum Albumin
HSDB Hazardous Substances Database
HSI hepatosomatic index
ICso half-maximal Inhibiting Concentration
lea inward calcium currents
ICR imprinting control region
IEF induction equivalency factor
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
XII
-------
IL-la interleukin
IL-6 interleukin 6
IRR incidence rate ratio
ITC isothermal titration calorimetry
IU international unit
IV intravenous
Ka adsorption rate constant
kg kilogram
KO knockout
Koc organic carbon water partitioning coefficient
Kow octanol-water partition coefficient
Kt affinity constant
L liter
LC50 Lethal concentration for 50% (statistical median) of animals
LC-ESI-
MS/MS liquid chromatography/electrospray ionization with tandem mass spectrometry
LC-MS liquid chromatography - mass spectrometry
LC-MS/MS liquid chromatography - negative electrospray tandem mass spectrometry
LD lactation day
LD50 Lethal dose for 50% (statistical median) of animals
LDH lactic dehydrogenase
LDL low density lipoprotein
L-FABP liver fatty acid binding protein
LI labeling index
LLOQ lower limit of quantification
LOAEL lowest observed adverse effect level
LOEC lowest observed effect concentration
LOQ Limit of Quantitation
LPS Lipopolysaccharide
m meter
MCLG Maximum Contaminant Level Goal
MDA malondialdehyde
Mdr2 multidrug resistance protein 2
ME malic enzyme
jig microgram
mg milligram
min minute
mL milliliter
|im micrometer
MOA mode of action
mol mole
MRL minimum reporting level
MRP multidrug resistance-associated protein
MTBE methyl tertiary-butyl ether
NAWQA National Water Quality Assessment
NDWAC National Drinking Water Advisory Council
ng nanogram
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
Xlll
-------
NA not applicable
ND not detected or not determined
NHANES The National Health and Nutrition Examination Survey
NTS sodium iodide symporter
NK natural killer
NMRI Naval Medical Research Institute
NOAEL no observed adverse effect level
NOEC no observed effect concentration
NPDWR National Primary Drinking Water Regulation
NRC National Research Council
NS no sample
NSP newborn screening program
NT not tested
OA octanoic acid
OAT organic anion transporter
OATp organic anion transporting peptide
OGWDW Office of Ground Water and Drinking Water
OR odds ratio
p probability
PB phenobarbital
PBDE polybrominated diphenyl ether
PBMC peripheral blood mononuclear cells
PBPK physiologically-based pharmacokinetic
PBS phosphate buffered saline
PCB polychlorinated biphenyl
PCNA proliferating cell nuclear antigen
PCoAO palmitoyl CoA oxidase
PFA perfluoroalkylate
PFC perfluorinated carboxylic acids
PFAA perfluoroalkyl acid
PFBA perfluorobutyric acid
PFBS perfluorobutane sulfonate
PFDA perfluorododecanoic acid
PFHS perfluorohexanesulfonic acid potassium salt
PFHxS Perfluorohexanesulfonic acid
PFOA Perfluorooctanoic acid
PFOC perfluorooctane
PFOS perfluoroocatane sulfonate
PFOSA perfluorooctane sulfamide
PFPA perfluoropropionic acid
PFTA perfluorotetradecanoic acid
pg picogram
PI proliferation index
PK pharmacokinetic
PND postnatal day
POD point of departure
POSF perfluorooctanesulfonyl fluoride
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
xiv
-------
pKa acid dissociation constant
PPAR peroxisome proliferator activated receptor
ppb parts per billion
ppm parts per million
ppt parts per trillion
mPSC miniature post-synaptic current
PTU propylthiouracil
PUFA polyunsaturated fatty acid
PWS public water system
PXR pregnane X receptor
Q flow in and out of tissues
Qfiic median fraction of blood flow to the filtrate
RBC red blood cell
Reg Det 2 Regulatory Determinations on the Second CCL
RfC reference concentration
RfD reference dose
RIA radio immunoassay
RNA ribonucleic acid
RSC relative source contribution
RSI renal-somatic index
RT-PCR reverse transcription polymerase chain reaction
RXRa retinoid X receptor alpha
SA serum albumin
SPC saponin compound
SD standard deviation
SDWA Safe Drinking Water Act
SIR standardized incidence ratio
SMR standardized mortality ratio
SOD superoxide dismutase
SRBC sheep red blood cells
STP sewage treatment plant
Syn 1 synapsin 1
SYP synaptophysin
T-AOC total antioxidation capability
Tmax time of maximum plasma concentration
T3 triiodothyronine
T4 thyroxine
ti/2 chemical half-life
Ti/2 elimination half-time
Tm transporter maximum
TAD target administered dose
TBG thyroxine-binding globulin
TC total cholesterol
TG triglycerides
TH tyrosine hydroxylase
TNFa tumor necrosis factor a
TNP trinitrophenol
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
xv
-------
TPO thyroid peroxidase
TRH thyrotropin releasing hormone
TSH thyroid stimulating hormone
TSHR thyroid stimulating hormone receptor
TT3 total triiodothyronine
TT4 total thyroxin
TIP time to pregnancy
TTR thyroid hormone transport protein, transthyretin
UCB umbilical cord blood
UCMR 3 Unregulated Contaminant Monitoring Rule 3
UF uncertainty factor
UGT1 uridine diphosphoglucuronosyl transferase
URAT urate transporter
U.S. EPA U.S. Environmental Protection Agency
USGS U.S. Geological Service
Vd volume of distribution
VLDL very low density lipoprotein
VOC volatile organic compound
WHAM weighted histogram analysis method
WT wild type
ww wet weight
WWTP waste water treatment plant
Wy Wyl4,648
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
xvi
-------
1.0 EXECUTIVE SUMMARY PFOS
Perfluorooctane sulfonate (PFOS) is a fluorinated organic compound with an eight-
carbon backbone and a sulfonate functional group. PFOS-related chemicals are used in a variety
of products including surface treatments for soil/stain resistance, surface treatments of textiles,
paper and metals and in specialized applications such as fire fighting foams. Because of strong
carbon-fluorine bonds, PFOS is stable to metabolic and environmental degradation and is
resistant to biotransformation. Data in humans and animals demonstrate ready absorption of
PFOS and distribution of the chemical throughout the body by noncovalent binding to plasma
proteins. Both experimental data and pharmacokinetic models show higher level of PFOS in
fetal serum and brain compared with the maternal compartments. PFOS is not readily eliminated
from humans as evidenced by the half-life of 5.4 years. In contrast, half-life values for the
monkey, rat, and mouse are 121 days, 48 days, and 37 days, respectively. The long half-lives
appear to be the result of resorption from the kidney. In other words after initial removal from
blood by the kidney, a substantial fraction of what would normally be eliminated in urine is
returned to the blood.
Peroxisome proliferation is usually associated with hepatic lesions in the rats, but some
uncertainties exist as to whether this is true for PFOS and if this is cause for concern in the
human population. Increased hepatic lipid content in the absence of a strong PPARa response is
a characteristic of exposure to PFOS. In two studies, mice administered PFOS showed
differential expression of proteins mainly involved in lipid metabolism, fatty acid uptake,
transport, biosynthetic processes, and response to stimulus. Many of the genes activated are
associated with nuclear receptors other than PPARa.
Epidemiology studies have examined occupational and residential populations at or near
large-scale PFOS production plants in the United States in an attempt to determine the
relationship between serum PFOS concentration and various health outcomes suggested by
standard animal toxicological studies. Exposures were mainly through contaminated drinking
water and to multiple PFCs. These studies found a positive association with increased PFOS
serum levels and an increase in total cholesterol, triglycerides, and uric acid in the general
population. In contrast, occupational studies did not indicate consistent associations between
PFOS and cholesterol and/or triglycerides in either cross-sectional surveys or in a longitudinal
analysis. Results are inconclusive or inconsistent for associations between increased serum
PFOS and affects on thyroid hormones and immunotoxicity.
In general population studies of effects on reproduction and development, the only
finding of note was a slight increase in the risk for low birth weight, however, this was not a
consistent finding across the studies.
In most animal studies with PFOS, short-term and chronic exposure resulted in an
increase in liver weight as at least one of the critical effects. Co-occurring effects in these
studies included decreased cholesterol, lower body weight, liver histopathology, and
developmental toxicity. In rat and monkey repeat-dosing studies (14 or 26 weeks), increased
liver weight was accompanied by decreased cholesterol and hepatocellular hypertrophy. As part
of a chronic bioassay, rats had low dose liver lesions with liver weight affected at higher doses.
The most severe effect observed in the longer-term studies was decreased pup survival in a one-
Perfluorooctane sulfonate (PFOS) - February 2014 1-1
Draft - Do Not Cite or Quote
-------
generation rat study at a LOAEL of 0.8 mg/kg/day. The LOAEL for decreased pup body weight
was 0.4 mg/kg/day in one- and two-generation studies. Developmental toxicity studies at
slightly higher doses support the concern for low dose-effects on pup survival. In a standard
developmental neurotoxicity study, male offspring showed increased motor activity and
decreased habituation on PND 17 following a maternal dose of 1 mg/kg/day. Two studies
provide evidence for immunological effects in mice.
U.S. EPA has selected 0.00003 mg/kg/day as the RfD for PFOS based on the consistency
of the response and with recognition of the use of developmental toxicity and liver weight as the
most sensitive endpoints for protection against co-occurring adverse effects. This value is the
outcome for modeled rat serum values for developmental. In the standard developmental
neurotoxicity study, male offspring showed increased motor activity and decreased habituation
on PND 17 following a maternal dose of 1 mg/kg/day in the absence of effects on pup body
weight. The human equivalent dose (HED) used as the basis for the RfD, was calculated from an
average serum concentration of 10.87 mg/L derived from the NOAEL of 0.3 mg/kg/day for
developmental neurotoxicity. A pharmacokinetic model was used to predict an area under the
curve (AUC) for the NOAEL and used to calculate an HEDNOAEL The total uncertainty factor
(UF) applied to the HEDNOAEL from the rat study was 30 which included a UF of 10 for
intrahuman variability, and a UF of 3 to account for toxicodynamic differences between animals
and humans. Comparable values derived from the HED for liver effects in rats and
developmental effects in mice are slightly higher than the RfD indicating that it will be
protective.
Under the EPA 2005 cancer guidelines, the evidence for the carcinogenicity of PFOS is
considered "suggestive of carcinogenicity, " but not sufficient to assess human carcinogenicity
potential. In a chronic oral toxicity and carcinogenicity study of PFOS in rats, liver, thyroid and
mammary fibroadenomas were identified. The biological significance of the mammary fibroadenomas
and thyroid tumors was questionable as a true dose-dependent response was not identified. The liver
tumors also had a questionable dose-response with slight but statistically significant increases only in
high-dose males and females. The liver tumors most found were adenomas (7/60 and 5/60 in high-
dose males and females vs. none in the controls of either sex); only one hepatocellular carcinoma was
found in a high-dose female. The genotoxicity data are uniformly negative. Human epidemiology
studies did not find a direct correlation between PFOS exposure and the incidence of carcinogenicity
in worker-based populations. Thus, the weight of evidence for the carcinogenic potential to humans of
these tumors was judged to be too limited to support a quantitative cancer assessment.
Perfluorooctane sulfonate (PFOS) - February 2014 1-2
Draft - Do Not Cite or Quote
-------
2.0 IDENTITY: CHEMICAL AND PHYSICAL PROPERTIES
Perfluorooctane sulfonate, commonly known as PFOS, and its salts are fluorinated
organic compounds and is part of the group of chemicals called perfluoroalkyl acids (PFAAs).
The two most widely known PFAAs have an eight-carbon backbone with either a sulfonate
(PFOS) or carboxylate (PFOA- perfluorooctanoic acid) attached (Lau et al., 2007). PFOS-
related chemicals are used in a variety of products including surface treatments for soil/stain
resistance, coating of paper as a part of a sizing agent formulation and in specialized applications
such as fire fighting foams. PFOS is produced commercially from perfluorooctanesulfonyl
fluoride (POSF) which is used primarily as an intermediate to synthesize other fluorochemicals.
POSF is manufactured through a process called Simons Electro-Chemical Fluorination (ECF) in
which an electric current is passed through a solution of anhydrous hydrogen fluoride and an
organic feedback of 1-octanesulfonyl fluoride causing the carbon-hydrogen bonds on molecules
to be replaced with carbon-fluorine bonds (OECD, 2002). PFOS can also be formed by the
degradation of other POSF-derived fluorochemicals.
Due to strong C-F bonds, PFOS is extremely stable and does not biodegrade in the
environment, making it very persistent. Because of this reason, most PFOS manufactured in the
United States was discontinued voluntarily by 3M in 2002. PFOS is soluble in fresh water at
approximately 519 mg/L. The solubility decreases significantly as the salt content of the water
increases. Because of the surface-active properties of PFOS, it forms three layers in
octanol/water making an n-octanol/water (Kow) partition co-efficient unable to be determined.
The potassium salt of PFOS has a low vapor pressure (OECD, 2002). No direct measurement of
the pKa of the acid has been located; however, the chemical is considered to have a low pKa.
The chemical structure is provided in Figure 2-1 and the physical properties for PFOS are
provided in Table 2-1.
Figure 2-1. Chemical Structure of PFOS
Perfluorooctane sulfonate (PFOS) - February 2014 2-1
Draft - Do Not Cite or Quote
-------
TABLE 2-1. Chemical and Physical Properties of PFOS
Property
Chemical Abstracts Registry (CAS) No.
EPA Pesticide Chemical Code
Chemical Formula
Molecular Weight
Color/Physical State
Boiling Point
Melting Point
Density
Vapor Pressure:
Henry's Law Constant
Kow
Koc
Solubility in Water
Information
2795-39-3*
C8F1703S
500.13
White powder
133°C@0.8kPa
> 400°C
3.31xlO'4Pa@20°C
Can not be measured**
519 mg/L in fresh water @ 25°C
12.4 mg/L in salt water @ 22-23°C
Sources: HSDB, 2009; OECD, 2002
"The CAS No. is for the potassium salt of PFOS which is the anion most commonly used in animal testing
"Because of the surface-active properties of PFOS, it forms three layers in octanol/water making this not measurable
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
2-2
-------
3.0 TOXICOKINETICS
Because of strong carbon-fluorine bonds, PFOS is stable to metabolic and environmental
degradation. It is not readily eliminated and can have a long half-life in humans and animals,
however, the toxicokinetic profile and the underlying mechanism for the chemical's long half-
life are not completely understood. In the case of another PFC, PFOA, transport families appear
to play role in absorption, distribution, and excretion and include organic anion transporters
(OATs), organic anion transporting peptides (OATps), multidrug resistance-associated proteins
(MRPs) and urate transporters (URAT). The transporters play a critical role in gastrointestinal
absorption, uptake by the tissues, and excretion via the kidney. Work is currently in progress to
determine if these same transporters are involved in PFOS toxicokinetics and preliminary data
appear to indicate that they are. Some preliminary inhibition studies suggest that PFOS has a
similar chain length dependent renal excretion and liver accumulation pattern as PFOA, and
would involve these same transporters.
Animal studies indicate that PFOS is well absorbed orally and distributes primarily in the
blood and liver. While PFOS can be a formed as a metabolite from other perfluocompounds,
PFOS itself does not undergo further metabolism after absorption takes place. PFAAs are known
to activate peroxisome proliferator activated receptor (PPAR) pathways by increasing
transcription of mitochondrial and peroxisomal lipid metabolism enzymes, sterol, and bile acid
biosynthesis and retinol metabolism genes. However, based on transcriptional activation of
many genes in PPARa-null mice, the effects of PFAAs involve more than activation of PPAR
(Andersen et al., 2008).
A summary of toxicokinetic data are provided in Appendix A, Table A.I and Table A.2.
3.1 Absorption
Absorption data are available for oral exposure in rats. While there are no absorption
studies available for humans that quantify the amounts absorbed relative to dose, extensive data
are available demonstrating the presence of PFOS in the serum. These data were reported in
Section 5.0 Biomonitoring Data.
The absorption process requires transport across the tissue interface with the external
environment. PFOS displays both hydrophobic and oleophobic properties indicating that
movement across the membrane surface is probably achieved with transporters rather than
simple diffusion.
3.1.1 Oral Exposure
Absorption in Animals
Rats
Following ingestion, PFOS is well absorbed. Three male rats were administered a single
dose of 4.2 mg/kg of PFOS-14C in solution; 3.45% of the total dose was found in the digestive
tract. The mean fecal excretion was 1.55% of the dose at 24 hours and 3.24% at 48 hours. At 24
hours, the mean sum of total carbon-14 in feces and digestive tract plus contents was 5% of the
Perfluorooctane sulfonate (PFOS) - February 2014 3-1
Draft - Do Not Cite or Quote
-------
dose. Some of this 5% likely represented systemically absorbed carbon-14 present either in the
digestive tract tissues or in the digestive tract contents as a result of excretion. The data from the
48 hour post dose group of rats were consistent with the 24 hour data. Thus, at least 95% of the
PFOS-14C dose was absorbed from solution after administration to non-fasted rats (Chang et al.,
2012).
3.1.2 Inhalation Exposure
An acute LCso study in rats indicated that PFOS absorption occurs by inhalation
exposure; however, pharmacokinetic data were not included (Rusch et al., 1979).
3.1.3 Dermal Exposure
No data are available on dermal absorption of PFOS.
3.2 Distribution
It has been suggested that PFOS is distributed within the body by non-covalently binding
to a plasma protein, most commonly, albumin. Binding studies are provided to help support this
hypothesis. Distribution data are provided only for rats following oral exposure. Indirect
distribution data are provided from analysis of PFOS in tissue and blood samples from studies
conducted in rats, monkeys and humans.
In humans, PFOS has been found to distribute mostly to the liver and blood, but has also
been identified in umbilical cord blood and breast milk. In humans, the ratio of PFOS identified
in the serum and liver tissue are similar, while in animals the amount found in the liver is higher
than that in the serum. In a study by Cui et al. (2009) bioaccumulation of PFOS was liver >
heart > kidney > whole blood > lung > testicle, brain and spleen in rats administered 5 or 20
mg/kg/day. The highest level of PFOS was found in the liver of the rats exposed to 20
mg/kg/day and was 648 ±17 |ig/g.
Binding Studies
The in vitro protein binding of PFOS in rat, monkey and human plasma at concentrations
of 1-500 ppm PFOS was investigated (Kerstner-Wood et al., 2003). The PFOS bound to plasma
protein in all three species at all concentrations with no sign of saturation (Table 3-1). When
incubated with human plasma protein, PFOS was highly bound (99.8 %) to albumin and showed
affinity for low density lipoproteins (LDLs, formerly beta-lipoproteins) (95.6%) with some
binding to alpha-globulins (59.4%) and gamma-globulins (24.1%).
Perfluorooctane sulfonate (PFOS) - February 2014 3-2
Draft - Do Not Cite or Quote
-------
TABLE 3-1. Percent (%) Binding of PFOS in Rat, Monkey and Human Plasma"
PFOS concentration (ppm)
1
10
100
250
500
Rat
~100b
99.8
99.7
99.5
99.0
Monkey
~100b
99.9
99.9
99.9
99.9
Human
99.4
99.9
99.9
99.9
99.9
a Data from Kerstner-Wood et al., 2003
b % binding values reported as —100 reflect a nonquantifiable amount of test article in the plasma water below quantifiable limit
(BQL) < 6.25 ng/mL
Zhang et al. (2009) used equilibrium dialysis, fluorophotometry, isothermal titration
calorimetry (ITC) and circular dichroism (CD) to characterize interactions between PFOS and
serum albumin (SA) and deoxyribonucleic acid (DNA). Solutions containing known amounts of
serum albumin or DNA were placed in dialysis tubing and suspended in solutions with varying
concentrations of PFOS. The solutions were allowed to equilibrate while measuring the decrease
in concentration of PFOS in the dialysate. During dialysis, the PFOS concentration decreased
reflecting its binding to the biopolymer within the dialysis bag. Based on the data, the SA could
bind up to 45 moles of PFOS per mole of protein and 0.36 moles per base pair of DNA. The
binding ratio increased with increasing PFOS concentrations and decreases in pH that would
promote protein and DNA denaturation. It is important to remember that these studies were
conducted in vitro and may not reflect in vivo situations.
The authors concluded that the interactions between SA and PFOS were the results of
surface electrostatic interactions between the sulfonate functional group and the positively
charged side chains of lysine and argenine. Hydrogen binding interactions between the negative
dipoles (fluorine) of the PFOS carbon-fluorine bonds could also play a role in the non-covalent
bonding of PFOS with SA. Intrinsic fluorescence analysis of tryptophan residues in SA
suggested a potential interaction of PFOS with tryptophan, an amino acid likely to be found in a
hydrophobic portion of SA. In the case of DNA, the authors postulated that the interaction with
PFOS occurred along the major or minor grooves of the double helix and was stabilized by the
hydrogen bonding and van der Waals interactions.
Serum albumin (SA) plays an important role in the transport of a number of endogenous
and exogenous compounds, such as fatty acids, bile acids, some medications and pesticides
(Zhang et al., 2009). Accordingly, changes in the conformation of SA could change its
transporting activity. Circular dichroism (CD) spectrometry was used to determine if PFOS
changed the conformation of the SA or DNA in solution. The results of both analyses indicated
conformational changes as a result of PFOS binding. However, the CD results did not
demonstrate whether there was a change in function as a result of the conformational change.
Accordingly, the authors investigated the effect of PFOS on the ability of serum albumin to
transport vitamin 82 (riboflavin). The study found that at normal physiological conditions, 1.2
mmol/L of PFOS decreased the binding ratio of serum albumin for riboflavin in vitro by > 30%.
Binding of five perfluoroalkyl acids, including PFOS, to human serum albumin was
investigated by using site-specific fluorescence (Chen and Guo, 2009). Intrinsic fluorescence of
trytophan-214 in human serum albumin was monitored upon addition of the perfluoroalkyl acids.
PFOS induced fluorescence quenching. A binding constant of 2.2 x lO4^!"1 and a binding ratio
of PFOS to human albumin of 14 were calculated. Human serum albumin also has two high-
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-3
-------
affinity drug binding sites which are known as Sudlow's drug Site I and Site II. Past
experiments have shown that two fluorescence probes, dansylamide (DA) and dansyl-L-proline
(DP) are specific for the two drug binding sites on human serum albumin. These two probes
emit negligible fluorescence, but after binding with albumin, fluorescence increases. The
titration of PFOS into human serum albumin pretreated with DA (site I), showed that at low
concentrations of PFOS (0.07 mM), DA emission increased as the PFOS concentration increased
until it was at 140% the original intensity. At the higher PFOS concentrations (0.7 to 4 mM),
however, the fluorescence dropped. The author speculated that the rise in fluorescence was
induced by the conformational changes of the protein after PFOS binds to it at a site different
from Site I and the decrease at higher concentrations was from displacement of DA by PFOS.
For the Site II, PFOS caused a fluorescence reduction that was quick at first but then became
more gradual making the possibility that PFOS was binding to this site with two different
affinities. The binding constant calculated at Site II was 7.6 x 106 M"1. These findings indicate
PFOS has binding sites that are similar to those identified for fatty acids.
Structure and the energy of binding sites were determined between PFOS and human
serum albumin (HSA) using molecular modeling (Salvalaglio et al., 2010). Calculations were
based on a compound approach docking, molecular dynamics simulations and estimating free
binding energies by adopting weighted histogram analysis method (WHAM)-umbrella sampling
and semiempirical methodology. The binding sites impacted were ones identified as human
serum albumin fatty acid binding sites. The PFOS binding site with the highest energy (-8.8
kcal/mol) was located near the tip of the Trp 214 binding site and the maximum number of
ligands that could bind to HSA for PFOS was 11. The most populated albumin binding site for
PFOS was dominated by van der Waals interactions. The author indicated that the number of
molecules adsorbed on HSA for PFOA was 9, compared to the 11 for PFOS, which may explain
why PFOS has a higher bioaccumulation than PFOA.
Weiss et al. (2009) screened the binding of several perfluorinated compounds, including
PFOS, to the thyroid hormone transport protein transthyretin (TTR) in a radioligand-binding
assay to determine if the compounds can compete with thyroxine (T4), the natural ligand of TTR.
Human TTR was incubated with 125I-labeled T4, unlabeled T4, and 10-10,000 nM competitor
(PFOS) overnight. The unlabeled T4 was used as a reference compound, and the levels of T4 in
the assay were close to the lower range of total T4 measured in healthy adults. PFOS had a high
binding potency to TTR. The 50% inhibition concentration was 940 nM. The authors concluded
that binding affinity for TTR did occur in perfluorinated compounds with peak binding in
compounds having at least an eight carbon length chain, such as PFOS.
Luebker et al. (2002) investigated the possibility that PFOS interferes with the binding
affinity of liver-fatty acid binding protein (L-FABP) which is an intracellular lipid-carrier
protein. This study was performed in vitro with a fluorescent fatty acid analogue 1 l-(5-dimethy-
laminoapthalenesulphonyl)-undecanoic acid (DAUDA). The concentration that can inhibit fifty
percent of specific DAUDA-L-FABP binding (IC50) was determined. PFOS demonstrated
inhibition of L-FABP in competitive binding assays; with 10 jiM PFOS added, 69% of specific
DAUDA-L-FABP binding was inhibited with the calculated ICso being 4.9 jiM.
Perfluorooctane sulfonate (PFOS) - February 2014 3-4
Draft - Do Not Cite or Quote
-------
3.2.1 Oral Exposure
Distribution in Humans
No studies are available in humans on administration of a controlled dose and PFOS
distribution. Olsen et al. (2003), however, sampled both liver and serum from cadavers for
PFOS. Both samples contained PFOS with good correlation between the samples from the same
subject. There was no difference in the PFOS concentrations identified in males and females or
between age groups. PFOS has been detected in both umbilical cord blood and breast milk
indicating that maternal transfer occurs (Apelberg et al., 2007; Von Ehrenstein et al., 2009;
Volkel et al., 2008). Karrman et al. (2010) also identified PFOS in postmortem liver samples
(n=12; 6 males and 6 females 27-79 years old) and in breast milk samples from healthy women
(n=10; females 30-39 years old) in Catalonia, Spain. The human samples indicate low levels in
the milk and good correlation between serum and hepatic levels.
Stein et al. (2012) compared perfluoroalkyl compound levels in maternal serum and
amniotic fluid. Concentrations of eight compounds were measured in paired samples from 28
women in their second trimester. PFOS (3.6-28.7 ng/mL) and three other compounds were
detected in all serum samples and PFOS was detected in nine amniotic fluid samples (0.2-1.8
ng/mL). The Spearman correlation coefficient was 0.76 for PFOS (p = 0.01) and the median
ratio of maternal serum: amniotic fluid concentration was 25.5:1. Based on simple regression,
PFOS was rarely detected in amniotic fluid until the serum concentration reached at least 5.5
ng/mL.
Harada et al. (2007) obtained cerebrospinal fluid (CSF) from seven patients (6 males and
1 female; ages 56-80) to evaluate the partitioning of PFOS between serum and the CSF. The
median concentration of PFOS in the serum was 18.4 ng/mL (0.018 ppm), compared to the
concentration in the CSF which was 0.10 ng/mL (0.0001 ppm). The CSF to serum ratio was 9.1
x 10"3. The levels identified indicate that PFOS does not easily cross into the adult blood-brain
barrier.
Distribution in Animals
Monkey
Seacat et al. (2002; further described under Section 4.2.3) administered 0, 0.03, 0.15 or
0.75 mg/kg/day potassium PFOS orally in a capsule by intragastric intubation to six young-adult
to adult cynomolgus monkeys/sex/group, except for the 0.03 mg/kg/day group which was 4/sex,
daily for 26 weeks (182 days) in a good laboratory practice (GLP) study, followed by a 52-week
recovery period. Levels of PFOS were recorded in the serum and liver. Serum PFOS
measurements demonstrate a linear increase with dosing duration in the 0.03 and 0.15 mg/kg/day
groups and a non-linear increase in the 0.75 mg/kg/day group. Levels in the high-dose group
appeared to plateau after about 100 days (14 weeks). Serum levels of PFOS decreased with
recovery in the two highest dosed groups. The average percent of cumulative dose of PFOS in
the liver ranged from 4.4 to 8.7% without any correlation to dose group or gender. The
concentration of PFOS in the liver decreased during the recovery period. Serum levels are
provided in Table 3-2.
Perfluorooctane sulfonate (PFOS) - February 2014 3-5
Draft - Do Not Cite or Quote
-------
TABLE 3-2. Average PFOS Level (ug/mL or ppm) in Serum of Monkeys"
Time
(weeks)
1
4
16
27
35
57
79
Group 1
0.0 mg/kg/day
Males
-------
TABLE 3-3. PFOS Levels in the Serum and Liver of Ratsa
Timepoint
(weeks)
0 ppm
M
F
0.5 ppm
(0.018-0.023
mg/k/day)
M
F
2 ppm
(0.072-0.099
mg/kg/day)
M
Serum PFOS levels (
0
14
53
105
106
-------
TABLE 3-4. Mean (± SD) daily PFOS Consumption and Tissue Residue Levels in Rats
Treated for 28 Days"
Parameter
PFOS
consumption
(mg/kg
bw/day)
Serum
(ugPFOS/g
serum)
Liver
(ugPFOS/g
liver)
Ratio
liver: serum
PFOS
Spleen
(ugPFOS/g
spleen)
Heart
(ug PFOS/g
heart)
0 mg/kg diet
M
0
0.47
±
0.27
0.79
±
0.49
2.04
±
1.39
0.27
±
0.36
0.10
±
0.14
F
0
0.95 ±
0.51
0.89 ±
0.44
1.30 ±
1.32
2.08 ±
4.17
1.42±
2.91
2 mg/kg diet
M
0.14 ±
0.02
0.95 ±
0.13
48.28 ±
5.81
51.34±
9.20
6.07 ±
1.85
4.67 ±
1.73
F
0.15 ±
0.02
1.50 ±
0.23
43.44
±6.79
29.99
±8.11
7.94 ±
3.76
6.54 ±
3.07
20 mg/kg diet
M
1.33 ±
0.24
13.45
±1.48
560.23
±
104.43
42.10
±9.20
45.27
±2.16
33.00
±3.44
F
1.43 ±
0.24
15.40
±1.56
716.55
±
59.15
46.81
±5.26
70.03
±
36.66
54.65
±
30.89
50 mg/kg diet
M
3.21 ±
0.57
20.93 ±
2.36
856.90
±
353.83
41.42 ±
16.95
122.51
±7.83
90.28 ±
4.95
F
3.73 ±
0.57
31.93±
3.59
596.75
±
158.01
20.23 ±
7.50
139.45
± 15.44
107.53
±6.24
100 mg/kg diet
M
6.34 ±
1.35
29.88 ±
3.53
1030.40 ±
162.80
35.23 ±
8.50
230.73 ±
11.47
154.13 ±
11.78
F
7.58 ±
0.68
43.20 ±
3.95
1008.59
±49.41
23.48 ±
1.98
294.96
± 26.66
214.45
± 17.58
1 Data from Table 1 on p.
SD = standard deviation
1531 in Curran et al., 2008
Ten three-month old male Sprague-Dawley rats/group were administered 0 (Milli-Q
water only), 5 or 20 mg/kg/day of PFOS by oral gavage for 28 days (Cui et al., 2009). Rats were
sacrificed after the exposure and blood and tissue samples obtained. Concentrations identified in
rat whole blood and various tissues at the end of the exposure are provided in Table 3-5. The
study indicated that the highest levels of PFOS were identified in the liver after 28 days of
exposure.
TABLE 3-5. Concentrations of PFOS in Male Rats' Whole Blood (ug/mL) and Various
Tissues (ug/g) After 28 Days3
Tissues
blood
liver
kidney
lung
heart
spleen
testicle
brain
Controls
ND
ND
ND
ND
ND
ND
ND
ND
5 mg/kg/day PFOS
72.0 ±25.7
345 ±40
93.9 ±13.6
46.6 ± 17.8
168 ± 17
38.5 ±11.8
39.5 ± 10.0
13.6 ± 1.0
20 mg/kg/day PFOS
No sample
648 ± 17
248 ± 26
228 ± 122
497 ± 64
167 ± 64
127 ± 11
146 ± 34
a Data from Table 1 in Cui et al., 2009.
ND = not detected
Yu et al. (2011) administered the following doses to approximately six female Wistar
rats/group are part of a study of PFOS on the thyroid:
1) vehicle (0.5% Tween 20),
2) PFOS at 0.2, 1.0 or 3.0 mg/kg,
3) propylthiouracil (PTU) at 10 mg/kg or
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-8
-------
4) PTU at 10 mg/kg and PFOS at 3.0 mg/kg once daily by gavage for 5 consecutive days.
Blood, bile and liver tissue were collected 24 hours after the last dose. The serum was used to
determine the level of PFOS as well as T4 (TT4) and T3 (TT3). PFOS levels in the serum as
well as the bile are provided in Table 3-6. The data demonstrate distribution to serum and bile
with a direct relationship to dose.
TABLE 3-6. Levels of PFOS in serum and bile of rats treated for 5 days"
PFOS (mg/kg bw)
0.0
0.2
1.0
3.0
Serum PFOS (mg/L)
-------
TABLE 3-7. PFOS Concentrations (Mean ± S.D.) in Samples From Pregnant Dams and
Fetuses (GD 21 only) in ug/mL (ppm) for Serum and Urine and ug/g for Liver and Feces"
Parameter
Serum
Liver
Urine
Feces
Dose
(mg/kg/day)
0.1
0.4
1.6
3.2
0.1
0.4
1.6
3.2
0.1
0.4
1.6
3.2
0.1
0.4
1.6
3.2
GDI
8.90 ±1.10
40.7 ± 4.46
160 ±12.5
318±21.1
-
-
-
-
0.05 ± 0.02
0.28 ±0.19
0.96 ±0.39
1.53 ±0.87
0.50 ±0.14
2.42 ±0.49
10.3 ±3.01
23.9 ±4.16
GD7
7.83 ±1.11
40.9 ±5.89
154 ±14.0
306 ±32.1
-
-
-
-
0.06 ±0.03
0.31 ±0.20
1.10 ±0.57
1.60 ±0.97
0.49 ±0.11
2.16 ±0.43
9.20 ±2.68
33.0 ±10.0
GDIS
8.81 ±1.47
41.4 ±4.80
156 ±25.9
275 ± 26.7
-
-
-
-
0.07± 0.04
0.53 ±0.23
0.36 ±0.35
0.52 ±0.28
0.66 ±0.10
2.93 ±0.62
11.1 ±3.28
29.5 ±8.92
GD21
Dams
4.52 ±1.15
26.2 ±16.1
136 ±86.5
155 ±39.3
29.2 ±10.5
107 ± 22.7
388 ±167
610 ±142
0.06 ±0.01
0.55 ±0.16
2.71 ±2.07
1.61 ±0.53
0.42 ±0.10
2.39 ±1.21
9.94 ±4.51
20.1 ±4.21
Fetuses
9.08
34.3
101
164
7.92
30.6
86.5
230
-
-
-
-
-
-
-
-
aData fromLuebkeretal.,2005a
- = no sample obtained
Twenty five female Sprague-Dawley rats/group were administered 0, 0.1, 0.3 or 1.0
mg/kg/day potassium PFOS by gavage from gestation day (GD) 0 through postnatal day (PND)
20 (Butenhoff et al., 2009). An additional 10 mated females were used as satellite rats to each of
the four groups and used to collect additional blood and tissue samples. Further details from this
study are provided in Section 4.2.4. Samples were taken from the dams, fetuses, and pups for
serum and tissue PFOS concentrations and the results reported by Chang et al. (2009). The
blood and tissue sampling results are provided in Table 3-8.
From GD 20 to PND 21, PFOS concentration in maternal serum, liver and brain
correlated with the daily doses administered. Maternal liver-to-serum PFOS ratios ranged from
1.8 to 4.9 while the maternal brain-to-serum ratios were 0.04 to 0.09 (Chang et al., 2009). From
GD 20 to PND 72, PFOS concentrations correlated well between fetal and pup serum, liver and
brain and the daily litter-matched maternal PFOS levels. Comparing results from dams and
fetuses/pups, liver PFOS concentrations were always higher for the dams than the respective
serum levels and the brain PFOS concentrations in the dams were always lower than that of
corresponding serum levels.
Based on the maternal and offspring data on GD 20, placental transfer of PFOS from rat
dams to developing fetuses does occur. Serum values were approximately 1-2 times greater in
the fetuses than in the dams at GD 20. The fetal liver PFOS concentration was less than that of
dams and the brain values were much higher; this is possibly due to the lack of development of
the blood-brain barrier at this stage of offspring development. PFOS serum concentrations in the
offspring were less during PND 4 and continued to drop through PND 72; however, based on the
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-10
-------
amounts still present in the pup serum lactational transfer of PFOS was occurring. At PND 72,
the males appeared to be eliminating PFOS more quickly as the serum values were lower than
those in the females; this difference was not observed at earlier time-points. In the liver, PFOS
was the greatest in the offspring at PND 4 but decreased significantly by PND 72. Values were
similar at all time-points between males and females. The level of PFOS in the brain was the
highest at GD 20 and had decreased by PND 21.
TABLE 3-8. Mean PFOS (± Standard Error) Concentrations in Serum, Liver and Brain
Tissue in Dams and Offspring"
Time
GD20b
PND4b
PND
21
PND
72
Dose
(mg/kg)
Control
0.1
0.3
1.0
Control
0.1
0.3
1.0
Control
0.1
0.3
1.0
Control
0.1
0.3
1.0
Serum PFOS (jig/mL)
Dam
-------
Ten pregnant Sprague-Dawley rats/group were administered 0, 0.1, 0.6 or 2.0 mg/kg/day
of PFOS by oral gavage in 0.5% Tween 80 from GD 2 to GD 21 (Zeng et al., 2011). On GD 21,
dams were monitored for parturition and the day of delivery was designated PND 0. On PND 0,
five pups/litter were sacrificed and the trunk blood, cortex and hippocampus were collected for
examination. The other pups were randomly redistributed to dams within the dosage groups and
allowed to nurse until PND 21, when they were sacrificed with the same tissues collected as
previously described. PFOS concentration in the hippocampus, cortex and serum increased in a
dose-dependent manner but overall was lower in all tissues on PND 21 when compared to PND
0. Levels of PFOS are included in Table 3-9.
TABLE 3-9. PFOS contents in serum, hippocampus and cortex of offspring (n=6)a
Time
PNDO
PND 21
Dose group (mg/kg/day)
control
0.1
0.6
2.0
control
0.1
0.6
2.0
Serum
fag/mL)
ND
1.50 ±0.43*
24.60 ±3.02**
45.69 ±4.77**
ND
0.37 ±1.12*
1.86 ±0.35**
4.26 ±1.73***
Hippocampus (jig/g)
ND
0.63 ±0.19*
7.43 ±1.62**
17.44 ±4.12**
ND
0.25 ±0.14*
1.59 ±0.78**
6.09 ± 1.30***
Cortex
(ng/g)
ND
0.39 ±0.09*
5.23 ± 1.58**
13.43 ±3. 89**
ND
0.06 ± 0.04*
1.03 ±0.59**
3.69 ±0.95***
aData from Table 2 in Zeng et al., 2011
ND- not detected
* p< 0.05 compared with control in the same day
** p< 0.05 compared with 0.1 mg/kg group in the same day
*** p< 0.05 compared with 0.6 mg/kg group in the same day
In Sprague-Dawley rats administered PFOS in 0.05% Tween (in deionized water) once
daily by gavage from GDI to GD21 at 0, 0.1 or 2.0 mg/kg/day. There was a postnatal decline in
the serum and brain PFOS levels between PND 0 and PND 21. PFOS concentrations were
higher in the serum when compared to the lung in offspring on both PND 0 and 21 (Chen et al.,
2012). See Table 3-10.
TABLE 3-10. Mean PFOS content in serum and lungs of rat offspring (n=6)a
Age
PNDO
PND 21
Treatment
0 mg/kg/day
0.1 mg/kg/day
2.0 mg/kg/day
0 mg/kg/day
0.1 mg/kg/day
2.0 mg/kg/day
PFOS in serum (jig/mL)
ND
1.7 ±0.35*
47.52 ±3.72*
ND
0.41 ±0.11*
4.46 ±1.82**
PFOS in lung (jig/g)
ND
0.92 ±0.04*
22.4 ±1.03*
ND
0.21 ±0.04*
3.16±0.11**
* Data from Table 2 in Chen et al., 2012
ND = not detected
* p< 0.05 or ** p< 0.01 compared with control
Mouse
In an immunotoxicity study (described in detail under Section 4.3.2. Immunotoxicity),
four to six C57BL/6 male mice/group were administered diets with 0 to 0.02% PFOS for 10
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-12
-------
days. Levels in the serum increased as the concentration increased (Qazi et al., 2009a). See
Table 3-11.
TABLE 3-11. Levels of PFOS (Means ± SE) in Mouse Serum Following Treatment
for 10 Days"
Dietary dose (% w/w)
PFOS (0)
PFOS (0.001%)
PFOS (0.005%)
PFOS (0.02%)
Number of mice
4
4
4
4
ppm
0.0287 ±0.01
50.8 ±2.5
96.7 ±5.2
340 ± 16
aData from study report by Qazi et al., 2009a
Adult male C57/BL6 mice (3 mice/group) were administered 35S-PFOS in the feed at a
low and high dose for 1, 3 and 5 days. The dose equivalents were 0.031 mg/kg/day in the low
dose group and 23 mg/kg/day in the high dose group. Tissue contents were determined by liquid
scintillation (Bogdanska et al., 2011). At 23 mg/kg/day after 5 days, mice had hypertrophy of
the liver, atrophy of fat pads and atrophy of epididymal fat when compared to the mice at 0.013
mg/kg/day at 5 days. To determine the amount of radioactivity recovered that was due to blood
in the tissues, the hemoglobin content was determined in all of the samples. By correcting for
PFOS in the blood, the actual tissue levels were then calculated.
At both doses and at all time-points, the liver contained the highest amount of PFOS. At
the low dose, the liver PFOS level relative to blood concentration increased with time, whereas
at the high dose, the ratio plateaued after three days. The autoradiography indicated that the
distribution within the liver did not appear to favor one area to a greater extent than any other.
The liver contained 40 to 50% of the recovered PFOS at the high dose. The authors
hypothesized that this could possibly reflect high levels of binding to tissue proteins.
In the high dose mice, the next highest level was found in the lungs. Distribution was
fairly uniform with some favoring of specific surface areas. The tissue: blood ratio for the lung
was greater than that for all other tissues except the liver. The lowest PFOS levels were in the
brain and fat deposits.
While the levels in Table 3-12 report the PFOS in the whole bone, when the authors did a
whole body autoradiogram of a mouse 48 hours after a single oral dose of 35S-PFOS (12.5
mg/kg), the results indicated that most PFOS was found in the bone marrow and not the calcified
bone. Levels for the kidney roughly equal those values observed in the blood at both
concentrations and all timepoints. See Table 3-12.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-13
-------
TABLE 3-12. Mean Concentration of PFOS (±SD) in Various Tissues of Mice"
Tissues
Iday
3 days
5 days
Dose of 0.013 mg/kg/day (PFOS in tissue reported as pmol/g)
Blood
Liver
Kidney
Lung
Whole bone
61(6)
114(13)**
38(19)
39 (29)
113(15)**
129 (41)#
343 (24)**#
65 (13)
88 (6)#
98 (24)
99 (21)
578 (39)**#
93(11)#
141 (10)*#
109 (6)
Dose of 23 mg/kg/day (PFOS in tissue reported as nmol/g)
Blood
Liver
Kidney
Lung
Whole bone
67(4)
246(31)**
62(3)
135(18)**
55 (6)*
171 (21)#
698 (71)**#
166 (8)#
336 (69)*#
155 (17)#
287 (9)#
1044(114)**#
233 (12)**#
445 (42)**#
207 (8)**#
aData from Tables 2 and 3 in Bogdanska et al., 2011
* significantly different (p< 0.05) than blood at the same time-point as evaluated by an independent t-test
**significantly different (p<0.01) than blood at the same time-point as evaluated by an independent t-test
#significantly different (p< 0.05) from the value for the same tissue at day 1 as determined by one-way analysis of variance
(ANOVA) followed by Duncan's test
Mouse- Distribution in Reproductive/Developmental Studies
Borg et al. (2010) administered a single dose of 12.5 mg/kg 35S-PFOS by intravenous
injection (n=l) or gavage (n=5) on gestation day (GD) 16 to C57B1/6 mouse dams. Distribution
of PFOS was determined in the dams/fetuses (GD 18 and 20) and pups (PND 1) by using whole-
body autoradiography and liquid scintillation counting. Distribution in the dams was similar
regardless of the route of exposure with the hepatic level being approximately four times greater
than the blood. At all timepoints in the dams, PFOS was most concentrated in the liver and
lungs. In the fetuses, the highest concentrations of PFOS were found in the kidneys and liver
and in pups on PND 1, PFOS was mostly concentrated in the lungs/liver. In dams, the
concentration of PFOS in the liver was approximately 4x and in the lung was approximately 2x
the blood concentrations, respectively. The distribution of PFOS in the kidneys was similar and
the amount in the brain was lower than that of the blood.
In the offspring at all timepoints, PFOS was homogeneously distributed in the liver at a
level 2.5x higher than maternal blood and 1.7x lower than maternal liver. In the fetuses on GD
18, values in the lungs were similar to the maternal lungs and this value increased by GD 20.
Pups on PND 1 had PFOS levels that were 3x higher in the lungs, compared to maternal blood
with a heterogeneous distribution. In the kidneys, the highest concentration of PFOS was
observed in the fetuses on GD 18 (3x higher than maternal levels) but then was similar to the
dams on GD 20 and PND 1. Levels in the brain were similar at all timepoints in the offspring and
higher than in the maternal brain, likely due to an immature brain-blood barrier. Select data are
provided in Table 3-13 and Figure 3-1.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-14
-------
TABLE 3-13. Ratios (means ± S.D.) between the concentrations of 35S-labeled PFOS in
various organs and blood of mouse dams, fetuses and pups versus the average concentration
in maternal blood"
Subject
Dams
Fetus on GD 18
Fetus on GD 20
Pups on PND 1
[35S-PFOS]organ/[35S-PFOS]matei.nalblood
Liver (n=6-8)
4 2** ±0.7
2.6** ±0.8
2.4** ±0.5
2.4* ±0.4
Lungs (n=5-6)
2.0* ±0.4
2.1* ±0.6
2.5** ±0.4
3.0** ±0.5
Kidneys (n=3-6)
0.9 ±0.1
2.8** ±0.3
1.4 ±0.2
1.0 ±0.5
Brain (n=6-9)
0.2** ±0.05
1.2 ±0.3
0.9 ±0.1
0.9 ±0.2
Blood (n=l-6)
1.0
2.3
1.1 ±0.04
1.7** ±0.3
a Data from Table 1 in Borg et al. (2010)
* Statistically significant (p<0.01) in comparison to maternal blood
** Statistically significant (p<0.001) in comparison to maternal blood
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-15
-------
w
o
m
(A) Liver
50-,
A
B A a
40- u
" B B
30- -r- m
r Jf-j —f~ •*
~J~| |M~ *** T
20~ "3T T""
10-
n
u I 1 1
GDIS GD20 PND1
Fetuses Fetuses Pups
(B)Lung
A 50"
^_
\}m 40-
II ^-
a 30-
• ^
o
£ 20-
^ £ 10-
o
A
b
c
A
a -fJ"
g — gl^_
ffl — P ^JOI
T • I
I u I 1
Maternal Maternal GD18 GD20 PND1
Liver Blood Fetuses Fetuses Pups
3
P
~TL~
B
I I
Maternal Maternal
Lungs Blood
50-1
40-
a 30-
to"
O
hj
O
i
.0
-^
O
i
(C) Kidney
A
B
25-,
20-
"o
a 15-1
w"
O
t(- 10-
(D) Brain
B
c
GD18 GD20 PND1
Fetuses Fetuses Pups
Maternal Maternal
Kidneys Blood
GD18 GD20 PND1
Fetuses Fetuses Pups
Maternal Maternal
Brain Blood
Filled symbols are representative after oral exposure; open after intravenous exposure.
A = p< 0.001 and a = p< 0.01, compared to maternal blood
B = p< 0.001 and b = p< 0.05, compared to maternal tissue
C = p< 0.001 and c = p< 0.05, comparing between fetuses/pups on GD20/PND1 with corresponding value
on GDIS;
Figure 3-1. Distribution of radiolabeled PFOS in dams and in fetuses/pups in the liver,
lung, kidney and brain
(Figure from Borg et al., 2010)
3.2.2 Inhalation and Dermal Exposure
No data on distribution following inhalation or dermal exposures were identified.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-16
-------
3.2.3 Other Routes of Exposure
Mice- Distribution in Reproductive/Developmental Studies
Male and female mice were administered PFOS by subcutaneous injection one time on
post-natal days (PNDs) 7, 14, 21, 28 or 35 (further study details provided in Section 4.2.5
Developmental/Reproductive Toxicity) at concentrations of 0 or 50 mg/kg bw (Liu et al., 2009).
Animals were killed 24 hours after treatment and the PFOS concentration levels obtained. The
percent distribution found in the blood, brain and liver are provided in Table 3-14. The
distribution shows that as the PND days increase, more PFOS is identified in the liver, and males
appear to accumulate slightly more than females.
TABLE 3-14. Percent Distribution (%) of PFOS in Mice After a 50 mg/kg Subcutaneous
Injection3
PND
7
14
21
28
35
Males
Blood
11.78 ±2.88
13.78 ± 1.52
9.85 ±2.74
9.89 ±2.94
13.33 ±0.89
Brain
5.04 ±1.49
1.61 ±0.80**
2.40 ±0.60**
0.85 ±0.19**
1.02 ±0.28**
Liver
14.84 ±4.01
26.50 ±7.36
51.35 ±11.06**
63.39 ±19.78**
73.68 ±6.86**
Females
Blood
10.77 ± 1.16
12.31 ±2.24
12.37 ±3.80
12.16 ±2.32
11.54 ± 1.28
Brain
4.17 ±1.17
3.26 ±0.58
2.14 ±0.38**
2.10 ±0.73**
0.90 ±0.23**
Liver
16.23 ±4.84
26.30 ±4.54
51.48 ±3.44**
51.05 ±10.59**
69.92 ±18.52**
a Data from Table 4 in Liu et al., 2009.
* statistically significant (p< 0.01)
Tissue Transport
As described earlier, Yu et al. (2011) administered PFOS to determine what transporters
were involved in hepatic uptake and to also determine the effect of transporters on thyroid
hormones. Approximately six female Wistar rats/group were administered 1) vehicle (0.5%
Tween 20), 2) PFOS at 0.2, 1.0 or 3.0 mg/kg, 3) propylthiouracil (PTU) at 10 mg/kg or 4) PTU
at 10 mg/kg and PFOS at 3.0 mg/kg once daily by gavage for 5 consecutive days. Blood, bile
and liver tissue were collected 24 hours after the last dose. Total mRNA as well as the mRNAs
for the following hepatic genes were isolated from the liver tissue: OATpl, OATp2, and MRP2.
Serum levels of PFOS were measured.
Exposure to 3.0 mg/kg of PFOS increased hepatic organic anion transporter OATp2
mRNA expression (1.43 times of control) and increased MRP2 approximately 1.80 and 1.69
times that of controls in the 1.0 and 3.0 mg/kg groups, respectively. No effect with treatment
was observed on OATpl. Studies of the role of transporters in PFOA tissue distribution are
rather extensive; that is not the case for PFOS. However, to the extent that it is like PFOA,
import and export tissue transporters are most likely important features controlling tissue
distribution and impacting pharmacokinetics.
3.3 Metabolism
No studies on the metabolism of PFOS were identified as it does not appear to be further
metabolized once absorbed. However, electrostatic interactions with biopolymers are indicated
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-17
-------
by the Kerstner-Wood et al. (2003) data on binding to plasma proteins plus the Zhang et al.
(2009) and Chen and Guo (2009) data from albumin-binding investigations.
3.4 Excretion
Humans
Harada et al. (2007) obtained serum and bile samples from patients (2 male and 2 female;
ages 63-76) undergoing gallstone surgery to determine the bile to serum ratio and biliary
resorption rate. The median concentration for PFOS in the serum was 23.2 ng/mL (0.023 ppm),
compared to the bile, 27.9 ng/mL (0.028 ppm). The biliary resorption rate was 0.97 which could
contribute to the long-half life in humans. Method of exposure to PFOS was unknown. See
Table 3-15.
TABLE 3-15. Estimation of Toxicokinetic Parameters for PFOSa
Bile
Participants
Median (n=4)
Serum levels- ng/mL (ppm)
23.2 (0.023)
Bile levels- ng/mL (ppm)
27.9 (0.028)
Bile to serum ratio
0.60
a Data from Tables 2 and 3 in Harada et al., 2007
Animals
Studies to determine the direct excretion routes are only available for rats. Most
excretion for PFOS occurs in the urine.
3.4.1 Oral Exposure
Ten male Sprague-Dawley rats (~ 9 weeks old)/group were administered 0, 5 or 20
mg/kg/day of either PFOA or PFOS by gavage once daily, 7 days a week for 4 weeks (Cui et al.,
2010). The dose groups were identified as the following: GO = ultrapure water; Gl= 5
mg/kg/day PFOA; G2 = 20 mg/kg/day PFOA; G3 = 5 mg/kg/day PFOS and G4 = 20 mg/kg/day
PFOS. Urine and fecal samples were obtained after the daily gavage by placing the rats in
metabolism cages for 24 hour intervals on the following days: prior to treatment (Day 0), Day 1
and Days 3, 5, 7, 10, 14, 18, 21, 24 and 28. Urine was collected three times daily and the volume
of the urine sample and weight of the fecal sample were recorded; samples were stored at -40°C
prior to analyzing. Target analytes were determined by using a high-performance liquid
chromatography-electrospray tandem mass spectrometry system with separation of PFOS and
PFOA achieved by the analytical column.
An upward trend of increased excretion was observed in the rats administered 5
mg/kg/day PFOS during the study; a similar trend was observed in the rats administered 20
mg/kg/day PFOS, but in the third week, mortalities occurred. By study day 24, there were only
2/10 rats in the 20 mg/kg/day group surviving. The range of PFOS excreted in urine by rats
treated with 20 mg/kg/day was 0.080 mg on day 1 to 0.673 mg on Day 14. In the feces, the
lowest amount of PFOS was in rats at 5 mg/kg/day on Day 1 (0.0015 mg) and the highest on Day
28 (0.355 mg). A similar trend in feces was observed in the rats treated with 20 mg/kg/day until
the deaths occurred; however, the fecal excretion reached a steady state after a maximum on day
18 (0.519 mg). This could have been the result of lower feces volume because the rats had
Perfluorooctane sulfonate (PFOS) - February 2014 3-18
Draft - Do Not Cite or Quote
-------
decreased food intake as well. The mean fecal excretion rates of PFOS between the two dose
groups was comparable as 1.2 and 1.3% of the oral doses were eliminated by fecal excretion in 5
and 20 mg/kg/day on day 1, respectively, indicating a majority of the dose was absorbed.
Overall, more PFOS was eliminated in the urine rather than the feces, but there was not a notable
difference in total excretion between the two PFOS dose groups. When the average elimination
rates (urinary, fecal and overall) of PFOA versus PFOS were compared, the amount of PFOA
being eliminated was higher than PFOS, especially on the first day. The elimination rates on the
first day were 2.6% and 2.8% in rats at 5 mg PFOS/kg/day and 20 mg PFOS/kg/day,
respectively. See Figure 3.2.
5 ma "Kg PFOS exposure group (G3)
M Bigflcg PFOS exposure group (G4)
J 14 21
Exposure Time (Day)
12
O>
E OS-I
c
O
Cfl
O
0.6
OJ
0.0
7 14 21
Exposure Time (Day)
No urine was available after day 18 in the 20 mg/kg/day group due to high mortality in this group.
* Statistically significant atp< 0.05 and ** p< 0.01
Figure 3-2. PFOS Contents in Urine, Feces and Overall Excretion in Male Rats Treated
for 28 Days
In a study by Chang et al (2012), three Sprague-Dawley rats/sex/timepoint were
administered 14C-PFOS as the potassium salt, one time by oral gavage at a dose of 4.2 mg/kg.
Urine, feces and tissues were collected in 24 and 48 hours and are presented below in Table 3-
16.
TABLE 3-16. Mean % (± SE) of 14C-K+PFOS in rats after a single dose of 4.2 mg/kga
Compartment
carcass
digestive tract
feces
urine
plasma
RBC
Total
% 14C of dose recovered
0-24 hr
79.0 ±1.8
3.58 ±0.23
1.55 ±0.15
1.57 ±0.25
11.02 ± 0.64 (estimated)*
2.29 ±0.18 (estimated)*
99.0
0-48 hr
94.2 ±5.1
3.32 ±0.12
3.24 ±0.08
2.52 ±0.31
10.01 ± 0.62(estimated)*
3.25 ±0.92 (estimated)*
116.5
"Data from Chang et al., 2012
*Mean body weight of 300g was used to estimate the red blood cell (RBC) and plasma volume.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-19
-------
3.4.2 Inhalation Exposure
In a case report, a 51-year old asymptomatic male researcher lived in a home with carpet
flooring that had been treated intermittently with soil/dirt repellants. The carpeting also had an
in-floor heating system under the carpets (Genuis et al., 2010). Because of his work the man
knew that he had an unusually high amount of PFCs in his serum, primarily
perfluorohexanesulfonic acid (PFHxS), PFOS and PFOA. The level of PFOS in his serum was
26 ng/g, the level in his urine was < 0.50 ng/mL and < 0.50 ng/g in sweat and stool samples. The
man began treatment with two bile acid sequestrants, cholestyramine (CSM) and saponin
compounds (SPCs) to see if they would lower the serum PFC levels. Stool samples were
evaluated for PFOS levels after administration of each compound. The concentration of PFOS
was increased after CSM treatment, suggesting that it may help with removing PFOS that gains
access to the GI tract with bile. The first stool sample after approximately 20 weeks of CSM
treatment showed PFOS levels of 9.06 ng/g and the second, 7.94 ng/g. The treatment with SPCs
did not increase the PFOS found in the stool. Serum levels of PFOS decreased to 15.6 ng/g after
12 weeks of treatment with CSM and to 14.4 ng/g after 20 weeks of treatment even though the
man's exposure at his home had not changed.
3.4.3 Dermal Exposure
No data on PFOS excretion following dermal exposures were identified.
3.4.4 Other Exposure Routes
At 89 days after a single intravenous (IV) dose of PFOS-14C of 4.2 mg/kg to male rats,
urinary excretion was 30.2 ± 1.5% of the total C-14 administered. Mean fecal excretion was
12.6 ± 1.2% (Chang et al., 2012). There was also evidence of enterohepatic circulation of PFOS.
3.5 Pharmacokinetic Considerations
3.5.1 Physiologically based models
Toxicokinetic models have been published as tools to estimate internal doses for humans,
monkeys, and rats that can accommodate half-life values that are longer than would be predicted
based on standard absorption, distribution, metabolism and excretion concepts. The underlying
assumption for all of the models is saturable resorption from the kidney filtrate, which
consistently returns a portion of the excreted dose to the systemic circulation and prolongs both
clearance from the body (e.g., extends half-life) and the time needed to reach steady state.
One of the earliest physiologically-based pharmacokinetic (PBPK) models (Andersen et
al., 2006) was developed for PFOS using two dosing situations in cynomolgus monkeys. In the
first, three male and three female monkeys received a single intravenous dose of potassium
PFOS at 2 mg/kg (Noker and Gorman, 2003). For oral dosing, groups of four to six male and
female monkeys were administered daily oral doses of 0, 0.03, 0.15 or 0.75 mg/kg PFOS for 26
weeks (Seacat et al., 2002).
This model was based on the hypothesis that saturable resorption capacity in the kidney
would possibly account for the unique half-life properties of PFOS across species. The model
Perfluorooctane sulfonate (PFOS) - February 2014 3-20
Draft - Do Not Cite or Quote
-------
structure (Figure 3-3; Andersen et al., 2006) was derived from a published model for glucose
resorption from the glomerular filtrate via transporters on the apical surface of renal tubule
epithelial cells.
input
(iv, oral)
Tissue
Compartment
n
Central
Compartment
(V,,; C,; fp,.*™)
Filtrate
Compartment
-------
liverblood partition coefficient and renal filtration. The author stated that development of a
human model was feasible.
Oral dose
I
Liver compartment
PL
IV
feces
QL
Central Compartment
(vol. of distr., free fraction of chemical in
plasma)
k13
k31
QFil
Tissue
compartment
Tm, Kt
Filtrate Compartment
(vol. of renal filtrate, renal filtrate rate,
saturable resorption)
storage
urine
Tm = transporter maximum, Kt= affinity constant and Q= flow in and out of tissues
Figure 3-4. Structure of model for PFOS in rats and monkeys
Loccisano et al. (2011) developed a PFOS PBPK model for monkeys based on the
Anderson et al. (2006) and Tan et al. (2008) models, and extrapolated it for use in humans
(Figure 3-5). The model reflects saturable renal absorption of urinary PFOS by the proximal
tubule of the kidney. This is represented in Figure 3-5 by the interactions between the plasma
and kidney plus the interaction of the filtrate compartment with both plasma and kidney. A
second route for PFOS resorption is represented by the gut plasma interaction which allows for
resorption of PFOS from bile secreted into the gastrointestinal tract.
The fraction of PFOS free in plasma and available for glomerular filtration was based on
data fit and was considered to decrease over time. Lacking primary data on transporter
resorption kinetics, the rate was based on the best fit to the plasma/urine data. Binding to serum
albumin allowed for less than a tenth of the plasma concentration to be available for glomerular
filtration. A storage compartment was added to the model between the filtrate compartment and
urine because PFOS appears in the urine at a slower rate than it disappears from the plasma.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-22
-------
Plasma
Free
fraction
-,,
QGut
QLiv
^
QFat
QSkn ^
QR
^
QKid
QFil
Gut
Liver
Fat
Skin
Rest of body
Kidney
/ V
Tm.Kt
Filtrate
\ /
Oral dose, drinking water
storage I
~~
kurine
urine
Tm = transporter maximum, Kt= affinity constant and Q= flow in and out of tissues
Figure 3-5. Structure of the PFOS PBPK model in monkeys and humans
Existing data sets for the cynomolgus monkey were used to develop the monkey model.
The IV data came from a single dose of 2 mg/kg (Noker and Gorman, 2003) wherein the
concentrations in plasma and urine were monitored for up to 161 days after dosing. The repeat-
dose oral data were those from Seacat et al. (2002) with exposures to 0, 0.03, 0.15 or 0.75 mg/kg
by capsule for 26 weeks with follow-up monitoring of plasma levels in two monkeys per group
at the two highest doses for a year after the cessation of dosing. Both data sets show that the
plasma and liver are the primary target tissues for PFOS. The model projections for the repeat
dose oral study were in good agreement with the Seacat et al. data for the 0.15 mg/kg dose, but
overestimated the plasma values for the 0.75 mg/kg/day dose. The model projected a sharper
rise in plasma levels with achievement of steady state more rapidly than indicated by the
experimental results.
Human data for PFOS are limited, although serum concentrations were collected from
retired workers (Olsen et al., 2007) and from residents (n=25) in Little Hocking, Ohio. The
structure of the human model was similar to that used for the monkeys. The fact that the serum
data applied to measurements made following uncertain exposure routes and uncertain exposure
durations presented a challenge in the assessment of model fit. The human half-life used for the
model (5.4 years) came from an occupational study (Olsen et al., 2007, see Section 3.5.2). No
measures of PFOS concentration were available for the drinking water at Little Hocking; thus,
the authors estimated the value that could account for the average population serum
concentration. The value for the drinking water was estimated to be 0.34 ppb. The model results
can be characterized as good when compared to the reported average serum measurements. The
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-23
-------
average daily exposure, consistent with the serum value, was estimated as 0.003 jig/kg/day
during the period from 1999 to 2000, and about 0.002 |ig/kg/day for the 2003 to 2009 time
period. The authors concluded that more data are needed on the kinetics of renal transporters,
intrahuman variability, and definitive information on exposures in order to refine the human
model.
Additional projections of human exposures consistent with measured average serum
levels from selected human populations have also been published (Egeghy and Lorber, 2011;
Thompson et al., 2010). Both papers used a first-order one-compartment model to assess PFOS
exposure from both an intake and body burden perspective using the following equations to
determine clearance (CL) with information on volume of distribution (Vd) and chemical half-life
(tvO-
CL = Vd x (In2 -1/2)
Human dose = average serum concentration x CL
Egeghy and Lorber (2011) estimated PFOS exposures from both intake and serum
measurements for both typical and contaminated scenarios for adults and children, using
available data from peer-reviewed publications. A range of intakes was estimated from the
PFOS serum concentrations reported by NHANES as well as published concentrations in various
media including dust, air, water, and foods. In the absence of human data, high and low
bounding estimates of 3 L/kg and 0.2 L/kg were used for volume of distribution. Total PFOS
intakes over all pathways were estimated to be 160 and 2200 ng/day for adults and 50 and 640
ng/day for children in typical and contaminated scenarios, respectively, with food ingestion
being the main exposure source in adults and food and dust ingestion being the two main sources
in children. Based on the model predictions, the range of intake of PFOS consistent with the
serum levels was 1.6 to 24.2 ng/kg-bw/day for adults assuming 70 kg.
Thompson et al. (2010) predicted PFOS concentration in blood serum as a function of
dose, elimination rate, and volume of distribution. The volume of distribution in this study, 0.23
L/kg bw, was adjusted by 35% from the calibrated data for PFOA in accordance with the
differences in PFOA and PFOS volumes of distribution calculated by Anderson et al. (2006).
The volume of distribution from PFOA was obtained by calibrating human serum and exposure
data collected from two communities in the Little Hocking, OH area (see Section 3.5.3).
Applying the volume of distribution and elimination rate values for PFOS calculated from the
Little Hocking population to serum data collected from members of the Australian population,
the predicted intake by the Australian population was calculated to be 21.7 to 3.6 ng/kg bw/day.
Loccisano et al. (2012a) utilized the saturable resorption hypothesis and pharmacokinetic
data from Chang et al. (2012), 3M (2009; unpublished) and Seacat et al. (2003) for adult
Sprague-Dawley Rats to develop the model depicted in Figure 3-6. The structure of the model is
similar to that for the monkey/human model depicted in Figure 3-5 but lacks the fat and skin
compartments and includes a storage compartment to accommodate fecal loss of unabsorbed
dietary PFOS as well as that from biliary secretions. Partition coefficients for liverplasma,
kidney: plasma, and rest of the body:plasma were derived from unpublished data on mice by
DePierre (2009; personal communication to authors), most of the other kinetic parameters were
based on values providing the best fit to the experimental data. The free fraction in plasma was
allowed to decrease with time suggesting a strong binding to serum proteins.
Perfluorooctane sulfonate (PFOS) - February 2014 3-24
Draft - Do Not Cite or Quote
-------
Oral, diet
faces
Figure 3-6. Structure of the PBPK Model for PFOS in the Adult Sprague-Dawley Rat
The agreement between the experimental data and the model output was generally good
but requires additional data from experimental studies on plasma binding and renal tubular
transporters to support further refinement of the parameters derived from model fit. In general,
liver and plasma concentrations after daily dosing were overestimated by a factor of about 2.
Male and female rats did not differ significantly in their ability to move PFOS from tissues to
urine or in resorption capability. PFOS appeared to have a greater capacity to bind to sites in the
liver than PFOA.
Loccisano et al. (2012b) expanded their adult Sprague-Dawley rat model described above
to cover gestational and lactational exposure to the fetus and pups. The data from Thibodeaux et
al. (2003) and Chang et al. (2012) for GD 0 to 20 were used in model development. Both studies
used multiple dose levels plus data on serum and selected tissue concentrations (liver, brain)
from the dams and fetus at one or more time points. The gestational model structure for the
dams is similar to Figure 3-6. The model was expanded to include the fetuses linked to the dams
by way of the placenta. Uptake from the placenta was described by simple diffusion; the fetal
plasma compartment was separate from the dams as was distribution to fetal tissues and amniotic
fluid. The model allowed for saturable binding of PFOS within the liver and to serum proteins.
Model performance was judged by its ability to predict 24-hour area under the curve (AUC) for
plasma, liver, and brain for both the fetus and dam. Brain data were only available from the
Cheng etal. (2012) study.
According to the model, liver concentrations for the dam are six to seven times greater
than those for the fetus and the brain levels for the fetus about eight times greater than those for
the dam. Model performance in comparison to the experimental data was judged to be good.
The model was used to project the maternal and fetal plasma levels expected at the doses
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-25
-------
employed in the Butenhoff et al. (2009), Luebker et al. (2005a,b) and Lau et al. (2003) studies as
depicted in Figure 3-7.
Reductions in maternal food/H2O
intake; reduced maternal wt. gain
(2 mg/kg & up) (Lau, et al ,2003)
1.6 mg'kg: Reduced neonatal
survival (Luebker. et al 2005).
10 mcj'kg: Significant
reduction in fetal weight;
significant increase in deft
palate and other
birth defects; all neonates
within 30-60 rnin. after
birth (Lau, etal2003
5 mgj: over
95% of neo nates
did not survive
PND1 (Lau, etal
20Q3}.'
1 mg'kg: LOAEL tor
developmental
neurotoxicity in male
pups after birth
(Butenhoff, et al
2009}
3.2 mg/kg: Reductions in implantation
sites, decreased gestation length, more
stillborn pups, increased number of pups
dead or dying on PNDs 1 -4 (Luebker, et al
2005); reduced serum thyraxine levels in
neonates (Yu, et al '2009;!.
Figure 3-7. Predicted Daily Average Concentration of PFOS in Maternal (black line) and
Fetal (gray line) Plasma at External Doses to the Dam
The lactational component of the Loccisano et al. (2012b) model allowed for PFOS
transport to neonates via mammary-tissues secretion and consequent ingestion by the pups. Pup
tissues included in the lactational model included the gut, liver, kidney and the remainder of the
body. A renal filtrate compartment linked to plasma and the kidney allowed for neonate PFOS
resorption. PFOS transfer to milk via the mammary gland was assumed to be controlled by
simple diffusion. Pup urine returned PFOS from the kidney filtrate to the dam.
The data of Luebker et al. (2005a,b) and Kuklenyik et al. (2004) were applied in model
development. The predicted milk:plasma ratio of 0.1 was in good agreement with the
experimental value (0.14) from Kuklenyik et al. (2004). The 24-hour AUC was used to evaluate
model predictions. At the beginning of lactation the levels in pup liver were greater than those
for the dam; at PND 14 they were equivalent and at PND 20 the levels for the dams were greater
than those for the pups. PFOS levels in the pup brain remained higher than those for the dams
throughout gestation and lactation but declined gradually across the duration of lactation. After
cessation of dosing PFOS levels in the livers of the dams declined gradually.
The authors acknowledged the lack of primary experimental data on PFOS transport and
potential transporters. Similarity to PFOA was assumed and PFOS was transparently described
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-26
-------
as lacking supporting transporter data. Additional research on PFOS binding to serum proteins
and liver tissues, its biliary excretion and resorption, plus information on renal resorption
transporters in dams and pups, is needed to accomplish further refinements to the published
model (Loccisano et al., 2012b).
3.5.2 Half-life data
Differences between species were observed when studies determining the half-life (Ti/2)
of PFOS in rats, mice, monkeys, and humans were performed. Gender differences in rats do not
appear to be as dramatic for PFOS as they are for PFOA (Loccisano et al., 2012a,b).
Humans
Occupational Population. Blood sampling was performed on retirees from the 3M plant in
Decatur, Alabama where PFOS was produced. These samples were taken approximately every 6
months over a 5-year period to predict the half-life of PFOS. Results ranged from approximately
4 years to 8.67 years (3M Company, 2000; Burris et al., 2002). Both of these studies exhibited
some deficiencies in how the samples were taken and the methods used.
More recently, Olsen et al. (2007) obtained samples from twenty-six (24 males and 2
females) retired fluorochemical production workers from the 3M Company in Decatur, Alabama
to determine the half-life of PFOS. Periodic serum samples (total of 7-8 samples) were collected
over a period of 5 years, stored at -80°C and at the end of the study, FtPLC-mass spectrometry
was used to analyze the samples. The study took place from 1998 to 2004. The mean number of
years worked at the plant was 31 years (range: 20 to 36 years), the mean age of the participants at
the initial blood sampling was 61 years (range: 55 to 75 years), and the average number of years
retired was 2.6 years (range: 0.4 to 11.5 years). The initial arithmetic mean serum concentration
of PFOS was 799 ng/mL [0.79 ppm] (range: 145 to 3490 ng/mL), and when samples were taken
at the end of the study, the mean serum concentration was 403 ng/mL [0.40 ppm] (range: 37 to
1740 ng/mL). Semi-log graphs of concentration versus time for each of the 26 individuals were
created and individual serum elimination half-lives were determined using first-order
elimination. The arithmetic and geometric mean serum elimination half-lives of PFOS were 5.4
years (95% CI, 3.9- 6.9) and 4.8 years (95% CI, 4.1-5.4), respectively.
General Population. Data were not found for estimation of the half-life of PFOS in the general
population.
Infants. Newborn Screening Programs (NSPs) collect whole blood as dried spots on filter paper
from almost all infants born in the United States. One hundred and ten of these NSPs collected
from infants born between 1997 and 2007 in the state of New York were analyzed for PFOS
(Spliethoff et al., 2008). The methods used for analysis were validated by using freshly drawn
blood from healthy adult volunteers. The mean whole blood concentration for PFOS was 0.81 to
2.41 ng/mL (0.00081 to 0.00241 ppm). The study grouped the blood spots by two different time-
points; those collected from 1999-2000 and from 2003-2004 which corresponded to the intervals
reported by NHANES. The PFOS concentrations decreased with a mean value of 2.43 ng/mL
(0.0024 ppm) reported in 1999-2000 and 1.74 ng/mL (0.0017 ppm) in 2003-2004. The study
authors also used regression slopes for natural log concentration versus years since 2000 to
calculate the half-life for PFOS. The calculated half-life for PFOS was 4.1 years.
Perfluorooctane sulfonate (PFOS) - February 2014 3-27
Draft - Do Not Cite or Quote
-------
Animal Data
A series of studies was performed to determine the pharmacokinetic parameters of PFOS
in rats, mice and monkeys following administration of single doses (Chang et al., 2012).
Another study provided half-life information from monkeys (Seacat et al., 2002). Serum half-
life estimates were obtained from short-term and long-term studies. Minimal gender-related
differences were observed in the species examined.
Monkeys
In the study by Chang et al. (2012), three male and three female monkeys were
administered a single IV dose of PFOS at 2 mg/kg and followed for 161 days. All monkeys were
observed twice daily for clinical signs and body weights were obtained weekly. Urine and serum
samples were taken throughout the study. There was no indication that elimination was different
from males versus females. Serum elimination half-lives ranged from 122-146 days in male
monkeys and 88-138 days in females. Mean values are shown in Table 3-17. The volume of
distribution values (Vd) suggest that distribution was predominately extracellular.
TABLE 3-17. PFOS pharmacokinetic data summary for monkeys3
Species
Cynomolgus
monkeys
Time
evaluated
after last dose
23 weeks
Route
IV
Sex
M
F
Amount
K+PFOS
(mg/kg)
2
2
Mean serum
Tin by sex
(days)
132.0 ±7
110.0 ±15
Mean
serum Tlfl
by species
(days)
120.8
Mean
serum Vd
by sex
(mL/kg)
202
274
a Data from Chang et al., 2012
Seacat et al. (2002) administered 0, 0.03, 0.15 or 0.75 mg/kg/day potassium PFOS orally
in a capsule by intragastric intubation to 6 young-adult to adult cynomolgus monkeys/sex/group,
except for the 0.03 mg/kg/day group which had 4/sex, daily for 26 weeks (182 days) in a GLP
study. Two monkeys/sex/group in the control, 0.15 and 0.75 mg/kg/day groups were monitored
for one year after the end of the treatment period for reversible or delayed toxicity effects. The
elimination half-life for potassium PFOS in monkeys was estimated from the elimination curves
as approximately 200 days. This value is consistent with that reported by Chang et al. (2012)
above.
Rats
Chang et al. (2012) conducted a series of pharmacokinetic studies in rats (Table 3-18).
First, a single oral dose of 4.2 mg ^C-K^PFOS/kg was administered to male Sprague-Dawley
rats (3/timepoint). Urine and fecal samples were collected for 24 and 48 hours. Interim
sacrifices to obtain plasma samples were obtained at 1, 2, 6, 12, 24, 48, 96 and 144 hours post-
dosing. In the next study, 3 rats/sex were administered 2.2 mg PFOS/kg once by oral gavage or
IV administration. The rats had a jugular cannula in place and serum samples from it were
obtained at 0.25, 0.5, 1, 2, 4, 8, 18 and 24 hours post-dosing. The Ti/2 values must be viewed
with caution because the blood samples were limited to a 24-hour post-dose observation period
in contrast to the 144-hour (6-day) period from the first study.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-28
-------
In a third study, serum uptake and elimination of PFOS were evaluated at two dose
levels: 2 mg/kg and 15 mg/kg. The PFOS was administered as a single oral dose in a 0.5%
Tween 20 vehicle to 3 rats/sex or 5/sex at the low and high dose, respectively. Periodic serum,
urine and fecal samples were taken for up to 10 weeks. Liver concentrations were evaluated at
termination. Half-life estimates (Table 3-18) did not differ significantly with dose, but there was
a difference by sex, with values for the males about half those for the females. There were also
gender related differences in the volume of distribution values. PFOS concentrations in the liver
exceeded those for paired serum concentrations.
The studies by Chang et al. (2012) described above are limited in that they each reflect
pharmacokinetic features associated with a single dose. In an unpublished study by 3M
(Butenhoff and Chang, 2007), 5 rats/sex were administered 1 mg/kg/day of PFOS orally for 28
days. Interim blood, urine and feces were obtained for up to 10 weeks. There was no effect on
body weight and PFOS elimination was more prominent in the urine than the feces. The
elimination of PFOS in this study approximated first order kinetics with a 'stair-stepping'
pattern. Using nonlinear, noncompartmental software for computation, the half-lives for males
ranged from 35 to 53 days and that for females from 33 to 55 days.
TABLE 3-18. PFOS pharmacokinetic data summary for Ratsa
Species
SD rats
SD rats
SD rats
SD rats
Time
evaluated
after last
dose
144 hrs
24hrs
10 weeks
10 weeks
Route
Oral
Oral
IV
Oral
Oral
Sex
M
M
F
M
F
M
F
M
F
Amount
K+PFOS
(mg/kg)
4.2
2.2
1x28
days
1x28
days
2
15
2
15
Mean
serum T1/2
by dose
(days)
8.2 ±1.5
3.1bc
1.9C
8.0C
5.6b
35-53
33-55
38.3 ±2.3
41.2±2.0
62.3 ±2.1
71. 1±
11.3
Mean
serum
Ti/2 by
sex
(days)
Mean
serum
Ti/2 by
species
(days)
Not determined due
to study design.
48.2
46.9
39.8
66.7
47.6
53.3
Mean
serum Vd
by dose
(mL/kg)
275
765 b
521
649
586b
-
-
1228
666
484
468
a Data from Chang et al., 2012 and Butenhoff and Chang, 2007 (unpublished)
bData reflected a single value derived from one rat only
°Within limits of the study design and a follow-up duration of only 24 hrs
NA= not available
Mice
CD-I male and female mice were administered PFOS as a single oral dose of 1 or 20
mg/kg (Chang et al., 2012). At designated times (2, 4, 8 hours and 1, 8, 15, 22, 36, 50, 64 and
141 days) post-dosing, four mice/sex were sacrificed and blood, kidneys, and liver samples were
obtained. Urine and feces were collected for each 24-hour period up until sacrifice. At the end
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-29
-------
of the observation period, the daily urinary and fecal excretion was less than 0.1% of the
administered dose. Results are shown in Table 3-19. Serum elimination values were similar for
males and females, independent of dose administered: distribution appeared to be mostly
extracellular.
TABLE 3-19. PFOS pharmacokinetic data summary for mice3
Species
CD-I
mice
Time
evaluated
after last
dose
20 weeks
Route
Oral
Sex
M
F
Amount
K^PFOS
(mg/kg)
1
20
1
20
Mean
serum
Ti/2 by
dose
(days)
42.8
36.4
37.8
30.5
Mean
serum
Ti/2 by
sex
(days)
39.6
34.2
Mean
serum
Ti/2 by
species
(days)
36.9
Mean
serum Vd
by dose
(mL/kg)
290.0
263.0
258.0
261.0
a Data from Chang et al., 2012
Table 3-20 summarizes the half-life data from the studies discussed above. Despite the
limitation that the half-life values from most animal studies were derived from administration of
only one dose (Chang et al., 2012), consistency was found in the half-lives for males and females
for the monkeys, rats, and mice. In rats, this is in contrast to the results observed for PFOA,
where there is a much longer half-life in males than in females. However, similar to PFOA, the
half-life of PFOS in humans is much greater than that in laboratory animals. A measure of PFOS
half-life in a retired worker population is 5.4 years (Olsen et al., 2007) compared with several
months in the laboratory animals.
The animal data summarized in Table 3-20 show fairly consistent half-life values
following single and multiple dosing regimens in both the rat and monkey probably due to the
relatively long follow-up in both species after the last dosing was given. In the rat, half-lives for
males and females were nearly identical at 48.2 and 46.9 days, respectively, after 28 days of
dosing and 10 weeks of follow-up (Butenhoff and Chang, 2007). These results for rats were
more consistent between sexes than those half-life values calculated after a single oral dose
(Chang et al., 2012). In monkeys similar half-life values were found after either a single
intravenous dose (Chang et al., 2012) or repeated oral dosing for 182 days (Seacat et al., 2002).
Half-life values for male and female monkeys from Chang et al. (2012) were calculated from the
serum concentrations measured over 23 weeks, while the value from Seacat et al. (2002) was
estimated from the elimination curves.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-30
-------
TABLE 3-20. Summary of Half-life Data
Source
Spliethoffetal.
2008
3M Co 2000
Olsen et al.
2007
Butenhoff and
Chang, 2007
Chang et al.
2012
Seacat et al.
2002
Human
4.1 years
4-8. 67 years
5 .4 years
Monkey
132 days (m)
110 days (f)
200 days (m/f)
Rat
48.2 days (m)
46.9 days (f)
39.8 days (m)
66.7 days (f)
Mouse
39.6 days (m)
34.2 days (f)
Strain
Infants
Occupational
Occupational
SD; 28 days
oral
SD; single oral
dose
CD-I; single
oral dose
Cynomolgus;
single IV dose
Cynomolgus;
oral, 1 82 days
3.5.3 Volume of Distribution Data
Humans
None of the available studies provide data for calibration of volume of distribution of
PFOS in humans. However, several researchers have attempted to characterize PFOS exposure
and intake in humans (Thompson et al., 2010; Egeghy and Lorber, 2011) through
pharmacokinetic modeling. As an integral part of model validation, the parameter for volume of
distribution of PFOS within the body was calibrated from the available data. In the models
discussed below, volume of distribution was defined as the total amount of PFOS in the body
divided by the blood or serum concentration.
Both research groups defined a volume of distribution for humans for use in a simple,
single compartment, first-order pharmacokinetic model (Thompson et al., 2010; Egeghy and
Lorber, 2011). The models developed were designed to estimate intakes of PFOS by young
children and adults (Egeghy and Lorber, 2011) and the general population of urban areas on the
east coast of Australia (Thompson et al., 2010). In both models, the volume of distribution was
calibrated using human serum concentration and exposure data from NHANES and assumed that
most PFOS intake is from contaminated drinking water. Thus, the value for volume of
distribution was calibrated so that model prediction of elevated blood levels of PFOS matched
those seen in the study population.
Thompson et al. (2010) used a the single compartment, 1st order pharmacokinetic model,
as described previously, to predict PFOS concentration in blood serum as a function of dose,
elimination rate, and volume of distribution. The volume of distribution was first obtained for
PFOA by calibrating human serum and exposure data. The volume of distribution for PFOS
(230 mL/kg) was adjusted from the calibrated PFOA data by 35% in accordance with the
differences in PFOA and PFOS volumes of distribution calculated by Anderson et al. (2006).
The original Anderson et al. (2006) model was developed from oral data in monkeys and
optimized a volume of distribution of 220 mL/kg for PFOS and 140 mL/kg for PFOA; thus, the
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
3-31
-------
volume of distribution in monkeys for PFOS was approximately 35% greater than that for PFOA
in the optimized models. Therefore, Thompson et al. (2010) used a volume of distribution of
230 mL/kg for humans in their model.
Egeghy and Lorber (2011) used high and low bounding estimates of 3000 mL/kg and 200
mL/kg for volume of distribution since data in humans were not available. The two separate
estimates of volume of distribution were used in a first-order, one compartment model to
estimate a range of intakes of PFOA. They concluded that the volume of distribution was likely
closer to the lower value based on a comparison of predicted modeled intake with estimates of
intakes based on exposure pathway analyses. Use of the lower value gave a modeled intake
prediction similar to that obtained by a forward-modeled median intake based on exposure
assessment. Thus, the authors concluded that the lower value of 200 mL/kg was more
appropriate for use in their analysis.
Both of the models described above used a volume of distribution calibrated from actual
human data on serum measurements and intake estimates. A calibration parameter obtained
from human studies, where constant intake was assumed and blood levels were measured, is
considered a more robust estimate for volume of distribution than that optimized within a model
developed from animal data.
Animals
The series of pharmacokinetic studies on rats, mice, and monkeys described above, also
included volume of distribution calculations (Chang et al., 2012). Values for all species were
calculated following a single oral or IV dose of PFOS. As discussed below, the volume of
distribution values reported for male and female monkey, female rats, and male and female mice
were reasonably similar. The reason for some much higher values found in male rats could not
be explained by the study authors.
For male and female cynomolgus monkeys, the volume of distribution was 202 and 274
mL/kg, respectively (Table 3-17), following a single IV dose of 2 mg/kg (Chang et al., 2012).
Animals in this pharmacokinetic study were evaluated up to 23 weeks after dosing and the
resulting volumes of distribution are similar to that calibrated from human data described above
for Thompson et al. (2010).
Volume of distribution findings by Chang et al. (2012) for rats are shown above in Table
3-18. Those values calculated from a follow-up duration of only 24 hours are not considered
reliable. In studies with a longer follow-up after dosing, the values for male rats were 275, 666,
and 1228 mL/kg and, for female rats, values were 468 and 484 mL/kg. The volume of
distribution was notably greater for male rats, with the exception of one value, than that of
females or other species including humans. As noted, the authors could not explain the higher
value for the male rat but concluded that the volume of distribution for monkeys, rats, and mice
is likely in the range of 200-300 mL/kg.
Data for mice (Chang et al., 2012) are shown in Table 3-20. For males and females the
volume of distribution was 263-290 mL/kg and 258-261 mL/kg, respectively, following a single
oral dose.
Perfluorooctane sulfonate (PFOS) - February 2014 3-32
Draft - Do Not Cite or Quote
-------
Pharmacokinetic models based on animal data described previously in this chapter
(Anderson et al., 2006; Tan et al., 2008) generally optimized the value for volume of distribution
based on model output. The original Anderson et al. (2006) model was developed using data
from Seacat et al. (2002) on serum PFOS concentrations in cynomolgus monkeys following oral
dosing. The volume of distribution in this model was 220 mL/kg.
Perfluorooctane sulfonate (PFOS) - February 2014 3-33
Draft - Do Not Cite or Quote
-------
4.0 HAZARD IDENTIFICATION
4.1 Human Effects
Studies on PFOS exposures are available from both occupational populations and the
general population. Several epidemiologic studies have recently reported associations with
PFOS and cholesterol, birth weight changes and various thyroid parameters. These studies are
discussed below.
4.1.1 Short-Term Studies and Case Reports
Intentional and Accidental Acute Ingestion
No studies of acute accidental or intentional human exposures to PFOS by ingestion were
identified.
Acute and Short-Term Inhalation Exposure
No studies of acute and short-term inhalation human exposures to PFOS were identified.
4.1.2 Long-Term and Epidemiological Studies
Several long-term epidemiological studies have been conducted to investigate possible
associations between PFOS exposures and various health outcomes. Occupational studies
conducted at the 3M Decatur, Alabama plant that produced PFOS are available as well as studies
from the C8 Health Project in Ohio and West Virginia. These studies and those on other
population groupings are described below. The studies that examined systemic endpoints of
toxicity other than cancer are described in Section 4.1.2.1 below where they are subcategorized
according to the endpoints examined.
4.1.2.1 Noncancer Systemic Toxicity Studies
Cholesterol, Lipid Homeostasis, Uric Acid, and Biochemical Toxicity Studies
Occupational Populations
Cross-sectional as well as a longitudinal analysis of medical surveillance data from the
3M Decatur, Alabama and Antwerp, Belgium plants were conducted to evaluate possible
associations between PFOS levels and hematology, clinical chemistry and hormonal parameters
(Olsen et. al., 1999; Olsen et. al, 2001b, 2001c). Male volunteers working at the Decatur plant in
1995, 1997 and 2000 and male and female Antwerp volunteers from a 2000 medical surveillance
study underwent clinical chemistry tests which evaluated hepatic enzyme activity, renal function
tests, thyroid activity and cholesterol levels. There were no consistent associations between
worker PFOS levels and any of the clinical chemistry tests in the 1995 and 1997 analyses. The
analysis of the data from the 2000 surveillance period indicated a positive association between
PFOS and T3, cholesterol, triglycerides and the activity of several hepatic enzymes among male
employees at the Decatur plant. However, there were many limitations to combining and
comparing the data from the various surveillance periods.
Perfluorooctane sulfonate (PFOS) - February 2014 4-1
Draft - Do Not Cite or Quote
-------
A longitudinal analysis of the above data was performed to determine whether
occupational exposure to fluorochemicals over time was related to changes in clinical chemistry
and lipid results in employees of the Antwerp and Decatur facilities (Olsen, et al., 2001c). The
medical surveillance data from 1995, 1997, and 2000 were analyzed using multivariable
regression. The three subcohorts included those who participated in two or more medical exams
between 1995 and 2000. When male employees from both plants were combined, no statistically
significant (p < 0.05) associations were observed over time between PFOS and serum cholesterol
or triglycerides. There were no significant associations between PFOS and changes over time in
high density lipoprotein (HDL), alkaline phosphatase, gamma-glutamyl transpeptidase (GGT),
aspartate aminotransferase (AST), or alanine transaminase (ALT) activities, total bilirubin, or
direct bilirubin.
General Population
Several studies on general population exposures to PFOS and other perfluorochemicals
have been generated by the C8 Health Project. The C8 Health Project was conducted in 2005 to
2006 to collect health data from approximately 69,000 residents in Ohio and West Virginia
living in the vicinity of a chemical plant producing PFOA. Public drinking water was
contaminated in six water districts surrounding the plant (>0.05 ng/mL of PFOA). The levels of
PFOS in this population were similar to those reported in the general U.S. population (median
20.2 ng/mL). Residents were eligible to participate in the study if they had consumed water from
any of the 6 water districts for at least one year prior to the study. Blood samples were collected
from the participants to determine PFOA and PFOS serum levels and clinical chemistry was
performed. Extensive questionnaires were administered as well.
Steenland et al. (2009) examined serum PFOS and PFOA levels and lipids among 46,294
residents, > 18 years old, participating in the C8 Health Project. The mean serum PFOS level
among participants was 22.4 ng/mL (0.022 ppm), range (0.25-759.2 ng/mL). Lipid outcomes
(total cholesterol, high density lipoprotein [HDL] cholesterol, low density lipoprotein [LDL]
cholesterol and triglycerides) were examined in relation to PFOS and PFOA serum levels. All
lipid outcomes, except for HDL, showed significant increasing trends with increasing PFOS
levels. The predicted increase in cholesterol from lowest to highest PFOS decile (0 to 60 ng/mL)
was 11-12 mg/dL. Logistic regression analyses indicate statistically significant increases in
cholesterol (> 240 mg/dL) with increasing PFOS serum levels. Cholesterol levels > 240 mg/dL
are characterized as high, medical intercession is recommended. The odds ratios across quartiles
for cholesterol > 240 mg/dL were 1.00, 1.14 [95% CL1.05, 1.23], 1.28 [95% CI: 1.19, 1.39] and
1.51 [(95% CI: 1.40, 1.64]). The cross-sectional design of this study, as well as the lack of
cumulative exposure measurements are limitations in the study design. In addition, the
mechanism by which perfluorinated compounds impact serum cholesterol is not yet understood.
Steenland et al. (2010) reported on another analysis of the C8 Health Project participants
> 20 years old (n= 54,951) for a possible association between PFOS (and PFOA) serum levels
and uric acid. Elevated uric acid is a risk factor for hypertension and may be an independent risk
factor for stroke. A statistically significant (p< 0.0001) trend was observed between increasing
PFOS levels (untransformed) and uric acid levels. A 0.2-0.3 |ig/dL increase in uric acid was
associated with an increase from the lowest to highest PFOS decile (10-50 ng/mL).
Hyperuricemia (> 6.0 mg/dL for women and > 6.8 mg/dL for men) risk by quintiles increased
Perfluorooctane sulfonate (PFOS) - February 2014 4-2
Draft - Do Not Cite or Quote
-------
slightly withPFOS levels (OR 1.00, 1.02, 1.11, 1.19 and 1.26). The serum of C8 study
participants included several PFCs; PFOA appeared to have a greater influence on uric acid
trends than PFOS in the models employed by Steenland et al. (2010).
Modest associations between PFOS and some lipids existed in children involved in the
C8 Health Project (Frisbee et al. 2010). The report stated that 12,476 community children < 18
years old that lived in the C8 Health Project communities were tested for total cholesterol, LDLs,
HDLs and fasting triglycerides. The mean level of PFOS was 22.7 ng/mL (0.023 ppm). PFOS
was significantly associated with increased total cholesterol, HDL-cholesterol and LDL-
cholesterol in a linear regression analysis after adjustment for co-variables. A statistically
significant increased risk of abnormal total cholesterol and LDL-cholesterol was also observed
between the first and fifth quintiles of PFOS serum concentrations. No trends were observed
with triglycerides. As with the other C8 project data, the authors acknowledge that the cross-
sectional nature of this study limits causal inference.
Gallo et al. (2012) investigated the correlation between serum PFOS levels and liver
enzymes in a total of 47,092 samples collected from members enrolled in the C8 Health Project.
The association of ALT, GOT, and direct bilirubin with PFOS was assessed using linear
regression models adjusted for various confounders. The In-transformed values of ALT were
significantly associated with In-transformed PFOS levels and showed a steady increase in fitted
levels of ALT per decile of PFOS, leveling off after approximately 30 ng PFOS/mL. Fitted
values of GGT showed no overall association with In-transformed PFOS levels. A positive
association was seen with direct bilirubin and PFOS levels in linear regression models but this
was not evident with logistic regression models. Limitations of the study include the cross-
sectional design and self-reported lifestyle characteristics. Only a small number of ALT values
were outside the normal range making the results difficult to interpret in terms of health.
Two studies used data from the U.S. National Health and Nutrition Examination Survey
(NHANES) to examine the relationship between perfluorinated chemicals, serum lipid
measurements and related conditions. One study analyzed NHANES 2003-2004 data and
examined cholesterol levels, obesity, and insulin resistance in relation to PFOS (and other PFCs)
(Nelson et al. 2010). The other study analyzed NHANES data from 1999-2000 and 2003-2004
in adolescents and adults and glucose homeostasis and metabolic syndrome (Lin et al. 2009).
Nelson et al. (2010) analyzed PFOS and three other perfluorinated chemicals and total
cholesterol, HDLs, non-HDL lipoproteins, and LDL. LDL was available only for a subsample of
the fasting population and was not measured directly, but was estimated by the Friedewald
formula as recommended by CDC. Homeostatic model assessment (HOMA) was used to assess
insulin resistance (calculated from fasting insulin and fasting glucose measurements collected in
NHANES). Body mass index (BMI) and waist circumference were used to measure body size.
Exclusion criteria included current use of cholesterol-lowering medications, participants over the
age of 80, pregnant/breastfeeding women or insulin use. After exclusion criteria, ~ 860
participants were included in the cholesterol analyses. The mean PFOS serum concentration for
participants 20-80 years old was 25.3 |ig/L (range, 1.4 to 392 |ig/L).
A positive association was identified between total serum cholesterol and serum PFOS
concentrations. When analyzed by PFOS serum quartiles, adults in the highest PFOS quartile
Perfluorooctane sulfonate (PFOS) - February 2014 4-3
Draft - Do Not Cite or Quote
-------
had total cholesterol levels of 13.4 mg/dL (95% CI, 3.8, 23.0) higher than those in the lowest
quartile. As expected, non-HDL cholesterol accounted for most of the total cholesterol.
Consistent trends were not observed for HDL or LDL. Adjusting the cholesterol models for
serum albumin produced similar results. Body weight and insulin resistance were not
consistently associated with serum PFOS levels.
Lin et al. (2009) analyzed the data on perfluorinated chemicals in NHANES 1999-2000
and 2003-2004 and glucose homeostasis and metabolic syndrome. The authors reported an
association between PFOS and metabolic syndrome; however, it was not reported how PFOS and
other perfluorinated chemicals in the analysis were combined for the NHANES sample years.
Effects of PFOS on plasma lipid levels in the Inuit population of Northern Quebec were
examined in a cross-sectional epidemiology study (Chateau-Degat et al., 2010). The relationship
between consumption of PFOS-contaminated fish and wild game with blood lipids was assessed
in 723 Inuit adults (326 man and 397 women). This traditional diet is also rich in n-3-
polyunsaturated fatty acids (n-3 PUFAs) which are known to have hypolipidemic effects;
therefore, the n-3 PUFAs were considered as a confounder in the analyses. Multivariate linear
regression modeling was used to evaluate the relationship of PFOS levels and blood lipids
including TC, HDL cholesterol, LDL cholesterol, and triacylglycerols. Plasma levels of HDL
cholesterol were positively associated with PFOS levels, even after adjustment for circulating
levels of n-3 PUFAs, but the other blood lipids were not associated with PFOS levels. The
geometric mean level of PFOS in plasma for women and men was 18.6 |ig/L.
Table 4-1 provides a summary of the results from the studies that examined the
relationship of serum PFOS with serum lipids and uric acid. The most consistent findings are
those for total cholesterol.
TABLE 4-1. Association of Serum PFOS with Serum Lipids and Uric Acid
Study
TC
VLDL
LDL
HDL
Non-
HDL
TG
UA
Occupational Population
Olsenetal.,2001c
<— >
NM
NM
<— >
NM
<— >
NM
General Population
Steenland et al., 2009;
2010
Frisbee et al., 2010
(children)
Chateau-Degat et al.,
2010
Nelson etal., 2010
t
t
<— >
t
NM
NM
NM
NM
t
t
<— >
<— >
<— >
t
t
<— >
NM
NM
NM
t
t
•^H>
<— >
NM
t
NM
NM
NM
1= positive association; \,= negative association; <-»= no association; TC= total cholesterol; VLDL=very low density lipoprotein;
LDL= low density lipoprotein; non-HDL= TC(VLDL,IDL, LDL)-HDL; HDL=high density lipoprotein; TG= triglycerides;
UA=uric acid; NM= not monitored
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-4
-------
4.1.2.2 Reproductive Hormones and Reproductive/Developmental Studies
Olsen et al. (2009) recently reviewed the epidemiological literature and identified six
studies examining potential associations between PFOS and human birth outcomes. Research
has focused on birth weight and other measures of fetal growth. Overall, in both the general and
occupational populations, inconsistent associations were identified between PFOS
exposure/levels and fetal birth weight or gestational age.
An occupational cohort study by Grice et al. (2007) examined the potential association
between PFOS exposure and adverse pregnancy outcomes in employees at a perfluorinated
chemical production facility in Decatur, Alabama. Current and former female employees of the
facility completed a questionnaire and provided a brief pregnancy history. The level of exposure
was categorized according to a job-specific exposure matrix. A total of 263 women participated
and reported 439 births of which there were 421 live births, 14 stillbirths and 4 no data. Birth
weight was adjusted for maternal age, smoking status and gravidity. No associations were
observed between PFOS exposure and pregnancy outcomes even when birth weight was adjusted
for maternal age, smoking status and gravidity.
A series of longitudinal, population-based studies was conducted in a subset of women
aged 25 to 35 enrolled in the Danish National Birth Cohort (DNBC) from March 1996 to
November 2002 (Fei et al., 2007, 2008a, 2008b, 2009, 2010a). A random sample of 1400
women was selected to investigate the association between blood levels of perfluorinated
chemicals and adverse reproductive and developmental outcomes in the women and their
children. Study data were collected by telephone interviews at 12 and 30 weeks of gestation and
approximately 6 and 18 months after birth. A food frequency questionnaire was filled out at
home during approximately week 25 of pregnancy. Maternal blood samples were taken in the
first and second trimester and from the cord blood in the infant just after birth. Only blood
results from the 1400 women in the first trimester were reported. Mean plasma PFOS levels by
age groups were: < 25 years- 38.6 ng/mL (0.039 ppm); 25-29 years- 36.8 ng/mL (0.037 ppm);
30-34 years- 33.9 ng/mL (0.034 ppm) and > 35 years- 33.0 ng/mL (0.033 ppm). Potential
confounders included: maternal age, maternal occupation and educational status, parity, pre-
pregnancy BMI, smoking/alcohol consumption during pregnancy, gestational weeks at blood
draw, child's sex, child's age at interview with mother, breast-feeding > 6 months (for 18-month
interview), out-of-home child care, hours mother spent with child/day, and home density (the
total number of rooms divided by the total number of people in the household).
Data from the DNBC were used to investigate the association between plasma levels of
PFOS in pregnant women, length of gestation, and the infant's birth weight (Fei et al., 2007).
The average PFOS levels in maternal plasma were 35.3 ng/mL (range, 6.4-106.7 ng/mL). The
data were adjusted for confounding factors that might also influence fetal growth or length of
gestation and analyzed by analysis of variance and linear regression using both continuous PFOS
concentrations, and categorical quartiles of PFOS. Elevated risk estimates were found for PFOS
levels and preterm birth, but the odds ratio was significant only for the third quartile of exposure.
No significant association was found between PFOS levels and length of gestation, low birth
weight, or small for gestational age.
Perfluorooctane sulfonate (PFOS) - February 2014 4-5
Draft - Do Not Cite or Quote
-------
When Fei et al. (2008a) investigated the association between PFOS levels in pregnant
women and their newborns and placental weight, birth length, and head and abdominal
circumference, he found maternal PFOS levels were not associated with any of the fetal growth
indicators when the lowest quartile was compared to the highest. In a stratified analysis of
PFOS, inverse associations were found with birth length for postterm and preterm infants and
with ponderal index in multiparous women and positive association with ponderal index in
nulliparous women. These associations were not statistically significant.
In the Fei et al. (2008b) report examining the association between plasma levels of PFOS
in pregnant women and the motor and mental development in their children, regression analysis
did not indicate any significant association between PFOS and Apgar score after adjustment for
potential confounders (odds ratio [OR], 1.20; 95% confidence interval [CI], 0.57-2.25). The
developmental measures examined in the infants included Apgar score of child at birth and
questionnaire responses about child development milestones at 6 and 18 months. Data from the
6 month interview did not show any association between PFOS levels and motor or mental
development. In children at 18 months, mothers with higher PFOS levels were slightly more
likely to report that their babies started sitting without support at a later age and "did not use
word-like sounds to tell what he/she wants."
Fei et al. (2009) used the same population of pregnant women to determine if there was
any association with PFOS levels and fecundity indicated by the time to pregnancy (TTP). In
women who had planned pregnancies (n=1240), there was a longer time to pregnancy (TTP)
with higher levels of PFOS (p<0.001). PFOS was also associated with irregular menstrual
periods (11.6% in the lowest quartile versus 14.2% in the upper three quartiles). The proportion
of women with infertility (TTP>12 months) was higher in the upper three quartiles of PFOS
versus the lowest quartile. These trends were significant. Women with longer TTP were also
older and had a history of spontaneous miscarriages or irregular menstrual cycles. The
biological mechanism by which PFOS may reduce fecundity is unknown.
Fei et al. (2010a) reported on the effects of PFOS on the length of breastfeeding. Self
reported data on the duration of breastfeeding was collected during the telephone interviews at 6
and 18 months after birth of the child. Higher levels of PFOS were significantly associated with
a shorter duration of breastfeeding. In multiparous women, the adjusted OR for weaning before
6 months is 1.20 [95% CI, 1.06-1.37]) for each 10 ng/mL increase in PFOS concentration in the
maternal blood and the increase was dose-related. A statistically significant positive trend was
observed for women having their first child, but no consistent association was found across
increasing blood levels of PFOS. The authors speculate that the observed associations may be
non-causally related to previous length of breastfeeding or to reduction of PFOS through
lactation.
Monroy et al. (2008) found no association between the maternal serum levels of PFOS
and the birth weight of the neonates in 101 pregnant women enrolled in a large cohort study,
Family Study, conducted at McMaster University Medical Center in Canada. PFOS was
measured in maternal serum at mid-pregnancy and delivery in 101 healthy pregnant women in
Ontario, Canada and in umbilical cord blood (UCB) from 105 babies. PFOS was detected in
100% of samples with mean levels of 18.3, 16.2, and 7.2 ng/L in maternal serum at 24-28 weeks,
maternal serum at delivery, and in UCB respectively. The concentration of PFOS in maternal
Perfluorooctane sulfonate (PFOS) - February 2014 4-6
Draft - Do Not Cite or Quote
-------
serum was significantly higher than in UCB. No significant association between levels of PFOS
in the maternal serum or UCB and the birth weight of the neonates was found. Maternal PFOS
levels were also not associated with maternal body mass index, gestational length, or gender.
A prospective cohort study was conducted on birth weight between July 2002 and
October 2005 at the Sapporo Toho Hospital in Hokkaido, Japan that included 428 native
Japanese women and their infants (Washino et al., 2009). Women enrolled were at 23-35 weeks
of gestation with a mean age of 30.5 years. Subjects reported on dietary habits, smoking status,
alcohol intake, caffeine intake, household income and educational level. At the time of the
questionnaire, a blood sample was obtained. Both PFOS and PFOA were analyzed in the blood
using liquid chromatography-tandem mass spectrometry coupled with solid-phase extraction.
The mean concentration in the women was 5.6 ng/mL (0.006 ppm) PFOS with every sample
having detectable PFOS. The highest PFOS concentration identified was 16.2 ng/mL (0.016
ppm). The results indicated that in utero exposure to PFOS negatively correlated with birth
weight in female infants only.
A cohort study on 252 pregnant women (> 18 years old) at 15-16 weeks gestation in the
city of Edmonton, Alberta, Canada was undertaken to examine a possible association between
perfluorinated chemicals, fetal growth and gestational age (Hamm et al., 2009). Serum samples
collected from December 2005 to June 2006 showed PFOS levels ranging from nondetectable to
35 ng/mL, with the mean and geometric mean being 9.0 ng/mL (0.009 ppm) and 7.4 ng/mL
(0.0074 ppm), respectively. Overall, there was no association with the level of PFOS and birth
weight or length of gestation. Mean birth weight was 3278 g (n = 83; PFOS < 6.1 ng/mL); 3380
g (n = 83; PFOS 6.1-10.0) and 3387 g (n = 86; PFOS > 10-35). Mean length of gestation for all
groups was 38 weeks; the pre-term delivery percentage was similar between groups.
Stein et al. (2009) presented on serum levels of PFOS and self-reported pregnancy
outcomes of a population of women (5,262 pregnancies; ages 15-55 years) in the Mid-Ohio
Valley in 2000-2006. These women were enrollees in the C8 Health Project, a community
health study of residents near a chemical plant that used PFOA in the manufacture of
fluoropolymers. Pregnancies within the 5 years preceding the exposure measurements were
analyzed. The mean level of PFOS in the serum of these women was 14.1 ng/mL (0.014 ppm).
There was no association between PFOS levels and miscarriages or pre-term births. PFOS was,
however, associated with an increased risk above the median (adjusted odds ratio =1.5: 95%
confidence interval: 1.1, 1.9) for low birth weight, and a dose-response relationship was reported
for the 50th-75th, 75th-90th and > 90th percentile serum PFOS exposure concentrations (adjusted
OR 1.3, 1.6, 1.8, respectively). PFOS was also weakly associated with pre-eclampsia (adjusted
odds ratio = 1.3, 95% confidence interval: 1.1, 1.7). The self-reported nature of pregnancy
outcomes in this study is a limitation.
Using the C8 Health Project data, blood samples from a population of women aged 18-65
years (n= 25, 957) were analyzed to determine if the onset of menopause, serum estradiol and the
amount of PFCs in the blood were inter-related (Knox et al., 2011). These data were cross-
sectional with no variable for the length of exposure. The mean PFOS level of all the women
was 17.6 ng/mL. Data were eliminated for those reporting undergoing a hysterectomy and
adjusted for age within the group, smoking, alcohol consumption, body mass index and exercise.
The analysis for menopause was determined upon three groups of women: childbearing (ages 30-
Perfluorooctane sulfonate (PFOS) - February 2014 4-7
Draft - Do Not Cite or Quote
-------
42), perimenopausal (ages >42 but < 51) and menopausal (ages > 51 or < 65). These same
groups were used for the estradiol concentrations except the childbearing group was extended to
include those > 18 years. For menopause, the odds of having experienced menopause in the
menopausal group exposed to PFOS showed a monotonic increase with all quintiles significantly
higher relative to the lowest (PFOS odds= 2.1, CI=1.6-2.8). Also, PFOS was negatively
associated with estradiol concentrations in all groups but significantly in the perimenapausal
group (P = -3.65; p< 0.0001) and menopausal group (P = -0.83; p< 0.007). The level of PFOS
was significantly higher in the set of women that had undergone a hysterectomy. While these
relationships were associated with PFOS, the authors still recommended caution when
interpreting results because the data are cross-sectional. Lopez-Espinosa et al. (2011) also used
the C8 Health Project data base to indicate that in 3076 boys and 2931 girls, ages 8-18 years,
there was a relationship of reduced odds of reaching puberty (raised testosterone) with increasing
PFOS levels, having a delay of 190 days between the highest and lowest quartile, and reduced
odds of postmenarche (138 days delay) in girls. This study suggested that PFOS exposure
correlated with a delay in puberty.
Joensen et al. (2009) investigated the possible associations between perfluoroalkyl acids
(PFAAs) and semen quality in 105 Danish men. The median PFOS serum level in men was 24.5
ng/mL. Men with high levels of combined levels PFOA/PFOS had a median level of 6.2 million
spermatozoa compared to 15.5 million in men with low PFOA/PFOS levels. There was no
significant association between testosterone levels and PFAA levels and no difference in PFAA
levels between high and low testosterone groups.
The relationships examined and outcomes from the studies that examined reproductive or
developmental endpoints as they related to serum PFOS concentrations are summarized in Table
4-2 that follows. For most endpoints no associations were identified. There is some evidence for
body weight effects in neonates (2 of 5 studies). Possitive associations were also noted for some
developmental endpoints and factors impacting fertility.
Perfluorooctane sulfonate (PFOS) - February 2014 4-8
Draft - Do Not Cite or Quote
-------
TABLE 4-2. Association of serum PFOS with reproductive and developmental
outcomes
Study
Grice et al.,
2007
Fei et al., 2007;
2008a; 2008b;
2009; 2010a
Monroy et al.,
2008
Washino et al.,
2009
Hamm et al.,
2009
Stein et al.,
2009
Knox et al.,
2011 and
Lopez-Espinosa
etal.,2011
Joensen et al.,
2009
(PFOA/PFOS
combined)
Population
Occupational
General
General
General
General
General
General
General
Outcome
<-»• (any
adverse)
<-»• (gestation
length)
<-»• (length of
breastfeeding)
<-»• (gestation
length)
NM
<-»• (gestation
length)
(miscarriage)
NM
NM
Measures
at birth
NM
<— >
(weight)
<-»• (size)
<-> (Apgar
score)
<— >
(weight)
t (low
weight
females
only)
<— >
(weight)
t (low
weight)
NM
NM
Growth/
Development
NM
<-»-(at6
months)
t (at 18
months; sitting
up later)
NM
NM
NM
NM
t (delayed
puberty)
NM
Fecundity/
Fertility
NM
t (time to
pregnancy)
t (infertility)
NM
NM
NM
NM
t (early
menopause)
t (lower
sperm count)
<— >
(testosterone)
1= positive association; \,= negative association; <-»= no association; NM = Not Monitored
4.1.2.3 Thyroid Effect Studies
The relationship between exposure to polyhalogenated compounds, including PFOS, and
thyroid hormone homeostasis was examined in a cross-sectional study of the adult Inuit
population of Nunavik, Quebec, Canada (Dallaire, 2009b). Concentrations of thyroid-
stimulating hormone (TSH), free thyroxine (fT4), total triiodothyronine (tT3) and thyroxine-
binding globulin (TBG) were measured in 623 individuals. Participants were given a survey to
indicate smoking status, frequency of alcohol use, medications taken, and dietary fish
consumption. Those using medication for thyroid disease and pregnant women were not
included in the results. The study detected PFOS in 100% of individuals with a mean plasma
PFOS concentration of 18 ng/ml (95% CI, 17-19 ng/ml). PFOS was negatively associated with
circulating levels of TSH, tT3 and TGB and positively associated with nv The results suggest
that human thyroid hormone levels may be affected by PFOS exposure. However, because the
majority of individuals had normal thyroid gland function, it is uncertain whether these
relationships are connected to thyroid disease.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-9
-------
National Health and Nutrition Examination Survey (NHANES) data from three
independent cross-sectional cycles (1999-2000; 2003-2004 and 2005-2006) were analyzed by
Melzer et al. (2010) to estimate associations between serum PFOA and PFOS concentrations and
thyroid disease in the general U.S. population. Overall, a total of 3,966 individuals > 20 years of
age (1900 men and 2066 women) were included. Of these, 292 women and 69 men reported
thyroid disease. The data showed that men with PFOS levels in the highest quartile > 36.8
ng/mL were more likely to report currently treated thyroid disease than men with PFOS levels in
the lowest quartile < 25.5 ng/mL (OR=2.68; 95% CI, 1.03-6.98; p=0.043). Women had lower
levels of PFOS than men and higher prevalence of thyroid disease, but serum PFOS
concentration was not significantly associated with treated thyroid disease. Further studies
measuring thyroid hormone levels in a larger sample population could clarify whether pathology,
changes in exposure, or altered pharmacokinetics can explain the association.
Bloom et al. (2010) examined the potential association between serum concentrations of
eight polyhalogenated compounds, including PFOS, and human thyroid function. Levels of
thyroid stimulating hormone (TSH) and free thyroxine (fT4) were measured in a subsample of
participants in the cross-sectional New York State Angler Cohort Study (27 men and 4 women).
A survey was conducted to determine smoking status, history of thyroid disease, medications
used, and dietary fish consumption. None of the participants reported a thyroid condition or the
use of thyroid medication. PFOS occurred at a high concentration compared to the other PFCs
measured with a mean concentration of 19.6 ng/mL (95% CI, 16.3-23.5). The results indicated
no significant association between PFOS serum concentration and thyroid hormone levels,
potentially due to the study's small sample size.
The potential relationship between PFOS exposure and thyroid disease was investigated
by Pirali et al. (2009) in a sample of 28 patients undergoing thyroid surgery (22 benign and 6
malignant) and a control group of 7 patients with no evidence of thyroid disease. PFOS was
detected in thyroid tissue in 100% of the 8 males and 20 females with thyroid disease, with a
median PFOS concentration of 5.3 ng/g, with no significant difference in levels between benign
and malignant patients. The median PFOS concentration (4.4 ng/g) in the healthy glands of the
control group was similar to that found in the diseased thyroid samples indicating that there was
no association between PFOS concentration and thyroid disease.
Maternal and umbilical cord blood concentrations of a number of fluorinated organic
compounds, including PFOS, were determined in 15 women (17-37 years of age) and their
newborns at Sapporo Toho Hospitals in Hokkaido, Japan from February to July 2003 (Inoue et
al., 2004). PFOS was detected in 100% of the maternal and cord blood samples with maternal
blood PFOS ranging from 4.9 to 17.6 ng/mL, and cord blood PFOS ranging from 1.6 to 5.3
ng/mL. Thyroid stimulating hormone (TSH) and free thyroxine (fT/t) levels in the infants
between days 4 and 7 of age were not related to cord blood PFOS concentration in this small
study.
Chan et al. (2011) used blood from 974 serum samples collected from women (mean age-
31.3 years) at 15-20 weeks gestation in 2005-2006 in Canada and measured thyroid hormones,
fT4 and the level of PFCs to determine if PFC levels were associated with hypothyroxinemia.
From the samples, there were 96 cases of hypothyroxinemia (identified as 'cases') and 175
controls used. The geometric mean for PFOS was 7.39 ng/mL. The mean free T4 levels were
Perfluorooctane sulfonate (PFOS) - February 2014 4-10
Draft - Do Not Cite or Quote
-------
7.7 pmol/L in the cases and 12.9 in the controls. The mean TSH concentrations were 0.69 mU/L
in the cases and 1.13 in the controls. Analysis by conditional logistic regression indicated that
the concentration of PFOS was not associated with hypothyroxinemia in this set of pregnant
women. For PFOS, the odds ratio for association of hypothyroxinemia with exposure to PFOS
was 0.88 with a 95% CI of 0.63-1.24.
The Dallaire et al (2009b) study provides the strongest evidence for effects on thyroid
hormones as illustrated in Table 4-3. The hormone endpoints were not monitored consistently in
a number of the other sudies of thyroid effects and where monitored exhibited no significant
association.
TABLE 4-3. Association of serum PFOS with the prevalence of thyroid disease and
thyroid hormone levels in studies of general and worker populations
Study
Dallaire et al., 2009b
Bloom etal., 2010
Melzer et al., 2010
Pirali et al., 2009
Inoue et al., 2004
Chan etal. ,2011
Population
General
General
General
General
Newborns
Pregnant
women
Thyroid Disease
<— >
<— >
<-»• (women)
t (men)
<— >
NM
<— >
TSH
1
<— >
NM
NM
<— >
<— >
T3
1
NM
NM
NM
NM
NM
T4
t
<— >
NM
NM
<— >
<— >
1= positive association; |= negative association; <-»= no association; NM = Not Monitored
4.1.2.4 Immunotoxicity
Okada et al. (2012) investigated the relationship between maternal PFOS concentration
and infant allergies and infectious diseases during the first 18 months of life as well as cord
blood IgE levels. The prospective birth cohort was based on infants delivered at the Sapporo
Toho Hospital in Sapporo, Hokkaido, Japan between July 2002 and October 2005. PFOS levels
were measured in maternal serum taken after the second trimester (n=343) and total IgE
concentration was measured in cord blood (n=231) at the time of delivery. Infant allergies and
infectious diseases were assessed in a maternal self-administered questionnaire at 18 months
post-delivery. Polynomial regression analyses, adjusted for potential confounders, were
performed on log-transformed data. Mean maternal PFOS concentration was 5.6 ng/mL and
cord blood IgE level was 0.62 lU/mL. No significant associations were observed between
maternal PFOS levels and cord blood IgE levels or incidence of food allergy, eczema, wheezing,
or otitis media in infants at 18 months of age. Limitations of the study include the small sample
size, potential selection bias of the population, and accurate diagnosis of disease in the infants.
The population from the Danish National Birth Cohort studies by Fei et al. (201 Ob) was
used to determine whether prenatal exposure to PFOS caused an increased risk of infectious
diseases leading to hospitalization in early childhood. No clear pattern was identified when
results were stratified by child's age of infection and the level of PFCs in the maternal blood.
Antibody responses to diphtheria and tetanus toxoids following childhood vaccinations
were assessed in context of exposure to perfluorinated compounds (Grandjean et al., 2012). The
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-11
-------
prospective study included a birth cohort of 587 singleton births during 1999-2001 from the
National Hospital in the Faroe Islands. Serum antibody concentrations were measured in
children at age 5 years prebooster, approximately 4 weeks after the booster, and at age 7 years.
Prenatal exposures to perfluorinated compounds were assessed by analysis of serum collected
from the mother during week 32 of pregnancy; postnatal exposure was assessed from serum
collected from the child at 5 years of age. Multiple regression analyses with covariate
adjustments were used to estimate the percent difference in specific antibody concentrations per
2-fold increase in PFOS concentration in both maternal and 5-year serum.
Maternal PFOS serum concentration was negatively associated with antidiphtheria
antibody concentration (-39%) at age 5 before booster. An effect was also found in comparison
of antibody concentrations at age 7 with serum PFOS concentrations at age 5 where a 2-fold
increase in PFOS was associated with a difference in diphtheria antibody of-28% (95% CI, -
46% to -3%). Additionally at ages 5 and 7, a small percentage of children had antibody
concentrations below the clinically protective level of 0.1 ITJ/mL. At age 5 the odds ratios of
antibody concentrations falling below this level for diphtheria were 2.48 (95% CI, 1.55 to 3.97)
compared with maternal and 1.60 (95% CI, 1.10 to 2.34) compared with age 5 serum PFOS
concentrations. For age 7 antibody levels correlated with age 5 PFOS serum concentrations,
odds ratios for inadequate antibody concentration were 2.38 (95% CI, 0.89 to 6.35) for diphtheria
and 2.61 (95% CI, 0.77 to 8.92) for tetanus.
The association between serum levels of perfluorinated compounds and childhood asthma
was investigated by Dong et al. (2013). The cross-sectional study included a total of 231
children aged 10-15 years with physician-diagnosed asthma and 225 age-matched non-asthmatic
controls. Between 2009 and 2010, asthmatic children were recruited from two hospitals in
Northern Taiwan, while the controls were part of a cohort population in seven public schools in
Northern Taiwan. Serum was collected for measurement often perfluorinated compounds,
absolute eosinophil counts, total IgE, and eosinophilic cationic protein. A questionnaire was
administered to asthmatic children to assess asthma control and to calculate an asthma severity
score (including frequency of attacks, use of medicine, and hospitalization) during the previous
four weeks. Associations of perflourinated compound quartiles with concentrations of
immunological markers and asthma outcomes were estimated using multivariable regression
models.
Nine often perfluorinated compounds were detectable in >84.4% of all children with
levels generally higher in asthmatic children compared with non-asthmatics. Serum
concentrations of PFOS in asthmatic and non-asthmatic children were 45.5±37.3 and 33.4±26.4
ng/mL, respectively; similar levels were measured for perfluorotetradecanoic acid with
concentrations of the remaining compounds much lower. The adjusted odds ratios for asthma
association with the highest versus lowest quartile levels were significantly elevated for seven of
the compounds. For PFOS the odds ratio was 2.63 (95%CI: 1.48, 4.69). In asthmatic children,
absolute eosinophil counts, total IgE, and eosinophilic cationic protein concentration were
positively associated with PFOS levels with a significant monotonic trend with increasing serum
concentration. None of these biomarkers was significantly associated with PFOS levels in non-
asthmatic children. Serum PFOA levels, as well as three other compounds, were significantly
associated with higher asthma severity scores.
Perfluorooctane sulfonate (PFOS) - February 2014 4-12
Draft - Do Not Cite or Quote
-------
4.1.2.5 Carcinogenicity Studies
Occupational Populations
Several analyses of various health outcomes have occurred on cohorts of workers at the
3M Decatur, Alabama plant (Mandel and Johnson, 1995; Alexander et al., 2003; Alexander and
Olsen, 2007). Cancer incidence and mortality have been examined periodically in these workers.
A cohort of 2083 workers employed for at least 1 year was examined. Workers were grouped
into three PFOS exposure categories: non-exposed, low exposed and high exposed. Cumulative
exposures were also estimated using a weighted approach based on biomonitoring data. The
geometric mean serum PFOS levels were 0.9 ppm for chemical plant employees and 0.1 ppm for
non-exposed workers.
A total of 145 deaths were identified with 65 of them in high-exposure jobs.
Standardized mortality ratios (SMRs) were calculated using the state of Alabama reference data
and when analyzing the entire cohort, SMRs were not elevated for most of the cancer types and
for non-malignant causes. SMRs that were above 1 (cancer of the esophagus, liver, breast,
urinary organs, bladder, and skin) were also elevated when the cohort was limited to any
employee ever employed in a high exposure job (except breast cancer). Only 2 or 3 deaths were
reported for each of these cause-specific categories and were not statistically significant, except
for bladder cancer. Three male employees in the cohort died of bladder cancer (0.12 expected).
All were employed at the Decatur plant for more than 20 years, and had worked in high exposure
jobs for at least 5 years. The SMR for bladder cancer for workers who were ever employed in a
high exposure job was 12.77 (0.23 expected, CI = 2.63 - 37.35). When the data were analyzed
for workers with >5 years of employment in a high exposure job, the SMR was 24.49. This
effect remained when the data were analyzed using county death rates.
While the 3 deaths from bladder cancer were greater than the expected number observed
in the general population, there were limitations observed in this study. The number of deaths
was small (especially for females in all categories), all death certificates were not located,
exposures to other chemicals were not accounted for, smoking status was not established and
animal studies do not indicate any bladder cancer.
Based on these results, 3M undertook another study of this cohort to identify bladder
cancer incidence (Alexander and Olsen, 2007). Cancer deaths were ascertained from death
certificates and via questionnaire for bladder cancer cases, year of diagnosis, and smoking
history. Eleven bladder cancer cases were identified; five deaths and 6 incident cases. Only 2 of
the 6 self-reported cases were confirmed with medical records. Five of the 6 cases had a history
of cigarette smoking. Standardized incidence ratios (SIR) were estimated for 3 exposure
categories and compared to US cancer rates. SIRs ranged from less than 1 to 2.72 but none were
statistically significant. The highest SIRs were for the lower exposure groups. The SIR was
1.74 (95% CI 0.64-3.79) for those ever employed in a high exposure job.
Grice et al. (2007) looked for association with PFOS exposure at the 3M Decatur,
Alabama plant to various malignant and benign disorders as well as adverse pregnancy outcomes
(results reported in Section 4.1.2.2). Current and past employees at the plant answered
questionnaires (n = 1400; 1137 male and 263 female) about any diagnosis of cancers or non-
Perfluorooctane sulfonate (PFOS) - February 2014 4-13
Draft - Do Not Cite or Quote
-------
cancerous conditions (including liver, kidney and gastrointestinal disease). Two exposure
models were used in this analysis. The first model grouped workers according to PFOS
exposure: unexposed (< 0.29 ppm), low (0.39-0.89 ppm) or high exposure (1.30-1.97 ppm). The
other model estimated cumulative exposures by using a weighted approach based on
biomonitoring data. Prostate, melanoma and colon cancer were the most frequently reported
malignancies. When cumulative exposure measures were analyzed, elevated odds ratios were
reported for both colon and prostate cancer, however, they did not reach statistical significance.
None of the other health conditions were positively associated with PFOS exposures
General Populations
A prospective Danish cohort study (1993-2006) compared plasma levels of PFOS and
PFOA and the incidence of cancer in 57,053 native Danish individuals (ages 50-65 years)
(Eriksen et al., 2009). The participants had no cancer diagnosis at the time of the enrollment.
The Danish Cancer Registry and the Danish Pathology Data Bank were used to identify 713,
332, 128 and 67 patients with prostate, bladder, pancreatic and liver cancer, respectively,
diagnosed 0-12 years after enrollment in the cohort. A comparison group reflecting the gender
ratio of the case group was also chosen from the original cohort (680 men, 92 women). Potential
confounders associated with each type of cancer were addressed in questionnaires administered
to the participants.
One-time, PFOS and PFOA blood samples were taken at recruitment. Cancer incidence
rate ratios (IRR) were estimated using Cox proportional hazards model, stratified by sex.
Median PFOS levels did not differ significantly between male and female cases (35.1 ng/mL,
males; 32.1 ng/mL, females) or controls. Adjusted IRRs for the 4 types of cancer based on
increasing PFOS quartiles were only above 1 for prostate and pancreatic cancer; none of them
were statistically significant, and no increasing trends were noted.
4.2 Animal Studies
A tabular summary of animal studies is provided in Appendix A, Table A.3.
4.2.1 Acute Toxicity
A limited number of acute studies are available for PFOS. The studies indicate an LD50
of 251 mg/kg and an LCso of 5.2 ppm in rats. PFOS caused no irritation in a dermal irritation
study although limited study details were available. An eye irritation study was also conducted
but few details were provided on effects observed.
Oral Exposure
Rat
Dean et al. (1978) exposed 5 CD rats/sex/dose by gavage to a single dose of 0, 100, 215,
464 or 1000 mg/kg of PFOS suspended in a 20% acetone/80% corn oil mixture. Rats were
observed for abnormal signs for 4 hours after exposure and then daily for up to 14 days. All rats
died in the 464 and 1000 mg/kg group and 3/10 rats died in the 215 mg/kg group. Clinical signs
observed included hypoactivity, decreased limb tone, and ataxia; necropsy results indicated
stomach distension, lung congestion and irritation of the glandular mucosa. Based on the
Perfluorooctane sulfonate (PFOS) - February 2014 4-14
Draft - Do Not Cite or Quote
-------
findings, the acute oral LDso was 233 mg/kg in males, 271 mg/kg in females and 251 mg/kg
combined.
Male Wistar rats (n=2-3/group) were administered a single oral dose of PFOS at 0, 125,
250 or 500 mg/kg and monitored for any neurological signs (Sato et al., 2009). Animals treated
with > 250 mg/kg had decreased body weight or delay of body weight gain during the 14 days
post-exposure. One of three rats in the 250 mg/kg group and both rats in the 500 mg/kg group
died.
Mice
Sato et al. (2009) also studied male ICR mice (n=2-3/group) administered a single oral
dose of PFOS at 0, 125, 250 or 500 mg/kg. Animals treated with > 250 mg/kg had decreased
body weight or delay of body weight gain during the 14 days post-exposure. One mouse in each
dose group died.
Inhalation Exposure
Rusch et al. (1979) exposed five Sprague-Dawley rats/sex/dose to PFOS dust (in air) at
concentrations of 0, 1.89, 2.86, 4.88, 6.49, 7.05, 13.9, 24.09 or 45.97 mg/L for 1 hour. Rats were
observed for abnormal signs prior to exposure, every 15-min during exposure, at removal from
the chamber, hourly for 4 hours after exposure and then daily for up to 14 days. The 45.97 mg/L
group was not used in determining the LCso as this portion of the study was terminated on day 2
due to high mortality; the 13.9 mg/L group was also not part of the calculation as this group was
terminated early due to mechanical problems. All rats died in the 24.09 mg/L group by day 6.
Mortality for the other groups was 0%, 10%, 20%, 80% and 80% in the 1.89, 2.86, 4.88, 6.49
and 7.05 mg/L groups, respectively. Clinical signs observed included emaciation, red material
around the nose or other nasal discharges, dry rales, breathing disturbances and general poor
condition; necropsy results indicated discoloration of the liver and lung. Based on the findings,
the acute inhalation LC50 was 5.2 mg/L (ppm).
Dermal/Ocular Exposure
Rabbit
The only dermal and ocular irritation PFOS studies were performed by Biesemeier and
Harris (1974) and were summarized in OECD (2002) with few details. In the dermal study, six
albino rabbits were treated by placing 0.5 grams of the test material on their intact or abraded
backs, covered and erythema and edema were scored after 24 and 72 hours. The primary
irritation score was zero indicating no irritation or edema. No information was provided on the
guidelines followed, sex of the animals, and the vehicle used. In the ocular study, six albino
New Zealand White rabbits, fitted with Elizabethan collars, were treated with one tenth of a
gram of the test substance instilled in one eye; the other eye was used as the untreated control.
Reaction to the test material was recorded at 1, 24, 48, and 72 hours after treatment; however, the
scale criteria were not presented or referenced. Scores were maximal at 1 hour and 24 hours
after treatment then decreased over the rest of the study; however, the raw data were not
provided.
Perfluorooctane sulfonate (PFOS) - February 2014 4-15
Draft - Do Not Cite or Quote
-------
4.2.2 Short-Term Studies
Oral Exposure
Rat
Five Crl:CD (SD) IGS BR rats/sex/dose level/interim necropsy were administered PFOS
in the diet at concentrations of 0, 0.5, 2.0, 5.0 or 20 ppm for 4 or 14 weeks (further discussion of
the 14-week results is provided in Section 4.2.3) as part of a larger 2-year cancer bioassay design
(Seacat et al., 2003). Doses were equivalent to 0, 0.05, 0.18, 0.37 and 1.51 mg/kg in males and
0, 0.05, 0.22, 0.47 and 1.77 mg/kg in females, respectively. Animals were observed twice daily
for mortality and morbundity with a clinical exam performed weekly. Body weight and food
consumption data were recorded weekly. Food efficiency was determined, and mean daily
intake of PFOS, cumulative dose and percentage of dose were identified in the liver and sera.
Blood and urine were obtained from 10 animals/sex/treatment during week 4 for clinical
chemistry, hematology and urinalysis evaluation. A thorough necropsy was performed at the end
of treatment and liver samples were collected for palmitoyl CoA oxidase (PCoAO) activity, cell
proliferation index (PI) and PFOS concentration analysis. Microscopic analysis of tissues was
performed on the control and high-dose animals. Analysis of PFOS in the liver and sera were
determined by HPLC-MS/MS and results were considered quantitative to ± 30%.
A summary of findings in the study are provided in Table 4-4. For the animals treated for
4 weeks, terminal body weight in the 20 ppm animals was decreased, although not statistically
significant. Absolute liver weight was not affected but relative (to body weight) liver weight was
significantly increased in the high dose males. Food consumption and food efficiency were only
decreased in the 20 ppm females. No treatment-related effects were observed on hematology or
urinalysis; female rats treated with 20 ppm had significant decreases in serum glucose and
increases in aspartate aminotransferase (AST). Analysis of PCoAO activity was weakly
increased (less than 2-fold) when compared to controls in the 20 ppm dose group males in one
laboratory and similar to controls in another laboratory analysis.
Perfluorooctane sulfonate (PFOS) - February 2014 4-16
Draft - Do Not Cite or Quote
-------
TABLE 4-4. Mean (± SD) Values for Select Parameters in Rats Treated for 4 Weeks"
Parameter
PFOS (mg/kg/day)
Males
Body wt (g)
Liver/body wt (%)
PCNA LI (%)
Glucose (mg/dL)
AST (IU/L)
PCoAO (lU/g)
0
323 ±34
3.6 ±0.2
0.042± 0.024
97 ±11
122 ± 26
9.0 ±2.2
0.05
315±16
4.1 ±0.4
0.038 ±0.014
97 ±5
146 ± 29
9.0 ±2.3
0.18
303 ± 25
3.9 ±0.2
0.069 ± 0.028
91±11
104 ± 23
7.0 ±4.0
0.37
309 ±19
3.5 ±0.3
0.043 ± 0.025
94 ±9
114 ±17
8.0 ±0.8
1.51
296 ±21
4.4* ±0.3
0.065 ± 0.029
84* ±5
131 ±20
6.0 ±1.4
Females
Body wt (g)
Liver/body wt (%)
PCNA LI (%)
Glucose (mg/dL)
AST (IU/L)
PCoAO (lU/g)
0
213±21
3.8 ±0.2
0.53± 0.032
114±7
123 ±28
5.0 ±1.5
0.05
192 ±11
3.7 ±0.2
0.055 ±0.015
ll±7b
120 ±37
6.0 ± 1.1
0.22
202 ±15
3.8 ±0.2
0.059 ±0.013
113 ±18
101 ± 12
37.0 ± 1.7
0.47
206 ± 29
3.7 ±0.4
0.097 ±0.036
109 ±11
112 ±24
2.0* ±1.1
1.77
193 ± 17
4.1 ±0.3
0.183 ±0.085
107 ±8
92 ± 16
4.0 ±1.1
a Data from Seacat et al., 2003
b Reviewer suspects this is a typo and should be 111 mg/dL as it was not marked significant and is not in the text.
* Statistically significant from controls, p< 0.05
PCNA LI= proliferating cell nuclear antigen labeling index
IU = international unit
Curren et al. (2008) conducted two 28-day studies in groups of 15 Sprague-Dawley
rats/sex/dose group. In both studies, the animals were administered 0, 2, 20, 50 or 100 mg
PFOS/kg diet which was equivalent to 0, 0.14, 1.33, 3.21 and 6.34 mg/kg/day, respectively, in
males and 0, 0.15, 1.43, 3.73 and 7.58 mg/kg/day, respectively, in females. In the first study
(Study 1), rats were assessed for changes in clinical chemistry, hematology, histopathology, and
gene expression. In Study 2, blood pressure, erythrocyte deformability and liver fatty acid
composition were assessed. Tissues were also analyzed for PFOS residues by liquid
chromatography negative electrospray tandem mass spectrometry (LC-MS/MS). Tissue residue
results showed a dose-dependent increase with most of the PFOS identified in the liver; values
for the PFOS residue levels are reported in Section 3.2 Distribution.
There were no treatment-related differences observed in hematology and urinalysis
parameters. Statistically significant (p < 0.05) decreases in body weight and food consumption
were observed in the males and females administered > 50 mg PFOS/kg diet. Food consumption
was also statistically decreased in males during one week of treatment in the 20 mg PFOS/kg diet
group. No differences in blood pressure measurements were observed across the groups. In red
blood cells from both males and females, deformability index values over a range of shear stress
levels were significantly lower relative to controls in animals exposed to 100 mg PFOS/kg diet.
Absolute and relative (to body weight) liver weight were statistically significantly
increased in the male and female rats at > 20 mg PFOS/kg diet. Relative (to body weight) liver
weight was also statistically increased in the 2 mg PFOS/kg diet females. Histopathological
changes were observed in the liver of the males treated with > 50 mg PFOS/kg diet and included
hepatocyte hypertrophy and an apparent increase in cytoplasmic homogeneity. Increased
hepatocyte hypertrophy and cytoplasmic homogeneity in the females was seen at dietary
concentrations > 50 mg PFOS/kg.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-17
-------
Both males and females showed a significant increase in expression of the gene for
peroxisomal acyl-coenzyme A oxidase (ACOX1) at concentrations > 50 mg/kg diet.
Cytochrome P-450 4A22 (CYP4A22) expression was increased 4-15% greater than controls in
the males in the > 20 mg/kg diet groups and 3-7% greater in the females administered > 50 mg
PFOS/kg diet. Liver fatty acid profiles showed increased total monounsaturated fatty acid levels
and decreased total polyunsaturated fatty acids. At the high doses, the serum levels of
conjugated bilirubin and total bilirubin were increased significantly. A total of 67 fatty acid
profiles were examined. The author stated that the profile changes were similar to those induced
by weak peroxisome proliferators. T4 and T3 levels were also decreased in males and females
with T4 levels being statistically significantly decreased at >20 mg PFOS/kg diet, when
compared to the control levels. Significant differences were observed in this study are provided
in Table 4-5.
TABLE 4-5. Mean (± SD) Values for Select Parameters in Rats Treated for 28 Days3
Parameter
PFOS (mg/kg diet)
0
2
20
50
100
Males
Final body wt (g)
Liver wt (g)
Liver/body wt (%)
Thyroid wt (g)
Conjugated bilirubin
(umol/L)
Total bilirubin
(umol/L)
Cholesterol (mmol/L)
Triglycerides
(mmol/L)
T4 (nmol/L)
T3 (nmol/L)
415.1 ±40.1
17.7 ±2.7
4.24 ±0.41
0.021± 0.004
0.57 ±0.18
2.75 ±0.63
2.54 ±0.63
1.74 ±0.93
80.94 ± 11.83
1.60 ±0.33
412.3 ±32.0
17.1 ±2.8
4.13 ±0.48
0.022 ± 0.005
0.65 ± 0.22
2.75 ± 0.89
2.46 ±0.55
1.92 ±0.78
66.97 ±14.75
1.81 ±0.19
386.2 ±25.9
18.4 ±3.2
4.75* ±0.67
0.020 ± 0.004
0.62 ±0.19
2.47 ±0.82
2.06 ±0.43
1.77 ±0.57
14.36* ±4.18
1.36 ±0.26
363.7* ±25.7
20.8* ±1.5
5.73* ±0.21
0.020 ± 0.003
0.75 ±0.27
2.55 ±0.91
1.63* ±0.31
1.00* ±0.42
12.88* ±2.67
1.29 ±0.26
327.0* ±21.6
21.7* ±2.3
6.64* ±0.41
0.021 ±0.055
2.13* ±0.44
4.01* ±0.87
0.31*±0.18
0.20* ±0.08
13.29* ±2.59
1.21* ±0.23
Females
Final body wt (g)
Liver wt (g)
Liver/body wt (%)
Thyroid wt (g)
Conjugated bilirubin
(umol/L)
Total bilirubin
(umol/L)
Cholesterol (mmol/L)
Triglycerides
(mmol/L)
T4 (nmol/L)
T3 (nmol/L)
247.2 ± 27.5
9.1 ± 1.5
3.64 ±0.38
0.016± 0.003
0.52 ±0.14
2.00 ±0.75
2.06 ±0.36
0.99 ±0.46
37.71 ± 15.41
1.83±0.17
251.2 ±13.1
10.2 ±1.2
4.06* ±0.39
0.017 ±0.004
0.47 ±0.14
1.67 ±0.43
2.02 ±0.51
1.68 ±0.99
32.39 ±10.40
1.72 ±0.14
245.9 ±10.5
11.0* ±1.2
4.45* ±0.40
0.018 ±0.003
0.49 ±0.17
1.51 ±0.54
1.66 ±0.28
1.11 ±0.70
19.62* ±2.49
1.75 ±0.27
217.6* ±15.1
11.2* ±1.2
5. 12* ±0.38
0.017 ±0.003
0.85* ±0.18
2.20 ±0.43
1.37* ±0.24
0.65 ±0.30
15.05* ±1.99
1.41* ±0.22
197.6* ± 10.4
12.2* ±1.4
6.24* ±0.67
0.018 ±0.005
2.60* ±0.73
4.69* ± 1.04
0.52* ±0.16
0.37* ±0.30
16.40* ±4.61
1.27* ±0.20
a Data from Tables 2-3 and 6-7 in Curran et al., 2008
* Statistically significant from controls, p< 0.05 or p < 0.05
Ten three-month old male Sprague-Dawley rats/group were administered by oral gavage
0 (Milli-Q water only), 5 or 20 mg/kg/day of PFOS for 28 days (Cui et al., 2009). Rats were
sacrificed after the exposure and blood and tissue samples obtained. All rats (10/10) administered
20 mg/kg/day of PFOS died by study day 26; at necropsy, rats had bleeding around the eye
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-18
-------
socket and nose, and tumescence/yellow staining at the urogenital region. Prior to death, rats
displayed significant weight loss and a decrease in food consumption when compared to
controls. Rats administered 5 mg/kg/day also had a significant decrease in body weight when
compared to controls at the study termination. Viscera indices were calculated including the
hepatosomatic (HSI), renal-somatic (RSI), and gonad-somatic (GSI) index to evaluate the
hyperplasia, swelling and/or atrophy of the organs and all three indices were statistically
significantly increased in all of the treated groups. The increases in the HSI and RSI showed a
dose dependency. Rats administered 20 mg/kg/day had swelling and discoloration of the liver
with hepatocyte hypertrophy and cytoplasmic vacuolation observed on histopathological exam.
Rats administered 20 mg/kg/day had congestion and thickened walls in the lungs with the
pulmonary congestion also observed in the 5 mg/kg rats. Based on the results, a no observed
adverse effect level (NOAEL) could not be identified and the lowest observed adverse effect
level (LOAEL) was 5 mg/kg/day in rats.
Mouse
Bijand et al. (2011) examined the molecular biology for the hepatic hyperlipidemia in
APOE*3-Leiden.CETP mice, a strain that exhibits human-like lipoprotein metabolism. The
experimental animals were fed a western-type diet with or without 3 mg PFOS/kg/day for
periods of 4 weeks. Plasma samples were collected via tail vein bleeding and analyzed for a
variety of lipid related endpoints including TC, triglycerides (TG), very low density lipoprotein
(VLDL) and HDL. Following terminal sacrifice the liver, heart, perigonadal fat, spleen, and
skeletal femoralis muscle were collected for analysis. Fecal samples were collected for
measurement of bile acids and neutral sterols.
PFOS was found to decrease hepatic VLDL production leading to increased retention of
triglycerides (steatosis) and hepatomegaly. As a consequence there was a decrease in plasma
free fatty acids and glycerol, the mass of perigonadal fat pad, and TG uptake by skeletal muscle.
Neutral sterols in the feces were not altered, but the presence of bile acids was decreased by
50%. Hepatic clearance of VLDL and HDL cholesterol were decreased primarily because of
impaired hepatic production and clearance of these lipoprotein complexes.
Impacted hepatic genes involved with lipid metabolism included those involved with
VLDL metabolism, fatty acid uptake and transport, fatty acid oxidation, and triglyceride
synthesis. Overall the genes upregulated were those involved with fatty acid uptake, transport
and catabolism; triglyceride synthesis, cholesterol ester synthesis; plus VLDL synthesis and
secretion. Genes involved with HDL synthesis, maturation and clearance plus bile acid
formation were down regulated. These changes are consistent with increased hepatic
hyperlipidemia, decrease in bile acid secretion, and serum hypolipidemia. Many of the genes
activated are associated with the nuclear pregnane X receptor (PXR) to a greater extent than
PPARa. Lipoprotein lipase activity and mRNA expressions were increased in the liver. This
enzyme facilitates removal of TGs from serum LDLs, and uptake into the liver and other organs
as free fatty acids and glycerol. Lipoprotein lipase activity in the liver is relatively low
compared to that of peripheral tissues.
Perfluorooctane sulfonate (PFOS) - February 2014 4-19
Draft - Do Not Cite or Quote
-------
Inhalation Exposure
No short-term inhalation animal exposure studies of PFOS were identified.
4.2.3 Subchronic Studies
There are three monkey studies of PFOA exposure, two with rhesus- and one with
cynomolgus- strains, and two rat subchronic studies. The study with cynomolgus monkeys was a
GLP study. There are no subchronic studies by dermal or inhalation routes of exposure with
PFOS.
Oral Exposure
Monkey
Two monkey studies were performed with rhesus monkeys (Goldenthal et al., 1978a and
1979). In the first study, 2 monkeys/sex/group were administered 0, 10, 30, 100 or 300
mg/kg/day of PFOS in distilled water by gavage. The study was terminated on day 20 as all of
the 300 mg/kg treated monkeys died beginning on day 4. Clinical signs of toxicity were
observed in all groups and included decreased activity, emesis, body stiffening, general body
trembling, twitching, weakness and convulsions. At necropsy, several of the 100 and 300
mg/kg/day monkeys had a yellowish-brown discoloration of the liver although there were no
microscopic lesions. A NOAEL or LOAEL was not set in the study.
In the second study, 2 rhesus monkeys/sex/group were administered 0, 0.5, 1.5 or 4.5
mg/kg/day of PFOS in distilled water by gavage for 90 days. All monkeys in the 4.5 mg/kg/day
group died or were euthanized in extremis by week 7 and exhibited decreased body weight, signs
of gastrointestinal tract toxicity (anorexia, emesis, black stool), decreased activity and marked to
severe rigidity and had a significant decrease in serum cholesterol. Histopathology of the 4.5
mg/kg/day monkeys showed diffuse lipid depletion in the adrenals (4/4), diffuse atrophy of the
pancreatic exocrine cells (3/4) and moderate diffuse atrophy of the serous alveolar cells (3/4).
All monkeys in the 0.5 and 1.5 mg/kg/day treated groups survived, but exhibited occasional
diarrhea, soft stools and anorexia. These clinical signs showed a dose-related increase, and 1/4
of the 1.5 mg/kg/day monkeys had low serum cholesterol. Body weight was decreased in males
and females at 1.5 mg/kg/day. There were no treatment-related effects observed in any of the 0.5
or 1.5 mg/kg/day monkeys at necropsy. Based on the findings, the LOAEL was 0.5 mg/kg/day
and the NOAEL could not be determined.
Seacat et al. (2002) administered 0, 0.03, 0.15 or 0.75 mg/kg/day of potassium PFOS
orally in a capsule by intragastric intubation to 6 young-adult to adult cynomolgus
monkeys/sex/group, except for the 0.03 mg/kg/day group (4 monkeys/sex), daily for 26 weeks
(182 days) in a GLP study. Two monkeys/sex in the control, 0.15, and 0.75 mg/kg/day groups
were monitored for one year post-exposure for reversible or delayed toxic effects. Monkeys
were observed twice daily for mortality, morbidity, clinical signs and qualitative food
consumption. Body weights were recorded pre-dosing and weekly, thereafter, and ophthalmic
examinations were performed pre- and post-treatment. PFOS levels were determined in serum
and liver tissue and hematology and clinical chemistry were performed. Urine and fecal analyses
were done and full histopathology performed at the scheduled sacrifice. Liver samples were also
Perfluorooctane sulfonate (PFOS) - February 2014 4-20
Draft - Do Not Cite or Quote
-------
obtained for hepatic peroxisome proliferation determination and immunohistochemistry was
performed by PCNA to look for cell proliferation. Selected results are shown in Table 4-6.
Two of the 0.75 mg/kg/day males died; one died on day 155 and one was found
moribund and sacrificed on day 179. The monkey that died had pulmonary necrosis and severe
acute recurrences of pulmonary inflammation as its cause of death. The specific cause of the
moribund condition was not established, however, the clinical chemistry results were suggestive
of hyperkalemia. Overall mean body weight gain was significantly (p < 0.05) less in the 0.75
mg/kg/day males and females after the treatment when compared to controls. The 0.75
mg/kg/day males and females lost 8 ± 8% and 4 ± 5%, respectively, while the control males and
females gained 14 ± 11% and 5 ± 5%, respectively. Mean absolute and relative (to body weight)
liver weight was increased significantly in the 0.75 mg/kg/day males and females.
Males and females at 0.75 mg/kg/day had lower total serum cholesterol beginning on day
91 (27-68% and 33-49% lower than controls) and lower high density lipoprotein cholesterol
beginning on day 153 (72-79% and 61-68% lower than controls) when compared to the control
values. This effect was reversible, however, as the total cholesterol levels were similar to
controls by week 5 during recovery and the total high density lipoprotein cholesterol was similar
to controls by week 9. Estradiol values were lower at 0.75 mg/kg in males and females on day
182; however, the data were highly variable and the study authors stated that the change was not
well understood. Total triiodothyronine (T3) values were significantly decreased and thyroid
stimulating hormone (TSH) was increased on day 182 in the high-dose monkeys but a true dose-
response was not observed and the monkeys had no sign of hypothyroidism.
Perfluorooctane sulfonate (PFOS) - February 2014 4-21
Draft - Do Not Cite or Quote
-------
TABLE 4-6. Mean (± SD) Values for Select Parameters in Monkeys Treated for 182 Days"
Parameter
PFOS (mg/kg/day)
Males
Body wt (g)
Body wt change (%)
Liver wt (g)
Liver/body wt (%)
Cholesterol (mg/dL)
HDL (mg/dL)
Total T3 (ng/dL)
TSH (uU/mL)
Estradiol (pg/mL)
0
3.7 ±0.7
14±11
54.9 ±8.1
1.6 ±0.2
152 ±28
63 ±11
146 ±19.8
0.55 ±0.44
23.0 ±11.5
0.03
3.9 ±0.6
16 ±8
62.1 ±5.3
1.7 ±0.3
110 ±17**
42 ±4**
145 ±18.0
0.56 ±0.10
24.1 ±14.2
0.15
3.3 ±0.3
8±7
57.3 ±5.5
1.8±0.1
147 ± 24
48 ±14
129 ±4.8
1.38 ±0.78
23.2 ±7.4
0.75
3.2 ±0.8
-8 ±8*
85.3 ±38.4
2.7 ±0.3*
48 ±19**
13 ±5**
76 ± 22**
1.43 ±0.25*
0.8 ±1.0**
Females
Body wt (g)
Body wt change (%)
Liver wt (g)
Liver/body wt (%)
Cholesterol (mg/dL)
HDL (mg/dL)
Total T3 (ng/dL)
TSH (nU/mL)
Estradiol (pg/mL)
0
3.0 ±0.4
5±5
51.1 ±9.4
1.8 ±0.2
160 ± 47
56 ±16
148 ±21.6
1.02 ±0.69
148.5 ±110.1
0.03
3.2 ±0.7
6±7
56.8 ±12.6
1.9 ±0.0
122 ± 22
42 ±9
139±11.5
2.01 ±2.09
125.2 ±101.2
0.15
3.1 ±0.5
4±5
57.0 ±3.1
2.1 ±0.2
129 ± 22
36 ±12**
116 ±16.8
1.33 ±1.13
70.6 ± 62.7
0.75
2.8 ±0.4
-4 ±5
75.3 ±13.3*
2.9 ±0.3*
82 ±15**
21 ±7**
99 ±16.8*
1.86 ±1.29
39.9 ±33.6
a Data from Seacat et al., 2002
Statistically significant from controls: *p<0.05; **p<0.01.
Hepatic peroxisome proliferation was measured by palmitoyl CoA oxidase activity and
was increased significantly in the 0.75 mg/kg/day females; however, it was less than the two-fold
increase typically indicating biological significance. There were no treatment-related effects on
cell proliferation in the liver, pancreas or testes when analyzed by proliferating cell nuclear
antigen immunohistochemistry cell labeling index. Two high dose males and one high-dose
female had 'mottled' livers on gross examination at sacrifice; this was also observed in the high-
dose male that died during the study. On histopathology, all high-dose females and 3/4 high-
dose males had centrilobular or diffuse hepatocellular hypertrophy.
Serum and liver samples collected during the study were analyzed for PFOS and animals
showed a dose-dependent increase in concentrations. Values decreased with recovery but never
returned to control levels. There was not any gender difference in the amount of PFOS identified
in the sera or liver. Based on the decreased body weight gain, decreased serum cholesterol,
increased absolute and relative liver weight and histopathological lesions in the liver, the
LOAEL in male and female monkeys treated with potassium PFOS was 0.75 mg/kg/day and the
NOAEL was 0.15 mg/kg/day. Serum concentrations associated with no adverse effect (0.15
mg/kg/day) were 82.6 ppm in males and 66.8 ppm in females.
Rat
Goldenthal (1978b) administered 0, 30, 100, 300, 1000 or 3000 ppm of PFOS in the diet
to five CD rats/sex/group for 90 days. Dietary levels were equivalent to 0, 2, 6, 18, 60 and 200
mg/kg/day, respectively. All rats at > 300 ppm exhibited emaciation, convulsions, hunched
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-22
-------
back, increased sensitivity to stimuli, reduced activity and red material around the nose/mouth
before their deaths starting on day 7. At 100 ppm body weights were decreased (-16.5%) as was
food consumption when compared to controls. Relative liver weight and relative/absolute liver
weight was significantly increased in the 100 ppm males and females, respectively. Both sexes
had significant increases in relative kidney weight. Three males and 2 females from the 100 ppm
group died. All rats survived at 30 ppm but there was a significant decrease in food consumption
(males) and significant increase in absolute and relative liver weight (females). All treated
animals had very slight to slight cytoplasmic hypertrophy of hepatocytes in the liver. Based on
the significant decrease in food consumption and increase in absolute and relative liver weight,
the LOAEL for PFOS in the rats was 30 ppm (2 mg/kg/day) and the NOAEL could not be
determined.
Five Crl:CD (SD) IGS BR rats/sex/dose level/interim necropsy were administered PFOS
in the diet at concentrations of 0, 0.5, 2.0, 5.0 or 20 ppm for 14 weeks as part of a larger 2-year
cancer bioassay design (Seacat et al., 2003). Doses were equivalent to 0, 0.03, 0.13, 0.34 and
1.33 mg/kg in males and 0, 0.04, 0.15, 0.40 and 1.56 mg/kg in females, respectively. Animals
were observed twice daily for mortality and morbundity with a clinical exam performed weekly.
Body weight and food consumption data were recorded weekly. Other parameters recorded
were: food efficiency, mean daily intake of PFOS, and cumulative/percentage of dose in the liver
and sera. Blood and urine were obtained from 10 animals/sex/treatment during week 14 for
clinical chemistry, hematology and urinalysis evaluation. A thorough necropsy was performed at
the end of treatment and liver samples were collected for palmitoyl CoA oxidase (PCoAO)
activity, cell proliferation index (PI) and PFOS concentration analysis. Microscopic analysis of
tissues was performed on the control and high-dose animals. Analysis of PFOS in the liver and
sera were determined by FtPLC-MS/MS and results were considered quantitative to ± 30%.
No effects were observed on body weight, food efficiency, urinalysis evaluation or
peroxisome proliferation (hepatic PCoAO was unchanged) at 14 weeks. All significant changes,
when compared to controls, were observed in the highest dose group. Food consumption was
decreased. Absolute and relative (to body weight) liver weights were increased significantly in
the males and males/females, respectively. All hematology parameters were similar to controls.
Clinical chemistry parameters that were significantly affected, compared to controls, included:
decreased serum cholesterol (males), increased alanine aminotransferase [ALT] (males) and
increased urea nitrogen (males/females). Select data are provided in Table 4-7.
Perfluorooctane sulfonate (PFOS) - February 2014 4-23
Draft - Do Not Cite or Quote
-------
TABLE 4-7. Mean (± SD) Values for Select Parameters in Rats Treated for 14 Weeks3
Parameter
PFOS (mg/kg/day)
Males
Body wt (g)
Liver wt (g)
Liver/body wt (%)
Seg. neutrophils (103/uL)
Glucose (mg/dL)
Cholesterol (mg/dL)
ALT (IU/L)
Urea nitrogen (mg/dL)
PCoAO (lU/g)
0
496 ± 56
15.5 ±1.1
3.2 ±0.3
1.1 ±0.4
102 ± 6.2
63 ±13
36 ±7
13 ±2
4.6 ±1.3
0.03
481 ±51
15.5 ±2.7
3.2 ±0.2
1.3 ±0.3
106 ±11
53 ±17
41±6
14 ±2
4.8 ±3.3
0.13
434 ±31
14.0 ±1.4
3.2 ±0.2
1.2 ±0.3
91 ±14
51 ±15
41±5
13 ±2
5.4 ±3.0
0.34
424 ± 44
18.8 ±3.0
3.6 ±0.3
1.2 ±0.4
99 ±9
57 ±7
44 ±14
14 ±1
1.8 ±1.8
1.33
470 ± 40
20.3*±2.2
4.3* ±0.4
1.6* ±0.4
95 ±10
37* ± 13
65* ±53
16* ±2
5.4 ±1.9
Females
Body wt (g)
Liver wt (g)
Liver/body wt (%)
Seg. neutrophils (103/uL)
Glucose (mg/dL)
Cholesterol (mg/dL)
ALT (IU/L)
Urea nitrogen (mg/dL)
PCoAO (lU/g)
0
284 ±39
9.3 ±1.6
3.3 ±0.2
1.0 ±0.5
106 ±12
75 ±15
34 ±2.4
12 ±2
1.8 ±1.6
0.04
298 ±41
9.2 ±1.3
3.1±0.1
1.0 ±0.5
106 ±9
88 ±27
36 ±9
13 ±2
3.0 ±2.6
0.15
266 ± 16
8.4 ±0.7
3.2 ±0.3
0.7 ±0.2
108 ±6
87 ±24
37 ±18
13 ±2
1.0 ±0.8
0.40
247 ± 18
8.7 ±1.0
3.5 ±0.3
0.9 ±0.6
95* ±8
70 ±13
34 ±5
14 ±3
1.6 ±2.6
1.56
249 ± 26
10.6 ±0.7
4.3* ±0.4
1.0 ±0.6
99 ±7
66 ±14
39 ±18
17* ±2
5.0 ±2.9
* Data from Table 1 in Seacat et al., 2003
* Statistically significant from controls, p< 0.05
Histopathological changes were not observed in the kidney; however, centrilobular
hepatocyte hypertrophy and mid-zonal to centrilobular vacuolization were observed in the livers
of the males and females. Based on the findings, the LOAEL for male and female rats
administered PFOS in the diet for up to 14 weeks was 20 ppm (1.33 mg/kg in males and 1.56
mg/kg in females) and the NOAEL was 5 ppm (0.34 mg/kg in males and 0.40 mg/kg in females).
4.2.4 Neurotoxicity
In vivo
Rat
Yang et al. (2009) determined the effect of PFOS on excitatory amino acids (EAAs) and
glutamine synthetase (GS) in the rat central nervous system. Five adult male Wistar rats/group
were administered one dose of 0, 12.5, 25 or 50 mg/kg bw PFOS by oral gavage. The animals
were sacrificed 5 days after administration. The EAAs analyzed in brain tissue were glutamate
(Glu), aspartate (Asp), glycine (Gly) and gamma-aminobutyric acid (GABA).
All treated rats had a significant (p<0.05) decrease in body weight, 15, 22 and 27% less
than the controls in the 12.5, 25 and 50 mg/kg groups, respectively. Among the EAAs, the Glu
content was significantly decreased at the high dose ([11% compared to controls; p<0.05) in the
hippocampus; no other significant differences were recorded. In the cortex, Glu was again the
only EAA affected with significant decreases at 25 (|33% compared to controls) and 50 (|47
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-24
-------
compared to controls) mg/kg. GS activity was significantly increased in the hippocampus at 25
and 50 mg/kg bw. The study had a LOAEL of 12.5 mg/kg/day in rats based on the decreased
body weight and indicated an effect on Glu and GS in the central nervous system starting at 25
mg/kg/day.
Male Wistar rats and ICR mice (n=2-3/group) were administered a single oral dose of
PFOS at concentrations of 0, 125, 250 or 500 mg/kg bw and monitored for any neurological
signs (Sato et al., 2009). Animals were checked daily for startle response, touch response, pain
response, righting reflex, visual placing, abdominal tone and limb tone. No neurological signs
were observed. However, when stimulated with ultrasound (44 kHz, 10 sec), tonic convulsions
occurred and 1/3 rats in the 250 mg/kg group, 2/2 rats in the 500 mg/kg group and 1 mouse in
each dose group died within 48 hours of the convulsions. No histopathological changes were
observed in the neuronal or glial cells of the cerebrum and cerebellum in rats killed 24 hours
after exposure. In these same rats, the highest concentration of PFOS was in the liver and the
lowest was in the brain. Rats administered 250 mg/kg bw one time did not show any differences
in the levels of catecholamines (norepinephrine, dopamine and serotonin) or amino acids
(glutamic acid, glycine and GABA) when compared to the controls at 24 and 48 hours post-
exposure.
In vitro
Slotkin et al. (2008) evaluated 10-250 |iM PFOS, PFOA, perfluorooctane sulfamide
(PFOSA) and perfluorobutane sulfonate (PFBS) in vitro in differentiated and undifferentiated
PC12 cells, a neurotypic cell line. In the study, inhibition of DNA synthesis, deficits in cell
numbers and growth, oxidative stress, cell viability and shifts in differentiation toward or away
from the dopamine (DA) and acetylcholine (ACh) neurotransmitter phenotypes were assessed.
No effects on cell size, cell number or neurocyte outgrowth were observed. PFOS decreased cell
viability at 250 jiM and promoted differentiation into the ACh phenotype at the expense of the
DA phenotype. The study suggests that the mechanisms of these neurotypic cell lines are not
similar between the tested perfluoroalkyl acids. The rank order of adverse effects was PFOSA >
PFOS > PFBS = PFOA.
Liao et al. (2009) assessed the neurotoxic effect of varying chain lengths of the
perfluorinated compounds on cultured Sprague-Dawley rat hippocampal neurons. Spontaneous
miniature post-synaptic currents (mPSCs) were recorded in gap-free mode from hippocampal
neurons at 8-15 days in vitro. The compounds tested included those in the carboxylic group:
perfluoropropionic acid (PFPA; C-3), perfluorobutyric acid (PFBA; C-4), perfluorooctanoic acid
(PFOA; C-8), perfluorododecanoic acid (PFDA; C-12) and perfluorotetradecanoic acid (PFTA;
C-14); those in the sulfonic group: perfluorobutane sulfonic (PFBS; C-4),
perfluorohexanesulfonic acid potassium salt (PFHS; C-6) and perfluorooctane sulfonate (PFOS;
C-8) and a nonfluorinated hydrocarbon, octanoic acid (OA; C-8). Also tested was 1 H-
perfluorooctane (PFOC; C-8). Testing showed frequency of mPSCs increased in proportion to
the increase in carbon length. Eight carbon PFOS had a statistically significant (p< 0.001)
increase in the mPSCs when compared to 4 carbon PFBS. Compounds with a carboxylic group
also had lower frequencies than those with sulfonic groups. PFOS significantly (p< 0.001)
increased the mPSC amplitude. The inward calcium currents (Ica) were recorded in the presence
or absence of compounds with a ramp depolarization pulse. Voltage values were recorded and
Perfluorooctane sulfonate (PFOS) - February 2014 4-25
Draft - Do Not Cite or Quote
-------
plotted versus the corresponding Ica every 5 mV and the resulting current-voltage relationship
curve established. All three sulfonic compounds increased the ICa with correlation to the chain
length with PFOS having the most effect (% increase not provided).
In the same study, the chronic effects of perfluorinated compounds on neuronal
development were evaluated by measuring neurite outgrowth and branching. In the sulfonic
compounds, only PFOS statistically suppressed the length of neurites (p < 0.001; 25% below that
of controls). The lengths of the longest neurites and sum length of neurites per neuron were not
affected with OA or PFOC but were reduced by PFOA and PFOS.
Overall, the adverse effects on cultured neurons increased as the chain length of the test
compound increased. The study also suggested that perfluorinated sulfonates exerted more
potent actions on neurons when compared to perfluorinated carboxylates. The study authors
hypothesized that this could occur because PFOS was more likely to be incorporated into the
lipid bilayer of cell membranes.
4.2.5 Developmental/Reproductive Toxicity
Rats, rabbits, and mice were all found to be affected in developmental/reproductive
studies of PFOS. Prenatal exposure to PFOS caused an increase in neonatal mortality when
dams were exposed to concentrations > 2 mg/kg/day. Effects were observed on gestation length,
birth weight, survival and developmental delays. Structural abnormalities observed include:
ocular lens abnormalities (questionable), decreased bone ossification, cleft palate, and
enlargement of the right atrium. Many specialized developmental studies (neurotoxicity,
immunotoxicity) have been conducted with PFOS.
Reproductive Effects
Rat
A two-generation reproductive study was conducted in Crl:CD(SD)IGS VAF rats with
five groups of 35 rats/sex/group administered 0, 0.1, 0.4, 1.6 or 3.2 mg/kg/day of PFOS by
gavage for six weeks prior to and during mating (Luebker et al., 2005b). Treatment in males
continued until one day prior to sacrifice and females were treated throughout gestation,
parturition and lactation.
FO Generation:
Parental animals (FO) were observed twice daily for clinical signs, and body weight and
food consumption monitored. Two sets of females in each dose group were treated and had
Caesarean-sections (C-section) performed on GD10; others delivered naturally and were killed
on lactation day (LD) 21. Typical reproductive parameters were monitored in the females. The
FO male rats were sacrificed and necropsied after the cohabitation interval, with the testes,
epididymides, prostate, and seminal vesicles weighed. All livers from adults were removed,
weighed and examined. Blood samples were collected from five male rats at sacrifice and five
female rats on LD 21 for pharmacokinetic analysis; livers of pups from the litters of these five
dams were also collected for analysis.
In the FO generation male rats, no effects were observed on mortality, treatment-related
clinical signs and the mating/fertility parameters evaluated. During pre-mating, decreases in
Perfluorooctane sulfonate (PFOS) - February 2014 4-26
Draft - Do Not Cite or Quote
-------
terminal body weight, body weight gain and food consumption occurred at 1.6 and 3.2
mg/kg/day in males. The only effect on weight of the organs evaluated was a significant
reduction in the absolute weight of the seminal vesicles (with fluid) and prostate in males
administered 3.2 mg/kg/day. In the FO generation female rats, there were no deaths and no
effects on the reproductive parameters measured in both dams sacrificed on GD 10 and those
allowed to deliver naturally. The FO dams administered > 0.4 mg/kg/day had localized alopecia
during pre-mating, gestation, and lactation, and a decrease in body weight and food
consumption.
Fl Generation:
The Fl generation pup viability was significantly reduced at 1.6 and 3.2 mg/kg/day, so
only the 0.1 and 0.4 mg/kg/day dose groups were carried into the second generation. Twenty-
five Fl rats/sex/dose were administered 0, 0.1, or 0.4 mg/kg/day of PFOS by oral gavage
beginning at weaning on post-natal day (PND) 22 and continuing through sacrifice. One
rat/sex/litter was tested in a passive avoidance paradigm at 24 days of age and one/sex/litter was
evaluated in a water-filled M-maze on PND 70. On PND 28, females were evaluated for vaginal
patency and on PND 34 males were examined for preputial separation. On PND 90, rats were
assigned within each dose group to cohabitation, and once confirmed pregnant, the females were
individually housed. The Fl generation male rats were sacrificed after mating, necropsied and
evaluated as described in the FO generation. All Fl generation females delivered naturally and
were sacrificed and necropsied on LD 21.
Mortality occurred in the Fl offspring of dams administered 1.6 or 3.2 mg/kg/day. At 1.6
mg/kg/day, over 26% of the pups were found dead between LDs 2-4. At 3.2 mg/kg/day, 45% of
the pups were found dead on LD 1 with 100% dead by LD 2. The dams dosed with 3.2
mg/kg/day also had a significant increase in stillborns and the viability index was 0% at 3.2
mg/kg/day and 66% at 1.6 mg/kg/day. The lactation index was 94.6% at 1.6 mg/kg/day. At 3.2
mg/kg/day, there were significant decreases in gestation length, number of implantation sites and
reductions in litter size. Statistically significant decreases in pup body weight were also
observed at the two highest doses. Additional adverse signs in pups at 3.2 mg/kg/day were a
high number of pups not nursing, not having milk present in the stomach, being stillborn and a
high incidence of maternal cannibalization of the pups.
In the Fl generation offspring, pups administered 3.2 mg/kg/day could be evaluated only
on LD 1 due to high mortality; all displayed decreased surface righting ability. The acoustic
startle reflex and air righting ability were both significantly reduced at 1.6 mg/kg/day. Similar
responses were seen for pinna unfolding and eye opening. The physical development delays
were transient; all pups were similar to controls by the end of the observation period. No delays
were observed in rats administered doses < 0.4 mg/kg/day.
F2 Generation:
Fl parental animals displayed no clinical signs or mortality. Fl male parents had a
transient decrease in food consumption but it was not affected in Fl females. The Fl dams
showed no effects in reproductive performance.
All F2 generation pups were sacrificed, necropsied, and examined on LD 21 as
previously described for the Fl generation pups. In the F2 generation pups, decreases in mean
pup body weights at 0.1 mg/kg/day on LDs 4 and 7 were transient; mean body weights were
Perfluorooctane sulfonate (PFOS) - February 2014 4-27
Draft - Do Not Cite or Quote
-------
similar to controls by LD 14. The pups at 0.4 mg/kg/day had significant decreases in body
weight on LDs 7-14; after LD 21, body weights remained lower than controls, but were not
statistically significant. No other treatment-related effects were observed.
Based on the decreases in body weight gain and food consumption, the LOAEL for both
the FO male and female parents was 0.4 mg/kg/day and the NOAEL was 0.1 mg/kg/day. For the
Fl parents, the NOAEL was 0.4 mg/kg/day and the LOAEL was not identified. For the Fl
offspring, the LOAEL was 1.6 mg/kg/day based on the significant decrease in the pup viability,
pup weight and survival; the NOAEL was 0.4 mg/kg/day. In the F2 generation offspring, the
LOAEL was 0.4 mg/kg/day, based on the significant decreases in mean pup body weight; the
NOAEL was 0.1 mg/kg/day.
Because of the significant reductions in pup viability observed at 1.6 and 3.2 mg/kg/day,
a cross-fostering study was conducted as a means of determining whether the effects observed in
pups were a result of in utero exposure to PFOS or as a result of exposure during lactation
(Luebker et al., 2005b). Twenty five female Sprague-Dawley rats/group were administered 0 or
1.6 mg/kg/day PFOS in 0.5% Tween-80 by gavage, beginning 42 days prior to mating to
untreated males, and continuing throughout gestation until LD 21. Parental females were
observed twice daily for viability and clinical observations were recorded. Maternal body weight
and food consumption were recorded. All maternal rats were sacrificed on LD 22 and gross
necropsy was performed; the number and distribution of implantation sites were recorded. After
parturition, litters were immediately removed from their respective dams and placed with either a
control- or PFOS-treated dam for rearing. This cross-fostering procedure resulted in four groups
as follows:
• control dams with litters from control dams (negative control)
• control dams with litters from PFOS-treated dams {in utero exposure only);
• PFOS-treated dams with litters from control dams (post-natal exposure only)
• PFOS-treated dams with litters from PFOS-treated dams (both in utero and post-natal
exposure).
There was no mortality or clinical signs associated with treatment in the dams. Mean
maternal body weight gain and food consumption at 1.6 mg/kg/day was reduced compared to
controls during premating and continuing throughout gestation, but not lactation. Reductions in
gestation length, the average number of implantation sites, delivered sizes, and live litter size
were observed in treated animals.
Live litter sizes were comparable between treated and control groups following cross-
fostering. On LDs 2-4, ~ 19% of the pups in the group exposed gestationally and lactationally
were either found dead or presumed cannibalized. Pup mortality for the negative control was
1.6%. For pups exposed only prenatally mortality was 9% while it was 1.1% for those exposed
during lactation only. Reductions in pup body weights on LD Iwere observed in groups exposed
both gestationally and lactationally and those with gestational exposure only. On LDs 4-21, pup
body weights were reduced in all exposed groups when compared to the negative control, with
the greatest decrease in the group with gestational and lactational exposures.
Perfluorooctane sulfonate (PFOS) - February 2014 4-28
Draft - Do Not Cite or Quote
-------
Two litters in the group with lactational exposure and one litter exposed during gestation
and lactation did not nurse. No milk was found in the stomachs of the pups found dead and
necropsied from the groups with in utero plus lactational exposures (100% of pups), in utero
exposures only (57% of pups), and lactational exposures only (87% of pups). Sex ratios and the
lactation index were comparable among all groups. Electron microscopic examination of the
livers revealed an increase in the number of peroxisomes in pups from dams treated with 1.6
mg/kg/day PFOS. No significant differences were observed between the negative control group
and the other groups following examination of pup lungs.
Serum PFOS concentrations in untreated dams ranged from below the limit of detection
(0.05 |ig/mL) to 5.34 jig/mL. Serum PFOS concentrations in the pups from the negative controls
were below the limit of detection. Serum PFOS concentrations in the pups from treated dams,
fostered with untreated dams (in utero exposures) ranged from 47.6 |ig/mL to 59.2 jig/mL.
Serum PFOS concentrations of treated dams ranged from 59.2 |ig/mL to 157 |ig/mL. Serum
PFOS concentrations in the pups from untreated dams, fostered with treated dams (lactational
exposure), ranged from below the limit of detection to 35.7 |ig/mL. Serum PFOS concentrations
in the pups from treated dams, fostered with treated dams (in utero plus lactational exposures),
ranged from 79.5 |ig/mL to 96.9 |ig/mL. These data indicate that exposure to PFOS can occur
both in utero and via milk from treated dams (3M Environmental Laboratory, 1999). The
accuracy of quantitation for the analyses was ±30%.
In conclusion, pups from control dams that were cross-fostered with PFOS-treated dams
(post-natal exposure only) had the same low mortality rate (1.1%) as pups from control dams
cross-fostered with control dams (1.6%; negative control). Mortality rates in the remaining two
groups (gestational exposures and gestational plus lactational exposures) were much higher at
9% and 19%, respectively. Although the study is limited, the data appear to indicate that
reduced pup survival is mainly a result of in utero exposure to PFOS and that post-natal exposure
via milk in conjunction with in utero exposure increases the risk of mortality. In contrast,
exposure during lactation alone from exposed dams had no significant affect on pup viability.
The dose-response curve for neonatal mortality in rat pups born to PFOS exposed dams
and the biochemical and pharmacokinetic parameters were investigated (Luebker et al., 2005a).
At 6 weeks prior to mating, female Crl:CD(SD)IGS VAF/Plus rats/dose were administered 0,
0.4, 0.8, 1.0, 1.2, 1.6 or 2.0 mg PFOS/kg bw/day by oral gavage. Dosing continued during the
mating interval and through gestation day (GD) 20 for dams assigned to C-section (eight dams in
the control, 1.6 and 2.0 mg/kg/day groups). Another group (-20 dams per dose group were
allowed to deliver naturally and nurse their pups through LD4. These dams and their pups were
killed on LD 5.
The dams in the C-section group were examined for the number of corpora lutea, number
of implantation sites, live/dead fetuses, early/late resorptions and placentas. Liver weights were
determined for the dams as well and their organs examined by gross necropsy. Fetuses were
pooled by litter and mean weight recorded. In the natural delivery dams, typical reproductive
and fetal parameters were measured and recorded. Biochemical parameters investigated in the
dams and litters included: serum lipids, glucose, mevalonic acid and thyroid hormones (total
[TT4] and free thyroxin [FT4], total [TT3] and free triiodothyronine [FT3] and thyroid
stimulating hormone [TSH]); milk cholesterol and liver lipids. Mevalonic acid was included as
Perfluorooctane sulfonate (PFOS) - February 2014 4-29
Draft - Do Not Cite or Quote
-------
it is a biomarker of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase activity.
Some chemicals that are inhibitors of this enzyme are known to cause developmental effects in
rats.
No mortality occurred and no effects were observed in reproductive parameters (corpora
lutea, implantations, dead fetuses/litter) in those C-sectioned. Mild decreases in body weight
were observed in dams at 1.6 and 2.0 mg/kg during gestation. Body weight gain in the dams was
affected statistically during lactation at > 0.8 mg/kg and food consumption showed a general
downward trend with increasing dose during pre-mating, gestation and lactation. For those
allowed to deliver normally, the fertility index, implantations per delivered litter, gestation index,
live births and delivered pups/litter were similar between treated and control dams. Based on the
decreased body weight gain, the LOAEL for the FO dams was 0.8 mg/kg/day and the NOAEL
was 0.4 mg/kg/day.
In the group sacrificed on LD5, a decrease in gestation length was observed at doses >
0.8 mg/kg and decreases in viability were observed starting at 0.8 mg/kg becoming statistically
significant at doses of 1.6 and 2.0 mg/kg. The viability indices were 97.3%, 93.1%, 88.8%,
81.7%, 49.3% and 17.1% at 0, 0.8, 1.0, 1.2, 1.6 and 2.0 mg/kg, respectively (Table 4-8). The
decrease in survival did not appear to be caused by a reduction in lipids, glucose utilization or
thyroid hormones as these parameters were either only slightly affected or similar between
treated and control animals. In all treated groups, pup body weight at birth and on postnatal day
5 was significantly less than that of controls. In one male and one female pup at 2.0 mg/kg/day,
the heart and thyroid were collected and examined microscopically. No lesions were found
when compared to the controls. The LOAEL for the Fl generation was 0.4 mg/kg/day,
respectively, based on decreased body weight and the NOAEL was not identified.
Several benchmark dose estimates (BMD5 and BMDL5) were run by the author and
provided in the study. They were as follows:
Effect on gestation length BMD5 = 0.45 mg/kg/day BMDL5 = 0.31 mg/kg/day
Birth weight effect BMD5 = 0.63 mg/kg/day BMDL5 = 0.39 mg/kg/day
Decreased pup wt (day 5) BMD5 = 0.39 mg/kg/day BMDL5 = 0.27 mg/kg/day
Pup weight gain (day 5) BMD5 = 0.41 mg/kg/day BMDL5 = 0.28 mg/kg/day
Decreased survival of pups
through day 5 BMD5 = 1.06 mg/kg/day BMDL5 = 0.89 mg/kg/day
Perfluorooctane sulfonate (PFOS) - February 2014 4-30
Draft - Do Not Cite or Quote
-------
TABLE 4-8. Fertility and Litter Observations in Dams Administered 0 to 2.0 mg
PFOS/kg/Daya
Fertility indexb (%)
Implantations per
delivered litter
Gestation length (days)
Gestation index0 (%)
Delivered pups/litter
Live births (%)
Dams with all pups dying
onLD 1-5
Viability indexd (%)
0.0
96.4
14.7 ±2.3
22.9 ±0.3
100
13.9 ±2.6
98.1
0
97.3
0.4
100.0
16.2 ±
1.8
22.6 ±
0.5
100
15.0±
2.3
97.0
0
97.6
0.8
89.5
15. 1±
2.2
22.5 ±
0.5*
100
14.5 ±
2.3
99.2
0
93.1
1.0
95.0
15.9 ±2.0
22.4 ±
0.6**
100
15.1 ±2.3
99.3
1
88.8
1.2
94.7
15.3 ±2.5
22.3 ±
0.5**
100
14.0 ±2.9
99.6
0
81.7
1.6
92.6
14.3 ±2.1
22.0 ±
0.0**
100
13.6 ±2.8
98.3
4
49.3**
2.0
96.4
14.4 ±1.9
22.2 ±
0.4**
100
13.3 ±2.5
99.6
14**
17.1**
a Data from Luebker et al., 2005a
b Number of dams pregnant/number of dams mated x 100
0 Number of dams with live offspring/number of pregnant dams x 100
d Number of live pups on day 5 postpartum/number of live births x 100
Statistically significant at * p < 0.05 or ** p < 0.01
Developmental Studies
Rat
Two older developmental studies were performed on rats. Gortner (1980) administered
0, 1, 5 or 10 mg/kg/day PFOS in corn oil by gavage to four groups of time-mated Sprague-
Dawley rats on gestation days (GD) 6-15 and then sacrificed them on GD 20. The maternal
LOAEL was 10 mg/kg/day based on significant decreases in body weight during GDs 12-20 and
the maternal NOAEL was 5 mg/kg/day. The developmental LOAEL was 1 mg/kg/day based on
abnormalities in the lens of the eye, not observed in the control fetuses. The eye abnormalities
were localized to the area of the embryonal lens nucleus and appeared to be a lack of complete
development of the primary lens fibers. The developmental NOAEL could not be determined.
The author added an amendment to the study, however, suggesting that the eye lesion was an
artifact created by sectioning and the results were not observed in the Wetzel (1983) study.
Therefore, the developmental lesion occurring in the eye is believed to be an artifact and not
related to PFOS.
In the second study, groups of 25 pregnant Sprague-Dawley rats were administered 0, 1,
5 or 10 mg/kg/day PFOS in corn oil by gavage on GDs 6-15 (Wetzel, 1983). In this study, the
maternal LOAEL was 5 mg/kg/day based on clinical signs observed at 5 and 10 mg/kg/day in
dams, including hunched posture, anorexia, uterine stains, bloody vaginal discharge, rough
haircoat, and decreased body weight. The maternal NOAEL was 1 mg/kg/day. Two dams in the
10 mg/kg/day group were found dead on GD 17, but information as to the cause of death was not
provided. The developmental LOAEL was 5 mg/kg/day based on a decrease in body weight in
fetuses; the NOAEL was 1 mg/kg/day. Statistically significant increases in the number of litters
containing fetuses with visceral abnormalities, delayed ossifications and skeletal abnormalities
were observed at 10 mg/kg/day.
A two-part developmental study with PFOS was performed in rats. Thibodeaux et al.
(2003) reported on the maternal and prenatal evaluation by administering 0, 1, 2, 3, 5 or 10
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-31
-------
mg/kg PFOS daily by gavage in 0.5% Tween-20 during gestational days (GDs) 2-20 to groups of
9-16 pregnant Sprague-Dawley rats. Maternal weight gain, food and water consumption and
serum clinical chemistries were monitored and recorded. Rats were euthanized on GD 21 and
uterine contents examined. At sacrifice, PFOS levels were measured in the serum and maternal
and fetal livers.
Maternal body weight, food consumption and water consumption were decreased
significantly (p< 0.0001) in a dose-dependent manner at doses > 2 mg/kg. A dose dependent
increase in the serum concentration of PFOS was observed with liver concentrations
approximately four times higher. Liver weight was not affected in the treated rats. Serum
chemistry showed significant decreases in cholesterol (| 14% compared to controls) and
triglycerides (| 34% compared to controls) at 10 mg/kg. Serum thyroxine (T4) and
triiodothyronine (Ts) were significantly decreased in all treated rats when compared to controls,
however, a feedback response on thyroid stimulating hormone (TSH) was not observed. The
number of implantations or live fetuses at term was not affected by treatment. There was a
decrease in fetal weight, and birth defects such as cleft palate, ventricular septal defect and
enlargement of the right atrium were observed at 10 mg/kg but the litter incidence rates were not
given. Benchmark dose estimates were provided for different parameters and were as follows:
maternal weight reduction (polynomial model) BMDs = 0.22 mg/kg and BMDLs = 0.15 mg/kg;
T4 effects on GD7 (Hill model) BMD5 = 0.23 mg/kg and BMDL5 = 0.05 mg/kg; fetal sternal
defects (logistic model) BMD5 = 0.31 mg/kg and BMDL5 = 0.12 mg/kg and fetal cleft palate
(logistic model) BMD5 = 8.85 mg/kg and BMDL5 = 3.33 mg/kg.
In the second part of the developmental study, the post-natal effects of in utero exposure
to PFOS were evaluated in the rat (Lau et al., 2003). Sprague-Dawley rats were administered 0,
1, 2, 3, 5 or 10 mg/kg PFOS in 0.5% Tween-20 by gavage on GDs 2-21. On GD 22, dams were
monitored for signs of parturition. The day after parturition was designated post natal day (PND)
1. The number of pups per litter, number of live pups in the litter daily and body weight were
monitored. All pups were weaned on PND 21 and separated by gender. Additional pregnant rats
were dosed in the same manner to 0, 1, 2, 3 or 5 mg/kg of PFOS and four pups from each litter
were sacrificed within 2-4 hours after birth and used to determine blood and liver PFOS
concentrations and thyroid hormone analysis. The other pups were used for additional PFOS
concentrations, thyroid hormone analysis and neurobehavioral tests.
In dams administered 10 mg/kg, the neonates became pale, inactive, and moribund within
30-60 minutes and all died. In 5 mg/kg dams, the neonates became moribund after 8-12 hours,
with 95% dying within the first 24 hours. A 50% fetal mortality was observed in dams
administered 3 mg/kg. Pups from dams treated with 2 mg/kg still had significant increases in
mortality but those from dams administered 1 mg/kg were similar to controls. No differences
were observed in liver weight in the rats. Pup body weight was affected with PFOS treatment
and was significantly decreased starting in dams administered 2 mg/kg. A slight but significant
delay in eye opening was observed in the rats (> 2 mg/kg), but no difference was observed in
onset of puberty. On PND 2, serum levels of both total T4 and free T4 were decreased
significantly in all the treated groups but total T4 recovered to levels similar to those of controls
by weaning. No changes were observed in serum TS or TSH. Choline acetyltransferase activity
in the prefrontal lobe which is sensitive to thyroid status was slightly reduced in rat pups, but
activity in the hippocampus was not. T-maze testing did not demonstrate any learning
Perfluorooctane sulfonate (PFOS) - February 2014 4-32
Draft - Do Not Cite or Quote
-------
deficiencies. Based on the findings, the developmental LOAEL is 2 mg/kg PFOS and the
NOAEL is 1 mg/kg. Benchmark dose estimates were provided for survival of the neonates on
PND 8 (NCTR model) and were as follows: BMD5 = 1.07 mg/kg and BMDL5 = 0.58 mg/kg.
Because of the high number of fetal deaths, a sub-study was performed with newborns
from the 5 mg/kg PFOS group being cross-fostered with control dams immediately after
parturition. Survival was monitored for 3 days. Cross-fostering the PFOS-treated rats (5 mg/kg)
with control dams did not increase survival, and all control pups fostered by PFOS treated dams
survived.
Grasty et al. (2003) investigated the critical window for prenatal exposure to PFOS by
administering it to pregnant rats by gavage at 25 mg/kg on four consecutive days (GDs 2-5, 6-9,
10-13, 14-17 or 17-20) or at 25 or 50 mg/kg on GDs 19-20. In those administered PFOS in the
4-day intervals, litter size at birth was unaffected but pup weight was decreased. Neonates died
after dosing in all time periods but the incidence of death increased as the time of dosing moved
closer to the end of gestation period. Mortality was 100% when administered on GD 17-20.
Most deaths occurred within 24 hours but all occurred by PND 4. In the two-day treatment,
survival of the pups was 98, 66 and 3% in the control, 25 and 50 mg/kg groups on PND 5.
Histological examination of the lungs showed differences in maturation between the control and
treated pups. Grasty et al. (2005) performed another study to determine if delayed lung
maturation was responsible for the deaths. The newborns were found to have thick alveolar
walls but had normal pulmonary surfactant profiles leading to doubt that lung maturation was the
cause of death.
While the lung maturation did not appear to be the cause of death, on-going studies
support that effects of PFOS on lung surfactant is still likely to be the reason neonatal deaths
occurred. PFOS has been shown to interact with dipalmitoylphosphatidylcholine (DPPC) which
is a major component of surfactant (Xie et al. 2007, 2010a and 2010b). As discussed in the
distribution section, Borg et al. (2010) found radiolabeled PFOS localized in the perinatal lung
on GD 18 after it was administered to the dams on GD 16. In these same pups, the PFOS levels
in the lungs were three-fold higher than that in the maternal blood on PND 1.
Chen et al. (2012) administered 0, 0.1 or 2.0 mg/kg/day PFOS to 10 pregnant Sprague-
Dawley rats/group in 0.05% Tween 80 in deionized water on GDs 1 to 21 by gavage. After
parturition (PND 0), pups were counted, weighed and 2 male and 2 female pups/litter were
randomly selected for sacrifice and serum and lung collection. Six offspring/litter were kept
until PND 21 when they were sacrificed for serum and lung collection. Lung tissue was used for
histopathological examination and assessing oxidative stress and extraction of cytoplasmic
protein. The serum and lungs were also analyzed for PFOS concentration. To determine the
maternal effect of PFOS on offspring survival, three additional groups of rats were treated as
previously described and the number of deaths/litter recorded until PND 4.
Body weight of the dams was decreased and postnatal mortality (by PND 3) increased
significantly (p< 0.05 and 0.01) at 2.0 mg/kg/day, when compared to the control litters. No
treatment-related findings were observed at 0.1 mg/kg/day. Postnatal mortality in the control,
0.1 and 2.0 mg/kg/day groups on PND 3 was approximately 4%, 3% and 23%, respectively. On
PND 0, PFOS concentrations in the serum (|ig/mL) were approximately 2x greater than that
Perfluorooctane sulfonate (PFOS) - February 2014 4-33
Draft - Do Not Cite or Quote
-------
found in the lung (|ig/g) at both 0.1 and 0.2 mg/kg/day. PFOS concentrations decreased in both
the serum and lungs on PND 21 but were still greater in the serum. PFOS was not detected in
controls at either timepoint. Histopathological changes were observed in the lungs of rats at 2.0
mg/kg/day on PND 0 that included marked alveolar hemorrhage, thickened interalveolar septum
and focal lung consolidation. At this same dose, the lungs also had alveolar hemorrhage,
thickened septum and inflammatory cell infiltration on PND 21. Numerous apoptotic cells were
also observed in the lungs of rats at 2.0 mg/kg/day. No abnormalities were observed on
examination of the control rats or at 0.1 mg/kg/day. An increase in biomarkers associated with
oxidative stress was also observed in the rats at 2.0 mg/kg/day. In the pups, there was an
increase in the level of malondialdehyde (MDA; 473% and 305% of controls on PND 0 and 21,
respectively) and a decrease in the level of glutathione (GSH) content. Superoxide dismutase
(SOD) activity declined also at 2.0 mg/kg/day. Cytochrome c (Cyt c) was increased in rats at 2.0
mg/kg/day which is an apoptogenic factor in the mitochondrial pathway. To determine the
intrinsic and extrinsic cell death pathway role in the apoptosis found in the rat offspring, caspase-
like activity was measured using specific chromogenic substrates. In the lungs of those
administered 2.0 mg/kg/day, caspase -3, -8 and -9 were activated markedly at both PND 0 and 21
when compared to the controls. No changes were observed in the 0.1 mg/kg/day substrates.
Overall, the results suggest that both the mitochondrial and the cell death receptor pathways were
activated by prenatal PFOS exposure.
To further investigate the effects of PFOS on the fetal lung, Ye et al. (2012) administered
0, 5, or 20 mg PFOS/kg/day by gavage in 0.5% Tween-20 to Sprague-Dawley rats on GDs 12-
18. Animals were sacrificed on GD 18.5 and the fetal lungs analyzed for histological lesions and
gene expression profiles. Maternal treatment with PFOS did not result in any apparent
microscopic changes in the fetal lung. However, gene expression profiling showed a dose-
dependent up-regulated expression of 21 genes at 5 mg/kg/day and of 43 genes at 20 mg/kg/day.
The genes included five PPARa target genes, four of which are involved in lipid metabolism; the
remaining up-regulated genes were involved in significant cytoskeletal, extracellular matrix
remodeling, and transporting and secreted proteins in the fetal lung.
Rabbit
Christian et al. (1999) administered 0, 0.1, 1.0, 2.5 or 3.75 mg/kg/day of PFOS in 0.5%
Tween-80 by gavage to 22 pregnant female New Zealand White rabbits/group on GDs 7-20.
Does were sacrificed on GD 29 and reproductive parameters measured. A satellite group of 3-5
pregnant rabbits/group were administered the same concentration and euthanized on GD 21 for
measurement of PFOS in blood and liver samples. Fetuses from the satellite does were removed,
examined grossly and samples pooled by litter. All animals had PFOS in serum and liver
although PFOS levels were higher in the liver. Maternal toxicity was observed at 1.0 mg/kg/day
and higher. At 2.5 mg/kg/day, one doe and at 3.75 mg/kg/day, nine does aborted between GDs
22-28. Mean maternal body weight was decreased significantly at doses >1.0 mg/kg/day. Based
on the findings, the maternal rabbit LOAEL was 1.0 mg/kg/day and the NOAEL was 0.1
mg/kg/day. The developmental LOAEL was 2.5 mg/kg/day based on decreased mean fetal body
weight and a reduction in the ossification of the sternum of the fetuses. The developmental
NOAEL was 1.0 mg/kg/day.
Perfluorooctane sulfonate (PFOS) - February 2014 4-34
Draft - Do Not Cite or Quote
-------
Mouse
As described for rats, a two-part developmental study with PFOS was performed also in
mice by Thibodeaux et al. (2003) and Lau et al. (2003). In the first study, groups of 20-29 CD-I
mice were administered 0, 1, 5, 10, 15 or 20 mg/kg PFOS during GDs 1-17 (Thibodeaux et al.,
2003). Maternal weight gain, food and water consumption and serum clinical chemistries were
monitored and recorded. Mice were euthanized on GD 18. Parameters as described for the rat
were recorded for the mice.
Maternal body weight gain was decreased significantly at 20 mg/kg/day. Food and water
consumption were also affected at the high dose. Increases in serum PFOS were comparable to
the rat but PFOS treatment increased the liver weight in a dose-dependent manner. T4 was
decreased but not as severely as in the rat and the effects of PFOS on the thyroid hormones were
not as pronounced as that seen in the rat. A significant increase in post-implantation losses was
observed in those administered 20 mg/kg/day and small reductions in fetal weight were observed
from dams in the 10 and 15 mg/kg/day groups. Birth defects such as cleft palate, ventricular
septal defect and enlargement of the right atrium were observed at > 10 mg/kg. Benchmark dose
estimates were provided for different parameters but for most, were much higher than those for
the rat. The estimates are as follows: maternal weight reduction (polynomial model) BMD5 =
15.15 mg/kg and BMDL5 = 3.14 mg/kg; maternal T4 effects on GD6 (Hill model) BMD5 = 0.51
mg/kg and BMDLs = 0.35 mg/kg; fetal sternal defects (logistic model) BMDs = 0.06 mg/kg and
BMDL5 = 0.02 mg/kg and fetal cleft palate (logistic model) BMD5 = 7.03 mg/kg and BMDL5 =
3.53 mg/kg.
In the second part of the developmental study, the post-natal effects of in utero exposure
were evaluated in the mouse (Lau et al., 2003). CD-I mice were administered 0, 1, 5, 10, 15 or
20 mg/kg of PFOS in 0.5% Tween-20 by gavage on GDs 1-17.
Most mouse pups from dams administered 15 or 20 mg/kg did not survive for 24 hours
after birth. Fifty percent mortality was observed at 10 mg/kg. Survival of pups in the 1 and 5
mg/kg treated dams were similar to controls. A significant (p< 0.0001) increase in absolute liver
weight was observed at > 5 mg/kg. A significant delay in eye opening was observed at > 5
mg/kg. Thyroid hormones were not affected as prominently as in the rat. The LOAEL for
mouse pups was 5 mg/kg and the NOAEL was 1 mg/kg. Benchmark dose estimates were
provided for survival of the neonates on PND 6 (NCTR model) and were as follows: BMDs =
7.02 mg/kg and BMDL5 = 3.88 mg/kg.
Ten pregnant ICR mice/group were administered 0, 1, 10 or 20 mg/kg of PFOS daily by
gavage from gestational day (GD) 1 to GD 17 or 18 (Yahia et al., 2008). Five dams/group were
killed on GD 18 for prenatal evaluation and histological examination of the maternal liver,
kidneys, lungs and brain; the other five were left to give birth. Body weight, food consumption
and water consumption were monitored in the dams. In the dams killed on GD 18, the gravid
uterus was removed and the number of live/dead fetuses as well as resorptions recorded. Fetal
weight was obtained and the bone, cartilage and skeletal morphology were examined. Four
pups/litter were killed immediately after birth for pulmonary examination.
Perfluorooctane sulfonate (PFOS) - February 2014 4-35
Draft - Do Not Cite or Quote
-------
All dams survived and exhibited no clinical signs. A statistically significant (p< 0.05 or
p< 0.01) decrease in body weight was observed in the dams administered 20 mg/kg/day
beginning on GD 10. Water consumption was increased. Maternal absolute liver weight
increased in a dose-dependent manner, significantly in the 10 (f 59%) and 20 (f 60%) mg/kg/day
groups.
All neonates in the 20 mg/kg/day dose group were born pale, weak, inactive and all died
within a few hours of birth. At 10 mg/kg/day, 45% of those born died within 24 hours. Survival
of the 1 mg/kg/day group was similar to that of controls. Neonatal weight was significantly
decreased at 10 and 20 mg/kg/day. In the fetuses from dams treated with 20 mg/kg/day, there
were large numbers of cleft palates (98.56%), sternal defects (100%), delayed ossification of
phalanges (57.23%), wavy ribs (84.09%), spina bifida occulta (100%) and curved fetus
(68.47%). Similar defects were observed in the fetuses from dams treated with 10 mg/kg/day
except at a lower incidence. See Table 4-9 below. The study did not distinguish or provide any
gender differences.
TABLE 4-9. Effects Observed in the Mice Administered PFOS from GD 0 to GD 17/1 8a
Effects
Number of dams
Total* of fetuses
Live fetuses (%)
Body weight of fetuses (g)
# of fetuses examined
Cleft palate (%)
Sternal defects (%)
Delayed ossification of
phalanges (%)
Wavy ribs (%)
Curved fetus (%)
Spina bifida occulta (%)
Survival rate at PND4 (%)
Control
5
80
98.75 ±1.25
1.49 ±0.01
60
0
0
0
0
3.55±2.11
0
98.18 ±1.82
1 mg/kg
5
76
98.88 ±1.12
1.46 ±0.01
44
1.96 ±1.96
15.77 ±0.99**
1.96 ±1.96
0
4.94 ±2.47
1.96 ± 1.96
100
10 mg/kg
5
79
96.85 ±1.97
1.41 ±0.01**
68
23.36 ±8.27**
52.44 ±2.79**
4.34 ±1.80
7.31 ±0.34*
33.38 ±8.47**
23. 13 ±3.94**
55.20 ±18.98*
20 mg/kg
5
71
90.06 ±3.02*
1.10 ±0.02**
60
98.56 ±1.44**
100**
57.23 ± 9.60**
84.09 ±2.56**
68.47 ±6.71**
100**
o**
a Data from Tables 2-3 in Yahia et al., 2008
* Statistically significant difference between control and treated groups, p< 0.05 or ** p< 0.01
Histopathological exam showed that all fetuses examined on GD 18 from dams treated
with 20 mg/kg were alive, had normal lung structures but mild to severe intracranial dilatation of
the blood vessels. Neonates from the 20 mg/kg treated dams had fetal lung atelectasis (partial or
complete collapse of the lung or a lobe of the lung) with reduction of alveolar space and
intracranial blood vessel dilatation when examined histopathologically. At 10 mg/kg, some pups
had normal lungs and some had severe or focal atelectasis. Three neonates from each of the five
dams treated with 10 mg/kg were examined, 27% had slight lung atelectasis and 87% had mild to
severe dilatation of the brain blood vessel. Based on the significant increase in liver organ
weight, the maternal LOAEL was 10 mg/kg/day and the NOAEL was 1 mg/kg/day. Based on
the abnormalities observed in the fetuses and decreased survival rate, the developmental LOAEL
was 10 mg/kg/day and the NOAEL was 1 mg/kg/day.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-36
-------
4.2.6 Specialized Developmental Studies
Hormonal Disruption
Rat
Yu et al. (2009a) fed adult pregnant Wistar rats (n=20/group) a control diet or a diet
containing 3.2 mg PFOS/kg feed. Treatment continued for both groups throughout gestation and
lactation. Dams were allowed to deliver naturally and on the day of delivery (PND 0), samples
were collected from two control litters and two PFOS treated litters. The remaining litters were
cross-fostered within 12 hours of birth to make the following groups:
1) litters from control dams fostered by control dams (CC, unexposed control; n=8),
2) litters from treated dams fostered by control dams (TC, prenatal exposure; n=8),
3) litters from control dams fostered by treated dams (CT, post-natal exposure; n=8) and
4) litters from treated dams fostered by treated dams (TT, prenatal + postnatal exposure;
n=10).
The pups were weaned on PND 21 and then fed the diet of their rearing dam. Pups were
weighed and killed on PNDs 0, 7, 14, 21 or 35. Serum thyroid hormone analysis was performed
and included total thyroxine (T4), total triiodothyronine (T3), reverse T3 (rT3) and hepatic
expression of genes involved in thyroid hormone (TH) transport, metabolism and receptors. The
genes associated with thyroid metabolism included type 1 deiodinase (DI01) and uridine
diphosphoglucuronosyl transferase 1 Al and 1A6 (UGT1 Al and UGT1A6). Those associated
with thyroid hormone transport included transthyretin (TTR). The genes for the thyroid hormone
receptors a and P (TRa and TRP) were also studied.
No mortality or clinical signs were observed in the dams. Offspring from PFOS treated
groups did not differ significantly from controls in body weight. Liver weights in pups from the
TT group were significantly increased on PND 21 and 35. As observed in other studies, levels of
PFOS in the dams and offspring were higher in the liver when compared to the serum. The
levels of PFOS in both the serum and liver increased with time in the pups exposed postnatally
(group CT) but decreased in those exposed only prenatally (group TC). The levels increased in
those treated both pre- and postnatally (group TT). These results indicate that PFOS can be
transferred by the placenta and by the milk.
The total T3 and rT3 were not affected by PFOS treatment of the pups. Pups in all
groups, except the controls, had significant (p< 0.05 or 0.01) decreases in total T4 on PNDs 21
and 35 with the response in the CT and TT groups being more severe. Pups in the TT group
(exposed pre- and postnatally) were also significantly T4 deficient at PND 14. For gene
expression, no statistically significant differences were observed between litters born to control
dams or litters born to treated dams on PND 0. The only significant finding in gene expression
at the other sacrifice time-points was a significant (p< 0.01) increase (1.5 times greater than the
controls) in TTR on PND 21 in the pups that had been treated both in the prenatal and postnatal
period (group TT). The author noted, however, that there are metabolic differences between rats
and humans that may make the rat's TH status more susceptible to exogenous exposures than
that of humans.
Perfluorooctane sulfonate (PFOS) - February 2014 4-37
Draft - Do Not Cite or Quote
-------
Developmental Immunotoxicity
Mouse
Keil et al. (2008) treated pregnant C57BL/6N females bred with male C3H/HeJ mice
with PFOS to evaluate developmental immunity in their inbred B6C3Fi offspring. The females
(10-12/group) were administered 0, 0.1, 1 or 5 mg/kg PFOS in 0.5% Tween-20 by gavage daily
on gestation day (GDs) 1-17. Pups remained with the dam for approximately 3 weeks with
immunotoxicity evaluations performed at 4 and 8 weeks. Body weight was recorded on the
dams during the study and pups after delivery. Organ weights (spleen, liver, thymus and uterus)
from the pups were recorded at sacrifice.
Natural killer (NK) cell activity was not altered in any pups at 4 weeks old. At 8 weeks,
however, NK cell activity was suppressed in males treated with 1 and 5 mg/kg/day (42.5% and
32.1% decreases compared to controls, respectively) and in females at 5 mg/kg/day (35.1%,
compared to controls). The positive control for NK cell activity produced the appropriate
response. The plaque-forming cell response for sheep red blood cell (SRBC) IgM production by
B cells was only assessed at 8 weeks and was significantly suppressed in the 5 mg/kg/day males
(53%); no effect was observed in the females. The only significant differences in lymphocyte
immunophenotypes was a 21% decrease in absolute numbers of B220+ cells in 4 week old
females in the 5 mg/kg/day group compared to controls; this effect was not observed at 8 weeks.
The other significant change was a 25% decrease in CD3+ and 28% decrease in CD4+
thymocytes at 5 mg/kg/day in males at the 8 week evaluation. Functional responses (nitrite
production) to LPS and interferon-gamma by peritoneal macrophages were not affected with
treatment in the 8 week old mice (not evaluated at 4 weeks). Based on the changes in the
immunotoxicity evaluations observed, the LOAEL in mice is 1 mg/kg/day in males and 5
mg/kg/day in females. The NOAEL is 0.1 mg/kg/day in males and 1 mg/kg/day in females.
Developmental Neurotoxicity
Rat
Twenty five female Sprague-Dawley rats/group were administered 0, 0.1, 0.3 or 1.0
mg/kg/day of potassium PFOS orally by gavage from gestation day (GD) 0 through postnatal day
(PND) 20 (Butenhoff et al., 2009). An additional 10 mated females/group were used to collect
additional blood and tissue samples. Offspring were monitored through PND 72 for growth,
maturation, motor activity, learning and memory, acoustic startle reflex, and brain weight.
There were no treatment-related effects on the pregnancy rates, gestation length, number
of implantation sites, number of pups born, % gender differences, birth to PND 4 survival, PND
4-21 survival, pup body weights through PND 72, and no gross internal findings were identified
at necropsy. A statistically significant, but not lexicologically significant, decrease in mean
maternal body weight and food consumption was observed in rats administered 1.0 mg/kg/day.
Based on results, the maternal toxicity NOAEL was 1.0 mg/kg/day and the LOAEL could not be
determined.
No treatment related effects were observed on functional observational battery (FOB)
assessments performed on PND 4, 11, 21, 35, 45 and 60. Male offspring from dams
administered 0.3 and 1.0 mg/kg/day had statistically significant (p< 0.05) increases in motor
Perfluorooctane sulfonate (PFOS) - February 2014 4-38
Draft - Do Not Cite or Quote
-------
activity on PND 17 but this was not observed on PND 13, 21 or 61. No effect on habituation
was observed in the 0.3 mg/kg/day males but it was decreased at 1.0 mg/kg/day. This effect was
not observed in the females. There were no effects in males or females on acoustic startle
reflexes or in the Biel swimming maze trials. Mean absolute and relative (to body weight) brain
weight and brain measurements (length, width) were similar between the control and treated
animals. Based on the increased motor activity observed with decreased habituation, the
LOAEL for developmental neurotoxicity in male rats was 1.0 mg/kg/day and the NOAEL was
0.3 mg/kg/day.
Ten pregnant Sprague-Dawley rats/group were administered 0, 0.1, 0.6 or 2.0 mg/kg/day
of PFOS by oral gavage in 0.5% Tween 80 from GD 2 to GD 21 (Zeng et al., 2011). On GD 21,
dams were monitored for parturition and the day of delivery was designated PND 0. On PND 0,
five pups/litter were sacrificed and the trunk blood, cortex and hippocampus were collected for
examination. The other pups were randomly redistributed to dams within the dosage groups and
allowed to nurse until PND 21, when they were sacrificed with the same tissues collected as
described for PND 0. PFOS concentration in the hippocampus, cortex and serum increased in a
dose-dependent manner but overall was lower in all tissues on PND 21 when compared to PND
0.
Astrocyte activation markers, glial fibrillary acidic protein (GFAP) and SI00 calcium
binding protein B, which are associated with morphological changes inside the cell, were
evaluated with immunohistochemistry. The number of GFAP positive cells was significantly
increased in the hippocampus and cortex of offspring from treated dams on PND 21. The protein
levels of GFAP in PND 21 offspring were also increased in the hippocampus and cortex on
Western Blot tests. The SI00 calcium binding protein B was increased in the offspring's
hippocampus and cortex on PND 21 in those from dams treated with 0.6 and 2.0 mg/kg/day.
In other tests, PFOS increased the mRNA expression of two inflammatory cytokines,
interleukin (IL-1P) and tumor necrosis factor (TNF)-a. The expression of IL-lp and TNF-a was
significantly increased compared to controls in all treated offspring in the hippocampus on PND
0 and in those from dams administered > 0.6 mg/kg on PND 21. In the cortex, IL-lp and TNF-a
were only significantly increased in the 0.6 mg/kg group and 2.0 mg/kg group, respectively, on
PND 0. On PND 21 in the cortex, IL-lp was increased in those administered > 0.6 mg/kg and
TNF-a only in the highest dosed group.
To determine the mechanisms leading to the inflammatory effect after PFOS exposure,
mRNA levels of three pro-inflammatory transcription factors in both brain tissues were
examined. The most increased effect was observed in the hippocampus on PND 0 with all
treated offspring having a significant increase in activation protein-1 (AP-1) and increases in
nuclear factor-KB (NF- KB) and cAMP response element-binding protein (CREB) observed in
the > 0.6 mg/kg groups. Two synaptic proteins, synapsin 1 (Syn 1) and synaptophysin (Syp)
were also affected; Syn 1 was decreased with PFOS exposure primarily in the hippocampus. Syp
was decreased in the hippocampus but increased in the cortex.
Perfluorooctane sulfonate (PFOS) - February 2014 4-39
Draft - Do Not Cite or Quote
-------
Mouse
Fuentes et al. (2007) treated 8-10 pregnant Charles River CD-I mice/group to 0 or 6
mg/kg of PFOS dissolved in 0.5% Tween-20 daily by gavage on gestation days (GDs) 12-18.
After treatment, mice were either left alone or restrained (immobilized) three times per day for
30 minutes each time to induce maternal stress. Maternal body weight and food and water
consumption were monitored. At birth, the length of gestation, number of live/dead pups, and
sex/weight of pups were recorded.
During the post-natal period, the body weight of the pups was recorded, landmarks for
development were monitored and neuromotor maturation tests (i.e. surface righting reflex,
forelimb grip strength) were conducted. At 3 months of age, the pups were tested in open-field
and rotarod to further assess development. The PFOS treatment had no effect on maternal body
weight or food/water consumption. Pups from dams treated with 6 mg/kg of PFOS had reduced
body weight on PND 4 and 8. Pups from dams exposed to 6 mg/kg of PFOS and having
exposure to restraint exhibited reduced activity in the open-field.
Ten-day old male neonatal NMRI mice (4-7/group) were exposed once to 0, 0.75 or 11.3
mg/kg bw of PFOS by oral gavage (Johansson et al., 2008). Spontaneous behavior (locomotion,
rearing and total activity) and habituation were examined in the mice at 2 and 4 months old.
Behavior was tested in an automated device equipped with horizontal infrared beams. Motor
activity was measured during a 60 minute period divided into three 20 minute sessions.
Locomotion, rearing and total activity were recorded.
No effects were observed on body weight. At 2 months old, mice exposed to 0.75 and
11.3 mg/kg bw of PFOS exhibited significant (p< 0.01) decreases in locomotion, rearing, and
total activity during the first 20 minutes compared to controls. After 60 minutes, activity was
significantly increased in the 11.3 mg/kg bw dose group when compared to controls. The
expected habituation response is for the highest activity pattern to occur in the first 20-minute
period not the last period. The same trend was observed at 4 months in the mice exposed to 11.3
mg/kg bw; at 4 months the responses in the 0.75 mg/kg bw dose group were similar to the
controls. Overall, a single PFOS treatment on PND 10 affected habituation even up to 4 months
of age for mice in the high dose group (11.3 mg/kg/day).
Johansson et al. (2009), administered a single oral dose of 0 (3 litters) or 11.3 (four
litters) mg of PFOS/kg bw to NMRI male mice (10 days old). The exact number of male mice in
each litter was not given. Sacrifice occurred 24 hours after treatment and the brain was
dissected. The cerebral cortex and hippocampus were homogenized to determine if PFOS
affected the protein levels of calcium/calmodulin-dependent protein kinase II (CaMKII), growth-
associated protein-43 (GAP-43), synaptophysin and tau, which are all proteins involved in
neuronal survival, growth and synaptogenesis change during the 'brain growth spurt' (BGS).
There were no acute toxic signs, and no treatment-related body weight differences. The
CaMKII and GAP-43 protein levels were both increased in the PFOS treated males in the
hippocampus; levels were increased 57% (p< 0.001) and 22% (p< 0.01), respectively, when
compared to controls. Protein values in the cerebral cortex were similar between the control and
treated mice. Synaptophysin protein levels were increased significantly (p< 0.001; | 48%) in the
Perfluorooctane sulfonate (PFOS) - February 2014 4-40
Draft - Do Not Cite or Quote
-------
hippocampus and (p< 0.01; f 59%) in the cerebral cortex in the treated mice. The tau protein
levels were increased significantly (p< 0.05; f 80%) in the cerebral cortex only when compared
to controls. Overall the study did support that the neuronal proteins were affected with just a
one-time treatment with PFOS.
4.2.7 Chronic Toxicity
Oral Exposure
Rat
A combined chronic toxicity/carcinogenicity GLP study was performed in compliance
with Good Laboratory Practice (GLP) in 40-70 male and female Crl:CD (SD)IGS BR rats
administered 0, 0.5, 2, 5 or 20 ppm of PFOS in the diet for 104 weeks (Thomford, 2002).
Interim sacrifice results after 14 weeks (Seacat et al., 2002) were described in Section 4.2.3.
Doses were equivalent to approximately 0, 0.018, 0.072, 0.184 and 0.765 mg/kg/day,
respectively, for males and 0, 0.023, 0.099, 0.247, and 1.10 mg/kg/day, respectively, for females.
A recovery group was administered the test substance at 20 ppm for 52 weeks and observed until
death. Five animals/sex in the treated groups were sacrificed during weeks 4, 14 and 53 and
liver samples were obtained for mitochondrial activity, hepatocellular proliferation rate, and
determination of palmitoyl-CoA oxidase activity. Liver weight was recorded only at weeks 14
and 53. The results from the 4-week and 14 week sacrifices are provided in Section 4.2.2 and
7.2.3 respectively. Serum and liver samples were obtained during and at the end of the study for
determination of PFOS concentration.
Survival was not affected by PFOS administration. Males and females administered 20
ppm had statistically significantly decreased mean body weight compared to controls during
weeks 9-37 and 3-101, respectively, but was similar to controls by week 105. The females at 20
ppm had decreased food consumption during weeks 2-44. At the weeks 14 and 53 sacrifice,
absolute and relative (to body weight) liver weights were significantly increased at 20 ppm in
males and relative (to body weight) liver weight was increased at 20 ppm in females. At week
53, organ weight data were given only for the control and 20 ppm groups such that a dose-
response could not be evaluated.
Nonneoplastic lesions in the liver are shown in Table 4-10. At sacrifice, males at 2 ppm
had a significant (p < 0.05) increase in hepatocellular centrilobular hypertrophy. In the males
and females at 5 and 20 ppm, there were significant (p < 0.05) increases in centrilobular
hypertrophy, centrilobular eosinophilic hepatocytic granules and centrilobular hepatocytic
vacuolation. The LOAEL for male rats was 2 ppm (0.072 mg/kg) and for female rats was 5 ppm
(0.247 mg/kg) based on the liver histopathology. The NOAEL for the males was 0.5 ppm (0.018
mg/kg) and 2 ppm (0.099 mg/kg) for females. No effects with treatment were observed on
hepatic palmitoyl-CoA oxidase activity or increases in proliferative cell nuclear antigen (PCNA)
at weeks 4 and 14 or bromodeoxyuridine (BrdU) at week 53. PFOS was identified in the liver
and serum samples of the treated animals and trace amounts were identified in the control
animals. Additional details from the study in regard to carcinogenicity are under Section 4.2.8
Carcinogenicity.
Perfluorooctane sulfonate (PFOS) - February 2014 4-41
Draft - Do Not Cite or Quote
-------
TABLE 4-10. Incidence of nonneoplastic liver lesions in rats
(number affected/total number)
Lesion
Oppm
0.5 ppm
2.0 ppm
5.0 ppm
20 ppm
Males
Centrilobular
hypertrophy
Eosinophilic
granules
Vacuolation
Single cell
necrosis
0/65
0/65
3/65
5/65
2/55
0/55
3/55
4/55
4/55*
0/55
6/55
6/55
22/55**
0/55
13/55**
5/55
42/65**
14/65*
19/65**
14/65*
Females
Centrilobular
hypertrophy
Eosinophilic
granules
Single cell
necrosis
2/65
0/65
7/65
1/55
0/55
6/55
4/55
0/55
6/55
16/55**
7/55**
6/55
52/65**
36/65**
15/65*
Data from Thomford, 2002
Significantly increased over control: *p<0.05; **p<0.01.
4.2.8 Carcinogenicity
Oral Exposure
Rat
As described above in Section 4.2.7, a combined chronic toxicity/carcinogenicity GLP
study was performed in which 40-70 male and female Crl:CD (SD)IGS BR rats were
administered 0, 0.5, 2, 5 or 20 ppm PFOS in the diet for 104 weeks (Thomford, 2002). Doses
were equivalent to approximately 0, 0.018, 0.072, 0.184 and 0.765 mg/kg/day, respectively, for
males and 0, 0.023, 0.099, 0.247, and 1.10 mg/kg/day, respectively, for females. A recovery
group was administered the test substance at 20 ppm for 52 weeks and observed until death.
Five animals/sex in the treated groups were sacrificed during weeks 4, 14 and 53, and liver
samples were obtained for mitochondrial activity, hepatocellular proliferation rate and
determination of palmitoyl-CoA oxidase activity. Serum and liver samples were obtained during
and at the end of the study to determine the concentration of PFOS in them.
Tumor incidence from the study is included in Table 4-11. A significant positive trend
(P=0.0276) was noted in the incidence of hepatocellular adenoma in male rats. This was
associated with a significant increase (P<0.05) in the high-dose group (7/60, 11.7%) over the
control (0/60, 0%). A significantly increased incidence was observed for thyroid follicular cell
adenoma in the high-dose recovery group (9/39, 23.1%) compared to the control group (3/60,
5%). There was also a significant increase in the combined thyroid follicular cell adenoma and
carcinoma in the high-dose males (10/39, 25.6%) compared to that of the control group (6/60,
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-42
-------
TABLE 4-11. Tumor Incidence (%)a
Tumors
0 ppm
0.5 ppm
2 ppm
5 ppm
20 ppm
20 ppm
recovery
MALES
Liver
hepatocellular
adenoma+
Thyroid
follicular cell
adenoma
follicular cell
carcinoma
combined
0 (0/60)
5.0 (3/60)
5.0 (3/60)
10.0 (6/60)
6.0 (3/50)
10.2 (5/49)
2.0 (1/49)
12.2 (6/49)
6.0 (3/50)
8.0 (4/50)
2.0 (1/50)
10.0 (5/50)
2.0 (1/50)
8.2 (4/49)
4.1 (2/49)
10.2 (5/49)
11.7** (7/60)
6.8 (4/59)
1.7 (1/59)
8.5 (5/59)
0 (0/40)
23.1** (9/39)
2.6 (1/39)
25.6 (10/39)
FEMALES
Liver
hepatocellular
adenoma+
hepatocellular
carcinoma
combined+
Thyroid
follicular cell
adenoma
follicular cell
carcinoma
combined
Mammary
Fibroma/
adenoma
carcinoma
combined
0 (0/60)
0 (0/60)
0 (0/60)
0 (0/60)
0(0/60)
0(0/60)
38.3 (23/60)
18.3(11/60)
48.3 (29/60)
2.0(1/50)
0 (0/50)
2.0 (1/50)
0 (0/50)
0 (0/50)
0 (0/50)
60.0** (30/50)
24.0 (12/50)
72.0* (36/50)
2.0 (1/49)
0 (0/49)
2.0 (1/49)
0 (0/49)
0 (0/49)
0 (0/49)
45.8 (22/48)
31.2(15/48)
64.6** (3 1/48)
2.0 (1/50)
0 (0/50)
2.0 (1/50)
4.0 (2/50)
2.0 (1/50)
6.0* (3/50)
52.04 (26/50)
22.0(11/50)
58.0 (29/50)
8.3** (5/60)
1.7 (1/60)
10.0** (6/60)
1.7(1/60)
0 (0/60)
1.7 (1/60)
25 (15/60)
23.3 (14/60)
40 (24/60)
5.0 (2/40)
0 (0/40)
5.0 (2/40)
2.5 (1/40)
0 (0/40)
2.5 (1/40)
40 (16/40)
10 (4/40)
42.5 (17/40)
"Data from Thomford, 2002.
Significant positive trend.
Significantly increased over the control: *p < 0.05; **p < 0.01.
In the females, significant positive trends were observed in the incidences of
hepatocellular adenoma (P=0.0153) and combined hepatocellular adenoma and carcinoma
(P=0.0057) at the terminal sacrifice. These cases were associated with significant increases in
the high-dose group (5/60, 8.3%, and 6/60, 10%) compared to the control (0/60, 0%). A
significant increase (P=0.0471) for combined thyroid follicular cell adenoma and carcinoma was
observed in the mid-high (5.0 ppm) group (3/50, 6%) compared to the control group (0/60, 0%).
Significant increases in mammary fibroadenoma/adenoma (30/50, 60%) and combined
mammary fibroadenoma/adenoma and carcinoma (36/50, 72%) were observed in the low-dose
(0.5 ppm) group compared to the respective controls (23/60, 38.3%; and 29/60, 48.3%). In
addition, the mid-high (5.0 ppm) group (31/48, 64.6%) exhibited a marginally significant
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-43
-------
(P=0.066) increase in the incidence of combined mammary fibroadenoma/adenoma/carcinoma
over the control group (29/36, 48.3%).
Because the tumor incidence does not indicate a dose response, the evidence of
carcinogenicity is suggestive but not definitive.
4.3 Other Key Data
4.3.1 Mutagenicity and Genotoxicity
Results of genotoxicity testing with PFOS are given in Tables 4-12 and 4-13. PFOS was
tested for mutation inductions in the Ames Salmonella/Microsome plate test and in the D4 strain
of Saccharomyces cerevisiae (Litton Bionetics, Inc., 1979). It was also tested in a Salmonella-
Escherichia colil Mammalian-microsome reverse mutation assay with and without metabolic
activation (Mecchi, 1999), an in vitro assay for chromosomal aberrations in human whole blood
lymphocytes with and without metabolic activation (Murli, 1999) and an unscheduled DNA
synthesis assay in rat liver primary cell cultures (Cifone, 1999). In all these assays, PFOS was
negative. In an in vivo mouse micronucleus assay, PFOS did not induce any micronuclei in the
bone marrow of Crl:CD-l BR mice (Murli, 1996). A 50% w/w solution of the
diethanolammonium salt of PFOS in water (T-2247 CoC) was also tested to determine if
induction of gene mutation in 5 strains of S. typhimurium and in S. cerevisiae strain D3 would
take place with and without metabolic activation (Simmon, 1978). The results were negative.
Overall, PFOS does not appear to be mutagenic.
TABLE 4-12. Genotoxicity of PFOS In Vitro
Species
(test system)
Salmonella strains
and D4 strain of
Saccharomyces
cerevisiae
Salmonella strains
and Escherichia coli
WP2uvr
5 strains of S.
typhimurium and S.
cerevisiae strain D 3
Human lymphocytes
Hepatocytes from
Fisher 344 male rat
End-point
Gene mutation
Gene mutation
Gene mutation
Chromosome
aberrations
DNA synthesis
With activation
negative
negative
negative
negative
Without activation
negative
negative
negative
negative
negative
Reference
Litton Bionetics,
Inc., 1979
Mecchi, 1999
Simmon, 1978
Murli, 1999
Cifone, 1999
TABLE 4-13. Genotoxicity of PFOS In Vivo
Species
(test system)
Crl:CD-lBRmice
End-point
Presence of micronuclei in
bone marrow
Results
negative
Reference
Murli, 1996
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-44
-------
4.3.2 Immunotoxicity
Human- in vitro
Eleven volunteers donated blood and peripheral blood mononuclear cells (PBMCs) were
isolated (Brieger et al., 2011). The PBMCs were incubated for 24, 48, or 72 hours and a human
promyelocytic leukemia cell line (HL-60) was incubated for 24 hours in normal culture medium
containing PFOS at the following concentrations: 3.9, 7.8, 15.6, 31.3, 62.5 and 125 |ig/mL.
Viability was determined after incubation by measuring neutral red uptake. No significant
reduction of viability was observed up to 125 |ig/mL; however, the highest concentration for
PFOS could not be evaluated due to limited solubility. Therefore, 100 |ig/mL was the highest
concentration used thereafter. The impact of PFOS on TNF-a and IL-6 release was also
determined in whole blood. PFOS reduced the release of the pro-inflammatory cytokine TNF-a
after lipopolysaccharide (LPS) stimulation. Natural killer cell activity was examined by culturing
PBMCs for 24 hours in the presence of 0, 1, 10 or 100 |ig/mL of PFOS with labeled K562 target
cells (labeled with 6-carboxyfluorecein succinimidyl ester [CFSE] diacetate). PFOS decreased
NK-cell-mediated killing of K562 cells, reducing NK-cell cytotoxicity by 32%, and was
statistically significant at 100 |ig/mL. This study suggests some effects on immunity in humans;
however the sample size used is small and the dose in which effects were observed are much
higher than that typically observed in humans.
Mouse
Peden-Adams et al. (2008) gave PFOS in Milli-Q water containing 0.5% Tween 20, daily
by gavage for 28 days to five adult male and female B6C3Fi mice/group. Equivalent daily
doses to the animals were 0, 0.00017, 0.0017, 0.0033, 0.017, 0.033, and 0.166 mg/kg/day,
respectively. Animals were euthanized at the end of treatment. Various immune parameters,
including lymphocytic proliferation, natural killer cell activity, lysozyme activity, antigen
specific IgM production, lymphocyte immunophenotypes, as well as serum PFOS concentrations
were determined after exposure.
The following were not affected with treatment: survival; behavior; body weight; spleen,
thymus, kidney, gonad and liver weights; and lymphocytic proliferation. Lysozyme activity was
not affected in males but increased significantly in females at 0.0033 and 0.166 mg/kg/day
compared to the control group; however, this did not follow a dose response. Natural killer cell
activity was increased significantly (p< 0.05) 2 to 2.5-fold in males at 0.017, 0.033 and 0.166
mg/kg/day, but was not affected in any of the females. Splenic T-cell immunophenotypes were
slightly affected in females but were significantly altered in males treated with > 0.0033
mg/kg/day. In both genders, thymus cell populations were less sensitive to PFOS. Male thymic
T-cell subpopulations were not affected with PFOS treatment and in females were increased only
at 0.033 and 0.166 mg/kg/day.
The sheep red blood cell (SRBC) plaque-forming cell response (IgM production) was
suppressed and demonstrated a dose response in males beginning at 0.0017 mg/kg/day and in
females at 0.017 mg/kg/day. In males it was suppressed by 52-78% and females by 50-74%.
Because IgM suppression can result from effects on both T- and B-cells, a T-cell independent
test was performed after the SRBC (T-cell dependent) test and it indicated suppression. An
additional group of female mice was treated with 0 or 0.334 mg/kg/day of PFOS orally for 21
Perfluorooctane sulfonate (PFOS) - February 2014 4-45
Draft - Do Not Cite or Quote
-------
days and challenged with a TI antigen [a trinitrophenyl (TNP) lipopolysaccharide (LPS)
conjugate]. Serum TNP-specific IgM liters were decreased after the TNP-LPS challenge with
serum levels of TNP-specific IgM significantly suppressed by 62% compared with controls.
Based on the IgM suppression observed in both the T-cell independent and dependent tests,
humoral immune effects can be attributed to B-cells, rather than T-cells. Serum levels of PFOS
were similar between males and females. Based on the results the LOAEL in mice is 0.0017
mg/kg/day in males and 0.017 mg/kg/day in females. The NOAELs are 0.00017 mg/kg/day in
males and 0.0033 mg/kg/day in females.
Qazi et al. (2009a) administered diets containing 0, 0.001%, 0.005% , 0.02% (40 mg/kg
bw/day), 0.05% (100 mg/kg bw/day), 0.1%, 0.25%, 0.5% or 1% PFOS and 0.02% PFOA for 10
days to four to six male (six to eight week old) C57B1/6 mice/group. The PFOS and PFOA were
dissolved in 20 mL of acetone prior to being mixed with the chow and then dried to allow the
odor of the acetone to dissipate prior to administration. At the end often days, mice were bled
for analysis of PFOA and PFOS, and then killed. Weights were obtained for the thymus, spleen,
liver and epididymal fat. The number of thymocytes and splenocytes were also measured and
checked for viability. Histology was also performed on the thymus and spleen.
Mice treated with dietary concentrations of > 0.02% (-40 mg PFOS/kg bw/day) PFOS
exhibited pronounced weight loss (>20%), a decrease in food consumption (> 40%), and lethargy
and were withdrawn from the experiment after 5 days of exposure. The author stated that this
was not due to taste aversion as it was observed when PFOS was administered intraperitoneally
or subcutaneously. The background levels of PFOS and PFOA were both similar in the control
mice; however, after administration of 0.02% in the diet, the serum level of PFOS was
approximately twice that of PFOA. Only the animals treated with 0.02% PFOS had a significant
decrease in total body weight and in the wet weights of the thymus, spleen and epididymal fat
pads compared to the controls. However, all three doses resulted in a significant increase
(p<0.05 or 0.01) in liver weight, compared to controls. Similar findings slightly more
pronounced were observed in mice administered PFOA. The mice administered 0.02% of PFOS
demonstrated a marked decrease in the total number of thymocytes (84% of controls) and
splenocytes (43% of controls) as well as having thymocytes and splenocytes reduced in size.
Finally, in the mice administered 0.02% PFOS or PFOA, the thymic cortex was small and devoid
of cells and the cortical/medullary junction was not distinguishable. No obvious histological
differences in the spleen of the mice administered any dose of PFOA or PFOS were observed.
In this same study, work was done to determine the role of PPARa which will be discussed in
Section 4.3.4. Briefly, results of that part of the study indicated that the immunomodulation was
partially dependent on PPARa.
Qazi et al. (2009b) also performed a study to see if exposure to PFOS influenced the cells
of the innate immune system. Four male C57B1/6 mice per dose were exposed to rat chow
supplemented with 0, 0.001 or 0.02% PFOS for 10 consecutive days. A second similar study
was performed to determine if the exposure to the chemicals influenced innate immune
responses to bacterial lipopolysaccharide (LPS). In this study, mice were exposed as described
and then on day 10, some mice were injected intravenously with 0.1 mL sterile saline containing
300 jig LPS (Escherichia coli), while others received the vehicle only. In the first study, mice
were bled directly after the 10 day exposure and in the second study mice were bled 2 hours after
Perfluorooctane sulfonate (PFOS) - February 2014 4-46
Draft - Do Not Cite or Quote
-------
administration of LPS. The spleen, thymus, epididymal fat and liver were collected as well as
peritoneal and bone marrow cells.
No effects were observed in any of the mice exposed to 0.001% PFOS. Exposure to
0.02% PFOS caused an increase in liver weight and a decrease in the weight of other organs and
overall body weight. Food consumption in these mice was also decreased 25% when compared
to control mice. The total intake of PFOS over the 10 days was approximately 6 mg and the total
concentration of PFOS in the serum was 340 ±16 |ig/mL (ppm). In the first study of the innate
immune system, the overall total number of white blood cells and lymphocytes were decreased;
however, neutrophil counts were similar to controls. The number of macrophages in the bone
marrow was increased but the peritoneal and splenic macrophages were not. The second study
also found that cells isolated from the peritoneal cavity and bone marrow, but not spleen, of mice
exposed to the high level of PFOS had enhanced levels of the pro-inflammatory cytokines, TNF-
a and IL-6 in response to stimulation by LPS. The levels of these cytokines in the serum were
not elevated. This study indicates that PFOS can have an effect on the immunotoxicity of mice.
Potassium PFOS suspensions were made with deionized water with 2% Tween 80 and
administered orally by gavage at doses of 0, 5, 20 or 40 mg/kg bw to twelve male (8-10 wks old)
C57BL/6 mice/group daily for 7 days (Zheng et al. 2009). Food consumption and body weight
were measured daily for 7 days. Mice were bled on the eighth day (24 hours after the last
treatment) and subsequently killed. The blood was analyzed for corticosterone and PFOS
concentration. Spleen, thymus, liver and kidneys were collected, weighed and the spleen and
thymus were processed into suspensions to look at functional immune endpoints and T-cell
immunophenotype determinations.
At 20 and 40 mg/kg bw/day starting about day 3, mean body weights were significantly
decreased compared to the controls. Food consumption decreased with treatment; food
consumption in the control animals on study day 0 was 5.9 ± 0.2 g and was similar during the
study. In the mice treated with 20 and 40 mg/kg bw/day, day one values were 5.8 ± 0.3g and 5.9
± 0.2 g and dropped to 2.8 ± 0.1 g and 1.6 ± 0.1 g on day 7, respectively.
At the end of treatment, the body weight, splenic and thymic weights were all decreased
at 20 and 40 mg/kg bw/day, compared to the controls and the liver weight increased by 34, 79
and 117% over controls at 5, 20 and 40 mg/kg bw/day, respectively. A dose-dependent increase
in PFOS was observed in the serum samples with no PFOS identified in the control mice; serum
corticosterone levels increased significantly in the mice treated with doses > 20 mg/kg/day.
Splenic and thymic cellularity was significantly decreased (p < 0.05) at 20 and 40 mg/kg bw/day;
cellularity in the spleen and thymus in the mice administered 40 mg/kg/day were decreased by
51 and 61%, respectively, compared to the control mice. To determine population changes in
functional cell types of spleen and thymic lymphocytes, CD4/CD8 marker analysis was
performed. Significant decreases in CD4+ and CD8+ cells were observed in both the spleen and
thymus in the mice administered > 20 mg/kg/day PFOS. A lactate-dehydrogenase (LDH) release
assay was performed to determine natural killer (NK) cell activity and this was decreased
significantly in the two highest dosed groups. The average NK-cell activity in control mice was
50.33 ± 4.08 with the activity at 20 mg/kg/day being 18.04 ± 1.42 and at 40 mg/kg/day, 13.08 ±
1.11. Finally, treatment in all groups of mice resulted in a significant suppression of the plaque-
forming cell response after 7 days of treatment; results were 63, 77 and 86% that of controls at 5,
Perfluorooctane sulfonate (PFOS) - February 2014 4-47
Draft - Do Not Cite or Quote
-------
20 and 40 mg/kg bw/day, respectively. Based on the increase in liver weight and the suppression
of the plaque-forming cell response, the LOAEL was 5 mg/kg bw/day in mice and the NOAEL
could not be determined.
In order to observe chronic effects of immunotoxicity, adult male C57BL/6 mice
(10/group) were administered PFOS delivered in de-ionized water with 2% Tween 80 daily by
gavage for 60 days to doses of 0, 0.008, 0.083, 0.417, 0.833 and 2.083 mg/kg/day (Dong et al.,
2009). Parameters similar to those described above for Zheng et al. (2009) were measured.
At sacrifice, mice treated with > 0.417 mg/kg/day had significantly lower body weight
compared to the control mice, as well as significant decreases in spleen, thymus and kidney
weight. Food consumption in the study was decreased in mice at 0.833 and 2.083 mg/kg/day.
Liver weight was increased significantly in all treated mice starting at 0.083 mg/kg/day. The
mean liver weight in control mice was 5.17 ± 0.12 g and the liver weights in the treated mice
were 5.21 ± 0.17 g (0.0.008 mg/kg/day), 5.78 ± 0.13 g (0.083 mg/kg/day), 6.67 ± 0.11 g (0.417
mg/kg/day), 8.17 ± 0.21 g (0.833 mg/kg/day) and 11.47 ± 0.12 g (2.083 mg/kg/day). Serum
corticosterone was similar to that in controls in the mice receiving 0.008, 0.083 or 0.417 mg
PFOS/kg/day but was decreased in those at the two higher doses. As in the shorter-term study,
thymic and splenic cellularity was decreased in a dose-dependent trend, with significant
decreases observed in mice receiving > 0.417 mg/kg/day. The CD4/CD8 marker analysis
performed on splenic and thymic lymphocytes demonstrated that the numbers of T cell and B
cell CD4/CD8 subpopulations were decreased starting at 0.417 mg PFOS/kg/day. Splenic NK
cell activity was increased significantly compared to controls (31.14 ± 1.93%) in the mice at
0.083 mg PFOS/kg/day (45.43 ± 4.74%) with marked decreases at 0.833 mg/kg/day (20.28 ±
2.51%) and 2.083 mg/kg/day (15.67 ± 1.52%). Based on the findings in the 60 day study, the
NOAEL was 0.008 mg/kg/day and the LOAEL was 0.083 mg/kg/day. The serum concentration
at the LOAEL was 7.132 mg/L.
Guruge et al. (2009) administered 0, 5 or 25 |ig/kg PFOS (0, 0.005 or 0.025 mg/kg,
respectively) in 30 female B6C3Fi mice/group for 21 days and then exposed them intranasally to
100 plaque forming units (pfu; in 30 jiL of phosphate buffered saline [PBS]) influenza A virus
suspension. Mice were observed for 20 days past inoculation. Concentrations of PFOS in the
plasma, spleen, thymus and lung all showed a dose dependent increase; however, there was not a
significant change in body or organ weights (spleen, thymus, liver, kidney and lung) of the
treated mice compared to the controls. Survival rate was significantly decreased in the mice at
25 |ig/kg PFOS after viral exposure. Survival rate in the mice on day 20 was 46% in the controls
and 17% in the high-dose group.
Male C57BL/6 (H-2b) mice (n=7) were administered PFOS in the diet at 0.005% (w/w)
for 10 days to determine the effect on the histology and immune status of the liver (Qazi et al.,
2010). Control mice were also used. There was no effect on body weight, food intake, thymus,
spleen or fat pad mass, serum levels of ALT or AST, hematocrit, hemoglobin or the numbers of
thymocytes and splenocytes. However, the liver mass was increased 1.6-fold when compared to
untreated controls although necrosis was not observed. Total serum cholesterol was also
decreased and there was a moderate increase in serum alkaline phosphatase (ALP). At the end of
the study, the total mean serum PFOS concentration for four mice was 125.8 |ig/mL. The livers
of the treated mice also displayed hypertrophic hepatocytes surrounding the central vein. PFOS
Perfluorooctane sulfonate (PFOS) - February 2014 4-48
Draft - Do Not Cite or Quote
-------
only increased one type of intrahepatic immune cells (TER119+) while a corresponding test with
0.002% PFOA increased all types of intrahepatic immune cells. PFOS treated mice, however,
had normal responses in the intrahepatic B and T cells by producing enough IgM. The treated
mice also had lower levels of the hepatic cytokines, TNF-a, IFN-y and IL-4, when compared to
the control mice and an increase in hepatic erythropoietin.
4.3.3 Physiological or Mechanistic Studies
4.3.3.1 Noncancer Effects
Hormone Disruption
Martin et al. (2007) administered 10 mg PFOS /kg to adult male Sprague-Dawley rats (n
= 5) for 1, 3, or 5 days by oral gavage and determined the impact of PFOS on hormone levels.
Blood was collected via cardiac puncture and the serum was analyzed for cholesterol,
testosterone, free and total T4, and total T3. RNA extracted from the livers was used for gene
expression profiling, genomic signatures, and pathway analyses to determine a mechanism of
toxicity.
Following a 1-day, 3-day, and 5-day dose, a significant decrease (p<0.05) was observed
in total T4 (~|47- 80%) and free T4 (~|60-82%). The total T3 was only significantly deceased
after day 5 (|~ 23%). Serum cholesterol was significantly decreased (p<0.05) after dosing for 3
and 5 days. Serum testosterone was similar to controls at all timepoints. PFOS treatment was
matched to hepatomegaly and hepatocellular hypertrophy. Genes associated with the thyroid
hormone release and synthesis pathway including Dio3, which catalyzes the inactivation of T3
and Diol, which deiodinates prohormone T4 to bioactivate T3. Treatment with PFOS caused
significant (p<0.05) Diol repression andDioS induction only on day 5.
Chang et al. (2007) tried to determine if the decrease of free thyroxine (fT4) often
observed in animals upon PFOS exposure was due to competition for carrier protein binding
interference. The study used equilibrium dialysis radioimmunoassay (ED-RIA) for fT4
measurements in in vitro and in vivo protocols. PFOS did not decrease serum total thyroxine
(TT4) or fT4 at concentrations up to 200 jiM in vitro. Female rats administered three daily 5
mg/kg oral doses of PFOS also had no changes to serum TSH and fT4 when checked by ED-
RIA. The authors suggested that further testing for thyroid hormone parameters should use a
reference method such as ED-RIA for determining serum fT4 as analog methods may falsely
appear to decrease free thyroid hormones.
Chang et al. (2008) also used a study on thyroid hormone status to determine if exposure
to PFOS in rats caused a competition with thyroxine for serum binding proteins. Three different
experimental designs were employed in this three-part study. In the first part, five to fifteen
female Sprague-Dawley rats/group were given either a single oral dose of vehicle (0.5% Tween
20 in distilled water; three groups) or 15 mg potassium PFOS/kg bw (three groups) suspended in
vehicle. Rats were killed at 2, 6 and 24 hours post-dosing and blood samples obtained. The
following thyroid parameters were measured: serum fT4, total thyroxine (TT4), triiodothyronine
(TT3), reverse triiodothyronine (rT3), and thyrotropin (TSH; measured at the 6 and 24 hour
timepoints only). PFOS concentrations in the blood and liver were also measured and the
following hepatic biochemical markers were measured: UDP-glucuronosyltransferase 1A
Perfluorooctane sulfonate (PFOS) - February 2014 4-49
Draft - Do Not Cite or Quote
-------
[UGT1A] family of mRNA transcripts (involved in glucuronidation and T4 turnover), malic
enzyme [ME] mRNA transcripts and ME activity (indicators of tissue response to thyroid
hormone).
Serum TT4 decreased significantly (p<0.05) compared to controls after 2 hours (|24%), 6
hours (|38%) and 24 hours (|53%). The TT3 and rT3 only decreased significantly at the 24
hours time-point, while fT4 was increased significantly at 2 and 6 hours (68 and 90% over
control, respectively) before becoming similar to that of controls at the 24 hour time-point.
Serum levels of PFOS were significantly (p<0.05) higher than controls at all time-points but
were the highest at 6 hours. A similar trend was observed with the concentration of PFOS in the
liver except the values were slightly less and continued to increase through the 24 hour time-
point. The ME and UGT1A mRNA transcripts were significantly increased (p< 0.05) only at the
2 hour time-point, compared to controls, and the ME activity was increased significantly only at
the 24 hour sampling.
In the second part of the study, Sprague-Dawley rats were injected intravenously with
either 9.3 jiCi (females; n=5/group) or 11 jiCi (males; n= 4/group) of 125I-T4 followed by a
single oral dose of either vehicle or 15 mg potassium PFOS/kg bw. Urine and feces were
collected for 24 hours after administration to determine the 125I elimination. At the end of the 24
hours, the animals were killed and liver and serum samples collected. Serum TT4 concentration
was decreased by 55% in the PFOS treated males and females compared to controls. There was
also a decrease in serum 125I in the treated males. Liver 125I radioactivity decreased by 40 and
30% in males and females, respectively, but the urine and feces 125I radioactivity increased with
the males exhibiting the most activity. This indicates a loss of thyroid hormones and increased
turnover.
In the last part of the study, adult male Sprague-Dawley rats (4-6/group) were
administered either vehicle only by gavage, 3 mg/kg bw of potassium PFOS suspended in
vehicle by gavage, 10 |ig/mL (10 ppm) propylthiouracil (PTU) in drinking water, or 10 ppm
PTU in drinking water + 3 mg PFOS/kg bw for 7 consecutive days. PTU is an inhibitor of
thyroid hormone synthesis. On days 1, 3, 7 and 8, TT4, TT3 and TSH were monitored and on
day 8, the pituitaries were removed and placed in static culture to assess thyrotropin releasing
hormone (TRH)-mediated release of TSH. During the days of dosing with PFOS, TSH levels
did not increase although TT4 and TT3 were decreased. Pituitary response to TRH-mediated
TSH release was not affected or lessened after the PFOS-only administration.
Results suggest that oral PFOS administration results in a transiently increased tissue
availability of thyroid hormones, increased turnover of T4 and a reduction in TT4, but does not
induce a typical hypothyroid state or alter the hypothalamic-pituitary-thyroid axis.
In the study by Curren et al. (2008; Section 4.2.2) where Sprague-Dawley rats
(15/sex/group) were administered 0, 2, 20, 50 or 100 mg PFOS/kg diet for 28 days, T4 and T3
levels were decreased. T4 levels were statistically significantly decreased at >20 mg PFOS/kg
diet, when compared to the control levels, in both males and females. T3 levels were decreased
significantly at >50 mg/kg diet in the females and 100 mg/kg diet in the males. There were no
treatment-related changes observed with absolute thyroid weight.
Perfluorooctane sulfonate (PFOS) - February 2014 4-50
Draft - Do Not Cite or Quote
-------
Yu et al. (2009a) fed adult pregnant Wistar rats (n=20/group) a control diet or a diet
containing 3.2 mg PFOS/kg feed. Treatment continued for both groups throughout gestation and
lactation. Dams were allowed to deliver naturally and on the day of delivery (PND 0), samples
were collected from two control litters and two PFOS treated litters. Litters were cross-fostered
to help determine if PFOS had more effect when administered prenatally, postnatally or both.
The total T3 and rT3 were not affected with PFOS treatment in the pups. Pups in all groups,
except the controls, had significant (p< 0.05 or 0.01) decreases in total T4 on PNDs 21 and 35.
Pups exposed pre- and postnatally were also significantly T4 deficient at PND 14.
Male Sprague-Dawley rats (8-10/group) were administered 0, 1.7, 5.0 or 15.0 mg/L
PFOS in drinking water for 91 days (Yu et al., 2009b). At the end of exposure, serum was
collected and analyzed for total thyroxine (T4), free thyroxine (fT4), total triiodothyronine (T3)
and thyrotropin (TSH) as well as liver and thyroid organ weights obtained. Also measured were
messenger RNA (mRNA) levels for two isoforms of uridine diphosphoglucuronosyl transferase
(UGT1A6 and UGT1A1) and type 1 deiodinase (DIO1) in liver, sodium iodide symporter (NTS),
TSH receptor (TSHR) and DIO1 in thyroid and activity of thyroid peroxidase (TPO).
No treatment-related effects were observed on body weight or thyroid absolute and
relative weight. Absolute and relative (to body wt) liver weights were increased significantly
(p< 0.05 or 0.01) in the rats administered 5 and 15 mg/L. Levels of the thyroid hormone activity
measured are in Table 4-14 and show that total T4 decreased in a significant dose-dependent
manner in the treated rats. Serum fT4 was only decreased at 5 mg/L, total T3 was only increased
at 1.7 mg/L, and there was no effect on TSH.
TABLE 4-24. Thyroid hormone levels in PFOS treated rats
Dose administered
mg/L
0
1.7
5.0
15.0
Total T3
(Hg/L)
0.29 ±0.04
0.48* ±0.08
0.23 ±0.05
0.23 ±0.03
Total T4
(Hg/L)
40.9 ±1.8
23. 9** ±1.3
16.4** ±5.4
8.5**± 1.6
Free T4
(pmol/L)
19.0 ±1.3
16.7 ±1.4
12.6* ±1.5
17.3 ±1.1
TSH
(IU/L)
0.72 ±0.30
0.67 ± 0.27
1.12 ±0.34
1.62 ±0.67
PFOS
(mg/L)
5
mg/L also lowered DIO1 mRNA in the liver when compared to controls. The DIO1 levels in the
thyroid increased in these same treatment groups by 1.8 and 2.9 fold, respectively, compared to
controls. PFOS treatment had no effect on NTS, TSHR or TPO activity.
Six female Wistar rats/dose were administered 0, 0.2, 1.0, or 3.0 mg/kg of PFOS by oral
gavage daily for 5 consecutive days (Yu et al., 2011). Groups of six were also administered
propylthiouracil at 10 mg/kg or PTU (10 mg/kg) + PFOS (3.0 mg/kg) in the same manner.
Serum and bile were evaluated for total T4 (TT4), total T3 (TT3), transthyretin and
thyroglobulin. Serum TT4 and TT3 both decreased significantly at 1.0 and 3.0 mg/kg for the TT4
(-63.7 and 58.9% of controls) and 3.0 mg/kg for the TT3 (~ 62.9% of the control value). The
values in bile were not affected and were similar to controls. Serum transthyretin and
thyroglobulin were also similar to controls. As stated earlier, Yu et al. also found that OATp2
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-51
-------
was increased significantly (143% compared to controls) in rats at 3.0 mg/kg indicating that this
may be involved in hepatic T4 uptake and could be one reason the serum TT4 was decreased.
This could also possibly be causing the TT3 decrease. Relative liver weight, absolute and
relative thyroid weight were all increased significantly with treatment of PFOS, PTU and PFOS
+ PTU. In the thyroid, PTU had the most effect followed by the PFOS/PTU mixture and then
the PFOS alone. In the liver, PFOS alone had the most effect.
PPAR activity
Studies have been conducted with PFOS in order to determine if it activates peroxisome
proliferator-activated receptors (PPARs). The PPARs are members of the nuclear hormone
receptor superfamily of ligand-activated transcription factors. These factors can alter gene
expression in response to endogenous and exogenous ligands and are associated with lipid
metabolism, energy homeostasis and cell differentiation. The three types, PPARa, P/5 or y, are
encoded by different genes, expressed in many tissues and have specific roles during
development as well as in the adult (Takacs and Abbott, 2007).
In vitro
Shipley et al. (2004) tested PFOS to determine if it activated human or mouse PPARa in
a COS-1 cell based luciferase reporter trans-activation assay. The COS-1 is a fibroblast-like cell
line derived from monkey kidney. Concentrations at 8, 16, 32, 64, 125, 250, 500 and 100 jiM
were tested. The COS-1 cells were transfected with either mouse or human PPARa expression
plasmid along with the reporter plasmid, pHD(x3)luc, which has three PPAR binding sites that
are linked to a minimal promoter controlling the gene for Firefly luciferase. Cells were also
coinfected with a plasmid encoding Renilla luciferase to serve as a control. A positive control,
Wy-14,643, was also used. In the experiments, PFOS activated both human and mouse PPARa.
The highest PFOS-activation was 4-6-fold and was similar to that obtained with the positive
control. The average ECso was 13 jiM in the mouse and 15 jiM in the human PPARa.
Both PFOS and PFOA were tested to determine whether they could activate peroxisome
proliferator-activated receptors (PPARs) with a transient transfection cell assay (Takacs and
Abbott, 2007). The Cos-1 cells were cultured in Dulbecco's Minimal Essential Medium
(DMEM) with fetal bovine serum in 96-well plates and transfected with mouse or human
PPARa, P/5 or y reporter plasmids. Transfected cells were then exposed to PFOS (1-250 jiM),
positive controls (known agonists and antagonists) or negative controls (DMEM, 0.1% water and
0.1% dimethyl sulfoxide). The positive control agonists and antagonists were WY-14643 and
MK-886, respectively, for PPARa and troglitazone and GW9662, respectively, for PPARy; only
the agonist L165,041 was used for PPAR.p/5. After treatment for 24 hours, activity was
measured using the Luciferase reporter assay. The agonist used for mouse and human PPARa
was WY-14,643 and it exhibited 15- and one-fold increase, respectively over the luciferase
response of the negative controls. The agonist for mouse and human PPARp/5, L165,041,
exhibited 28- and 13-fold increases in the luciferase response, respectively, compared to the
negative controls. Finally, troglitazone, the agonist for mouse and human PPARy, increased the
luciferase response 3- and 2-fold over the negative controls, respectively. The antagonists
showed appropriate inhibitory responses with maximum inhibition of agonist activity of 90 and
60% for mouse and human PPAR a, respectively, and 47 and 45% for mouse and human PPARy.
Perfluorooctane sulfonate (PFOS) - February 2014 4-52
Draft - Do Not Cite or Quote
-------
In this study, PFOS activated the mouse PPARa with a significant (p< 0.01) 1.5-fold
increase in activity at 120 jiM PFOS, compared to the negative control. PFOS did not
significantly increase activity in the human PPARa construct. PFOS activated the mouse
PPARP/5 but not the human PPARp/5 construct. It did not activate the mouse or human PPARy
construct. Table 4-15 shows summary data. The authors concluded that PFOA activated PPARa
more than PFOS and the mouse was more responsive than the human. PFOA and PFOS both
activated mouse but not human PPAR.p/5 and neither chemical activated human or mouse
PPARy.
TABLE 4-35. Summary of PFAA Transactivation of Mouse and Human PPARa, p/5 and ya
PPAR isoform
a
P/5
Y
PFAA
PFOA
PFOS
PFOA
PFOS
PFOA
PFOS
Mouse LOECb
10 nM
120 nM
40 nM
20 nM
NA
NA
Human LOECb
30 nM
NA
NA
NA
NA
NA
a Data from Table 1 in Takacs and Abbott, 2007
b LOEC = lowest concentration (|iM) at which there was a significant difference compared to the negative control (p< 0.05)
Wolf et al. (2008) tested PFAAs, including PFOS, to determine if mouse and human
PPARa activity could be induced in transiently transfected COS-1 cell assays. COS-1 cells were
transfected with either a mouse or human PPAR-a receptor-luciferase reporter plasmid and after
24-hours were exposed to either negative controls (water or 0.1% dimethyl sulfoxide [DMSO]),
a positive control (WY14,643) or PFOS at 1-250 |iM. At the end of 24-hours of exposure, the
luciferase activity was measured. The positive and negative controls had the expected results. A
lowest observed effect concentration (LOEC) and no observed effect concentration (NOEC) was
determined. In the study, the mouse PPARa was more responsive than the human. The NOEC
for PFOS was 60 jiM in the mouse and 20 jiM in humans; the LOEC was 90 jiM (48.4 |ig/mL)
in the mouse and 30 jiM (16.2 jig/mL) in humans.
In a study similar to that above but including additional PFAAs, Wolf et al. (2012)
incubated transfected cells with PFAAs at concentrations of 0.5 to 100 jiM, vehicle (water or
0.1% DMSO as negative control) or with 10 jiM WY14,643 (positive control). Assays were
performed with 3 identical plates per compound per species with 9 concentrations/plate and 8
wells/concentration. Cell viability was assessed using the Cell Titer Blue cell viability kit and
read in a fluorometer. The positive and negative controls had the expected results. All PFAAs
significantly induced human and mouse PPARa. The study also provided the C20max which was
the concentration at which a PFAA produced 20% of the maximal response elicited by the most
active PFAA. For PFOS, this was 94 jiM in mouse PPARa and 262 jiM in human PPARa. For
comparison, PFOA was 6 jiM and 7 jiM, respectively.
Ishibashi et al. (2011) assessed the transactivation potencies of the Baikal seal
peroxisome proliferator-activated receptor a (BS PPARa) with various PFCs using an in vitro
reporter gene assay. They tested eight perfluoroalkyl carboxylates and two perfluoroalkyl
sulfonates, including PFOS. As found in the two Wolf studies above, the authors found that the
number of perfluorinated carbons was one of the factors determining the transactivation
potencies of the BS PPARa, and the carboxylates were more potent than the sulfonates with the
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-53
-------
same carbon numbers. The transactivation potencies were measured between the compounds by
estimating the PFOA induction equivalency factor (IEF) which was the ratio of the 50% effective
concentration of PFOA to the concentration of each compound that can induce the response
corresponding to 50% of the maximal effect of PFOA. The IEF would then be 1 and efficacy (%
of PFOA) would be 100%. For PFOS, the IEF was 0.26 and the efficacy was 45%.
In vivo
Rats
Martin et al. (2007) administered PFOS to male Sprague-Dawley rats by oral gavage at
doses of 0 or 10 mg/kg/day for one, three or five consecutive days. Clinical chemistry,
hematology, histopathology and gene expression profiling of the livers from three rats/group
were performed. Body weight was not affected with treatment but relative liver weight increased
after 5 days of treatment. In gene expression, PFOS exhibited peroxisome proliferator-activated
receptor alpha agonist-like effects on genes associated with fatty acid homeostasis. Exposure
also caused down-regulation of cholesterol biosynthesis genes. PFOS caused significant type 1
iodothyronine deiodinase (Diol) repression and Dio3 induction on day 5 of exposure which
corresponded to decreases of T3 only on day 5 and total and free T4 decreases. Diol deiodinates
thyroxine (T4) to bioactivate T3 and Dio3 catalyzes the inactivation of T3. PFOS was poorly
correlated with peroxisome proliferators in the global gene expression patterns and indicated
weak matches with hepatotoxicity related signatures and weak correlation to PPARa agonist
treatment. Expression of HMG-CoA reductase was significantly upregulated and cholesterol
biosynthesis was downregulated in a manner suggesting a mechanism distinct from the statins.
The authors suggested a link between PFOS, PPAR, and thyroid hormone homeostasis based on
work by Miller et al. (2001) who observed decreased serum T4 and T3 levels and increased
hepatic proliferation following exposure to peroxisome proliferators. They also noted that PFOS
exhibited similarities to compounds that induce xenobiotic metabolizing enzymes through
PPARy and CAR.
Wang et al. (2010) dosed albino Wistar female rats in the feed with 3.2 mg/kg of PFOS
from gestation day (GD) 1 to weaning (postnatal day [PND] 21). Pups were allowed access to
the treated feed until PND 35. Pups on PND 2 were divided into groups: pups born to treated
dams fostered by control dams; pups born to control dams fostered by treated dams; pups born to
control dams fostered by other control dams and pups born to treated dams fostered by other
treated dams. This was done to determine if prenatal or lactational exposure had more effect on
altering gene expression. Gene expression changes were examined on PNDs 1, 7 and 35.
Significant effects were observed on genes involved in neuroactive ligand-receptor interaction,
calcium signaling pathways, cell communication, the cell cycle and peroxisome proliferator-
activated receptor (PPAR) signaling. Transthyretin (TTR) which is a serum and cerebrospinal
fluid carrier of thyroxine (T4) was decreased after PND 1. Based on analysis of PFOS in the
serum, the half-life of PFOS in the neonates was approximately 14 days, and overall, prenatal
exposure altered gene expression more than lactational exposure.
In a 4-week study in rats, the hepatic effects of PFOS, Wyl4,648 and phenobarbital (PB)
were compared (Elcombe et al., 2012). Groups of 30 male Sprague-Dawley rats were
administered either 20 ppm PFOS, 100 ppm PFOS, 50 ppm Wyl4,648 or 500 ppm PB in the diet
Perfluorooctane sulfonate (PFOS) - February 2014 4-54
Draft - Do Not Cite or Quote
-------
ad libitum for either 1, 7, or 28 days. Control animals received only diet ad libitum for the
duration of the study. Ten animals from each group were sacrificed on days 2, 8, and 29 for
evaluation of liver weights, peroxisome proliferation, enzyme induction, cell proliferation,
apoptosis and other clinical and pathological parameters. The study showed that PFOS exhibits
the combined effects of Wyl4,643 and PB, behaving as a combined peroxisome proliferator and
"phenobarbital-like" enzyme inducer. The data suggested that PFOS may activate not only
PPARa, but also CAR and PXR.
Mice
To assess PPAR involvement in developmental effects of PFOS, male and female
129Sl/Svlm wild type (WT) and PPARa knockout (KO) mice were bred and pregnancy
confirmed (Abbott et al., 2009). The females (n= 8-20 dams/group) were administered either
vehicle (0.5% Tween-20) or PFOS by gavage on GDs 15-18; the WT mice were administered
4.5, 6.5, 8.5 or 10.5 mg/kg/day PFOS and the KO mice, 8.5 or 10.5 mg/kg/day. Dams and pups
were observed daily and pups were weighed on postnatal day (PND) 1 and 15. Eye opening was
recorded on PNDs 12-15. Dams and pups were killed on PND 15 and body and liver weights
were recorded and serum collected.
Maternal body weight, maternal body weight gain, and reproductive parameters measured
included implantation sites, total number of pups at birth, and the percent litter loss from
implantation to birth. Pup body weight and pup body weight gain were not affected with
treatment in either the KO or WT mice. PFOS exposure had no effect on absolute or relative (to
body weight) liver weight in any of the dams. PFOS exposure at 10.5 mg/kg/day caused a
significant increase in relative liver weight (sexes were combined) in both strains of pups.
Survival of the pups was affected with treatment. Most post-natal deaths occurred between
PNDs 1 and 2. Survival of the WT pups was significantly (p< 0.001) decreased and was 65% ±
10 (n=16), 45% ± 14 (n=8), 55% ± 6 (n=7), 43% ± 9 (n=20) and 26% ± 9 (n=17) in the control,
4.5, 6.5, 8.5 and 10.5 mg/kg/day groups, respectively. Survival of the KO pups was significantly
(p< 0.001) decreased and was 84% ± 9 (n=12), 56% ± 12 (n=13) and 62% ± 8 (n=14) in the
control, 8.5 and 10.5 mg/kg/day groups, respectively.
Post-natal development was also affected in the WT and KO pups. On PND 13, open
eyes were reported in 44% of the control pups and none in the 8.5 mg/kg/day WT group. In the
KO mice, open eyes were reported in 23% of the 10.5 mg/kg pups on PND 14 and 59% of the
controls. All serum samples (pups and adults) showed a linear relationship between the amount
of PFOS administered and the amount found in the serum, with levels in treated groups being
significantly increased compared to the controls. As the results from the WT and KO pups were
similar, the author concluded that PFOS-induced neonatal lethality and delayed eye opening
were not dependent on the PPARa activation.
In another developmental study a PFOS solution with 0.5% Tween-20 was administered
to timed-pregnant CD-I mice by oral gavage at 0, 5 or 10 mg/kg/day from GD 1-17 (Rosen et
al., 2009). Five dams per group were euthanized at 'term' and three fetuses per litter were
collected for preparation of total RNA from liver and lung. Additional liver and lungs were
collected from 2 more fetuses/litter for histological examination.
Perfluorooctane sulfonate (PFOS) - February 2014 4-55
Draft - Do Not Cite or Quote
-------
Treatment with PFOS had no effect on body weight, general appearance or litter size.
Hematoxylin and eosin stained sections from treated and control fetal livers showed eosinophilic
granules characteristic of peroxisome proliferation in the PFOS treated dose groups. A
transcriptional response occurred but it was not as prolific as it had been in an earlier PFOA
study. At 5 mg/kg/day, PFOS had 753 fully annotated genes in the fetal liver. PFOS up-
regulated a number of markers for PPARa activity in the fetal liver. In the fetal lungs, regulation
only occurred in a limited group of genes including: Cyp4al4, enoyl-Coenzyme A hydratase
(Ehhadh) and fatty acid binding protein 1 (Fabpl). The pathways or functional groups
significantly enriched by PFOS included: fatty acid metabolism in the fetal liver and lung,
xenobiotic metabolism, peroxisome biogenesis, cholesterol biosynthesis, bile acid biosynthesis
and metabolism of glucose and glycogen. Overall, there was little difference in the
transcriptional changes made by PFOS when compared to the changes activated by PFOA except
for the up-regulation of CypSal 1 and Cyp3a25 occurring in PFOS treated fetal livers.
Taken together, these studies suggest PPARa-independent mechanism for PFOS-induced
neonatal mortality. While transcription changes induced by PFOS in the fetal mouse liver and
lung were related to activation of PPARa and were similar to profiles induced by PFOA (Rosen
et al., 2009), neonatal mortality occurs in PPARa-null mice treated with PFOS but not with
PFOA (Abbott et al., 2009). Thus neonatal toxicity observed with maternal PFOS administration
may not be a result of the transcriptional alterations.
Treatment of mice at a dietary level of 0.05% (500 ppm) PFOS for 5 days caused
increases in relative liver weight and activities of palmitoyl-CoA oxidase, catalase and other
peroxisomal enzymes (Sohlenius et al., 1993). Induction of peroxisome proliferation and
peroxisomal enzyme activities in the liver were also found in rats exposed to 0.02% (200 ppm)
PFOS in the diet for 2 weeks (Ikeda et al., 1987).
Qazi et al. (2009a) tested the effects of 0, 0.005% or 0.02% PFOS on wild-type and
PPARa-null 129/Sv mice. Dietary administration of 0.02% PFOS for 10 days resulted in a
significant increase in liver weight and a reduction in the weight of the spleen in both the wild-
type and null mice; the thymus and epididymal fat pad were both decreased in the wild-type mice
only. The wild-type mice administered 0.02% PFOS in the diet had a pronounced decrease in
the total number of thymocytes and splenocytes as well as a decrease in the size of the ones
present. In the knock-out mice, there was a reduction in the total number of thymocytes and
subpopulations that was partially or almost totally attenuated; effects on splenocytes were mostly
eliminated. There were no effects in the wild-type or knock-out mice administered 0.005%. The
study indicated that the immunomodulation was partially dependent on PPARa.
Changes in gene expression were examined in wild type and PPARa-null mice
administered PFOS by gavage at 0, 3, or 10 mg/kg/day for seven days (Rosen et al., 2010). At
sacrifice liver tissues were processed for histopathology and total RNA; microarray analysis was
conducted using Affymetrix GeneChip 430_2 mouse genome arrays. Liver weight was
increased at 10 mg/kg/day in both wild type and null mice. Overall gene expression showed
dose-related changes in wild type mice while the number of transcripts influenced by PFOS in
null mice was similar between the dose groups. This finding suggests robust PPARa-
independent effects in null mice.
Perfluorooctane sulfonate (PFOS) - February 2014 4-56
Draft - Do Not Cite or Quote
-------
In wild type mice, PFOS altered the expression of PPARa-regulated genes including
those involved in lipid metabolism, peroxisome biogenesis, proteasome activation, and
inflammatory response. Altered PPARa-independent genes included those associated with
xenobiotic metabolism in both wild type and null mice. PFOS caused induction of a constitutive
androstane receptor (CAR) inducible gene, Cyp2blO, indicating the likelihood that PFOS also
activates CAR. In null mice, changes induced by PFOS included up-regulation of genes in the
cholesterol biosynthesis pathway and modest down-regulation of genes associated with oxidative
phosphorylation and ribosome biogenesis (Figure 4-1). Unique in null mice, PFOS up-regulated
Cyp7al, an important gene related to bile acid/cholesterol homeostasis.
Fatty acid metabolism
Proteasome activation
Peroxisome biogenesis
Inflammatory response
Xenobiotic metabolism
Cholesterol biosyn thesis
Oxidative phosphorylation
Ribosome biogenesis
Fold change
2 1.5 1 -1.5 -2
Figure 4-1. Functional categories of genes modified by PFOS in wild type and null mice.
Tan et al. (2012) conducted a dose-response study of hepatic proteonomic responses
following exposure of male Kunming mice (5/dose group) to PFOS at levels of 0.1, 1.5 or 5
mg/kg/day by i.p. injection for 7 days. Twenty-four hours after the last dose, the animals were
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
4-57
-------
sacrificed and the livers harvested, weighed, and preserved in liquid nitrogen. Body weight was
recorded at study initiation and immediately before sacrifice.
Liver tissues were pooled for each dose group and homogenized for proteonomic
analysis. The liver proteins were extracted and grouped using iTRAQ labeling guidelines,
digested with trypsin, and labeled with iTRAQ reagent. The iTRAQ proteonomic analysis is a
novel, MS-based approach for the relative quantification of proteins. It relies on the
derivatization of primary amino groups in intact proteins using isobaric tags for relative and
absolute quantitation (Wiese et al., 2007). The tryptic peptides were separated using reverse
phase liquid chromatography, identified following LC-MS/MS analysis, and correlated to intact
proteins based on peptide structures.
Treatment with PFOS caused a slight deficit in body weight for the high dose group and a
significant dose-related increase in liver weight for the two highest dose groups. The iTRAQ
process identified 1502 unique proteins; 71 showed a greater than 1.5-fold change in expression.
Sixty-two proteins showed increased expression and 9 showed decreased expression. Figure 4-2
illustrates the impact of the PFOS exposure on identified proteins as associated with
subcompartments within the liver cells compared to the proteins in the reference data base.
Enrichment was greatest for peroxisomes and endoplasmic reticulum, mitochondrial, and cell
membrane proteins. Relative to biochemical processes, Figure 4-2 shows that the majority of
enriched proteins were involved with lipid metabolism, transport, biosynthetic processes,
catabolic processes, and carbohydrate metabolic processes.
Perfluorooctane sulfonate (PFOS) - February 2014 4-58
Draft - Do Not Cite or Quote
-------
• F-xjiccI
• llfTO
10 15 X
Number of prolcins
Figure 4-2. Function distribution and category enrichment analysis of the differential
proteins.
Top: cellular component; Bottom: biological process
npro: the number of proteins belongs to one category in the proteome database
Expect: the number of proteins having an ontology annotation in the reference database.
Sixteen proteins were identified that showed dose-response for the increase in expression.
Four of these were related to peroxisomal beta-oxidation, four were related to CYP-450
aromatase activity, and three had transferase activity including GSTmuS and GSTmu6. A GTP
binding protein (GTP sar-lb) also displayed a dose-related response. One of the remaining four
proteins exhibiting dose-response, cysteine sulfmic acid decarboxylase, is the rate limiting
enzyme in taurine production and has been proposed as a biomarker for hepatocarcinogenesis.
In the developmental study by Butenhoff et al. (2009), quantitative reverse transcription
polymerase chain reaction (RT-PCR) to obtain mRNA transcript data for the control and 1.0
mg/kg/day dose groups were recorded on GD 20 in the dams and fetuses and PND 21 in the male
pups. Results for this part of the study were reported by Chang et al. (2009). Statistically
significant changes included: Cyp2b2 levels in dams and male pups were higher (f 2.8-fold and
1.8-fold, respectively) than controls on GD 20 and PND 21; mean acyl CoA (ACoA) and
Cyp4al levels in male pups were greater (| 1.5-fold and 2.1-fold, respectively) than those of
controls and the mean Cyp7al was lower (| 3.5-fold) than those of controls. This suggests
induction of PPARa as well as hepatic constitutive androstane receptor (CAR). Transcripts
possibly related to thyroid status were all similar between the treated dams and pups and the
controls.
Perfluorooctane sulfonate (PFOS) - February 2014 4-59
Draft - Do Not Cite or Quote
-------
Oxidative Damage
Liu et al. (2009) conducted a study in which 3-6 male and female KD mice/group were
administered one subcutaneous injection of 0 or 50 mg PFOS/kg bw on postnatal days (PNDs) 7,
14, 21, 28 or 35. The study was done in an attempt to determine the effects of treatment on the
oxidation-antioxidation system by measuring maleic dialdehyde (MDA) content, superoxide
dismutase (SOD) activity and total antioxidation capability (T-AOC). Animals were killed 24
hours post-treatment, and blood was collected as well as liver and brain removed and weighed.
No treatment-related effects were observed on body weight. Relative (to body weight)
liver weight was significantly increased (p< 0.01 or p<0.05) when compared to controls in both
males and females at most time-points. The levels of MDA in the brain and liver and SOD
activity were similar between treated mice and the controls at most time-points. On PNDs 7 and
21 in the treated males, brain SOD activity was significantly lower when compared to controls
by 19 and 13%, respectively. Liver SOD activity was lower (j 19%) in the treated females on
PND 14 when compared to controls. Male brain T-AOC was decreased at all stages of post-natal
development but only significantly at PND 21 (J, 15%). Male liver T-AOC was decreased
significantly at PND 7 (|25%) and 14 (|27%). Female brain T-AOC had no significant
differences from controls and the liver T-AOC was decreased only at PND 21 (J, 15%). The
study also demonstrated that distribution increased in the liver and lessened in the blood and
brain with post natal development in both the males and females. On PND 7, PFOS
concentrations were 11.78, 5.04 and 14.84% in the male mouse blood, brain and liver,
respectively. On PND 28, the PFOS concentrations were 9.89, 0.85 and 63.39% in the male
mouse blood, brain and liver, respectively. A similar trend was observed in the females. The
study suggested that oxidative damage from PFOS can occur, is more prominent in the younger
neonates and is slightly more pronounced in the males.
Gap Junctional Intercellular Communication (GJIC)
Gap junctions are found in the cell plasma membrane and formed by proteins that
connect and form an intercellular connection that allows a direct exchange of chemicals from the
interior of one cell to that of adjacent cells without passage into the extracellular space. The
GJIC is considered to be essential in maintaining healthy cells and thus disruptions are thought to
cause abnormal cell growth, including tumor formation. They also appear to be linked to some
neurological, reproductive or endocrine abnormalities.
Hu et al. (2002) tested PFOS exposure in vitro and in vivo to determine if disruption to
the GJIC could possibly be a mechanism for the effects observed with PFOS exposure. The
study exposed a rat epithelial cell line (WB-F344), a dolphin kidney epithelial cell line (CDK) as
well as exposing Sprague-Dawley rats orally to PFOS for 3 days and 3 weeks. GJIC effects were
measured using the scrape loading dye technique. The in vitro cell lines were exposed to PFOS
at concentrations of 0, 3.1, 6.25, 12.5, 50, 100 or 160 |iM for 30 minutes. PFOS inhibited GJIC
rapidly in a dose-dependent method starting at 12.5 jiM, but it was reversible once exposure
ended. Four to six rats/sex/group were exposed to 0 or 5 mg/kg PFOS by gavage for either 3 or
21 days. GJIC was significantly reduced in the liver tissue after 3 days of exposure. Inhibition
also occurred in rats exposed for 21 days without an increase in magnitude of the inhibition. No
differences were observed between the male and female rats.
Perfluorooctane sulfonate (PFOS) - February 2014 4-60
Draft - Do Not Cite or Quote
-------
Mitochondrial Function
Starkov and Wallace (2002) isolated mitochondria from the livers of adult male Sprague-
Dawley rats and used them to measure mitochondrial membrane potential and oxygen
consumption when exposed to PFOS. PFOS appeared to be a weak mitochondrial toxicant. At
higher concentrations, PFOS caused a small increase in resting respiration rate and slightly
decreased the membrane potential. The observed effects were attributed to a slight increase in
nonselective permeability of the mitochondria membranes caused by the surface-active property
of the compound.
4.3.4 Structure-Activity Relationship
In vitro
Bjork and Wallace (2009) performed a study to see if the PPARa agonism was relevant
in human cell lines and whether effects differed with various chain lengths. Primary rat and
human hepatocytes and HepG2/C3 A hepatoma cells were exposed to 25 jiM PFAAs for 24 hours
to determine the structure-activity relationships across various chain lengths. The concentration
was the maximum concentration not causing cell injury in any of the cell lines. The PFAAs
tested included perfluorinated carboxylic acids with carbon chain lengths of 2 to 8 and
perfluorinated sulfonic acids with chain lengths of 4 to 8.
The PFAAs stimulated mRNA expression of either acyl CoA oxidase (Acox) or acyl
CoA thioesterase (Cte-rats or Acot 1-humans) only in rat hepatocytes and within both series and
transcripts; the degree of stimulation of gene expression increased with increasing carbon
number. Maximum stimulation of Acox gene expression was 3-fold over control for PFOS;
maximum stimulation for Cte/Acot 1 gene expression was 4-fold for PFOS. PFOS did not cause
any significant stimulation of Acox or Cte/Acot 1 gene expression in human hepatocytes. The
Cyp4al 1 gene was not expressed or stimulated by any of the PFAAs in HepG2/C3A cells.
However, this gene expression was stimulated by PFAA exposure in both rat and human
hepatocytes with the perfluorocarboxylates indicating a chain-length dependent structure activity
relationship. Maximum gene expression stimulation was in the longer carbon PFAAs but the
variability was large with little statistical difference between the 6 and 8 carbon molecules.
Study results suggested that the PPARa related gene expression by PFAAs was induced in
primary rat hepatocytes, increased with carbon chain length and appeared to be greater in the
carboxylic acids (such as PFOA) when compared to the sulfonates (such as PFOS). There also
was not any induction of peroxisome-related fatty acid oxidation gene expression (Acox and
Cte/Acot 1) in either primary or transformed human liver cells in culture. This suggests that the
PPARa mediated peroxisome proliferation observed in rodent liver may not be relevant as an
indicator of human risk.
4.4 Hazard Characterization
Human information from biomonitoring as well as more long-term epidemiology studies
is available, however, the actual dose of exposure is not known. Controlled dosing studies are
available in a variety of species including monkey, rats, rabbits, and mice. The main toxicity
endpoints observed are discussed below for both humans and animals with some possible known
or speculated modes of action provided.
Perfluorooctane sulfonate (PFOS) - February 2014 4-61
Draft - Do Not Cite or Quote
-------
4.4.1 Synthesis and Evaluation of Major Noncancer Effects
Liver Effects, Cholesterol and Uric Acid
Human
Few studies are available where the effects of PFOS in the liver of humans were
examined. Olsen et al. (2003) sampled liver tissue and serum from cadavers and found
correlation between the samples; Karrman et al. (2010) also identified PFOS in human hepatic
tissue in twelve subjects. Biomonitoring studies performed at the 3M Decatur, Alabama plant
(Olsen et al., 1999; Olsen et al., 2001b, 2001c) identified occasional differences in hepatic
clinical chemistry values but there were no reported increases in hepatic disease and or hepatic
carcinogenicity associated with them.
Several studies examining potential associations between PFOS exposures and
cholesterol and other lipid measurements are available. Occupational studies did not reveal
consistent associations between PFOS and cholesterol and triglycerides in either cross-sectional
surveys or in a longitudinal analysis (Olsen et al., 1999; Olsen et al., 2001b, 2001c). However,
community studies and an analysis of NHANES data have shown positive relationships of PFOS
with total cholesterol (Steenland et al., 2009; Nelson et al., 2010; Lin et al., 2009), LDL, and
triglycerides (Steenland et al., 2009). A statistically significant positive association was also
reported in a community study between PFOS and uric acid levels (Steenland et al., 2010).
Elevated uric acid is a risk factor for hypertension and may be an independent risk factor for
stroke, diabetes, and metabolic syndrome.
Animal
PFOS, when absorbed, is primarily found in the liver tissue. In monkeys, rats and mice,
PFOS levels in the liver showed a dose-dependent increase that was consistently greater than
serum levels (Goldenthal et al., 1978a; Seacat et al., 2002; Thomford, 2002; Curran et al., 2008;
Liu et al., 2009). Chang et al. (2009) also identified PFOS levels in the liver of offspring as early
as GD 20 and Stein et al. (2012) measured PFOS in amniotic fluid supporting placental transfer.
In experimental studies, increased absolute liver weight was observed in monkeys
exposed to 0.75 mg/kg/day for 182 days (Seacat et al., 2002), in rats at >1.33 mg/kg/day for 14
weeks (Curran et al., 2008; Seacat et al., 2003), and in rats at >0.77 mg/kg/day for 53 weeks
(Thomford, 2002). As part of a chronic bioassay, rats were administered PFOS in the diet for up
to 104 weeks (Thomford, 2002). Liver weight was increased in males and females at the highest
dietary concentration after both 14 and 53 weeks. Liver weight data were not collected at the
104 week sacrifice.
Histopathological lesions of the liver were observed in rats and monkeys. Lesions were
found in rats at 1.33-1.56 mg/kg/day after 14 weeks (Seacat et al., 2003), in rats at 0.072-0.247
mg/kg/day after 104 weeks (Thomford, 2002), and in monkeys at 0.75 mg/kg/day after 53 weeks
(Seacat et al., 2002). Liver lesions were similar in both species and included centrilobular
hypertrophy and vacuolation after the subchronic and chronic exposures with eosinophilic
granules also observed after chronic duration. In these studies, no evidence of peroxisome
proliferation was found in either species.
Perfluorooctane sulfonate (PFOS) - February 2014 4-62
Draft - Do Not Cite or Quote
-------
Rat and monkey studies demonstrated a decrease in cholesterol levels and high density
lipoprotein cholesterol at 0.75 mg/kg/day when compared to the controls. Serum concentrations
increased during recovery. Male rats had decreased serum cholesterol at 14-weeks at a dose of
about 1.4 mg/kg/day. Increased hepatic lipid content in the absence of a strong PPAR-a
response is a characteristic of exposure to PFOS.
As discussed above in Section 4.3.3.1, mice administered PFOS showed differential
expression of proteins mainly involved in lipid metabolism, transport, biosynthetic processes,
and response to stimulus (Tan et al., 2012) and in genes involved in cholesterol biosynthesis and
xenobiotic metabolism (Rosen et al., 2010). More specific investigations into the genes involved
in lipoprotein metabolism were conducted by Bijand et al. (2011) as described below. In
addition, the nuclear hormone receptors CAR and PXR have been shown to be activated in mice
(Bijand et al., 2011; Rosen et al., 2010) and rats (Elcombe et al., 2012). Taken together, these
studies consistently show an effect on expression of genes involved in lipid metabolism and
cholesterol transport and biosynthesis following in vivo PFOS exposure.
To further examine PFOS-specific effects on lipid metabolism, Bijand et al. (2011)
examined the molecular biology of hepatic hyperlipidemia in APOE*3-Leiden.CETP mice, a
strain that exhibits human-like lipoprotein metabolism. Details of the experimental procedure
were given in Section 4.2.2. Animals fed 3 mg/kg/day for 4 weeks had decreased hepatic VLDL
production leading to increased retention of triglycerides and hepatomegaly, with concomitant
decreased hepatic clearance of VLDL and HDL cholesterol. Fecal bile acid content was
decreased by 50%.
Overall the genes upregulated were those involved with fatty acid uptake, transport and
catabolism; triglyceride synthesis, cholesterol ester synthesis; plus VLDL synthesis and
secretion. Genes involved with HDL synthesis, maturation and clearance plus bile acid
formation were down regulated. Lipoprotein lipase activity and mRNA expression, both
normally low in the liver, were increased.
Many of the genes activated are associated with the nuclear PXR receptor to a greater
extent than PPARa. Lipoprotein lipase activity facilitates removal of TGs from serum LDLs,
and uptake into the liver and other organs as free fatty acids and glycerol.
Developmental/Reproductive Toxicity
Human
Studies evaluating the reproductive and developmental health in humans exposed to
PFOS have been performed in both occupational settings (Grice et al., 2007) and in the general
population (Inoue et al., 2004; Apelberg et al., 2007; Fei et al., 2007, 2008a, 2008b, 2010a and
2010b; Monroy et al., 2008; Washino et al., 2009). No adverse correlations were found in the
occupational workers with birth outcome. In general population studies, the most frequent
associations were those for lower birth weight but this was not a consistent finding in all of the
studies. PFOS has been detected in human amniotic fluid samples indicating that the chemical
crosses the placenta. The median ratio of maternal serum:amniotic fluid concentration was
Perfluorooctane sulfonate (PFOS) - February 2014 4-63
Draft - Do Not Cite or Quote
-------
25.5:1 and PFOS was rarely detected in amniotic fluid until the serum concentration reached at
least 5.5 ng/mL (Stein et al., 2012).
Animal
Increased pup mortality was observed when rat dams were treated only during gestation
as part of developmental toxicity studies (Chen et al., 2012; Thibodeaux, 2003; Lau et al., 2003).
Chen et al. (2012) found increased mortality, decreased body weight and histopathological
changes in the lungs (alveolar hemorrhage, thickened interalveolar septum) in rat offspring from
dams treated with 2.0 mg/kg/day from GD 1 to 21. No effects were observed in those
administered 0.1 mg/kg/day. Developmental delays were found in rat offspring at a lower dose
than that affecting survival (1 mg/kg/day; Butenhoff et al., 2009) and in mice at a slightly higher
dose (5 mg/kg/day; Thibodeaux et al., 2003; Lau et al., 2003).
Rat dams were treated with PFOS for 63 or 84 days in a one- or two-generation
reproductive study, respectively (Luebker et al., 2005a,b). No changes in maternal liver weight
were observed on either protocol. The most sensitive endpoint was decreased pup body weight
with reduced survival also observed at higher maternal doses. A NOAEL for pup body weight
effects was 0.1 mg/kg/day in the two-generation study (Luebker et al., 2005b); this dose was not
tested in the one-generation study (Leubker et al., 2005a) where the LOAEL was 0.8 mg/kg/day
for decreased pup survival, decreased maternal body weight and decreased gestation length. A
0.4 mg/kg/day dose was a NOAEL in the one generation and LOAEL for decreased body weight
gain in the two generation study.
To help characterize the mechanism of PFOS induced neonatal mortality, Grasty et al.
(2003) examined critical windows of exposure by treating rats with a high dose of PFOS (25
mg/kg/day) for a four-day period during various stages of pregnancy. Mortality was highest
when treatment occurred on gestation days (GD) 17-20 identifying late gestation as the sensitive
window for neonatal death. In a subsequent experiment, exposure to 50 mg/kg/day of PFOS on
GD 19 and 20 alone was sufficient to produce almost 100% mortality to pups at birth.
Studies by Grasty et al. (2003; 2005) and Chen et al. (2012) describe significant
histological and morphometric differences in the lungs between control and PFOS-treated
newborn pups, suggesting that lung maturation and pulmonary surfactant interactions are
potential modes of action (MOAs). Changes in lung morphology were noted in rat pups, but
prenatal exposure to PFOS did not affect lung phospholipids or alter the expression of marker
genes for alveolar differentiation associated with lung maturation (Grasty et al., 2005). Chen et
al. (2012) found that PFOS caused oxidative stress and cell apoptosis in the lungs of offspring
from mothers treated with 2.0 mg/kg/day during GDs 1-21.
Currently, the leading hypothesis for the MOA of PFOS-induced neonatal mortality is
that PFOS interacts directly with components of natural lung surfactants (Grasty et al., 2005; Xie
et al., 2010a, b). PFOS interacts with the major phosphatidylcholine components of pulmonary
surfactants and cell membranes and, therefore, has the potential to alter the dynamic properties of
lung surfactant (Xie et al., 2010a). PFOS partitions into phospholipid membranes to increase
membrane fluidity in several cell types (Xie et al., 2010b). This high tendency of PFOS to
partition into phosphatidylcholine lipid bilayers is consistent with its resemblance to medium
Perfluorooctane sulfonate (PFOS) - February 2014 4-64
Draft - Do Not Cite or Quote
-------
chain fatty acids and may be responsible for interfering with the normal physiological function of
pulmonary surfactant.
Immunotoxicity
Human
Limited data, which focus on infants and children, are inconclusive on the potential
immunotoxicity from PFOS exposure to humans. No significant associations were observed
between maternal PFOS levels and cord blood IgE levels or incidence of food allergy, eczema,
wheezing, or otitis media in infants at 18 months of age (Okada et al., 2012). In another study,
maternal serum PFOS concentration, measured at week 32 of pregnancy, was negatively
associated with anitdiphtheria antibody concentration in their children at 5 years of age. The
odds ratio was increased for a 5-year old child to have inadequate antibody concentrations to
diphtheria when compared to both maternal PFOS and the child's age 5 PFOS serum
concentrations. At age 7, lower antibody concentrations to diphtheria and tetanus were
correlated with higher serum PFOS levels at age 5 (Grandjean et al., 2012).
Animal
Peden-Adams et al. (2008) and Dong et al. (2009) both identified immunotoxicity in male
mice following exposure to 0.0017 mg/kg and 0.083 mg/kg, respectively. In the Peden-Adams
study, IgM suppression occurred after 28 days of treatment although there were not any overt
signs of toxicity. Further investigation found that the IgM suppression was observed in both the
T-cell independent and dependent tests making the humoral immune effects caused by B-cells.
In the Dong et al. study, there was an increase in splenic natural killer (NK) cell activity after 60
days, when PFOS serum concentrations were approximately 7.1 mg/L. Gurunge et al. (2009)
found a decrease in survival in mice exposed to 0.025 mg/kg of PFOS after exposure to influenza
A virus. Inbred B6C3Fi male mice offspring had a decrease in NK cell activity starting at 1
mg/kg/day after 8 weeks of treatment.
Qazi et al. (2009a) reported that 0.02% PFOS in the diet for 10 days in wild-type and
PPARa-null 129/Sv knock-out mice caused a pronounced decrease in the total number of
thymocytes and splenocytes as well as a decrease in the size of the ones present in wild-type
mice. Knock-out mice had a reduction in the total number of thymocytes that less than that seen
in the wild-type mice. Effects on splenocytes were mostly eliminated in knock-out mice. The
study thus indicated that the immunomodulation was partially dependent on PPARa.
Mechanisms that could cause these effects other than PPAR- activation are not known but are a
topic of investigation.
Neurotoxicity
Human
No epidemiology studies of neurotoxicity as associated with PFOS exposure were identified.
Perfluorooctane sulfonate (PFOS) - February 2014 4-65
Draft - Do Not Cite or Quote
-------
Animal
In animals effects were observed on excitatory amino acids in the central nervous
systems of rats when administered 25 mg/kg/day of PFOS one time (Yang et al., 2009).
Butenhoff et al. (2009) and Wang et al. (2010) both looked for developmental neurotoxicity
effects in Sprague-Dawley and Wistar rats, respectively. Butenhoff found significant increases
in motor activity of male offspring at one time point (PND 17) and decreased habituation in the
1.0 mg/kg/day males. Wang found that pre-natal exposure to 3.2 mg/kg/day of PFOS in the feed
had some effect on gene expression involved in neuroactive ligand-receptor interaction, calcium
signaling pathways and PPAR signaling. Transthyretin (CSF carrier of T4) was also decreased.
Zeng et al. (2011) also found PFOS administered to pregnant rats as low as 0.1 mg/kg from
gestation day 2 to 21 caused significant increases of PFOS in the brain (hippocampus and cortex)
of the offspring with effects on inflammatory markers and transcription factors. Two-month old
mice exposed to 0.75 mg/kg PFOS (Johansson et al., 2008) displayed abnormal habituation
responses in motor activity testing.
Thyroid Effects
Human
Several epidemiological studies included the evaluations of circulating thyroid hormones.
Olsen et al. (2003) reported no significant associations of TSH, TT4 and fT4 with serum PFOS
among production workers although a positive association with T3 was apparent. PFOS was
detected in maternal and cord blood samples in a susceptible population in Japan with no
significant correlations between PFOS concentration and TSH or fT4 (Inoue et al., 2004). Pirali
et al. (2009) found no relationship between intrathyroidal concentrations of PFOS and
underlying thyroid diseases in patients. Dallaire et al. (2009b) examined a population of
Canadian Inuit adults and noted that PFOS at mean plasma levels of 18.3 ng/mL was negatively
associated with TSH, T3 or thyroid binding protein (TBG), but positively associated with fT4.
The findings of Bloom et al. (2010) in a non-occupational cohort in the U.S. with mean serum
PFOS concentrations of 19.6 ng/mL were similar.
Melzer et al. (2010) examined NHANES records for thyroid disease and serum PFOS
levels in the U.S. general population. The only significant association was for men with thyroid
disease with PFOS levels > 36.8 ng/mL compared to men with PFOS levels < 25.5 ng/mL
Animal
Several animal models have described changes in thyroid hormone levels after
administration of PFOS. In Cynomolgus monkeys treated with 0.03, 0.15 or 0.75 mg/kg/day of
PFOS for 26 weeks, Seacat et al. (2002) saw significant reductions of total triiodothyronine (T3)
(-50%), and a less consistent effect in total thyroxine (TT4, females only). This was more
pronounced at the end of exposure period in the high-dose group but a dose-response was not
observed and no evidence of hypothyroidism was seen. Thyroid-stimulating hormone (TSH)
levels were variable during the study, but increased two-fold in the high-dose group at the end of
exposure. PFOS-induced alterations of thyroid hormones were also seen in the adult rat models
(Thibodeaux et al., 2003; Martin et al., 2007; Yu et al., 2009b; Yu et al., 2011); however, in
Perfluorooctane sulfonate (PFOS) - February 2014 4-66
Draft - Do Not Cite or Quote
-------
contrast to the monkey model, most reductions involved circulating TT4, instead of T3. In all
animal studies, the changes in T3 and rT4 failed to activate the hypothalamic-pituitary-thyroid
(HPT) feedback mechanism to produce significant elevations of serum TSH.
Typically in pregnancy, serum rT4 will decrease by 70% while TSH will increase 3-fold.
Exposure of pregnant rats to PFOS exacerbated both of these hormonal reactions without further
elevating the levels of TSH (Thibodeaux et al., 2003). The effective dose of PFOS on TT4 was 1
mg/kg/day, which corresponded to maternal serum concentrations of 14-26 ug/mL. A similar
effect of PFOS on serum TT4 was also seen in the pregnant mouse model, although this rodent
species appears to be much less sensitive than the rat, with significant changes noted only at the
high dose (20 mg/kg/day, corresponding to 114-261 ug/mL) (Thibodeaux et al., 2003).
In utero exposure to PFOS led to postnatal mortality in rat neonates, in a dose-dependent
fashion (Lau et al., 2003). Among the surviving pups, the ontogenetic increases of serum TT4
during the first two weeks of life were delayed or attenuated. No significant changes were noted
in the ontogenetic rises of T3 or TSH. Luebker et al. (2005a) observed significant dose-related
reductions of TT4 (and to a lesser extent, T3) on postnatal day 5 when >0.4 mg/kg/day of PFOS
was administered in the diet during gestation. No effect was observed on serum TSH.
Histological and morphometric evaluations of the fetal and neonatal thyroid glands indicated
normal number and size distribution of follicles, and normal follicular epithelial cell heights and
colloid areas, despite the PFOS-induced TT4 deficits (Chang et al., 2009).
In addition to the evaluation of PFOS's effects on serum TT4, several studies have
examined the levels of circulating free T4 (fT4) (Thibodeaux et al., 2003; Lau et al., 2003,
Luebker et al., 2005a; Yu et al., 2011). In these studies, fT4 was reduced after PFOS
administration when measured by analog radioimmunoassays (RIA). However, when the fT4
was measured by an equilibrium dialysis step prior to the standard RIA (ED-RIA), fT4 levels in
the PFOS-treated rats were comparable to those of controls (Luebker et al., 2005a).
Mechanisms underlying the PFOS-induced hypothyroxinemia are still under active
investigation, but do not likely involve altered de novo biosynthesis of the hormones or
compromised integrity of the HPT axis. Yu et al. (2009b) reported no significant effects of
PFOS on the sodium iodide symporter gene expression (for iodide uptake) or thyroid peroxidase
activity in the thyroid gland. Chang et al. (2008) showed that release of TSH from the pituitary
in response to ex vivo TRH stimulation was not altered by PFOS exposure.
Weiss et al. (2009) demonstrated that perfluorinated chemicals (including PFOS) are
capable of competing with T4 and displacing the hormone from binding to the human thyroid
hormone transport protein transthyretin (TTR). In fact, PFOS ranks the second highest in
binding potency among all perfluorinated compounds examined, although its TTR binding
potency is only one-fifteenth of that for T4. Hence, a plausible scenario can be constructed to
account for the hypothyroxinemic responses in the PFOS-treated animals. PFOS in circulation
competes with T4 and displaces the hormone from binding to TTR (the primary thyroid hormone
transport protein in the rat), initially leading to a transient elevation of fT4 (within 6 h), and a
brief compensatory decrease of TSH. Concomitantly, hepatic metabolism of the hormone by
UGT1A is enhanced (presumably in response to the transient elevation of the free hormone),
which results in an increase of hormonal clearance and urinary excretion of iodide. As the fT4
Perfluorooctane sulfonate (PFOS) - February 2014 4-67
Draft - Do Not Cite or Quote
-------
level returns to normal range subsequently (within 24 h), a new equilibrium is reached between
normal complements of fT4 and TSH, but a net reduction of total T4 (resulted from protein
binding displacement and metabolism).
A lack of significant change in TSH receptor gene expression in the thyroid gland is also
consistent with the transient nature of TSH depression (Yu et al., 2009b). Thus, so long as fT4
levels are maintained, TSH levels will remain within the normal range, despite the repeated
displacement of T4 from TTR by PFOS, which results in net loss of TT4. Because the extent of
T3 binding to TTR is less than that of T4, the deficits of serum T3 caused by PFOS exposure are
smaller than those of T4. Maintenance of fT4 levels despite the PFOS-induced deficits of TT4 is
indirectly supported by a general lack of thyroid hormone-specific responses in the rat (Lau et
al., 2003; Chang et al. 2008), suggesting that the functional thyroid status has not been
compromised significantly by subchronic exposure to the chemical.
Several possibilities may account for the differential findings of thyroid hormone
disruption between animal models and human biomonitoring data. First, hypothyroxinemia was
observed in adult monkeys and rodents only when serum PFOS reached the 70-90 ug/mL (ppm)
range. Pregnant rats and neonatal rats appeared to be more sensitive, exhibiting TT4 depression
when serum PFOS reached about 20 and 40 ug/mL, respectively. However, serum PFOS in the
general populations is estimated to be 15-30 ng/mL (ppb), about three orders of magnitude lower
than the effective body burden for thyroid hormone disruption in animal models. Secondly, TBG
(rather than TTR as in rodents) is the major thyroid hormone transporter in humans. Although
PFOS can bind to human TTR and effectively displaces T4 as illustrated in the rat model, its
binding affinity to TBG is unknown and may be lower than that to TTR (suggested by the PBDE
findings reported by Cao et al., 2010), thus leading to a weaker thyroidal effect of the chemical
in humans. Weiss et al. (2009) have shown that most of these compounds are capable of binding
to human TTR, and therefore effective in displacing T4 from its transporter proteins. Indeed,
these investigators used a T4-EQ approach to estimate the combined thyroidal effects of multiple
perfluorinated compounds, and suggested a margin of safety of 503 for European adults and 306
for North American adults.
Synthesis and Evaluation of Carcinogenic Effects
4.4.2 Synthesis and Evaluation of Carcinogenic Effects
A positive association between PFOS exposure and the incidence of cancer was not
identified in occupational studies (Alexander et al., 2003; Alexander and Olsen, 2007; Grice et
al., 2007), and a study of the general population (Eriksen et al., 2009). The only chronic
toxicity/carcinogenicity study in animals was a rat study (Thomford, 2002). This study was
'suggestive' of carcinogenicity. Increased incidence of hepatocellular adenomas in the male and
female rats and combined adenomas/carcinomas in the females were observed, but did not
display a clear dose-related response. All genotoxicity studies including an Ames test,
mammalian-microsome reverse mutation assay, an in vitro assay for chromosomal aberrations,
an unscheduled DNA synthesis assay, and mouse micronucleus assay were negative.
Epidemiological studies in occupational and general populations did not support any increases in
the incidence of carcinogenicity with exposure to PFOS.
Perfluorooctane sulfonate (PFOS) - February 2014 4-68
Draft - Do Not Cite or Quote
-------
4.4.3 Mode of Action and Implications in Cancer Assessment
Short-term genotoxicity assays suggested that PFOS is not a DNA-reactive compound
Therefore, the induction of tumors by PFOS is probably due to non-genotoxic mechanisms.
Induction of peroxisome proliferation has been suggested as the mode of action for an increasing
number of non-genotoxic carcinogens that induce liver tumors upon chronic administration to
rats and/or mice (Rao and Reddy, 1996; Ashby et al., 1994). The liver-expressed peroxisome
PPARa regulates the transcription of genes involved in peroxisome proliferation, cell cycle
control, apoptosis, and lipid metabolism. Mode of action (MOA) analysis suggested that the
liver tumors induced by PFOS in rats are not clearly related to PPARa activation.
The mode of action (MOA) for PPARa-agonist induced liver tumors has been
hypothesized to involve four key causal key events:
• PPARa activation
• Cell proliferation/decreased apoptosis
• Preneoplastic foci
• Clonal Expansion of foci
• Liver tumors
The data for PFOS are adequate to support the some but not all of the key events in the PPARa-
agonist induced tumorigenic MOA.
Peroxisome proliferation can lead to oxidative stress and may contribute to the mode of
action by causing indirect DNA damage and leading to mutations, or by stimulating cell
proliferation. Information that would help establish that a chemical is inducing liver tumors via a
PPARa agonist MOA includes in vitro evidence of PPARa agonism (i.e., evidence from an in
vitro receptor assay), in vivo evidence of an increase in number and size of peroxisomes,
increases in the activity of acyl CoA oxidase, and hepatic cell proliferation.
Treatment of rats (Ikeda et al., 1987; Thomford, 2002), and mice (Sohlenius et al., 1993)
caused increases in relative liver weight and increased activities of peroxisomal enzymes.
Several studies demonstrated that PFOS activates mouse and human PPARa in a luciferase
reporter ^ram'-activation assay in COS-1 or 3T3-L1 cells and in a rat liver cell model where the
induction of endogenous PPARa target genes was monitored (Shipley et al., 2004; Takacs and
Abbott, 2007). Studies by Martin et al (2007) and Rosen et al (2010) And Tan et al (2012)
demonstrated that PFOS exposure is associated with gene expression patterns and induction of
proteins associated with PPARa but also to CAR and other cellular receptors. Maloney and
Waxman (1999) found that PPARa is activated by endogenous cellular fatty acids which
suggests that displacement of endogenous ligands from liver fatty acid binding protein may be
one mechanism by which PFOS induces peroxisome proliferation.
Although there are some data supporting the ability of PFOS to activate PPARa, data are
generally lacking for increased cell proliferation, another of the key events in a peroxisome
proliferator-induced hepatocarcinogenesis. No increase in cell proliferation in the liver was
detected in the subchronic study (Seacat et al., 2003) or the cancer bioassay (Thomford, 2002) of
PFOS. No studies were identified wherein evidence of preneoplastic foci were observed in the
liver. Liu et al. (2009) studied biomarkers for oxidative stress in the liver and brain in KD mice.
Perfluorooctane sulfonate (PFOS) - February 2014 4-69
Draft - Do Not Cite or Quote
-------
Levels of MDA did not differ between controls and exposed animals; SOD activity was lower
than that observed in the controls.
Other possible MO As for carcinogenicity have been explored including mitochondrial
biogenetics and gap junctional intercellular communication (GJIC). While PFOS was shown to be a
weak toxicant to isolated mitochondria (Starkov and Wallace, 2002), it inhibited GJIC in a dose-
dependent manner in two cell lines and in liver tissue from rats exposed orally (Hu et al., 2002).
However, these are not clearly defined MO As and their importance relative to PFOS exposure is not
certain. The results from genotoxicity studies conducted with PFOS were negative.
PFOS was tested for its ability to induce mutation/genotoxicity in a number of in vitro
and in vivo assay systems and it was neither mutagenic or genotoxic when tested under these
conditions (see Section on Mutagenicity and Genotoxicity).
4.4.4 Weight of Evidence Evaluation for Carcinogenicity
Based on the available evidence, the data are inadequate to support a PPARa-linked
MOA for the liver and thyroid adenoma's observed by Thomford (2002) in the chronic two year
bioassay in Crl:CD(SD)IGS BR rats. Although liver adenomas were significantly increased in
males and females at the highest dose, a dose-response pattern was not observed although the test
for trend was positive (P=0.03). In males the incidence of thyroid tumors was elevated only in
the high-dose, recovery group males exposed for 52 weeks, where it was about 3 times greater
than the incidence in rats given the same dose for 104 weeks. As was the case for the liver
tumors, the thyroid adenoma data did not show a direct response to dose. The classification for
of PFOS under the U.S. EPA Guidelines for Carcinogen Risk Assessment (U.S. EPA, 2005a) is
currently consistent with the suggestive evidence of carcinogenic potential descriptor.
4.4.5 Potentially Sensitive Populations
In humans, single blood samplings of different populations within the United States do
not support major gender differences. Gender differences could not be determined by those
exposed by occupational exposure as the majority of those tested were males. Evidence from
animal studies does not suggest major differences between genders in the amount of PFOS
identified in the serum and liver tissue of animals or in the toxicity. In the monkey studies and
most developmental rat studies, there do not appear to be any differences between the males and
females after administration of PFOS. However, in the chronic/carcinogenicity study in rats, the
male rats do appear to be slightly more sensitive to liver toxicity. Also males had lower NOAEL
and LOAEL values in the animal studies involved in determining immunology function.
Animal studies clearly show that developmental exposure of rats or mice to PFOS
administered during gestation results in rapid, dose-dependent effects on neonatal survival (Lau
et al., 2003; Luebker et al., 2005b). Additional long term effects on postnatal growth and delays
in developmental landmarks (eye opening, pinna unfolding, surface righting, air righting) occur
in surviving rat pups. The mechanistic cause of this developmental toxicity is unknown, but
investigations of several potential modes of action are summarized here. Generally, there is a
lack of consistency among the epidemiological studies regarding potential associations between
PFOS levels during pregnancy and developmental birth outcomes. Some studies indicate a
potential impact on birth weight, but this finding is not consistent across studies.
Perfluorooctane sulfonate (PFOS) - February 2014 4-70
Draft - Do Not Cite or Quote
-------
To help characterize the mechanism of PFOS induced neonatal mortality, Grasty et al.
(2003) examined critical windows of exposure by treating rats with a high dose of PFOS (25
mg/kg/day) for a four-day period during various stages of pregnancy. Neonatal mortality
occurred after all treatment periods, but the incidence of neonatal death increased when exposure
occurred later in gestation. Mortality was highest when treatment occurred on gestation days
(GD) 17-20 identifying late gestation as the sensitive window for neonatal death.
The evidence for reproductive and developmental effects of PFOS in humans is unclear.
The strength of many of the studies is reliance on direct biomonitoring data to assess exposure,
therefore minimizing misclassification (Olsen et al., 2009). Birth weight was measured by
medical records, birth certificates, or maternal recall within several years of birth leading to good
accuracy and reliability of the data, but other birth outcome measures may be subject to a greater
degree of error. A number of methodological differences across the investigations may affect the
comparability of the risk estimates and could account for the inconsistencies between the studies.
Differences in the reported data make comparisons difficult because some reported log-
transformed data and others did not (Olsen et al., 2009). No two studies reported statistically
significant correlations for the same birth outcome and associations were sometimes specific for
birth gender or parity. The subjects of the studies differed substantially in nationality, ethnicity,
and pre-pregnancy BMI. Several investigators attempted to correct for potential confounders for
birth weight. Parity was found to be a significant confounder. Low response rates may also be a
factor affecting the representativeness of the population sample.
Perfluorooctane sulfonate (PFOS) - February 2014 4-71
Draft - Do Not Cite or Quote
-------
5.0 DOSE-RESPONSE ASSESSMENT
5.1 Dose-Response for Noncancer Effects
A Reference Dose (RfD) or Reference Concentration (RfC) is used as a benchmark for
the prevention of long-term toxic effects other than carcinogen!city. RfD/RfC determination
assumes that thresholds exist for toxic effects, such as cellular necrosis, significant body or organ
weight changes, blood disorders, etc. The RfD is expressed in terms of milligrams per kilogram
per day (mg/kg-day) and the RfC is expressed in milligrams per cubic meter (mg/m3). The RfD
and RfC are estimates (with uncertainties spanning perhaps an order of magnitude) of the daily
exposure to the human population (including sensitive subgroups) that is likely to be without an
appreciable risk of deleterious effects during a lifetime.
5.1.1 RfD Determination
While the database and number of studies is robust for PFOS, quantifying the dose
response is challenging. PFOS is a unique chemical as the toxicokinetics are still not fully
understood, the half-life differences between species are great (>100X difference between rats
and humans), true exposure data in humans are lacking, and a difference in effects has been
observed between animal studies and human epidemiology studies.
Human Data. In humans, data have been obtained for occupational and general
population exposure scenarios. Some studies have monitored similar populations over time to
determine the trend observed. In general, PFOS levels in the serum of the general population
have decreased since production was stopped in the United States. Some limitations associated
with the epidemiology studies include the small population of persons evaluated, the lack of
control over other factors that may be contributing to the effects observed, and concurrent
exposures to other perfluorinated chemicals which were measured in serum. In most cases, the
findings are suggestive and not conclusive of an effect. Pathways of exposure in the general
population appear to be from drinking water, food (especially fish/seafood), and some
environmental exposures (i.e. carpets, house dust).
Some human epidemiology studies found an association with increased PFOS serum
levels and an increase in total cholesterol in adults (Chateau-Degat et al., 2010; Nelson et al.,
2010; Steenland et al., 2009) as well as children (Frisbee et al., 2010). Steenland et al. (2009)
also found a correlation between increased PFOS and higher triglycerides. Occupational studies
did not indicate consistent associations between PFOS and cholesterol and/or triglycerides in
either cross-sectional surveys or in a longitudinal analysis (Olsen et al., 1999; Olsen et al.,
2001b,2001c). A statistically significant positive association was also reported between PFOS
and uric acid levels in a community study (Steenland et al., 2010). A number of the studies in
the general population were associated with the C8 project in which the mean serum level of
PFOS was approximately 22.4 ng/mL or derived from NHANES data in which the mean serum
level was approximately 25.3 |ig/L (25.3 ng/mL).
The relationship between PFOS and an increase in thyroid hormones was examined in
human populations (Bloom et al., 2010; Dallaire et al., 2009a; Chateau-Degat et al., 2010;
Meltzer et al., 2010) with inconsistent results. While a significant increase in free T4 was
observed in subjects with higher PFOS levels, thyroid function appeared to be either normal or
Perfluorooctane sulfonate (PFOS) - February 2014 5-1
Draft - Do Not Cite or Quote
-------
not affected. The mean serum concentration observed in these populations was approximately
20 ng/mL. Limited data, which focus on infants and children, are inconclusive with regard to the
potential immunotoxicity from PFOS exposure (Okada et al., 2012; Grandjean et al., 2012; Dong
etal., 2013).
Studies of the impact of PFOS on reproductive and developmental health have been
conducted in both occupational settings (Grice et al., 2007) and for the general population (Inoue
et al., 2004; Apelberg et al., 2007; Fei et al., 2007, 2008a, 2008b, 2010a and 2010b; Monroy et
al., 2008; Washino et al., 2009). The researchers focused on endpoints of birth weight and other
measures of fetal growth. No significant effects were found in the occupational workers relative
to birth outcome. In general population studies, the only finding of note was a slight increase in
the risk for low birth weight, however, this was not a consistent finding across the studies. Mean
serum concentrations in pregnant females participating in these studies ranged from 14 to 30
ng/mL.
Animal Data. Adequate studies were available for short-term, subchronic, chronic,
developmental and reproductive parameters in rats, mice, rabbits and primates. Subchronic,
chronic, and reproductive toxicity animal studies, all with exposure duration greater than 60
days, have been summarized in Table 5-1. Shorter duration studies which focused on
immunotoxicity endpoints and developmental toxicity studies are summarized in Table 5-2.
Perfluorooctane sulfonate (PFOS) - February 2014 5-2
Draft - Do Not Cite or Quote
-------
TABLE 5-1. NOAEL/LOAEL and Effects for Longer-term Duration Studies of PFOS
Species
monkey
monkey
rat
rat
rat
rat
rat
rat
mouse
Study Duration
90 days
182 days (6 months)
90 days
98 days (14 weeks)
2 generation
(84 days; 12 weeks)
1 generation
(females only)
63 days
1 generation
(females only)
63 days
728 days
(104 weeks; 2 yrs)
60 days
NOAEL
(mg/kg/day)
ND
0.15
ND
0.40 (f)
0.34 (m)
0.1
0.4
ND
0.099 (f)
0.018 (m)
0.008
LOAEL
(mg/kg/day)
0.5
0.75
2.0
1.56 (f)
1.33 (m)
0.4
0.8
0.4
0.247 (f)
0.072 (m)
0.083
Critical Effect(s)
diarrhea, anorexia
I cholesterol
I body wt gain
t liver wt; histopath.
t liver wt
I food consumption
t liver wt
J, cholesterol (m)
t ALT (m) and BUN
(m/f)
t liver hypertrophy
J, food consumption
J, adult body wt gain
1 pup body wt
1 maternal wt gain
I gestation length
I pup survival
I pup body weight
liver
histopathological
changes
t liver wt
t splenic natural
killer cell activity
Reference
Goldenthal et
al, 1979
Seacat et al.,
2002
Goldenthal et
al., 1978b
Seacat et al.,
2003
Luebker et al.,
2005b
Luebker et al.,
2005a
Luebker et al.,
2005a
Thomford,
2002
Dong et al.,
2009
ND= not determined
BUN = blood urea nitrogen
Seacat et al. (2002) treated monkeys for up to 6 months and found increased liver weight
and centrilobular or diffuse hepatocellular hypertrophy at 0.75 mg/kg/day, but no clear evidence
of peroxisomal or cell proliferation. Hepatic peroxisome proliferation, measured by palmitoyl
CoA oxidase activity, was increased significantly in the females at 0.75 mg/kg/day; however, the
magnitude was less than the two-fold increase typically indicating biological significance. There
were no treatment-related effects on cell proliferation in the liver, pancreas or testes when
analyzed by proliferating cell nuclear antigen immunohistochemistry cell labeling index. At the
highest dose, 0.75 mg/kg/day, monkeys also had decreased cholesterol and 2/6 males died. At
the concentration with no effects observed (0.15 mg/kg/day), the serum concentration was 83
Hg/mL in males and 67 |ig/mL in females. At the dose where effects were present (0.75
mg/kg/day), the serum concentrations were 173 |ig/mL in males and 171 |ig/mL in females. In
the 3-month study with monkeys, Goldenthal et al. (1979) found anorexia and clinical signs, but
no changes in liver weight at 0.5 mg/kg/day.
As part of a chronic bioassay, rats were administered PFOS in the diet with an interim
sacrifice after 14 weeks (Seacat et al., 2003) or continued for up to 104 weeks (Thomford, 2002).
Liver weight was increased in males and females at the highest dietary concentration after both
14 and 53 weeks, but a dose- and time-response could not be evaluated because data for the
lower dose groups were not reported at week 53 and no liver weight data were reported for any
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-3
-------
group at study termination after 104 weeks. An increase in liver weight was noted in the
subchronic study by Goldenthal et al. (1978b) at a slightly higher dose (2 mg/kg/day) than that
seen in Seacat et al. (2003) but a NOAEL was not determined in the older study. In mice, Dong
et al. (2009) found a significant increase in liver weight after a 60-day exposure to a dose of
0.083 mg/kg/day.
Histopathological lesions of the liver were observed at doses of 1.33 mg/kg/day in males
and 1.56 mg/kg/day in females after 14 weeks (Seacat et al., 2003) and at 0.072 mg/kg/day in
males and 0.247 mg/kg/day in females after 104 weeks (Thomford, 2002). Liver lesions
included centrilobular hypertrophy and vacuolation after the subchronic and chronic exposures
with eosinophilic granules also observed after chronic duration. No evidence of peroxisome
proliferation was found during either phase of the study. No effect levels in males and females
were 0.34 mg/kg/day and 0.40 mg/kg/day, respectively, after 14 weeks and 0.018 mg/kg/day and
0.099 mg/kg/day, respectively, after 104 weeks.
Rat dams were treated with PFOS for 63 or 84 days in a one- or two-generation
reproductive study, respectively (Luebker et al., 2005a,b). No changes in maternal liver weight
were observed with either protocol. The most sensitive endpoint was decreased pup body weight
at 0.4 mg/kg/day in both the one- and two-generation studies with reduced pup survival observed
in the one generation study at higher maternal doses of >0.8 mg/kg/day. A NOAEL for pup
body weight effects was 0.1 mg/kg/day in the two-generation study; the one-generation study
(Leubeker et al., 2005a) lacked a NOAEL for decreased pup body weight because it was
impacted at the lowest dose tested (0.4 mg/kg/day). Offspring survival was affected in a dose-
related manner in the one-generation study with a biologically significant decrease attained at 0.8
mg/kg/day and statistical significance reached at 1.6 mg/kg/day. In the two generation study
(Luebeker et al 2005b), FI offspring viability was markedly impacted at a dose of 1.6 mg/kg/day
resulting in discontinuation of that dose for production of the F2 generation. The highest dose
used during production of the F2 generation was 0.4 mg/kg/day which was a NOAEL for effects
on survival. Thus, the most sensitive endpoint was decreased offspring body weight which
occurred at a lower dose than that resulting in reduced pup survival and was similar to the dose
causing maternal toxicity.
Some effects on thyroid-related parameters were noted in animals, but there did not
appear to be any increase in hypothyroid or hyperthyroid disorders. In the Seacat et al. monkey
study (2002), a significant reduction of total triiodothyronine (T3) and increased TSH were
observed that were more pronounced at the end of exposure period in the high-dose group.
However, a dose-response was not observed and no evidence of hypothyroidism was seen.
PFOS-induced alterations of thyroid hormones were also seen studies on adult rats (Thibodeaux
et al., 2003; Martin et al., 2007; Yu et al., 2009b; Yu et al., 2011); however, most reductions
involved circulating TT4, instead of T3. In two studies (Martin et al., 2007; Yu et al., 2009b)
when PFOS serum levels were at 88 ug/mL, TT4 levels were reduced by 75-79%, suggesting
effects of PFOS on serum TT4 are directly related to endogenous concentrations of the chemical.
In all animal studies, however, the changes in T3 and TT4 failed to activate the hypothalamic-
pituitary-thyroid (HPT) feedback mechanism to produce significant elevations of serum TSH.
Across the range of longer-term studies the lowest LOAEL is 0.072 mg/kg/day for
histopathological changes in the liver of male Sprague-Dawley rats following a 104-week (2-
Perfluorooctane sulfonate (PFOS) - February 2014 5-4
Draft - Do Not Cite or Quote
-------
year) exposure (Thomford, 2002). Histological changes observed included centrilobular
hypertrophy and centrilobular vacuolization in the hepatocytes, likely associated with PFOS or
lipid accumulation. Significant increases in absolute and relative liver weights were not noted.
The LOAEL for comparable effects in females was about 3 times higher. Increases in liver
weight were observed in male and female rats with shorter durations of exposure. After 14
weeks, Seacat et al. (2003) reported increased absolute and relative liver weights in male and
absolute liver weight in female Sprague-Dawley rats, accompanied by centrilobular hypertrophy
and decreased cholesterol levels at a dose of 1.33 mg/kg/day for the males and 1.56 mg/kg/day
for the females. Goldenthal et al. (1978b) identified 2 mg/kg/day as a LOAEL for increased
absolute and relative liver weights in rats treated for 90 days. In monkeys, increased relative
liver weight and decreased cholesterol were seen at a LOAEL of 0.75 mg/kg/day administered
for six months (Seacat et al., 2002).
In the Dong et al. (2009) study, an increase in splenic natural killer (NK) cell activity and
increased liver weight were seen in male mice after 60 days of treatment with 0.083 mg/kg/day;
resulting PFOS serum concentrations were approximately 7.1 mg/L. No other studies of an
immunological endpoint with a comparable exposure duration were identified.
The most severe of the effect observed in the longer-term studies was the decrease pup
survival in the one-generation study by Luebker et al. (2005a) in SD rats at a LOAEL of 0.8
mg/kg/day. The LOAEL for the less serious effect of decreased pup body weight was 0.4
mg/kg/day in one- and two-generation studies. The short-term studies compiled in Table 5-2
below support the concern for low dose-effects on pup survival. The NOAEL for liver effects in
male rats (0.072 mg/kg/day) appears to be protective for other effects from PFOS exposures.
TABLE 5-2. NOAEL/LOAEL Data for Short-term Oral Studies of PFOS
Species
Rat
Rat
Rat
Rat
Mouse
Mouse
Study Duration
28 days
GD 1-21
GD 2-20
GD 1-2 1+20 days
postnatal
GD 1-17
28 days
NOAEL
(mg/kg/day)
ND(f)
0.14(m)
0.1
1.0
0.3
1.0
0.00017 (m)
0.0033 (f)
LOAEL
(mg/kg/day)
0.15 (f)
1.33 (m)
2.0
2.0
1.0
5.0
0.0017 (m)
0.017 (f)
Critical Effect(s)
t relative liver wt
t mortality;
histopathological
changes to lungs
(pups)
J, dam and pup bwt
J, pup survival
t motor activity in
male pups PND 17
t liver wt, dams and
pups; delayed eye
opening
| SRBC plaque-
forming cell response
Reference
Curranet al.,
2008
Chenetal.,
2012
Thibodeaux et
al., 2003 ;Lau
etal.,2003
Butenhoff et
al., 2009
Thibodeaux et
al., 2003 ; Lau
etal.,2003
Peden-Adams
etal.,2008
Similar to the decreased offspring survival described in the reproductive toxicity studies
(Luebker et al., 2005a), increased pup mortality was observed when rat dams were treated only
during gestation as part of developmental toxicity studies (Chen et al., 2012; Thibodeaux et al.,
2003; Lau et al., 2003). Chen et al. (2012) found increased mortality, decreased body weight and
histopathological changes in the lungs (alveolar hemorrhage, thickened interalveolar septum) in
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-5
-------
rat offspring from dams treated with 2.0 mg/kg/day from GD 1 to 21. No effects were observed
in those administered 0.1 mg/kg/day. Thibodeaux et al. (2003) and Lau et al. (2003) both found
decreased maternal and pup weight gain, but no effects on maternal liver weight, when dams
were dosed at 2 mg/kg/day from GD 2 to 20.
Developmental neurotoxicity was found in rat offspring at a lower dose than that
affecting survival (1 mg/kg/day; Butenhoff et al., 2009) and developmental delays were observed
in mice at a slightly higher dose (5 mg/kg/day; Thibodeaux et al., 2003; Lau et al., 2003). In the
standard developmental neurotoxicity study by Butenhoff et al. (2009), male offspring showed
increased motor activity and decreased habituation on PND 17 following a maternal dose of 1
mg/kg/day; no effects on body weight were reported. Evaluating postnatal effects of in utero
exposure in the mouse, Lau et al. (2003) reported increased liver weight and delayed eye opening
in offspring from dams treated with 5 mg/kg/day.
Several studies (Lau et al., 2003; Thibodeaux et al., 2003; Curran et al., 2008) found
increased liver weight generally concurrent with other endpoints. Rats treated for 28 days
showed dose-related decreased body weight and increased liver weight with statistical
significance beginning at 1.33 mg/kg/day for males and 0.15 mg/kg/day for females (Curran et
al., 2008). Treatment of mice during gestation resulted in increased maternal liver weight and
developmental delays in the offspring at 5 mg/kg/day (Thibodeaux et al., 2003; Lau et al., 2003).
Peden-Adams et al. (2008) identified immunotoxicity in male mice exposed to 0.0017
mg/kg/day. IgM suppression occurred after 28 days of treatment although no overt signs of
toxicity or effects on liver weight were observed at any dose. The only effect at a LOAEL less
than that from the Thomford (2002) chronic study was a decrease (52 to 78%) in the SRBC
plaque-forming cell response in male mice in the study by Peden-Adams et al. (2008) with a
LOAEL of 0.0017 mg/kg/day and an NOAEL of 0.00017 mg/kg/day. The number of animals per
dose groups in this study was small (n=5) suggesting the need for additional research to confirm
the NOAEL and LOAEL for this endpoint.
In the process of RfD development, all relevant endpoints must be considered within the
context of the database as a whole. As part of that analysis, data from the mouse do not appear
to be the best choice on which to base the risk assessment of PFOS. The in vitro measures of
immunocompetence on mice may not be relevant to the human experience and limited human
data from epidemiology studies are inconclusive regarding the immunotoxicity of PFOS in
humans. Values associated with immunotoxicity endpoints in mice were markedly lower, by
several orders of magnitude, than those from developmental and liver endpoints in monkeys and
rats (Tables 5-1 and 5-2), but are not supported by other immunotoxicity studies in the database
(Keil et al., 2008; Zhang et al., 2009; Dong et al., 2009; Qazi et al., 2010). Developmental
delays occurred in mice at higher doses than those affecting development and survival in rats.
Thus, the mouse does not appear to be the most sensitive species for evaluation of the potential
developmental toxicity of PFOS.
Perfluorooctane sulfonate (PFOS) - February 2014 5-6
Draft - Do Not Cite or Quote
-------
5.1.1.1 Benchmark Dose Approach
As a second step in the dose-response analysis, benchmark modeling of dose-response for
key endpoints was evaluated. Endpoints considered as critical effects in several studies included
decreased offspring body weight, reduced pup survival, liver histopathology, and liver weight
changes. For the developmental endpoints, benchmark dose estimates were calculated and
published by the authors of the studies. Liver endpoints were modeled as described below based
on the available data.
Benchmark dose estimates for developmental endpoints in rats were published in a two-
part developmental toxicity study (Thibodeaux et al., 2003; Lau et al., 2003) and in a one-
generation study in which only dams were treated (Luebker et al., 2005a). Protocols varied
slightly between studies, but dams were treated throughout gestation and the offspring evaluated
either on GD 21 or by postnatal day 8 in each study. As recommended for continuous, normally
distributed developmental toxicity endpoints, a shift in the distribution of 0.67 standard
deviations was used as the benchmark response, which represents approximately an extra 5% of
the individual values being greater than approximately the 99th percentile or about an extra 5%
less than approximately the 1st percentile of the distribution in controls (U.S. EPA, 1995).
Goodness-of-fit information was used by the authors to choose the best model for each data set.
The estimated values from the modeling are reported as the BMD5 and its lower 95% confidence
limit, the BMDLs as shown in Table 5-3.
TABLE 5-3. Benchmark Dose Modeling for a 5% Increased Risk of Developmental
Toxicity in Rats
Maternal
treatment
GD 2-20
GD 2-21
6 weeks prior
to mating
until lactation
day 4
Best fit
Model
Polynomial
Hill
Logistic
Logistic
NCTR
Not stated
Not stated
Not stated
Not stated
Not stated
BMDS
mg/kg/day
0.22
0.23
0.31
8.85
1.07
0.45
0.63
0.39
0.41
1.06
BMDL5
mg/kg/day
0.15
0.05
0.12
3.33
0.58
0.31
0.39
0.27
0.28
0.89
Effects Modeled
Decreased maternal body
weight.
Decreased maternal T4.
Fetal sternal defects.
Fetal cleft palate
Reduced neonatal survival on
postnatal day 8
Reduced gestation length
Decreased pup birth weight
Decreased pup weight on day 5
Decreased pup weight gain
Reduced pup survival on day 5
Reference
Thibodeaux
etal.,2003
Lau et al.,
2003
Luebker et
al., 2005a
Benchmark dose estimates were not published in the two-generation study (Luebker et
al., 2005b) which had the lowest NOAEL, 0.1 mg/kg/day, for decreased pup body weight and a
LOAEL of 0.4 mg/kg/day. Significantly reduced body weight was noted in the F2 pups on
lactation days 7 and 14 but not on days 1, 4, or 21. Data presented included mean pup
weight/litter but number of litters was not reported. As a consequence benchmark dose analyses
could not be run on these data.
Benchmark dose estimates for maternal and developmental endpoints were also published
for mice (Thibodeaux et al., 2003; Lau et al., 2003) treated similarly to the protocol for rats. The
resulting estimates for offspring survival and malformations were approximately an order of
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-7
-------
magnitude greater than those for the rat and are not considered further as potential points of
departure (PODs) for use in RfD development.
The incidence data for liver lesions were modeled by EPA using benchmark dose
software (BMDS) v. 2.1.2 for dichotomous variables. Goodness-of-fit information (p value and
Akaike's Information Criterion [AIC]) was used to choose the best model for each data set. In
the 2-year study in rats (Thomford, 2002), dose-related increased incidences were observed for
hepatocellular centrilobular hypertrophy in males and females and centrilobular hepatocytic
vacuolation in males (as shown in Table 4-10). These lesions are consistent with the high
uptake/storage of PFOS by the liver. A 10% extra risk for each endpoint was selected as the
benchmark response for the analysis. Results are shown in Table 5-4 and the modeling output is
included in Appendix B. Figure 5-1 provides the graphic results from these datasets. Because
the incidence of centrilobular, eosinophilic, hepatocytic granules was markedly increased in
males and females only at the highest dose, these data were not modeled. In males, the BMDLio
for both hepatocellular hypertrophy and hepatocyte vacuolization are comparable (0.033 and
0.028 mg/kg/day, respectively). Based on hepatocyte hypertrophy females are less responsive to
chronic exposure to PFOS than males by a factor of about 2; thus, the data for females were not
considered further for derivation of the RfD.
TABLE 5-4. Benchmark Dose Modeling for a 10% Increased Incidence of Liver Lesions in
Rats
Sex
Best fit
Model
BMD10
mg/kg/day
BMDL10
mg/kg/day
Notes
Reference
Hepatocellular centrilobular hypertrophy
Males
Females
Log-Probit
(no
restriction)
Log-Probit
(no
restriction)
0.0521299
0.0981083
0.0326765
0.0680339
2-year dietary administration
2-year dietary administration
Thomford,
2002
Thomford,
2002
Centrilobular hepatocytic vacuolation
Males
Log-Probit
(no
restriction)
0.0931649
0.0278419
2-year dietary administration
Thomford,
2002
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-8
-------
LogProbit Model with 0.95 Confidence Level
LogProbit
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
14:5710/032012
Males - hypertrophy
LogProbit Model with 0.95 Confidence Level
LogProbit
BMDL 3MD
0.1 0.2 0.3 0.4 0.5 0.6
13:0301/292013
Females - hypertrophy
LogProbit Model with 0.95 Confidence Level
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
LogProbit
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
18:5610/042012
Males - vacuolation
Figure 5-1. BMDS graphic output from selected model runs; data from Thomford, 2002.
For the studies wherein liver weight was a critical effect, the data were also modeled
using BMDS v. 2.1.2 for a continuous dataset with modeled variance where applicable. A 10%
increase in absolute liver weight was chosen as the benchmark response for the initial analysis.
This endpoint is considered to be a biomarker for systemic exposure in rodents when the
chemical is an activator of PPAR-a, rather than a biomarker of adversity. Although the PFOS
data support an increase in the expression of proteins involved in peroxisomal fatty acid
catabolism (Tan et al., 2012) and lipid transport and metabolism (Bijland et al., 2011), those
effects may not totally be a reflection of PPAR-a activation. Compared to PFOA, PFOS appears
to be a relatively weak activator of the PPAR-a receptor (Shipley et al., 2004; Wolf et al.,
2008,2012; Ishibashi et al., 2011).
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-9
-------
Results of the liver weight analyses are shown in Table 5-5 where goodness-of-fit
information (p value and AIC) was used to chose the best model for each data set. The modeling
output is included in Appendix B; Figure 5-2 provides the graphic results from these datasets.
Liver weight data were available in the male and female monkey (Seacat et al., 2002), the male
rat (Seacat et al., 2003), and the male mouse (Dong et al., 2009). However, the data for the male
mouse did not adequately fit any model as indicated by values of p<0.1 and are not included
here. For each data set, the BMD (10% increase in liver weight) and the lower-bound confidence
limit on the BMD (BMDL) are provided.
TABLE 5-5. Benchmark Dose Modeling for a 10% Increase in Liver Weight
Species
Best fit
Model
BMD10
mg/kg/day
BMDL10
mg/kg/day
Notes
Reference
Monkey
Monkey
male
Monkey
female
Power
(modeled
variance)
Exponential
Model 2
0.60743
0.03338792
0.0147931
0.0207989
Increased absolute liver
weight; 26 weeks
Increased absolute liver
weight; 26 weeks
Seacat et al.,
2002
Seacat et al.,
2002
Rat
Rat male
Exponential
Model 5
0.280426
0.0585612
Increased absolute liver wt;
98 days
Seacat et al.,
2003
The benchmark dose, liver weight response in male (BMDLio = 0.015 mg/kg/day) and
female (BMDLio = 0.021 mg/kg/day) monkeys after 26-weeks of exposure is slightly more
sensitive than both the liver weight and histopathology responses in male rats. Given that
monkeys are less responsive than rodents to PPAR-a activation, the liver weight BMDLs for the
monkey suggest that the increase, in this case, may not simply be a reflection of a classic PPAR-
a response.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-10
-------
Power Model with 0.95 Confidence Level
S 100
09:3301/292013
0.1 0.2 0.3 0.4 0.5 0.6 0.7
dose
Male monkey
Exponential Model 2 with 0.95 Confidence Level
100
90
80
70
eo
50
40
30
Exponential
10:1701/292013
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
dose
Female monkey
Exponential Model 5 with 0.95 Confidence Level
11:21 01/292013
Male rat
0.6 0.8
dose
Figure 5- 2. BMDS graphic output from liver weight model runs; data from Seacat et al.,
2002, 2003
5.1.1.2 Pharmacokinetic Model Approach
The pharmacokinetics of PFOS are not well understood but are considered to be roughly
similar to that of the more extensively studied PFOA (Andersen et al., 2006; Loccisano et al.,
2011, 2012a,b). Pharmacokinetic studies with PFOS have measured a serum half-life of
approximately 48 days in rats (Butenhoff and Chang, 2007) and approximately 121 days in
monkeys (Chang et al., 2012). Compared with a half-life of months in animals, the human PFOS
half-life of roughly five years (Olsen et al., 2007) is clearly inconsistent with simple allometric
scaling across species. The reasons for the contrast between the multi-year PFOS elimination
half-life in humans and the shorter half-lives of laboratory animals are not fully understood.
In the process of developing pharmacokinetic models for PFOS in monkeys and rats,
several assumptions had to be made by the authors in order to fit the model outcome to the
available pharmacokinetic data (Anderson et al., 2006; Tan et al., 2008; Loccisano et al., 2011,
2012a,b). Parameters for transporter characteristics were assumed to be the same for both sexes
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-11
-------
consistent with no observed differences in elimination; a storage compartment was added
because PFOS appears in the urine of male rats at a slower rate than it disappears from plasma.
To account for the concentrations of PFOS found in the liver at long times (72-99 days) after
dosing, description of saturable binding in the liver was used in the adult rat model (Loccisano et
al., 2012a), and a time-dependent description of the free fraction of chemical in the central
(plasma) compartment was added to describe possible changes in binding or elimination that
may be occurring over time.
It is hypothesized that the pharmacokinetics of perfluorinated compounds, including
PFOS, are driven by transporters, notably in kidney and liver (Kudo, 2006). All published
pharmacokinetic models assume that perfluoroalkyl acids are rapidly eliminated into the urine,
but then reabsorbed into the body by transporters before they can be excreted. Whenever the
ability of transporters to reabsorb the compounds is saturated, however, the compounds can be
rapidly eliminated.
Currently, no data characterizing the uptake kinetics of PFOS by renal organic anion
transporters are available. The long half-life of PFOS in both the male and female rat, the
monkey, and humans, suggests that it is most likely reabsorbed back into the plasma by a renal
transport process, similar to the documented process for PFOA. However, because no significant
difference is observed for the elimination half-life between male and female rats (Chang et al.,
2012), it can be assumed that the transporters responsible for reabsorption are similar in both
sexes, and different in some way from those for PFOA. Differences could be due to either the
kinetic interactions of the chemicals with the same transporters or differences in transporters.
Studies characterizing the uptake and reuptake of PFOS by renal transporters will help in
refinement of these parameters.
Because the exposures of interest typically involve repeated doses, the saturable renal
resorption model (Andersen et al., 2006, Section 3.5.1) was used to describe how PFOS can
reach steady state faster than the elimination half-life would indicate. Using a simpler, linear
pharmacokinetic model (e.g. the one-compartment model) would result in estimated exposures
due to repeated doses that are much greater than actually seem to occur (Seacat et al., 2003).
Despite the uncertainties described above, application of this model to PFOS yielded good
agreement between the experimental data and the model simulations.
The biological basis for the saturable resorption model parameters is uncertain (Lou et al.
2009), particularly with respect to the idea that glomerular filtration alone is responsible for the
movement of PFOS into the renal filtrate given that PFOS is highly protein bound (Kerstner-
Wood et al., 2003) and the interactions with the renal transporters are unknown. However, the
saturable resorption model functions well in describing the internal doses in studies of PFOS
(Andersen et al., 2006; Loccisano et al., 2011, 2012a,b).
The saturable resorption model is an empirical model similar to the commonly used one-
and two-compartment models. Because it is an empirical model, extrapolation between species
is difficult. Resorption parameters cannot be directly linked to the properties of the relevant
serum transport proteins or membrane transporters. Therefore the saturable resorption model
parameters must be estimated using species-specific pharmacokinetic data.
Perfluorooctane sulfonate (PFOS) - February 2014 5-12
Draft - Do Not Cite or Quote
-------
Pharmacokinetic data (serial blood concentrations following treatment with known
quantities of PFOS) were collected for three species: cynomolgus monkeys (Chang et al., 2012;
Seacat et al., 2002), Sprague-Dawley rat (Chang et al., 2012), and CD1 mouse (Chang et al.,
2012). Given that the available pharmacokinetic studies were not necessarily designed with the
expectation of non-linear pharmacokinetics, many parameters associated with saturable
resorption should be expected to be uncertain. In the absence of ideal data, the question of
parameter values becomes one of what range of values are consistent with the data, rather than
which single value is most consistent with the data. Considering a range of possible values that
might equally well explain the data can be addressed with Bayesian statistical analysis, which
determines the distributions of parameter values that are likely to reproduce the observed data
(Gelman et al., 2004).
In this case, a non-hierarchical model for parameter values was applied in which there is
a single value shared by all individuals of the same species and gender. Body-weight and
treatment (number and magnitude of doses) are the only parameters that may vary between
individuals. Body-weight was only varied when animal-specific body-weights were available.
Bayesian analysis allows formal inclusion of prior knowledge in the form of set
distributions on the parameters being estimated (Gelman et al., 2004). Given that empirical
pharmacokinetic parameters can have a wide range of values, vague, bounded prior distributions
are appropriate (Wambaugh et al., 2008). For all estimated parameters, the prior knowledge was
distributed log-normally. This constrained the parameters to positive values. The mean and
variances used are given in Table 5-6.
TABLE 5-6. Description of prior distributions used.
Parameter
ka
vcc
k12
Rv2:Vl
T
A maxc
kT
free
Pec
v^
Mean
1
1
1
1
1
1
0.01
1
1
Variance
1000
1000
1000
0.5
1000
1000
0.5
1000
1000
Bounds
Lower
1/6
10-io
10-io
10-io
10-io
10-io
10-io
0.01
10-io
Upper
105
105
105
100
105
105
1
105
10
Parameters were log-normally distributed with the mean and variance listed. Bounds were used to reduce the time spent sampling
in areas thought to have low probability, and were expanded when large posterior mass at the bounds indicated that the bounds
were too narrow.
The deep tissue compartment of the Andersen et al. (2006) model is characterized in
terms of the rates to and from that compartment (ki2 and k2i, respectively). This corresponds to a
volume of distribution V2 = ki2*Vi/k2i, so that the ratio of the volume of the second
compartment to the first is Rv2:vl = ki2/k2i. In order to enforce the assumption that the primary
(serum) compartment contains a significant portion of the PFOS, the volume of the deep tissue
was constrained to be no more than 100 times greater than the volume of distribution of the
serum. For this reason the ratio of the two volumes is estimated, rather than the rate from the
second compartment to the first compartment. The rate of flow from the deep tissue back to the
serum was calculated as k2i = ki2/Rv2:vi.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-13
-------
Bayesian analysis was performed using Markov Chain Monte Carlo (Gelman et al.,
2004). The distribution of parameter values were considered to "burned in" (i.e., true draws
from the posterior combination of the prior distributions and the available data and independent
of the starting values) when they passed the Heidelberger and Welch Stationarity test
(Heidelberger and Welch, 1983) as implemented in the Coda Package (Best et al., 1995) for R (R
Development Core Team, 2010).
The estimated parameters (median and 95% interval), presented in Table 5-7, generally
appear plausible. The volumes of distributions were all less than 1 L/kg BW, and varied between
species. The parameters for the male mouse were extremely uncertain, reflecting a combination
of the limited amount of data (two single dose treatments) and perhaps inappropriateness of the
model used. However, data to support the model assumptions are needed. Parameters such as
the flow to and volume of the filtrate compartment were found to be very uncertain, as in
previous studies (Andersen et al. 2006). The median fraction of blood flow to the filtrate (Qfiic)
was physiological (less than or equal to the fraction of blood flow to the kidney) for the male and
female rats, but appears too high for the mouse and monkeys, which might be explained by either
a lack of sufficient data or perhaps the role of secreting transporters in the kidney. Although the
physical interpretation is moot, the form of the saturable resorption model seems appropriate.
Perfluorooctane sulfonate (PFOS) - February 2014 5-14
Draft - Do Not Cite or Quote
-------
TABLE 5-7. Pharmacokinetic parameters used in the Andersen et al. (2006) model.
Parameter
BWb
Cardiac
Outpuf
ka
vcc
ki2
Rv2:Vl
T
-'-maxc
kT
Free
Pec
V^
Units
kg
L/h/kga74
1/h
L/kg
1/h
Unitless
M/h
M
Unitless
L/h
L/kg
CD1 Mouse
Female"
0.02
8.68
1.16(0.617
- 42400)
0.264 (0.24
- 0.286)
0.0093
(2.63e-10 -
38900)
1.01 (0.251
- 4.06)
57.9 (0.671
- 32000)
0.0109
(1.44e-05 -
1.45)
0.00963
(0.00238 -
0.0372)
0.439
(0.0125 -
307)
0.00142
(4.4e-10 -
6.2)
CD1 Mouse
Male3
0.02
8.68
433.4 (0.51
-803.8
0.292 (0.268
-0.317
2976 (2.8e-
10 - 4.2e4)
1.29(0.24-
4.09)
I.le4(2.1-
7.9e4)
381 (2.6e-5
-2.9e3)
0.012
(0.0024 -
0.038)
27.59 (0.012
-283)
0.51 (3.5e-
10-6.09)
Sprague-
Dawley Rat
Female"
0.203
12.39
4.65 (3.02 -
1980)
0.535 (0.49 -
0.581)
0.0124 (3. le-
10 - 46800)
0.957 (0.238 -
3.62)
1930(4.11 -
83400)
9.49 (0.00626
-11100)
0.00807
(0.00203 -
0.0291)
0.0666 (0.0107
-8.95)
0.0185 (8.2e-
07-7.34)
Sprague-
Dawley Rat
Male3
0.222
12.39
0.836 (0.522
-1.51)
0.637 (0.593
- 0.68)
0.00524
(2.86e-10 -
43200)
1.04 (0.256 -
4.01)
1.34e-06
(1.65e-10 -
44)
2.45 (4.88e-
10 - 60300)
0.00193
(0.000954 -
0.00249)
0.0122
(0.0101 -
0.025)
0.000194
(1.48e-09 -
5.51)
Cynomolgus
Monkey Male
and Female"
3.42
19.8
132 (0.225 -
72100)
0.303 (0.289 -
0.314)
0.00292 (2.59e-
10 - 34500)
1.03 (0.256 -
4.05)
15.5 (0.764 -
4680)
0.00594 (2.34e-
05-0.0941)
0.0101 (0.00265
- 0.04)
0.198(0.012-
50.5)
0.0534(l.le-07-
8.52)
Means and 95% confidence interval from Bayesian analysis are reported. For some parameters the distributions are quite wide,
indicating uncertainty in that parameter (i.e., the predictions match the data equally well for a wide range of values).
aData sets modeled for the mouse and rat were from Chang et al., 2012 and for the monkey from Seacat et al., 2002 and Chang et
al.,2012.
a Average body weight for species — individual-specific body weights
b Cardiac outputs obtained from Davis & Morris, Pharmaceutical Res 10, 1093, 1993
For each study with a toxicological endpoint and NOAEL/LOAEL, the time-integrated
serum concentration (area under the curve or AUC) was determined for the exposure duration
investigated in that study. Generally, it was assumed that animals were observed at the end of
dosing. The data for studies in the rats are summarized in Table 5-8. For the Butenhoff et al.
(2009) study two different AUCs were calculated - gestational only (for the male offspring
endpoint) and gestational plus twenty days postnatal (for the maternal endpoint). This separation
of the two exposures neglects lactational transfer of compound, which was not modeled.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-15
-------
TABLE 5-8. Predicted final serum concentration and time integrated serum concentration
(AUC) for different treatments of rat.
Study
Curran et al.
2008
Curran et al.
2008
Seacat et al.
2003
Seacat et al.
2003
Butenhoffet
al. 2009 and
Chang et al.
2009
Butenhoffet
al. 2009 and
Chang et al.
2009
Thibodeaux
et al. 2003
and Lau et
al. 2003
Luebker et
al. 2005b
Luebker et
al. 2005a
Species /
Strain
Male
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Male Rat/
Crl:CD(SD)
IGSBR
Female Rat/
Crl:CD(SD)
IGSBR
Rat/Sprague-
Dawley
Rat/Sprague-
Dawley
Rat/Sprague-
Dawley
Rat/Crl:CD
(SD)IGS
VAF/Plus
Rat/Crl:CD
(SD)IGS
VAF/Plus
Study
Duration
And Type
28 days
28 days
98 days
98 days
Gestation
(22 Days)
Gestation
(21 Days)
+ 20 Days
Postnatal
GDs 2-20
(19 days)
6 wks
prior to
mating
through
gestation
and
lactation
84 days
6 wks
prior to
mating
through
gestation;
63 days
Oral
Doses
mg/kg/day
0.14
1.33
3.21
6.34
0.15
1.43
3.73
7.58
0.03
0.13
0.34
1.33
0.04
0.15
0.40
1.56
0.1
0.3
1
0.1
0.3
1
1
2
3
5
10
0.1
0.4
1.6
3.2
0.4
0.8
1.0
1.2
1.6
2.0
Measured
Serum
Concentration
mg/L
No data
No data
No data
No data
No data
No data
No data
No data
4.04 (0.80)
17.1 (1.22)
43.9 (4.9)
148(14)
6.96 (0.99)
27.3 (2.3)
64.4 (5.5)
223 (22)
1.722(0.068)
6.245 (0.096)
26.630 (3.943)
3.159(0.081)
8.981 (0.275)
30.480(1.294)
19.69a
44.33a
70.62a
79.39a
189.4a
4.52(1.15)
26.2(16.1)
136(86.5)
155(39.3)
NT
NT
NT
NT
NT
NT
Species /
Strain Used
for
Prediction
Male
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Male
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Predicted
Final Serum
Concentration
mg/L
4.88(0.148)
46.4(1.4)
112(3.39)
221 (6.7)
6.88(0.212)
65.5 (2.03)
170 (5.76)
344(18.3)
2.64(0.103)
10.5 (0.408)
26.3(1.02)
105 (4.08)
4.24 (0.0867)
17(0.345)
42.3 (0.882)
168(6.38)
3.7(0.121)
11.1(0.364)
37(1.21)
6.49(0.172)
19.5(0.515)
64.8(1.74)
32.4(1.1)
64.7 (2.2)
97(3.31)
161 (5.69)
321 (15.6)
9.65 (0.202)
38.5(0.814)
153(4.74)
303(17.5)
38.5(0.814)
76.9(1.75)
96.1 (2.32)
115(2.99)
153(4.74)
191 (7.07)
Predicted
AUC mg/L*h
1840(56.9)
17400 (540)
42100(1300)
83100(2580)
2500(85.9)
23800 (820)
62100 (2170)
126000 (5090
3970 (128)
15900(507)
39600 (1270)
158000
(5070)
5780 (128)
23100(511)
57700 (1280)
230000
(6390)
1060 (37.8)
3170(114)
10600 (378)
3570(111)
10700 (333)
35600(1110)
8010 (293)
16000 (585)
24000 (879)
40000 (1470)
79800(3190)
9100(235)
36400 (942)
145000
(4100)
288000
(11200)
36400 (942)
72600(1910)
90700 (2410)
109000
(2940)
145000
(4100)
181000
(5460)
Numbers in parentheses indicate standard deviation
GD = gestation day; NT = not tested
aThibodeaux et al. (2003) data available only in a graph; values obtained from author.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-16
-------
The data on the results from studies in mice and the monkey are provided in Tables 5-9
and 5-10, respectively. The Lau et al. (2003) data on mice are representative of the impact of
PFOS on developmental endpoints as described in Table 5-2 although the duration of this study
is relatively short at 17 days. The Seacat et al. (2002) study on monkeys is a long term (6
months) multiple dose study of systemic toxicity in which the LOAEL for effects on liver weight
and cholesterol was accompanied by death of 2/6 monkeys.
TABLE 5-9. Predicted final serum concentration and time integrated serum concentration
(AUC) for the mouse.
Study
Lauet
al. 2003
Species /
Strain
Female
Mouse/CD-
1
Study
Duration
And
Type
GD 1-17
(17 days)
Administered
Doses
mg/kg/day
1
5
10
15
20
Measured
Final Serum
Concentration
mg/L
NT
NT
NT
NT
NT
Species /
Strain
Used for
Prediction
Female
Mouse /
CD1
Predicted
Final Serum
Concentration
mg/L
62.7 (2.28)
287 (19.2)
421 (107
484 (186)
532 (257)
Predicted
AUC
mg/L*h
14700
(533)
70900
(2560)
123000
(14500)
156000
(32100)
180000
(49900
Numbers in parentheses indicate standard deviation
GD = gestation day; NT = not tested
TABLE 5-10. Predicted final serum concentration and time integrated serum
concentration (AUC) for the monkey.
Study
Seacat
etal.
2002
Species /
Strain
Monkey /
Cynomol-
gus
Study
Durat-
ion
And
Type
182 days
Adminis-
tered Doses
mg/kg/day
0.03
0.15
0.75
Measured
Final Serum
Concentrat-
ion mg/L
F: 13.2(1.4)
M: 15.8(1.4)
F: 66.8 (10.8)
M: 82.6 (25.2)
F: 171 (22)
M: 173 (37)
Species /
Strain Used
for
Prediction
Monkey /
Cynomol-
gus
Predicted
Final Serum
Concentrat-
ion
mg/L
7.58(0.16)
32.9 (0.557)
86.7(2.14)
Predict-
ed AUC
mg/L*h
22100
(382)
102000
(1530)
332000
(6450)
Numbers in parentheses indicate standard deviation
M = male; F = female
The AUC for the LOAEL or NOAEL of each data set can be used to determine an
average serum concentration by dividing it by the duration of the study in days with adjustment
for the number of hours in a day. Averaging the terminal serum concentrations for the duration
of exposure is important because of the variability in the times of exposure across the studies
(17-182 days). The following equation is used for the conversion:
(days)
Average Serum Concentration = AUC (mg/L*h) x 1 day/24 hours + exposure duration
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-17
-------
For example, in a case where the AUC was 30,000 mg/L*h and the study duration was 90
days, the Average Serum Concentration would be calculated as follows:
Average Serum Concentration = 30,000 mg/L*h H- (90 days x 24 h/day) = 13.89 mg/L
This calculation has the advantage of normalizing the serum concentration across the
exposure durations to generate a uniform metric for internal dose in situations where the dosing
durations varied and serum measurements were taken immediately prior to sacrifice. The
averaged serum concentration is a hybrid of the AUC and the maximum serum concentration.
As applied to the database for PFOS, average serum concentration appears to be a stable
reflection of internal dosimetry.
Table 5-11 provides the AUC from the model, the dosing duration from each of the
modeled studies, and the resultant average serum concentration. Internal doses associated with
developmental toxicity were 18.06 and 24.07 mg/L for reduced pup body weight (Luebker et al.,
2005a,b), 36.18 mg/L for changes in motor activity (Butenhoff et al., 2009), and 35.09-48.02
mg/L for pup survival (Lau et al., 2003; Luebker et al., 2005a). In comparison, internal doses
associated with increased liver weight were 67.18-76.01 mg/L (Seacat et al., 2002; 2003). Thus,
the internal doses associated with the developmental and liver effect levels (LOAELs) differ by
much less than an order of magnitude (18.06 mg/L to 76.01 mg/L) while the corresponding AUC
values differ by more than an order of magnitude (16000 mg/L*h to 332000 mg/L*h).
The internal doses associated with no adverse effects on developmental and liver
endpoints (NOAELs) were very similar with overlapping ranges; the average serum
concentrations ranged from 4.51-24.07 mg/L for developmental endpoints (Butenhoff et al.,
2009; Lau et al., 2003; Luebker et al., 2005a,b) and from 16.84-23.35 mg/L for liver weight
changes (Seacat et al., 2002; 2003). Despite the similarity in average serum concentrations, the
AUC values differ by two orders of magnitude (8010 mg/L*h to 102000 mg/L*h). Given the
differences in external doses, the projected serum levels are proportionally quite similar.
Perfluorooctane sulfonate (PFOS) - February 2014 5-18
Draft - Do Not Cite or Quote
-------
TABLE 5-11. Average Serum concentrations Derived from the AUC and the duration of
Dosing
Study
Seacat et al., 2002
monkey t liver
weight
Luebker et al.,
2005b | rat pup
body weight
Luebker et al.,
2005 a | rat pup
body weight
Luebker et al.,
2005 a J, maternal
weight, pup
survival
Lau et al., 2003
J, rat pup survival
Butenhoff etal.,
2009 rat |DNT
Seacat et al., 2003
male frat liver
weight
Dosing
duration
days
182
84
63
63
19
41
98
NOAEL
mg/kg/day
(AUC
mg/L*h)
0.15
(102000)
0.1
(9100)
None
0.4
(36400)
1.0
(8010)
0.3
(10700)
0.34
(39600)
NOAEL
(Av serum
mg/L)
23.35
4.51
None
24.07
17.56
10.87
16.84
LOAEL
mg/kg/day
(AUC
mg/L*h)
0.75
(332000)
0.4
(36400)
0.4
(36400)
0.8
(72600)
2.0
(16000)
1.0
(35600)
1.33
(158000)
LOAEL
(Av serum
mg/L)
76.01
18.06
24.07
48.02
35.09
36.18
67.18
Table 5-11 identifies 4.51 and 10.87 mg/L as the lowest average serum concentrations
which were associated with a NOAEL for offspring effects; the LOAELs were associated with
decreased pup body weight (Luebker et al., 2005b) and increased motor activity in male pups
(Butenhoff et al., 2009). Average serum values for no increases in liver weight in monkeys
(Seacat et al., 2002) and male rats (Seacat et al., 2003) are very similar to the average serum
value in Butenhoff et al. (2009). Thus, it appears that the NOAELs are consistent across gender,
species, and treatment with respect average serum concentration. Assuming that mode of action
and susceptibility to toxicity do not vary and that pharmacokinetics alone explains variation, it is
reasonable to expect similar concentrations to cause similar effects in humans.
The EPA model employed here to generate the predicted AUC values that became the
basis for the average serum concentrations shown in Table 5-11, does not include a gestational or
lactational component. However the results are in good agreement with those of Loccisano et al.
(2012b) from their gestational and lactational model. Comparison of the average maternal serum
concentrations calculated for developmental endpoints (Butenhoff et al., 2009; Lau et al., 2003;
Luebker et al., 2005a) with those depicted graphically in Figure 3-7 (from Loccisano et al.,
2012b), demonstrates good agreement between the two models. For example the LOAEL of 1
mg/kg/day for developmental neurotoxicity (Butenhoff et al., 2009) yields a calculated average
maternal serum of 36.18 mg/L as seen in Table 5-11 which is very similar to the approximately
25 mg/L for the dams that can be estimated from the graph (Loccisano et al., 2012b). The
slightly higher value calculated from the EPA model may be due to the longer dosing interval, 41
days, used by Butenhoff et al. (2009) versus GD 20 levels presented graphically by Loccisano et
al. (2012b). Fetal PFOS serum concentration on GD 20 was published by Chang et al. (2009),
but because the EPA model predicts maternal values, not fetal, a direct comparison to the fetal
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-19
-------
plasma predicted by Loccisano et al. (2012b; Figure 3-7) cannot be made. However, despite the
limitations in the fetal data, values generated by the EPA model can be accepted with reasonable
confidence that the predicted AUC values accurately represent maternal levels during gestational
and lactational exposures.
The predicted serum concentrations can be converted into an oral equivalent dose at
steady state by recognizing that, at steady state, clearance from the body must equal dose to the
body. Clearance can be calculated if the rate of elimination (derived from half-life) and the
volume of distribution are both known.
A reliable measure of half-life in humans is from a retired worker population followed for
five years. Olsen et al. (2007) calculated the PFOS half-life in this former worker population as
5.4 years (see Section 3.5.2). Thompson et al. (2010) gives a volume of distribution of 0.23 L/kg
bw (see Section 3.5.3). These values combined give a clearance of S.lxlO"5 L/kg bw/day as
determined by the following equation:
CL = Vd x (In 2 -H ti/2) = 0.23 L/kg bw/day x (0.693 - 1971 days) = 0.000081 L/kg bw/day
Where:
Vd = 0.23 L/kg
In 2 = 0.693
ti/2 = 1971 days (5.4 years x 365 days/year = 1971 days)
These values combined give a clearance of 8.1 x 10"5 L/kg bw/day.
Scaling the derived average concentrations (in mg/L) for the NOAELs and LOAELs in
Table 5-9 gives predicted oral human equivalent doses (HEDs) in mg/kg bw/day for each
corresponding serum measurement. The FLED values are the predicted human oral exposures
necessary to achieve serum concentrations equivalent to the NOAEL or LOAEL in the animal
toxicity studies. Note that this scaling uses linear human kinetics in contrast to the non-linear
phenomena observed at high doses in animals.
Thus, FLED = average serum concentration (in mg/L) x CL
Where:
Average serum is from model output in Table 5-11
CL = 0.000081 L/kg bw/day
The resulting FLED values are shown in Table 5-12. Endpoints considered as critical
effects in multiple studies include offspring growth and survival, liver weight changes, and liver
histopathology.
Perfluorooctane sulfonate (PFOS) - February 2014 5-20
Draft - Do Not Cite or Quote
-------
TABLE 5-12. Human Equivalent Doses Derived from the Modeled Animal Average Serum
Values
Study
Seacatetal.,
2002 monkey
t liver weight
Luebker et al.,
2005b
Luebker et al.,
2005a| rat pup
body weight
Luebker et al.,
2005a
J, maternal
weight pup
survival
Lau et al., 2003
I rat pup
survival
Butenhoffet
al., 2009 t rat
DNT
Seacatetal.,
2003 tmale rat
liver weight
Dosing
duration
days
182
84
63
63
19
41
98
NOAEL
mg/kg/d
0.15
0.1
None
0.4
1.0
0.3
0.34
NOAEL
Av serum
mg/L
23.35
4.51
None
24.07
17.56
10.87
16.84
RED
mg/kg/d
0.0019
0.00037
0.0019
0.0014
0.00088
0.0014
LOAEL
mg/kg/d
0.75
0.4
0.4
0.8
2.0
1.0
1.33
LOAEL
Av serum
mg/L
76.01
18.06
24.07
48.02
35.09
36.18
67.18
RED
mg/kg/d
0.0062
0.0015
0.0019
0.0039
0.0028
0.0029
0.0054
5.1.1.3 RfD Quantitation
Several acceptable points of departure (PODs) can be used in the process of identifying
the POD for RfD development:
• NOAEL or LOAEL values
• Lower 95% confidence bounds on the BMD (BMDLs), and
• Human Equivalent Doses (HED).
Studies that have more than one POD for the same NOAEL or LOAEL are summarized
in Table 5-13. All studies identified a NOAEL for PFOS except for the endpoint of offspring
growth as measured by body weight in the one-generation study by Luebker et al. (2005a). The
developmental effects of reduced pup body weight and survival occurred in the absence of
changes in maternal liver weight. The calculated HED values associated with no adverse effects
on developmental and liver endpoints (NOAELs) were very similar with a range of 0.00088-
0.0019mg/kg/day.
Modeling of dose-response to identify a BMD and BMDL was successful for most
studies. All benchmark models targeted a 10% increase in liver weight or histopathology
incidence or a 5% decrease in offspring body weight and survival. Most of the studies were
amenable for derivation of HED based on average serum measurements from the
pharmacokinetic model because dose and species-specific serum values were available for model
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-21
-------
development. Entries in Table 5-13 provide at least three potential PODs for consideration; four
studies provide five values.
TABLE 5-13. RfD Point of Departure Options from the PFOS Animal Studies
Studies
Seacat et al.,
2002
Seacat et al.,
2003
Thomford,
2002
Thibodeaux
etal.,2003
Lau et al.,
2003
Butenhoff et
al., 2009
Luebker et
al., 2005b
Luebker et
al., 2005a
Luebker et
al., 2005a
NOAEL
(mg/kg/
day)
0.15
0.34
0.018
1.0
1.0
0.3
0.1
None
0.4
LOAEL
(mg/kg/
day)
0.75
1.33
0.072
2.0
2.0
1.0
0.4
0.4
0.8
BMDL10
(mg/kg/
day)
0.015
0.059
0.033
0.12*
0.58*
-
-
0.27*
0.89*
RED
NOAEL
(mg/kg/
day)
0.0019
0.0014
-
0.0014
0.00088
0.00037
-
0.0019
RED
LOAEL
(mg/kg/
day)
0.0062
0.0054
-
0.0028
0.0029
0.0015
0.0019
0.0039
Endpoint
Increased liver weight.
Increased liver weight.
Liver hypertrophy.
Decreased maternal and
pup body weight
(LOAEL); fetal sternal
defects (BMDL5)
Reduced pup survival.
Increased motor activity in
male pups on PND 17.
Decreased pup body
weight.
Decreased pup body
weight.
Reduced pup survival.
' The value provided for developmental endpoints is the BMDL05 rather than the BMDL10.
As explained previously, human data have identified significant relationships between
serum levels and specific indicators of adverse health effects but lack the exposure information
for dose-response modeling. For this reason none of the human studies provided an appropriate
POD for RfD derivation. The pharmacokinetically-modeled average serum values from the
animal studies are restricted to the animal species selected for their low dose response to oral
PFOS intakes. Extrapolation to humans adds a layer of uncertainty that must be accommodated
in deriving the RfD.
Each of the POD values represented in Table 5-13 requires a different quantification
approach. Thus, EPA has systematically examined the impact of POD on outcome through three
sets of calculations as follows:
• Derivation from the NOAEL values;
• Derivation from the BMDL values;
• Derivation from FLED values derived from modeled average serum values.
NOAEL PODs: Table 5-14 provides the potential RfDs derived from the seven studies which
identified a NOAEL. These studies also provided a LOAEL which is critical for dose-response
characterization. The NOAEL is used as the POD according to Agency policy instead of the
LOAEL because lower uncertainty is applied to the NOAEL when calculating the RfD. The
lowest NOAEL is from a chronic bioassay in the rat (Thomford, 2002) and is the dose which did
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-22
-------
not cause microscopic lesions in the liver. Because it is a chronic study the uncertainty applied
in derivation of the RfD is less than that used in the other rat studies and the monkey study. The
NOAEL from Thomford (2002) is also protective of co-critical developmental toxicity effects
observed in other studies.
TABLE 5-14. The Impact of Quantification Approach on the RfD outcome for the PODs
from the available NOAELs
Study
Seacat et al., 2002
monkey
Seacat et al., 2003 rat
Thomford, 2002 rat
Thibodeaux et al.,
2003 rat
Lau et al., 2003 rat
Butenhoff etal.,
2009 rat
Luebker et al., 2005b
rat
Luebker et al., 2005a
rat
POD dose
mg/kg/day
0.15
0.34
0.018
1.0
1.0
0.3
0.1
0.4
UFH
10
10
10
10
10
10
10
10
UFA
48
123
123
123
123
123
123
123
UFL
1
1
1
UFS
10
10
1
10
10
10
10
10
UFD
1
1
1
UFtotal
4800
12300
1230
12300
12300
12300
12300
12300
Potential
RfD
mg/kg/day
0.00003
0.00003
0.00001
0.00008
0.00008
0.00002
0.000008
0.00003
Uncertainty Factor (UF) Application
UFH A ten-fold adjustment is assigned to account for intrahuman variability and applied for all
PODs.
UFA. Determination of the interspecies uncertainty factor for the NOAEL requires application of
the equation for first order kinetics in order to determine the pharmacokinetic adjustment
associated with differences in half-life between humans and rats or monkeys. Pharmacokinetic
studies with PFOS have measured a serum half-life of approximately 48 days in rats (Butenhoff
and Chang, 2007) and approximately 121 days in monkeys (Chang et al., 2012). The equation
also utilizes a volume of distribution component, which for humans has been calibrated as 230
mL/kg (Thompson et al., 2010). This volume of distribution is similar to those reported for
monkeys, female rats, and mice in pharmacokinetic studies (Chang et al., 2012) and utilized in
pharmacokinetic models on monkeys (Anderson et al., 2006).
The equation that describes first order kinetics is as follows (Medinsky and Klaassen, 1996):
CL = Vd x (In 2 - ti/2)
Where:
Vd = 0.23 L/kg
Ln 2 = 0.693
Half-life = 48 days for rats; 121 days for monkeys; and 1971 days for humans
CLrat = 0.23 L/kg x (0.693 - 48 days) = 0.0033 L/kg/day
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-23
-------
CLmonkey = 0.23 L/kg x (0.693 - 121 days) = 0.0013 L/kg/day
CLhuman = 0.23 L/kg x (0.693 - 1971 days) = 0.000081 L/kg/day
The ratio of CLratto CLhuman (0.0033 L/kg/day - 0.000081 L/kg/day = 41) is used as the
pharmacokinetic adjustment for differences between these species. The total UFA requires an
additional 3-fold factor for species differences in pharmacodynamics (41x3 = 123).
The ratio of CLmonkeyto CLhuman (0.0013 L/kg/day - 0.000081 L/kg/day = 16) is used as the
pharmacokinetic adjustment for differences between these species. The total UFA requires an
additional 3-fold factor for species differences in pharmacodynamics (16 x 3 = 48).
UFL A UF of 1 was used for the NOAEL-derived PODs following Agency policies.
UFS A ten-fold factor was applied to the NOAEL PODs to account for studies with less than
lifetime exposure.
In all cases the uncertainty factor for the strength of the database (UFo) is 1. The data base
for oral PFOS exposure studies is essentially complete although mechanistic questions relative to
the MOA have not yet been fully elucidated.
BMDL PODs. Potential PODs from benchmark dose estimates are listed in Table 5-13. The
lowest, species-specific benchmark dose values are those from the Seacat et al. (2002; 2003)
studies in the male monkey and male rat and the Thomford (2002) study in the male rat. For the
Seacat et al. (2002; 2003) studies the effect modeled was a 10% change in liver weight. For the
Thomford (2002) study the effect modeled was a 10% increase in incidence of hepatocellular
centrilobular hypertrophy. The BMDL estimates for the liver effects are lower than those
calculated for developmental effects, indicating that liver endpoints are protective of other co-
critical endpoints. The values for developmental endpoints are included in Table 5-15 for
comparison.
Perfluorooctane sulfonate (PFOS) - February 2014 5-24
Draft - Do Not Cite or Quote
-------
TABLE 5-15. The Impact of Quantification Approach on the RfD Outcome for the
BMDLs from liver and developmental endpoints
POD
BMDL10
(monkey Seacat et al.,
2002)
BMDL10
(rat Seacat et al., 2003)
BMDL10
(rat Thomford, 2002)
BMDL05
(rat Thibodeaux et al.,
2003)
BMDL05
(rat Lau et al., 2003)
BMDL05
(rat Luebker et al.,
2005a)
Dose
mg/kg/d
0.015
0.059
0.033
0.12
0.58
0.27
UFH
10
10
10
10
10
10
UFA
48
123
123
123
123
123
UFL
1
1
1
1
1
1
UFS
10
10
1
10
10
10
UFD
1
1
1
1
1
1
UFtotal
4800
12300
1230
12300
12300
12300
Potential
RfD
mg/kg/day
0.000003
0.000005
0.00003
0.00001
0.00005
0.00002
Uncertainty Value Application
The UFn, UFA, UF§ and UFo values are assigned as described for the NOAEL data in Table 5-
14.
UFL A UF of 1 was used for the BMDL-derived PODs following Agency policies.
HED PODs. The pharmacokinetic (PK) HEDs derived from Seacat et al. (2002; 2003), Lau et
al. (2003), Butenhoff et al. (2009), and Luebker et al. (2005a,b) were each examined as the
potential basis for the RfD (Table 5-16). Each of these studies contained a NOAEL from which
the HED could be derived. The outcomes for potential RfD values are quite similar
demonstrating the ability of the model to normalize the animal data across species, gender, and
exposure duration. Co-critical effects of liver lesions and developmental toxicity resulted in
similar HED values which yielded similar potential RfD values.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-25
-------
TABLE 5-16. The Impact of Quantification Approach on the RfD Outcomes for the HEDs
from the Pharmacokinetic Model Average Serum Values
POD
r K-rlh/JJ monkey Seacat et
al., 2002
PK-HED rat Seacat et al., 2003
PK-HED rat Lau et al., 2003
PK-HED rat Butenhoff et al.,
2009
PK-HED ratLuebkeretal.,
2005b
PK-HED rat Luebker et al.,
2005a
Value
mg/kg/day
0.0019
0.0014
0.0014
0.00088
0.00037
0.0019
UFH
10
10
10
10
10
10
UFA
3
3
3
3
3
3
UFL
1
1
1
1
1
1
UFS
1
1
1
1
1
1
UFD
1
1
1
1
1
1
UFtotal
30
30
30
30
30
30
Potential
RfD
mg/kg/day
0.00006
0.00005
0.00005
0.00003
0.00001
0.00006
Uncertainty Value Application
The UFH, UFL, and UFD values are assigned as described for the NOAEL data in Table 5-14.
The UFA is 3 for each study because the HED was derived using the steady state serum values
from the model to account for pharmacokinetic differences between animals and humans. The 3-
fold factor is applied to account for toxicodynamic differences between the animals and humans
The UFsis 1 bacause the point of departure is based on steady state serum concentrations.
RfD Selection
Based on the consistency of the response and with recognition of the use of
developmental toxicity as the sensitive endpoint, the 0.00003 mg/kg/day outcome is selected as
the RfD for PFOS. This value is the outcome for the modeled rat serum for developmental
neurotoxicity (Butenhoff et al., 2009) and supported by the slightly higher 0.00005 and 0.00006
mg/kg/day values for increases in liver weight and other developmental effects. Thus, co-
occurring critical endpoints are protected by the chosen RfD.
In the standard developmental neurotoxicity study by Butenhoff et al. (2009), male
offspring showed increased motor activity and decreased habituation on PND 17 following a
maternal dose of 1 mg/kg/day; no effects on pup body weight were reported. Animal data
consistently show higher PFOS levels in fetal tissues, including brain, as compared to maternal
tissues (Chang et al., 2009; Borg et al., 2010) resulting in the higher modeled predictions for fetal
versus maternal levels (Loccisano et al., 2012b; see Figure 3-7). Use of the developmental
toxicity endpoint is directly relevant to human health because in utero and lactational exposures
have been demonstrated. PFOS has been measured in the blood of newborns (Spliethoff et al.,
2008), in human breast milk (So et al., 2006; Karrman et al., 2010; Tao et al., 2008), and in
serum samples from children aged 2-12 (Olsen et al., 2002b). A human epidemiology found no
association with maternal PFOS levels and motor or mental development of their children; the
mean maternal serum concentration was approximately 35 mg/L (Fei et al., 2008b). The FED
used as the basis for the RfD, was calculated from an average serum concentration of 10.87 mg/L
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
5-26
-------
derived from the NOAEL of 0.3 mg/kg/day for developmental neurotoxicity (Butenhoff et al.,
2009; Table 5-11). This further supports the selected RfD as being protective of adverse human
health risks.
Use of the developmental neurotoxicity and liver weight endpoints as the basis for the
RfD confers protection against co-occurring adverse effects on offspring body weight and
survival. The selected RfD is supported by the results from the modeled serum value for rat
offspring body weight (Luebker et al., 2005a) and survival (Lau et al., 2003). These supporting
values are slightly greater than the recommended RfD, indicating that the RfD is protective of
co-occurring effects on the offspring. For these co-occurring effects, both the modeled values
were similar (0.00005-0.00006 mg/kg/day) adding confidence to the selected RfD.
The most conservative potential RfD was derived from modeled serum for decreased pup
body weight from the two-generation study (Luebker et al., 2005b). While this represents the
most sensitive endpoint, it was described by the authors as transient because the body weights
were significantly lower than those for controls for 2 of 5 time points monitored with the weight
differences before and after those two time points not significantly different. Benchmark data
for the body weight data to frame the NOAEL POD versus the LOAEL POD were not available.
Other modeled results are in good agreement with those from a gestational and lactational model
(Loccisano et al., 2012b) but cumulative effects from continuous exposure over multiple
generations is unknown and not accounted for by the present model.
5.1.2 RfC Determination
For development of an RfC, an adequate subchronic inhalation study must be available for
review. At this time, data are insufficient for PFOS to develop an RfC. The only inhalation study
available is an acute lethality inhalation study in rats (Rusch et al., 1979); no inhalation data are
available in humans. Therefore, an RfC for PFOS cannot be derived.
5.2 Dose-Response for Cancer Effects
In a chronic oral toxicity and carcinogen!city study of PFOS in rats, liver, thyroid and
mammary fibroadenomas were identified (Thomford, 2002). The biological significance of the
mammary fibroadenomas and thyroid tumors were questionable as a true dose-dependent response
was not identified. Mammary fibroadenomas were increased only in low-dose females and the
incidence rate showed a significant decreasing trend with dose. The incidence of thyroid follicular cell
adenomas was increased in the high-dose males administered PFOS for 52 weeks followed by basal
diet until natural death, but was not increased in high-dose males given PFOS for 104 weeks. The
liver tumors also had a questionable dose-response with slight but statistically significant increases
only in high-dose males and females. The liver tumors most found were adenomas (7/60 and 5/60 in
high-dose males and females vs. none in the controls of either sex); only one hepatocellular carcinoma
was found in a high-dose female. No evidence of cell proliferation or peroxisome proliferation was
found as measured by hepatic palmitoyl-CoA oxidase activity, PCNA, and BrdU.
Human epidemiology studies did not find a direct correlation between PFOS exposure and the
incidence of carcinogenicity in worker-based populations. Several studies looked at worker-related
toxicities of PFOS including some on cancer incidence (Mandel and Johnson, 1995; Alexander et al.,
2003; Alexander and Olsen, 2007). A slight increase in bladder cancer was identified in workers;
Perfluorooctane sulfonate (PFOS) - February 2014 5-27
Draft - Do Not Cite or Quote
-------
however, the outcome was not adjusted to account for other possible causes such as cigarette smoking
practices (Mandel and Johnson, 1995; Alexander et al., 2003). Based on these results, 3M undertook
another study of the same cohort to examine bladder cancer incidence (Alexander and Olsen, 2007)
and identified eleven bladder cancer cases with five deaths and 6 incident cases. Only 2 of the 6 self-
reported cases were confirmed with medical records and five of the 6 cases had a history of cigarette
smoking. Standardized incidence ratios (SIR) were estimated for 3 exposure categories and compared
to US cancer rates. SIRs ranged from less than 1 to 2.72 but none were statistically significant.
Grice et al. (2007) also looked for association between PFOS exposure the incidence of cancer
at the 3M Decatur plant and found that prostate, melanoma and colon cancer were the most frequently
reported malignancies, but none reached statistical significance. In the general population, Eriksen et
al. (2009) compared plasma levels of PFOS to the incidence of cancer in 57,000 Danish individuals
and found no statistically significant trends.
As discussed in Section 4.4.3., in animals, the induction of peroxisome proliferation has been
suggested to be the mode of action (MOA) for the tumors observed with PFOS. Evidence of this
MOA can be identified by an increase in the number of peroxisomes, increases in the activity of CoA
oxidase activity and hepatic cell proliferation as described by Rao and Reddy, (1996) and Ashby et al.,
(1994). While a number of short-term studies in rats and mice (Sohlenius et al., 1993; Ikeda et al.,
1987; 3M Company, 2004) have shown that PFOS is capable of inducing peroxisome proliferation,
longer-term studies in monkeys and rats have not (Seacat et al., 2002; 2003; Thomford, 2002).
Positive evidence for peroxisome proliferation was found in mice administered a dietary level
of 500 ppm PFOS for 5 days (Sohlenius et al., 1993) and in rats given 200 ppm in the diet for two
weeks (Ikeda et al., 1987). In contrast, no increases in hepatic cell proliferation (a precursor for tumor
development) were detected in a rat subchronic study (Seacat et al., 2003), the cancer bioassay in rats
(Thomford, 2002), or in monkeys administered PFOS for six months (Seacat et al., 2002). Also, in
studies by Wolf et al. (2008 and 2012) and Ishibashi et al. (2011), peroxisome proliferation was
activated less by the perfluorinated sulfonates, including PFOS, when compared to the perfluorinated
carboxylates and mouse PPARa was more activated than human suggesting possible involvement of
other MO As. In contrast to PFOA, PFOS was poorly correlated with peroxisome proliferators in a
study of global PPARa gene expression patterns following exposure (Martin et al., 2007). The weight
of evidence for the carcinogenic potential to humans of these tumors was judged to be too limited to
support a quantitative cancer assessment.
Perfluorooctane sulfonate (PFOS) - February 2014 5-28
Draft - Do Not Cite or Quote
-------
6.0 REFERENCES
3M Company. 2000. Determination of serum half-lives of several fluorochemicals, June 8, 2000,
3M Company. FYI-0700-1378, 8(e) Supplemental Submission, 8EHQ-0373/0374. (as
cited in OECD 2002)
3M Company. 2004. Comparative molecular biology of perfluorooctanesulfonate (PFOS, T-
6295); N-Ethyl perfluorooctanesulfonamido ethanol (N-EtFOSE, T-6316); N-ethyl-
perfluorooctanesulfonamide (N-EtFOSA, T-6868); perfluorooctanesulfonamido acetate
(FOSAA, T-7071); and/or perfluorooctanesulfonamide (FOSA, T-7132) in rats and
guinea pigs following oral dosing. Final Report, July 16, 2004, AR-226-1813.
3M Environmental Laboratory. 1999. Analytical laboratory report on the determination of the
presence and concentration of potassium perfluorooctanesulfonate (CAS No. 2795-39-3)
in the serum of Sprague-Dawley rats exposed to potassium perfluorooctanesulfonate via
gavage, Laboratory Report No. U2779, Requestor Project No. 3M Tox 6295.13. Study
initiation date: June 10, 1999. Completion at signing (2/11/00). Sample analysis
completion June 28, 1999. Study identification number: FACT Tox-108. (as cited in
OECD 2002)
3M Environmental Laboratory. 200 la. Oral (gavage) pharmacokinetic study of PFOS in rats,
analytical laboratory report, determination of the presence and concentration of
perfluorooctanesulfonate (PFOS) in serum, liver, urine and feces samples. 3M Medical
Department Study T-6295.12, Argus In-Life Study #418-013, Analytical Report FACT
TOX-110, 3M Laboratory Request No. U2849, May 4, 2001. (as cited in OECD 2002)
3M Environmental Laboratory. 2001b. Oral (gavage) pharmacokinetic recovery study of PFOS
in rats, analytical laboratory report, determination of the concentration of
perfluorooctanesulfonate (PFOS) in the serum, liver, urine and feces of Crl:CDBR
VAF/Plus rats exposed to PFOS via gavage, 3M Environmental Laboratory Report No.
TOX-111, Laboratory Request No. U2994, 3M Ref.No.T-6295.14, May 4, 2001. (as
cited in OECD 2002)
Abbott, B.D., C.J. Wolf, K.P. Das, R.D. Zehr, J.E. Schmid, A.B. Lindstrom, M.J. Strynar, and C.
Lau. 2009. Developmental toxicity of perfluorooctane sulfonate (PFOS) is not dependent
of expression of peroxisome proliferator activated receptor-alpha (PPARa) in the mouse.
Reprod. Toxicol. 27: 258-265.
Alexander, B.H., G.W. Olsen, J.M. Burris, J.H. Mandel, and J.S. Mandel. 2003. Mortality of
employees of a perfluorooctanesulfonyl fluoride manufacturing facility. Occup. Environ.
Med. 60:722-729.
Alexander, B.H. and G.W. Olsen. 2007. Bladder cancer in perfluorooctanesulfonyl fluoride
manufacturing workers. Ann. Epidemiol. 17:471-478. (as cited in Olsen et al. 2009)
Perfluorooctane sulfonate (PFOS) - February 2014 6-1
Draft - Do Not Cite or Quote
-------
Andersen, M., HJ. Clewell, Y.-M. Tan, J.L. Butenhoff, and G.W. Olsen. 2006. Pharmacokinetic
modeling of saturable, renal absorption of perfluoroalkylacids in monkeys- probing the
determinants of long plasma half-lives. Toxicology 227:156-164.
Andersen, M.E., J.L. Butenhoff, S.-C. Chang, D.G. Farrar, G.L. Kennedy, C. Lau, G.W. Olsen,
and K.B. Wallace. 2008. Review: Perfluoroalkyl acids and related chemistries-
toxicokinetics and modes of action. Tox. Sci. 102:3-14.
Apelberg, B.J., L.R. Goldman, A.M. Calafat, J.B. Herbstman, Z. Kuklenyik, J. Heidler, L.L.
Needham, R.U. Halden, and F.R. Witter. 2007. Determinants of fetal exposure to
polyfluoroalkyl compounds in Baltimore, Maryland. Environ. Sci. Technol. 41:3891-
3897.
Ashby, J., A. Brady, C.R. Elcombe, B.M. Elliot, J. Ishmael, J. Odum, J.D. Tugwood, S. Kettle,
and I.F.H. Purchase. 1994. Mechanistically-based human hazard assessment of
peroxisome proliferator-induced hepatocarcinogenesis. Hum Exp Toxicol 13 (Suppl. 2),
S1-S117.
Best, N.G., M.K. Cowles, and K. Vines. 1995. CODA: convergence diagnosis and output
analysis for Gibbs sampling output.
Biesemeier, J.A. and D.L. Harris. 1974. Eye and skin irritation report on sample T-l 117. Report.
Project No. 4102871, WARF Institute Inc. (as cited in OECD 2002)
Bijland, S., P.C.N. Rensen, EJ. Pieterman, A.C.E. Maas, J.W. vanderHoorn, MJ. vanErk, L.M.
Havekes, K.W. van Dijk, S.-C. Chang, EJ. Ehresman, J.L. Butenhoff, and H.M.G.
Princen. 2011. Perfluoroalkyl sulfonates cause alkyl chain length-dependent hepatic
steatosis and hypolipidemia mainly by impairing lipoprotein production in APOE*3-
Leiden CETP mice. Toxicol. Sci. 123:290-303.
Bjork, J.A. and K.B. Wallace. 2009. Structure-activity relationships and human relevance for
perfluoroalkyl acid-induced transcriptional activation of peroxisome proliferation in liver
cell cultures. Toxicol. Sci. 111:89-99.
Bloom, M.S., K. Kannan, H.M. Spliethoff, L. Tao, K.M. Aldous, and J.E. Vena. 2010.
Exploratory assessment of perfluorinated compounds and human thyroid function.
Physiol. andBehav. 99:240-245.
Bogdanska, J., D. Borg, M. Sunstrom, U. Bergstrom, K. Halldin, M. Abedi-Valugerdi, A.
Bergman, B. Nelson, J. DePierre, and S. Nobel. 2011. Tissue distribution of 35S-labelled
perfluorooctane sulfonate in adult mice after oral exposure to a low environmentally
relevant dose or a high experimental dose. Toxicology 284:54-62.
Borg, D., J. Bogdanska, M. Sundstrom, S. Nobel, H. Hakansson, A. Bergman, J.W. DePierre, K.
Halldin, and U. Bergstrom. 2010. Tissue distribution of 35S-labelled perfluorooctane
Perfluorooctane sulfonate (PFOS) - February 2014 6-2
Draft - Do Not Cite or Quote
-------
sulfonate (PFOS) in C57B1/6 mice following late gestational exposure. Reprod. Toxicol.
30:550-557.
Brieger, A., N. Bienefeld, R. Hasan, R. Goerlich, and H. Haase. 2011. Impact of perfluorooctane
sulfonate and perfluorooctanoic acid on human peripheral leukocytes. Toxicology in
Vitro 25(4): 960-968.
Burris, J.M., J.K. Lundberg, G.W. Olsen, C. Simpson and J. Mandel. 2002. Determination of
serum half-lives of several fluorochemicals. Interim Report #2. 3M Medical Dept. (as
cited in OECD 2002)
Butenhoff, J. and S. Chang. 2007. ADME Study of perfluorooctane sulfonate in rats. ST-179
Final Report, 3M Strategic Toxicology Laboratory, 3M Center, 270-3 S-06, St. Paul, MN.
34pp.
Butenhoff, J., G. Costa, C. Elcombe, D. Farrar, K. Hansen, H. Iwai, R. Jung, G. Kennedy, Jr., P.
Lieder, G. Olsen, and P. Thomford. 2002. Toxicity of ammonium perfluorooctanoate in
male Cynomologus monkeys after oral dosing for 6 months. Toxicol. Sci. 69:244-257.
(As cited in Andersen et al., 2006)
Butenhoff, J., G.L. Kennedy, Jr., P.M. Hindliter, P.H. Lieder, R. Jung, K.J. Hansen, G.S.
Gorman, P.E. Noker, and P.J. Thomford. 2004. Pharmacokinetics of perfluorooctanoate
in Cynomologus monkeys. Toxicol. Sci. 82:394-406. (As cited in Andersen et al., 2006)
Butenhoff, J.L., D.J. Ehresman, S.-C. Chang, G.A. Parker, and D.G. Stump. 2009. Gestational
and lactational exposure to potassium perfluorooctanesulfonate (K+PFOS) in rats:
developmental neurotoxicity. Reprod. Toxicol. 27:319-330.
Calafat, A., L.-Y. Wong, Z. Kuklenyik, J.A. Reidy, and L.L. Needham. 2007. Polyfluoroalkyl
chemicals in the U.S. population: data from the National Health and Nutrition
Examination Survey (NHANES) 2003-2004 and comparisons withNHANES 1999-2000.
Environ. Health Perspect. 115:1596-1602.
Cao, J., Y. Lin, L.-H. Guo, A.-Q. Zhang, Y. Wei, and Y. Yang. 2010. Structure-based
investigation on the binding interaction of hydroxylated polybrominated diphenyl ethers
with thyroxine transport proteins. Toxicology 277:20-28.
Chan, E., I. Burstyn, N. Cherry, F. Bamforth, and J.W. Martin. 2011. Perfluorinated acids and
hypothyroxinemia in pregnant women. Environ. Res. 111:559-564.
Chang, S.-C., J.R. Thibodeaux, M.L. Eastvold, D.J. Ehresman, J.A. Bjork, J.W. Froehlich, C.
Lau, R.J. Singh, K.B. Wallace, and J.L. Butenhoff. 2007. Negative bias from analog
methods used in the analysis of free thyroxine in rat serum containing
perfluorooctanesulfonate (PFOS). Toxicology 234:21-33.
Perfluorooctane sulfonate (PFOS) - February 2014 6-3
Draft - Do Not Cite or Quote
-------
Chang, S.-C., J.R. Thibodeaux, M.L. Eastvold, DJ. Ehresman, J.A. Bjork, J.W. Froehlich, C.
Lau, RJ. Singh, K.B. Wallace, and J.L. Butenhoff. 2008. Thyroid hormone status and
pituitary function in adult rats given oral doses of perfluorooctanesulfonate (PFOS).
Toxicology 243:330-339.
Chang, S.-C., DJ. Ehresman, J.A. Bjork, K.B. Wallace, G.A. Parker, D.G. Stump, and J.
Butenhoff. 2009. Gestational and lactational exposure to potassium
perfluorooctanesulfonate (K+PFOS) in rats: toxicokinetics, thyroid hormone status and
related gene expression. Reprod. Toxicol. 27:387-399.
Chang, S.-C., P.E. Noker, G.S. Gorman, SJ. Gibson, J.A. Hart, DJ. Ehresman, and J.L.
Butenhoff. 2012. Comparative pharmacokinetics of perfluorooctanesulfonate (PFOS) in
rats, mice and monkeys. Reprod. Toxicol. 33:428-440.
Chateau-Degat, M.-L., D. Pereg, R. Dallaire, P. Ayotte, S. Dery, and E. Dewailly. 2010. Effects
of perfluorooctanesulfonate exposure on plasma lipid levels in the Inuit population of
Nunavik (Northern Quebec). Environ. Res. 110:710-717.
Chen, Y.-M. and L.-H. Guo. 2009. Fluorescence study on site-specific binding of perfluoroalkyl
acids to human serum albumin. Arch. Toxicol. 83:255-261.
Chen, T., L. Zhang, J.-q. Yue, Z.-q. Lv, W. Xia, Y.-j. Wan, Y.-y. Li, and S.-q. Xu. 2012. Prenatal
PFOS exposure induces oxidative stress and apoptosis in the lung of rat off-spring.
Reprod. Toxicol. 33:538-545.
Christian, M.S., A.M. Hoberman, andR.G. York. 1999. Oral (stomach tube) developmental
toxicity study of PFOS in rabbits. Argus Research Laboratories Inc. Protocol No. 418-
012, January 1999. FYI-0500-01378 (as cited in OECD 2002)
Cifone, M.A. 1999. Unscheduled DNA synthesis in rat liver primary cell cultures with PFOS.
Final Report. Covance Study No. 207840447. Covance Laboratories Inc., Vienna, VA
22182 (as cited in OECD 2002)
Corley, R.A., A.L. Mendrala, F.A. Smith, D.A. Staats, M.L. Gargas, R.B. Conolly, M.E.
Andersen, and R.H. Reitz. 1990. Development of a physiologically based
pharmacokinetic model for chloroform. Toxicol. Appl. Pharmacol. 103:512-527. (As
cited in Andersen, 2006)
Cui, L., Q.-f Zhou, C.-y. Liao, J.-j. Fu, and G.-b. Jiang. 2009. Studies on the toxicological
effects of PFOA and PFOS on rats using histological observation and chemical analysis.
Arch. Environ. Contam. Toxicol. 56:338-349.
Cui, L., C.-y. Liao, Q.-f. Zhou, T.-m. Xia, Z.-j. Yun, and G.-b. Jiang. 2010. Excretion of PFOA
and PFOS in male rats during a subchronic exposure. Arch. Environ. Contam. Toxicol.
58:205-213.
Perfluorooctane sulfonate (PFOS) - February 2014 6-4
Draft - Do Not Cite or Quote
-------
Curran, I, S.L. Hierlihy, V. Listen, P. Pantazopoulos, A. Nunnikhoven, S. Tittlemier, M. Barker,
K. Trick, and G. Bondy. 2008. Altered fatty acid homeostasis and related toxicologic
sequelae in rats exposed to dietary potassium perfluorooctanesulfonate (PFOS). J.
Toxicol. Environ. Health, Part A 71:1526-1541.
Dallaire, R., P. Ayotte, D. Pereg, S. Dery, P. Dumas, E. Langlois, and E. Dewailly. 2009a.
Determinants of plasma concentrations of perfluorooctanesulfonate and brominated
organic compounds inNunavik Inuit adults (Canada). Environ. Sci. Technol. 43:5130-
5136.
Dallaire, R., E. Dewailly, D. Pereg, S. Dery, and P. Ayotte. 2009b. Thyroid function and plasma
concentrations of polyhalogenated compounds in Inuit adults. Environ. Health Perspect.
117:1380-1386.
Dean, W.P., D.C. Jessup, G. Thompson, G. Romig and D. Powell. 1978. Fluorad fluorochemical
surfactant FC-95 acute oral toxicity (LD50) study in rats. Study No. 137-083,
International Research and Development Corporation (as cited in OECD 2002)
Dong, G.-H., Y.-H. Zhang, L. Zheng, W. Liu, Y.-H. Jin, and Q.-C. He. 2009. Chronic effects of
perfluorooctane sulfonate exposure on immunotoxicity in adult male C57BL/6 mice.
Arch. Toxicol. 83:805-815.
Dong, G.-H., K.-Y. Tung, C.-H. Tsai, M.-M. Liu, D. Wang, W. Liu, Y.-H. Jin, W.S. Hsieh, Y.L.
Lee, and P.-C. Chen. 2013. Serum polyfluoroalkyl concentrations, asthma outcomes,
and immunological markers in a case-control study of Taiwanese children. Environ.
Health Perspect. Online 8 January 2013. http://dx.doi.org/10.1289/ehp.1205351
Egeghy, P. and M. Lorber. 2011. An assessment of the exposure of Americans to
perfluorooctane sulfonate: a comparison of estimated intakes with values inferred from
NHANES data. J. Expos. Sci. Environ. Epidemiol 21:150-168.
Elcombe, C.R., B.M. Elcombe, J.R. Foster, S.-C. Chang, D.J. Ehresman, and J.L. Butenhoff
2012. Hepatocellular hypertrophy and cell proliferation in Sprague-Dawley rats from
dietary exposure to potassium perfluorooctanesulfonate results from increased expression
of xenosensor nuclear receptors PPARa and CAR/PXR. Toxicology 293:16-29.
Eriksen, K., M. S0rensen, J.K. McLaughlin, L. Lipworth, A. Tj0nneland, K. Overvad, and O.
Raaschou-Nielsen. 2009. Perfluorooctanoate and perfluorooctanesulfonate plasma levels
and risk of cancer in the general Danish population. J. Nat. Cancer Instit. 101:605-609.
Fei, C., J.K. McLaughlin, R.E. Tarone, and J. Olsen. 2007. Perfluorinated chemicals and fetal
growth: a study within the Danish National Birth Cohort. Environ. Health Perspect
115:1677-1682.
Perfluorooctane sulfonate (PFOS) - February 2014 6-5
Draft - Do Not Cite or Quote
-------
Fei, C., J.K. McLaughlin, R.E. Tarone, and J. Olsen. 2008a. Fetal growth indicators and
perfluorinated chemicals: a study in the Danish National Birth Cohort. Am. J. Epidemiol.
168:66-72.
Fei, C., J.K. McLaughlin, L. Lipworth, and J. Olsen. 2008b. Prenatal exposure to
perfluorooctanoate (PFOA) and perfluorooctane sulfonate (PFOS) and maternally
reported developmental milestones in infancy. Environ. Health Perspect. 116:1391-1395.
Fei, C., J.K. McLaughlin, L. Lipworth, and J. Olsen. 2009. Maternal levels of perfluorinated
chemicals and subfecundity. Human Reprod. 1:1-6.
Fei, C., J.K. McLaughlin, L. Lipworth, and J. Olsen. 2010a. Maternal concentrations of
perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) and duration of
breastfeeding. Scan. J. Work Environ. Health 36:413-421.
Fei, C., J.K. McLaughlin, L. Lipworth, and J. Olsen. 2010b. Prenatal exposure to PFOA and
PFOS and risk of hospitalization for infectious diseases in early childhood. Environ Res
110:773-777.
Frisbee, S.J., A. Shankar, S.S. Knox, K. Steenland, D.A. Savitz, T. Fletcher, and A. Ducatman.
2010. Perfluorooctanoic acid, perfluorooctanesulfonate, and serum lipids in children and
adolescents: results from the C8 health project. Arch. Pediat. Adolescent Med. 164:860-
869.
Fuentes, S., M.T. Colomina, P. Vicens, N. Franco-Pons, and J. Domingo. 2007. Concurrent
exposure to perfluorooctane sulfonate and restraint stress during pregnancy in mice:
effects on post-natal development and behavior of the offspring. Toxicol. Sci. 98:589-
598.
Gallo, V., G. Leonard!, B. Genser, M.-J. Lopez-Espinosa, S.J. Frisbee, L. Karlsson, A.M.
Ducatman, and T. Fletcher. 2012. Serum perfluorooctanoate (PFOA) and
perfluorooctane sulfonate (PFOS) concentrations and liver function biomarkers in a
population with elevated PFOA exposure. Environ. Health Perspect. 120:655-660.
Gelman, A., J.B. Carlin, H.S. Stern, and D.B. Rubin. 2004. Bayesian data analysis, 2nd edition.
Chapman and Hall/CRC.
Genuis, S.J., D. Birkholz, M. Ralitsch and N. Thibault. 2010. Review paper: human
detoxification of perfluorinated compounds. Public Health 124: 367-375.
Goldenthal, E.I., D.C. Jessup, R.G. Geil, and J.S. Mehring. 1978a. Ninety-day subacute rhesus
monkey toxicity study. Study No. 137-092, International Research and Development
Corporation, Mattawan, MI. FYI-0500-1378. (as cited in OECD 2002)
Perfluorooctane sulfonate (PFOS) - February 2014 6-6
Draft - Do Not Cite or Quote
-------
Goldenthal, E.I., D.C. Jessup, R.G. Geil, N.D. Jefferson, and R.J. Arceo. 1978b. Ninety-day
subacute rat study. Study No. 137-085, International Research and Development
Corporation, Mattawan, MI. FYI-0500-1378. (as cited in OECD 2002)
Goldenthal, E.I., D.C. Jessup, R.G. Geil and. J.S. Mehring. 1979. Ninety-day subacute rhesus
monkey toxicity study. Study No. 137-087, International Research and Development
Corporation, Mattawan, MI. FYI-0500-1378. (as cited in OECD 2002)
Gortner, E.G. 1980. Oral teratology study of FC-95 in rats. Safety Evaluation Laboratory and
Riker Laboratories, Inc. Experiment No.: 0680TR0008, December 1980. FYI-0500-
01378. (as cited in OECD 2002)
Grandjean, P., E.W. Andersen, E. Budtz-J0rgensen, F. Nielsen, K. M01bak, P. Weihe, and C.
Heilmann. 2012. Serum vaccine antibody concentrations in children exposed to
perfluorinated compounds. JAMA 307:391-397.
Grasty, R.C., B.E. Grey, C.S. Lau, and J.M. Rogers. 2003. Prenatal window of susceptibility to
perfluorooctane sulfonate-induced neonatal mortality in the Sprague-Dawley rat. Birth
Defects Research (Part B) 68:465-471.
Grasty, R.C., J.A. Bjork, K.B. Wallace, D.C. Wolf, C. Lau, and J.M. Rogers. 2005. Effects of
prenatal perfluorooctane sulfonate exposure on lung maturation in the perinatal rat. Birth
Defects Research (Part B) 74:405-416.
Grice, M., B. Alexander, R. Hoffbeck, and D. Kampa. 2007. Self-reported medical conditions in
perfluorooctanesulfonyl fluoride manufacturing workers. J. Occupat. Environ. Med.
49:722-729.
Guruge, K.S., H. Hikono, N. Shimada, K. Murakami, J. Hasegawa, L.W.Y. Yeung, N.
Yamanaka, and N. Yamashita. 2009. Effect of perfluorooctane sulfonate (PFOS) on
influenza A virus-induced mortality in female B6C3F1 mice. The J. Toxicol. Sci. 34:687-
691.
Hamm, M., N.M. Cherry, E. Chan, J. Martin, and I. Burstyn. 2009. Maternal exposure to
perfluorinated acids and fetal growth. J. Expos. Sci. Environ. Epidemiol. 1-9.
Harada, K.H., S. Hashida, T. Kaneko, K. Takenaka, M. Minata, K. Inoue, N. Saito, and A.
Koizumi. 2007. Biliary excretion and cerebrospinal fluid partition of perfluoroctanoate
and perfluorooctane sulfonate in humans. Environ. Toxicol. Pharmacol. 24:134-139.
Haug, L.S., C. Thomsen, and G. Becher. 2009. Time trends and the influence of age and gender
on serum concentrations of perfluorinated compounds in archived human samples.
Environ. Sci. Technol. 43:2131-2136.
Perfluorooctane sulfonate (PFOS) - February 2014 6-7
Draft - Do Not Cite or Quote
-------
Heidelberger, P. and P.D. Welch. 1983. Simulation run length control in the presence of an
initial transient. Opns. Res. 31:1109-1144.
HSDB (Hazardous Substances Data Bank). 2009. Perfluorooctane sulfate. U.S. National
Library of Medicine, Bethedsa, MD. Available from: http://toxnet.nlm.nih.gov/cgi-
bin/sis/htmlgen?HSDB
Hu, W., P.D. Jones, B.L. Upham, I.E. Trosko, C. Lau, and J.P. Giesy. 2002 Inhibition of gap
junctional intercellular communication by perfluorinated compounds in rat liver and
dolphin kidney epithelial cell lines in vitro and Sprague-Dawley rats in vivo. Toxicol.
Sci. 68:429-436.
Ikeda, T., K. Fukuda, I. Mori, M. Enmoto, T. Komai, and T. Suga. 1987. Induction of
cytochrome P-450 and peroxisome proliferation in rat liver by perfluorinated
octanesulfonic acid. In : Peroxisomes in Biology and Medicine. (H.D. Fahimi and H.
Sies, Eds.), Springer Verlag, New York, 304-308.
Inoue, K., F. Okada, R. Ito, S. Kato, S. Sasaki, S. Nakajima, A. Uno, Y. Saijo, F. Sata, Y.
Yoshimura, R. Kishi, and H. Nakazawa. 2004. Perfluorooctane sulfonate (PFOS) and
related perfluorinated compounds in human maternal and cord blood samples: assessment
of PFOS exposure in a susceptible population during pregnancy. Environ. Health
Perspect. 112:1204-1207.
Ishibashi, H., E.-Y. Kim, and H. Iwata. 2011. Transactivation potencies of the Baikal seal (Pusa
sibiricd) peroxisome proliferator-activated receptor a by perfluoroalkyl carboxylates and
sulfonates: estimation of PFOA induction equivalency factors. Environ. Sci. Technol.
45:3123-3130.
Joensen, U.N., R. Bossi, H. Leffers, A.A. Jensen, N. Skakkebaek, andN. J0rgensen. 2009. Do
perfluoroalkyl compounds impair human semen quality? Environ. Health Perspect
117:923-927.
Johansson, N., A. Fredriksson, and P. Eriksson. 2008. Neonatal exposure to perfluorooctane
sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural defects in
adult mice. Neurotoxicol. 29:160-169.
Johansson, N., P. Eriksson, and H. Viberg. 2009. Neonatal exposure to PFOS and PFOA in mice
results in changes in proteins which are important for neuronal growth and
synaptogenesis in the developing brain. Toxicol. Sci 108:412-418.
Karrman, A., J.L. Domingo, X. Llebaria, M. Nadal, E. Bigas, B. van Bavel, and G. Lindstrom.
2010. Biomonitoring perfluorinated compounds in Catalonia, Spain: concentrations and
trends for human liver and milk samples. Environ. Sci. Pollut. Res. 17:750-758.
Perfluorooctane sulfonate (PFOS) - February 2014 6-8
Draft - Do Not Cite or Quote
-------
Kato, K., L.-Y. Wong, L.T. Jia, Z. Kuklenyik, and A.M. Calafat. 2011. Trends in exposure to
polyfluoroalkyl chemicals in the U.S. population: 1999-2008. Environ. Sci. Tech.
45:8037-8045.
Keil, D., T. Mehlmann, L. Butterworth, and M. Peden-Adams. 2008. Gestational exposure to
perfluorooctane sulfonate suppresses immune function in B6C3F1 mice. Toxicol. Sci.
103:77-85.
Kerstner-Wood, C., L. Coward, and G. Gorman. 2003. Protein binding of perfluorohexane
sulfonate, perfluorooctane sulfonate and perfluorooctanoate to plasma (human, rat and
monkey) and various human-derived plasma protein fractions. Southern Research
Institute, Study ID 9921.7.
Kim, S.-K., K.T. Lee, C.S. Kang, L. Tao, K. Kannan, K.-R. Kim, C.-K. Kim, J.S. Lee, P.S. Park,
Y.W. Yoo, J.Y. Ha, Y.-S. Shin, and J.-H. Lee. 2011. Distribution of perfluorochemicals
between sera and milk from the same mothers and implications for prenatal and postnatal
exposure. Environ. Pollut. 159:169-174.
Knox, S.S., T. Jackson, B. Javins, SJ. Frisbee, A. Shankar, and A.M. Ducatman. 2011.
Implications of early menopause in women exposed to perfluorocarbons. J. Clin.
Endocrin. Metab. 96:1747-1753.
Kudo, N. 2006. Mechanisms for bioaccumulation of perfluorinated detergents and their risk
assessment. Advan. Pharmaceut. Sci. 22:21-26.
Kuklenyik, Z., J.A. Reich, J.S. Tully, L.L. Needham, and A.M. Calafat. 2004. Automated solid-
phase extraction and measurement of perfluorinated organic acids and amides in human
serum and milk. Environ. Sci. Technol. 38:3698-3704.
Lau, C., J.R. Thibodeaux, R.G. Hanson, J.M. Rogers, B.E. Grey, M.E. Stanton, J.L. Butenhoff,
and L.A. Stevenson. 2003. Exposure to perfluorooctane sulfonate during pregnancy in rat
and mouse. II: postnatal evaluation. Toxicol. Sci. 74:382-392.
Lau, C., K. Anitole, C. Hodes, D. Lai, A. Pfahles-Hutchens, and J. Seed. 2007. Perfluoroalkyl
acids: a review of monitoring and toxicological findings. Toxicol. Sci. 99:366-394.
Liao, C., T. Wang, L. Cui, Q. Zhou, S. Duan, and G. Jiang. 2009. Changes in synaptic
transmission, calcium current, and neurite growth by perfluorinated compounds are
dependent on the chain length and functional group. Environ. Sci. Technol. 43:2099-
2104.
Lin, C.-Y., P.-C. Chen, Y.-C. Lin, and L.-Y. Lin. 2009. Association among serum perfluoroalkyl
chemicals, glucose homeostasis, and metabolic syndrome in adolescents and adults.
Diabetes Care 32:702-707.
Perfluorooctane sulfonate (PFOS) - February 2014 6-9
Draft - Do Not Cite or Quote
-------
Litton Bionetics, Inc. 1979. Mutagenicity evaluation of T-2014 CoC in the Ames
Salmonella/Microsome plate test. Final Report. LBI Project No. 20838 (as cited in OECD
2002)
Liu, L., W. Liu, J. Song, H. Yu, Y. Jin, K. Oami, I. Sato, N. Saito, and S. Tsuda. 2009. A
comparative study on oxidative damage and distributions of perfluorooctane sulfonate
(PFOS) in mice at different postnatal developmental stages. The J. Toxicol. Sci. 34:245-
254.
Loccisano, A.E., J.L. Campbell, M.E. Andersen, andHJ. Clewell. 2011. Evaluation and
prediction of pharmacokinetics of PFOA and PFOS in the monkey and human using a
PBPK model. Reg. Toxicol. Pharmacol. 59:157-175.
Loccisano, A.E., J.L. Campbell, J.L. Butenhoff, M.E. Andersen, andHJ. Clewell. 2012a.
Comparison and evaluation of pharmacokinetics of PFOA and PFOS in the adult rat
using a physiologically based pharmacokinetic model. Reprod. Toxicol. 33:452-467.
Loccisano, A.E., J.L. Campbell, J.L. Butenhoff, M.E. Andersen, andHJ. Clewell. 2012b.
Evaluation of placental and lactational pharmacokinetics of PFOA and PFOS in the
pregnant, lactating, fetal and neonatal rat using a physiologically based pharmacokinetic
model. Reprod. Toxicol. 33:468-490.
Lopez-Espinosa, M.-J., T. Fletcher, B. Armstrong, B. Genser, K. Dhatariya, D. Mondal, A.
Ducatman, and G. Leonard. 2011. Association of perfluorooctanoic acid (PFOA) and
perfluorooctane sulfonate (PFOS) with age of puberty among children living near a
chemical plant. Environ. Sci. Tech. 45:8160-8166.
Lou, L, J.F. Wambaugh, C. Lau, R.G. Hanson, A.B. Lindstrom, MJ. Strynear, R.D. Zehr, R.W.
Setzer, and H.A. Barton 2009. Modeling single and repeated dose pharmacokinetics of
PFOA in mice. Toxicol. Sci. 107:331-341.
Luebker, DJ. KJ. Hansen, N.M. Bass, J.L. Butenhoff, and A.M. Seacat. 2002. Interactions of
fluorochemicals with rat liver fatty acid-binding protein. Toxicology 176:175-185.
Luebker, D.J., R.G. York, K J. Hansen, J.A. Moore, and J.L. Butenhoff. 2005a. Neonatal
mortality from in utero exposure to perfluorooctanesulfonate (PFOS) in Sprague-Dawley
rats: dose-response and biochemical and pharmacokinetic parameters. Toxicology
215:149-169.
Luebker, D.J., M.T. Case, R.G. York, J.A. Moore, K J. Hansen, and J.L. Butenhoff. 2005b. Two-
generation reproduction and cross-foster studies of perfluorooctanesulfonate (PFOS) in
rats. Toxicology 215:126-148.
Maloney, E.K. and DJ. Waxman. 1999. trans-Activation of PPARa and PPARy by structurally
diverse environmental chemicals. Toxicol. Appl. Pharmacol. 161:209-218.
Perfluorooctane sulfonate (PFOS) - February 2014 6-10
Draft - Do Not Cite or Quote
-------
Mandel J. and R. Johnson. 1995. Mortality study of employees at 3M plant in Decatur, Alabama.
Minneapolis: Division of Environmental and Occupational Health, School of Public
Health, University of Minnesota, (as cited in OECD 2002)
Martin, M.T., RJ. Brennan, W. Hu, E. Ayanoglu, C. Lau, H. Ren, C.R. Wood, J.C. Gorton, RJ.
Kavlock, and DJ. Dix. 2007. Toxicogenomic study of triazole fungicides and
perfluoroalkyl acids in rat livers predict toxicity and categorizes chemicals based on
mechanisms of toxicity. Toxicol. Sci. 97:595-613.
Mecchi, M.S. 1999. Salmonella- Escherichia colil- Mammalian-microsome reverse mutation
assay with PFOS. Final Report. Covance Study No.: 20784-0-409. Covance Laboratories,
Vienna VA 22182 (as cited in OECD 2002)
Medinsky, M.A. and C.D. Klaassen. 1996. Toxicokinetics. In: Cassarett andDoull's
Toxicology The Basic Science of Poisons, 5th Ed., C.D. Klaassen, ed. pp. 187-198.
Melzer, D., N. Rice, M. Depledge, W. Henley, and T. Galloway. 2010. Association between
serum perfluorooctanoic acid (PFOA) and thyroid disease in the NHANES study.
Environ. Health Perspect. 118:686-692.
Miller, R.T., L.A. Scappino, S.M. Long, and J.C. Gorton. 2001. Role of thyroid hormones in
hepatic effects of peroxisome proliferators. Toxicol. Pathol. 29, 149-155.
Monroy, R., K. Morrison, K. Teo, S. Atkinson, C. Kubwabo, B. Stewart, and W. Foster. 2008.
Serum levels of perfluoroalkyl compounds in human maternal and umbilical cord blood
samples. Environ. Res. 108:56-62.
Murli, H. 1996. Mutagenicity test on T-6295 in an in-vivo mouse micronucleus assay. Final
Report. CHV Study No.: 17403-0-455. Corning Hazelton Inc. (CHV) Vienna, VA 22182.
(as cited in OECD 2002)
Murli, H. 1999. Chromosomal aberrations in human whole blood lymphocytes with PFOS. Final
Report. Covance Study No.: 2784-0-499. Covance Laboratories Inc., Vienna, VA 22182.
(as cited in OECD 2002)
Nelson, J.W., E.E. Hatch, and T.F. Webster. 2010. Exposure to polyfluoroalkyl chemicals and
cholesterol, body weight, and insulin resistance in the general US population. Environ.
Health Perspect. 118:197-202.
Noker, P. and G. Gorman. 2003. A pharmacokinetic study of potassium perfluorooctanesulfonate
in the Cynomolgus monkey. In Southern Research Institute Study ID: 9921.6. USEPA
public docket, administrative record AR-226-1356. (As cited in Andersen et al., 2006)
OECD (Organization for Economic Co-operation and Development). 2002.
ENV/JM/RD(2002)17/FINAL: Hazard assessment of perfluorooctane sulfonate (PFOS)
Perfluorooctane sulfonate (PFOS) - February 2014 6-11
Draft - Do Not Cite or Quote
-------
and its salts. Report of the Environmental Directorate, Joint Meeting of the Chemicals
Committee and the Working Party on Chemicals, Pesticides and Biotechnology, Co-
operation on Existing Chemicals, Paris, November 21, 2002
Okada, E., S. Sasaki, Y. Saijo, N. Washino, C. Miyashita, S. Kobayashi, K. Konishi, Y.M. Ito, R.
Ito, A., Nakata, Y. Iwasaki, K. Saito, H. Nakazawa, and R. Kishi. 2012. Prenatal
exposure to perfluorinated chemicals and relationship with allergies and infectious
diseases in infants. Environ. Res. 112:118-125.
Olsen, G.W., P.W. Logan, C.A. Simpson, KJ. Hansen, J.M. Burris, M.M. Burlew, J.C.
Schumpert, J.H. Mandel. 1999. Fluorochemical exposure assessment of Decatur chemical
and film plant employees. Final Report, 3M Medical Department. FYI-0500-01378,
August 11, 1999. (As cited in OECD 2002)
Olsen, G.W., P.W. Logan, C.A. Simpson, J.M. Burris, M.M. Burlew, J.K. Lundberg, and J.H.
Mandel. 2001a. Descriptive summary of serum fluorochemical levels among employee
participants of the year 2000 Decatur fluorochemical medical surveillance program. Final
Report. March 19, 2001. (as cited in OECD 2002)
Olsen, G.W., M.M. Burlew, J.M. Burris, and J.H. Mandel. 2001b. A cross-sectional analysis of
serum perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA) in relation to
clinical chemistry, thyroid hormone, hematology and urinalysis results from male and
female employee participants of the 2000 Antwerp and Decatur fluorochemical medical
surveillance program. Final Report. 3M Medical Department, (as cited in OECD 2002)
Olsen, G.W. M.M. Burlew, J.M. Burris, and J.H. Mandel. 2001c. A longitudinal analysis of
serum perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA) levels in
relation to lipid and hepatic clinical chemistry test results from male employee
participants of the 1994/95, 1997 and 2000 fluorochemical medical surveillance program.
3M Final Report, (as cited in OECD 2002)
Olsen, G.W., J.M. Burris, J.K. Lundberg, K.J. Hansen, J.H. Mandel, and L.R. Zobel. 2002a.
Identification of fluorochemicals in human sera. I. American Red Cross adult blood
donors. Final Report; February 25, 2002. (as cited in OECD 2002)
Olsen, G.W., J.M. Burris, J.K. Lundberg, K.J. Hansen, J.H. Mandel, and L.R. Zobel. 2002b.
Identification of fluorochemical s in human sera. III. Pediatric participants in a Group A
Streptococci clinical trial investigation. Final Report; 3M Company, Medical Dept. (as
cited in OECD 2002)
Olsen, G., K. Hansen, L. Stevenson, J. Burris, and J. Mandel. 2003. Human donor liver and
serum concentrations of perfluorooctanesulfonate and other perfluorochemicals. Environ.
Sci. Technol. 37:888-891.
Perfluorooctane sulfonate (PFOS) - February 2014 6-12
Draft - Do Not Cite or Quote
-------
Olsen, G.W., J.M. Burns, DJ. Ehresman, J.W. Froehlich, A.M. Seacat, J.L. Butenhoff, and L.R.
Zobel. 2007. Half-life of serum elimination of perfluorooctanesulfonate,
perfluorohexanesulfonate and perfluorooctanoate in retired fluorochemical production
workers. Environ. Health Perspect. 115:1298-1305.
Olsen, G.W., D.C. Mair, T.R. Church, M.E. Ellefson, W.K. Reagen, T.M. Boyd, R.M. Herron, Z.
Medhdizadehkashi, J.B. Nobiletti, J.A. Rios, J.L. Butenhoff, and L.R. Zobel. 2008.
Decline in perfluorooctane sulfonate and other polyfluoroalkyl chemicals in American
Red Cross adult blood donors, 2000-2006. Environ. Sci. Technol. 42:4989-4995.
Olsen, G.W., J.L. Butenhoff, and L.R. Zobel. 2009. Perfluoroalkyl chemicals and human fetal
development: an epidemiologic review with clinical and toxicological perspectives.
Reprod. Toxicol. 27:212-230.
Peden-Adams, M.M., J.M. Keller, J.G. EuDaly, J. Berger, G.S. Gilkeson, and D.E. Keil. 2008.
Suppression of humoral immunity in mice following exposure to perfluorooctane
sulfonate. Toxicol. Sci. 104:144-154.
Pirali, B., S. Negri, S. Chytiris, A. Perissi, L. Villani, L. La Manna, D. Cottica, M. Ferrari, M.
Imbriani, M. Rotondi, and L. Chiovato. 2009. Perfluorooctane sulfonate and
perfluorooctanoic acid in surgical thyroid specimens of patients with thyroid disease.
Thyroid 19:1407-1412.
Qazi, M.R., Z. Xia, J. Bogdanska, S.-C. Chang, D.J. Ehresman, J.L. Butenhoff, B.D. Nelson,
J.W. DePierre, and M. Abedi-Valugerdi. 2009a. The atrophy and changes in the cellular
compositions of the thymus and spleen observed in mice subjected to short-term exposure
to perfluorooctane sulfonate are high-dose phenomena mediated in part by peroxisome
proliferator-activated receptor-alpha (PPARa). Toxicology 260:68-76.
Qazi, M.R., J. Bogdanska, J.L. Butenhoff, B.D. Nelson, J.W. DePierre, and M. Abedi-Valugerdi.
2009b. High-dose, short-term exposure of mice to perfluorooctanesulfonate (PFOS) or
perfluorooctanoate (PFOA) affects the number of circulating neutrophils differently, but
enhances the inflammatory responses of macrophages to lipopolysaccharide (LPS) in a
similar fashion. Toxicology 262:207-214.
Qazi, M.R., M.R. Abedi, B.D. Nelson, J.W. DePierre, and M. Abedi-Valugerdi. 2010. Dietary
exposure to perfluorooctanate or perfluorooctane sulfonate induces hypertrophy in
centrilobular hepatocytes and alters the hepatic immune status in mice. Int.
Immunopharmacol. 10:1420-1427.
R Development Core Team. 2010. R: a language and environment for statistical computing,
ISBN 3-900051-07-0. http://www.R-proiect.org
Rao, M.S. and J.K. Reddy. 1996. Hepatocarcinogenesis of the peroxisome proliferators. Ann.
N.Y. Acad. Sci. 804:573.
Perfluorooctane sulfonate (PFOS) - February 2014 6-13
Draft - Do Not Cite or Quote
-------
Rosen, M.B., I.E. Schmid, K.P. Das, C.R. Wood, R.D. Zehr, and C. Lau. 2009. Gene expression
profiling in the liver and lung of perfluorooctane sulfonate-exposed mouse fetuses:
comparison to changes induced by exposure to perfluorooctanoic acid. Reprod. Toxicol.
27:278-288.
Rosen, M.B., J.R. Schmid, J.C. Gorton, R.D. Zehr, K.P. Das, B.D. Abbott, and C. Lau. 2010.
Gene expression profiling in wild-type and PPARa-null mice exposed to Perfluorooctane
sulfonate reveals PPARa-independent effects. PPAR Res. 2010, Article ID 794739, 23
pp. doi: 10.1155/2010/794739.
Rusch, G.M., W.E. Rinehart, and C.A. Bozak. 1979. An acute inhalation toxicity study of T-
2306 CoC in the rat. Project No. 78-7185. Bio/dynamics Inc. (as cited in OECD 2002)
Salvalaglio, M., I. Muscionico, and C. Cavallotti. 2010. Determination of energies and sites of
binding of PFOA and PFOS to human serum. J. Phys. Chem. 114:14860-14874.
Sato, I, K. Kawamoto, Y. Nishikawa, S. Tsuda, M. Yoshida, K. Yaegashi, N. Saito, W. Liu, and
Y. Jin. 2009. Neurotoxicity of perfluorooctane sulfonate (PFOS) in rats and mice after
single oral exposure. The J. Toxicol. Sci. 34:569-574.
Seacat, A.M., P.J. Thomford, KJ. Hansen, G.W. Olsen, M.T. Case, and J.L. Butenhoff. 2002.
Subchronic toxicity studies on perfluorooctanesulfonate potassium salt in cynomolgus
monkeys. Toxicol. Sci. 68:249-264.
Seacat, A.M., P.J. Thomford, K.J. Hansen, L.A. Clemen, S.R. Eldridge, C.R. Elcombe, and J.L.
Butenhoff. 2003. Sub-chronic dietary toxicity of potassium perfluorooctanesulfonate in
rats. Toxicology 183:117-131.
Shipley, J.M., C.H. Hurst, S.S. Tanaka, F.L. DeRoos, J.L. Butenhoff, A.M. Seacat, and D.
Waxman. 2004. Trans-activation of PPARa and induction of PPARa target genes by
perfluorooctane-based chemicals. Toxicol. Sci. 80:151-160.
Simmon, V.F. 1978. In-vitro microbiological mutagenicity assays of 3M company compounds
T-2247 CoC and T-2248 CoC. Final Report. SRI Project: LSC-4442-016. SRI
International, Menlo Park, CA 94025. (as cited in OECD 2002)
Slotkin, T., E. MacKillop, R. Melnick, K. Thayer, and F. Seidler. 2008. Developmental
neurotoxicity of perfluorinated chemicals modeled in vitro. Environ. Health Perspect.
116:716-722.
So, M.K., N. Yamashita, S. Taniyasu, Q. Jiang, J.P. Giesy, K. Chen, and P.K.S. Lam. 2006.
Health risks in infants associated with exposure to perfluorinated compounds in human
breast milk from Zhoushan, China. Environ. Sci. Technol. 40:2924-2929.
Perfluorooctane sulfonate (PFOS) - February 2014 6-14
Draft - Do Not Cite or Quote
-------
Sohlenius, A.-K., A.M. Eriksson, C. Hogstrom, M. Kimland, and J.W. DePierre. 1993.
Perfluorooctane sulfonic acid is a potent inducer of peroxisomal fatty acid-p oxidation
and other activities known to be affected by peroxisome proliferators in mouse liver.
Pharmacol. Toxicol. 72:90-93.
Spliethoff, H.M., L. Tao, S.M. Shaver, K.M. Aldous, K.A. Pass, K. Kannan, and G.A. Eadon.
2008. Use of newborn screening program blood spots for exposure assessment: declining
levels of perfluorinated compounds in New York state infants. Environ. Sci. Technol.
42:5361-5367.
Starkov, A.A. andK.B. Wallace. 2002. Structural determinations of fluorochemical-induced
mitochondrial dysfunction. Toxicol. Sci. 66:244-252.
Steenland, K., S. Tinker, S. Frisbee, A. Ducatman, and V. Vaccarino. 2009. Association of
perfluorooctanoic acid and perfluorooctane sulfonate with serum lipids among adults
living near a chemical plant. Am. J. Epidemiol. 170:1268-1278.
Steenland, K., S. Tinker, A. Shankar, and A. Ducatman. 2010. Association of perfluorooctanoic
acid (PFOA) and perfluorooctane sulfonate (PFOS) with uric acid among adults with
elevated community exposure to PFOA. Environ. Health Perspect. 118:229-233.
Stein, C.R., D.A. Savitz, and M. Dougan. 2009. Serum levels of perfluorooctanoic acid and
perfluorooctane sulfonate and pregnancy outcome. Am. J. Epidemiol 170:837-846.
Stein, C.R., M.S. Wolff, A.M. Calafat, K. Kato, and S.M. Engel. 2012. Comparison of
polyfluororalkyl compound concentration in maternal serum and amniotic fluid: a pilot
study. Reprod. Toxicol. 34:312-316.
Takacs, M.L. and B.D. Abbott. 2007. Activation of mouse and human peroxisome proliferator-
activated receptors (a ,P/5,y) by perfluorooctanoic acid and perfluorooctane sulfonate.
Toxicol. Sci. 95:108-117.
Tan, Y.-M., HJ. Clewell, III, and M.E. Andersen. 2008. Time dependencies in
perfluorooctylacids disposition in rat and monkeys: a kinetic analysis. Toxicol. Lett.
177:38-47.
Tan, F., Y. Jin, W. Liu, X. Quan, J. Chen, and Z. Liang. 2012. Global liver proteome analysis
using iTRAQ labeling quantitative proteomic technology to reveal biomarkers in mice
exposed to perfluorooctane sulfonate (PFOS). Environ. Sci. Technol. 46:12170-12177.
Tao, L., J. Ma, T. Kunisue, E.L. Libelo, S. Tanabe, and K. Kannan. 2008. Perfluorinated
compounds in human breast milk from several Asian countries and in infant formula and
dairy milk from the United States. Environ. Sci. Technol. 42:8597-8602.
Perfluorooctane sulfonate (PFOS) - February 2014 6-15
Draft - Do Not Cite or Quote
-------
Thibodeaux, J.R., R.G. Hanson, J.M. Rogers, B.E. Grey, B.D. Barbee, J.H. Richards, J.L.
Butenhoff, L.A. Stevenson, and C. Lau. 2003. Exposure to perfluorooctane sulfonate
during pregnancy in rat and mouse. I: maternal and prenatal evaluations. Toxicol. Sci.
74:369-381.
Thomford, PJ. 2002. 104-week dietary chronic toxicity and carcinogenicity study with
perfluorooctane sulfonic acid potassium salt (PFOS; T-6295) in rats. Final Report, 3M T-
6295 (Covance Study No. 6329-183), Volumes I-IX, 4068 pages, January 2, 2002. 3M,
St. Paul, MN.
Thompson, I, M. Lorber, L.-M.L. Toms, K. Kato, A.M. Calafat, and J.F. Mueller. 2010. Use of
pharmacokinetic modeling to characterize exposure of Australians to perfluorooctanoic
acid and perfluorooctane sulfonate. Environ. Int. 36:392-397.
U.S. EPA (United States Environmental Protection Agency). 1986a. Guidelines for the health
risk assessment of chemical mixtures. Fed. Reg. 51(185):34014-34025.
U.S. EPA (United States Environmental Protection Agency). 1986b. Guidelines for mutagenicity
risk assessment. Fed. Reg. 51(185):34006-34012.
U.S. EPA (United States Environmental Protection Agency). 1988. Recommendations for and
documentation of biological values for use in risk assessment. EPA 600/6-87/008.
Available from: National Technical Information Service, Springfield, VA; PB88-
179874/AS.
U.S. EPA (United States Environmental Protection Agency). 1991. Guidelines for developmental
toxicity risk assessment. Fed. Reg. 56(234):63798-63826.
U.S. EPA (United States Environmental Protection Agency). 1994a. Interim policy for particle
size and limit concentration issues in inhalation toxicity studies. Fed. Reg.
59(206):53799.
U.S. EPA (United States Environmental Protection Agency). 1994b. Methods for derivation of
inhalation reference concentrations and application of inhalation dosimetry. EPA/600/8-
90/066F. Available from: National Technical Information Service, Springfield, VA;
PB2000-500023, and http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 1995. Use of the benchmark dose
approach in health risk assessment. U.S. Environmental Protection Agency. EPA/630/R-
94/007. Available from: National Technical Information Service, Springfield, VA; PB95-
213765, and http://www.epa.gov/iris/backgrd.html
Perfluorooctane sulfonate (PFOS) - February 2014 6-16
Draft - Do Not Cite or Quote
-------
U.S. EPA (United States Environmental Protection Agency). 1996. Guidelines for reproductive
toxicity risk assessment. Fed. Reg. 61(212):56274-56322. Available from:
http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 1998. Guidelines for neurotoxicity
risk assessment. Fed Reg 63(93):26926-26954. Available from:
http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2000a. Science Policy Council
Handbook: peer review. 2nd edition. Prepared by the Office of Science Policy, Office of
Research and Development, Washington, DC. EPA 100-B-OO-OOl. Available from:
http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2000b. Benchmark dose technical
guidance document [external review draft]. EPA/630/R-00/001. Available from:
http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2000c. Supplemental guidance for
conducting for health risk assessment of chemical mixtures. EPA/630/R-00/002.
Available from: http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2002. A review of the reference
dose and reference concentration processes. Risk Assessment Forum, Washington, DC;
EPA/630/P-02/0002F. Available from: http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2005a. Guidelines for carcinogen
risk assessment. Risk Assessment Forum, Washington, DC; EPA/630/P-03/001B.
Available from: http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2005b.Supplemental guidance for
assessing susceptibility from early-life exposure to carcinogens. Risk Assessment Forum,
Washington, DC; EPA/630/R-03/003F. Available from:
http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2006a. Peer review workshop
report on "Draft framework for assessing health risks of environmental exposures to
children". National Center for Environmental Assessment, Washington, DC; EPA/600/R-
06/123. Available from: http://www.epa.gov/iris/backgrd.html
U.S. EPA (United States Environmental Protection Agency). 2006b. A framework for assessing
health risks of environmental exposures to children. National Center for Environmental
Perfluorooctane sulfonate (PFOS) - February 2014 6-17
Draft - Do Not Cite or Quote
-------
Assessment, Washington, DC; EPA/600/R-05/093F. Available from:
http://www.epa.gov/iris/backgrd.html
Volkel, W., O. Genzel Boroviczeny, H. Demmelmair, C. Gebauer, B. Koletzko, D. Twardella, U.
Raab, and H. Fromme. 2008. Perfluorooctane sulphonate (PFOS) and perfluorooctanoic
acid (PFOA) in human breast milk: results of a pilot study. Int. J. Hyg. Environ. Health
211:440-446.
Von Ehrenstein, O., S. Fenton, K. Kato, Z. Kuklenyik, A. Calafat, and E. Hines. 2009.
Polyfluoroalkyl chemicals in the serum and milk of breast-feeding women. Reproduct.
Toxicol. 27:239-245.
Wambaugh, J., H.A. Barton, and R.W. Setzer. 2008. Comparing models for perfluorooctanoic
acid pharmacokinetics using Bayesian analysis. J. Pharmacokinet. Pharmacodyn.
35:683-712.
Wang, F., W. Liu, Y. Jin, J. Dai, W. Yu, X. Liu, and L. Liu. 2010. Transcriptional effects of
prenatal and neonatal exposure to PFOS in developing rat brain. Environ. Sci. Technol.
44:1847-1853.
Washino, N., Y. Saijo, S. Sasaki, S. Kato, S. Ban, K. Koishi, R. Ito, A. Nakata, Y. Iwasaki, K.
Saito, H. Nakazawa, and R. Kishi. 2009. Correlations between prenatal exposure to
perfluorinated chemicals and reduced fetal growth. Environ. Health Perspect. 117:660-
667.
Weiss, J.M., P.L. Andersson, M.H. Lamoree, P.E.G. Leonards, S.P.J. van Leeuwen, and T.
Hamers. 2009. Competitive binding of poly-and perfluorinated compounds to the thyroid
hormone transport protein transthyretin. Toxicol. Sci. 109:206-216.
Wetzel, L.T. 1983. Rat teratology study, T-3351, final report. Hazelton Laboratories America,
Inc. Project No. 154-160, December 19, 1983 (8EHQ-0399-374) (as cited in OECD
2002)
Wiese, S., K.A. Reidegeld, H.E. Meyer, and B. Wardheild. 2007, Protein labeling by iTRAQ: A
new tool for quantitative mass spectrometry in proteome research. Proteomics Feb 7
(3):340-350.
Wilhelm, M., J. Angerer, H. Fromme, and J. Holzer. 2009. Contribution to the evaluation of
reference values for PFOA and PFOS in plasma of children and adults from Germany.
Int. J. Hyg. Environ. Health 212:56-60.
Wolf, C.J., M.L. Takacs, I.E. Schmid, C. Lau, and B.D. Abbott. 2008. Activation of mouse and
human peroxisome proliferator- activated receptor alpha by perfluoroalkyl acids of
different functional groups and chain lengths. Toxicol. Sci. 106:162-171.
Perfluorooctane sulfonate (PFOS) - February 2014 6-18
Draft - Do Not Cite or Quote
-------
Wolf, C., J. Schmid, C. Lau, and B. Abbott. 2012. Activation of mouse and human peroxisome
proliferator- activated receptor-alpha (PPARa) by perfluoroalkyl acids (PFAAs); further
investigation of C4-C12 compounds. Reprod. Toxicol. 33:546-551.
Xie, W., I. Kania-Korwel, P.M. Bummer, and H.-J. Lehmler. 2007. Effect of potassium
perfluorooctanesulfonate, perfluorooctanoate and octanesulfonate on the phase transition
of dipalmitoylphosphatidycholine (DPPC) bilayers. Biocim. Biophys. Acta. 1768:1299-
1808.
Xie, W., G.D. Bothun, and H.-J. Lehmler. 2010a. Partitioning of perfluorooctanoate into
phosphatidylcholine bilayers is chain length-independent. Chem. Phys. Lipids 163:300-
308.
Xie, W., G. Ludewig, K. Wang, and H.-J. Lehmler. 2010b. Model and cell membrane
partitioning of perfluorooctanesulfonate is independent of the lipid chain length. Colloids
and Surfaces B: Biointerfaces 76:128-136.
Yahia, D., C. Tsukuba, M. Yoshida, I. Sato, and S. Tsuda. 2008. Neonatal death of mice treated
with perfluorooctane sulfonate. The J. Toxicol. Sci. 33:219-226.
Yang, X., L. Wang, W. Sun, Z. Xue et al. 2009. [Effects of perfluorooctane sulfonate on amino
acid neurotransmitters and glutamine synthetase in rats.] J. Hyg. Res. 38:19-21. (Chinese
article with translated abstract and ORNL colleague translation)
Ye, L., B. Zhao, K. Yuan, Y. Chu, C. Li, C. Zhao, Q.-Q. Lian, and R.-S. Ge. 2012. Gene
expression profiling in fetal rat lung during gestational perfluorooctane sulfonate
exposure. Toxicol. Lett. 209:270-276.
Yu, W.-G, W. Liu, Y.-H. Jin, X.-H. Liu, F.-Q. Wang, L. Liu, and S.F. Nakayama. 2009a.
Prenatal and postnatal impact of perfluorooctane sulfonate (PFOS) on rat development: a
cross-foster study on chemical burden and thyroid hormone system. Environ. Sci.
Technol. 43:8416-8422.
Yu, W.-G., W. Liu, and Y.-H. Jin. 2009b. Effects of perfluorooctane sulfonate on rat thyroid
hormone biosynthesis and metabolism. Environ. Toxicol. Chem. 28:990-996.
Yu, W.-G., W. Lu, L. Liu, and Y.-H. Jin. 2011. Perfluorooctane sulfonate increased hepatic
expression of OAPT2 and MRP2 in rats. Arch. Toxicol. 85:613-621.
Zeng, H.-c., L. Zhang, Y.-y. Li, Y.-j. Wang, W. Xia, Y. Lin, J. Wei, and S.-q. Xu. 2011.
Inflammation-like glial response in rat brain induced by prenatal PFOS exposure.
Neurotoxicol. 32:130-139.
Perfluorooctane sulfonate (PFOS) - February 2014 6-19
Draft - Do Not Cite or Quote
-------
Zhang, X., L. Chen, X.-C. Fei, Y.-S. Ma, and H.-W. Gao. 2009. Binding of PFOS to serum
albumin and DNA: insight into the molecular toxicity of perfluorochemicals. BMC
Molecular Biology 10:16.
Zheng, L., G.-H. Dong, Y.-H. Jin, and Q.-C. He. 2009. Immunotoxic changes associated with a
7-day oral exposure to perfluorooctane sulfonate (PFOS) in adult male C57BL/6 mice. Arch.
Toxicol. 83:679-689.
Perfluorooctane sulfonate (PFOS) - February 2014 6-20
Draft - Do Not Cite or Quote
-------
APPENDIX A: Summary of Data
Perfluorooctane sulfonate (PFOS) - February 2014 A-l
Draft - Do Not Cite or Quote
-------
TABLE A.l. PFOS Toxicokinetic Information
Species
human1
human
human
human
human
human6
monkey
monkey7
rat8
rat8
Dose
NA
NA
NA
NA
NA
NA
0.15
mg/kg/day
for 26 weeks
with 52
week
recovery
0.75
mg/kg/day
for 26 weeks
with 52
week
recovery
0.03
mg/kg/day
for 104
weeks
0.4
mg/kg/day
Route of
exposure
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
capsule
capsule
diet
diet
Effects
observed
NA
None
observed
t total
cholesterol
t total
cholesterol
t total
cholesterol
t incidence of
thyroid
disease
None
observed
t liver wt
I cholesterol
and body wt
None
observed
t liver
histopath.
PFOS in liver
Gig/g)
M
F
0.19
NS
NS
NS
NS
NS
NS
NS
wkO: 11.0
wk 10: 23.8
wk 105: 7.83
wk 0:47.6
0.27
NS
NS
NS
NS
NS
NS
wk 0:8.71
wk 10: 19.2
wk 105:
12.9
wk 0:83.0
PFOS in blood
(ppm)
M
(seru
0.015
NS
F
m)
ppm
NS
0.022
0.023
0.023
0.037
(serum)
wk 1:4.60
wk 35: 84.5
wk79: 19.1
(serum)
wk 1:21.0
wk35: 181
wk 79: 41.1
(serum)
wk 0:0.091
wk 14: 4.04
wk!05: 1.31
(serum)
wk 0:7.57
0.018
NS
(serum)
wk 1:3.71
wk 35: 74.7
wk 79: 21.4
(serum)
wk 1:20.4
wk35: 171
wk 79: 41.4
(serum)
wkO: 1.61
wk 14: 6.96
wk 105: 4.35
(serum)
wkO: 12.6
PFOS in brain
Gig/g)
M
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
F
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-2
-------
TABLE A.l. PFOS Toxicokinetic Information
Species
rat8
rat9
rat9
rat9
rat10
(m only)
rat10
(m only)
rat11
rat11
Dose
for 104
weeks
1.5
mg/kg/day
for 104
weeks
2 mg/kg diet
for 28 days
20 mg/kg
diet for 28
days
100 mg/kg
diet for 28
days
5 mg/kg for
28 days
20 mg/kg for
28 days
0.4 mg/kg-
42 days
prior to
cohabitation
through GD
21
1.6 mg/kg-
42 days
prior to
Route of
exposure
diet
diet
diet
diet
drinking
water
drinking
water
oral
gavage
oral
gavage
Effects
observed
lesions
t body and
liver wt
t
hepatocellular
adenoma
None
observed
|T4 and t
liver wt.
| body wt, T4>
T3and
cholesterol
t hepatocyte
hypertrophy
J, body wt
10/10 died
(day 26)
hepatic
hypertrophy
PFOS in liver
(ng/g)
M
wk 10: 358
wk 105: 70.5
wk 0:282
wk 10: 568
wk 105: 189
48.28
560.23
1030.40
345
648
NS
NS
F
wk 10: 370
wk 105:
131
wk 0: 373
wk 10: 635
wk 105:
381
43.44
716.55
1008.59
NS
NS
GD21:
dams = 107
fetuses =
30.6
GD21:
dams = 388
fetuses =
PFOS in blood
(ppm)
M
wk 14: 43.9
wk 105: 22.5
(serum)
wk 0:41.8
wk 14: 148
wk 105: 69.3
(serum)
0.95 ug/g serum
(serum)
13.45 ug/g serum
(serum)
29.88 ug/g serum
(whole blood)
72.0
(whole blood)
NS
NS
NS
F
wk 14: 64.4
wk 105: 75.0
(serum)
wk 0:54.0
wk 14: 223
wk 105: 233
(serum)
1.50 ug/g serum
(serum)
15.40 ug/g serum
(serum)
43.20 ug/g serum
NS
NS
(serum)
GD 1: 8.90
GD 7: 7.83
GD21:
dams = 26.2
fetuses = 34.3
(serum)
GD 1: 160
GD 7: 154
PFOS in brain
(ng/g)
M
NS
NS
NS
NS
NS
NS
NS
NS
F
NS
NS
NS
NS
NS
NS
NS
NS
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-3
-------
TABLE A.l. PFOS Toxicokinetic Information
Species
rat11
rat12
rat12
Dose
cohabitation
through GD
21
3.2mg/kg-
42 days
prior to
cohabitation
through GD
21
0.1 mg/kg-
GDOto
PND20
1.0 mg/kg-
GDOto
PND20
Route of
exposure
oral
gavage
oral gavage
oral gavage
Effects
observed
None
observed in
dams or
offspring
t motor
activity and J,
habituation in
male
offspring
PFOS in liver
Gig/g)
M
NS
PND21:
Offspring =
5.98
PND72:
Offspring =
0.98
PND21:
Offspring =
44.89
PND72:
Offspring =
7.17
F
86.5
GD21:
dams = 610
fetuses =
230
GD20:
Dams =
8.35
Offspring =
3.21
PND21:
Dams = NS
Offspring =
5.28
PND72:
Dams = NS
Offspring =
0.80
GD20:
Dams =
48.88
Offspring =
20.03
PND21:
Dams = NS
Offspring =
41.23
PND72:
Dams = NS
Offspring =
PFOS in blood
(ppm)
M
NS
(serum)
PND21:
Offspring = 1.73
PND72:
Offspring = 0.04
(serum)
PND21:
Offspring = 18.61
PND72:
Offspring = 0.56
F
GD21:
dams = 136
fetuses =101
(serum)
GD 1:318
GD 7: 306
GD21:
dams = 155
fetuses =164
(serum)
GD20:
Dams= 1.72
Offspring =3. 91
PND21:
Dams = 3.16
Offspring = 1.77
PND72:
Dams = NS
Offspring = 0.21
(serum)
GD20:
Dams = 26.63
Offspring =3 1.46
PND21:
Dams = 30.48
Offspring = 18.01
PND72:
Dams = NS
Offspring = 1.99
PFOS in brain
Gig/g)
M
NS
PND21:
Offspring
= 0.22
PND21:
Offspring
= 2.62
F
NS
GD20:
Dams =
0.15
Offspring
= 1.23
PND21:
Dams =
NS
Offspring
= 0.23
GD20:
Dams =
0.99
Offspring
= 12.98
PND21:
Dams =
NS
Offspring
= 2.70
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-4
-------
TABLE A.l. PFOS Toxicokinetic Information
Species
mouse13
Dose
50 mg/kg
one time
Route of
exposure
SQ injection
Effects
observed
t liver wt.
and signs of
oxidative
damage
PFOS in liver
M
PND 7: 15%
PND 21:
51%
PND 35:
74%
F
7.2
PND 7:
16%
PND 21:
52%
PND 35:
70%
PFOS in blood
(ppm)
M
PND 7: 12%
PND 21: 10%
PND 35: 13%
F
PND 7: 11%
PND 21: 12%
PND 35: 12%
PFOS in brain
M
PND 7:
5%
PND 21:
2%
PND 35:
1%
F
PND 7:
4%
PND 21:
2%
PND 35:
1%
NS = no sample obtained or recorded; NA = not applicable
'Olsen et al. 2003 6 Melzer et al. 2010
2Karrman et al. 2010 7Seacat et al. 2002
3Steenland et al. 2009 8Thomford 2002
9Curran et al. 2008
"Luebkeretal. 2005a
12Chang et al. 2009
13Liu et al. 2009
4Frisbee et al. 2010
5Nelsonetal. 2010
10,
'Cui et al. 2009
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-5
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Species/Strain
Exposure
duration
Dose
(mg/kg/day)
PFOS concentration
fag/mL)
M
F
NOAEL
(mg/kg)
LOAEL
(mg/kg)
Critical Effect
LIVER EFFECTS
Curran et al.
2008
Seacat et al.
2003
Seacat et al.
2002
Rat/Sprague-Dawley
15/sex/group
diet
Rat/ Crl:CD(SD) IGS
BR
5/sex/group
diet
Monkey /Cynomolgus
6/sex/group
capsule
28 days
98 days
(14 weeks)
182 days
(26 weeks)
followed by
52 week
recovery
0
M:0.14;F:
0.15
M: 1.33; F:
1.43
M:3.21;F:
3.73
M:6.34;F:
7.58
0
0.035
0.14
0.35
1.4
0
serum (values
expressed as
ug/g serum)
0.47
0.95
13.45
20.93
29.88
serum
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Thomford,
2002
Species/Strain
Rat/Crl:CD
(SD)IGS BR
diet
Exposure
duration
104 weeks
Dose
(mg/kg/day)
0.03
0.15
0.75
0
0.018-0.023
0.072-0.099
PFOS concentration
fag/mL)
M
wk 35: 0.05
wk 79: 0.02
wk 1:0.869
wk!6: 11.2
wk27: 15.9
wk35:NS
wk 79: ND
wk 1:4.60
wk 16: 56.2
wk 27: 68.1
wk 35: 84.5
wk79: 19.1
wk 1:21.0
wk 16: 189
wk 27: 194
wk35: 181
wk 79: 41.1
serum
wkl:
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Species/Strain
Exposure
duration
Dose
(mg/kg/day)
0.184-0.247
0.765-1.1
PFOS concentration
fag/mL)
M
wk53:NS
wk 105: 7.60
wk 1:7.57
wk 14: 43.9
wk53:NS
wk 105: 22.5
wk 1:41.8
wk 14: 148
wk53: 146
wk 105: 69.3
F
wk53:NS
wk 105: NS
wkl: 12.6
wk 14: 64.4
wk53:NS
wk 105: 75.0
wkl: 54.0
wk 14: 223
wk53:220
wk 105: 233
NOAEL
(mg/kg)
LOAEL
(mg/kg)
Critical Effect
THYROID EFFECTS
Chang et al.
2008
Curran et al.
2008
Rat/Sprague-Dawley
5-15 females/group
oral gavage
Rat/Sprague-Dawley
15/sex/group
diet
Single dose
28 days
0
15
0
M:0.14;F:
0.15
No males
dosed
serum (values
expressed as
ug/g serum)
0.47
0.95
13.45
serum
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Species/Strain
Exposure
duration
Dose
(mg/kg/day)
M: 1.33; F:
1.43
M:3.21;F:
3.73
M:6.34;F:
7.58
PFOS concentration
fag/mL)
M
20.93
29.88
F
15.40
31.93
43.20
NOAEL
(mg/kg)
LOAEL
(mg/kg)
Critical Effect
DEVELOPMENTAL EFFECTS
Butenhoff et
al. 2009 and
Chang et al.
2009
Rat/Sprague-Dawley
25 females/group
oral gavage
GD 0- PND
20
Offspring
monitored
through PND
72
0
0.1
serum
PND 21:
M Offspring
= 1 mg/kg
Maternal:
1.0
Male
offspring: 1.0
Female
Offspring: ND
Maternal: J, body wt
Male offspring:
based on J,
habituation response
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-9
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Thibodeaux et
al. 2003 and
Species/Strain
Rat/Sprague-Dawley
16-25 females/group
Exposure
duration
CDs 2-20
Dose
(mg/kg/day)
0.3
1.0
PFOS concentration
fag/mL)
M
PND21:
M Offspring
= 5.05
PND72:
M Offspring
= 0.12
PND21:
M Offspring
= 18.61
PND72:
M Offspring
= 0.56
F
Dams = NA
F Offspring
= 0.21
GD20:
Dams = 6.25
Fetus =
10.45
PND21:
Dams = 8.98
F Offspring
= 5.25
PND72:
Dams = NA
F Offspring
= 0.56
GD20:
Dams =
26.63
Fetus =
31.46
PND21:
Dams =
30.48
F Offspring
= 18.01
PND72:
Dams = NA
F Offspring
= 1.99
serum on GD 21
NOAEL
(mg/kg)
Maternal
= 1
LOAEL
(mg/kg)
Maternal= 2
Critical Effect
Maternal: J, body wt
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-10
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Lau et al.
2003
Species/Strain
oral gavage
Exposure
duration
Dose
(mg/kg/day)
0
1
2
o
J
5
10
PFOS concentration
fag/mL)
M
F
Dam: 0.24
Newborn: 0.188
Dam: 19.6
Newborn: 35.9
Dam: 45.0
Newborn: 71.9
Dam: 71.9
Newborn: 86.5
Dam: 80.6
Newborn: 108.2
Dam: 189.9
Newborn: NS
NOAEL
(mg/kg)
Develop
mental
= 1
LOAEL
(mg/kg)
Developmental
= 2
Critical Effect
Developmental:
based on J, survival
and developmental
delays
BMDL5
corresponding to
maternal dose for
survival of rat pups
at PND 8 was 0.58
mg/kg
BMDL5
corresponding to
fetal sternal defects
was 0.1 2 and cleft
palates was 3.33
mg/kg
REPRODUCTIVE
Luebker et al.
2005a and
2005b
Rat/Crl:CD (SD)IGS
VAF/Plus
oral gavage
6 wks prior
to mating
through
gestation and
lactation
across 2
generations
Due to high
number of
pups dying,
only 0,0.1
and 0.4
administered
to 2nd
0
0.1
0.4
NS
serum
NS
Dams
GD 1: 8.90
GD 7: 7.83
GD 15: 8.81
GD 21: 4.52
Fetus GD 21:
9.08
Dams
GD 1:40.7
GD 7: 40.9
FO dams
0.4
Fl pups:
0.4
FO dams 0.8
Fl pups: 0.8
FO dams:
Decreased body wt
Fl pups:
Decreased gestation
length
BMDL5 for
decreased survival of
pups through LD5
was 0.89 mg/kg/day
BMDL5 for pup
weight gain was 0.28
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-ll
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Species/Strain
Exposure
duration
generation
Dose
(mg/kg/day)
1.6
3.2
PFOS concentration
fag/mL)
M
F
GD 15: 41.4
GD 21: 26.2
Fetus GD 21:
34.3
Dams
GD 1: 160
GD 7: 154
GD 15: 156
GD21: 136
Fetus GD 21:
101
Dams
GDI: 318
GD 7: 306
GD 15: 275
GD21: 155
Fetus GD 21:
164
NOAEL
(mg/kg)
LOAEL
(mg/kg)
Critical Effect
mg/kg/day
IMMUNOTOXICITY
Dong et al.
2009
Mouse/B6C3Fl
10 m/dose
oral gavage
60 days
0
0.0083
0.083
0.42
0.83
2.08
serum
0.048
0.674
7.132
21.638
65.426
120.67
NS
0.0083
0.083
Based on f splenic
natural killer cell
activity
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-12
-------
TABLE A.2. Key Studies Used With Effects Related to Serum Values (Condensed Version)
Study
Peden-Adams
et al. 2008
Species/Strain
Mouse/B6C3Fl
5m/5f
oral gavage
Exposure
duration
28 days
Dose
(mg/kg/day)
0
0.0017
0.0017
0.0033
0.017
0.033
0.166
PFOS concentration
fag/mL)
M
serum
0.012
0.018
0.092
0.131
NS
NS
NS
F
serum
0.017
NS
0.088
0.123
0.666
NS
NS
NOAEL
(mg/kg)
Males
0.00017
Females
0.0033
LOAEL
(mg/kg)
Males 0.00 17
Females
0.017
Critical Effect
Males: J, plaque
forming cell
response
Females: J, plaque
forming cell
response
a LOQ or LLOQ are below limits that can be quantified
NS = no sample; NA= not applicable
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-13
-------
TABLE A.3. Summary of Animal Studies with Exposure to PFOS
Method of
exposure
oral gavage
oral gavage
oral
(capsule)
oral gavage
oral gavage
oral gavage
inhalation
oral (in
diet)
oral (in
diet)
oral gavage
oral (in
diet)
Length of
study
20 days
90 days
182 days
single dose
single dose
single dose
thyroid
hormone
activity
1 hour
28 days
28 days
28 days
90 days
Species
monkey
monkey
monkey
rat
rat
rat
rat
rat
rat
rat
rat
Concentration
0, 10, 30, 100 or 300
mg/kg/day
2 monkeys/sex/dose
0,0.5, 1.5 or 4.5
mg/kg/day
2 monkeys/sex/dose
0,0.03, 0.15 or 0.75
mg/kg/day
4-6 monkeys/sex/dose
0,100, 215, 464 or
1000 mg/kg
5 rats/sex/dose
0, 12.5, 25 or 50 mg/kg
5 male rats/dose
0 or 15 mg/kg
5/15 female rats/group
0,1.89,2.86,4.88,6.49,
7.05, 13.9, 24.09 or
45.97 ppm
5 rats/sex/dose
0,0.05,0.18, 0.37 or
1.51 mg/kg/day- males
0, 0.05, 0.22, 0.47 or
1.77mg/kg/day-
females
(0, 0.5, 2, 5 or 20 ppm)
5 rats/sex/dose
0, 2, 20, 50 or 100 mg
PFOS/kg diet
0, 5 or 20 mg/kg/day
10 males/dose
0,2,6, 18, 60 or 200
mg/kg/day
5 rats/sex/dose
Results
NOAEL= NA
LOAEL= 10 mg/kg/day from
deaths at all doses
NOAEL= NA
LOAEL= 0.5 mg/kg/day
based on diarrhea and
anorexia
NOAEL= 0.15 mg/kg/day
LOAEL= 0.75 mg/kg/day
from J, body weight, t liver
wt and jcholesterol
LD50 = 251 mg/kg
(combined)
NOAEL= NA
LOAEL= 12.5 mg/kg based
on I body weight
Total thyroxine (TT4)-
significant J, at 2, 6 and 24
hrs
Triiodothyronine (TT3) and
reverse triiodothyronine
(rT3)- significant J, at 24 hrs
Free thyroxine- significant |
at 2 and 6 hrs; normal at 24
hrs
LC50 = 5.2 ppm
NOAEL = 0.37 mg/kg/day in
males and 0.47 mg/kg/day in
females
LOAEL =1.51 mg/kg/day in
males and 1.77 mg/kg/day in
females, based on J, body wt
(m/f) and J, food
consumption (f)
NOAEL = 2 mg PFOS/kg in
males and NA in females
LOAEL = 20 mg PFOS/kg
in males and 2 mg/kg in
females based on f relative
liver wt
NOAEL= NA
LOAEL= 5 mg/kg/day based
on 1 body wt and lung
congestion
NOAEL= NA
LOAEL= 2 mg/kg/day, from
t liver wt, J, food
consumption
Reference
Goldenthal
etal. 1978a
Goldenthal
et al. 1979
Seacat et al.
2002
Dean et al.
1978
Yang et al.
2009
Chang et al.
2008
Rusch et al.
1979
Seacat et al.
2003
Curran et
al. 2008
Cui et al.
2009
Goldenthal
etal. 1978b
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-14
-------
TABLE A.3. Summary of Animal Studies with Exposure to PFOS
Method of
exposure
oral (in
diet)
oral gavage
oral gavage
oral gavage
oral gavage
oral gavage
Length of
study
98 days
GD 0 to PND
20a
developmental
neurotoxicity
study
GD 6-15
developmental
study
GD2-21
GD2-21
GD 1-21
Species
rat
rat
rat
rat
rat
rat
Concentration
0,0.5, 2.0, 5.0 or 20
ppm
0,0.03, 0.13, 0.34 or
1.33 mg/kg/day- males
0,0.04,0.15, 0.40 or
1.56 mg/kg/day-
females
5 rats/sex/dose
0,0.1, 0.3 or 1.0
mg/kg/day
25 females/dose
0, 1, 5 or 10 mg/kg/day
25 pregnant rats
0, 1, 2, 3, 5 or 10 mg/kg
0,0. 1,0.6 or 2.0 mg/kg
0,0.1, or 2.0 mg/kg/day
Results
NOAEL = 0.34 mg/kg/day in
males and 0.40 mg/kg/day in
females
LOAEL =1.33 mg/kg/day in
males and 1.56 mg/kg/day in
females, based on f liver wt
(m) and t relative liver wt
(m/f)
Maternal
NOAEL= 1 mg/kg/day
LOAEL= NA
Developmental
NOAEL= 0.3 mg/kg/day
LOAEL= 1 mg/kg/day based
on t motor activity
Maternal
NOAEL= 1 mg/kg/day
LOAEL= 5 mg/kg/day based
on clinical signs
Developmental
NOAEL= 1 mg/kg/day
LOAEL= 5 mg/kg/day based
on J, body wt
Maternal
NOAEL= 1 mg/kg
LOAEL= 2 mg/kg based on
J, body wt
Developmental
NOAEL= 1 mg/kg
LOAEL= 2 mg/kg based on
J, survival
BMDL5 corresponding to
maternal dose for survival of
rat pups at PND 8 was 0.58
mg/kg
Offspring
NOAEL = can't be
determined
LOAEL= 0.1 mg/kg based
on changes in the cortex and
hippocampus (astrocyte
activation markers, pro-
inflammatory transcription
factors)
Offspring
NOAEL= 0.1 mg/kg/day
LOAEL = 2.0 mg/kg/day
based on histopathological
Reference
Seacat et al.
2003
Butenhoff
et al. 2009
Wetzel
1983
Thibodeaux
et al. 2003
and Lau et
al. 2003
Zengetal.,
2011
Chenetal.,
2012
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-15
-------
TABLE A.3. Summary of Animal Studies with Exposure to PFOS
Method of
exposure
oral gavage
oral gavage
oral (diet)
oral gavage
Length of
study
6 wks prior to
mating and
Males- 22 days
Females-
through
gestation,
parturition and
lactation
reproductive
study
6 wks prior to
mating and
continued
through mating,
gestation and
lactation day 4
reproductive
study
104 weeks
1 time
developmental
Species
rat
rat
rat
mouse
Concentration
0,0.1,0.4, 1.6 or 3.2
mg/kg/day
35 rats/sex/dose
0,0.4,0.8. 1.0, 1.2, 1.6
and 2.0 mg/kg/day
20-28 dams/dose
0,0.03, 0.1, 0.4 or 1.5
mg/kg/day
0,0.5, 2, 5 or 20 ppm
40-70 males and
females
0,0.75 or 11.3 mg/kg
4-7 males/group
Results
changes in lungs, J, body wt
and t mortality
FO (m/f) parents
NOAEL= 0.1 mg/kg/day
LOAEL= 0.4 mg/kg/day
based on J, bwt gain/food
consumption
Fl (m/f) parents
NOAEL = 0.4 mg/kg
LOAEL = 1.6 mg/kg/day
Fl offspring
NOAEL= 0.4 mg/kg/day
LOAEL= 1.6 mg/kg/day
based on J, viability, body wt
F2 offspring
NOAEL= 0.1 mg/kg/day
LOAEL= 0.4 mg/kg/day
based on J, body wt
FO dams
NOAEL= 0.4 mg/kg/day
LOAEL= 0.8 mg/kg/day
based on J, bwt gain
Fl offspring
NOAEL= not identified
LOAEL= 0.4 mg/kg/day
based on J, pup body weight
BMDL5 estimates for
decreased gestation length
was 0.31 and viability was
0.89 mg/kg/day
Males NOAEL= 0.5 ppm
(0.03 mg/kg)
Males LOAEL= 2 ppm (0.1
mg/kg) based on liver
histopathology
Females NOAEL= 2 ppm
(0.1 mg/kg)
Females LOAEL= 5 ppm
(0.4 mg/kg) based on liver
histopathology
Suggestive of
carcinogenicity
Mice at both concentrations
showed J, activity and t
neuroprotein levels in the
Reference
Luebker et
al., 2005b
Luebker et
al., 2005a
Thomford,
2002
Johansson
et al. 2008
and 2009
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-16
-------
TABLE A.3. Summary of Animal Studies with Exposure to PFOS
Method of
exposure
oral gavage
oral gavage
oral gavage
oral gavage
oral gavage
oral gavage
Length of
study
neurotoxicity
7 days
immunotoxicity
study
GD 1-17
developmental
immunotoxicity
GD 1-17
GD 12-18
developmental
GD 0-17/18
developmental
28 days
immunotoxicity
Species
mouse
mouse
mouse
mouse
mouse
mouse
Concentration
0, 5, 20 or 40 mg/kg
12 male mice/dose
0, 0.1, 1 or 5 mg/kg
10-12 female mice/dose
0, 1,5, 10, 15 or 20
mg/kg
0 or 6 mg/kg/day
8-10 mice/dose
0,1, 10 or 20
mg/kg/day
10 mice/dose
0,0.005,0.05,0.1,0.5,
1, or 5 mg/kg
5 mice/dose
Results
hippocampus
NOAEL= NA
LOAEL= 5 mg/kg based on
t liver wt and suppression of
the plaque forming cell
response
Males NOAEL= 0.1 mg/kg
Males LOAEL= 1 mg/kg
based on J.NK cell activity
Females NOAEL= 1 mg/kg
Females LOAEL= 5 mg/kg
based on J.NK cell activity
Maternal
NOAEL= 1 mg/kg
LOAEL= 5 mg/kg based on
t liver wt
Developmental
NOAEL= 1 mg/kg
LOAEL= 5 mg/kg based on
t liver wt, delayed eye
opening
BMDL5 corresponding to
maternal dose for survival of
mouse pups at PND 6 was
3. 88 mg/kg
Maternal
NOAEL= 6 mg/kg/day
LOAEL= NA
Developmental
NOAEL= NA
LOAEL= 6 mg/kg/day
based on J, body wt
Maternal
NOAEL= 1 mg/kg/day
LOAEL= 10 mg/kg/day,
based on f liver organ wt.
Developmental
NOAEL= 1 mg/kg/day
LOAEL= 10 mg/kg/day,
based on fetal abnormalities
and I survival
Males NOAEL= 0.005
mg/kg
Males LOAEL= 0.05 mg/kg
based on J, plaque forming
cell response
Females NO AEL= 0.1
Reference
Zheng et al.
2009
Keil et al.
2008
Thibodeaux
et al. 2003
and Lau et
al. 2003
Fuentes et
al. 2007
Yahia et al.
2008
Peden-
Adams et
al., 2008
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-17
-------
TABLE A.3. Summary of Animal Studies with Exposure to PFOS
Method of
exposure
oral gavage
dermal
ocular
oral gavage
Length of
study
60 days
immunotoxicity
single dose
single dose
GD 7-20
developmental
Species
mouse
rabbit
rabbit
rabbit
Concentration
0, 0.008, 0.083, 0.417,
0.833 or 2.083 mg/kg
10 male mice/group
0.5 g*
(no data on gender)
0.5 g*
(no data on gender)
0,0.1, 1.0, 2.5 or 3.75
mg/kg/day
22 females/dose
Results
mg/kg
Females LOAEL= 0.5 mg/kg
based on J, plaque forming
cell response
NOAEL= 0.008 mg/kg/day
LOAEL= 0.083 mg/kg based
on t splenic NK cell activity
and t liver weight
No irritation
Exact score not provided
except maximal score at 1
and 24 hrs
Maternal
NOAEL= 0.1 mg/kg/day
LOAEL= 1 mg/kg/day based
on I body wt.
Developmental
NOAEL= 1 mg/kg/day
LOAEL= 2.5 mg/kg/day
based on J, fetal body wt and
I in sternum ossification
Reference
Dong et al.
2009
Biesemeier
and Harris
1974
Biesemeier
and Harris
1974
Christian et
al. 1999
*Exact dose not provided; NA= not applicable; could not be determined
a GD = gestation day and PND = post natal day
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
A-18
-------
APPENDIX B
Contents: Benchmark Dose Modeling Output Files 10% increased incidence of liver lesions and
10% increase in liver weight for:
1) Thomford, 2002: male rat - hepatocellular centrilobular hypertrophy;
2) Thomford, 2002: female rat - hepatocellular centrilobular hypertrophy;
3) Thomford, 2002: male rat - centrilobular hepatocytic vacuolation;
4) Seacat et al., 2002: male monkey - increased liver weight;
5) Seacat et al., 2002: female monkey - increased liver weight;
6) Seacat et al., 2003: male rat - increased liver weight.
Perfluorooctane sulfonate (PFOS) - February 2014 B-1
Draft - Do Not Cite or Quote
-------
1) Thomford, 2002: male rat - hepatocellular centrilobular hypertrophy
Probit Model. (Version: 3.2; Date: 10/28/2009)
Wed Oct 03 14:57:49 2012
BMDS Model Run
The form of the probability function is:
P[response] = Background
+ (1-Background) * CumNorm(Intercept+Slope*Log(Dose) ) ,
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = Incidence
Independent variable = Dose
Slope parameter is not restricted
Total number of observations = 5
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
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0
intercept = 0.551536
slope = 0.619893
Asymptotic Correlation Matrix of Parameter Estimates
(*** The model parameter(s) -background have been estimated at a
boundary point, or have been specified by the user, and do not appear in the
correlation matrix )
intercept
intercept 1
slope 0.79
slope
0.79
1
Variable
Background
Intercept
Slope
Estimate
0
0.600166
0.637003
Std. Err.
NA
0.163276
0.0848112
Parameter Estimates
95.0% Wald Confidence Interval
Lower Conf. Limit Upper Conf. Limit
0.280151
0.470776
0.92018
0.80323
NA - Indicates that this parameter has hit a bound implied by some ineguality
constraint and thus has no standard error.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-2
-------
Model
Full model
Fitted model
Reduced model
AIC:
Analysis of Deviance Table
Log(likelihood)
-102.179
-104.472
-161.64
212.944
# Param's
5
2
1
Deviance Test d.f.
4.58658
118.923
P-value
0.2047
<.0001
Goodness of Fit
Dose
0.0000
0.0180
0.0720
0.1840
0.7650
Est. Prob.
0.0000
0.0251
0.1410
0.3163
0. 6662
Expected
0.000
1.378
7.755
17.395
43.305
Observed
0.000
2.000
4.000
22.000
42.000
Size
65
55
55
55
65
Scaled
Residual
0.000
0.536
-1.455
1.335
-0.343
=4.31
d.f. = 3
P-value = 0.2303
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0521299
BMDL = 0.0326765
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-3
-------
2) Thomford, 2002: female rat - hepatocellular centrilobular hypertrophy;
Probit Model. (Version: 3.2; Date: 10/28/2009)
Tue Jan 29 13:03:16 2013
BMDS Model Run
The form of the probability function is:
P[response] = Background
+ (1-Background) * CumNorm(Intercept+Slope*Log(Dose) ) ,
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = Incidence
Independent variable = Dose
Slope parameter is not restricted
Total number of observations = 5
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
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0.0307692
intercept = 0.921158
slope = 0.875963
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background 1 0.037 0.25
intercept 0.037 1 0.78
slope 0.25 0.78 1
Variable Estimate
Background 0.0242361
Intercept 1.11217
Slope 1.03103
Parameter Estimates
95.0% Wald Confidence Interval
Std. Err. Lower Conf. Limit Upper Conf. Limit
0.0139844 -0.0031728 0.0516449
0.200749 0.718705 1.50563
0.140636 0.755385 1.30667
Model
Full model
Fitted model
Reduced model
Analysis of Deviance Table
Log(likelihood)
-93.9538
-94.0971
-167.248
# Param's
5
3
1
Deviance Test d.f.
0.286555
146.589
P-value
0.8665
<.0001
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-4
-------
AIC: 194.194
0
0
0
0
0
Dose
.0000
.0180
.0720
.1840
.7650
Est
0.
0.
0.
0.
0.
. Prob.
0242
0254
0776
2812
8033
Expe
1.
1.
4,
15,
52,
Gooc
scted
.575
.399
.271
.464
.215
Ines
0
2
1
4
16
52
s of Fit
bserved
.000
.000
.000
.000
.000
Size
65
55
55
55
65
Sc
Res
0.
-0.
-0.
0.
-0.
aled
idual
343
341
136
161
067
ChiA2 = 0.28 d.f. = 2 P-value = 0.8681
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0981083
BMDL = 0.0680339
Perfluorooctane sulfonate (PFOS) - February 2014 B- 5
Draft - Do Not Cite or Quote
-------
3) Thomford, 2002: male rat - centrilobular hepatocytic vacuolation;
Probit Model. (Version: 3.2; Date: 10/28/2009)
Thu Oct 04 08:56:32 2012
BMDS_Model_Run
The form of the probability function is:
P[response] = Background
+ (1-Background) * CumNorm(Intercept+Slope*Log(Dose) ) ,
where CumNorm(.) is the cumulative normal distribution function
Dependent variable = incidence
Independent variable = dose
Slope parameter is not restricted
Total number of observations = 5
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
User has chosen the log transformed model
Default Initial (and Specified) Parameter Values
background = 0.0461538
intercept = -0.349312
slope = 0.465561
Asymptotic Correlation Matrix of Parameter Estimates
background intercept slope
background 1 -0.2 0.33
intercept -0.2 1 0.63
slope 0.33 0.63 1
Parameter Estimates
95.0% Wald Confidence Interval
Variable Estimate Std. Err. Lower Conf. Limit Upper Conf. Limit
Background 0.0409871 0.022081 -0.00229092 0.0842651
Intercept -0.470429 0.188805 -0.84048 -0.100379
Slope 0.341758 0.108985 0.128151 0.555365
Analysis of Deviance Table
Model Log(likelihood) # Param's Deviance Test d.f. P-value
Full model -112.104 5
Fitted model -112.911 3 1.61515 2 0.4459
Reduced model -124.265 1 24.3222 4 <.0001
AIC:
Perfluorooctane sulfonate (PFOS) - February 2014 B- 6
Draft - Do Not Cite or Quote
-------
0
0
0
0
0
Dose
.0000
.0180
.0720
.1840
.7650
Est
0.
0.
0.
0.
0.
. Prob.
0410
0723
1229
1821
3163
Expe
2,
3,
6,
10,
20,
Gooc
scted
.664
.976
.759
.013
.559
Ines
0
3
3
6
13
19
s of Fit
bserved
.000
.000
.000
.000
.000
Size
65
55
55
55
65
S
Re
0
-0
-0
1
-0
caled
sidual
.210
.508
.312
.044
.416
ChiA2 = 1.66
d.f. = 2
P-value = 0.4357
Benchmark Dose Computation
Specified effect = 0.1
Risk Type = Extra risk
Confidence level = 0.95
BMD = 0.0931649
BMDL = 0.0278419
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-7
-------
4) Seacat et al., 2002: male monkey - increased liver weight;
Power Model. (Version: 2.16; Date: 10/28/2009)
Tue Jan 29 09:33:47 2013
BMDS Model Run
The form of the response function is:
Y[dose] = control + slope * dose^power
Dependent variable = Mean
Independent variable = Dose
The power is restricted to be greater than or equal to 1
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 = 6.06377
rho = 0
control = 54.9
slope = 18.8575
power = 0.447473
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter (s) -power have been estimated at a
boundary point, or have been specified by the user, and do not appear in the
correlation matrix )
lalpha
rho
control
slope
lalpha
1
-1
-0.16
0.9
rho
-1
1
0.14
-0.91
control
-0.16
0.14
1
-0.11
slope
0.9
-0.91
-0.11
1
Variable
lalpha
rho
control
slope
power
Estimate
-32.9753
8.99283
58.3909
4773.05
18
Parameter Estimates
95.0% Wald Confidence Interval
Std. Err.
21.1498
5.16599
1.82406
2967.06
NA
Lower Conf. Limit
-74.4282
-1.13233
54.8158
-1042.28
Upper Conf. Limit
8.47765
19.118
61.966
10588.4
NA - Indicates that this parameter has hit a bound implied by some inequality
constraint and thus has no standard error.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B- 8
-------
Table of Data and Estimated Values of Interest
Dose N Obs Mean Est Mean Obs Std Dev Est Std Dev Scaled Res.
0
0.03
0.15
0.75
54.9
62.1
57.3
85.3
58.4
58.4
58.4
85.3
8.1
5.3
5.5
38.4
6.05
6.05
6.05
33.3
-0.999
1.23
-0.361
-7.12e-008
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
Al
A2
A3
fitted
R
Log(likelihood)
-50.652100
-39.523352
-40.563551
-41.316982
-53.538667
# Param's
5
8
6
4
2
AIC
111.304201
95.046703
93.127102
90.633964
111.077334
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
Test 1
Test 2
Test 3
Test 4
-2*log(Likelihood Ratio) Test df
28.0306
22.2575
2.0804
1.50686
p-value
<.0001
<.0001
0.3534
0.4707
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is less than .1.
model appears to be appropriate
A non-homogeneous variance
The p-value for Test 3 is greater than .1. The modeled variance appears
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-9
-------
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 = 0.1
Risk Type = Estimated standard deviations from the control mean
Confidence level = 0.95
BMD = 0. 60743
BMDL = 0.0147931
Perfluorooctane sulfonate (PFOS) - February 2014 B-10
Draft - Do Not Cite or Quote
-------
5) Seacat et al., 2002: female monkey - increased liver weight;
Exponential Model. (Version: 1.7; Date: 12/10/2009)
Tue Jan 29 10:17:21 2013
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^d}]
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.
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))
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
Initial Parameter Values
rho)
Variable
Model 2
Model 3
Model 4
Model 5
Inalpha
rho
a
b
c
d
-7.12771
2.78482
53.6357
0.452798
-7.12771
2.78482
53.6357
0.452798
-7.12771
2.78482
48.545
2.77794
1.6287
-7.12771
2.78482
48.545
2.77794
1.6287
1
Variable
Model 2
Parameter Estimates by Model
Model 3 Model 4
Model 5
Inalpha
rho
a
b
c
d
-1.92335
1.54391
53.6802
0.451454
-2.1108
1.58965
53.8771
0.479691
1.06864
-1.59973
1.46555
53.4554
0.00050114
1083.43
-2.44766
1.6722
54.2396
0.0042198
721.127
1.30686
Table of Stats From Input Data
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-ll
-------
Dose
Obs Mean
Obs Std Dev
0 3
0.03 4
0.15 4
0.75 4
Model Dose
2 0
0.03
0.15
0.75
3 0
0.03
0.15
0.75
4 0
0.03
0.15
0.75
5 0
0.03
0.15
0.75
51.1
56.8
57
75.3
Estimated Values
Est Mean
53.68
54.41
57.44
75.31
53.88
54.46
57.21
75.35
53.46
54.33
57.8
75.2
54.24
54.55
56.82
75.37
9.4
12.6
3.1
13.3
of Interest
Est Std
8.273
8.36
8.717
10.74
8.276
8.347
8. 681
10.8
8.296
8.394
8.785
10.65
8.29
8.33
8. 618
10.92
Scaled Residual
-0.5402
0.5712
-0.1012
-0.002181
-0.5812
0.5608
-0.04889
-0.008551
-0.4918
0.5896
-0.1832
0.01899
-0.656
0.5391
0.04191
-0.01305
Other models for which likelihoods are calculated:
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + log(mean(i)) * rho)
Model R: Yij = Mu + e(i)
Var{e(ij)} = Sigma/x2
Likelihoods of Interest
Model
Log(likelihood)
DF
AIC
Al
A2
A3
R
2
3
4
5
-40.
-36.
-39.
-45.
-40.
-40.
-40.
-40.
,44292
,89929
,91492
,74039
,49219
,49157
,51305
,50101
5
8
6
2
4
5
5
6
90.88584
89.79858
91.82985
95.48078
88.98438
90.98314
91.0261
93.00202
Additive constant for all log-likelihoods = -13.78. This constant added
to the above values gives the log-likelihood including the term that does not
depend on the model parameters.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-12
-------
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 4: Does Model 2 fit the data? (A3 vs. 2)
Test 5a: Does Model 3 fit the data? (A3 vs 3)
Test 5b: Is Model 3 better than Model 2? (3 vs. 2)
Test 6a: Does Model 4 fit the data? (A3 vs 4)
Test 6b: Is Model 4 better than Model 2? (4 vs. 2)
Test la: Does Model 5 fit the data? (A3 vs 5)
Test 7b: Is Model 5 better than Model 3? (5 vs. 3)
Test 7c: Is Model 5 better than Model 4? (5 vs. 4)
Tests of Interest
Test
Test 1
Test 2
Test 3
Test 4
Test 5a
Test 5b
Test 6a
Test 6b
Test 7a
Test 7b
Test 7c
-2*log(Likelihood Ratio) D. F.
17.68 6
7.087 3
6.031 2
1.155 2
1.153 1
0.001241 1
1.196 1
-0.04172 1
1.172 0
-0.01888 1
0.02408 1
p-value
0.007077
0.06917
0.04901
0.5614
0.2829
0.9719
0.2741
N/A
N/A
N/A
0.8767
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 less than .1.
consider a different variance model.
You may want to
The p-value for Test 4 is greater than .1. Model 2 seems
to adeguately describe the data.
The p-value for Test 5a is greater than .1. Model 3 seems
to adeguately describe the data.
The p-value for Test 5b is greater than .05. Model 3 does
not seem to fit the data better than Model 2.
The p-value for Test 6a is greater than .1. Model 4 seems
to adeguately describe the data.
The p-value for Test 6b is less than .05. Model 4 appears
to fit the data better than Model 2.
Degrees of freedom for Test 7a are less than or egual to 0.
The Chi-Sguare test for fit is not valid.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-13
-------
The p-value for Test 7b is less than .05. Model 5 appears
to fit the data better than Model 3.
The p-value for Test 7c is greater than .05. Model 5 does
not seem to fit the data better than Model 4.
Benchmark Dose Computations:
Specified Effect = 0.100000
Risk Type = Estimated standard deviations from control
Confidence Level = 0.950000
BMD and BMDL by Model
Model BMD BMDL
2 0.0338792 0.0207989
3 0.0415751 0.0208011
4 0.0286094 0.0165794
5 0.0629252 0.00584753
Perfluorooctane sulfonate (PFOS) - February 2014 B-14
Draft - Do Not Cite or Quote
-------
6) Seacat et al., 2003: male rat - increased liver weight.
Exponential Model. (Version: 1.7;
Tue Jan 29 11:21:33 2013
Date: 12/10/2009)
BMDS Model Run
Variable
Inalpha
rho(S)
a
b
c
d
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.
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]))
rho is set to 0.
A constant variance model is fit.
Total number of dose groups = 5
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
Initial Parameter Values
Model 2
1.35789
0
15.3355
0.226423
(S) = Specified
Model 3
1.35789
0
15.3355
0.226423
1
Model 4
1.35789
0
13.3
1.66321
1.60263
Model 5
1.35789
0
13.3
1.66321
1.60263
1
Variable
Inalpha
rho
a
b
c
Model 2
1.7164
0
15.4453
0.216725
Parameter Estimates by Model
Model 3 Model 4 Model 5
1.7164
0
15.4453
0.216725
1.64166
0
14.8246
1.73553
1.42089
1.43222
0
15
2.97948
1.35333
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-15
-------
18
Table of Stats From Input Data
Dose N Obs Mean Obs Std Dev
0
0.03
0.13
0.34
1.33
15.5
15.5
14
18.8
20.3
1.1
2.7
1.4
3
2.2
Estimated Values of Interest
Model
2
3
4
5
Dose
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
.03
.13
.34
.33
0
.03
.13
.34
.33
0
.03
.13
.34
.33
0
.03
.13
.34
.33
Est Mean
15.
15.
15.
16.
20.
15.
15.
15.
16.
20.
14.
15.
16.
17.
20.
18
20
45
55
89
63
61
45
55
89
63
61
82
14
08
61
44
15
15
15
.8
.3
Est
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
2,
Std
.359
.359
.359
.359
.359
.359
.359
.359
.359
.359
.272
.272
.272
.272
.272
.046
.046
.046
.046
.046
Scaled Residual
0.
-0.
-1
-0.
0.
-0.
-1
-0.
0.
0.
-2
1
-0.
0.
0.
-1
0518
0437
.788
2.06
2895
0518
0437
.788
2.06
2895
6646
3531
.052
.175
1414
5463
5463
.093
1.225e-007
-6.774e-008
Other models for which likelihoods are calculated:
Model Al: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma/x2
Model A2: Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3: Yij = Mu(i) + e(ij)
Var{e(ij)} = exp(lalpha + log(mean(i)) * rho)
Model R: Yij = Mu + e(i)
Var{e(ij)} = Sigma/x2
Model
Likelihoods of Interest
Log(likelihood) DF
AIC
Al
A2
A3
R
-29.47369
-26.27122
-29.47369
-40.48438
6
10
6
2
70.94737
72.54245
70.94737
84.96876
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-16
-------
-33.95502
-33.95502
-33.02072
-30.40279
73.91004
73.91004
74.04144
70.80558
Additive constant for all log-likelihoods = -22.97. 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 4: Does Model 2 fit the data? (A3 vs. 2)
Test 5a: Does Model 3 fit the data? (A3 vs 3)
Test 5b: Is Model 3 better than Model 2? (3 vs. 2)
Test 6a: Does Model 4 fit the data? (A3 vs 4)
Test 6b: Is Model 4 better than Model 2? (4 vs. 2)
Test 7a: Does Model 5 fit the data? (A3 vs 5)
Test 7b: Is Model 5 better than Model 3? (5 vs. 3)
Test 7c: Is Model 5 better than Model 4? (5 vs. 4)
Tests of Interest
Test
Test 1
Test 2
Test 3
Test 4
Test 5a
Test 5b
Test 6a
Test 6b
Test 7a
Test 7b
Test 7c
-2*log(Likelihood Ratio) D. F.
p-value
28.43 8
6.405 4
6.405 4
8.963 3
8.963 3
-1.421e-014 0
7.094 2
1.869 1
1.858 1
7.104 2
5.236 1
0.0003996
0.1709
0.1709
0.02979
0.02979
N/A
0.02881
0.1716
0.1728
0.02866
0.02213
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.
variance appears to be appropriate here.
The modeled
The p-value for Test 4 is less than .1. Model 2 may not adeguately
describe the data; you may want to consider another model.
The p-value for Test 5a is less than .1. Model 3 may not adeguately
describe the data; you may want to consider another model.
Degrees of freedom for Test 5b are less than or egual to 0.
Perfluorooctane sulfonate (PFOS) - February 2014
Draft - Do Not Cite or Quote
B-17
-------
The Chi-Square test for fit is not valid.
The p-value for Test 6a is less than .1. Model 4 may not adequately
describe the data; you may want to consider another model.
The p-value for Test 6b is qreater than .05. Model 4 does
not seem to fit the data better than Model 2.
The p-value for Test 7a is qreater than .1. Model 5 seems
to adequately describe the data.
The p-value for Test 7b is less than .05. Model 5 appears
to fit the data better than Model 3.
The p-value for Test 7c is less than .05. Model 5 appears
to fit the data better than Model 4.
Benchmark Dose Computations:
Specified Effect = 0.100000
Risk Type = Estimated standard deviations from control
Confidence Level = 0.950000
BMD and BMDL by Model
Model BMD BMDL
2 0.0699374 0.0498937
3 0.0699374 0.0498937
4 0.0213759 0.00939927
5 0.280426 0.0585612
Perfluorooctane sulfonate (PFOS) - February 2014 B-18
Draft - Do Not Cite or Quote
------- |