vvEPA
United States     Office of Water EPA 822-R-16-002
Environmental Protection Mail Code 4304T May 2016
Agency
   Health Effects Support
        Document for
 Perfluorooctane Sulfonate
           (PFOS)
Perfluorooctane sulfonate (PFOS) -May 2016

-------
                       Health Effects Support Document

                                        for

                          Perfluorooctane Sulfonate (PFOS)
                         U.S. Environmental Protection Agency
                               Office of Water (43 04T)
                         Health and Ecological Criteria Division
                               Washington, DC 20460

 http://www.epa.gov/dwstandardsregulations/drinking-water-contaminant-human-health-effects-
                                    information.
                         EPA Document Number: 822-R-16-002
                                     May 2016
Perfluorooctane sulfonate (PFOS) -May 2016

-------
                                    BACKGROUND

    The Safe Drinking Water Act (SDWA), as amended in 1996, requires the Administrator of
the U.S. Environmental Protection Agency (EPA) to establish a list of unregulated
microbiological and chemical contaminants known or anticipated to occur in public water
systems and that might require control in the future through national primary drinking water
regulations. The SDWA also requires the Agency to make regulatory determinations on at least
five contaminants on the Contaminant Candidate List (CCL) every 5 years. For each contaminant
on the CCL, before EPA makes a regulatory determination, the Agency needs to obtain sufficient
data to conduct analyses on the extent to which the contaminant occurs and the risk it poses to
populations via drinking water. Ultimately, this information will assist the Agency in
determining the most appropriate course of action in relation to the contaminant (e.g., developing
a regulation to control it in drinking water, developing guidance, or deciding not to regulate it).

    The PFOS health assessment was initiated by the Office of Water, Office of Science and
Technology in 2009. The draft Health Effects Support Document for Perfluorooctane Sulfonate
Acid (PFOS) was completed in 2013 and released for public comment in February 2014. An
external peer-review panel meeting was held on August 21 and 22, 2014. The final document
reflects input from the panel as well as public comments received on the draft document. Both
the peer-reviewed draft and this document include only the sections of a health effects support
document (HESD) that cover the toxicokinetics and health effects of PFOS.  If a decision is made
to regulate the contaminant, this document will be expanded.

    One of the challenges inherent in conducting this assessment was the wealth of experimental
data published before and during its development. This section provides a synopsis of the
approach used in identifying and selecting the publications reflected in the final assessment.

    Data were identified through the following:
    Monthly/bimonthly literature searches conducted by EPA library staff (2009-2015) and New
Jersey Department of Environmental Protection library staff (2012-2015).

    •   Papers identified by EPA internal and external peer reviewers.
    •   Papers identified through public comments on the draft assessments.
    •   Papers submitted to EPA by the public.

    In mid-2013, the EPA library  searches were expanded to cover other members of the
perfluorocarboxylic acids (C-4 to C-12) and sulfonate families (C-4, C-6, C-8). Appendix A
describes the literature search strategy used by the libraries. Through the literature search,
documents were identified for retrieval, review,  and inclusion in the HESD using the following
criteria:

    •   The study examines a toxicity endpoint or population not examined by studies  already
       included in the draft document.
    •   Aspects of the  study design such as the size of the population exposed or quantification
       approach make it superior to key studies already included in the draft document.
    •   The data contribute substantially to the weight of evidence for any of the toxicity
       endpoints covered by the draft document.
    •   Elements of the study design merit its inclusion in the draft document based on its
       contribution to the mode of action or the quantification approach.
Perfluorooctane sulfonate (PFOS) -May 2016                                                iii

-------
   •   The study elucidates the mode of action for any toxicity endpoint or toxicokinetic
       property associated with PFOS exposure.
   •   The effects observed differ from those in other studies with comparable protocols.

In addition to each publication being evaluated against the criteria above, the relevance of the
study to drinking water exposures and to the U.S. population also were considered.

   The studies included in the final draft were determined to provide the most current and
comprehensive description of the toxicological properties of PFOS and the risk it poses to
humans exposed to it in their drinking water.  Appendix B summarizes the studies evaluated for
inclusion in the FtESD following the August 2014 peer review and identifies those selected for
inclusion in the final assessment. Appendix B includes epidemiology data that provide a high-
level summary of the outcomes across the studies evaluated.

   Development of the hazard identification  and dose-response assessment for PFOS has
followed the  general guidelines for risk assessment forth by the National Research Council
(1983) and EPA's Framework for Human Health Risk Assessment to Inform Decision Making
(USEPA 2014a). Other EPA guidelines used  in the development of this assessment include the
following:
   •   Guidelines for the Health Risk Assessment of Chemical Mixtures (USEPA 1986a)
   •   Guidelines for Mutagenicity Risk Assessment (USEPA 1986b)
   •   Recommendations for and Documentation of Biological Values for Use in Risk
       Assessment (USEPA 1988)
   •   Guidelines for Developmental Toxicity Risk Assessment (USEPA 1991)
   •   Interim Policy for Particle Size and Limit Concentration Issues in Inhalation Toxicity
       Studies (USEPA 1994a)
   •   Methods for Derivation of Inhalation Reference Concentrations and Application of
       Inhalation Dosimetry (USEPA 1994b)
   •   Use of the Benchmark Dose Approach in Health Risk Assessment (USEPA 1995)
   •   Guidelines for Reproductive Toxicity Risk Assessment (USEPA 1996)
   •   Guidelines for Neurotoxicity Risk Assessment (USEPA 1998)
   •   Science Policy Council Handbook: Peer Review (2nd edition) (USEPA 2000a)
   •   Supplemental Guidance for Conducting Health Risk Assessment of Chemical Mixtures
       (USEPA 2000b)
   •   A Review of the Reference Dose and Reference Concentration Processes (USEPA 2002)
   •   Guidelines for Carcinogen Risk Assessment (USEPA 2005a)
   •   Supplemental Guidance for Assessing Susceptibility from Early-Life Exposure to
       Carcinogens (USEPA 2005b)
   •   Science Policy Council Handbook: Peer Review (USEPA 2006a)
   •   A Framework for Assessing Health Risks of Environmental Exposures to Children
       (USEPA 2006b)
   •   Highlights of the Exposure Factors Handbook (USEPA 2011)
   •   Benchmark Dose Technical Guidance Document (USEPA 2012)
   •   Child-Specific Exposure Scenarios Examples (USEPA 2014b)
Perfluorooctane sulfonate (PFOS) -May 2016                                               iv

-------
                  AUTHORS, CONTRIBUTORS, AND REVIEWERS

Joyce Morrissey Donohue, Ph.D. (Chemical Manager)
Office of Water
U.S. Environmental Protection Agency, Washington, D.C.

Amal Mahfouz, Ph.D. (Chemical Manager, pre-retirement).
Office of Water
U.S. Environmental Protection Agency, Washington, D.C.

Tina Moore Duke, M.S. (previously with Office of Water, U.S. Environmental Protection
Agency)

John Wambaugh, Ph.D.
Office of Research and Development
U.S. Environmental Protection Agency, Research Triangle Park, NC

The following contractor authors supported the development of this document:

Dana F. 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

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.

                        CONTRIBUTORS AND REVIEWERS

Internal Contributors and Reviewers

Office of Water, U.S. Environmental Protection Agency
     Elizabeth Doyle, Ph.D. (retired)
     Edward Hackett

Office of Research and Development, U.S. Environmental Protection Agency
     Glinda Cooper, Ph.D.
     Barbara Glenn, Ph.D.
     Erin Hines, Ph.D.
     Christopher Lau, Ph.D.
     Matthew Lorber, Ph.D.
     Jaqueline Moya
     Linda Phillips, Ph.D.
Perfluorooctane sulfonate (PFOS) -May 2016

-------
     Paul White, Ph.D.
     Michael Wright, Sc.D.

Office of Chemical Safety and Pollution Prevention, U.S. Environmental Protection Agency
     E. Laurence Lib el o
     Andrea Pfehales-Hutchens, Ph.D.
     Tracy Williamson
     David Lai, Ph.D. (retired)
     Jennifer Seed, Ph.D. (retired)

Office of Children's Health Protection, U.S. Environmental Protection Agency
     Gregory Miller

Office of Land and Emergency Management, U.S. Environmental Protection Agency

External Reviewers

James Bruckner, Ph.D.
Department of Pharmacology and Toxicology
University of Georgia, Athens, GA

Deborah Cory-Slechta, Ph.D.
Department of Environmental Medicine
University of Rochester Medical Center, Rochester, NY

Jamie DeWitt, Ph.D.
Pharmacology and  Toxicology
East Carolina University, Greenville, NC

Jeffrey Fisher, Ph.D.
Biochemical Toxicology, National Center for Toxicological Research
U.S. Food and Drug Administration, Jefferson, AK

William Hayton, Ph.D.
College of Pharmacy  (Emeritus)
The Ohio State University, Columbus, OH

Matthew Longnecker, M.D., Sc.D.
Biomarker-based Epidemiology Group
National Institute of Environmental Health Sciences, Research Triangle Park, NC

Angela Slitt, Ph.D.
Biomedical and Pharmaceutical Sciences
University of Rhode Island, Kingston,  RI
Perfluorooctane sulfonate (PFOS) -May 2016                                                vi

-------
                                     CONTENTS

BACKGROUND	iii
ABBREVIATIONS AND ACRONYMS	xiii
EXECUTIVE SUMMARY	ES-1
1.   IDENTITY: CHEMICAL AND PHYSICAL PROPERTIES	1-1
2.   TOXICOKINETICS	2-1
  2.1     Absorption	2-1
    2.1.1    Oral Exposure	2-1
    2.1.2    Inhalation Exposure	2-2
    2.1.3    Dermal Exposure	2-2
  2.2     Distribution	2-2
    2.2.1    Oral Exposure	2-4
    2.2.2    Inhalation and Dermal Exposure	2-15
    2.2.3    Other Routes of Exposure	2-15
  2.3     Metabolism	2-16
  2.4     Excretion	2-17
    2.4.1    Oral Exposure	2-17
    2.4.2    Inhalation Exposure	2-19
  2.5     Pharmacokinetic Considerations	2-20
    2.5.1    Pharmacokinetic models	2-20
    2.5.2    Half-life data	2-30
    2.5.3    Volume of Distribution Data	2-34
  2.6     Toxicokinetic Summary	2-36
3.   HAZARD IDENTIFICATION	3-1
  3.1     Human Effects	3-1
    3.1.1    Long-TermNoncancer Epidemiological Studies	3-2
      3.1.1.1   Serum Lipids and Cardiovascular Diseases	3-2
      3.1.1.2   Liver Enzymes and Liver Disease	3-10
      3.1.1.3   Biomarkers of Kidney Function and Kidney Disease	3-11
      3.1.1.4   Reproductive Hormones and Reproductive/Developmental Studies	3-13
      3.1.1.5   Thyroid Effect Studies	3-30
      3.1.1.6   Immunotoxicity	3-36
      3.1.1.7   Other Effects	3-41
      3.1.1.8   Summary and conclusions from the human epidemiology studies	3-41
    3.1.2    Carcinogenicity Studies	3-44
      3.1.2.1   Summary and Conclusions from the Human Cancer Epidemiology Studies ...3-49
  3.2     Animal Studies	3-49
    3.2.1    Acute Toxicity	3-50
    3.2.2    Short-Term Studies	3-51
    3.2.3    Subchronic Studies	3-56
    3.2.4    Neurotoxicity	3-60

Perfluorooctane sulfonate (PFOS) -May 2016                                             vii

-------
    3.2.5    Developmental/Reproductive Toxicity	3-62
    3.2.6    Specialized Developmental/Reproductive Studies	3-73
    3.2.7    Chronic Toxicity	3-78
    3.2.8    Carcinogenicity	3-79
  3.3     Other Key Data	3-81
    3.3.1    Mutagenicity and Genotoxicity	3-81
    3.3.2    Protein binding	3-82
    3.3.3    Immunotoxicity	3-83
    3.3.4    Physiological or Mechanistic Studies of Noncancer Effects	3-89
    3.3.5    Structure-Activity Relationship	3-102
    3.3.6    ToxCast Assays	3-102
  3.4     Hazard Characterization	3-104
    3.4.1    Synthesis and Evaluation of Major Noncancer Effects	3-104
      3.4.1.1   Liver Effects, Cholesterol, and Uric Acid	3-104
      3.4.1.2   Developmental/Reproductive Toxicity	3-107
      3.4.1.3   Immunotoxicity	3-109
      3.4.1.4   Neurotoxicity	3-110
      3.4.1.5   Thyroid Effects	3-111
    3.4.2    Synthesis and Evaluation of Carcinogenic Effects	3-113
    3.4.3    Mode of Action and Implications in Cancer Assessment	3-114
    3.4.4    Weight of Evidence Evaluation for Carcinogenicity	3-114
    3.4.5    Potentially Sensitive Populations	3-115
4.   DOSE-RESPONSE ASSESSMENT	4-1
  4.1     Dose-Response for Noncancer Effects	4-1
    4.1.1    RfD Determination	4-1
      4.1.1.1   Pharmacokinetic Model	4-7
      4.1.1.2   RfD Quantification	4-14
    4.1.2    RfC Determination	4-17
  4.2     Dose-Response for Cancer Effects	4-17
5.   REFERENCES	5-1
Appendix A: Literature Search Strategy Developing the Search	A-l
Appendix B: Studies Evaluated Since August 2014	B-l
Appendix C: Summary  of Data	C-l
Perfluorooctane sulfonate (PFOS) -May 2016                                               viii

-------
                                     TABLES
Table 1-1. Chemical and Physical Properties of PFOS	1-2
Table 2-1. Mean % (± SE) of 14C-K+PFOS in Rats After a Single Dose of 4.2 mg/kg	2-2
Table 2-2. Percent (%) Binding of PFOS to Human Plasma Protein Fractions	2-2
Table 2-3. Average PFOS Level (|ig/mL or ppm) in Serum Of Monkeys	2-6
Table 2-4. Levels of PFOS in Serum and Bile of Rats Treated for 5 Days	2-7
Table 2-5. Mean (± SD) Daily PFOS Consumption and Tissue Residue Levels in Rats
         Treated for 28 Days	2-7
Table 2-6. Concentrations of PFOS in Male Rats' Whole Blood (|ig/mL) and Various
         Tissues (|ig/g) After 28 Days	2-8
Table 2-7. PFOS Levels in the  Serum and Liver of Rats	2-8
Table 2-8. Mean Concentration of PFOS (±  SD)  in Various Tissues of Mice	2-9
Table 2-9. Levels of PFOS (Means ± SE) in Mouse Serum Following Treatment for 10
         Days	2-10
Table 2-10. PFOS Concentrations (Mean ± Standard Deviation [SD]) in Samples from
         Pregnant Dams and Fetuses (GD 21 only) in |ig/mL (ppm) for Serum and Urine
         and |ig/g for Liver andFeces	2-11
Table 2-11. Mean PFOS (± Standard Error)  Concentrations in Serum, Liver, snd Brain
         Tissue in Dams and Offspring	2-12
Table 2-12. PFOS Contents in Serum, Hippocampus, and Cortex of Offspring (n = 6)	2-13
Table 2-13. Mean PFOS Content in Serum and Lungs of Rat Offspring (n = 6)	2-14
Table 2-14. Ratios (Means ± SD) 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	2-14
Table 2-15. Percent Distribution (%) of PFOS in Mice After a 50 mg/kg Subcutaneous
         Injection	2-16
Table 2-16. Pharmacokinetic Parameters from Wambaugh et al. (2013) Meta-Analysis of
         Literature Data	2-30
Table 2-17. PFOS Pharmacokinetic Data Summary for Monkeys	2-32
Table 2-18. PFOS Pharmacokinetic Data Summary for Rats	2-33
Table 2-19. PFOS Pharmacokinetic Data Summary for Mice	2-33
Table 2-20. Summary of Half-Life Data	2-34
Table 3-1. Association of Serum PFOS with Serum Lipids	3-6
Table 3-2. Summary of Epidemiology Studies of PFOS and Liver Enzymes	3-10
Table 3-3. Summary of Epidemiology Studies of PFOS and Measures of Kidney Function ... 3-12
Table 3-4. Summary of Epidemiology Studies of PFOS and Pregnancy Outcomes	3-14
Table 3-5. Summary of Epidemiology Studies of PFOS and Fetal Growth	3-16

Perfluorooctane sulfonate (PFOS) -May 2016                                              ix

-------
Table 3-6. Summary of Epidemiology Studies of PFOS and Thyroid Effects	3-34
Table 3-7. Summary of Epidemiology Studies of PFOS and Immune Suppression
         (Infectious Disease and Vaccine Response)	3-38
Table 3-8. Summary of PFOS Epidemiology Studies of Cancer	3-47
Table 3-9. Mean (± SD) Values for Select Parameters in Rats Treated for 4 Weeks	3-52
Table 3-10. Mean (± SD) Values for Select Parameters in Rats Treated for 28 Days	3-53
Table 3-11. Mean (± SD) Values for Select Parameters in Monkeys Treated for 182 Days	3-58
Table 3-12. Mean (± SD) Values for Select Parameters in Rats Treated for 14 Weeks	3-60
Table 3-13. Fertility and Litter Observations in Dams Administered 0 to 2.0 mg
         PFOS/kg/day	3-66
Table 3-14. Effects Observed in the Mice Administered PFOS from GD 0 to GD 17/18	3-72
Table 3-15. Incidence of Nonneoplastic Liver Lesions in Rats (Number Affected/Total
         Number)	3-79
Table 3-16. Tumor Incidence (%)	3-80
Table 3-17. Genotoxicity of PFOS in vitro	3-82
Table 3-18. Genotoxicity of PFOS in vivo	3-82
Table 3-19. Summary of SRBC and NK Cell Findings in Mice after PFOS Exposure	3-89
Table 3-20. Thyroid Hormone Levels in PFOS Treated Rats	3-91
Table 3-21. Summary of PFAS  Transactivation of Mouse and Human PPARa, P/5, and y	3-94
Table 4-1. NOAEL/LOAEL and Effects for Longer-Term Duration Studies of PFOS	4-4
Table 4-2. NOAEL/LOAEL Data for Short-Term Oral Studies of PFOS	4-6
Table 4-3. Predicted Final Serum Concentration and Time Integrated Serum Concentration
         (AUC)  for Different Treatments of Rat	4-9
Table 4-4. Predicted Final Serum Concentration and Time Integrated Serum Concentration
         (AUC)  for the Mouse	4-9
Table 4-5. Predicted Final Serum Concentration and Time Integrated Serum Concentration
         (AUC)  for the Monkey	4-10
Table 4-6. Average serum Concentrations for the Duration of Dosing	4-10
Table 4-7. Comparison of Average Serum Concentration and Steady-State Concentration	4-12
Table 4-8. Human  Equivalent Doses Derived from the Modeled Animal  Average
         Serum Values	4-14
Table 4-9. POD Outcomes for the HEDs from the Pharmacokinetic Model Average Serum
         Values	4-15
Table B-1. PFOS Epi Papers—Post Peer Review (Retrieved and Reviewed)	B-1
Table B-2. PFOA Post Peer Review Animal Toxicity  Studies	B-3
Table B-3. Toxicokinetics: Post Peer Review	B-5
Perfluorooctane sulfonate (PFOS) -May 2016

-------
Table B-4. Association of Serum PFOS with Serum Lipids and Uric Acid	B-6
Table B-5. Association of Serum PFOS with Reproductive and Developmental Outcomes	B-8
Table B-6. Association of PFOS Level with the Prevalence of Thyroid Disease and Thyroid
          Hormone Levels	B-ll
Table B-7. Association of Serum PFOS with Markers of Immunotoxicity	B-12
Table C-l. PFOS Toxicokinetic Information	C-2
Table C-2. Summary of Animal Studies with Exposure to PFOS	C-7
Perfluorooctane sulfonate (PFOS) -May 2016

-------
                                      FIGURES
Figure 1-1. Chemical Structure of PFOS	1-1
Figure 2-1. Distribution of Radiolabeled PFOS in Dams and in Fetuses/Pups in the
          Liver, Lung, Kidney, and Brain (Figure from Borg et al. 2010)	2-15
Figure 2-2. PFOS Contents in Urine, Feces, and Overall Excretion in Male Rats Treated
          for 28 Days	2-19
Figure 2-3. Schematic for a Physiologically-Motivated Renal Resorption Pharmacokinetic
          Model	2-21
Figure 2-4. Structure of Model for PFOS in Rats and Monkeys	2-22
Figure 2-5. Structure of the PFOS PBPK Model in Monkeys and Humans	2-22
Figure 2-6. Structure of the PBPK Model for PFOS in the Adult Sprague-Dawley Rat	2-25
Figure 2-7. Predicted Daily Average Concentration of PFOS in Maternal (Black Line) and
          Fetal (Gray Line) Plasma at External Doses to the Dam	2-26
Figure 2-8. PBPK Model Structure for Simulating PFOA and PFOS Exposure During
          Pregnancy in Humans (Maternal, Left; Fetal, Right)	2-27
Figure 3-1. Functional Categories of Genes Modified by PFOS in Wild-Type and Null
          Mice	3-98
Figure 3-2. Function Distribution and Category Enrichment Analysis of the Differential
          Proteins	3-99
Perfluorooctane sulfonate (PFOS) -May 2016                                              xii

-------
                        ABBREVIATIONS AND ACRONYMS
ACh
ADHD
ALP
ALT
ANOVA
AP
AST
AUC
BMD
BMDL
BMI
BUN
°C
CAR
CAS
CASRN
CCL
CD
CFSE
CHMS
CI
CL
CoA
Cone.
Css
CSF
CSM
Cte
d
DA
DAUDA
DCDQ
DIO1
DIO3
dL
DMEM
DMSO
DNA
DNBC
DP
DR
E
EAA
ECso
ECF
ED
acetylcholine
attention deficit hyperactivity disorder
alkaline phosphatase
alanine transaminase
analysis of variance
activation protein
aspartate aminotransferase
area under the curve
benchmark dose
benchmark dose - lower 95th percentile confidence bound
body mass index
blood urea nitrogen
degrees Celsius
constitutive androstane receptor
Chemical Abstracts Service
Chemical Abstracts Service Registry Number
Chemical Contaminants List
circular dichroism
6-carboxyfluorescein succinimidyl ester
Canadian Health Measures Survey
confidence interval
clearance
coenzyme A
Concentration
Steady-state concentration
cerebrospinal fluid
cholestyramine
acyl CoA thioesterase
day
dansylamide
1 l-(5-dimethylaminoapthalenesulphonyl)-undecanoic acid
Developmental Coordination Disorder Questionnaire
type 1 deiodinase
type 3 deiodinase
deciliter
Dulbecco's Minimal Essential Medium
dimethyl sulfoxide
deoxyribonucleic acid
Danish National Birth Cohort
dansyl-L-proline
dose rate
estradiol
excitatory  amino acid
half maximal effective concentration
Electro-Chemical Fluorination
equilibrium dialysis
Perfluorooctane sulfonate (PFOS) -May 2016
                                                               Xlll

-------
eGFR
EMM
EPA
FABP
FAI
FR
FSH
FT
FT3
FT4
g
GABA
GAP-43
GD
GFAP
GFR
GGT
GI
GJIC
GLP
Glu
GS
HDL
HED
HESD
HL-60
HMG-CoA
HOMA
HPT
HPLC/MS/MS
h
HSDB
HSI
ICa
ICso
ICR
IgE
IL
INUENDO
IQR
IRR
IU
IV
JVow
Kt
kg
KO
L
LEW
estimated glomerular filtration rate
Estimated Marginal Mean
U.S. Environmental Protection Agency
fatty acid binding proteins
free androgen index
fecundability ratio
Follicle-stimulating hormone
free testosterone
free triiodothyronine
free thyroxin
gram
gamma-aminobutyric acid
growth-associated protein-43
gestation day
glial fibrillary acidic protein
glomerular filtration rate
gamma-glutamyl transpeptidase
gastrointestinal
gap junction intercellular communication
good laboratory practice
glutamate
glutamine synthetase
high density lipoprotein
human equivalent dose
Health Effects Support Document
human promyelocytic leukemia cell line
3-hydroxy-3-methylglutaryl coenzyme A
homeostatic model assessment
hy pothal ami c-pituitary-thy roi d
High-performance liquid chromatography/tandem mass spectrometry
hour
Hazardous Substances Database
hepatosomatic index
inward calcium currents
half-maximal Inhibiting Concentration
imprinting control region
Immunoglobulin E
interleukin
Biopersistent Organochlorines in Diet and Human Fertility study
interquartile range
incidence rate ratio
international unit
intravenous
octanol-water partition coefficient
affinity constant
kilogram
knockout
liter
low birth weight
Perfluorooctane sulfonate (PFOS) -May 2016
                                                               xiv

-------
LCso
LC/MS/MS
LD
LDso
LDL
L-FABP
LI
LIFE
LLOQ
LOAEL
LOEC
LOQ
LPS
m
MDA
ME
US
mg
min
mL
mmol
jimol
MOA
mol
MRP
NA
NCEH1
ND
ng
NHANES
NIS
NJDEP
NK
nmol
NMRI
NOAEL
NOEC
NR1H3
NS
NSP
NT
OAT
OATp
OR
P
PB
PBDE
PBMC
PBPK
Lethal concentration for 50% (statistical median) of animals
liquid chromatography/tandem mass spectrometry
lactation day
Lethal dose for 50% (statistical median) of animals
low density lipoprotein
liver fatty acid binding protein
labeling index
Longitudinal Investigation of Fertility and the Environment
lower limit of quantitation
lowest observed adverse effect level
lowest observed effect concentration
Limit of quantitation
Lipopolysaccharide
meter
malondialdehyde
malic enzyme
microgram
milligram
minute
milliliter
millimole
micromole
mode of action
mole
multidrug resistance-associated protein
not applicable
Neutral Cholesterol Ester Hydrolase 1
not detected or not determined
nanogram
National Health and Nutrition Examination Survey
sodium iodide symporter
New Jersey Department of Environmental Protection
natural killer
nanomole
Naval Medical Research Institute
no observed adverse effect level
no observed effect concentration
Nuclear Receptor Subfamily 1, Group H, Member 3
no sample
newborn screening program
not tested
organic anion transporter
organic anion transporting peptide
odds ratio
probability
phenobarbital
polybrominated diphenyl ether
peripheral  blood mononuclear cells
physiologically-based pharmacokinetic
Perfluorooctane sulfonate (PFOS) -May 2016
                                                               xv

-------
PCB                polychlorinated biphenyl
PCNA              proliferating cell nuclear antigen
PCoAO             palmitoyl CoA oxidase
PFAS               perfluoroalkyl substance
PFBA               perfluorobutyric acid
PFBS               perfluorobutane sulfonate
PFHxS              Perfluorohexanesulfonic acid
PFNA               perfluorononanoic acid
PFOA               Perfluorooctanoic acid
PFOS               perfluoroocatane sulfonate
PFOSA             perfluorooctane sulfamide
pg                  picogram
PI                  proliferation index
PK                 pharmacokinetic
pKa                acid dissociation constant
pmol                picomole
PND                postnatal day
POD                point of departure
POSF               perfluorooctanesulfonyl fluoride
PPAR               peroxisome proliferator activated receptor
ppb                 parts per billion
ppm                parts per million
mPSC               miniature post-synaptic current
mRNA              messenger ribonucleic acid
PTU                propylthiouracil
PUFA               polyunsaturated fatty acid
PXR                pregnane X receptor
Q                  flow in and out of tissues
RBC                red blood cell
RfC                reference concentration
RfD                reference dose
RIA                radio immunoassay
RNA                ribonucleic acid
RR                 rate ratio
RSI                renal-somatic index
SD                 standard deviation
SDQ                Strengths and Difficulties Questionnaire
SDWA              Safe Drinking Water Act
SHBG              sex hormone-binding globulin
SIR                standardized incidence ratio
SMR                standardized mortality ratio
SOD                superoxide dismutase
SPC                saponin compound
SRBC               sheep red blood cells
Syn 1                synapsin 1
Syp                synaptophysin
T                  total testosterone
T-AOC              total antioxidation capability
T3                  triiodothyronine

Perfluorooctane sulfonate (PFOS) -May 2016                                              xvi

-------
T4
tl/2
Tl/2
Tm
TBG
TC
TG
TH
TNF-a
TNP
TPO
TPOAb
TRH
TSH
TSHR
TT3
TT4
TIP
TTR
TUNEL
UCB
UF
UGT
UK
U.S.
Vd
VLDL
WHO
thyroxine
chemical half-life
elimination half-time
transport maximum
thyroxine-binding globulin
total cholesterol
triglycerides
thyroid hormone
tumor necrosis factor-a
trinitrophenol
thyroid peroxidase
thyroid peroxidase antibody
thyrotropin releasing hormone
thyroid stimulating hormone
thyroid stimulating hormone receptor
total triiodothyronine
total thyroxin
time to pregnancy
thyroid hormone transport protein, transthyretin
Terminal deoxynucleotidyl transferase dUTP nick end labeling
umbilical cord blood
uncertainty factor
uridine diphosphoglucuronosyl transferase
United Kingdom
United States
volume of distribution
very low density lipoprotein
World Health Organization
Perfluorooctane sulfonate (PFOS) -May 2016
                                                               XVII

-------
                               EXECUTIVE SUMMARY

   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 firefighting 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 serum
albumin and other plasma proteins. Both experimental data and pharmacokinetic models show
higher levels of PFOS in fetal serum and brain compared with the maternal compartments. PFOS
is not readily eliminated from humans as evidenced by the estimated average half-life values of
4.1-8.67 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 saturable resorption from
the kidney. In other words, after initial PFOS removal from blood by the kidney, a  substantial
fraction of what would normally be eliminated in urine is resorbed from the renal tubules and
returned to the blood. A number of published toxicokinetic models use saturable resorption as a
basis for predicting serum values in animals and humans, including one developed  by the
U.S. Environmental Protection Agency  (EPA) to support this assessment.

   Peroxisome proliferation as a result of binding to and activation of peroxisome  proliferator-
activated receptor-alpha (PPARa), is usually associated with hepatic lesions in the rat, but some
uncertainties exist as to whether this is true for liver effects induced by PFOS. 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 by PFOS are associated with nuclear receptors other than
PPARa.

   Numerous  epidemiology studies have examined occupational populations at large-scale
PFOS production plants in the United States and a residential population living near a PFOA
production facility in  an attempt to determine the relationship between serum PFOS
concentration and various health outcomes. Epidemiology data report associations between
PFOS exposure and high cholesterol and reproductive and developmental parameters. The
strongest associations are related to serum lipids with increased total cholesterol and high density
lipoproteins  (HDLs).  Data also suggest  a correlation between higher PFOS levels and decreases
in female fecundity and fertility, in addition to decreased body weights in offspring, and other
measures of postnatal growth. Several human epidemiology studies evaluated the association
between PFOS and cancers including bladder, colon, and prostate, but these data present a small
number or cases and some are cofounded by failure to adjust for smoking. The associations for
most epidemiology endpoints are mixed. While mean serum  values are presented in the human
studies, actual  estimates of PFOS exposure (i.e., doses/duration) are not currently available.
Thus, the serum level at which the effects were first manifest and whether the serum had
achieved steady state  at the point the effect occurred cannot be determined. It is likely that some
of the human exposures that contribute to serum PFOS values come from PFOS derivatives or
precursors that break  down metabolically to PFOS. These compounds may originate from PFOS
in diet and materials used in the home, thus, there is potential for confounding. Additionally,
most of the subjects of the epidemiology studies have many perfluoroalkyl substances (PFAS),
other contaminants, or both in their blood. Taken together, the weight of evidence for human


Perfluorooctane sulfonate (PFOS) - May 2016                                             ES-1

-------
studies supports the conclusion that PFOS exposure is a human health hazard. At this time, EPA
concludes that the human studies are adequate for use qualitatively in the identification hazard
and are supportive of the findings in laboratory animals.

   Short-term and chronic exposure studies in animals demonstrate increases in liver weight
consistently at doses generally > 0.5 milligrams per kilogram per day (mg/kg/day). Co-occurring
effects in these studies include decreased cholesterol, hepatic steatosis, lower body weight, and
liver histopathology.

   One and two generation toxicity studies also show decreased pup  survival and body weights.
Additionally,  developmental neurotoxicity studies show increased motor activity and decreased
habituation and increased escape latency in the water maze test following in utero and lactational
exposure to PFOS. Gestational and lactational exposures were also associated with higher serum
glucose levels and evidence of insulin resistance in adult offspring. Limited evidence suggests
immunological effects in mice.

   EPA derived a reference dose (RfD) for PFOS of 0.00002 mg/kg/day based on decreased
neonatal rat body weight from the two-generation study by Luebker et al. (2005b). A
pharmacokinetic model was used to predict an area under the curve (AUC) for the no observed
adverse effect level (NOAEL) and used to calculate a human equivalent dose (HED)NOAEL. The
total uncertainty factor (UF) applied to the FEDNOAEL 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. The FLED for effects on pup body weight in the two generation
study is supported by comparable values derived  from the lowest observed adverse effect level
for the same effect in the one-generation study and the NOAEL for effects seen in a
developmental neurotoxicity study.

   Applying  the U.S. EPA Guidelines for Carcinogen Risk Assessment, there is suggestive
evidence of carcinogenic potential for PFOS (USEPA 2005a). 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 linear response to dose was not observed. The liver tumors also showed a
slight, but statistically-significant increase 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, respectively, versus
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. Although one worker cohort found an increase in bladder cancer, smoking
was a major confounding factor, and the standardized incidence ratios were not significantly
different from the general population. Other worker and general population studies found no
statistically-significant trends for any cancer type. Thus, the weight of evidence for the
carcinogenic potential to humans was judged to be too limited to support a quantitative cancer
assessment.
Perfluorooctane sulfonate (PFOS) - May 2016                                             ES-2

-------
1.  IDENTITY: CHEMICAL AND PHYSICAL PROPERTIES

   Perfluorooctane sulfonate, commonly known as PFOS, and its salts are fluorinated organic
compounds and are part of the group of chemicals called perfluoroalkyl substances (PFAS). The
two most widely known PFAS have an eight-carbon backbone with either a sulfonate (PFOS) or
carboxylate (perfluorooctanoic acid, PFOA) 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 firefighting
foams. PFOS is produced commercially from perfluorooctanesulfonyl fluoride (POSF), which is
primarily used 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). This process yields a mixture of linear
and branched chain isomers (Beesoon and Martin 2015). The isomer  ratio is about 70% linear
and 30% branched chain. Yu et al. (2015) measured the isomer profiles of drinking water
samples collected from 10 locations in China and found that the levels of the branched isomers
accounted for 31.8% to 44.6% of the PFOS present using limits of quantification (LOQ) that
ranged from 0.04 to 0.06 nanograms per liter (ng/L). Some systems had 1-methyl and 6-methyl
isomers that were > 2% of the total. Levels of the other isomers were lower. Isomer
concentrations are important because half-life decreases as the percentage of branched isomers
increases.

   A second process for preparing PFOS is called telomerization. It produces linear chains and
was the favored process in the United States until the  time 3M voluntarily ceased production in
2002 (Beesoon et al. 2011). PFOS can also be formed in the environment by the degradation of
other POSF-derived fluorochemicals such as N-methyl or N-ethyl perfluorooctane sulfonamides
(PFOSAs) often referred to as precursors.

   Because of strong carbon-fluorine bonds, PFOS is stable to metabolic and environmental
degradation. It is a solid at room temperature and has  a low vapor pressure. Because of the
surface-active properties of PFOS, it forms three layers in octanol/water making determination of
an n-octanol/water partition coefficient (Kow) impossible. No direct measurement of the acid
dissociation constant (pKa) of the acid has been located; however, the chemical is considered to
have a low pKa and exist as a highly dissociated anion. The chemical structure is provided in
Figure 1-1, and the physical properties for PFOS are provided in Table 1-1.
                                                 V  V
                             F   F   F    'F   F    F   F    F
                               Source: Environment Canada (2006)

                        Figure 1-1. Chemical Structure of PFOS

The branched chain isomers have a 7 carbon linear chain with methyl groups located on carbons
1, 3, 4, 5, or 6 (Beesoon and Martin 2015).

Perfluorooctane sulfonate (PFOS) - May 2016                                               1-1

-------
                     Table 1-1. Chemical and Physical Properties of PFOS
Property
Chemical Abstracts Service
Registry Number (CASRN)
Chemical Abstracts Index
Name
Synonyms
Chemical Formula
Molecular Weight (grams
per mole [g/mol])
Color/Physical State
Boiling Point
Melting Point
Vapor Pressure
Henry's Law Constant
KOW
organic carbon water
partitioning coefficient (Koc)
Solubility in Water
Half-life in Water
Half -life in Air
PFOS, acidic form*
1763-23-1
1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,8-
heptadecafluoro-1-octanesulfonic acid
Perfluorooctane sulfonic acid;
heptadecafluoro-1 -octane sulfonic acid;
PFOS acid
CsHFnOsS
500.13
White powder
(potassium salt)
258-260 °C
No data
2.0 x 10"3 milligrams Mercury (mm Hg)
at 25 °C (estimate)
Not measureable
Not measurable
2.57
680 mg/L
Stable
Stable
Source




Lewis (2004); Hazardous Substances
Database (HSDB) (2012); SRC (2016)
OECD (2002)
SRC (2016)

HSDB (2012)
ATSDR(2015)
EPS A (2008); ATSDR (2015)
Higgins and Luthy (2006)
OECD (2002)
UNEP (2006)
UNEP (2006)
Notes: *PFOS is commonly produced as a potassium salt (CASRN 2795-39-3). Properties specific to the salt are not included.
This CASRN given are for linear PFOS, respectively, but the toxicity studies are based on a mixture of linear and branched, and
thus the RfD applies to the total linear and branched.
Perfluorooctane sulfonate (PFOS) - May 2016
1-2

-------
2.  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 perfluorinated compound (PFAS),
PFOA, membrane transporter families appear to play an important role in absorption,
distribution, and excretion. The transporter families identified for PFOA include organic anion
transporters (OATs), organic anion transporting peptides (OATps), multidrug resistance-
associated proteins (MRPs), and urate transporters. Transporters play a critical role in
gastrointestinal absorption, uptake by the tissues, and excretion via the kidney. Limited data are
available regarding the transporters and PFOS, however the toxicokinetic properties of PFOS
suggest facilitated transport functions in tissue uptake and renal resorption. Hepatic OATpl,
OATp2, and MRP2 messenger ribonucleic acid (mRNA) respond to PFOA exposure in a dose-
related manner.  Some inhibition studies  suggest that PFOS  with its similar chain length, renal
excretion properties and liver accumulation could involve the same transporters. However,
transporter-specific data related to PFOS are minimal.

   Animal studies indicate that PFOS is well-absorbed orally and distributes primarily to the
blood and liver. While PFOS can form as a metabolite from other perfluorinated compounds,
PFOS itself does not undergo further metabolism after absorption takes place. PFAS are known
to activate peroxisome proliferator activated receptor (PPAR) pathways by increasing
transcription of mitochondrial and peroxisomal lipid metabolism, as well as sterol and bile acid
biosynthesis based on transcriptional activation of many genes in PPARa-null mice, the effects
of PFAS involve more than activation of PPAR receptors (Andersen et al. 2008). A summary of
toxicokinetic  data are provided in Appendix C, Table C-l.

2.1   Absorption

   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 likely to be associated with transporters rather than
simple diffusion. Unfortunately no information on the interaction of PFOS with intestinal, lung,
or skin transporters in mammals was identified.

   While there are no absorption studies available that quantify absorption in humans, extensive
data on serum PFOS demonstrate uptake from the environment but not the exposure route.
Studies that provide the basis for human half-life estimates  rely on changes in serum levels over
time. Section  2.5.2 of this document provides serum levels measured in humans.

2.1.1   Oral Exposure

   Chang et al. (2012) administered a single dose of 4.2 milligrams per kilogram (mg/kg) of
PFOS-14C in solution to 3 male rats. At 48 hours after dosing, 3.32% of the total dose was found
in the digestive tract and 3.24% in the feces, indicating that most of the dose had been absorbed
with some of the unabsorbed material excreted in fecal matter (Table 2-1).
Perfluorooctane sulfonate (PFOS) - May 2016                                               2-1

-------
    Table 2-1. Mean % (± SE) of 14C-K+PFOS in Rats after a Single Dose of 4.2 mg/kg
Compartment
carcass
digestive tract
feces
urine
plasma
red blood cell (RBC)
Total
% 14C of dose recovered
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
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
Source: Data from Chang et al. 2012
Note: *A mean body weight of 300g was used to estimate the red blood cell (RBC) and plasma volume.

2.1.2   Inhalation Exposure

    An acute lethal concentration for 50% (statistical median) of animals (LCso) study in rats
indicates that PFOS absorption occurs after inhalation exposures. However, pharmacokinetic
data were not included in the published report (Rusch et al. 1979). The analytical methods for
measuring PFOS in animals were limited at the time the study was conducted.

2.1.3   Dermal Exposure

    No data are available on  dermal absorption of PFOS.

2.2  Distribution

    PFOS is distributed within the body by non-covalently binding to plasma proteins, most
commonly albumin.  The in vitro protein binding of PFOS in rat, monkey, and human plasma at
concentrations of 1-500 parts per million (ppm) PFOS was investigated by Kerstner-Wood et al.
(2003). PFOS was bound to plasma protein in all three species at all concentrations with no sign
of saturation (99.0-100%). When incubated with separate human-derived plasma protein
fractions, 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%). Low levels of binding to alpha-2-macroglobulin and
transferrin were measured when the protein concentrations  were approximately 10% of
physiological concentration (Table 2-2).

       Table 2-2. Percent (%) Binding of PFOS to Human Plasma Protein Fractions
Fraction
Albumin
Gamma-globulin
Alpha-globulin
Fibrinogen
Alpha-2-macroglobulin
Transferrin
LDLs
~ 10% Physiological
Concentration (Cone.)
99.0
6.3
49.9
<0.1
12.5
7.2
90.1
100% Physiological Cone.
99.8
24.1
59.4
<0.1
<0.1
<0.1
95.6
Source: Data from Kerstner-Wood et al. 2003.
Perfluorooctane sulfonate (PFOS) - May 2016
2-2

-------
   Zhang et al. (2009) used equilibrium dialysis, fluorophotometry, isothermal titration
calorimetry and circular dichroism (CD) to characterize interactions between PFOS and serum
albumin 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 change
in the PFOS concentration in the dialysis solution. During dialysis, the PFOS concentration in
the solution decreased reflecting its ability to cross the dialysis membrane and bind to the
biopolymer within the dialysis bag. Based on the data, the serum albumin 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 decreasing solution pH (i.e., capable of
promoting protein and DNA denaturation), thus providing an increased number of binding sites.
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 serum albumin and PFOS were the
results of surface electrostatic interactions between the sulfonate functional group and the
positively charged side chains  of lysine and arginine. 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 serum albumin. Intrinsic fluorescence analysis of tryptophan
residues in serum albumin suggested a potential interaction of PFOS with tryptophan, an amino
acid likely to be found in a  hydrophobic portion of the albumin. 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 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 conformation could change its transporting activity. CD
spectrometry was used to determine if PFOS changed the conformation of the albumin 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 transport
function as a result of the conformational change.

   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 indicative of binding. A binding constant of 2.2 x lO4^!"1
and a binding ratio of PFOS to human albumin of 14 moles PFOS/mole albumin were calculated.

   Human serum albumin has two high-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. Alone these two probes emit negligible fluorescence; 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 mmol), DA emission increased as the
PFOS concentration increased until it was at 140% the original intensity. At the higher PFOS
concentrations (0.7-4 mmol), 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 Site II, PFOS caused a fluorescence reduction that was quick
at first, but then became more gradual suggesting the possibility that PFOS was binding to this

Perfluorooctane sulfonate (PFOS) - May 2016                                              2-3

-------
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 PFOS binding sites were determined for human serum albumin
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 the weighted histogram analysis method 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 kilocalories per mole
[kcal/mol]) was located near the tip of the tryptophan-214 binding site, and the maximum
number of ligands that could bind to human serum albumin for PFOS was 11. The most
populated albumin binding site for PFOS was dominated by van der Waals interactions. The
author indicated that eleven PFOS molecules were adsorbed on  the surface of the albumin.

    PFOS binding to bovine serum albumin was evaluated using electrospray ionization mass
spectrometry by D'Alessandro et al. (2013). Using this approach, the estimate for the maximum
number of PFOS binding sites was also 11, but the data on collision-induced PFOS removal was
more consistent with 7 binding sites. Two of the potential binding sites (Sudlow's sites I and II)
are binding sites for a number of pharmaceuticals.

    D'Alessandro et al. (2013) also examined whether PFOS could prevent binding of ibuprofen
to its Sudlow II site and whether it was also able to displace bound ibuprofen. The study showed
that PFOS competes with ibuprofen for its site when the PFOS:ibuprofen ratio is > 0.5 moles: 1
mole. In addition, when the binding site is occupied by PFOS, ibuprofen is unable to bind. Zhang
et al. (2009) conducted a similar study of the impact of PFOS on the ability of serum albumin to
bind 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%. These data suggest that PFOS can alter the pharmacokinetics and pharmacodynamics of
medicinal and natural substances that share a common site on albumin.

    Beesoon and Martin (2015) examined differences in the binding of the linear and branched
chain isomers to serum albumin and human serum proteins. The linear PFOS molecule was
found to bind more strongly to calf serum albumin than the branched chain isomers. When
arranged in order of increasing binding the order was 3m < 4m < 1m < 5m  < 6m  (iso) 
-------
1.0 and 3.0 mg/kg groups, respectively. No effect with treatment was observed on OATpl. No
additional information on PFOS tissue transport was identified.

Humans. In humans, PFOS distributes mostly to the liver and blood. Olsen et al. (2003a)
sampled both liver and serum from cadavers for PFOS. There was a good correlation between
samples from the same subject. There was no difference in the PFOS concentrations identified in
males and females or between age groups. Karrman et al. (2010) identified PFOS in postmortem
liver samples (n = 12; 6 males and 6 females 27-79 years old) with a mean concentration of
26.6 ng/g tissue.

    Perez et al. (2013) collected tissue  samples from 20 adult subjects (aged 28-83) who had
been living in Catalonia, Spain for 10 years and died of a variety of causes. Autopsies and tissue
collection (liver, kidney, brain lung, and bone) were carried out in the first 24 hours after death.
The tissues were analyzed for 21 perfluorinated compounds. PFOS was present in 90% of the
samples but could be quantified in only 20% (median 1.9 ng/g). PFOS accumulated primarily in
the liver (104 ng/g),  kidney (75.6 ng/g), and lung (29.1 ng/g), and it was low in brain (4.9 ng/g)
and bone (not detected) based on the mean wet weight tissue concentration. Detection levels
varied with the tissue evaluated.

    Stein et al. (2012) compared PFAS levels in maternal serum and amniotic fluid paired
samples from 28 females in their second trimester of pregnancy. PFOS (0.0036-0.0287 |ig/mL)
was detected in all serum samples and in nine amniotic fluid samples (0.0002-0.0018 jig/mL).
The Spearman correlation coefficient between the serum and amniotic fluid levels was 0.76 and
is significant (p =  0.01), indicating a direct relationship between the levels in blood and amniotic
fluid. The median ratio of maternal serum:amniotic fluid concentration was 25.5:1. Based on a
simple regression between the levels in each compartment, PFOS was rarely detected in amniotic
fluid unless the serum concentration was > 0.0055  jig/mL.

    Harada et al. (2007) obtained cerebrospinal fluid (CSF) from seven patients (6 males and
1 female; aged 56-80) to evaluate the partitioning of PFOS between serum and the CSF. The
median concentration of PFOS in the serum was 0.0184 |ig/mL, compared to the concentration
in the CSF (0.00010 |ig/mL).  The CSF to serum ratio was 9.1 x 10'3.  The levels identified
indicate that PFOS does not easily cross the adult blood-brain barrier.

    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) identified  PFOS in breast  milk samples from healthy females (n = 10;  females 30-39
years old). The levels in milk (mean 0.12 ng/mL) were low compared to liver levels.

Animals

Monkey. 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 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). Serum and tissues were collected at the time of sacrifice.  The dosing was followed
by a 52-week recovery period in  2 animals in the control, 0.15 and 0.75 mg/kg/day groups.
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
Perfluorooctane sulfonate (PFOS) - May 2016                                              2-5

-------
dosed groups. The average percent of the cumulative dose of PFOS in the liver at the end of
treatment ranged from 4.4% to 8.7% with no difference by dose group or gender. The
concentration of PFOS in the liver decreased during the recovery period. Serum levels are
provided in Table 2-3.

          Table 2-3. Average PFOS Level (ug/mL or ppm) in Serum of Monkeys
Time
(weeks)

1
4
16
27
35
57
79
Group 1
0.0 milligram
(mg)/kilogram (kg)/day
Males

-------
          Table 2-4. 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 (microgram
[ug]/milliliter [mL])

-------
    Table 2-6. Concentrations of PFOS in Male Rats' Whole Blood (ug/mL) and Various
                                Tissues (ug/g) after 28 Days
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
Source: Data from Table 1 in Cui et al. 2009.
Note: ND = not detected

    A combined chronic toxicity/carcinogenicity good laboratory practice (GLP) study was
performed in 40-70 male and female Sprague-Dawley Crl:CD (SD)IGS BR rats administered 0,
0.5, 2, 5, or 20 ppm of PFOS for 104 weeks (Thomford 2002/Butenhoff et al. 20121). Doses
were approximately 0, 0.018-0.023, 0.072-0.099, 0.184-0.247 and 0.765-1.1 mg/kg/day. A
recovery group was administered the test substance at 20 ppm for 52 weeks and observed until
sacrifice at 106 weeks. Serum and liver samples were obtained during and at the end of the  study
to determine the concentration of PFOS. Dose-dependent increases in the PFOS level in the
serum and liver were observed, with values slightly higher in females. Further study details are
described in section 3.2.7 Chronic Toxicity. Levels of PFOS identified in the liver and serum are
included in Table 2-7.

                  Table 2-7. PFOS Levels in the Serum and Liver of Rats
Timepoint
(weeks)
0 ppm PFOS
(0 mg/kg/day)
M
F
0.5 ppm
(0.024-0.029
mg/kg/day)
M
F
2 ppm
(0.098-0.120
mg/kg/day)
M
F
5 ppm
(0.242-0.299
mg/kg/day)
M
F
20 ppm
(0.984-1.251
mg/kg/day)
M
F
Serum PFOS levels (jig/mL)
0
14
53
105
106

-------
Mouse. 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 to 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 2-8 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 2-8).

        Table 2-8. Mean Concentration of PFOS (± SD) in Various Tissues of Mice
Tissues
1 day
3 days
5 days
Dose of 0.013 mg/kg/day (PFOS in tissue reported as picomole [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 nanomole [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)"*
Source: Data from Tables 2 and 3 in Bogdanska et al. 2011
Notes: '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
Perfluorooctane sulfonate (PFOS) - May 2016
2-9

-------
   In an immunotoxicity study, four to six C57BL/6 male mice/group were administered diets
with 0% to 0.02% PFOS for 10 days. Levels in the serum increased as the concentration
increased (Table 2-9) (Qazi et al. 2009a).

Table 2-9. Levels of PFOS (Means ± SE) in Mouse Serum Following Treatment for 10 Days
Dietary level (% 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
Source: Data from study report by Qazi et al. 2009a

Distribution during Reproduction and Development

   The availability of distribution data from pregnant females plus animal pups and neonates is
a strength of the PFOS pharmacokinetic database, because it helps to identify those tissues
receiving the highest concentration of PFOS during development. For this reason the information
on tissue levels during reproduction and development are presented separately from those that
are representative of other life stages.

Humans. T. Zhang et al. (2013) recruited pregnant females for a study to examine the
distribution of PFOS between maternal blood, cord blood, the placenta, and amniotic fluid.
Thirty two females from Tianjin, China volunteered to take part in the study. Samples were
collected at time of delivery. Maternal ages ranged from 21 to 39 years, and periods ranged from
35 to 37 weeks. It was the first child for 26 of the females and the second child for 6. The study
yielded 31 maternal whole blood samples, 30 cord blood samples, 29 amniotic fluid samples, and
29 placentas. The maternal blood contained variable levels of 10 PFAS, 8 acids and 2 sulfonates.
The mean maternal blood concentration was highest for PFOS (14.6 ng/mL), followed by PFOA
(3.35 ng/mL). In both cases, the mean was greater than the median, indicating a distribution
skewed toward the higher concentrations.

   PFOS was found in all fluids/tissues sampled. It was transferred to the amniotic fluid to a
lesser extent than PFOA based on their relative proportions in the maternal blood and cord blood
(21% versus 58%, respectively). Compared to the mean PFOS value in maternal blood, the mean
levels in the cord blood, placenta, and amniotic fluid were 21%, 56%, and 0.14% of the mean
levels in the mother's blood, respectively. The correlation coefficients between the maternal
PFOS blood levels and placenta, cord blood, and amniotic fluid levels were good (0.7 to 0.9),
and the relationships were statistically-significant (p < 0.001).

Rat. To determine the dose-response curve for neonatal mortality in rat pups born to PFOS
exposed dams and to investigate associated biochemical and pharmacokinetic parameters, 5
groups of 16 female Sprague-Dawley Crl:CD(SD)IGS VAF/Plus rats each were administered 0,
0.1, 0.4, 1.6, or 3.2 mg PFOS/kg bw/day by oral gavage beginning 42 days prior to cohabitation
and continuing through gestation day (GD) 14 or 20 (Luebker et al. 2005b). Eight rats from each
group were randomly chosen and sacrificed on GD 15, followed by Caesarean removal of the
pups. All remaining animals were sacrificed and C-sectioned on GD 21. Urine and feces were
collected overnight from dams on the eve of cohabitation day 1 and during GDs 6-7, 14-15, and
20-21. Serum samples were collected just prior to cohabitation and on GD 7, GD 15, and GD 21.
Fetal liver and blood samples were obtained on GD 21 and pooled by litter.
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-10

-------
   The urine, feces, and liver of the control animals all contained PFOS at small concentrations.
In treated rats, the highest concentration of PFOS was in the liver. Serum levels in the dams for
each dose were consistent between GD 1 and GD 15, indicating achievement of steady state prior
to conception (Table 2-10). The GD 21 levels in the dams had dropped below those observed
earlier in the pregnancy. Serum levels in the GD,21 fetuses were higher than those in the dams.
In contrast, the liver levels in the dams on GD 21 were about three times higher than in the
fetuses. Fecal excretion was greater than urinary excretion by the dams.

   Table 2-10. PFOS Concentrations (Mean ± Standard Deviation [SD]) 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

-
-
-
-

-
-
-
-
Source: Data from Luebker et al. 2005b
Note: - = no sample obtained

    This same study also included a subset of dams allowed to litter naturally and dosed through
lactation day (LD) 4. Liver and serum samples were collected from dams and pups on LD 5. In
this sampling, serum PFOS levels were similar between the dam and offspring, but the liver
values were now higher in the neonates than in their dams.

    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 GD 0 through postnatal day (PND) 20. An
additional 10 mated females served as satellite rats to each of the four groups and were used to
collect additional blood and tissue samples. Further details from this study are provided in
section 3.2.6 as reported in Butenhoff et al. (2009). 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 2-11.
Perfluorooctane sulfonate (PFOS) - May 2016
2-11

-------
    Table 2-11. Mean PFOS (± Standard Error) Concentrations in Serum, Liver, and Brain
                                 Tissue in Dams and Offspring
Time
GD20a



PND4a



PND21



PND72



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

-------
   Based on the maternal and offspring data on GD 20, there is placental transfer of PFOS from
rat dams to developing fetuses. Serum values were approximately 1-2 times greater in the
fetuses than in the dams at GD 20. The concentration of PFOS in fetal liver 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 lower than those for the dams on PND 4 and continued to drop through PND 72.
However, based on the concentrations still present in the neonate 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 and decreased
significantly by PND 72. Liver values were similar at all time-points between males and females.
On GD 20,  the brain levels for the pups were ten-fold higher than those for the dam. The levels
in pup brain gradually declined between PND 4 and PND 21.

   In a study by Zeng et al. (2011) 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.
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 2-12.

   Table 2-12. PFOS Contents in Serum, Hippocampus, and Cortex of Offspring (n = 6)
Time
PNDO




PND 21



Dose group (mg/kg/day)
Control
0.1
0.6
2.0

Control
0.1
0.6
2.0
Serum (jig/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 (jig/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*"
Source: Data from Table 2 in Zeng et al. 2011
Notes: 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

    Sprague-Dawley rats were administered PFOS in 0.05% Tween (in deionized water) once
daily by gavage from GD  1 to GD 21 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 2-13).
Perfluorooctane sulfonate (PFOS) - May 2016
2-13

-------
       Table 2-13. Mean PFOS Content in Serum and Lungs of Rat Offspring (n = 6)
Age
PNDO



PND21


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 Qig/g)
ND
0.92 ±0.04*
22.4 ± 1.03*

ND
0.21 ±0.04*
3.16±0.11"
Source: Data from Table 2 in Chen et al. 2012
Notes: ND = not detected
* p < 0.05 compared with control
** p < 0.01 compared with control

Mouse. Borg et al. (2010) administered a single dose of 12.5 mg/kg 35S-PFOS by intravenous
injection (n = 1) or gavage (n = 5) on GD 16 to C57B1/6 dams. Using whole-body
autoradiography and liquid scintillation, counting distribution of PFOS was determined for the
dams/fetuses (GD 18 and 20) and the neonates on PND 1. Distribution in the dams was similar
regardless of the route of exposure, with the hepatic level being approximately four times greater
than the serum. Maternal PFOS levels were highest in the liver and lungs at all timepoints. In
dams, the concentration of PFOS in the liver was approximately 4 times and in the lung was
approximately 2 times 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 fetuses,
the highest concentrations of PFOS were found in the kidneys and liver. In the fetuses on GD 18,
values in the lungs were similar to the maternal lungs, and this value increased by GD 20. In the
kidneys, the highest concentration of PFOS was observed in the fetuses on GD 18 (3 times
higher than maternal levels)

   In the offspring at all timepoints, PFOS was homogeneously distributed in the liver at a level
2.5 times higher than maternal blood and 1.7 times lower than maternal liver. In pups on PND 1,
PFOS was mostly concentrated in the lungs and liver. Pups on PND 1  had PFOS levels that were
3 times higher in the lungs, compared to maternal blood with a heterogeneous distribution. In the
kidneys, the levels in pups on PND 1 were similar to their dams despite being higher on GD 18.
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 2-14 and
Figure 2-1.

   Table 2-14. Ratios (Means ± SD) 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]maternal blood
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 = 1-6)
1.0
2.3
1.1 ±0.04
1.7" ±0.3
Source: Data from Table 1 in Borg et al. 2010
Notes: 'Statistically-significant (p < 0.01) in comparison to maternal blood
**Statistically-significant (p < 0.001) in comparison to maternal blood
Perfluorooctane sulfonate (PFOS) - May 2016
2-14

-------
                   (A) Liver
   (B)Lung

40-
"o»
~3>
a 30-
U)
o
!£ 20-
V)
" 10-
0-
A
B A a
• B B
i i V
•i- -I*
TD B

GD18 GD20 PND1
Fetuses Fetuses Pups
A ™~
T. 40-
i ^ ~-^
a 30-
» 5?
O
£ 20-
w
f|^ " 10-
o
A
b
c
A
a -f^
m j^ 	
i "•&
i

I u I i I
Maternal Maternal GD18 GD20 PND1
Liver Blood Fetuses Fetuses Pups

a
4^
j_
^
i i
Maternal Maternal
Lungs Blood
ou-
40-
"Si
a 30-
o
fc 20^
F
10-
0
\*^ f I*1V«I I^f J
A
B
"•" c c
^*
GD18 GD20 PND1
Fetuses Fetuses Pups



* ^
i \
Maternal Maternal
Kidneys Blood
                                                 20-
                                                 15-
                                               to
                                               o
                                                 10-
                                                  5-
                                                                 (D) Brain
B
c
                                                     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 GD 20/PND 1 with corresponding value on GD 18;

    Figure 2-1. Distribution of Radiolabeled PFOS in Dams and in Fetuses/Pups in the
                             Liver, Lung, Kidney, and Brain
                                 (Figure from Borg et al. 2010)

2.2.2   Inhalation and Dermal Exposure

   No data on distribution following inhalation or dermal exposures were identified.

2.2.3   Other Routes of Exposure

   Male and female mice were administered PFOS by subcutaneous injection one time on PNDs
7, 14, 21, 28, or 35 at concentrations of 0 or 50 mg/kg bodyweight (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 2-15. The
distribution shows that beyond PND 14 the levels in the liver are approximately two to four
times greater than those found on PND 7.
Perfluorooctane sulfonate (PFOS) - May 2016
                         2-15

-------
   Table 2-15. Percent Distribution (%) of PFOS in Mice after a 50 mg/kg Subcutaneous
                                       Injection
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"
Source: Data from Table 4 in Liu et al. 2009.
Note: "Statistically significant from PND 7 (p < 0.01)
2.3  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
by the Kerstner-Wood et al. (2003) data on binding to plasma proteins, in addition to the Zhang
et al. (2009) and Chen and Guo (2009) data from albumin-binding investigations. PFOS binding
to other serum and intracellular proteins also occurs.

   Weiss et al. (2009) screened the binding of PFOS to the thyroid hormone transport protein
transthyretin (TTR) in a radioligand-binding  assay to determine if it could compete with
thyroxine (T4), the natural ligand of TTR. Human TTR was incubated with 125I-labeled T4,
unlabeled T4, and 10-10,000 nanomoles (nmol) 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 nmol. The authors concluded that PFOS demonstrates an
affinity to TTR and had a greater affinity than the compounds with shorter chain lengths.

   Luebker et al. (2002) investigated the possibility that PFOS could interfere with the binding
affinity of liver-fatty acid binding protein (L-FABP), an intracellular lipid-carrier protein, with
long chain fatty acids (e.g., palmitic and oleic acid). This study was performed in vitro with a
fluorescent fatty acid analogue  ll-(5-dimethylaminoapthalenesulphonyl)-undecanoic acid
(DAUDA). The concentration that can inhibit fifty percent of specific DAUDA-L-FABP binding
(half-maximal Inhibiting Concentration, or ICso) was determined. PFOS demonstrated inhibition
of L-FABP in competitive binding assays; with 10 micromoles ((imols) PFOS added, 69% of
specific DAUDA-L-FABP binding was inhibited with the calculated ICso of 4.9 jimol.

   L. Zhang et al. (2013) cloned the human L-FABP gene and used it to produce purified
protein for evaluation of the binding of PFOA and other PFASs. Nitrobenzoxadizole-labeled
lauric acid was the fluorescent substrate used in the displacement assays. ICso values and
dissociation constants were generated for the PFAS studied. Oleic and palmitic acids served as
the normal substrates for L-FABP binding. The nitrobenzoxadiazole labeled lauric acids
indicated that there were two distinct binding sites for fatty acids in human FABP, with the
primary site having a 20-fold higher affinity than the secondary site. The ICso value for PFOS
was 3.3 ±1  ^mol, suggesting that it has a higher binding affinity than PFOA.
Perfluorooctane sulfonate (PFOS) - May 2016
2-16

-------
2.4  Excretion

2.4.1   Oral Exposure

Humans. Urinary excretion of PFOS in humans is impacted by the isomeric composition of the
mixture present in blood and the gender/age of the individuals. The half-lives of the branched
chain PFOS isomers are shorter than those for the linear molecule, an indication that renal
resorption is less likely with the branched chains.

   Y. Zhang et al. (2013) determined half-lives for PFOA isomers based on paired serum
samples and early morning urine samples collected from healthy volunteers in two large Chinese
cities. Half-lives were determined using a one compartment model and an assumption of first
order clearance. The volume of distribution (Vd) applied in the analysis as determined by
Thompson et al. (2010) was 170 mL/kg. Clearance was estimated from the concentration in urine
normalized for creatinine and assuming excretion of 1.2 and 1.4 L/day of urine and 0.9 and
1.1 mg creatinine/day for males and females, respectively. The mean half-life for the sum of all
PFOS isomers in younger females (n = 20) was 6.2 years (range: 5.0-10 years), while that for all
males and older females (n = 66) was 27 (range: 14-90 years); the medians were 6.0 and 18
years, respectively.

   The mean half-life values for the six branched chain isomers of PFOS were lower than the
value for the linear chain with the exception of the 1-methyl heptane sulfonate, suggesting that
resorption transporters may favor uptake of the linear chain and 1-methyl branched chain over
the other isomers. Older females and males have longer half-lives than young females,
suggesting the importance of monthly menstruation as a pathway for excretion (Y. Zhang et al.
2013). The mean half-life for the 1-methylheptane sulfonate in the males and older females
(n = 43) was considerably greater than that for the sum of all isomers (90 years versus 27 years).
For males and older females there were considerable inter-individuals differences, with 100-fold
differences between the minimum and maximum values among the males and older females
compared to < 10-fold differences for the younger females.

   T. Zhang et al. (2014) derived estimates for PFOA's urinary excretion rate using paired urine
and blood samples from 54 adults (29 male, 25 female) in the  general population and 27
pregnant females in Tainjin, China The age range for the general population was 22-62 and that
for the pregnant females was 21-39. Urinary excretion was calculated based on the concentration
in the urine times volume of urine, wherein a urinary volume of 1200 mL/day was applied to all
females and 1600 mL/day applied to all males. Urine samples were first draw morning samples.
Total daily intakes for PFOS were calculated from the concentration in blood using first order
assumptions, a half-life of 5.4 years (Olsen et al. 2007) and a volume of distribution of
170 mL/kg (Thompson et al. 2010; Egeghy and Lorber 2011). Urinary elimination rate was then
calculated from the urine levels and the modeled total daily intake. Total daily intake, and thus
the urinary elimination rate, was not calculated for pregnant females due to the highly variable
blood levels of PFOS during pregnancy. PFOS was detected in the blood samples for all
participants but only for 48% of the urine samples from the general population and 11% of
samples from the pregnant females. Unfortunately the urinary PFOS was below detection for
most of the females in the study.

   The calculated geometric mean total daily intake for PFOS was 89.2 ng/day for the adult
general population, resulting in a daily urinary excretion rate of 16% of the estimated intake;
there was no significant difference between males and females. From the limited number of urine

Perfluorooctane sulfonate (PFOS) - May 2016                                             2-17

-------
samples available, the urine:blood ratio was lower for pregnant females than nonpregnant
females (0.0004 versus 0.0013) suggesting other removal pathways such as placenta and cord
blood. There was a difference between the younger menstruating females (21-50 years versus
51-61 years), with a higher ratio for the younger females (0.0018 versus 0.0006). There is no
indication that data were collected from the participants relative to menstruation status on the day
of blood and urine collection. There was a significant difference between PFOS urinary excretion
in older adults compared to younger adults (p = 0.015), with a higher elimination rate in the
younger adults compared to the older age group.

   Wong et al. (2014) looked at the role of menstrual blood as an excretory pathway to explain
the shorter half-life of PFOS in females than males. They fit a population-based pharmacokinetic
model to six cross-sectional National Health and Nutrition Examination Survey (NHANES) data
sets (1999-2012) for males and females. They concluded that menstruation could account for
about 30% of the elimination half-life difference between females and males. Verner and
Longnecker (2015) suggested a need to consider the nonblood portion of the menstrual fluid and
its albumin content in the Wong et al. (2014) estimate for the menstrual fluid volume. A yearly
estimate for serum loss of 868 mL/year by Verner and Longnecker (2015) compared to the
432 mL/year estimate of Wong et al. (2014) suggests that the menstrual fluid loss can account
for > 30% of the difference in the elimination half-life between females and males.

   Harada et al. (2007) obtained serum and bile samples from patients (2 male and 2 female;
aged 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 fact that the levels in bile concentrations are
higher than in serum is supportive of bile as a route of excretion. The biliary resorption rate was
0.97, which could contribute to the long half-life in humans. Method of exposure to PFOS was
unknown.

   Biliary excretion in humans and the potential for resorption from bile discharged to the
gastrointestinal (GI) tract is supported by the Genuis  et al. (2010) self-study of the potential for
cholestyramine to lower the levels of PFAS in blood. Ingestion of 4 g/day cholestyramine (a bile
acid sequestrant) in three doses for 20 weeks decreased the PFOS serum levels from 23 ng/g
serum to 14.4 ng/g serum.

Animals. 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 and feces were collected after 24 and 48 hours. The amounts recovered in urine and feces
were approximately equivalent at each time point: 1.57% and 1.55%, respectively, at 24 hours
and 2.52% and 3.24%, respectively, at 48 hours.

   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: Group (G) 0 = 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 degrees Celsius (°C) prior to analyzing. Target analytes were
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-18

-------
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 and a similar trend was observed in the rats administered 20 mg/kg/day
PFOS. However, in the third week, mortalities occurred. By study day 24, there were only 2 out
of 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 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 steady state could have been the result of lower feces volume because the rats
had 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
the 5 mg/kg/day and 20 mg/kg/day groups 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).
                    S mo/kg PFOS exposure group (G3)
               1.2
                                               1.2
                                             O)
                                             E 0.9
                                             o
                                             O
                                             8
                                               0.6
       OJ
               0.0
                                               0.0
            20 mo/Kg PFOS exposure group (G4)
          (d)
          -•-Urlrt*
          -•—Feces
          —Overall
                       7     14     21
                       Exposure Time (Day)
28
7     14     21
Exposira Time (Day)
           Notes: No urine was available after day 18 in the 20 mg/kg/day group due to high mortality in this group.
           'Statistically-significant at p < 0.05
           "Statistically-significant at p < 0.01
  Figure 2-2. PFOS Contents in Urine, Feces, and Overall Excretion in Male Rats Treated
                                      for 28 Days

   Five groups of 16 female Crl:CD(SD)IGS VAF/Plus rats each were administered 0, 0.1, 0.4,
1.6, or 3.2 mg PFOS/kg bw/day by oral gavage beginning 42 days prior to cohabitation and
continuing through GD 14 or 20 (Luebker et al. 2005b). Urine and feces were collected
overnight from dams on the eve of cohabitation day 1 and during GDs 6-7, 14-15, and 20-21.
The concentrations in the feces were consistently about 5 times greater than in the urine (see
Table 2-10).

2.4.2   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
Perfluorooctane sulfonate (PFOS) - May 2016
                                            2-19

-------
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 PFASs 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 it was < 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 PFAS 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.

2.5  Pharmacokinetic Considerations

2.5.1   Pharmacokinetic models

   Toxicokinetic models that can accommodate half-life values that are longer than would be
predicted based on standard absorption, distribution, metabolism, and excretion concepts have
been published as tools to estimate internal doses for humans, monkeys, and rats. 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 account for the unique half-life properties of PFOS across species. The model structure
(Figure 2-3; 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.

   The model was parameterized using the body weight and urine output for cynomolgus
monkeys (Butenhoff et al. 2002, 2004) and a cardiac output of 15 liters (L)/hour (h)-kg from the
literature (Corley et al. 1990). Other parameters were  assumed or optimized to fit the best for
monkeys. In the intravenous time course data, some time and/or dose-dependent changes
occurred in distribution of PFOS between the blood and tissue compartments, and these changes
were less noticeable in the females, therefore, only the female data were used. The simulation
captured the overall time  course scenario but did not provide good correspondence with the
initial rapid loss from plasma and the apparent rise in  plasma concentrations over the first
20 days. For the oral dosing, the 0.15 mg/kg dose simulation was uniformly lower, and the
0.75 mg/kg dose simulation was higher than the data.  When compared to PFOA, PFOS had a
longer terminal half-life and more rapid approach to steady-state with repeated oral
administration.

Perfluorooctane sulfonate (PFOS) - May 2016                                            2-20

-------
                                          Tissue
                                        Compartment
                                          \ "tissue)
                            input
                          (iv, oral)
                                          Central
                                        Compartment
                                        
-------
                                                Oral dose
                                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
           Notes: Tm = transport maximum, Kt = affinity constant, and Q = flow in and out of tissues

               Figure 2-4. Structure of Model for PFOS in Rats and Monkeys
         IV-
Plasma
Free
fraction

.,,
QGut
,,
QLiv
-,,
QFat
^
QSkn
	 3>
QR
.
QKid
QFil
Gut
N f
Liver

Fat

Skin

Rest of body

Kidney
J \
Tm,Kt
Filtrate
N '
                                                            Oral dose, drinking water
                                          storage

                                              kurine

                                            urine
              Notes: Tm = transport maximum, Kt = affinity constant, and Q = flow in and out of tissues

         Figure 2-5. Structure of the PFOS PBPK Model in Monkeys and Humans
Perfluorooctane sulfonate (PFOS) - May 2016
                                                              2-22

-------
   Existing data sets for the cynomolgus monkey were used to develop the monkey model. The
IV data came from monkeys administered a single dose of 2 mg/kg, and the concentrations in
plasma and urine were monitored for up to 161 days after dosing (Noker and Gorman 2003). The
repeat-dose oral data were 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. (2002) 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 (Loccisano et al. 2011). 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
2.5.2). No measures of PFOS concentration were available for the drinking water at Little
Hocking, so 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 parts per billion (ppb).
The model results can be characterized as good when compared to the reported average serum
measurements. The average daily exposure, consistent with the serum value,  was estimated as
0.003 |ig/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 in order to refine the human model more data are
needed on the kinetics  of renal transporters and intrahuman variability, as well as definitive
information on exposures.

   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 Vd and chemical half-life (ty2).

                                  CL = Vd x (In2 H- t%)

                    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 2,200 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 a 70 kg body weight.
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-23

-------
   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 Andersen 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, Ohio area (see section 2.5.3). 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 1.7 to 3.6 ng/kg bw/day.

   Fabrega et al. (2014) adapted the Loccisano et al. (2011) model to include compartments for
the brain and lung and remove the skin. They applied the adjusted model to humans by using
intake and body burden data from residents in Tarragona County, Spain. Food and drinking
water were the major vehicles  of exposure. Body  burden information came from blood samples
from 48 residents, and tissue burdens came from 99 samples of autopsy tissues. The adjusted
model over-predicted serum levels by a factor of about two for PFOS but under-predicted the
levels in both liver (slightly) and kidney (by a factor of about 4).

   The authors also looked at the value of using partition coefficients from human tissues in
place of the Loccisano et al. (2011) rat data. The PFOS simulation values were closer to the
human experimental data when using the human partition coefficients values for liver, brain, and
kidney but not for the lung PFOS results. However, the Loccisano et al.  (2011) model
demonstrated better performance overall. The authors suggested that both saturable resorption
and variations in protein binding are important parameters for pharmacokinetic models. With the
exception of serum albumin, the existing models have not considered protein binding constants
within tissues. Even though the use of human partition coefficients improved the steady state
predictions overall for tissues there were still considerable differences between the experimental
values and the predictions for both models.

   Loccisano et al. (2012a) utilized the saturable resorption hypothesis and pharmacokinetic
data from Chang et al. (2012),  3M Environmental Laboratory (2009), and Seacat et al. (2003) for
adult Sprague-Dawley rats to develop the model depicted in Figure 2-6.  The structure of the
model is similar to that for the monkey/human model depicted in Figure 2-5 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) through personal communication to authors (Loccisano et al. 2012a); 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.

   The agreement between the experimental data and the model output was 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 two. 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-24

-------
                          Oral, diet
                                                                feoes
   Figure 2-6. Structure of the PBPK Model for PFOS in the Adult Sprague-Dawley Rat

   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 GDs 0 to 20 were used in model development. Both studies
used multiple dose levels in addition to data on serum and selected tissue concentrations (liver,
brain) from the dams and fetuses at one or more time points. The gestational model structure for
the dams is similar to Figure 2-6. 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  Chang et al. (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 employed in
the Butenhoff et al. (2009), Luebker et al. (2005a, 2005b), Yu et al. (2009a), and Lau et al.
(2003) studies as depicted in Figure 2-7.
Perfluorooctane sulfonate (PFOS) - May 2016
2-25

-------
                                    5 mg/kg: over
                                    95% of neo nates
                                    did not survive
                                    PND1 (Lau. et al
                                    2003).
10 mg/kg: Significant
reduction in fetal weight:
significant increase in cleft
palate and other
birth defects: all neortates
died within 30-60 min. after
birth (Lau. et a!2003
                    •1 mg/kg: LOAEL fa-
                    developmental
                    neurotoxicity in male
                    pups after birth
                    (Butenhoff. et al
                    2009)
           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).
   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 thyroxine levels in
 neon at es(Yu.etal 2009).
 Figure 2-7. 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-tissue 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.

    Loccisano et al. (2013) extended their model development to cover humans during pregnancy
and lactation, building on the work done with rodents and recognizing the limitations of the
human data available for evaluating the model predictions. Figure 2-8 illustrates the structure of
the model used. The basic structure was derived from the rat model discussed above. Some of the
key features of the model are summarized below:

    •   The fetus is exposed via the placenta through simple bidirectional diffusion.
    •   Transfer rates to the fetus from the amniotic fluid are governed by bidirectional diffusion.
    •   Transfer from the fetal  plasma to tissues is flow-limited.
    •   Maternal plasma is directly linked to the milk compartment and considered to be flow
       limited; only the free fraction in plasma is transferred to maternal milk.
    •   The neonate is exposed to PFOS only via maternal milk for the first 6 months
       postpartum.
    •   The infant in the model is treated as one compartment with a volume of distribution.
Perfluorooctane sulfonate (PFOS) - May 2016
                                    2-26

-------











Plasma
(free

available
for uptake
into












h 	 1
QGut

£
^
QLIv
€ 	
QFat


•y
noin


QSk

^


_,
•s.
no
UK
__,
^
rwiH

QFII

fiut

4




Fat


Placenta


Qltin


Mammary



Rest of body



Kidney

A
Filtrate

r- Oral dose
Drinking 1






ktransl

if'

ktrans2













Tm.Kt
N storage


H2O










plasma
(free
fraction)












	 »








Amniotic fluid I
I
* ^
ktrsnsS ktrsns^

w

Rest of fetal body
QRFet











urine

           T
                    IV
   Figure 2-8. PBPK Model Structure for Simulating PFOA and PFOS Exposure During
                   Pregnancy in Humans (Maternal, Left; Fetal, Right)

   Limitations to the model are acknowledged and attributed primarily to lack of data to support
a more mechanistic approach. Physiological parameters applicable to a pregnant or lactating
woman, the fetus, and the nursing infant were obtained from a variety of referenced publications.

   In order to obtain a plasma value at the time of conception, the model was run until it reached
a prepregnancy steady state concentration. The model predicted 30 years as the exposure
necessary to reach steady state (1 x 10"4 to 2 x 10"3 |ig/kg bw/day) for the general female
population. The model performance simulations for PFOS were run using an exposure of
1.35 x  icr3 ng/kg bw/day. Projections were developed for maternal plasma, fetal plasma, infant
plasma, and maternal milk. Agreement between the observed concentrations (|ig/L) and the
predicted values was considered satisfactory if the predicted value was within 1% of the
observed value. Model output was compared to maternal and fetal plasma values at delivery or at
specific time points, and for the infant plasma and milk data where available.  Predicted
maternal:fetal plasma (cord blood) concentration ratios were more variable for PFOS than PFOA
in the comparisons to the published data. The projections for fetal internal dose were reasonable,
and there was good agreement between the model and the available human lactation data. When
modeled, the maternal plasma was 14 jig/L at conception, slowly decreased across the gestation
period, and increased slightly at delivery. For the most part the modeled results fell within ± 1
SD of the observed data.
Perfluorooctane sulfonate (PFOS) - May 2016
2-27

-------
   During lactation there was a gradual, very-slight, decline in maternal plasma across the six
months of lactation. Thereafter, plasma values slowly increased and stabilized at about 12.5 jig/L
at six months postpartum. The fetal plasma was about 6.5 jig/L at the start of gestation, and
declined to about 5.5 |ig/L at the time of delivery. Maternal plasma values are about twice those
for the fetus. During the lactation period, the infant plasma increased in a linear fashion to a
terminal value of about 13 |ig/L. Milk concentrations declined very slightly across the lactation
period with an initial concentration of 0.16 |ig/L and a final value of 0.15 |ig/L. These
concentrations were estimated from the graphic data presentation.

   The projections for PFOS differed from those for PFOA in several respects. Most
importantly maternal and fetal plasma values were similar for PFOA but for PFOS, maternal
levels were approximately two-fold greater than fetal levels. Compared with PFOS, there was a
much greater decline in maternal PFOA plasma values during lactation accompanied by a
comparable decline in the PFOA concentration in milk. The increase in infant plasma across the
lactation period was comparable for PFOA and PFOS with the concentration at 6 months
postpartum about 2.5 times higher  than that at 1 month.

   The authors compared the human pregnancy lactation model results to published  data, and
they identified several important research needs as follows:

   •  Are there differences in the transporter preferences and transfer rates for the individuals
       PFASs? Do those differences correlate with half-life differences?
   •  Are there qualitative or quantitative differences between the transporters favored by
       PFOS compared to PFOA?
   •  What physiological factors  influence  clearance for the mother, the fetus, and the infant
       during gestation and lactation?
   •  Are placental transport processes active, facilitated, or passive?

These research needs are more pronounced for PFOS than PFOA, because the information
supporting renal resorption and tissue uptake via membrane transporters for PFOS is very
limited. Most models infer that PFOS and PFOA are similar based on their half-lives rather than
on published research on transporters.

   The authors acknowledged the  lack of primary experimental data on PFOS transport and
potential transporters. Similarity to PFOA was assumed in model development, and PFOS was
transparently described as lacking  supporting transporter data.  The authors concluded that
additional research on PFOS binding to serum proteins and liver tissues, its biliary excretion and
resorption, and information on renal resorption transporters in dams and pups are needed to
accomplish further refinements to the published model (Loccisano et al. 2012b, 2013).

   Building on the work of other researchers, Wambaugh et al. (2013) developed and published
a pharmacokinetic model to support the development of an EPA reference dose for PFOS. The
model was applied to data from studies conducted in monkeys, rats,  or mice that demonstrated an
assortment of systemic, developmental, reproductive, and immunological effects. A saturable
renal resorption pharmacokinetic (PK) model was again used. This concept has played  a
fundamental role in the design of all of the published PFOS models summarized in this section.
In this case, an oral dosing version of the original model introduced by Andersen et al. (2006)
and summarized early in section 2.5.3 was selected for having the fewest number of parameters
that would need to be estimated. A unique feature of the Wambaugh et al. (2013) approach was
to use a single model for all species in the toxicological studies in order to examine the


Perfluorooctane sulfonate (PFOS) - May 2016                                             2-28

-------
consistency in the average serum values associated with effects and with no effects from 13
animal studies of PFOS. The model structure is that depicted in Figure 2-3 with minor
modifications.

   Wambaugh et al. (2013) placed bounds on the estimated values for some parameters of the
Andersen et al. (2006) model to support the assumption that serum carries a significant portion of
the total PFOS body load. The Andersen et al. (2006) model is a modified two-compartment
model in which a primary compartment describes the serum and a secondary deep tissue
compartment acts as a specified tissue reservoir. Wambaugh et al. (2013) constrained the total
volume of distribution to a value of not > 100 times that in the serum. As a result, the ratio of the
two volumes (serum versus total) was estimated in place of establishing a rate of transfer from
the tissue to serum.

   A nonhierarchical model for parameter values was assumed. Under this assumption a single
numeric value represents all individuals of the same species, gender,  and  strain. The gender
assumption was applied to the  monkeys and mice, while male and female rats were treated
separately because of the established gender toxicokinetic differences. Body weight, the number
of doses, and magnitude of the doses were the only parameters to vary. In place of external
doses, serum concentrations as measured at the time of euthanasia were used as the metric for
dose magnitude. Measurement errors  were assumed to be log-normally distributed. Table 2-16
provides the estimated and assumed PK parameters applied in the Wambaugh et al. (2013) model
for each of the species evaluated.

   The PK data that supported the analysis were derived from two PFOS PK in vivo studies. The
monkey PK data were derived from Seacat et al. (2002) and Chang et al. (2012). Data for the rats
(male/females) and mice were both from Chang et al. (2012). The data were analyzed within a
Bayesian framework using a Markov  Chain Monte Carlo sampler implemented as an R package
developed by EPA to allow predictions across species, strains, and genders and identify serum
levels associated with the no observed adverse effect level  (NOAEL) and lowest observed
adverse effect level (LOAEL) external doses. The model chose vague, bounded prior
distributions on the parameters being  estimated allowing them to be significantly informed by
the data. The values were assumed to be log normally distributed constraining each parameter to
a positive value.

   The model predictions were evaluated by comparing each predicted final serum
concentration to the serum value in the supporting animal studies. The predictions were generally
similar to the experimental values. There were no systematic differences between the
experimental  data and the model predictions across species, strain,  or gender, and median model
outputs uniformly appeared to  be biologically plausible despite the uncertainty reflected in some
of the 95th percentile credible intervals. The application of the model outputs in the derivation of
a human RfD is the focus of section 4 of this document.
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-29

-------
  Table 2-16. Pharmacokinetic Parameters from Wambaugh et al. (2013) Meta-Analysis of
                                       Literature Data
Parameter
Bodyweightb
Cardiac
Output0
*a
Fcc
ku
Rv2:Vl
T
i maxc
^T
Free
gfilc
Vfl\C
Units
kg
L/h/kga74
L/h
L/kg
L/h
Unitless
umol/h
umol
Unitless
Unitless
L/kg
CD1 Mouse (F)a
0.02
8.68
1.16(0.617-
42,400)
0.264 (0.24-
0.286)
0.0093 (2.63 x
e-10- 38,900)
1.01 (0.251-
4.06)
57.9 (0.671-
32,000)
0.0109 (1.44 xe-5
-1.45)
0.00963
(0.00238-
0.0372)
0.439 (0.0125-
307)
0.00142 (4.4 x
e-10-6.2)
CD1 Mouse
(M)a
0.02
8.68
433.4(0.51-
803.8)
0.292 (0.268-
0.317)
2,976 (2.8 x e-10
-4.2 x e4)
1.29 (0.24-4.09)
l.lxe4(2.1-7.9
xe4)
381 (2.6 xe-5-
2.9 x e3)
0.012 (0.0024-
0.038)
27.59 (0.012-
283)
0.51 (3.5 xe-10
-6.09)
Sprague-
Dawley Rat (F)a
0.203
12.39
4.65 (3.02-
1,980)
0.535 (0.49-
0.581)
0.0124 (3. Ix
e40 -46,800)
0.957 (0.238-
3.62)
1,930(4.11-
83,400)
9.49 (0.00626-
11,100)
0.00807
(0.00203-
0.0291)
0.0666 (0.0107-
8.95)
0.0185 (8.2 xe'7
-7.34)
Sprague-
Dawley Rat
(MT
0.222
12.39
0.836 (0.522-
1.51)
0.637 (0.593-
0.68)
0.00524 (2.86 x
e-10-43,200)
1.04 (0.256-
4.01)
1.34xe-6(1.65x
e-10-44)
2.45 (4.88 xe-10
-60,300)
0.00193
(0.000954-
0.00249)
0.0122(0.0101-
0.025)
0.000194 (1.48 x
e-9-5.51)
Cynomolgus
Monkey
(M/F)a
3.42
19.8
132 (0.225-
72,100)
0.303 (0.289-
0.314)
0.00292 (2.59 x
e-10-34,500)
1.03 (0.256-
4.05)
15.5 (0.764-
4,680)
0.00594 (2.34 x
e-5-0.0941)
0.0101
(0.00265-0.04)
0.198(0.012-
50.5)
0.0534 (l.lx
e-7-8.52)
Notes: Means and 95% confidence interval (in parentheses) 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
b Average bodyweight for species:individual-specific bodyweights
c Cardiac outputs obtained from Davies and Morriss 1993
Qfiic = median fraction of blood flow to the filtrate
Tmax = time of maximum plasma concentration
M = male; F = female


2.5.2  Half-life data

   Differences between species were observed in studies determining the elimination half-life
(Ti/2) of PFOS in rats, mice, monkeys, and humans. Gender differences in rats do not appear to
be as dramatic for PFOS as they are for PFOA (Loccisano et al. 2012a, 2012b).

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 sample collection and methods.
Perfluorooctane sulfonate (PFOS) - May 2016
2-30

-------
   More recently, Olsen et al. (2007) obtained samples from 26 retired fluorochemical
production workers (24 males and 2 females) from the 3M plant in Decatur, Alabama to
determine the half-life of PFOS. Periodic serum samples (total of 7-8 samples per person) were
collected over a period of 5 years, stored at -80 °C, and at the end of the study, High-
performance liquid chromatography/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: 20-36 years), the mean age of the participants at the initial blood sampling was 61 years
(range: 55-75 years), and the average number of years retired was 2.6 years (range:  0.4-11.5
years). The initial arithmetic mean serum concentration of PFOS was 0.799 |ig/mL (range:
0.145-3.490 jig/mL), and when samples were taken at the end of the study the mean serum
concentration was 0.403  |ig/mL (range: 0.037-1.740  jig/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% confidence interval [CI]: 3.9-6.9 years) and
4.8 years (95% CI: 4.1-5.4 years), respectively.

General Population. No data on the half-life of PFOS in the general population were identified.

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 the NSPs collected in
the state of New York from infants born between 1997 and 2007 were analyzed for PFOS
(Spliethoff et al. 2008). The  analytical methods were  validated by using freshly drawn blood
from healthy adult volunteers. The mean whole blood concentration for PFOS ranged from
0.00081 to 0.00241 |ig/mL. The study grouped the blood spots by two different time-points;
those collected in 1999-2000 and in 2003-2004, which corresponded to the intervals reported by
NHANES. The PFOS concentrations decreased with  a mean value of 0.00243 |ig/mL reported in
1999-2000 and 0.00174 |ig/mL in 2003-2004. The study authors determined the half-life of
PFOS using the regression slopes for natural log blood concentrations versus the year 2000 and
after. The calculated half-life for PFOS was 4.1 years.

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 administered PFOS for 26 weeks (Seacat et
al. 2002). 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 of 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 122-146 days in male monkeys
and 88-138 days in females. Mean values are shown  in Table 2-17. The Vd values suggest that
distribution was predominately extracellular.
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-31

-------
             Table 2-17. PFOS Pharmacokinetic Data Summary for Monkeys



Species
Cynomolgus
monkeys

Time
evaluated
after last dose
23 weeks




Route
IV




Sex
M

F

Amount
K+PFOS
(mg/kg)
2

2

Mean serum
Ti/2 by sex
(days)
132.0 ±7

110.0 ±15
Mean serum
Tiflby
species
(days)
120.8

Mean
serum Vd
by sex
(mL/kg)
202

274
Source: Data from Chang et al. 2012
M = male; F = female
    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 1 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 2-18).
First, a single oral dose of 4.2 mg 14C-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 should 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.

    In a third study, serum uptake and elimination of PFOS were evaluated at two dose levels: 2
mg/kg and 15 mg/kg. 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 2-18) 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 35-53 days and that for females ranged 33-55 days.
Perfluorooctane sulfonate (PFOS) - May 2016
2-32

-------
               Table 2-18. PFOS Pharmacokinetic Data Summary for Rats
Species
SD rats
SD rats
SD rats
SD rats
Time
evaluated
after last
dose
144 hours
24 hours
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
1 x 28 days
1 x 28 days
2
15
2
15
Mean
serum Tin
by dose
(days)
8.2 ±1.5
3.1ab
1.9b
8.0b
5.6a
35-53
33-55
38.3 ±2.3
41.2 ±2.0
62.3 ±2.1
71. 1± 11.3
Mean
serum
Tinby
sex (days)

Mean
serum TIB
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 a
521
649
586 a
-
-
1,228
666
484
468
Source: Data from Chang et al. 2012 and Butenhoff and Chang 2007
Notes: aData reflected a single value derived from one rat only
b Within limits of the study design and a follow-up duration of only 24 hours
NA= not available
M = male; F = female

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 of
the observation period, the daily urinary and fecal excretion was < 0.1% of the administered
dose. Results are shown in Table 2-19. Serum elimination values were similar for males and
females, independent of dose administered (distribution appeared to be mostly  extracellular).

               Table 2-19. PFOS Pharmacokinetic Data Summary for Mice




Species



CD-I mice



Time
evaluated
after last
dose



20 weeks






Route



Oral






Sex









Amount
K+PFOS
(mg/kg)
1

20


1
20
Mean
serum
Tinby
dose
(days)
42.8

36 4


37.8
30.5

Mean
serum
Tinby
sex (days)





7/1 i
34. z
Mean
serum
Tinby
species
(days)



36.9



Mean
serum Vd
by dose
(mL/kg)
290.0

263 0


258.0
261.0
Source: Data from Chang et al. 2012
M = male; F = female

   Table 2-20 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
Perfluorooctane sulfonate (PFOS) - May 2016
2-33

-------
half-life in a retired worker population is 5.4 years (Olsen et al. 2007), compared with several
months in the laboratory animals.

                         Table 2-20. Summary of Half-Life Data
Source
Spliethoffetal. 2008
3M Company 2000
Olsen et al. 2007
Butenhoff and Chang 2007
Chang etal. 2012
Seacat et al. 2002
Human
4.1 years
4-8.67 years
5. 4 years
ND
ND
ND
ND
ND
Monkey
ND
ND
ND
ND
ND
ND
132 days (M)
110 days (F)
200 days (M/F)
Rat
ND
ND
ND
48.2 days (M)
46.9 days (F)
39.8 days (M)
66.7 days (F)
ND
ND
ND
Mouse
ND
ND
ND
ND
ND
39.6 days (M)
34.2 days (F)
ND
ND
Strain
Infants
Occupational
Occupational
SD; 28 days oral
SD; single oral
dose
CD-I; single oral
dose
Cynomolgus;
single IV dose
Cynomolgus;
oral, 182 days
Note: ND = No Data
M = male; F = female

    The animal data summarized in Table 2-20 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 male and female monkeys, half-life values were similar for 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.

2.5.3   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. 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 using a simple, first-order,
one-compartment 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 it was assumed that
most PFOS intake was 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.
Perfluorooctane sulfonate (PFOS) - May 2016
2-34

-------
   Thompson et al. (2010) used a first-order, one-compartment 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 Andersen et al.  (2006). The
original Andersen 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 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 3,000 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 an exposure
assessment. The authors concluded that the lower value of 200 mL/kg was appropriate for 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 Chang et al. (2012) series of pharmacokinetic studies on rats,  mice, and monkeys
described above, included volume of distribution calculations. 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 monkeys, female rats, and male and female
mice were reasonably similar.

   The volume of distribution was 202 and 274 mL/kg, for male and female cynomolgus
monkeys, respectively (Table 2-17), following a single IV dose of 2 mg/kg (Chang  et al. 2012).
Animals were evaluated up to 23 weeks after dosing, and the resulting volumes of distribution
are similar to the 230 mL/kg calibrated from human data by Thompson et al. (2010) described
above.

   The Chang et al. (2012) volume of distribution findings for rats are in Table 2-18. 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 than that of female rats  or other species including  humans, with
the exception of one value. 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                            2-35

-------
   Data for mice (Chang et al. 2012) are shown in Table 2-19. For males and females the
volume of distribution was 263-290 mL/kg and 258-261 mL/kg, respectively, following a single
oral dose.

   Pharmacokinetic models based on animal data described previously in this section generally
optimized the value for volume of distribution based on model output. The original Andersen 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.

2.6  Toxicokinetic Summary

   Uptake and egress of PFOS from cells is largely regulated by transporters in cell membranes
based on data collected for PFOA, a structurally similar chemical. On the basis of the tissue
concentrations found in the pharmacokinetic studies (Cui et al. 2009; Curran et al. 2008), PFOS
is absorbed from the gastrointestinal tract, as indicated  by the  serum measurements in treated
animals, and distributed to  the tissues. The highest tissue concentrations are usually those in the
liver. Post mortem tissues samples collected from 20 adults in Spain found PFOS in liver,
kidney, and lung (Perez et al. 2013). The levels in brain and bone were low. In serum, PFOS is
electrostatically bound to albumin occupying up to eleven sites (Weiss et al. 2009). Linear PFOS
chains display stronger binding than branched chains (Beesoon and Martin 2015). Binding
causes a change in the conformation of serum albumin  (Weiss et al. 2009) thereby changing its
affinity for the endogenous compounds also transported by serum albumin. PFOS binds to other
serum proteins including immunoglobulins and transferrin (Kerstner-Wood et al. 2003). It is not
metabolized, thus any effects observed in toxicological studies are not the effects of metabolites.

   Electrostatic interactions with proteins are an important toxicokinetic feature of PFOS.
Studies demonstrate binding or interactions with nuclear receptors (e.g., PPARa), transport
proteins (e.g., transthyretin [TTR]), FABP), and enzymes (Luebker et al. 2002; Ren et al. 2015;
L. Wang et al. 2014; Weiss et al. 2009; Wolf et al. 2008; L. Zhang et al. 2013, 2014).  Saturable
renal resorption of PFOS from the glomerular filtrate via transporters in the kidney tubules is
believed to be a major contributor to the  long half-life of this compound. No studies were
identified on specific renal  tubular transporters for PFOS, but  many are available for PFOA. All
toxicokinetic models for PFOS and PFOA are built on the concept of saturable renal resorption
first proposed by Anderson et al. (2006). Some PFOS is removed from the body with bile (Chang
et al. 2012; Harada et al. 2007), a process that is also transporter-dependent. Accordingly, the
levels in fecal matter represent both unabsorbed material and that discharged with bile.

   The arithmetic mean half-life in humans for occupationally exposed workers (Olsen et al.
2007) was 5.4 years (95% CI: 3.9-6.9 years). Half-lives from  animals include 120.8 days for
monkeys, 33-35 days for male and female  Sprague-Dawley rats, and 36.9 days for male and
female CD1 mice (Chang et al. 2012). The half-life differences between male and female rats
observed for PFOA were not observed with PFOS. This indicates a lack of sex-related
differences in renal excretion in rats and  implies that the renal excretion and/or resorption
transporters for PFOS differ from those for PFOA. No  comprehensive studies of PFOS
transporters in humans or laboratory animals were identified.
Perfluorooctane sulfonate (PFOS) - May 2016                                             2-36

-------
3.  HAZARD IDENTIFICATION

   The Hazard Identification section provides a summary and synthesis of the data from a large
number of human epidemiology studies accompanied by studies in laboratory animals designed
to identify both the dose-response and critical effects that result from exposures to PFOS and to
examine the mode of action leading to toxicity.

3.1   Human Effects

   There is a substantial body of research on the adverse effects of PFOS in both humans and
animals. The human database lacks data on acute effects and short term exposures, but it
includes many epidemiology studies. The database of human studies is large, in part, due to the
extensive research program conducted by the C8 Science Panel on residents of communities in
Ohio and West Virginia that were impacted by PFOA discharges from the DuPont Washington
Works plant in Parkersburg, West Virginia. The purpose of the C8 Health Project is to assess if
there are any probable links between PFOA (and PFOS) exposure and disease. During the period
August 2005-July 2006, about 69,000 study participants were identified. Eligible participants
included those who had consumed drinking water for at least one year up to and including
December 4, 2004 from the (1) Lubeck  and Mason County water districts in West Virginia;
(2) the Belpre, Little Hocking, Tuppers  Plains-Chester, and Pomeroy water districts in Ohio; or
(3) private water source within the geographical boundaries of the public water sources.  The
participants (n = 69,030; 33,242 males,  35,788 females; aged < 10 to 70 years and older) donated
a blood sample, filled out an extensive questionnaire, and received $400 in compensation.
Although the project was designed to examine the impact of PFOA on health effects among
residents of the impacted community, the serum was analyzed for other perfluorochemicals,
including PFOS. Medical records were used to validate diseases reported by participants. The C8
Science panel studies were funded by DuPont under a consent decree. Some of the studies
evaluated the impact of PFOS (or PFOA) on outcome.

   Commercial use of PFOS and other PFASs began over 60 years ago, resulting in global
release of this family of compounds. As a result, population monitoring of serum is widespread
and has supported multiple epidemiological investigations of the general population within the
United States and abroad. Occupational epidemiology  studies are available from 3M, a U.S.
manufacturer of PFOS. Studies investigating the association between PFOS levels and health
effects in the U.S. general population have also been conducted using the NHANES data set. The
NHANES examined representative members of the U.S.  population through their surveys
focusing on different health topics. These studies consist of an interview (demographic,
socioeconomic, dietary, and medical questions) and examination (medical including blood and
urine collection, dental, and physiological parameters).

   A study by Jain (2014) examined the influence of diet and other factors on the levels of
serum PFOS and other PFASs using the NHANES 2003-2004, 2005-2006, and 2007-2008 data.
Significantly higher serum PFOS levels were found in males (0.020 jig/mL) compared to
females (0.014 jig/mL). There was a significant decreasing trend in serum PFOS concentration
between 2003 and 2008.  There was a positive association of PFOS with increases in  serum
cholesterol (p < 0.01) and serum albumin (p < 0.01) in the 5,591 records used for the assessment.
Intakes of meat and fish were positively associated with serum PFOA (p < 0.01).
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-1

-------
3.1.1   Long-Term Noncancer Epidemiological Studies

3.1.1.1  Serum Lipids and Cardiovascular Diseases

Occupational studies. Cross-sectional, as well as a longitudinal analyses 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 200la, 200Ib, 2003b). In the cross-sectional study, male (n = 215) and
female (n = 48) volunteers working at the Decatur plant and male (n = 206) and female (n = 49)
volunteers working at the Antwerp pi ant underwent clinical chemistry tests to evaluate hepatic
enzyme activity, renal function, thyroid activity, and cholesterol levels. Data on employees from
both plants appeared to be combined for the regression analyses; however, it was not clear
whether females were included or whether the analyses only included males. The mean PFOS
level in all employees from the Decatur and Antwerp plants was 1.40 |ig/mL (range: 0.11-
10.06 |ig/mL) and 0.96 |ig/mL (range: 0.04-6.24 jig/mL), respectively. Positive significant
associations were reported between serum PFOS and cholesterol (probability [p] = 0.04) and
between serum PFOS and triglycerides (p = 0.01); similar results were found for PFOA. Age was
also significant in both analyses. Alcohol consumed per day was significant in the cholesterol
model, while body mass index (BMI) and cigarettes smoked per day was significant for
triglycerides. PFOS was positively associated with alkaline phosphatase (ALP). Hepatic enzymes
and bilirubin were not associated with PFOA. However, there were many limitations to
combining and comparing the data from the two plants.

    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 lipids
(Olsen et al. 2001b, 2003b). The medical surveillance data from 175 individuals who had
participated in two or more medical exams in 1995, 1997, and 2000 were  analyzed using
multivariable regression. Mean PFOS levels at the beginning and end of the surveillance period
were 2.62 |ig/mL and 1.67 |ig/mL, respectively, in Decatur employees and 1.87 |ig/mL and
1.16 |ig/mL, respectively, in Antwerp employees. 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 HDL, ALP, gamma-glutamyl transpeptidase (GGT), aspartate
aminotransferase (AST), or alanine transaminase (ALT) activities, total bilirubin, or direct
bilirubin. PFOA was positively associated with cholesterol and triglycerides in  the Antwerp
employees.

High-exposure community studies. The C8 Health Project conducted in 2005-2006 on
approximately 69,000 residents in Ohio and West Virginia evaluated general population
exposures to PFOS and other perfluorochemicals. Public drinking water was contaminated in six
water districts surrounding the plant (> 0.05 ng/mL of PFOA). 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. The levels of PFOA were  elevated, however, levels  of PFOS in this
population were similar to those reported in the general U.S. population (median 0.02 |ig/mL).

    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 0.022 jig/mL, with a range of 0.00025-0.7592 jig/mL. Lipid outcomes

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-2

-------
(total cholesterol, HDL cholesterol, LDL cholesterol, and triglycerides) were examined in
relation to PFOS and PFOA serum levels. All lipid outcomes, except for FIDL, showed
significant increasing trends with increasing PFOS levels (similar for PFOA). The predicted
increase in cholesterol from lowest to highest PFOS decile was 11-12 mg/deciliter (dL). Logistic
regression analyses indicate statistically-significant incidence of hypercholesterolemia
(> 240 mg/dL) with increasing PFOS serum levels. Cholesterol levels > 240 mg/dL are
characterized as high, and medical intercession is recommended. The odds ratios (ORs) across
quartiles for cholesterol > 240 mg/dL were 1.00, 1.14 (95% CI:  1.05-1.23), 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.

   Frisbee et al. 2010 evaluated 12,476 children < 18 years old  who lived in the C8 Health
Project communities for total cholesterol, LDLs, HDLs, and fasting triglycerides. The mean level
of PFOS was 0.023 |ig/mL. 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 high total cholesterol [OR 1.6 (1.4-1.9)]
and LDL-cholesterol [OR 1.6 (1.3-1.9)] was also observed between the first and fifth quintiles of
PFOS serum levels. No trends were observed with triglycerides. Total cholesterol, LDL, and
triglycerides were also positively associated with serum PFOA concentration. As with the other
C8 project data, the authors acknowledge that the cross-sectional nature of this study limits
causal  inference.

   A cohort of 521 adult members of the C8 Health Project was evaluated for an association
between changes in serum PFOS levels and changes in serum LDL-cholesterol, HDL-
cholesterol, total cholesterol, and triglycerides over a 4.4-year period (Fitz-Simon et al. 2013).
Linear regression models were fit to the logarithm (base 10) of ratio change in each serum  lipid
measurement in relation to the logarithm of ratio change in PFOS. Mean serum PFOS
concentration decreased by approximately one-half between baseline (0.023 ± 0.014 |ig/mL) and
follow-up  (0.011 ± 0.007 |ig/mL). No  corresponding changes in serum lipids were found.
However, those individuals with the greatest declines in serum PFOS had a tendency for a  slight
decrease in LDL-cholesterol. Similar results were found with PFOA.

   A subset of 290 individuals in the C8 Health Project was evaluated for evidence that PFOS
exposure can influence the transcript expression of genes involved in cholesterol metabolism,
mobilization, or transport (Fletcher et al. 2013). Ribonucleic acid (RNA) was extracted from
whole blood samples taken from 144 males and 146 females aged 20-60 years;  serum collected
at the same time was used to measure PFOS concentration. The  association between candidate
gene expression levels and PFOS levels was assessed by multivariable linear regression with
adjustments  for confounders. A positive association was seen between PFOS and a transcript
involved in cholesterol mobilization (Neutral Cholesterol Ester Hydrolase 1  [NCEH1];
p = 0.018), and a negative relationship with a transcript involved in cholesterol transport
(Nuclear Receptor Subfamily 1, Group H, Member 3 [NR1H3];  p = 0.044). When sexes  were
analyzed separately, PFOS was positively associated with expression of genes involved in
cholesterol mobilization and transport in females (NCEH1 and Peroxisome Proliferator-
Activated Receptor alpha [PPARa]; p  = 0.003 and 0.039, respectively), but no effects were
evident in  males. Similar associations were also found for PFOA.

General population studies. Nelson et al. (2010) used NHANES 2003-2004 data to analyze
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

Perfluorooctane sulfonate (PFOS) - May 2016                                              3-3

-------
was not measured directly, but was estimated by the Friedewald formula2 as recommended by
Centers for Disease Control and Prevention. Homeostatic model assessment (HOMA) was used
to assess insulin resistance (calculated from fasting insulin and fasting glucose measurements
collected in NHANES). 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 females or insulin use. After exclusion criteria, approximately
860 participants were included in the analyses. The mean PFOS serum concentration for
participants 20-80 years old was 0.025 |ig/mL (range: 0.0014-0.392 |ig/mL).

   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 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. Similar results were found for PFOA.

   Lin et  al. (2009) explored associations of serum lipid levels with NHANES PFOA data from
1999-2000 and 2003-2004. Serum HDL was inversely associated with serum PFOS
concentration OR ((95% CI): 1.61 (1.15-2.26), p < 0.05). Triglycerides did not show an
association with PFASs.

   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 females). 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 total cholesterol (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 females and males was 0.019 jig/mL.

   Eriksen et al. (2013) examined the association between plasma PFOS levels and total
cholesterol levels  in a middle-aged Danish population. This cross-sectional study included
663 males and 90  females aged 50-65 years who were enrolled in the Danish Diet, Cancer and
Health cohort. Generalized linear models were used to analyze the association between PFOS
and total cholesterol levels and adjusted regression analyses were performed. The mean plasma
PFOS level was 0.0361 jig/mL. A significant,  positive association was found between PFOS
(and PFOA) and total cholesterol such that in the fully adjusted model, a 4.6 mg/dL (95% CI:
0.8-8.5) higher concentration of total cholesterol was found per interquartile range of plasma
PFOS. The quartiles of PFOS used in the analyses were not defined and no comparison was
made for cholesterol levels between the highest and lowest PFOS quartile.
2 Friedewald formula: [LDL-cholesterol] = [total cholesterol] - [HDL-cholesterol] - [triglycerides/5]. All values are
expressed in mg/dL units.

Perfluorooctane sulfonate (PFOS) - May 2016                                              3-4

-------
   A cross-sectional study of 891 pregnant females evaluated the association between plasma
PFOS levels and plasma lipids (Starling et al. 2014). Six other perfluoroalkyl substances were
also quantified and evaluated. The females were a cohort of the Norwegian Mother and Child
Cohort Study, and the majority of blood samples were drawn during weeks 14-26 of gestation.
Weighted multiple linear regression was used to estimate the association between PFOS level
and each lipid level. The median plasma PFOS level was 0.013 |ig/mL. No association was
observed between PFOS and triglycerides. PFOS was positively associated with total cholesterol,
HDL-cholesterol, and LDL-cholesterol, although confidence intervals were broad for all
associations. Each In-unit increase in PFOS was associated with an increase of 8.96 mg/dL
(95% CI: 1.70-16.22) in total cholesterol and for each interquartile range (IQR)-unit increase in
the In-PFOS concentration, total  cholesterol increased by 4.25 mg/dL (95% CI: 0.81-7.69). With
HDL-cholesterol, each IQR-unit increase in In-PFOS was associated with an increase of
2.08 mg/dL (95% CI: 1.12-3.04). For LDL-cholesterol, each IQR-unit shift in In-PFOS was
associated with a change of 3.07 mg/dL LDL (95% CI: -0.03-6.18). Five of the seven PFASs
studied were positively associated with HDL cholesterol, and all seven had elevated HDL
associated with the highest quartile.

   Fisher et al. (2013) examined the association of plasma PFAS levels, including PFOS, with
metabolic function and plasma lipid levels. This cross-sectional study included 2,700
participants, aged  18-74 years (approximately 50% male), in the Canadian Health Measures
Survey. Multivariate linear and logistic regression models were used for analyses of associations
between PFOS levels and cholesterol outcomes, metabolic syndrome, and glucose homeostasis.
The geometric mean PFOS concentration was 0.0084 ± 0.002 jig/mL. In weighted analyses, no
association was found between PFOS (or PFOA) and total cholesterol, HDL- and LDL-
cholesterol, and metabolic syndrome and glucose homeostasis parameters. Hypercholesterolemia
(cholesterol greater than 240 mg/dL), was associated with PFOS exposure in unadjusted analyses
of this cohort.

   Multiple epidemiologic studies have evaluated  serum lipid status in association with PFOS
concentration (Table 3-1). These studies provide support for an association between PFOS and
small increases in total cholesterol in the general population at mean serum levels of 0.0224-
0.0361 |ig/mL (Frisbee et al. 2010; Nelson et al. 2010; Eriksen et al. 2013).
Hypercholesterolemia, (clinically defined as cholesterol greater than 240 mg/dL), was associated
with PFOS exposure in a Canadian cohort (Fisher et al. 2013) and in the C8 cohort (Steenland et
al. 2009). Cross-sectional occupational studies demonstrated an association between PFOS and
total cholesterol (Olsen et al. 2001a, 2001b, 2003b). Evidence for associations between other
serum lipids and PFOS is mixed, including HDL cholesterol, LDL, very low density lipoprotein
(VLDL), non-HDL cholesterol, and triglycerides. The  studies on serum lipids in association with
PFOS serum concentrations are largely cross-sectional in nature and were largely conducted in
adults, but some studies exist on children and pregnant females.  The location of these cohorts
varied from the U.S. population including NHANES volunteers, to the Avon cohort in the United
Kingdom (UK), to Scandinavian countries. Limitations to these studies include the frequently
high correlation between PFOA and PFOS exposure; not all studies control for PFOA in study
design. Studies also included populations with known elevated exposure to other environmental
chemicals including PFOA in the C8 population or polybrominated diphenyl ethers (PBDEs) and
other persistent organic compounds among the Inuit population. Overall, the epidemiologic
evidence supports an association between PFOS and increased total cholesterol.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-5

-------
                                Table 3-1. Association of Serum PFOS with Serum Lipids
Reference and Study
Details
PFOS Level
(Hg/mL)

Total Cholesterol (TC)
Low Density
Lipoprotein (LDL)
High Density
Lipoprotein (HDL)

Triglycerides (TG)
Occupational Populations
Olsenetal. 2001a,2003b
Cross-sectional from
manufacturing plant
workers
n = 263 (Decatur)
n = 255 (Antwerp)
Olsenetal. 200 lb,2003b
Longitudinal; ~ 5 years
n= 175
(Decatur and Antwerp
combined for analysis)










Mean 1 .40 Decatur

Mean 0.96 Antwerp



Mean
2.62
(baseline)
1 .67 (follow-up)
(Decatur)

1.87 (baseline)
1.16
(follow-up)
(Antwerp)
PFOS Quartiles
Ql: 0.04-0.42
Q2: 0.43-0.81
Q3: 0.82-1 .68
Q4: 1. 69-10.06 ppm
Beta =0.010 (95% CI)
(-0.005, 0.025)




TC by quartile of PFOS
mean (SD):
Ql:214(41)
Q2: 214 (43)
Q3:215(39)
Q4: 222 (44)









NM





NM














No association





HDL by quartile of PFOS
mean (SD):
Ql:54(15)
Q2:47(ll)
Q3:48(13)
Q4:48(15)









Beta = 0.025 (95% CI) (-0.015, 0.065)





TG by quartile of PFOS mean (SD):
Ql:131(95)
Q2: 155 (102)
Q3: 169(123)
Q4: 177 (123)
p<0.05Q4vQl









General Populations with high environmental exposure to other PFASs
Steenland et al. 2009
Cross-sectional (C8),
Logistic regression
analysis, 2005-2006 n =
46,294
Age: 1 8-80 yrs (not
taking cholesterol-
lowering medications)
Mean duration: not
provided
Linear regression,
quartiles and continuous
Mean 0.022

Quartiles of PFOS
(ng/mL):
Ql: 0-13.2
Q2: 13.3-19.5
Q3: 19.6-28.0
Q4:>28.1




Odds Ratio (95% CI) for
high cholesterol by 1 IQR
increase in PFOS
Ql: l(referrant)
Q2: 1.14(1.05,1.23)
Q3: 1.28(1.19,1.39)
Q4: 1.51(1.40,1.64)

Beta 0.02660
(SD 0.00140)
[log PFOS and lipids]

Nearly monotonic increase
in association with PFOS

Beta 0.04176
(SD 0.00221)
[log PFOS and lipids]






Null associations

Beta 0.00355
(SD 0.00173)
[log PFOS and lipids]







Increased

Beta 0.01 998
(SD 0.00402)
[log PFOS and lipids]







Perfluorooctane sulfonate (PFOS) - May 2016
3-6

-------
Reference and Study
Details
Fitz-Simonetal. 2013
Longitudinal (C8);
n=521
Duration: 4.4 years
Within-individual changes
in PFOS & lipids over
time, 2005-2006 versus
2010 serum
concentrations.
Linear regression fit to log
of ratio change in lipid in
relation to change in
PFOS
Nelson etal. 2010
Cross- sectional
(NHANES), USA. n =
860 (20-80 yrs old)
Linear regression analysis
for PFOS and serum lipids




Chateau-Degat et al. 2010
Cross-sectional, Inuit
population (Quebec).
PFOS effect on total
lipids. Effect modification
of n-3 PUFAs, which can
be hypolipidemic
n=723
Multiple linear regression
modeling
PFOS Level
(Hg/mL)
0.023 (baseline)
0.011 (follow-up)

TertilesofPFOSng/ml
(ratio follow
up/baseline)
Tl:<0.4
T2: 0.4-0.54
T3:>0.54




0.025

Serum PFOS by
quartile
Ql: 1.4-13.6
Q2: 13.8-19.7
Q3: 19.8-28.1
Q4: 28.2-392.0


0.019

Geometric mean (95%
CI) ug/L
Women:
16.8(15.8-17.8)
Men:
20.4(19.1-21.8)



Total Cholesterol (TC)
Geometric mean (mg/dL):
baseline, follow-up

192.5,192.8

Percent decrease (95% CI) in
lipid per halving PFOS
3.20(1.63^.76)





TC by PFOS Quartile
(mg/dl):

Ql: 198.6
Q2:201.6
Q3: 202
Q4: 205.7

Beta 0.27
(95% CI; 0.05-0.48)
Adjusted models
R2, Beta (p value)

0.17,0.0009(0.086)






Low Density
Lipoprotein (LDL)
Geometric mean (mg/dL):
baseline, follow-up

107.8, 109.2

Percent decrease (95% CI)
in lipid per halving PFOS
4.99 (2.46-7.44)





LDL by PFOS Quartile
(mg/dl):

Ql: 113.6
Q2: 116.4
Q3: 113.4
Q4: 123.1

Beta 0.12 (95% CI; -0.17-
0.41)
Adjusted models
R2, Beta (p value)

0.17, -0.0020 (0.242)






High Density
Lipoprotein (HDL)
Geometric mean (mg/dL):
baseline, follow-up

48.6,47.2

Percent decrease (95% CI)
in lipid per halving PFOS
1.28 (-0.59-3. 12)





HDL by PFOS Quartile
(mg/dl):

Ql: 54.3
Q2: 56.0
Q3: 52.7
Q4 : 55.2

Beta 0.02 (95% CI; -0.05-
0.09)
Adjusted models
R2, Beta (p value)

Women:
0.12,0.0042(0.001)
Men: 0.12, 0.0016 (< 0.001)





Triglycerides (TG)
Geometric mean (mg/dL): baseline,
follow-up

144.1, 146.9

Percent decrease (95% CI) in lipid per
halving PFOS
2.49 (-2.88-7.57)





NM









Adjusted models
R2, Beta (p value)

Women:
0.20, -0.0014(0.04)
Men: 0.16, -0.0009 (0.162)




Perfluorooctane sulfonate (PFOS) - May 2016
3-7

-------
Reference and Study
Details
Eriksenetal. 2013 Cross-
sectional, Middle aged
Danish population
n = 753 (663 men and 90
women)
Generalized linear models
used for analysis


Fisher etal. 2013
Cross-sectional, 2007-
2009, Canadian Health
Measures Survey (CHMS)
Cycle l.n = 2700 (aged
18-74)
Used multivariate linear
and logistic regression
models to assess
associations between
PFOS and serum lipids.


PFOS Level
(Hg/mL)
0.036








0.0084













Total Cholesterol (TC)
Differences in TC (mg/dl)
per 1 IQR increase
Beta (95% CI):

Total population:
3.7(0.1,7.3)
Women:
11. 7 (-0.2, 23.6)
Men: 2.9 (-0.9, 6.7)
Unadjusted OR for high
cholesterol compared to Ql
of PFOS exposure:

OR (95% CI)
Ql:Referrent
Q2: 1.12(0.89,1.41)
Q3: 1.15(0.91,1.45)
Q4: 1.66(1.32,2.09)
p trend = 0.03

Null effects in adjusted
model
Low Density
Lipoprotein (LDL)
NM








Null effects












High Density
Lipoprotein (HDL)
NM








Null effects













Triglycerides (TG)
NM








NM












Children and Adolescents
Frisbee etal. 2010
Cross-sectional (C8,
children)
GLM Analysis,
n = 12,476
Differences of Estimated
Marginal Mean (EMM)
between Ql and Q5 and
regression analysis for Q
trend
Geigeretal. 2014a
Cross- sectional,
NHANES, 1999-2008,
dyslipidemia (TC, LDL,
HDL, TG).
n =815 (Age < 18)
Multivariate regression
analysis.
n=815


0.023









Tl:<12.1
T2: 12.1-21.8
T3:>21.8ppb








Differences in Estimated
Marginal Mean (EMM),
Beta (SE), p for trend:

Age 1 to< 12: 5.5,
1. 3 (0.3), < 0.001

Age 12 to < 18: 9.5,
2.1 (0.4), < 0.001

TC (mg/dL) association with
PFOS by tertiles

Tl: 1
T2: 1.73 (-2. 89, 6. 36)
T3: 3.91 (-1.32, 9.14)
p trend: 0.15

log transformed PFOS
Beta 0.04 (95% CI: 0.00-
0.08)
Differences in EMM,
B(SE) p for trend:

Age 1 to< 12: 3.4,
0.9 (0.3), .002

Age 12 to < 18: 7.5,
1.7 (0.2), < 0.001


Association between
PFOS and LDL:

Tl: 1 (referent)
T2: 0.49 (-3.41, 4. 38)
T3: 4.59 (-0.17, 9.35)
P trend: 0.0632

log transformed PFOS
Beta 2.83 (95% CI: 0.03-
5.37)
Differences in EMM,
B(SE), p for trend:

Age 1 to< 12: 1.6,
0.3(0.1), 0.007

Age 12 to < 18: 1.5,
0.4(0.1), 0.001


Association between PFOS
and HDL:

Tl: 1
T2: 2.86 (0.44, 5.28)
T3: 1.11 (-0.93, 3.15)
P trend: 0.2931




Differences in EMM, B(SE), p for trend:

Age 1 to< 12: 2.8,
0.1 (1.4), 0.99

Age 12 to <1 8:2.8,
-0.1(1.0), 0.90



Association between PFOS and TG:

Tl: 1
T2:-8.13(-15.50, -0.77)
T3: -8. 89 (-15.67, -2.11)
P trend: 0.0126

log transformed PFOS
Beta -3.90 (95% CI: -7.72 to -0.08)


Perfluorooctane sulfonate (PFOS) - May 2016
3-S

-------
Reference and Study
Details
Lin et al. 2009
Cross- sectional,
NHANES, 1999-2000,
2003-2004. Adolescents
and adults aged > 12 yrs
n= 3,685
Maisonet et al. 2015
Avon Longitudinal Study
of Parents and Children.
Prenatal PFOS compared
to serum lipids in female
offspring.
n= Ill(age7),n=88
(age 15)
Timmermann et al. 2014
Danish children, aged 8-
10 years old. Linear
regression models. 1997.
n = 499
PFOS Level
(Hg/mL)
Mean (SEM) Log
PFOS
12 to < 20 yrs olds:
3.11(0.05)ng/mL
20 yrs old and older:
3.19(0.04)ng/mL
Mean (SD)
22.2(11.4)mg/dl
Median 41.5 ng/mL
Total Cholesterol (TC)
NM
Non-linear associations of
TC with PFOS.
Null findings in normal
weight children.
Low Density
Lipoprotein (LDL)
NM
Non-linear associations of
LDL with PFOS.
Null findings in normal
weight children.
High Density
Lipoprotein (HDL)
OR (95% CI), p
1.61(1. 15-2.26), p< 0.05
in those 20 yrs or older
Null findings
Null findings in normal
weight children.
Triglycerides (TG)
Null findings
Null findings
Null findings in normal weight children.
In overweight children, 10 ng increase
PFOS/mL plasma associated with 8.6%
(95% CI: 1.2%-16.5%)higher
triglyceride concentrations
Pregnant Women
Starling etal. 2014
Cross-sectional (maternal
at 14-26 weeks gestation),
Norwegian Mother and
Child Cohort (MoBa)
2003-2004.
n=891
0.013
Quartiles (ng/mL):
Ql:< 10.31
Q2: 10.31-13.03
Q3: 13.04-16.59
Q4: > 16.60
B (95% CI) PFOS (ng/ml)
and TC (mg/dL).
Ql: Referrent
Q2: -3.35 (-10.34, 3.64)
Q3: 3.06 (-4.93, 11.05)
Q4: 7.59 (-0.42, 15.60)
TC change per IQR change
in PFOS: 4.25 (0.81, 7.69)
B (95% CI) PFOS (ng/ml)
and LDL(mg/dL).
Ql: referrent
Q2: -3.23 (-9.28, 2. 83)
Q3: 2.60 (~4.49, 9.70)
Q4: 5.51 (-1.62, 12.64)
LDL change per IQR
PFOS change: 3.07
(-0.03,6.18)
B (95% CI) PFOS (ng/ml)
and HDL (mg/dL).
Ql: Referrent
Q2: 1.96 (-0.39, 4.31)
Q3: 2.49 (0.00, 4.97)
Q4: 4.45 (2.04, 6. 86)
HDL change per IQR
change in PFOS: 2.08 (1.12,
3.04)
B (95% CI) PFOS (ng/ml) and TG
(mg/dL).
Ql: Referrent
Q2: 0.00 (-0.06, 0.07)
Q3: -0.03 (-0.10, 0.05)
Q4: 0.00 (-0.07, 0.07)
TG change per IQR PFOS change:
-0.01 (-0.04, 0.02)
NM = Not Measured
    Perfluorooctane sulfonate (PFOS) - May 2016
3-9

-------
    Some of the studies that examined serum LDL and HDL cholesterol also found significant
increases these measures. Neither of these lipoprotein complexes is a stand-alone indicator for
cardiovascular decrease risk. Rather, it is the relationship across the lipoprotein complexes
within the same individuals that is important with HDLs considered as protective and LDLs a
biomarker for potential atherosclerosis. Relatively few studies of triglycerides noted a significant
increase with the  serum PFOS levels.

3.1.1.2  Liver Enzymes and Liver Disease

    Cross-sectional studies and longitudinal studies of PFOS and liver enzymes in various
populations are described below and summarized in Table 3-2.

         Table 3-2. Summary of Epidemiology Studies of PFOS and Liver Enzymes
     Reference and Study Details
 PFOS Level
               Results
 Lin etal. 2010
 n = 2,216 adults (1,076 men and 1,140
 women)
 Age: > 20 years old
 Data from 1999-2000 and 2003-2004
 NHANES
 Regression models used to analyze data
 and adjust for confounders	
Mean levels
Women: 0.0222
Men: 0.0274
Linear regression coefficients (standard
error), p-value (adjusted for age, gender,
race, lifestyle, measurement data, etc.)
ALT (U/L): 1.01 (0.53), 0.066 (slight pos.
association)
GOT (U/L): 0.01 (0.03), p = 0.81
Total bilirubin (umol): 0.30 (0.24), p = 0.22
 Gallo etal. 2012
 n = 47,092
 Data from those enrolled in C8 Health
 Project
 Linear and logistic regression models
 used.
Mean level: 0.0233
Linear regression coefficients, (partial R2)
Ln-ALT: 0.020, 95% CI: 0.014-0.026
(< 0.001)

Raised ALT in logistic regression odds ratio,
(p-value)
OR: 1.13, 95% CI: 1.07-1.18 (p< 0.001)

GOT: no association
Direct bilirubin: less consistent results
   Lin et al. (2010) investigated the association between low-dose serum PFOS (along with
three other individual PFAS) and liver enzymes in the adult population of the United States by
analyzing data from the 1999-2000 and 2003-2004 NHANES. The study population included
2,216 adults (1,076 males, 1,140 females) older than 20 years who were not pregnant or nursing;
had fasted > 6 hours at the time of examination; were negative for hepatitis B or C virus; had
body weight, height, educational attainment, and smoking status data available; and had serum
tests for PFAS, liver function, and metabolic syndrome. Regression models were used to analyze
the data and adjust for confounding factors. Mean PFOS levels were 0.0274 and 0.0222 |ig/mL
for males and females, respectively.

   Serum PFOS concentration was divided into quartiles. Unadjusted liver enzymes, serum
ALT, and log-GGT increased across quartiles of PFOS (p < 0.03), but total bilirubin showed no
trend. The linear regression models were adjusted for:

   •  Age, gender, and race/ethnicity.
   •  Lifestyle (smoking status, drinking status, education level).
   •  Biomarker data (BMI, metabolic syndrome,  iron saturation status, insulin resistance).
Perfluorooctane sulfonate (PFOS) - May 2016
                                                     3-10

-------
   In the fully adjusted model, a slight positive association was found between serum PFOS
concentration and serum ALT (p = 0.066). A positive association was also found between serum
PFOA concentration and serum ALT and PFOA concentration and serum GOT.  Data
interpretation was limited by the cross-sectional study design, and the fact that other
environmental chemicals (possible covariates or explanatory variables) and medication use were
not included in the study.

   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, GGT, and direct bilirubin with PFOS was assessed using linear regression
models adjusted for age, physical activity, body mass index, average household income,
education level, race, alcohol consumption, and cigarette smoking. Median PFOS level was
0.0233 |ig/mL with an interquartile range of 0.0137-0.0294 jig/mL. The In-transformed values
of ALT were significantly associated with In-transformed PFOS levels (and PFOA) and showed
a steady increase in fitted levels of ALT per decile of PFOS, leveling off after approximately
0.030 jig 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.

   The epidemiological data supporting liver damage based on serum ALT and GGT as reported
by Gallo et al. (2012) are not strong enough to support an association of serum PFOS alone with
liver damage in humans, because in most of the epidemiology studies the serum contains a
mixture of PFASs and possibly other exogenous chemicals.

3.1.1.3  Biomarkers of Kidney Function and Kidney Disease

   Epidemiology studies of PFOS and kidney function and biomarkers in various populations
are described below and summarized in Table 3-3.

   Shankar  et al. (2011) used data from the NHANES to determine whether there was a
relationship between serum PFOS levels and chronic kidney disease. A total  of 4,587 adult
participants (51.1% females) with PFOS measurements available from the 1999-2000 and 2003-
2008 cycles  of the survey were examined. Chronic kidney disease was defined as glomerular
filtration rate (GFR) < 60 mL/minute (min)/1.73 m2. Serum PFOS levels were categorized into
quartiles: quartile 1 = < 0.012 |ig/mL; quartile 2 = 0.012-0.019 |ig/mL; quartile  3 = 0.019-0.030
|ig/mL; quartile 4 = > 0.030 jig/mL. The multivariable odds ratio for chronic kidney  disease for
individuals in quartile 4 was 1.82 (95% CI: 1.01-3.27; p for trend = 0.019) compared with
individuals in quartile 1. This association was shown to be independent for confounders of age,
sex, race/ethnicity, body mass index, diabetes, hypertension, and serum cholesterol level.
However, the authors noted that because of the cross-sectional nature of the study, the possibility
of reverse causality could not be excluded. A low GFR would diminish the removal of PFOS
from serum for excretion by the kidney, thus increasing the serum PFOS levels.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-11

-------
 Table 3-3. Summary of Epidemiology Studies of PFOS and Measures of Kidney Function
  Reference and Study Details
   PFOS levels (jig/mL)
                 Results
 Shankaretal. 2011
 USA, NHANES
 n = 4587 adults
 PFOS from 1999-2000 and
 2003-2008
Quartiles, ug/mL, n

1:< 0.012 ug/mL, 1,152
2: 0.012-0.019 ug/mL, 1,151
3: 0.019-0.030 ug/mL, 1,137
4: > 0.030 ug/mL, 1,147
Estimated glomerular filtration rate (eGFR)
Chronic kidney disease defined as eGFR < 60
mL/minute/1.73 m2

Quartile, OR (95% CI)
1: Referent
2: 1.12(0.64, 1.99)
3: 1.53(0.87,2.67)
4: 1.82(1.01,3.27)
p = 0.02

Logistic regression adjusting for age, gender,
race/ethnicity, education, smoking, alcohol,
BMI, systolic blood pressure, diastolic blood
pressure, diabetes, serum total cholesterol and
glycohemoglobin
Adjustment for PFOS did not alter association
with PFOA

Multivariate regression of association PFOS
with eGFR among subjects with and without
chronic kidney disease
P (SE) with -1.8 (0.8) and without -3.2 (0.6)
chronic kidney disease (p < 0.05)	
 Steenlandetal. 2010
 USA, C8 Health Project
 participants
 n = 54,591(> 20 yrs old)
Mean: 0.0234 ±0.0161
Increased predicted uric acid of 0.2 to 0.3
ug/dL with increasing deciles of PFOS.

Odds Ratio, p-value
Hyperuricemia (> 6.0 mg/dL for women and >
6.8 mg/dL for men):
1.00
1.02 (95% CI: 0.95-1.10), p < 0.0001
1.11(95%CI: 1.04-1.20), p< 0.0001
1.19 (95% CI: 1.11-1.27), p< 0.0001
1.26 (95% CI: 1.17-1.35), p < 0.0001

Trend for increase uric acid more prominent
with PFOA
 Children
 Watkinsetal. 2013
 USA, C8 Health Project
 participants
 n = 9,660 (1 to < 18 yrs old)
Median: 0.020
P (95% CI) change in unit eGFR (mL/min/1.73
m2) per In serum PFOS,
-1.10 (-1.66 to -0.53), p = 0.0001
Linear regression adjusting for age, gender,
race, smoking, and household income.	
 Geigeretal. 2014b
 USA, NHANES
 n= 1644 (12-18 yrs old)
Mean: 0.018 ±0.005
Multivariable-adjusted OR (95% CI) between
PFOS and hypertension

Quartile 1: 1 (referent)
Quartile 2: 0.99 (0.55, 1.78)
Quartile 3: 0.73 (0.36, 1.61)
Quartile 4: 0.77 (0.37, 1.61)
p = 0.36
Log transformed PFOS = 0.83 (0.58, 1.19)
Perfluorooctane sulfonate (PFOS) - May 2016
                                                                 3-12

-------
    Steenland et al. (2010) reported on another analysis of the C8 Health Project participants
> 20 years old (n = 54,591) 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. The mean PFOS level was 0.0234 ± 0.0161 jig/mL. 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 (0.010-0.050 |ig/mL). Hyperuricemia (> 6.0 mg/dL for females and
> 6.8 mg/dL for males) risk by quintiles increased slightly with PFOS levels (OR 1.00, 1.02,
1.11, 1.19, and 1.26). The serum of C8 study participants included several PFASs; PFOA
appeared to have a greater influence on uric acid trends than PFOS in the models employed by
Steenland et al. (2010).

Children. Watkins et al. (2013) evaluated the cross-sectional association between PFOS
exposure and kidney function among children aged 1 to <18 years (mean 12.4 ±3.8 years)
enrolled in the C8 Health Project. A total of 9,660 participants had data available on serum PFOS
(median = 0.020 jig/mL), serum creatinine, and height, which were used to calculated an
estimated glomerular filtration rate (eGFR). Linear regression was used to evaluate the
association between quartiles of measured serum PFOS concentration and eGFR. A shift from
the lowest to the highest quartile of measured, natural log-transformed concentrations of PFOS
in serum [IQR In-(PFOS) = 0.64] was associated with a decrease in eGFR of 1.10 mL/min/1.73
m2 (95% CI: -1.66 to -0.53; p = 0.0001) adjusting for age, sex, race, smoking  status, and house-
hold income. With increasing quartile of serum PFOS concentrations, eGFR decreased mono-
tonically with a decrease of 2.3, 2.6, and 2.9 mL/min/1.73 m2 for the second, third, and fourth
quartile of serum PFOS,  respectively, compared  with the lowest quartile (p for trend across
quartiles = 0.0001).

   Geiger et al. (2014b) used data from the NHANES to determine whether there was a
relationship between serum PFOS levels and hypertension in children. A total of 1,655
participants (aged 12-18 years) with PFOS measurements available from the 1999-2000 and
2003-2008 cycles of the survey were examined.  Blood pressure was measured to determine the
presence of hypertension and linear regression modeling was used to study the association
between increasing quartiles of serum PFOS and mean changes in systolic and diastolic blood
pressures. Mean PFOS level was 0.018 ± 0.005 jig/mL. No association was found between
serum PFOS levels and hypertension in either unadjusted or multivariable-adjusted analyses.
Compared with the lowest quartile, the multivariable-adjusted odds ratio (95% confidence
interval) of hypertension in the highest quartile of exposure was 0.77 (0.37-1.61) (p-trend >
0.30).

3.1.1.4  Reproductive Hormones and Reproductive/Developmental Studies

   Many of the studies of PFOS focused on pregnancy-related outcomes, including measures of
fetal growth retardation,  puberty, and other developmental endpoints, as well as pregnancy-
related hypertension, preeclampsia, and gestational diabetes. Reproductive outcomes such as
measures affecting fertility were also evaluated. Within each section, the discussion is divided
into occupational exposure studies (if applicable) and general population studies. Epidemiology
studies of PFOS and pregnancy-related outcomes in various populations are described below and
summarized in Table 3-4.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-13

-------
     Table 3-4. Summary of Epidemiology Studies of PFOS and Pregnancy Outcomes
             Study
                                 PFOS level
                                  Results
 Stein et al. 2009
 United States (C8 Health Project)
 n = 5,262 pregnancies
 Self-reported pregnancy outcomes
 in mid-Ohio Valley in 2000-2006.
Median: 0.014
OR (95% CI), preeclampsia
per IQR(lnPFOS) increase in PFOS: 1.1 (0.9, 1.3)
                < 50th percentile
                > 50th percentile
                < 50th percentile
                50th-<75th percentile
                 75th_90th percentiie
                > 90th percentile
                       1.0 (referent)
                       1.3 (1.1, 1.7)
                       1.0 (referent)
                       1.3 (1.0, 1.7)
                       1.1 (0.8, 1.6)
                       1.6(1.2,2.3)
 Darrowetal. 2013
 United States (C8 Health Project)
 n = 1,630 live births from 1,330
 women after January 1, 2005
Geometric mean:
0.0132
Pregnancy induced hypertension
OR (95% CI) per log unit increase in PFOS: 1.47 (1.06,
2.04)
By quintile:
 Ql up to 0.0086 ug/mL
 Q2 0.0086-< 0.0121
 Q3 0.0121-< 0.0159
 Q4 0.0159-< 0.0214
 Q5> 0.0214
 Ql up to 0.0086 ug/mL
First pregnancy after
 PFOS measure
                                                                     1.0 (referent)
                                                                     1.46(0.69,3.11)
                                                                     2.71(1.33,5.52)
                                                                     2.21 (1.07, 4.54)
                                                                     1.56 (0.72, 3.38)
                                                                     1.0 (referent)
                                                                     2.02(1.11,3.66)
 Zhang etal. 2015
 n = 258 women as part of LIFE
 study. Blood samples taken during
 2005-2009.
Mean: 0.0131
with gestational
diabetes and
0.012 without
Gestational diabetes
OR (95% CI) associated with SD increment of
preconception PFOS log-transformed concentration
OR 1.13 (0.75, 1.72) (fully adjusted for age, BMI,
smoking, etc.)	
Pregnancy-related Outcomes. Stein et al. (2009) examined serum levels of PFOS and self-
reported pregnancy outcomes of a population of females (5,262 pregnancies; aged 15-55 years)
in the mid-Ohio Valley in 2000-2006. These females 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 females was 0.014 jig/mL. There was
no association between PFOS levels and miscarriages. PFOS was associated with preeclampsia
(adjusted odds ratio = 1.3; 95% CI:  1.1-1.7). Similarly, PFOA was not associated with
miscarriage and only weakly associated with preeclampsia. The self-reported nature of
pregnancy outcomes is a recognized limitation with uncertain impact on study results.

   Darrow et al. (2013; 2014) analyzed pregnancy outcomes for the five years following
enrollment in the C8 Health Project. Among the 69,030 females who provided serum for PFOS
measurement in 2005-2006, 32,354 provided follow-up interviews on reproductive histories.
After exclusions, 1,630 singleton live births from 1,330 females after January 1, 2005 were
linked to birth records to identify outcomes of preterm birth (i.e., < 37 gestational weeks),
pregnancy-induced hypertension, low birth weight (LEW) (i.e., < 2500 grams), and birth weight
among full-term infants (Darrow et  al. 2013). Effects on fetal growth measures are described in
that section below. Another subset of 1,129 females with a total of 1,438 pregnancies was
evaluated  for an association between PFOS levels and miscarriage (Darrow et al. 2014). The
baseline mean PFOS level for these females was 0.016-0.017 jig/mL. Confounders that were
adjusted in each model for every outcome in the 2013 Darrow et al. study included maternal age,
Perfluorooctane sulfonate (PFOS) - May 2016
                                                        3-14

-------
educational level, smoking status, parity, BMI, self-reported diabetes, time between conception
and serum measurement. Parity was excluded, and race was included in the miscarriage analysis
(Darrowetal. 2014).

   An increased risk of pregnancy-induced hypertension was detected per log unit increase in
PFOS (OR = 1.47; 95% CI: 1.06-2.04) and PFOA (OR = 1.27; 95% CI: 1.05-1.55). Although
monotonicity was not evident, consistently increased odds were found across all upper PFOS
(OR range: 1.46-2.72) and PFOA (OR range: 2.39-3.43) quintiles.

   The odds of miscarriage per each log unit increase in PFOS was 1.21 (95% CI: 0.94-1.55)
for all reported prospective pregnancies and 1.34 (95% CI:  1.02-1.76) when restricted the
analysis to each woman's first pregnancy. Miscarriage results were comparable across all PFOS
quintiles in the primary analysis (OR range: 1.34-1.59) and those restricted to first pregnancy
(OR range: 1.68-1.94). PFOA was not associated with miscarriage and was not a confounder of
the observed association with PFOS. To address the potential for reverse causality related to
PFAS levels decreasing from prior pregnancies, analyses were restricted to nulliparous and
nulligravid females.  Adjusted odds ratios were higher across all four quintiles for nulliparous
(OR range: 1.88-3.08) and nulligravid females (OR range: 2.04-3.73). These studies represent
prospective assessment of PFASs in relation to adverse pregnancy outcomes, which address
some of the limitations in the available cross-sectional studies. The impact of measurement error
resulting from unknown critical exposure windows and the time lag (> 99% of births were within
3 years) between the estimated conception date and the serum collection is unclear in these
studies.

   Preconception serum levels of PFOS (and other PFASs) were evaluated in females
attempting pregnancy in relation to risk of developing gestational diabetes (Zhang et al. 2015).
The 258 participants were members of the Longitudinal  Investigation of Fertility and the
Environment (LIFE) study with blood samples taken during 2005-2009. The ORs and 95%  CIs
of gestational diabetes associated with each SD increment of preconception serum PFOS
concentration (log-transformed) (and six other PFAS) were estimated with the use of logistic
regression after adjusting  for age, pre-pregnancy body mass index, smoking, and parity, each
conditional on the number of times a woman had been pregnant. Preconception mean serum
PFOS levels were 0.0131  |ig/mL in females with gestational diabetes and 0.012 |ig/mL in
females without gestational diabetes (p-value for mean difference = 0.10). A positive association
was found between PFOS and risk of gestational diabetes in the fully adjusted model (OR =
1.13; 95% CI: 0.75-1.72). PFOA was the only PFAS that was significantly associated with
developing gestational diabetes in this analysis.

Fetal Growth. Many different measures of fetal growth can be used in epidemiology studies.
Birth weight is widely available (as it is routinely collected in medical records and birth
certificates). LEW (defined as < 2500 g) can be a proxy  measure for preterm birth (particularly
when accurate gestational age dating is not available). Other measures of fetal growth, such  as
small for gestation age, tend to more accurately reflect fetal growth retardation. Epidemiology
studies of PFOS and fetal growth are described below and summarized in Table 3-5.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-15

-------
               Table 3-5. Summary of Epidemiology Studies of PFOS and Fetal Growth
            Study
       PFOS level (jig/mL)
                      Results
Grice et al. 2007
United States (C8 Health Project)
n = 263 women reporting 429
births
Self-reported pregnancy
outcomes in workers associated
with perfluorinated chemical
production factory.
Exposure to PFOS was based on job
assignment and varied

Never exposed: 0.11-0.29 ppm
Low exposure: 0.39-0.89 ppm
High exposure: 1.30-1.97 ppm
No association between PFOS exposure and mean birth
weight
Regression coefficients for birth weight compared to never-
exposed pregnancies, 95% CI (adjusted for maternal age,
smoking, gravidity)
                                    Ever exposed, low exposure
                                    Ever exposed, high exposure
                                    High exposure, > 1 yr
                                    Low or high exposure, > 1 yr
                                    Ever exposed, low or high
                              -0.08 (-0.25, 0.09)
                               0.07 (-0.14, 0.28)
                               0.11 (-0.11, 0.33)
                              -0.03 (-0.19, 0.13)
                              -0.05 (-0.20, 0.11)
Apelberg et al. 2007
United States (Baltimore)
n = 293 newboms born between
November 2004 and March 2005
Cord blood samples
Geometric mean: 0.005
Change in birth weight (g) per log unit increase (95% CI)
-69 (-149, 10)
Fei et al. 2007
n = 1,400 women and their
infants randomly selected from
the group enrolled in the DNBC
Mean: 0.035
LEW

OR (95% CI) for LEW by quartile
                                   Ql 0.0064 to 0.026 ug/dL
                                   Q2 0.026 to 0.033  ug/dL
                                   Q3 0.033 to 0.043  ug/dL
                                   Q4 > 0.043 ug/dL
                                   Trend: p = 0.13	
                         1.0 (referent)
                         3.5(0.37,31.16)
                         6.0(0.73,49.34)
                         4.8(0.56,41.16)
Andersen et al. 2010
n = 1,010 women and their
infants randomly selected from
the group enrolled in the DNBC
Median: 0.0334 (range: 0.0064-
0.1067)
PFOS concentrations per each 0.001 ug/mL increase
inversely associated with:
birth weight in girls: Beta = -3.2; 95% CI: -6.0 to -0.3
weight at 12 months in boys: Beta = -9.0; 95% CI:
-15.9 to-2.2
Monroy et al. 2008
n = 101 pregnant women as part
of a larger cohort study
conducted at McMaster
University Medical Center
Mean:
0.0183 in maternal serum (24-28 wks)
0.0162 in maternal serum at delivery
0.0072 in umbilical cord blood
No association between PFOS levels and infant birth weight

Change in PFOS per g change in birth weight
Beta = 0.000853 (p = 0.73)
Washino et al. 2009
Japan
n = 428 women and their infants
between July 2002 and October
2005
Mean: 0.006
Change in birth weight per log unit increase (95% CI)
For all: Beta = -149 g (-297.0, -0.5)
For female infants: Beta = -269.4 g (-465.7, -73.0)
Hammetal. 2010
Canada
n = 252 women with blood
samples taken between December
2005 and June 2006
Mean: 0.009
Change in birth weight per Ln unit increase (95% CI)
31.3 g (-43.3, 105.9)
Stein et al. 2009
Mean: 0.014
OR (95% CI), birth weight < 5.5 Ibs.
per IQR(lnPFOS) increase in PFOS: 1.3 (1.1, 1.6)
                                                                  < 50th percentile
                                                                  > 50th percentile
                                                                  < 50th percentile
                                                                  50th-<75th percentile
                                                                  75th_90th percentiie
                                                                  > 90th percentile
                                                            1.0 (referent)
                                                            1.5(1.1,1.9)
                                                            1.0 (referent)
                                                            1.3(0.9,1.8)
                                                            1.6(1.1,2.3)
                                                            1.8(1.2,2.8)
   Perfluorooctane sulfonate (PFOS) - May 2016
                                                                              3-16

-------
         Study
     PFOS level (jig/mL)
                  Results
Darrowetal. 2013
United States (C8 Health Project)
n = 1,630 live births from 1,330
women after January 1, 2005
Geometric mean: 0.0132
LEW
OR (95% CI) per LEW (< 2,500 g) per log unit increase:
1.12(0.75,1.67)
By quintile:
                   1.0 (referent)
                   1.48(0.71,3.08)
                   1.23(0.57,2.65)
                   1.31(0.59,2.94)
                   1.33(0.60,2.96)
                                                     Ql up to 0.0086
                                                     Q2 0.0086-< 0.0121
                                                     Q30.0121-< 0.0159
                                                     Q4 0.0159-< 0.0214
                                                     Q5> 0.0214
                                                     First pregnancy after
                                                     PFOS measure
                                               0.97(0.61,1.54)
      An occupational cohort study by Grice et al. (2007) examined the relationship between PFOS
   exposure and self-reported 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 females participated
   (participation rate = 73%) and reported 439 births, of which there were 421 live births,
   14 stillbirths, and 4 with missing outcome data. The birth weight models of single births were
   adjusted for maternal age, smoking status, and gravidity. No associations were detected between
   PFOS exposure and the pregnancy outcomes that were examined (i.e., stillbirth and mean birth
   weight).

      Apelberg et al. (2007) measured PFOS in the cord blood of 293 newborns (singleton births
   without congenital anomalies) born November 26, 2004 through March 16, 2005 at Johns
   Hopkins Hospital in Baltimore, Maryland. Maternal and infant data, including maternal birth
   cohort, social class, place of residence, past pregnancies, insurance type, BMI, age, race,
   education, marital status, parity, gestational age, smoking status, and infant sex were collected
   from the hospital database and forms filled out at time of delivery. PFOS was found in > 99% of
   the cord blood  samples (geometric mean 0.005, range < level of detection [0.2]-0.035  jig/mL).
   PFOS concentrations were evenly distributed across larger maternal age categories. The non-
   smoker and passively exposed individuals (5.2 ng/mL) had higher mean PFOS levels than
   smokers (4.1 ng/mL), as did Asians (6.5 ng/mL) and Blacks (5.2 ng/mL) compared to
   Caucasians (4.5 ng/mL). No associations were observed between PFOS and maternal age,
   gestational age, BMI, or various socioeconomic measures (e.g., education, insurance, marital
   status, living in Baltimore City). Birth weight, head circumference, and ponderal index were
   inversely associated with both cord PFOS and PFOA levels. For example, large deficits in mean
   birth weight per one In-unit increase were found for both PFOS (P = -69; 95% CI: -149-10) and
   PFOA (P = -104 g; 95% CI: -213-5).

      A series of longitudinal, population-based studies was conducted in a subset of 91,827
   females aged 25-35 enrolled in the Danish National Birth Cohort (DNBC) from March 1996 to
   November 2002 (Andersen et al. 2013; Fei et al. 2007, 2008a, 2008b, 2009, 2010a). This
   prospective birth cohort was comprised of a random sample of 1,400 females who were recruited
   through general practitioners around weeks 6-12 of gestation to investigate the association
   between blood levels of perfluorinated chemicals and adverse reproductive and developmental
   outcomes in the females and their children.  This subset was sampled from 43,035 females with
   singleton live births without congenital malformation who provided the first blood sample
   between gestational weeks 4 and 14 and who responded to all four telephone interviews. Study
   data were collected by telephone interviews at 12 and 30 weeks of gestation, approximately
   Perfluorooctane sulfonate (PFOS) - May 2016
                                                               3-17

-------
6 and 18 months after birth, and when the children were 7 years of age. 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 infant cord blood was sampled
just after birth. Only blood results from the 1,400 females in the first trimester were reported.
Mean plasma PFOS levels by age groups were: < 25 years: 0.039 |ig/mL; 25-29 years:
0.037 ng/mL; 30-34 years: 0.034 |ig/mL and > 35 years: 0.033 |ig/mL.

   Potential confounders for which adjustments were made  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 per day, and home density (the total  number of rooms divided by the
total number of people in the household). Although dietary data were available for at least  80%
of the births, it is unclear why some of these studies did not examine these data as confounders
(e.g., Fei et al. 2009). Although the DNBC had a low participation rate (31%), a previous study
of various exposures in relation to three different outcomes (preterm birth, small-for-gestational-
age,  infancy and antepartum stillbirth) did not provide any evidence of non-participation bias
(Nohr et al. 2006).

   Using  data from the DNBC, Fei et al. (2007) investigated the association between plasma
levels of PFOS in pregnant females, length of gestation, preterm birth (i.e., < 37 gestational
weeks), and infant birth weight. The average PFOS levels in maternal plasma were 0.035|ig/mL
(range: 0.0064-0.107 jig/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 PFOS quartiles. No associations
between PFOS and birth weight were found. PFOA concentrations based on the continuous
exposure measures were inversely associated with birth weight (P= -10.6; 95% CI: -20.8 to
-0.5) following adjustment for confounding (unadjusted P= -20.5; 95%  CI: -31.5 to -9.6).
Although  most were not statistically-significant, ORs for preterm birth were consistent in
magnitude (OR range: 1.43-2.94) across both the upper three PFOS and  PFOA quartiles.
Consistently elevated ORs were also detected (OR range:  3.39-6.00) for LEW across the upper
three PFOS and PFOA quartiles, but all of these analyses were limited by very small cell sizes
given low incidence of these outcomes. Although these ORs often lacked statistical significance
due to low statistical power, the elevated odds detected between PFOS levels and various
outcomes  including preterm delivery and LEW warrant further research, especially given the
potential generalizability limitations of this low-risk study population.

   Fei et  al. (2008a) also investigated the association between PFOS levels and placental
weight, birth length, and head and abdominal circumference in the DNBC study population.
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 post-term and pre-term infants and with ponderal
index (relationship between mass and height) in multiparous females. In  nulliparous females the
association was positive. These associations were not statistically-significant.

   Andersen et al. (2010) examined the association between maternal plasma PFOS
concentration and offspring weight, length, and  BMI at 5 and 12 months of age from participants
in the DNBC. The mothers (n = 1,010) reported the information during an interview and weight
and length measurements were used to calculate BMI. Median maternal plasma PFOS level was
0.0334 |ig/mL with a range of 0.0064-0.1067 |ig/mL. PFOS concentrations (per each 0.001

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-18

-------
|ig/mL increase) were inversely associated with birth weight in girls (P= -3.2; 95% CI: -6.0 to
-0.3), weight at 12 months in boys (P= -9; 95% CI: -15.9 to -2.2), and BMI at 12 months in
boys (P= -0.017; 95% CI: -0.028 to -0.005) in models adjusted for maternal age, parity,
prepregnancy BMI, smoking, gestational age at blood draw, socioeconomic status, and
breastfeeding. Similar inverse associations were found with PFOA only in boys.

   Monroy et al. (2008) examined the relationship between the maternal serum levels of PFOS
and PFOA and infant birth weight from neonates born to 101 pregnant females enrolled in a
large cohort study, the Family Study, conducted at McMaster University Medical Center in
Ontario, Canada. Linear regression analyses were adjusted for parity, gestational length, BMI,
gender, and smoking status as confounding factors. PFOS was measured in maternal serum from
24-28 weeks  of gestation and at delivery and in umbilical cord blood (UCB) from 105 babies.
PFOS was detected in all of the collected samples with mean levels of 0.0183, 0.0162, and
0.0072 |ig/mL in maternal serum at 24-28 weeks, maternal serum at delivery, and in UCB,
respectively. The concentration of PFOS in maternal serum was significantly higher than in UCB
(mean ratio of UCB/maternal serum at delivery was 0.45). No statistically-significant
associations were detected between levels of PFOS in the maternal serum or UCB and infant
birth weight. Maternal PFOS levels were also not associated with maternal body mass index,
gestational length, or gender. Results were similar for PFOA.

   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
females and their infants (Washino et al. 2009). Females enrolled were at 23-35 weeks of
gestation with a mean age of 30.5 years. Exclusion criteria included maternal pregnancy-induced
hypertension, diabetes mellitus, fetal  heart failure, and multiple births (i.e., restricted to
singletons). A self-administered questionnaire survey after the second trimester of pregnancy
was used by the subjects to report dietary habits, smoking status, alcohol consumption, caffeine
intake, household income, and educational level. Other potential confounding factors collected
from medical records included prepregnancy BMI, pregnancy complications, gestational age,
infant sex, parity, infant disease, birth weight, and birth size. A blood sample was collected for
measurement of PFOS and PFOA during the second trimester when the questionnaire was
administered or after pregnancy for anemic mothers. The mean concentration of PFOS in the
females was 0.006 |ig/mL with detection in 100% of samples. The highest PFOS concentration
identified was 0.016 |ig/mL. The results indicated that large reductions in mean birth weight
(P = -149 g; 95% CI: -297.0 to -0.5) were detected for each log-10 change in maternal serum
PFOS exposure, especially among female infants (P = -269.4 g; 95% CI: -465.7 to -73.0).
Large birth weight deficits were also  detected per each unit increase in PFOA for both males
(-68.1  g; 95% CI: -246.2-110.0) and females (-76.7 g; 95% CI:  -234.7-81.3), with an overall
change in mean birth weight of 75 grams (95% CI: -191.8-41.6).

   A cohort study on pregnant females (>  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. 2010). The study
population included 252 pregnant females who elected to undergo a second trimester prenatal
triplescreen at 15-16 weeks of gestation for Down's syndrome, trisomy 18, and open spina
bifida.  This population was restricted to mothers > 18 years of age who gave birth to live
singletons without evidence of malformations, and who delivered at greater than or equal to
22 weeks of gestation. Serum samples collected from December 2005 to June 2006 during the
second trimester had PFOS levels ranging from nondetectable to 0.035 |ig/mL, with the mean
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-19

-------
and geometric mean being 0.009 |ig/mL and 0.0074 |ig/mL, respectively. Potential confounders
included maternal age, maternal weight, maternal height, maternal smoking status, maternal race,
gravida, gestational age at the time of serum collection, infant sex, infant birth weight, and infant
gestational age at birth. Overall, there was no association with the level of PFOS and birth
weight or length of gestation. Mean birth weight increased slightly by increasing PFOS tertiles
(3,278 g for < 0.006 jig/mL; 3,380 g for 0.006-0.010 |ig/mL; 3,387 g for > 0.010-0.035 jig/mL).
The mean length of gestation for all groups was 38 weeks; the preterm delivery percentage was
similar between groups. Similar associations were found for other PFASs, which were correlated
with serum PFOS  including PFOA (Spearman correlation coefficient = 0.52) and
perfluorohexane sulfonate (Spearman correlation coefficient = 0.54).

   In addition to the pregnancy-related outcomes discussed previously, Stein et al. (2009)
examined fetal growth outcomes among females enrolled in the C8 Health Project. Pregnancies
within the 5 years  preceding the exposure measurements were analyzed. The mean level of
PFOS in the serum of these females was 0.014 |ig/mL at the time of measurement. There was no
association between PFOS levels and preterm births. PFOS was, however, associated with an
increased risk above the median (adjusted odds ratio = 1.5: 95% CI: 1.1-1.9) for LEW, and a
dose-response relationship was reported for the 50th-?5th, 75th-90th and > 90th percentile serum
PFOS exposure concentrations (adjusted ORs = 1.3, 1.6, and 1.8, respectively). Similarly, PFOA
was not associated with LEW and preterm birth. The self-reported nature of pregnancy outcomes
is a recognized limitation with uncertain impact on study results. Although this 5-year window
was intended  to ensure that measured PFAS values at the time of study enrollment reflected
exposure level at the time of pregnancy, this could have  resulted in exposure misclassification
given changes in maternal PFAS levels that could have occurred between the time of serum
collection and pregnancy and lactation because measures had been implemented  to decrease
population exposures.

   Darrow et al. (2013, 2014) analyzed pregnancy outcomes for the 5 years after enrollment in
the C8 Health Project. Among the 69,030 females who provided serum for PFOS measurement
in 2005-2006, 32,354 provided follow-up interviews on reproductive histories. After exclusions,
1,630 singleton live births from 1,330 females after January 1, 2005 were linked  to birth records
to identify outcomes of preterm birth (i.e., < 37 gestational weeks), LEW, and birth weight
among full-term infants (Darrow et al. 2013). Another subset of 1,129 females with a total of
1,438 pregnancies was evaluated for an association between PFOS levels and miscarriage
(Darrow et al. 2014). The baseline mean PFOS level for these females was 0.016-0.017 jig/mL.
Confounders that were adjusted in each model for every outcome in the 2013 Darrow et al. study
included maternal age, educational level, smoking status, parity, EMI, self-reported diabetes, and
time between conception and serum measurement. Parity was excluded and race  was included in
the miscarriage analysis (Darrow et al. 2014). Maternal serum PFOS levels were not associated
with preterm birth or LEW. An inverse association was found between PFOS and mean birth
weight in full-term infants (-29 g per log unit increase; 95% CI: -66-7). PFOA was not
associated with mean birth weight, and therefore was not a confounder of this association. These
studies represent prospective assessments of PFASs in relation to adverse pregnancy outcomes
thereby avoiding some of the limitations of the cross-sectional studies. The impact of
measurement error resulting from unknown critical exposure windows and the time lag (> 99%
of births were within 3 years) between the estimated conception date and the serum collection is
unclear.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-20

-------
   Preeclampsia is a condition where the pregnant female is hypertensive because of reduced
renal excretion associated with a decrease in GFR. Preecampsia is often accompanied by LEW
(Whitney et al. 1987). Morken et al. (2014) used a subset of the Norwegian Mother and Child
Cohort to evaluate the relationship between GFR and fetal size. Participants included 470
preeclamptic patients and 483 non-preeclamptic females; plasma creatinine measured during the
second trimester was used to estimate GFR. For the overall cohort, for each mL/min increase in
GFR, infant weight at birth increased 0.73-0.83 g depending on the method used to calculate
GFR. The increases in body weight with increased GFR were greater, and statistically-
significant, in females with preeclampsia. Differences were not statistically-significant for the
nonpreeclamptic group. Morken et al. (2014) was not a study of perfluorochemicals, and there
were no serum measurements of any PFAS. However because PFOA/PFOS serum levels are
expected to be higher with a lower GFR, the finding stimulated examination of the GFR as it
relates to serum PFAS levels and the LEW identified in the epidemiology studies (Vesterinen et
al. 2014; Verneretal. 2015).

   Evidence for an inverse association between PFAS levels and birth weight raised the
question of reverse causality linked to maternal GFR. PFOS excretion by the kidney is
dependent, in part, by the GFR. Conditions that result in impairment of GFR (and, thus,
increased serum PFOS) and are also related to fetal growth restriction could result in a
confounded observation of an association between PFOS and decreased birth weight. Vesterinen
et al. (2014), using the Navigation Guide systematic review methods, examined evidence
pertaining to the relation between fetal growth and maternal GFR. They identified relevant
studies that met the Navigation Guide criteria for inclusion in the analysis; none included
consideration of PFOS or other PFASs. All studies were rated as low or very low quality leading
to the conclusion that data were inadequate to determine an association between fetal growth and
GFR.

   Verner et al. (2015) modified the PK model of PFOS during pregnancy by Loccisano et al.
(2013)  described in section 2.5.1 to evaluate the association between GFR, serum PFOS levels
and birth weight. When low GFR was accounted for in the model simulations, the reduction in
birth weight associated with increasing serum PFOS was less than that found by the author's
meta-analysis of the same data. This finding suggests that a portion of the association between
prenatal PFOS and birth weight could be confounded by maternal GFR differences within the
populations studied. The true association for each 1 ng/mL increase in PFOS could be closer to a
2.72 g reduction (95% CI: -3.40 to -2.04) in body weight compared to the 5.00 g reduction
(95% CI: -21.66 to -7.78) predicted by meta-analysis of the epidemiology data without a
correction for low GFR.

Other Developmental Effects. Fei et al. (2010a) reported on the effects of PFOS and PFOA on
the length of breastfeeding. Self-reported data on the duration of breastfeeding were collected
during the telephone interviews at 6 and  18 months after birth of the child. Statistically-
significant higher levels of PFOS  were associated with a shorter duration of breastfeeding
following adjustment for confounding. This is an expected consequence because PFOS is
transferred from the mother during breast feeding; thus, the shorter the lactation period the
greater the proportion of the serum PFOS at the time of birth remains with the mother. A
20% increase risk for the mother in weaning before 6 months was noted in both primiparous
[OR= 1.20; 95% CI: 1.04-1.37] and multiparous females, [OR= 1.20; 95% CI: 1.06-1.37]) for
each 0.010 |ig/mL increase in PFOS concentration in the maternal blood.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-21

-------
   A dose-response relationship was noted only among multiparous females (OR range:
1.55-2.64) based on categorical PFOS exposures, as only the highest PFOS quartile showed an
elevated effect estimate [OR = 1.52; 95% CI: 0.89-2.60]) among primiparous females. For
analyses based on termination of exclusive breastfeeding before 4 months, associations were
only seen among multiparous females for both PFOS and PFOA exposures. Given that the
associations between length of breastfeeding and PFOA and PFOS exposures were largely only
seen among multiparous females, reverse causality is a possible explanation since reductions of
current PFOS and PFOA levels may have resulted from longer lactation periods for previous
children.

   Andersen et al. (2013) evaluated the association between maternal plasma PFOS levels and
the children's body mass index, waist circumference, and risk of being overweight at 7 years of
age. From the subset of 1,400 randomly selected females from the DNBC who provided blood
samples during their first trimester, only those children with weight and height information
(n = 811) or waist measurements (n = 804) at age 7 years were included in the analysis. Maternal
plasma PFOS levels were evaluated as both continuous and categorical exposures. Maternal
PFOS concentrations were inversely associated with all of the children's anthropometric
endpoints, but statistical significance was not attained and a dose-response relationship was not
observed. Neither maternal PFOS nor PFOA levels were associated with anthropometric
measures in either boys or girls at age 7 in this prospective birth cohort.

   A case-cohort study from the DNBC population was used  to evaluate the relationship
between prenatal PFAS exposure and the risk of congenital cerebral palsy (Liew et al. 2014).
From a source population of 83,389 mother-child pairs, 156 cases of cerebral palsy were
identified and matched to 550 randomly selected controls (including 440 boys).  Stored maternal
plasma samples collected in early or mid-pregnancy were analyzed for 16 PFAS; six compounds
were quantifiable in > 90% of the samples. For the cerebral palsy cases and matched controls,
median maternal PFOS levels were 0.0289 and 0.0276 |ig/mL, respectively, for boys and 0.0275
and 0.0262 jig/mL, respectively, for girls. A statistically-significant increased risk of developing
cerebral palsy in boys (rate ratio [RR] = 1.7; 95% CI:  1.0-2.8) was detected per  each natural-log
unit increase in maternal PFOS level. A dose-response relationship between cerebral palsy and
categorical PFOS exposures was detected in boys. Positive associations were also found with
PFOA and perfluoroheptanesulfonate (PFHpS), and the results for PFOS remained unchanged
after adjusting for multiple PFAS in the regression models. No association was found between
any PFAS level and risk of cerebral palsy in girls, although this analysis was much more limited
by smaller numbers.

   Fei and Olsen (2011) examined the association between prenatal PFOS (and PFOA) exposure
and behavior or coordination problems in children aged 7 enrolled in the DNBC study.
Behavioral problems were assessed using the Strengths and Difficulties Questionnaire (SDQ),
and coordination problems were assessed using the Developmental Coordination Disorder
Questionnaire (DCDQ) completed by the mothers. A total of 787 mothers completed the SDQ
and 537 completed the DCDQ for children aged 7.01-8.47 years (mean age 7.15 years). The
mean maternal PFOS concentration was 0.036 jig/mL, and PFOS levels were divided into
quartiles: 
-------
score, conduct problems, or peer problems. Odds ratios adjusted for different outcomes were
adjusted for the following confounders: parity, maternal age, pre-pregnancy BMI, smoking and
alcohol consumption during pregnancy, sex of the child, breastfeeding, birth year, housing
density, gestational age at blood draw, and parental behavioral problem scores during their
childhood. Overall, no associations between behavioral or coordination problems in children 7
years of age and prenatal PFOS (and PFOA) exposure were found.

   A prospective birth cohort study called INUENDO3 was designed to examine biopersistent
organochlorines in diet and human fertility (H0yer et al. 2015b). Pregnant females were enrolled
between May 2002 and February 2004 with a total of 1,106 mother-child pairs at follow-up
between January 2010 and May 2012 when the children were 7-9 years old. The study
population consisted of 526 pairs from Greenland, 89 from Poland, and 491 from Ukraine. Since
maternal blood samples for measurement of plasma PFOS levels were taken any time during
pregnancy,  median gestational age at time of collection varied by country (range: 23-33).
Behavior of children was assessed with SDQ score, and logistic regression models were used in
the analyses of PFOS tertiles and behavioral problems. Motor development was assessed with
DCDQ score, and linear regression was used for analyses. All analyses were performed on the
entire cohort,  as well as by country although not all analyses could be performed on the Polish
subset due to the small number of cases. Analyses were adjusted for the following potential
confounders: maternal cotinine level during pregnancy, maternal alcohol consumption at
conception, maternal age at pregnancy, gestational age at blood-sampling, and child gender.

   The median maternal plasma PFOS level was 0.01 |ig/mL for the combined population and
0.02, 0.005, and 0.008 |ig/mL for the pregnant females from Greenland, Ukraine, and Poland,
respectively. No associations were found between PFOS (and PFOA) levels and motor
development score. Total SDQ score was not associated with PFOS levels; however, PFOS
concentrations were associated with higher total SDQ score only in Greenland. The highest
PFOS tertile was associated with a 0.5 point higher hyperactivity scores in the combined analysis
in Greenland (0.3) and Poland (1.3), but no association was found in Ukraine. The adjusted OR
for hyperactive behavior in the combined analysis was 1.4 (95% CI: 0.4-4.9) for the highest
tertile compared to the lowest PFOS tertile, with comparable results found for Greenland and
Ukraine. Although statistical adjustment in the regression models included country of
participant, inter-country differences complicate interpretation of the study results especially
given variability in exposure data collection periods and vastly different participation rates (e.g.,
37% in Poland and 86% in Greenland). In addition to the potential for selection and information
biases, the unknown critical exposure window(s), including the impact of unmeasured post-natal
exposures, for these outcomes increases the uncertainty of these study results.

   Fei et al. (2008b) examined the association between plasma levels of PFOS in pregnant
females and the motor and mental development in their children. The developmental measures
examined in the infants included Apgar score of child at birth and maternal reported
questionnaire responses about child development milestones at 6 and 18 months. Using linear
regression,  no significant association between PFOS  and Apgar score was observed after
adjustment for potential confounders (OR = 1.20; 95% CI: 0.57-2.25). Although these data were
limited by maternal reporting of the outcome data, there was no association between PFOS levels
and motor or mental development as reported in the questionnaire at 6 months. In children at
18 months,  mothers with higher PFOS levels were slightly more likely to report that their babies
3 Biopersistent Organochlorines in Diet and Human Fertility study.

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-23

-------
started sitting without support at a later age and "did not use word-like sounds to tell what he/she
wants." No statistically-significant associations were found with PFOA.

   Hoffman et al. (2010) examined the associations between perfluorochemicals, including
PFOS, and diagnosis of attention deficit hyperactivity disorder (ADHD) using the NHANES data
from 1999-2000 and 2003-2004. The study population included 571 children aged 12-15 years
including those who had been diagnosed as having ADHD (n = 48) and/or taking ADHD
medications (n = 21). Various potential confounders were considered, including birth weight,
admittance to a neonatal intensive-care unit, socioecomonic status, health insurance coverage,
having a routine health care provider, preschool attendance, and lead exposure. NHANES
sample cycle, age, sex, race/ethnicity, living with a smoker, and maternal smoking were adjusted
for in the logistic regression models. The median serum PFOS levels were 0.023 |ig/mL and
ranged from 0.002 to 0.09 ug/mL. Serum PFOS was positively associated with parental report of
ADHD (OR = 1.03, 95% CI:  1.01-1.05). The adjusted odds ratio per each 1 ug/L increase in
serum PFOA for parental report of ADHD and ADHD medication use was 1.05 (95% CI: 1.02-
1.08). Both PFOA and perfluorohexane sulfonate were also positively associated with parentally-
reported ADHD. Data interpretation were limited by the cross-sectional study design, other
potential confounders (e.g., alcohol consumption) that were not included in the available data,
and measurement error resulting from using current PFOS levels as proxy measures of
etiologically relevant exposures.

   In a prospective study, Halldorsson et al. (2012) examined prenatal exposure to PFASs,
including PFOS, and the risk of being overweight at 20 years of age. A birth cohort consisting of
965 singleton pregnancies were recruited from a midwife center in Aarhus, Denmark. Maternal
PFOS levels were measured in serum samples collected during week 30 of gestation for
assessment of in utero PFOS  exposure and offspring anthropometry at 20 years of age. Among
the 965 study subjects, 915 of their offspring were located and 665 agreed to participate.  The
median PFOS concentration was 0.0215 ± 0.0019 |ig/mL with quartiles of 0.016 ± 0.0056,
0.0202 ± 0.0057, 0.0236 ± 0.0068, and 0.0285 ± 0.0021 |ig/mL. Four PFASs, including PFOA,
PFOS, PFOSA, and perfluorononanoic acid (PFNA) exhibited sufficient contrasts to examine
quartiles of exposure; while eight of the other  quantified PFASs did not. PFOS was positively
associated with female offspring BMI at 20 years. Maternal PFOS concentrations were not
associated with offspring anthropometry at 20 years. Associations of PFOS and other variables
including smoking status; waist circumference; or insulin, leptin, or adiponectin concentrations
at 20 years were not reported. Therefore, possible confounding cannot be assessed. Study
strengths include a high rate of participation (69%) in the offspring  analysis and for sample
collection from the original cohort (72%).

   The relationship between maternal PFOS (and PFOA) levels and prevalence of offspring
overweight and waist-to-height ratio > 0.5 was investigated in a subset of the INUENDO
(biopersistent organochlorines in diet and human fertility) prospective birth cohort (H0yer et al.
2015a). Pregnant females were enrolled between May 2002 and February 2004 with a total of
1,022 mother-child pairs at follow-up between January 2010 and May 2012 when the children
were 7-9 years old. The study population consisted of 531 pairs from Greenland and 491 from
Ukraine. The maternal blood  samples for measurement of plasma PFOS levels were taken at a
mean gestational age of 24 weeks, but there was a substantial range of collection windows in
both Greenland (5-42 weeks) and Ukraine (9-40 weeks). The child's weight and height were
measured and used to calculate BMI. All analyses were performed on the entire cohort as well as
by country.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-24

-------
   The median maternal plasma PFOS level was 0.0202 |ig/mL in the pregnant females from
Greenland and 0.0050 |ig/mL in the pregnant females from Ukraine. No associations were found
between PFOS (and PFOA) levels and risk of being overweight in the combined analysis or in
Ukraine. No associations were observed between PFOS and BMI score in either country. In the
combined analysis, an association was detected for having waist-to-height ratio > 0.5 and the
continuous (per each In-unit increase) exposure (RR = 1.38, 95% CI: 1.05-1.82). Comparable
results were noted for PFOA also and waist-to-height ratio > 0.5 in the combined analysis (H0yer
et al. 2015a), although this was not statistically-significant (RR =  1.30, 95% CI: 0.97-1.74).

Reproductive Outcomes in Females. Using the C8 Health Project data, blood samples from a
population of females aged 18-65 years (n = 25,957) were analyzed to determine whether the
onset of menopause, levels of serum estradiol, and the amount of PFAS in the blood were inter-
related (Knox et al. 2011). These data were cross-sectional, with a one-time serum measurement
collected for participants. The mean PFOS level of all the females was 0.018 jig/mL. The
analyses of menopause excluded participants who reported undergoing a hysterectomy. Logistic
regression models were adjusted for age, smoking,  alcohol consumption, BMI, and exercise. The
analysis for menopause was determined upon three groups of females: childbearing (aged 30-
42), perimenopausal (aged > 42-51) and menopausal (aged > 51- < 65). These same groups
were used for the estradiol concentrations except the childbearing group was extended to include
those > 18 years; exclusions for this analyses included pregnant females, females with a full
hysterectomy, or females taking hormones, fertility drugs, or selective estrogen receptor
modulators.

   Among females aged 51-65, statistically-significant ORs for menopause were detected
across PFOS quintiles, including a monotonic dose-response relationship. Similar results were
found with PFOA quintiles (OR range: 1.5-1.7).  Although dose-response relationships were not
evident, consistent ORs for menopause were detected among the perimenopausal age group, as
well for both PFOS and PFOA exposures (OR range:  1.2-1.4). Inverse associations were
detected between estradiol concentrations and PFOS in the perimenopausal group (P  = -3.65;
p < 0.0001) and menopausal group  (P = -0.83; p <  0.007). Serum PFOA and estradiol
concentrations were not associated. Despite the contaminated water supplies, the PFOS exposure
levels were comparable to those from NHANES  and likely represented general population levels.
A study limitation was the one-time serum measurement and cross-sectional study design; thus,
exposure misclassification is likely despite long half-lives reported for PFAS. The level of PFOS
was significantly higher in the set of females that had undergone a hysterectomy. Menopause and
having undergone a hysterectomy, therefore, may be associated with increased serum PFAS due
to the loss of menstruation as a route for removing PFOS with the associated menstrual blood
loss.  Thus, reverse causation cannot be ruled out as an alternative explanation for the study
findings.

   Lopez-Espinosa et al. (2011) evaluated the relationship between pubertal timing and PFOS
levels among 2,931 girls and 3,076 boys  aged 8-18 years from the C8 study. A high proportion
of available participants provided serum biomarkers among both boys (66%) and girls (67%).
The median serum PFOS level was 0.018 |ig/mL among these female participants, and exposures
were examined continuous and categorical (quartiles) variables. Pubertal development was based
on hormone levels (total > 50 ng/dL and free > 5 pg/mL testosterone in boys and estradiol
> 20 pg/mL in girls) or onset of menarche. although participant age at survey and time of day of
blood sampling were the only confounders that were identified and adjusted for, other covariates
considered as potential confounders included BMI z-score, height annual household family
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-25

-------
income, ethnicity, ever smoking, and ever alcohol consumption. A reduced odds of having
reached puberty was found with increasing PFOS levels, with girls having a difference of
138 days between the highest and lowest PFOS quartile. A reduced odds of postmenarche was
found for both PFOS (138 days of delay) and PFOA (130 days of delay).

   Christensen et al. (2011) used data from a prospective cohort study in the United Kingdom to
conduct a nested case-control study examining the association between age at menarche and
gestational exposure to perfluorinated chemicals including PFOS and PFOA. The study
population from the Avon Longitudinal Study of Parents and Children included single-birth
female subjects who had completed at least two puberty staging questionnaires between the ages
of 8 and 13 years and whose mothers provided at least one prenatal serum sample. If more than
one serum sample were available, the earliest sample provided was used for analysis. The study
does not provide information as to when samples were collected. The females were divided into
two groups, including those who experienced menarche prior to age 11.5 years (n = 218 cases),
and a sample of those who experienced menarche after age 11.5 (n = 230  controls) from the
5,756 female offspring enrolled in the Avon study. Confounders including the mother's pre-
pregnancy BMI, age at delivery, age at menarche, educational level, and the child's birth order
and ethnic background were included in linear and logistic regression models used to analyze the
data. The median maternal serum PFOS concentrations were 0.019 and 0.02 |ig/mL for the early
menarche and non-early menarche groups, respectively.

   Although not statistically-significant, decreased adjusted odds ratios for earlier age at
menarche were found for the prenatal PFOS examined as a continuous [OR = 0.68; 95%
CI: 0.40-1.13] and the categorical [OR = 0.83; 95% CI: 0.56-1.23] exposure dichotomized as
the median value (0.0198 |ig/mL). Results were null for the continuous PFOA exposure measure
and slightly elevated for the categorical exposure [OR = 1.29; 95% CI: 0.86-1.93] above the
median value of 0.0037 |ig/mL. The limitations of the study included having a small sample size,
using a single maternal gestational serum sample for perfluorinated chemical measurement, and
the self-reported nature of some covariates including menarche status and age at menarche.

   The relationship between prenatal exposure to PFOS (and PFOA) and female and male
reproductive function was evaluated in 343 females and 169 males whose mothers participated in
an Aarhus, Denmark cohort in 1988-1989 (Kristensen et al. 2013; Vested et al. 2013). Maternal
blood samples were collected during week 30 of gestation. Follow-up was initiated in 2008 when
the offspring were approximately 20 years old. Median serum PFOS level was 0.0211 |ig/mL for
the mothers with daughters evaluated. Potential confounders adjusted for included maternal
smoking during pregnancy, social class, and daughter's BMI. No statistically-significant
association was found between prenatal exposure to PFOS and age of menarche. In adjusted
regression analysis, daughters from mothers in the highest PFOA tertile had a later age at
menarche compared with those in the lowest tertile. No statistically-significant relationships
were found between PFOS (or PFOA) exposure and cycle length, reproductive hormone levels,
and number of follicles assessed by ultrasound (Kristensen et al. 2013). Study limitations
included retrospective collection of some health outcome data, such as age of the menarche,
which was queried 2-10 years afterward.

   Fei et al. (2009) evaluated associations with PFOS levels and fecundity as indicated by the
time to pregnancy (TTP) in the DNBC study population. In females who had a planned
pregnancy (n = 1,240), there was a longer 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 exposure quartiles). The proportion of females with infertility

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-26

-------
(TTP > 12 months) was higher in the upper three quartiles of PFOS versus the lowest quartile.
These trends were statistically-significant. In females who had planned pregnancies (n = 1,240),
there was a longer TTP with higher levels of PFOS (p < 0.001). Females 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. Both TTP and
infertility were also positively associated with serum PFOA levels. The selection of females who
gave birth among only those with planned or partly planned pregnancies may limit study
generalizability. Selection bias is also possible if excluded fertile females who did not plan their
pregnancy had differentially higher or lower PFAS exposures. Additional analyses of unplanned
pregnancies actually resulted in stronger association between PFAS levels and TTP.

    Participants enrolled in the Maternal-Infant Research on Environmental Chemicals Study, a
Canadian pregnancy and birth cohort, were evaluated for an association between serum  PFOS
levels (as well as PFOA and PFHxS) and TTP (Velez et al. 2015). Females (n = 1,743) recruited
from prenatal clinics across 10 Canadian cities between 2008 and 2011 (39% participation rate)
were included in this analysis if they provided a first trimester blood sample collected between
6 and 14 gestational weeks. Infertility was defined as having a TTP of > 12 months or requiring
infertility treatment for the current  pregnancy. The geometric mean plasma PFOS level was
0.00459 |ig/mL. No statistically-significant associations with fecundity were observed, although
an increased risk was observed for  infertility (OR = 1.14;  95% CI:  0.98-1.34) per one SD
increased in PFOS. In contrast, statistically-significant associations were detected for infertility
and reduced fecundity and both PFOA and PFHxS.

Reproductive Outcomes in Males. Lopez-Espinosa et al. (2011) also included 3,076 boys aged
8-18 years  from the C8 database in their analysis, with a high proportion of available
participants providing serum biomarkers (66%).  The median serum PFOS level was 0.020
|ig/mL among these male participants. Pubertal development was based on hormone levels (total
> 50 ng/dL and free > 5 pg/mL testosterone). Reduced odds of reaching puberty in boys (i.e.,
raised testosterone) was detected with increasing PFOS (delay of 190 days between the  highest
and lowest  quartile).

    Reproductive function and other reproductive endpoints also were evaluated in the sons of
the mothers who participated in the Aarhus, Denmark cohort (Kristensen et al. 2013). The
median (25th-75th percentile) serum PFOA level was 0.0212 |ig/mL (0.017.4-0.026.5 ng/mL) for
the mothers with sons who were evaluated. PFOS was not associated with any outcome of
reproductive function analyzed with multivariable regression models. No associations were
found between PFOS (and PFOA) levels and percentage of progressive sperm, sperm
morphology, semen volume, or testicular volume. Monotonic exposure-response relationships
were detected for in utero PFOA exposure and sperm concentration, total sperm count, and
percentage  of progressive spermatozoa (based on the computer-assisted semen analysis), and
positive associations for follicle-stimulating hormone (FSH) and luteinizing hormone (LH) levels
were associated with PFOA (Vested et al. 2013).

    Joensen et al. (2009) investigated the relationship between PFAS and semen quality in a
cross-sectional study of 105 Danish males. The study participants were recruited in 2003 from a
sample of 546 males from a compulsory medical examination for all young Danish males being
considered  for military service. They represented the individuals with the lowest and highest
testosterone levels in that study population. Nine PFAS were measured from frozen, archived
(5 years) samples, while the semen samples were collected during the 2003 examination.
Confounders adjusted for in the various regression models included duration of abstinence and

Perfluorooctane  sulfonate (PFOS) - May 2016                                             3-27

-------
time between ejaculation and semen analysis. The median PFOS serum level in the 105 study
participants was 0.025 jig/mL. Males with high combined levels of PFOA/PFOS had a median
level of 6.2 million morphologically normal spermatozoa compared to 15.5 million in males with
low PFOA/PFOS levels (p = 0.030).

   There was no statistically significant association between testosterone levels and PFAS
exposures and no difference in PFAS levels between high and low testosterone groups. To
address previous study limitations and expand the generalizability of the findings, a later study
by Joensen et al. (2013) was conducted to investigate the associations between serum PFOS
concentration and reproductive hormones and semen quality. Study participants included a
random sample of 247 healthy young Danish males (mean age 19.6 years) recruited in 2008-
2009 from the same study population. Serum samples were analyzed for PFOS, as well as total
testosterone (T), estradiol (E), sex hormone-binding globulin (SFffiG), LH, FSH, and inhibin-B.
Semen samples were collected the same morning as the blood samples, and self-administered
questionnaires were also completed by the study participants. Confounders adjusted for in the
various regression models included time to semen analysis, abstinence time, BMI, and smoking.
The mean PFOS level was 0.0085 jig/mL. Inverse associations were detected for PFOS and
various outcomes including T, calculated free T (FT), free androgen index (FAI),  and ratios of
T/LH, FT/LH, and FAI/LH (all p-values < 0.05). PFOS was also inversely associated with
estradiol, T/E ratio, and inhibin-B/FSH ratio, and positively associated with SFffiG, LH, FSH,
and inhibin-B, although statistical significance was not attained. No associations were detected
between PFOS levels and any semen quality parameters. Study strengths included improved
generalizability due to the random selection of subjects from the general population and a higher
participation rate was (30%) compared to other population-based semen quality studies.

   The relationship between serum PFOS exposures and 35 semen quality parameters was
evaluated in 462 males enrolled in the LIFE Study cohort (Buck Louis et al. 2015). The males
were recruited from 501 couples discontinuing contraception for the purposes of becoming
pregnant and residing in 16 counties from Michigan and Texas. Forty-two percent of eligible
couples enrolled in the study, and the 462 males provided at least one semen sample. Linear
mixed models were adjusted for age, BMI,  smoking, abstinence time, sample age, and study site.
The study participants had a mean age of 31.8 years and mean PFOS levels were 0.017 |ig/mL
for Michigan residents and 0.021 |ig/mL for Texas residents. Statistically-significant associations
were detected between PFOS exposures and for a lower percentage of sperm with coiled tails; no
associations were found for any other endpoint. In total, six PFAS (including PFOS) were
associated with changes in 17 semen quality endpoints. Study strengths included improved
generalizability, since participants were from the general population and had a higher
participation rate (42%) compared to other population-based semen quality studies. A key study
limitation of this and many of these types of epidemiology studies is the uncertainty related to
the critical exposure window(s) relative to timing of the collected samples and the multiple
comparisons (n = 245) that were examined.

   Raymer et al. (2012) conducted a cross-sectional study of the relationships between PFAS
and semen quality and reproductive hormones. The study population included 256 males
recruited between 2002 and 2005 from Duke University Medical  Center's IVF Clinic.
Reproductive health questionnaires were administered to participants. Blood and semen samples
were used to detect PFAS and were both collected at the time of evaluation. Linear and logistic
regression models were used to calculate effect estimates and were adjusted for age, period  of
abstinence, and tobacco use. The average PFOS levels in plasma were 0.0374 |ig/mL and
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-28

-------
0.0008 |ig/mL in semen. The strongest correlations detected between PFAS and hormones were
between plasma PFOS and LH (r = 0.12), plasma PFOA and LH (r = 0.16), plasma PFOS and
triiodothyronine (r = 0.14), as well as semen PFOS and FSH (r = 0.13). No statistically-
significant associations were detected between PFOS and PFOA concentrations and reproductive
hormones or different semen quality outcomes. The older population (mean age = 42 years) may
limit comparability with previous studies and generalizability of study findings.

   The INUENDO prospective birth cohort study of persistent organic pollutants and fertility
was used to examine the relationship between PFAS and semen quality parameters (Toft et al.
2012). The study population included 588 males (97%) from Greenland, Poland and Ukraine
who provided a semen sample among the underlying 607 male partners of 1710 pregnant
females. PFOS levels were quantified from serum samples; these were categorized into tertiles
and also examined as continuous exposures. Linear regression models and categorical analyses
were adjusted for the following potential confounders: age, abstinence time, spillage, smoking,
urogenital infections, BMI, and country of origin. For the categorical analysis combining the
three cohorts, compared to the first tertile, the percent of normal sperm cells was decreased in the
upper two serum PFOS tertiles with a decrease of 22% (95% CI: l%-44%) and 35% (95% CI:
4%-66%) in the second and  third PFOS tertiles, respectively. Exposure-response relationships
were detected for the overall population based on the continuous PFOS exposure data, although
this was only evident among the Polish and Ukrainian populations. No other associations
between PFOS exposure and semen  quality parameters were noted. The variable participation
rates across study sites and potential for participation bias (i.e., if participation was related to
fertility status and exposure levels) complicate interpretation of these results. The cross-sectional
nature of this study also limits the ability to draw causal inference from these types of studies,
especially since temporality  could not be established some of the study population based on the
timing of the blood and semen samples (e.g., nearly 60% of the Greenland samples were
collected approximately a year before the semen samples).

Summary. Fetal growth retardation was examined through measures including mean birth
weight, LEW, and small for  gestational age. Mean birth weight examined as a continuous
outcome was the most  commonly examined endpoint for epidemiology studies of serum/cord
PFOS exposures. Although three studies were null (Fei  et al. 2008a; Hamm et al. 2010; Monroy
et al. 2008), birth weight deficits ranging from 29 to 149 grams were detected in five studies
(Apelberg et al. 2007; Chen et al. 2015; Darrow et al. 2013; Maisonet et al. 2012; Washino et al.
2009). Larger reductions (from 69 to 149 grams) were noted in three of these studies (Apelberg
et al. 2007; Chen et al.  2015; Washino et al.  2009) on the basis of per unit increases in
serum/cord PFOS exposures, while the lone categorical data showed an exposure-response
deficit in mean birth weight up to 140 grams across the PFOS tertiles (Maisonet et al. 2012).
Two (Chen et al. 2015; Whitworth et al. 2012) out of four (Fei et al. 2007; Hamm et al. 2009)
studies of SGA and serum/cord PFOS exposures showed some suggestion of increased ORs
(range: 1.3-2.3), while three (Chen et al. 2012; Fei et al. 2007; Stein et al. 2009) out of four
(Darrow et al. 2014) studies  of LEW showed increased risks (OR range: 1.5-4.8). Although a
few of these  studies showed  some suggestion of dose-response relationships across different fetal
growth measures (Fei et al. 2007; Maisonet et al. 2012;  Stein et al. 2009), study limitations,
including the potential  for exposure misclassification, likely precluded the ability to adequately
examine exposure-response patterns. While there is  some uncertainty in the interpretation of the
observed association between PFOS and birth weight given the potential impact of low GFR, the
available information indicates that the association between PFOS exposure and birth weight for
the general population  cannot be ruled out. In humans with low GFR (which includes females

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-29

-------
with pregnancy induced hypertension or preeclampsia) the impact on body weight is likely due
to a combination of the low GFR and the serum PFOS.

   A small set of studies observed an association with gestational diabetes (preconception serum
PFOS; Zhang et al. 2015), pre-eclampsia (Stein et al. 2009) and pregnancy-induced hypertension
(Darrow et al. 2013) in populations with serum PFOS concentrations of 0.012 - 0.017 ug/mL.
Zhang et al. (2015) and Darrow et al. (2013) used a prospective assessment of adverse pregnancy
outcomes in relation to PFASs which addresses some of the limitations the available cross-
sectional studies. Associations with these outcomes and serum PFOA also were observed.

   Although there was some suggestion of an association between PFOS exposures and semen
quality parameters in a few studies (Joensen et al. 2009; Toft et al. 2012), most studies were
largely null (Buck Louis et al. 2015; Ding et al. 2013; Joensen et al. 2013; Raymer et al. 2012;
Specht et al. 2012; Vested et al. 2013). For example, morphologically abnormal sperm associated
with PFOS were detected in three (Buck Louis et al. 2015; Joensen et al. 2009; Toft et al. 2012)
out of nine (Buck Louis et al. 2015; Ding et al. 2013; Joensen et al. 2013; Raymer et al. 2012;
Specht et al. 2012; Vested et al. 2013) studies.

   Small increased odds of infertility was found for PFOS exposures in  studies by J0rgensen et
al. (2014)  [OR= 1.39; 95% CI: 0.93-2.07] and Velez etal. (2015) [OR = 1.14; 95% CI: 0.98-
1.34]. Although one study was null (Vestergaard et al. 2012), PFOS exposures were associated
with decreased fecundability ratios (FRs), indicative of longer time to pregnancy, in studies by
Fei et al. (2009) [FR = 0.74 (95% CI: 0.58-0.93) and in studies by J0rgensen et al. (2014)
[FR = 0.90; 95% CI: 0.76-1.07]. Whitworth et al. (2012) data suggested that reverse causality
may explain their observation of subfecundity odds of 2.1 (95% CI: 1.2-3.8) for the highest
PFOS quartile among parous  females, but a reduced odds among nulliparous females (OR = 0.7;
95% CI: 0.4-1.3). A recent analysis of the pooled DNBC study samples found limited evidence
of reverse causality with an overall FR of 0.83 (95% CI: 0.72-0.97) for PFOS exposures, as well
as comparable ratios for parous (0.86; 95% CI: 0.70-1.06) and nulliparous (0.78;  95% CI: 0.63-
0.97) females (Bach et al. 2015). The same authors reported an increased infertility OR of 1.75
(95% CI: 1.21-2.53) and OR for parous (OR =1.51; 95% CI: 0.86-2.65) and nulliparous
(OR= 1.83; 95% CI: 1.10-3.04) females. Although there remains some concern over the
possibility of reverse causation explaining some previous study results, these collective findings
indicate a consistent association with fertility and fecundity measures and PFOS exposures.

3.1.1.5  Thyroid Effect Studies

Occupational Populations. In the cross-sectional study described above for production workers,
thyroid hormone (TH) levels were also measured in male (n = 215) and female (n = 48)
volunteers working at the Decatur, Alabama plant and male (n = 206) and female (n = 49)
volunteers working at the Antwerp, Belgium plant (Olsen et al. 2001a). The mean PFOS level in
all employees from the Decatur and Antwerp plants was 1.40 |ig/mL (range: 0.11-10.06 jig/mL)
and 0.96 |ig/mL (range: 0.04-6.24 jig/mL), respectively. No significant associations were found
for quartile of PFOS level and thyroid-stimulating hormone (TSH), serum thyroxine (T4), free
thyroxine (FT4), triiodothyronine (T3), and thyroid hormone binding ratio.

General Population. 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 et al. 2009). Those using
medication for thyroid disease and pregnant females were not included in the study.

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-30

-------
Concentrations of TSH, 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 consumption, medications taken, and dietary fish consumption. The
study detected PFOS in 100% of individuals, with a mean plasma PFOS concentration of
0.018 ng/mL (95% CI: 0.017-0.019 |ig/mL). PFOS was negatively associated with circulating
levels of TSH, TT3, and TBG and positively associated with FT4. The results suggest that
human thyroid hormone levels could be affected by PFOS exposure. However, because the
majority of individuals were reported by the authors to have normal thyroid gland function and
the thyroid hormone levels were in the normal range,  it is uncertain that these relationships are
connected to thyroid disease or are a reflection of hormone variability in the human population.

   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 (1,900 males and 2,066 females) were included. Of
these, 292 females and 69 males reported thyroid disease. Overall mean PFOS levels were
0.025 |ig/mL for males and 0.019 |ig/mL for females. The data showed that males with PFOS
levels in the highest quartile  > 0.037 |ig/mL  were more likely to report currently treated thyroid
disease than males with PFOS levels in the lowest two quartiles combined, < 0.026 |ig/mL (OR
= 2.68; 95% CI: 1.03-6.98; p = 0.043). Females had lower levels of PFOS than males and higher
prevalence of thyroid  disease, but serum PFOS concentration was not significantly associated
with treated thyroid disease.  With PFOA, the opposite was found, with females in the highest
quartile, but not males, more likely to report 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. Thyroid hormone levels were
not reported by Melzer et al.  (2010).

   Another study of 1,181 members of NHANES for survey years 2007-2008 and 2009-2010
examined  the association between serum PFOS levels (and 12 other PFASs) and thyroid
hormone levels (Wen  et al. 2013). Multivariable linear regression models were used with serum
thyroid measures as the dependent variable and individual natural log-transformed PFAS
concentration as a predictor along with confounders. The geometric mean serum PFOS level was
0.0142 |ig/mL. No associations between PFOS level and thyroid hormones were found in males
and females. However in 23  individuals defined as subclinical hypothyroid (TSH above normal
range), a 1-unit increase in natural log-PFOS was positively associated with hypothyroidism (OR
= 3.03; 95% CI: 1.14-8.07 in females; OR= 1.98; 95% CI: 1.19-3.28 for males; both p < 0.05).

   Webster et al. (2015) also used NHANES 2007-2008 data from 1,525 adults to explore the
contribution of PFOS  exposure to those with risk factors for thyroid disease, low iodide status
and/or high thyroid peroxidase antibody (TPOAb). Webster et al. 2015  saw that people with both
elevated TPOAb and low iodide (those at risk for thyroid insufficiency) were more susceptible to
PFOS associated disruption of thyroid hormone concentrations than were people without these
two risk factors.

   Bloom et al. (2010)  examined the potential association between serum concentrations of
eight polyhalogenated compounds, including PFOS, and human thyroid function. Levels of TSH
and FT4 were measured in a subsample of participants in the cross-sectional New York State
Angler Cohort Study (27 males and 4 females). 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

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-31

-------
high concentration compared to the other PFASs measured with a mean concentration of
0.0196 ng/mL (95% CI: 0.0163-0.0235). The results indicated no significant association
between PFOS serum concentration (or PFOA) and thyroid hormone levels, potentially due to
the study's small sample size.

   The relationship between thyroid biomarkers and serum levels of PFOS, PFOA, and other
persistent organic pollutants was investigated in older adults (Shrestha et al. 2015). Levels of
TSH, FT4, T4, and T3 were measured in 51 males and 36 females with a mean age of 63.6 years.
None of the participants had thyroid disease or were taking thyroid medication. Covariates in the
analysis included age, sex, education level, polychlorinated biphenyl (PCB) and PBDE exposure,
smoking status, and alcohol consumption. The mean PFOS serum level was 0.0366 ± 0.023 |ig/mL
for all participants.  In both unadjusted and adjusted models, PFOS was significantly (p < 0.05 or
0.01) and positively associated with FT4 and T4; a possible dose-response was not evaluated in
this small sample.

   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, and 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.

   A cross-sectional study of 903 pregnant females  evaluated the association between plasma
PFOS levels and plasma TSH (Wang  et al. 2013). Twelve other perfluoroalkyl substances were
also quantified and evaluated. The females were a cohort of the Norwegian Mother and Child
Cohort Study, and the blood samples were drawn at  approximately week 18 of gestation. The
median PFOS concentration was 0.013 |ig/mL with an interquartile range of 0.010-0.017 |ig/mL.
A trend was  observed for increasing TSH across PFOS quartiles, with females in the third and
fourth quartiles having significantly higher TSH levels compared with the first quartile. After
adjustment, each 0.001 |ig/mL increase in PFOS  concentration was associated with a 0.8% (95%
CI: 0.1%-1.6%) rise in TSH. The odds ratio of having an abnormally high TSH, however, was
not increased. The plasma levels of other perfluoroalkyl substances were not related to TSH
levels.

   Expanding on the above study, Berg et al. (2015) investigated the association between a
number of perfluoroalkyl substances,  including PFOS, and  TSH, T3, T4, free triiodothyronine
(FT3),  and FT4. A subset of 375 females on the Norwegian Mother and Child Cohort Study with
blood samples at about gestational week 18 and at 3  days and 6 weeks  after delivery were
included. Seven compounds were detected in > 80% of the  blood samples with PFOS present in
the greatest concentration. The median PFOS level was 0.00803  |ig/mL and the females were
assigned to quartiles based on the first blood sample at week 18 of gestation. After adjustment
for covariates (parity, age, thyroxin binding capacity, BMI), TSH was positively associated with
PFOS.  Females in the highest quartile had significantly higher mean TSH at all three time points
compared to females in the first quartile. No associations were found between PFOS and the
other thyroid hormone levels.

   Maternal and umbilical cord blood concentrations of a number of fluorinated organic
compounds,  including PFOS, were determined in 15 females (17-37 years of age) and their

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-32

-------
newborns at Sapporo Toho Hospitals in Hokkaido, Japan from February 2003 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 0.0049 to 0.0176 |ig/mL, and cord blood PFOS ranging from
0.0016 to 0.0053 |ig/mL. TSH and FT4 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 in 2005-2006 from females
in Canada (mean age 31.3  years) at 15-20 weeks gestation and measured thyroid hormones, FT4
and the level of PFAS to determine whether PFAS levels were associated with hypothyroxinemia.
From the samples, there were 96 identified as cases of hypothyroxinemia and 175 identified as
controls. The cases had normal TSH concentrations and free T4 concentrations in the lowest 10th
percentile (< 8.8 pmol/L).  The controls had normal TSH concentrations and free T4
concentrations between the 50th and 90th percentiles (12-14.1 pmol/L). The geometric mean for
PFOS was 0.0074 |ig/mL.  The mean free T4 levels were 7.7 pmol/L in the cases and 12.9 in the
controls. The mean TSH concentrations were 0.69 milli-Units/L in the cases and 1.13 in the
controls. Analysis by conditional logistic regression indicated that the concentration of PFOS (or
PFOA) was not significantly associated with hypothyroxinemia. 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.

   A similar study of 152 Canadian females evaluated maternal serum PFOS levels (and
PFHxS,  PFNA, PFOA) for associations with thyroid hormone levels during the early second
trimester of pregnancy, weeks 15-18 (Webster et al. 2014). Mixed effects linear models were
used to examine associations between PFOS levels and FT4, total T4, and TSH; associations
were made for all females  and separately for females with high levels of thyroid peroxidase
antibody, a marker of autoimmune hypothyroidism. Median serum PFOS was 0.0048 jig/mL. No
associations were found between levels of PFOS  (or PFOA and PFHxS),  and thyroid hormone
levels in females with normal antibody levels. PFNA was positively associated with TSH.
Clinically elevated thyroid peroxidase antibody levels were found in 14 (9%)  of the study
population. In the females with high antibody levels, PFOS, PFNA, and PFOA were strongly and
positively associated with TSH. An IQR increase in maternal PFOS concentrations was associated
with a 69% increase in maternal TSH compared to the median TSH level. PFNA and PFOA
concentrations were associated with 46% and 54% increases, respectively, in maternal TSH.

   Numerous epidemiologic studies have evaluated thyroid hormone levels, thyroid disease, or
both in association with serum PFOS concentrations (Table 3-6). These epidemiologic studies
provide  limited support for an association between PFOS exposure and incidence or prevalence
of thyroid disease, and they include large studies  of representative samples of the general U.S.
adult population (Melzer et al. 2010; Wen et al. 2013). These highly powered studies reported
associations between PFOS exposure (serum PFOS concentrations) and thyroid disease but not
thyroid hormone status. Melzer et al. (2010) studied thyroid disease with  medication and Wen et
al. (2013) studied subclinical thyroid disease. In studies of pregnant females, PFOS was
associated with increased TSH levels (Berg et al. 2015; Wang et al. 2013; Webster et al. 2014).
Thyroid function can be affected by iodide sufficiency and by autoimmune disease. Pregnant
females  testing positive for the anti-thyroid peroxidase (TPO) biomarker  showed a positive
association with PFOS and TSH (Webster et al. 2014). An association with PFOS and TSH and
T3 was found in a subset of the NHANES population with both low iodide status and positive
anti-TPO antibodies (Webster et al. 2015). These studies used anti-TPO antibody levels as an
indication of stress to the thyroid system, not a disease state. Thus, the association between
PFOS and altered thyroid hormone levels is stronger in people at risk for thyroid insufficiency.
Perfluorooctane sulfonate (PFOS) - May 2016                                            3-33

-------
            Table 3-6. Summary of Epidemiology Studies of PFOS and Thyroid Effects
          Study
                             PFOS level
                             TSH
                                  T3
                                                        T4
Olsenetal. 200la
n = 263 Decatur, AL plant
n = 255 Antwerp, Belgium
plant	
Decatur plant: 1.4
Antwerp plant: 0.96
No effect observed.
No effect observed.
No effects observed.
Dallaire et al. 2009
Canada
n = 623 (adult Inuit
population)
Adjusted for sex, age, BMI,
education, lipids and smoking
0.018
Adjusted Beta = -0.102 (p
<0.05)
Adjusted Beta =
-0.017 (p< 0.05)
Adjusted Beta =0.014
(p<0.05)
Melzeretal. 2010
n = 3,966 adults,
> 20 yrs old
NHANES (1999-2000;
2003-2004 and 2005-2006)
0.025 (men)
0.019 (women)

Men (ug/mL)
Ql: 0.0003-0.018
Q2: 0.0182-0.0255
Q3: 0.0256-0.0367
Q4: 0.0368-0.435
Similar cut-points in
women
Self-Reported on thyroid disease, with medication use (fully adjusted);
OR (95% CI), p-value
Men
Ql: 1 (referent)
Q2: 0.43 (0.17, 1.08), p = 0.073
Q3: 0.95 (0.34, 2.70), p = 0.926
Q4: 1.89 (0.72, 4.93), p = 0.190
Q4 vs Q1&2: 2.68 (1.03, 6.98), p = 0.043
            Women
            Ql: 1 (referent)
            Q2: 1.05 (0.55, 2.00), p = 0.89
            Q3: 0.81 (0.44, 1.51), p = 0.496
            Q4: 1.31 (0.72, 2.36), p = 0.269
            Q4vsQl&2: 1.27(0.82,1.97)
Wen etal. 2013
United States, NHANES
2007-2008,2009-2010
n= 1,181 adults, aged > 20
yrs
Linear regression, adjusted,
with sampling weights
0.0142
Subclinical hypothyroidism (fully adjusted); OR (95% CI), p-value
Men                            Women
1.98 (1.19, 3.28), p < 0.05          3.03 (1.14, 8.07)

No associations between serum PFOS and thyroid hormones.
 Webster etal. 2015
n= 1,525 adults
NHANES (2007-2008)

Results are on those with
high TPOAb and low iodine-
n = 26
Geometric mean:
0.014
% difference in serum
thyroid hormones for each
IQ ratio increase in PFOS
(95% CI), p-value (n = 26)

17.1(6.6, 28.7), p< 0.05
% difference in serum
thyroid hormones for
each IQ ratio increase
in PFOS (95% CI),
p-value (n = 26)

4.7 (3.9,5.5), p< 0.05
% difference in serum
thyroid hormones for
each IQ ratio increase in
PFOS (95% CI), p-value
(n = 26)
                                                                                            0.05
Bloom etal. 2010
n = 31 adults, subset of New
York Angler Cohort study
0.0196
Log-PFOS and log-TSH,
(95% CI), p-value

Beta = 0.04 (-0.52,0.59), p
= 0.90
NM
Log-PFOS and log-FT4,
(95% CI), p-value

Beta = 0.03 (-0.17,
0.10),p = 0.62
Shrestha etal. 2015
n = 87 adults (mean age of
64)
United States
Geometric mean:
0.036
Log-PFOS and log-TSH,
(95% CI), p-value

Beta = 0.129 (-0.02, 0.28),
p= 0.09
Log-PFOS and log-T3,
(95% CI), p-value

Beta = 2.631 (-2.25,
7.5 l),p= 0.29
Log-PFOS and log-FT4,
(95% CI), p-value
Beta = 0.054 (0.002,
0.11),p = 0.04

Log-PFOS and log-T4,
(95% CI), p-value
Beta = 0.766 (0.33,
1.21), p = 0.001	
Pirali et al. 2009
n = 28 patients undergoing
thyroid surgery
n = 7 control group
5.3 ng/g thyroid
tissue

No association with
PFOS concentration
and thyroid disease
NM
NM
NM
  Perfluorooctane sulfonate (PFOS) - May 2016
                                                                                3-34

-------
Study
Wang etal. 2013
n = 903 women
Norway (from case-control
study of subfecundity in the
Norwegian Mother and
Child Cohort Study; cases
and controls combined)
Blood sample (mean 1 8
weeks pregnancy)
Berg etal. 2015
n = 375 women in the
Norwegian Mother and
Child Cohort Study
Blood samples at week 18,
and 3 days/6 weeks post-
delivery
Inoue et al. 2004
n = 15 women (17-37 yrs
old)
Japan
Chan etal. 2011
n = 96 identified as cases of
hypothyroxinemia
n= 175 controls
Canada (2005-2006)
Webster etal. 2014
n = 1 52 women
Canada
Blood samples taken during
weeks 15-18 of pregnancy
PFOS level
(jig/mL)
Median: 0.013
Median: 0.00803
(ug/mL)
Ql: 0.0003-0.0057
Q2: 0.0058-0.008
Q3: 0.0081-0.011
Q4: 0.0111-0.0359
0.0016-0.0053
(cord blood)
0.0049-0.0176
(maternal blood)
Geometric mean:
0.0074
Median: 0.0048
TSH
PFOS and In-TSH (95%
CI)
Beta= 0.012 (0.005, 0.019)
PFOS and In-TSH mLU/L
(95% CI), p-value
Ql: 1 (referent)
Q2: 0.18 (0.06, 0.31), p =
0.11
Q3: 0.26 (0.13, 0.40), p =
0.03
Q4:0.35 (0.21, 0.50), p =
0.00
No correlation between
PFOS and TSH
Association of
hypothyroxinemia with
PFOS exposure, OR (95%
CI), adjusted
OR =0.88 (0.63, 1.24)
Beta per IQR PFOS and
TSH, (95% CI, p-value)
Normal TPOAb
0.07 (-0.06, 0.2), p = 0.28
High TPOAb
0.9 (0.2, 2), p = 0.02
[IQR PFOS = 0.0033
ug/mL]
T3
NM
No association
NM
NM
NM
T4
NM
No association
No correlation between
PFOS and free T4
No association
Beta per IQR PFOS and
free T4, (95% CI), p-
value
Normal TPOAb
0.05 (-0.1, 0.2), p = 0.58
High TPOAb
-0.7 (-2, 0. 3), p = 0.18
[IQR PFOS = 0.0033
ug/mL]
   In people without diagnosed thyroid disease or without biomarkers of thyroid disease,
thyroid hormones (TSH, T3, or T4) show mixed effects across cohorts. Studies of thyroid disease
and thyroid hormone concentrations in children and pregnant females found mixed effects. TSH
was the indicator most frequently associated with PFOS in studies of pregnant females. In cross-
sectional studies where thyroid hormones were measured in association with serum PFOS,
increased TSH was associated with PFOS exposure in the most cases (Berg et al. 2015; Wang et
al. 2013; Webster et al. 2014), but this association was null in a smaller study with 15
participants (Inoue et al. 2004).

   A case-control study of hypothyroxinemia (normal TSH and low free T4) in pregnant
females (Chan et al. 2011), did not show associations of disease with PFOS exposure; in most
other thyroid diseases, T4 and its compensatory TSH co-vary. In children from the C8 cohort,
increasing PFOS was associated with increased T4 in children aged 1 to 17 years (Lopez-
Espinosa et al. 2011); PFOS was not associated with hypothyroidism. A small South Korean
study examined correlations between maternal PFASs during pregnancy and fetal thyroid
Perfluorooctane sulfonate (PFOS) - May 2016
3-35

-------
hormones in cord blood (Kim et al. 2011). PFOS was associated with increased fetal TSH and
with decreased fetal T3 (Kim et al. 2011). Studies of pregnant females show associations
between TSH and PFOS, and studies in children show mixed results.

3.1.1.6  Immunotoxicity

Immune suppression

   Immune function, and specifically immune system suppression, can affect numerous health
outcomes, including risk of common infectious diseases (e.g., colds, flu, otitis media) and some
types of cancer. The World Health Organization (WHO) guidelines for immunotoxicity risk
assessment recommend measures of vaccine response as a measure of immune effects, with
potentially important public health implications (WHO 2012).

   Okada et al. (2012) investigated the relationship between maternal PFOS concentration (and
PFOA) and otitis media (and allergic conditions), as well as cord blood Immunoglobulin E (IgE)
levels during the first 18 months of life. 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. Infectious diseases and infant allergies were assessed through a self-administered
questionnaire in mothers at 18 months post-delivery. Polynomial regression analyses, adjusted
for potential confounders, were performed on log-transformed data. Mean maternal PFOS
concentration was 0.0056 |ig/mL and cord blood IgE level was 0.62 international units (IU)/mL.
No significant associations were observed between maternal PFOS levels (or PFOA) and cord
blood IgE levels or incidence of otitis media, wheeze, food allergy, or eczema in infants  at 18
months of age.

   The population from the DNBC studies  evaluated 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. Information was collected by telephone interview. No clear
pattern was identified when results were stratified by child's age at the time of hospitalization for
an infectious disease and the level of PFASs in the maternal blood, although effect modification
by sex was indicated (i.e., associations were seen in girls but not in boys). Hospitalizations
among girls increased with higher prenatal PFOS concentration (incidence rate ratio [IRR] for
trend across PFOS quartiles = 1.18, 95% CI: 1.03-1.36). Mean maternal plasma levels were
0.0353 |ig/mL, with a range of 0.0064-0.107 |ig/mL.

   Two other studies, described below, examined reported history of colds and gastroenteritis in
children (up to age 3 years) (Granum et al. 2013) or colds and flu in adults (Looker et al. 2014).
Neither study reported associations with PFOS concentration.

   Three studies have examined response to one or more vaccine (e.g., measured by antibody
titer) in relation to higher exposure to PFOS in children  (Grandjean et al. 2012;  Granum  et al.
2013) or adults  (Looker et al. 2014); the latter study was conducted in the high-exposure C8
community population (Table 3-7).
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-36

-------
   Antibody responses to diphtheria and tetanus toxoids following childhood vaccinations were
assessed in context of exposure to perfluorinated compounds (Grandjean et al. 2012). The
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 (geometric mean 0.0273 |ig/mL; IQR 0.0232-0.0331);
postnatal exposure was assessed from serum collected from the child at 5 years of age (geometric
mean 0.0167 |ig/mL; IQR 0.0135-0.0211). 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 inversely associated with antidiphtheria antibody
concentration (-39%) at age 5 before booster. In addition, an association of antibody
concentrations at age 7 was found with serum PFOS concentrations at age 5. 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 lU/mL.  At age 5, the odds ratios of antibody concentrations
falling below this level for diphtheria were 2.48 (95% CI: 1.55-3.97) compared with maternal
and 1.60 (95% CI: 1.10-2.34) compared with age 5 serum PFOS concentrations. For age 7
antibody levels associated with age 5 PFOS serum concentrations, odds ratios for inadequate
antibody concentration were 2.38 (95% CI: 0.89-6.35) for diphtheria and 2.61 (95% CI: 0.77-
8.92) for tetanus. Models were adjusted for maternal serum PCB concentration. Similar
associations were also observed with PFOA concentrations.

   The effects of prenatal exposure to perfluorinated compounds on vaccination responses and
clinical health outcomes in early childhood were investigated in a subcohort of the Norwegian
Mother and Child Cohort Study (Granum et al. 2013). A total  of 56 mother-child pairs, for whom
both maternal blood samples at delivery and blood samples from the children at 3 years of age,
were evaluated. Antibody liters specific to measles, rubella, tetanus, and influenza were
measured  as these vaccines are part of the Norwegian Childhood Vaccination Program. Serum
IgE levels were also measured. Mean maternal plasma PFOS concentration was 0.0056 |ig/mL at
delivery; the PFOA level was 0.0011 |ig/mL and PFNA and PFHxS were below the limit of
quantitation. PFOS levels in the children were not measured. A slight, but significant, inverse
relationship between maternal PFOS level and anti-rubella antibodies in children at 3 years was
found (P = -.08 [95% CI: -0.14 to -0.02]). No associations were found with PFOS or any
perfluorinated compound and antibody levels to the other vaccines.

   A cohort of 411 adult members of the C8 Health Project was evaluated in 2010 to determine
whether there was an association between serum PFOS levels and antibody response following
vaccination with an inactivated trivalent influenza vaccine (Looker et al. 2014). A prevaccination
serum sample was collected at the time of vaccination and a post-vaccination serum sample was
collected 21 ± 3 days later. The geometric mean serum PFOS level was 0.0083 |ig/mL (95% CI:
0.0077-0.0091),  and participants were divided into quartiles for analyses. Vaccine response, as
measured  by geometric mean antibody titer rise, was not affected by PFOS exposure.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-37

-------
       Table 3-7. Summary of Epidemiology Studies of PFOS and Immune Suppression
                              (Infectious Disease and Vaccine Response)
 Reference and Study Details |    PFOS level
                                             Results
                                        General Population: Children
 Okadaetal. 2012
 Japan, birth cohort study, July
 2002-October 2005 enrollment;
 follow-up to 18 months; n = 343
 Log-transformed PFOS in blood
 after second trimester
 Logistic regression adjusting for
 maternal age, maternal
 educational level, parity, infant
 gender, breast-feeding period,
 environmental tobacco smoke at
 18 months, day care attendance,
 period of blood sampling.	
Mean 0.0056
ug/mL
Incidence otitis media 17.8% (n = 61)
OR (95% CI) n
Overall 1.40 (0.33, 6.00) n = 343
Males 1.38(0.18, 10.60) n= 169
Females 1.43 (0.17, 12.30) n= 174
 Feietal. 201 Ob
 Denmark, birth cohort study,
 1996-2002, follow-up through
 2008; Number hospitalizations
 219 girls, 358 boys
 Maternal blood sample median 8
 weeks gestation
 Poisson regression adjusting for
 parity, maternal age, pre-
 pregnancy BMI, breastfeeding,
 smoking during pregnancy, socio-
 occupational status, home density,
 child's age, gender of child,
 sibling age difference, gestational
 age at blood draw, birth year, and
 birth season.
Mean 0.0353
ug/mL
Quartiles
Ql 0.0064-0.026
Q2 0.0261-0.0333
Q3 0.0334-0.0432
Q4 > 0.433
Adjusted IRR for hospitalization for infectious diseases by gender,
IRR (95% CI) n
         Overall
              Ql     1.0n=147
              Q2     0.93(0.71, 1.21) n= 142
              Q3
              Q4
         Trend 1.00
         Girls
              Ql
              Q2
              Q3
              Q4
         Trend 1.18
         Boys
              Ql
              Q2
              Q3
              Q4
         Trend 0.90
                     0.90(0.68, 1.18) n= 136
                     1.00(0.76, 1.32) n= 152
                   (0.91, 1.09)

                     1.0n=39
                     1.14(0.73, 1.791) n = 48
                     1.61 (1.05, 2.47) n = 67
                     1.59 (1.02, 2.49) n = 65
                   (1.03,1.36)

                     1.0n=108
                     0.80(0.57, 1.13) n= 94
                     0.61(0.42, 0.89) n = 69
                     0.77(0.54, 1.12) n= 87
                   (0.80, 1.02)
 Grandjeanetal. 2012
 Faroe Islands
 Birth cohort, follow-up to age 7
 yrs
 n=587
 Age 5 pre-booster (e.g., tetanus,
 diphtheria) and 4 weeks after
 booster and age 7
 PFOS in 3rd trimester blood
 sample and in child (age 5)
 Linear regression, adjusted for
 sex, age, birth weight, maternal
 smoking, breastfeeding, and PCBs
 [and time since booster for post-
 booster analysis]
Geometric mean
Maternal sample
0.027 ug/mL

Child's sample
0.0167 ug/mL
                        Tetanus
                   -10.1 (-31.9, 18.7)
                    -2.3 (-28.6, 33.6)
                    35.3 (-3.9, 90.6)
Log PFOS and Log antibody Beta (95% CI) [% change in antibody
titer per 2-fold increase in PFOS]
                                             Diptheria
                                        -38.6 (-54.7,-16.9)
                                         -20.6 (-37.5, 0.9)
                                         -19.7 (-41.8, 10.7)

                                         -10.0 (-32.6, 20.0)

                                             Diptheria
                                         -16.0 (-34.9, 8.3)
                                         -15.5 (-31.5, 4.3)
                                         -27.6 (-45.8, -3.3)
Maternal PFOS
Pre-booster
Post-booster
Year 7
Year 7
(adjusted for age 5)    33.1 (1.5, 74.6)
                   Child's PFOS
                   Pre-booster
                   Post-booster
                   Year 7
                   Year 7
                   (adjusted for age 5)
                        Tetanus
                   -11.9 (-30.0, 10.9)
                   -28.5 (-45.5,-6.1)
                    -23.8 (-44.3, 4.2)
                                                                      -11.4 (-30.5, 12.8)     -20.6 (-38.2, 2.1)
                                                   Similar results seen with PFOA
Perfluorooctane sulfonate (PFOS) - May 2016
                                                                          3-38

-------
Reference and Study Details
Granum et al. 201 3
Norway
Birth cohort, Norwegian Mother
and Child Cohort Study
n = 56 with maternal blood at
delivery and child blood samples
at 3 yrs
Linear regression, considered
potential confounders
PFOS level
Mean 0.0056
Hg/mL







Results
Beta (95% CI ) (p- value), PFOS and antibody titer
Rubella -0.08 (-0.14, -0.02) (p = 0.007)
Measles -0.05 (-0.10, 0.01) (p = 0.09)
Tetanus -0.002 (-0.03, 0.02) (p = 0.87)
Hib -0.16 (-1.02, 0.70) (p = 0.71)
Similar results for other PFASs



General Population: Adults
Looker etal. 2014
C8 Health Project, West Virginia
2005-2005 enrollment and
baseline blood sample and
questionnaires; 2010 follow-up n
= 411 with pre- vaccination blood
sample - flu vaccination - 21 day
post vaccination blood sample
Linear regression: antibody titer
rise
Logistic regression:
seroconversion and seroprotection
Considered possible confounders,
retained in final model: age,
gender, mobility (# addresses),
and history of previous influenza
vaccination







Geometric mean
0.0083 ng/mL

Ql: 0.001-0.0058
Q2: 0.0059-0.0092
Q3: 0.0093-0.0145
Q4: 0.0147-0.0423

















Percentage positive) OR (95% CI), by influenza strain:
Seroconversion Seroproection
(4-fold increase (antibody titer 1 :40
in antibody titer) following vaccine)

Influenza B (62%) (66%)
PFOS continuous 1.17(0.63,2.17) 0.85(0.44,1.64)
Ql 1.0 (referent) 1.0 (referent)
Q2 0.72(0.39,1.33) 0.67(0.35,1.25)
Q3 0.81 (0.42, 1.53) 0.82 (0.42, 1.59)
Q4 0.87 (0.43, 1 .74) 0.73 (0.36, 1 .47)

A/HlNa (84%) (96%)
PFOS continuous 1.10(0.51,2.37) 0.93(0.23,3.71)
Ql 1.0 (referent) 1.0 (referent)
Q2 0.97(0.44,2.14) 0.55(0.13,2.37)
Q3 0.78 (0.35, 1.75) 1.81 (0.32, 10.22)
Q4 0.94(0.38,2.31) 1.26(0.24,6.61)
A/H3N2 (65%) (84%)
PFOS continuous 1.17(0.63,2.15) 0.63(0.26,1.49)
Ql 1.0 (referent) 1.0 (referent)
Q2 1.08(0.59,1.97) 0.85(0.38,1.88)
Q3 1.10(0.59,2.06) 1.09(0.47,2.56)
Q4 1.41 (0.72, 2.78) 0.56 (0.24, 1.28)
Asthma

   Humblet et al. (2014) evaluated a cohort from NHANES to investigate children's PFAS
serum levels, including PFOS, and their association with asthma-related outcomes. Sera were
analyzed for 12 PFAS with focus on PFOA, PFOS, PFHxS, and PFNA. A total of 1,877 children
12-19 years old with at least one serum sample available were included. Asthma and related
outcomes were self-reported. Median serum PFOS levels were 0.017 |ig/mL for those ever
having asthma and 0.0168 |ig/mL for children without asthma. In the multivariable adjusted
model, a doubling of PFOS level was inversely associated with the odds of ever having asthma
(OR = 0.88, 95% CI: 0.74-1.04), but statistical significance was not attained. PFOA was
significantly associated with asthma and no associations were found between the other PFASs
and outcome.

   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,
Perfluorooctane sulfonate (PFOS) - May 2016
3-39

-------
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 4
weeks. Associations of perfluorinated 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 0.0455 ± 0.0373 and 0.0334 ± 0.0264
|ig/mL, respectively; similar levels were measured for perfluorotetradecanoic acid with much
lower concentrations of the remaining six perfluorinated carboxylated and two sulfonates
sulfonates. The adjusted odds ratios for asthma association with the highest versus lowest
quartile levels were significantly elevated for seven of the PFAS 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
PFOS levels, as well as three other compounds, were significantly associated with higher asthma
severity scores.

   A summary of the studies that examined the relationship between PFOS serum levels and
markers of immunotoxicity in humans is presented in Table 3-7. A few studies have evaluated
associations with measures indicating immunosuppression. Two studies reported decreases in
response to one or more vaccines in children aged 3, 5, and 7 years (e.g., measured by antibody
titer) in relation to increasing maternal serum PFOS levels during pregnancy or at 5 years of age
(Grandjean et al. 2012; Granum et al. 2013). Decreased rubella and mumps antibody
concentrations in relation to serum PFOS concentration were found among 12-19 year old
children in the NHANES, particularly among seropositive children (Stein et al.  2015). A third
study of adults found no associations with antibody response to influenza vaccine (Looker et al.
2014).  In the three studies examining exposures in the background range among children (i.e.,
general population exposures, geometric means < 0.02 |ig/ml), the associations  with PFOS were
also seen with other correlated PFAS, complicating conclusions specifically for PFOS. No clear
associations were  reported between prenatal PFOS exposure and incidence of infectious disease
among children (Fei et al. 2010b; Okada et al. 2012), although an elevation in risk of
hospitalizations for an infectious disease was found among girls suggesting an effect at the
higher maternal serum levels  measured in the Danish population (mean maternal plasma levels
were 0.0353 |ig/mL).

   With regard to other immune dysfunction, serum PFOS levels were not associated with risk
of ever having had asthma among children in the NHANES with median levels  of 0.017 |ig/mL
(Humblet et al. 2014).  A study among children in Taiwan with higher serum PFOS
concentrations (median with and without asthma 0.0339 and 0.0289  jig/mL, respectively) found
higher odds ratios for physician-diagnosed asthma with increasing serum PFOS quartile (Dong et
al. 2013). Associations also were found for other PFASs. Among asthmatics, serum PFOS was
also associated with higher severity scores, serum total IgE, absolute eosinophil counts and
eosinophilic cationic protein levels.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-40

-------
3.1.1.7  Other Effects

   Metabolic syndrome is a combination of medical disorders and risk factors that increase the
risk of developing cardiovascular disease and diabetes. Lin et al. (2009) investigated the
association between serum PFOS (plus three other PFASs) and glucose homeostasis and
metabolic syndrome in adolescents (12-20 years) and adults (> 20 years) by analyzing the 1999-
2000 and 2003-2004 NHANES data. The National Cholesterol Education Program Adult
Treatment Panel III guidelines were used to define adult metabolic syndrome and the modified
guidelines were used to define adolescent metabolic syndrome. The study population included
1,443 subjects (474 adolescents, 969 adults) at least 12 years of age who had a morning
examination and triglyceride measurement. There were 266 male and 208 female adolescents
and 475 male and 493  female adults. Multiple linear regression and logistic regression models
were used to analyze the data. Covariates included age, sex, race, smoking status, alcohol
consumption, and household income. Log-transformed PFOS concentration was 3.11 ng/mL and
3.19 ng/mL for adolescents and adults, respectively. In adults, serum PFOS concentration was
associated with increased p-cell function (P coefficient 0.15, p < 0.01). Serum PFOS
concentration was not  associated with metabolic syndrome, glucose concentration, homeostasis
model of insulin resistance, or insulin levels in adults or adolescents.

3.1.1.8  Summary and conclusions from the human epidemiology studies

   Numerous epidemiology studies have been conducted evaluating occupational PFOS
exposure and environmental PFOS exposure including a large community highly-exposed to
PFOA (the C8 Health Project) and background exposures in the general population in several
countries. Occupational and general populations have evaluated the association of PFOS
exposure to a variety of health endpoints. Health outcomes assessed include blood lipid and
clinical chemistry profiles, thyroid effects, immune function, reproductive effects, pregnancy-
related outcomes, fetal growth and developmental outcomes, and cancer.

Serum Lipids. Multiple epidemiologic studies have evaluated serum lipid status in association
with PFOS concentration (Table 3-1).  These studies provide support for an association between
PFOS and small increases in total cholesterol. Hypercholesterolemia, which  is clinically defined
as cholesterol > 240 mg/dL, was associated with PFOS exposure in a Canadian cohort (Fisher et
al. 2013) and in the C8 cohort (Steenland et al. 2009). Cross-sectional occupational studies
demonstrated an association between PFOS and total cholesterol (Olsen et al. 2001a,  2001b,
2003b). Evidence for associations between other serum lipids and PFOS is mixed, including
HDL cholesterol, LDL, VLDL, and non-HDL cholesterol, as well as triglycerides. The studies on
serum lipids in association with PFOS serum concentrations are largely cross-sectional in nature
and were largely conducted in adults, but some studies exist on children and pregnant females.
The location of these cohorts varied from the U.S. population including NHANES volunteers, to
the Avon cohort in the UK, to Scandinavian countries. Limitations to these studies include the
frequently high correlation between PFOA and PFOS exposure; not all studies control for PFOA
in study design. Also studied were populations with known elevated exposure to other
environmental chemicals including PFOA in the C8 population and PBDEs and other persistent
organic chemicals in the Inuit population.

Liver. Cross-sectional and longitudinal studies evaluated PFOS and liver enzymes in adults. Lin
et al. (2010) looked at  data from the NHANES, which is representative of the U.S. national
population, and Gallo et al. (2012) reported an analysis of data from the C8 Health Project,
reflective of a highly-exposed community. Both  studies saw a slight positive association between

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-41

-------
serum PFOS levels and increased serum ALT values. The association between PFOS levels and
increased serum GOT levels was less defined and overall did not appear to be affected. Total or
direct bilirubin showed no association with PFOS in either study. In the Gallo et al. (2012) study,
the cross-sectional design and self-reported lifestyle characteristics are limitations to the study,
and while both studies showed a trend, it was not large in magnitude.

Kidney. Shankar et al. (2011) and Watkins et al. (2013) analyzed sub-sets or the entire
population for an association between PFOS  serum levels and either kidney disease or
biomarkers that may be associated with kidney function. Shankar et al. (2011) used NHANES
data and showed a positive association between increasing levels of PFOS and chronic kidney
disease, as defined as an eGFR of < 60 mL/min/1.73 m2. The odds ratio for chronic kidney
disease at > 0.030 |ig/mL of PFOS was 1.82 (95% CI: 1.01-3.27), and while the possibility of
reverse causality could not be excluded, the association between PFOS and eGFR when
examined in those with and without chronic kidney disease supports an effect. Watkins et al.
(2013) evaluated C8 Health Project children to look at PFOS levels and kidney function in
children, as defined as decreased eGFR, and found a dose-related trend: the decrease was
1.10 mL/min/1.73 m2 (95% CI: -1.66 to -0.53). Geiger et al. (2014b) found no association in
children between serum PFOS levels and hypertension. Steenland et al. (2010) evaluated C8
Health Project adults and found a positive association between PFOS  serum levels and an
increase in uric acid with odds ratios increasing from 1.02 to 1.26 with each decile. Overall,
studies do suggest an association between chronic kidney disease, as defined by estimated
glomerular filtration rate; however, reverse causality cannot be excluded.

Fertility, Pregnancy, and Birth Outcomes. Fetal growth retardation was examined through
measures including mean birth weight, LEW, and small for gestational age. Mean birth weight
examined as a continuous outcome was the most commonly examined endpoint for
epidemiology studies of serum/cord PFOS exposures. Although three studies were null (Fei et al.
2008b; Hamm et al. 2010; Monroy et al. 2008), birth weight deficits ranging 29-149 grams were
detected in five studies (Apelberg et al. 2007; Chen et al. 2015; Darrow et al. 2013; Maisonet et
al. 2012; Washino et al. 2009). Larger reductions (69-149 grams) were noted in three of these
studies (Apelberg et al. 2007; Chen et al. 2015; Washino et al. 2009) based on per unit increases
in serum/cord PFOS exposures, while the lone categorical data showed an exposure-response
deficit in mean birth weight up to 140 grams  across the PFOS tertiles  (Maisonet et al. 2012).
Two (Chen et al. 2015; Whitworth et al. 2012) out of four (Fei et al. 2007; Hamm et al. 2009)
studies of SGA and serum/cord PFOS exposures showed some suggestion of increased ORs
(range: 1.3-2.3), while three (Chen et al. 2012; Fei et al. 2007; Stein et al. 2009) out of four
(Darrow et al. 2014) studies of LEW showed increased risks (OR range: 1.5-4.8). Although a
few of these studies showed some suggestion of dose-response relationships across different fetal
growth measures (Fei et al. 2007; Maisonet et al. 2012; Stein et al. 2009), study limitations,
including the potential for exposure misclassification, likely precluded the ability to adequately
examine the exposure-response pattern.

   Recent data also indicate an association between low maternal GFR and infant birth weight,
supporting GFR as  a confounder in epidemiology studies (Morken et al. 2014; Verner et al.
2015). In such cases the increased serum PFOS could be the result of the developmental
milestone rather than a cause. However, while a proportion of the association between prenatal
PFOS and birth weight may be confounded by low maternal GFR, a direct effect of PFOS on
neonatal weight cannot be entirely dismissed based on the available data.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-42

-------
   A small set of studies observed an association with gestational diabetes (Zhang et al. 2015,
preconception serum PFOS), pre-eclampsia (Stein et al. 2009) and pregnancy-induced
hypertension (Darrow et al. 2013) in populations with serum PFOS concentrations of 0.012-
0.017 ug/mL. Zhang et al. (2015) and Darrow et al.  (2013) used a prospective assessment of
adverse pregnancy outcomes in relation to PFASs which addresses some of the limitations in the
available cross-sectional studies. Associations with these outcomes and serum PFOA also were
observed.

   Although there was  some suggestion of an association between PFOS exposures and semen
quality parameters in a few studies (Joensen et al. 2009; Toft et al. 2012), most studies were
largely null (Buck Louis et al. 2015; Ding et al. 2013; Joensen et al. 2013; Vested et al. 2013;
Raymer et al. 2012;  Specht et al. 2012; Vested et al. 2013). For example, morphologically
abnormal sperm associated with PFOS were detected in three (Buck Louis et al. 2015; Joensen et
al. 2009;  Toft et al. 2012) out of eight (Buck Louis et al. 2015; Ding et al. 2013; Joensen et al.
2013; Raymer et al.  2012; Specht et al. 2012; Vested et al. 2013) studies.

   Small increased  odds of infertility was found for PFOS exposures in studies by J0rgensen et
al. (2014) [OR= 1.39; 95% CI: 0.93-2.07] and Velez etal. (2015) [OR = 1.14; 95% CI: 0.98-
1.34]. Although one study was null (Vestergaard et al. 2012), PFOS exposures associated with
decreased FRs, indicative of longer time to pregnancy, were noted in studies by Fei et al. (2009)
[FR = 0.74 (95% CI: 0.58-0.93) and in studies by J0rgensen et al. (2014) [FR = 0.90; 95% CI:
0.76-1.07]. Whitworth et al. (2012) data suggested that reverse causality may explain their
observation of subfecundity odds of 2.1 (95% CI: 1.2-3.8) for the highest PFOS quartile among
parous females, but a reduced odds among nulliparous females (OR = 0.7; 95%  CI: 0.4-1.3). A
recent analysis of the pooled DNBC study samples did not find strong evidence of differences by
parity status with an overall fecundability ratio of 0.83 (95% CI: 0.72-0.97) for  PFOS exposures,
as well as comparable ratios for parous (0.86; 95% CI: 0.70-1.06) and nulliparous (0.78; 95% :
0.63-0.97) females (Bach et al. 2015). The same authors reported an increased infertility OR of
1.75  (95% CI: 1.21-2.53) and OR for parous (OR =1.51; 95% CI: 0.86-2.65) and nulliparous
(OR= 1.83; 95% CI: 1.10-3.04) females. Although there remains some concern over the
possibility of reverse causation explaining some previous study results, these collective findings
indicate a consistent association with fertility and fecundity measures and PFOS exposures.

Thyroid. Numerous epidemiologic studies have evaluated thyroid hormone levels and/or thyroid
disease in association with serum PFOS concentrations. These epidemiologic studies provide
limited support for an association between PFOS exposure and incidence or prevalence of
thyroid disease, and  include large studies of representative samples of the general U.S. adult
population (Melzer et al. 2010; Wen et al. 2013). These highly powered studies  reported
associations between PFOS exposure (serum PFOS  concentrations) and thyroid disease but not
thyroid hormone status. Melzer et al. (2010) studied thyroid disease with medication and Wen et
al. (2013) studied subclinical thyroid disease. Thyroid function can be affected by iodide
sufficiency and by autoimmune disease. People testing positive for the anti-TPO biomarker
showed associations with PFOS and TSH or T4 (Webster et al. 2014); this association was
stronger in people with both low iodide status and positive anti-TPO antibodies  (Webster et al.
2015). These studies used anti-TPO antibody levels as an indication of stress to  the thyroid
system, not a disease state. Thus, the association between PFOS and altered thyroid hormone
levels is stronger in people at risk for thyroid insufficiency. In people without diagnosed thyroid
disease or without biomarkers of thyroid disease, thyroid hormones (TSH, T3, or T4) show
mixed effects across cohorts.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-43

-------
Immune Function. A few studies have evaluated associations with measures indicating
immunosuppression. Two studies reported decreases in response to one or more vaccines in
children aged 3, 5, and 7 years (e.g., measured by antibody titer) in relation to increasing prenatal
serum PFOS levels or at 5 years of age (Grandjean et al. 2012; Granum et al. 2013). Decreased
rubella and mumps antibody concentrations in relation to serum PFOS concentration were found
among 12-19 year old children in the NHANES, particularly among seropositive children (Stein
et al. 2015). A third study of adults found no associations with antibody response to influenza
vaccine (Looker et al. 2014). In the three studies examining exposures in the background range
among children (i.e., general population exposures, geometric means < 0.02 jig/ml), the
associations with PFOS were also seen with other correlated PFASs, complicating conclusions
specifically for PFOS.

   No clear associations were reported between prenatal PFOS exposure and incidence of
infectious disease among children (Fei et al. 2010b; Okada et al.  2012), although an elevation in
risk of hospitalizations for an infectious disease was found among girls suggesting an effect at
the higher maternal serum levels measured in the Danish population (mean maternal plasma
levels were 0.0353 |ig/mL). With regard to other immune dysfunction, serum PFOS levels were
not associated with risk of ever having had asthma among children in the NHANES with median
levels of 0.017 |ig/mL (Humblet et al. 2014).  A study among children in Taiwan with higher
serum PFOS concentrations (median with and without asthma 0.0339 |ig/mL and 0.0289 jig/mL,
respectively) found higher odds ratios for physician-diagnosed asthma with increasing serum
PFOS quartile (Dong et al. 2013). Associations also were found for other PFASs. Among
asthmatics, serum PFOS was also associated with higher severity scores, serum total IgE,
absolute eosinophil counts and eosinophilic cationic protein levels.

3.1.2   Carcinogenicity Studies

Occupational Exposure. Several analyses of various health outcomes have occurred on cohorts
of workers at  the 3M Decatur, Alabama plant (Alexander et al. 2003; Alexander and Olsen 2007;
Mandel and Johnson 1995). Cause-specific mortality was examined in a cohort of 2,083 workers
employed for at least 1 year among workers grouped into three PFOS exposure categories: non-
exposed, low  exposed, and high exposed.  Exposure classifications were determined using PFOS
serum concentrations measured in a  subset of workers linked to specific jobs and work histories.
Cumulative exposures were also estimated by applying a weight to each of the exposure
categories and multiplying by the number of years of employment for that job for each
individual. The geometric mean serum PFOS levels were 0.941 |ig/mL for chemical plant
employees and 0.136 |ig/mL for non-exposed workers. Results of these studies are summarized
in Table 3-8.

   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 included cancer of the esophagus, liver, breast,
urinary organs, bladder, and skin. However, the number of cases was very small (1-3), resulting
in wide confidence intervals. The SMRs for these causes (except breast cancer) were also
elevated when the cohort was limited to the 65 employees ever employed in a high exposure job.
The SMR for bladder cancer was 4.81 (95% CI: 0.99-14.06). Three male  employees in the
cohort died of bladder cancer (0.62 expected). All were employed at the Decatur plant for
> 20 years and had worked in high exposure jobs for at least 5 years. The  SMR for bladder

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-44

-------
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 three deaths from bladder cancer were greater than the expected number observed
in the general population, the small number of deaths (especially for females in all categories)
precludes a definitive conclusion regarding an association with PFOS exposure. In addition, six
death certificates were not obtained, and smoking status was not known for the cohort increasing
the uncertainty with regard to the estimated risk.

   Based on these results, another study of this cohort was conducted to evaluate 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 six incident cases. Only two of the six self-
reported cases were confirmed with medical records. Five of the six incident cases had a history
of cigarette smoking. Standardized incidence ratios (SIR) were estimated for the three exposure
categories described for the mortality study and compared to U.S. cancer rates. SIRs were 0.61,
2.26, and 1.74 for the nonexposed, ever low, and ever high exposure categories, respectively.
Rate ratios by cumulative exposure index were increased in the higher categories (5 to < 10 and
> 10) when using either the U.S. population rates or an internal referent population,  however the
number of cases were few and confidence intervals were wide including the null. These results,
while suggestive of an elevated risk of bladder cancer, were not conclusive.

   Grice et al. (2007) evaluated associations between PFOS exposure at the 3M Decatur,
Alabama plant and various malignant or benign tumors reported by the same study group
evaluated by Alexander and Olsen (2007). Current and past employees at the plant answered
questionnaires (n = 1,400; 1,137 male and 263 female) about diagnosis of cancers or non-
cancerous conditions. Data were analyzed by PFOS exposure category: unexposed
(< 0.29 |ig/mL), low (0.39-0.89 jig/mL),  or high exposure (1.30-1.97 |ig/mL) and by categories
of estimated cumulative exposure using the same weighted approach described in the previous
studies of this cohort. 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.
Length of follow-up may not have been adequate to detect cancer incidence in this cohort as
approximately one-third of the participants had worked < 5 years in their jobs, and only 41.7%
were employed > 20 years.

C8 Health Project Community. Members of the C8 Health Project,  47,151 cancer-free adults
and 203 cases, were evaluated for an association between serum PFOS levels and incidence of
colorectal cancers (Innes et al. 2014). This cross-sectional study compared serum PFOS (and
PFOA) levels at enrollment with diagnosis of primary colorectal cancer. Serum PFOS levels
ranged from < 0.0005 to 0.759 jig/mL, with an average of 0.0234 |ig/mL. A concentration-
related inverse relationship was found between PFOS level and diagnosis of colorectal cancer
with OR = 0.24 (95% CI 0.16, 0.37; highest to lowest quartile, p for trend < 0.00001). An inverse
association was also found between PFOA and colorectal cancer.

General Population. A subset of females enrolled in the DNBC was evaluated for an
association between plasma PFOS levels  (as well as 15 other perfluoroalkylated substances)
measured during pregnancy and risk of breast cancer during a follow-up period of 10-15 years

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-45

-------
(Bonefeld-J0rgensen et al. 2014). A total of 250 females diagnosed with breast cancer were
matched for age and parity with 233 controls. The mean PFOS level in the controls was
0.0306 |ig/mL while levels in the cases were divided into quintiles ranging from < 0.0204 up to
> 0.0391  |ig/mL. No association was found between PFOS levels and breast cancer risk in
logistic regression models adjusted for age at blood draw, BMI before pregnancy, gravidity, use
of oral contraceptives, age at menarche, smoking, alcohol consumption, maternal education and
physical activity. A weak positive Relative Risk (1.04; 95% CI: 0.99-1.08) was found only with
perfluorooctane- sulfonami de.

   These same researchers had previously observed a borderline significant positive association
with PFOS levels and breast cancer (adjusted OR = 1.03, 95% CI: 1.001-1.07) in a small cohort
from Greenland (Bonefeld-J0rgensen et al. 2011). Logistic regression models were adjusted for
age, BMI, total number of full-term pregnancies, breastfeeding, menopausal status, and serum
cotinine, but the unadjusted results that included the entire  study group were not different.
Median serum PFOS levels were 0.0456 |ig/mL (range: 0.0116-0.124 |ig/mL) among 31  breast
cancer patients and 0.0219 |ig/mL (range: 0.0015-0.172 jig/mL) among 98 controls. A weak
positive odds  ratio of 1.03 (95% CI: 1.00-1.05) was also found for the sum of
perfluorosulfonated compounds which included PFOS along with perfluorohexane sulfonate and
perfluorooctane sulfonamide.

   Eriksen et al. (2009) examined the association between plasma PFOS concentration and the
risk of cancer in the general Danish population. The study population was chosen from
individuals (50-65 years of age) who had enrolled in the prospective Danish cohort Diet, Cancer,
and Health study between December 1, 1993 and May 31,  1997. The Danish Cancer Registry
and Danish Pathology Data Bank were used to identify cancer patients diagnosed between
December 1, 1993, and July 1, 2006. The cancer patients (n = 1,240) consisted of 1,111 males
and 129 females whose median age was 59 years having prostate cancer (n = 713), bladder
cancer (n = 332), pancreatic cancer (n = 128), and liver cancer (n = 67). The individuals
(n = 772) in the subcohort comparison group were randomly chosen from the cohort study and
consisted of 680 males and 92 females whose median age was 56 years.  The participants
answered a questionnaire upon enrollment in the cohort study, and data on known  confounders
were obtained from the questionnaires. The plasma PFOS concentrations, based on blood
samples provided at enrollment (1993-1997) for cancer patients were as follows: males
0.0351 |ig/mL, females 0.0321 jig/mL, prostate cancer 0.0368 |ig/mL, bladder cancer
0.0323 |ig/mL, pancreatic cancer 0.0327 |ig/mL, and liver cancer 0.0310 |ig/mL. The plasma
PFOS concentrations for the subcohort comparison group were 0.0350, 0.0293, and
0.0343 |ig/mL for the males, females, and combined, respectively. Incidence rate ratios, crude
and adjusted for confounders, did not indicate an association between plasma PFOS
concentration and bladder, pancreatic, or liver cancer in models adjusting for potential
confounders. For prostate cancer, increased odds ratios 30% above the comparison group for
quartiles 2 through 4 were observed, but there was no increasing trend in the analysis using
PFOS concentration as a continuous variable. The plasma PFOS levels in the population were
lower than those observed in occupational cohorts.

   Hardell et al.  (2014) investigated an association between prostate cancer and levels of PFAS
in whole blood. Patients with newly diagnosed prostate cancer (n = 201) had median PFOS
levels of 0.009 |ig/mL, while the control group (n = 186) had a median level of 0.0083 |ig/mL.
PFOS levels, which were measured 1-3 years after cancer diagnosis, were not associated with
higher risks of prostate cancer in logistic regression models adjusted for age, BMI, and year of
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-46

-------
 blood sampling, or when analyzed according to Gleason score (pathology grade) and prostate-
 specific antigen. A significantly higher risk for prostate cancer was found for a group with PFOS
 levels above the median and a first-degree relative with prostate cancer indicating a potential
 genetic risk factor.

     A small study found no differences in blood PFOS levels between cancer and non-cancer
 patients; the types of cancer in the patients were not defined. Vassiliadou et al.  (2010) found
 median  serum PFOS concentrations among 40 cancer patients (0.0113 |ig/mL,  males;
 0.008 |ig/mL, females) were similar to two control groups (0.0105 and 0.0137 jig/mL, males;
 0.007 and 0.0085  |ig/mL, females).

     Results of the cancer epidemiology studies in the highly exposed and general populations are
 summarized in Table 3-8.
                 Table 3-8. Summary of PFOS Epidemiology Studies of Cancer
   Reference and Study Details
      Analysis Group
          Relative Risk Estimates
                                     Occupational Exposure Studies
Alexander et al. 2003
Fluorochemical production, Decatur,
Alabama
Film plant and chemical plant
employees (current, retired and former),
n = 2,083, follow-up through 1998
83% male, median age 50.9 yrs at
follow-up, median 13.2 yrs of
employment
Mortality
Comparisons by exposure group
classified using matrix of work history
(1961-1997) and job-specific serum
PFOS concentration: No exposure, low
and high potential workplace exposure;
Cumulative exposure level based on
exposure category weight (1,3, or 10)
and years spent in specific jobs
Mortality through 1998

All (Alabama referent)

Non-exposed jobs
(0.11-0.29 ug/ml)
Low exposure jobs
(0.39-0.89 ug/ml)
High exposure jobs
(1.30-1.97 ug/ml)

All (Alabama referent)

Non-exposed jobs
(0.11-0.29 ug/ml)
Low exposure jobs
(0.39-0.89 ug/ml)
High exposure jobs
(1.30-1.97 ug/ml)
Liver Cancer
SMR (95% CI)
1.61(0.20, 5.82) (n = 2)

No cases

3.94 (0.10,21.88) (n=  1)

2.00(0.05,11.01)(n=  1)

Bladder Cancer
SMR (95% CI)
4.81(0.99,14.06)(n= 3)

No cases

No cases

12.77(2.63, 37.35) (n= 3)
Alexander and Olsen 2007; Grice et al.
2007)
Fluorochemical production, Decatur,
Alabama
Film plant and chemical plant
employees, n = 1,400 of 2,083 who
completed questionnaire in 2002 and
188 decedents since mortality analysis.
495 declined; participation 73.9% of
eligible, 43,739 person-years of follow-
up. 81.2% male, Incidence (via
questionnaire) with confirmation by
physician for some
Incidence through 2002

All (U.S. population referent)
Non-exposed jobs
(0.11-0.29 ug/ml)
Low exposure jobs
(0.39-0.89 ug/ml)
High exposure jobs
(1.30-1.97 ug/ml)
Ever low or high
Low or high (> 1 yr)
Non-exposed jobs
Ever low or high
Low or high (> 1 yr)
High(>lyr)
Bladder cancer (2 of 6 reported confirmed; 5 deaths)
SIR (95% CI) (n cases)
1.28 (0.64,2.29) (n = 11)
0.61 (0.07,2.19) (n = 2)

2.26 (0.91,4.67) (n = 7)

1.74(0.64, 3.79) (n = 6)

1.7(0.77, 3.22) (n = 9)
1.31 (0.48,2.85) (n = 6)
Colon cancer (12 of 22 reported confirmed)
OR, 95% CI, (n cases)
1.0 (n= 8)
1.21 (0.51,2.87) (n= 15)
1.37 (0.57, 3.30) (n= 14)
1.69 (0.68,4.17) (n = 7)
 Perfluorooctane sulfonate (PFOS) - May 2016
                                                                 3-47

-------
   Reference and Study Details
      Analysis Group
           Relative Risk Estimates
                                    Non-exposed jobs
                                    Ever low or high
                                    Low or high (> 1 yr)
                                    High (> 1 yr)
                              Prostate cancer (22 of 29 reported confirmed)
                              OR (95% CI) (n cases)
                              1.0 (n= 10)
                              1.34 (0.62,2.91) (n= 19)
                              1.36 (0.61, 3.02) (n= 16)
                              1.08 (0.44,2.69) (n = 9)
                                           General Population Studies
Bonefeld-Jergensen et al. 2014
Denmark; case-control study nested in
prospective cohort; DNBC,  1996-2002,
follow-up to 2010. 250 women with
breast cancer identified using cancer
registry (mean age at blood draw 30.4
yr) and 233 controls (mean age at blood
draw 29.6 yr), frequency matched on
age and parity, selected at random from
cohort at baseline. PFOS (and other
perfluorochemicals) in blood drawn
between gestation weeks 6 and 14.
Mean serum PFOS in controls
0.031 ug/mL; correlation PFOS
and PFOA 0.69

Continuous PFOS
Quintiles
<0.02
0.02-0.025
0.025-0.030
0.030-0.039
> 0.039
Breast Cancer
Adjusted RR (95% CI) (n cases)
0.99(0.98,1.01) (n = 221)

1.0(n = 42)
1.51 (0.081, 2.71) (n= 52)
1.51 (0.82,2.84) (n = 49)
1.13 (0.59,2.04) (n = 43)
0.90(0.47,1.7) (n= 35)
Hardell etal. 2014
Denmark; case-control study
Prostate cancer cases from hospital
admissions, 2007-2011 (n = 201,
participation 79%, median age 67 yr);
population-based controls matched on
age geographical location (n = 186,
participation 54%); Blood sampling for
perfluorinated alkyl acids 2007-2011
Median blood PFOS in cases
0.009 ug/mL, controls 0.0083
ug/mL
Prostate Cancer
Adjusted RR (95% CI) (n cases)
1.0(0.60, 1.5) (n= 109)
Bonefeld-Jergensen et al. 2011
Greenland, case-control study
Inuit women with breast cancer
registered at hospital (n = 31, 80% of
all cases) in 2000-2003 (median age 50
yr). Age and district-matched
(frequency) controls selected from
cross-sectional biomonitoring study (n
= 115, median age 54 yr)
Median serum PFOS (range)
Cases: 0.0456 ug/mL
(0.0116-0.124)
Controls: 0.0219 ug/mL
(0.0015-0.172)
Breast Cancer
OR (95% CI), p-value, (n cases/n controls)
Unadjusted 1.01 (1.003,1.02), p = 0.02, (98 cases/31
controls)
Adjusted 1.03 (1.001, 1.07), p = 0.05, (69 cases/9
controls)
Eriksen et al. 2009
Denmark Diet, Cancer and Health
Study; enrolled December 1, 1993-
May 31,1997; cancer diagnoses
between December 1,1993-July 1,
2006. 1,240 cancer cases (1,111 male,
129 female), median age 59 years
compared to 772 participants selected
at random from cohort, median age 56
years. Analysis using Cox proportional
hazards model stratified by sex (IRR)
Plasma PFOS concentrations at
enrollment; range: 0.001-0.131
ug/mL.
Quartiles PFOS


Ql
Q2
Q3
Q4
Trend per 10 ng/mL increase


Ql
Q2
Q3
Q4
Trend per 10 ng/mL increase


Ql
Q2
Q3
Q4
Trend per 10 ng/mL increase
IRR (95% CI)
Bladder Cancer (n = 332)
1.0
0.76(0.50,1.16)
0.93(0.61,1.41)
0.70(0.446,1.07)
0.93(0.83,1.03)
Liver Cancer (67)
1.0
0.62(0.29,1.33)
0.72(0.33,1.56)
0.59(0.27,1.27)
0.97(0.79,1.19)
Prostate Cancer (n = 713)
1.0
1.35(0.97,1.87)
1.31(0.94,1.82)
1.38(0.99,1.93)
1.05(0.97,1.14)
 Perfluorooctane sulfonate (PFOS) - May 2016
                                                                        3-48

-------
3.1.2.1  Summary and Conclusions from the Human Cancer Epidemiology Studies

   A small number of epidemiology studies of PFOS exposure and cancer risk are available.
While these studies do report elevated risk of bladder and prostate cancers, limitations in design
and analysis preclude the ability to make definitive conclusions. While an elevated risk of
bladder cancer mortality was associated with PFOS exposure in an occupational study
(Alexander et al. 2003), a subsequent study to ascertain cancer incidence in the cohort observed
elevated but statistically insignificant incidence ratios that were 1.7- to 2-fold higher among
workers with higher cumulative exposure (Alexander and Olsen 2007). The risk estimates lacked
precision because the number of cases was small. Smoking prevalence was higher in the bladder
cancer cases, but the analysis did not control for smoking because data were missing for
deceased workers, and therefore positive confounding by smoking is a possibility. Mean PFOS
serum levels were 0.941 jig/mL. No elevated bladder cancer risk was observed in a nested case-
control study in a Danish cohort with plasma PFOS concentrations at enrollment of 0.001-
0.1305 |ig/mL (Eriksen et al. 2009).

   Elevated odds ratios for prostate cancer were reported for the occupational cohort examined
by Alexander and Olsen (2007) and the Danish population-based cohort examined by Eriksen et
al. (2009). However, the confidence intervals included the null, and no association was reported
by another case-control study in Denmark (Hardell et al. 2014). A case-control study of breast
cancer among Inuit females in Greenland with similar serum PFOS levels to those of the Danish
population (0.0015-0.172 |ig/mL) reported an association of low magnitude that could not be
separated from other perfluorsulfonated acids, and the association was not confirmed in a Danish
population (Bonefeld-J0rgensen et al. 2011, 2014). Some studies evaluated associations with
serum PFOS concentration at the time of cancer diagnosis, and the impact of this potential
exposure misclassification on the estimated risks is unknown (Bonefeld-J0rgensen et al. 2011;
Hardell et al. 2014). No associations were adjusted for other perfluorinated chemicals in serum in
any of the occupational and population-based studies.

3.2   Animal Studies

   Acute and short-term studies in rats and mice provide data on lethality, systemic toxicity,
neurotoxicity, and mode of action. Subchronic studies in monkeys and rats found decreased body
weight, increased liver weight accompanied by microscopic lesions, and decreased serum
cholesterol. The most prominent microscopic lesion of the liver in both monkeys and rats was
centrilobular hepatocellular hypertrophy. In a chronic bioassay, rats had decreased body weight,
increased liver weight with microscopic lesions, and an increased incidence of hepatocellular
adenomas. Effects on development and reproduction were found in both rats and mice, including
increased neonatal  mortality. Other developmental and reproductive toxicity effects included
decreased gestation length, lower birth weight, and developmental delays. Postnatal effects of
gestational and lactational exposure included evidence of developmental neurotoxicity, changes
in thyroid and reproductive hormones,  altered lipid and glucose metabolism, and decreased
immune function. Each of these studies is described in detail below, and a tabular summary of
the animal studies is provided in Appendix C, Table C-2.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-49

-------
3.2.1   Acute Toxicity

    The few available acute toxicity studies of PFOS indicate a lethal dose for 50% (statistical
median) of animals (LDso)of 251 mg/kg and an LCso of 5.2 ppm in rats (Dean et al. 1978; Rusch
et al. 1979). PFOS caused no irritation in a dermal irritation study although limited study details
were available (OECD 2002). An eye irritation study was also conducted but few details were
provided on effects observed (OECD 2002).

Oral Exposure

    Dean et al. (1978) exposed 5 CD rats/sex/dose by gavage to a single dose of 0, 100, 215, 464,
or 1,000 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 1,000 mg/kg group, and 3 of 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 findings, the acute
oral LDso was 233 mg/kg in males, 271 mg/kg in females, and 251 mg/kg combined.

    Male Wistar rats and male ICR mice (n = 2-3 per 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 of both species 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. One mouse in each dose group died. No neurological
signs were observed. 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 did not show any differences in the levels of catecholamines
(norepinephrine, dopamine, and serotonin) or amino acids (glutamic acid, glycine, and gamma-
aminobutyric acid [GABA]) when compared to the controls at 24 and 48 hours post-exposure.

Inhalation Exposure

    Rusch et al. (1979) exposed Sprague-Dawley rats (5/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 in the 24.09 mg/L group died 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 LCso was 5.2 mg/L (ppm).

Dermal/Ocular Exposure

    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 and covered. Erythema and edema were scored after 24 and 72 hours. The primary

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-50

-------
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. The
raw data were not provided in the OECD  (2002) report.

3.2.2   Short-Term Studies

   Short-term oral toxicity studies in rats and mice included data on lethality, body weight, liver
weight, and histopathology, as well as serum lipids. Body weight was decreased and liver weight
increased at > 2 mg/kg/day in rats. Higher doses resulted in hepatocyte hypertrophy and
decreased cholesterol in rats and mice.  Mechanistic studies in mice indicate changes suggestive
of hepatic hyperlipidemia or fatty liver disease.

Oral Exposure

Rat. Forty to seventy Sprague-Dawley  Crl:CD (SD) IGS BR rats/sex/dose were administered
PFOS in the diet at concentrations of 0, 0.5, 2.0, 5.0, or 20 ppm as part of a long term  chronic
cancer bioassay (Seacat et al. 2003).  Five animals per dose group were sacrificed for interim
necropsies at 4 weeks. 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. Animals were observed twice daily for mortality
and moribundity, 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/dose during week 4 for clinical chemistry, hematology, and
urinalysis evaluation. A thorough necropsy  was performed on five animals/ sex/dose at the end
of 4 weeks of treatment and liver samples were collected for palmitoyl CoA oxidase (PCoAO)
activity, liver weight, 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 is provided in Table 3-9. 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 liver weight was increased in the
high dose males and females; the increase was significant only for males. Food consumption and
food efficiency were decreased only  in the 20 ppm females.  No treatment-related effects were
observed on hematology or urinalysis; male rats treated with 20 ppm had significant decreases in
serum glucose. Analysis of PCoAO activity was weakly increased (< 2-fold) when compared to
controls in the 20 ppm dose group males in  one laboratory and similar to controls in another
laboratory analysis. The 20 ppm (1.5 mg/kg/day) dose group was a LOAEL for males following
a 4 week exposure.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-51

-------
     Table 3-9. 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±7a
120 ±37
6.0 ±1.1
0.22
202 ±15
3.8 ±0.2
0.059 ±0.013
113 ±18
101 ±12
3.0 ±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
Source: Data from Seacat et al. 2003
Notes: a 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

    Curran et al. (2008) conducted two 28-day studies in groups of 15 Sprague-Dawley
rats/sex/dose. 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, or 6.34 mg PFOS/kg body weight/day,
respectively,  in males and 0, 0.15, 1.43, 3.73, or 7.58 mg/kg body weight/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 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 2.2, 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 week 3 of treatment in the 20 mg PFOS/kg diet
group. No differences in blood pressure measurements were observed across the groups.
Deformability index values in red blood cells over a range of shear stress levels were
significantly lower in both males and females exposed to 100 mg PFOS/kg diet, relative to
controls.

    Absolute  and  relative liver weights were statistically-significantly increased in the male and
female rats at > 20 mg PFOS/kg diet. Relative 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 > 50 mg PFOS/kg diet.
Perfluorooctane sulfonate (PFOS) - May 2016
3-52

-------
   Both males and females showed a significant increase in expression of the gene for peroxisomal
acyl-coenzyme A oxidase at concentrations > 50 mg PFOS/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. A total of 67 fatty acid profiles were examined. The authors stated that
the profile changes were similar to those induced by weak peroxisome proliferators.

   At the high doses,  the serum levels of conjugated bilirubin and total bilirubin were increased
significantly. Serum cholesterol was significantly decreased for males and females at > 50 mg
PFOS/kg diet. Serum 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 as observed in this study are provided in Table 3-10.

    Table 3-10. Mean (± SD) Values for Select Parameters in Rats Treated for 28 Days
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
Source: Data from Tables 2-3 and 6-7 in Curran et al. 2008
Note: Statistically-significant from controls, p < 0.05 or p < 0.05

    The LOAEL was the 20 mg/kg dietary level (males: 1.33 mg PFOS/kg/day; females: 1.43 mg
PFOS/kg/day) for a significant increase in absolute (females) and relative (males and females)
liver weights and significant decrease in serum T4 (males and females). The NOAEL was the 2
mg/kg diet level (0.14-0.15 mg PFOS/kg/day).
Perfluorooctane sulfonate (PFOS) - May 2016
3-53

-------
    Ten three-month old male Sprague-Dawley rats/group were administered 0 (Milli-Q water
only), 5, or 20 mg/kg/day PFOS by oral gavage for 28 days (Cui et al. 2009). Rats were
sacrificed after exposure, and blood and tissue samples were 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  socket and nose and yellow staining in 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 index (HSI), renal-somatic index (RSI), and gonad-somatic 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 LOAEL of 5
mg/kg/day in rats was identified based on a significant decrease in body weight, dose-related
effects in the liver and pulmonary congestion. A NOAEL could not be identified.

Mouse.  The variability in the serum lipid profiles in humans suggests that response  to PFOS
exposure could be impacted by individual physiological differences and that environmental
factors such as diet might contribute to intraspecies variability in response. L. Wang et al.  (2014)
reported on the differences in response of male BALB/c mice (4-5 weeks old) administered
PFOS (0, 5, or 20 mg/kg) for 14 days while concurrently given diets that varied in fat [regular fat
(RF) versus high fat (HF) content]. The high fat diet contained 10% more lard and 3% more
cholesterol than the regular fat diet. Liver and serum responses were evaluated after a 14 day
exposure period. The data were for the endpoints monitored were presented graphically.

    Following PFOS exposure, there was an increase in liver fat content in both groups and a
decrease in liver glycogen in rats on both diets. For the mice on the regular fat diet,  the addition
of PFOS led to a significant increase in liver fat content (an approximately two-fold increase).
For the  mice on the high fat diet, the addition of PFOS caused a slight a slight and nonsignificant
increase in the liver fat content.

    The fat content of the diet alone was associated with significantly higher serum  levels  of
glucose, HDL cholesterol, LDL cholesterol, total cholesterol, and triglycerides. The differences
were significant for glucose, albumin, and total cholesterol (p < 0.01). For glucose,  cholesterol,
HDL, and LDL, the serum levels declined as the dose of PFOS increased; for triglycerides the
levels increased at a dose of 5 mg PFOS/kg/day and decreased at 20 mg PFOS/kg/day. PPARa
expression at the end of 14 day PFOS treatment increased for the RF group, but it decreased for
the HF groups (significant for the high dose).

    The authors examined the expression of several genes involved with lipid metabolism
(CPT1A and CYP7A1). CPT1A plays a role in transport of fatty acid into the mitochondria for
beta oxidation, and CYP7A1 is involved with the transformation of cholesterol into bile acids.
The high fat diet alone increased the expression of both genes. On the RF diet, the exposure to
PFOS was associated with a significant dose-related increase in CPT1A expression, whereas for
the high fat diet plus PFOS there was a significant decrease in expression. For CYP7A1
expression there was no significant impact of PFOS with the RF diet, whereas with the high fat
diet there was a highly significant decrease in expression with PFOS. The study demonstrates a
clear influence of diet alone on the liver and lipid profile of the treated mice, combined with

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-54

-------
some dose-related differences in the responses to PFOS exposure. The data support a possible
role for PFOS in inhibiting pathways for cholesterol metabolism and export of liver lipids and
identify some PFOS associated liver responses that are independent of PPARa activation.

   A 21-day study by Wan et al. (2012) examined mechanistic aspects related to the role of
PFOS in leading to hepatic steatosis in male CD-I mice (4/dose). Animals were given PFOS in
corn oil by gavage at doses of 0, 1, 5, or 10 mg/kg/day with sacrifice after 3, 7, 14, or 21  days.
Liver weights were significantly (p < 0.05) increased for the highest two dose groups across the
duration of the study and only at day 7 for the 1  mg/kg/day dose. The size of the liver was
significantly increased (p < 0.0003) at 5 and 10 mg/kg/day and a yellowish coloration of the
tissues was visually apparent. Histologically there was microvesicular steatosis on day 14 and
macrovesicular steatosis on day 21 at 10 mg/kg/day. The level of liver triglycerides was
significantly (p < 0.001) increased compared to control for the 5 and 10 mg/kg/day dose groups.

   The Wan et al. (2012) study included a series of mechanistic components to investigate the
mode of action for the effects observed. Both mRNA and protein expression for fatty acid
translocase and lipoprotein lipase were significantly increased for the 10 mg/kg/day dose. Levels
of mRNA in adipose tissue from the fat pad were not increased for either enzyme. Export of liver
lipids appeared to decrease, leading to lower serum LDL/VLDL levels on days 14 and 21. The
change correlated with increased liver weight and decreased expression of liver apolipoprotein
B-100  (apob). By day 21, apob expression was significantly decreased (p < 0.001) even in the
low dose group. Formation of hepatic VLDLs requires apob; the VLDLs are carriers of liver
triglycerides and other lipids from liver to serum.

   The authors also examined total hepatic P  oxidation, peroxisomal P oxidation, and
mitochondrial P oxidation using d31 palmitic acid. The results of this assay indicated that the
PFOS was primarily responsible for a decrease in mitochondrial P oxidation as monitored on  day
14. While total and peroxisomal P oxidation were slightly, but significantly, increased (p < 0.01)
at 10 mg/kg/day, mitochondrial P oxidation was markedly decreased (p < 0.05 or 0.01) in all
dose groups. Transcripts for mRNA for  peroxisomal acyl-CoA oxidase, Cyp 4al4, and acyl-Co
A dehydrogenase were significantly increased in the 5  and 10 mg/kg/day  dose groups, suggesting
breakdown of long chain fatty acids by peroxisomes. Increases in peroxisomal oxidation  in the
absence of increased mitochondrial beta oxidation can lead to accumulation of fatty acids in the
liver (steatosis). The LOAEL identified  for this study is 5 mg/kg/day. At 1 mg/kg/day there was
increased liver weight in the absence of histopathological correlates. The  1 mg/kg/day dose is
accordingly a NOAEL. The authors concluded that the hepatic changes observed in mice were
similar to those associated with nonalcoholic fatty liver disease in humans and were not totally a
reflection  of PPARa activation.

   Bijland 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 containing 0.25% cholesterol, 1% corn oil,
and 14% bovine fat for 4 weeks with or without 3 mg PFOS/kg/day. The diet  contained 0.25%
cholesterol, 1% corn oil, and 14% bovine fat. Plasma samples were collected via tail vein
bleeding and analyzed for a variety of lipid related endpoints including TC, triglycerides, VLDL,
and FIDL. 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-55

-------
    Significant decreases in triglycerides (-50%), total cholesterol (-60%), HDL (-74%), and
non-HDL (-60%) were found in mice given PFOS compared with controls. VLDL was also
significantly less than that of controls, but the level was only presented graphically. Radiolabeled
VLDL-like emulsion particles showed the plasma half-life of VLDL was reduced by 52% in
PFOS treated mice compared with controls accompanied by significantly increased uptake by
liver, heart, and muscle. VLDL production by the liver was markedly decreased (-87%) in
treated animals. Liver weight and hepatic triglyceride content were significantly greater
(p < 0.0001) and perigonadal fat pad weight was significantly less (p < 0.05) in PFOS treated
mice compared to those of controls. Thus, 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 and the mass of
perigonadal fat pad. 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.

    Compared with the controls, PFOS treated animals had 3,986 differentially expressed genes.
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 (1- to 2-fold) were those involved with fatty acid uptake and
transport and catabolism; triglyceride  synthesis; cholesterol storage; and VLDL synthesis. Genes
involved with HDL synthesis, maturation, clearance, and bile acid formation and secretion were
downregulated (1-fold for most genes to almost 4-fold for genes involved in secretion). 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, as well as 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.

3.2.3  Subchronic Studies

    Three monkey studies of oral PFOS exposure (two with rhesus- and one with cynomolgus-
strains) and two rat subchronic  studies are available.  The study with cynomolgus monkeys was a
GLP study. There are no subchronic studies by dermal or inhalation routes of exposure with
PFOS. In monkeys, clinical signs of toxicity were observed at 0.5 mg/kg/day, while lower body
weight, increased liver weight with hepatocellular hypertrophy, and decreased serum cholesterol
occurred at 0.75 mg/kg/day. Rats given 1.3-1.6 mg/kg/day had increased liver weight with
hepatocyte hypertrophy and decreased cholesterol.

Oral Exposure

Monkey. Two monkey studies were performed with rhesus monkeys (Goldenthal et al. 1978a
and 1979). In the first study, 2 monkeys/sex/dose 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; deaths were also observed at all
lower doses, but whether it was one or both of the animals was not stated. 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
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-56

-------
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 determined for this study.

   In the second study, 2 rhesus monkeys/sex/dose 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 they 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/dose,
except for the 0.03 mg/kg/day group (4 monkeys/sex), daily  for 26 weeks (182  days) in a GLP
study. Two monkeys per sex in the control, 0.15, and 0.75 mg/kg/day groups were monitored for
1 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 obtained for hepatic peroxisome
proliferation determination and immunohistochemistry was performed by PCNA to look for cell
proliferation. Selected results are shown in Table 3-11.

   Two of the 0.75 mg/kg/day males died; one died on day 155 and one was found moribund and
was 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 (lost 8  ± 8% and 4 ± 5%, respectively) after the treatment when
compared to controls (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% [males] and 33%-49% [females] 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 TSH was increased on  day  182 in the high-dose monkeys, but a true dose-response
was not observed and the monkeys had no indication of clinical hypothyroidism (TSH values
within reference range, no hyperlipidemia, and no thyroid gland histopathological lesions).

Perfluorooctane sulfonate (PFOS) - May 2016                                              3-57

-------
  Table 3-11. 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 (nU/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
Source: Data from Seacat et al. 2002
Notes: 'Statistically-significant from controls: *p<0,
** Statistically-significant from controls: p < 0.01.
05
   Hepatic peroxisome proliferation was measured by PCoAO activity and was increased
significantly in the 0.75 mg/kg/day females; however, the increase was not dose-related and it
was < two-fold. 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.
All females and 3/4 males at the high-dose 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 |ig/mL in males and 66.8 |ig/mL in females. Serum concentrations
associated with adverse effects (0.75 mg/kg/day) were 173 |ig/mL in males and 171  |ig/mL in
females.
Perfluorooctane sulfonate (PFOS) - May 2016
                                                3-58

-------
Rat. Goldenthal et al. (1978b) administered 0, 30, 100, 300, 1,000, or 3,000 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 died starting on day 7 after exhibiting
emaciation, convulsions, hunched back, increased sensitivity to stimuli, reduced activity, and red
material around the nose/mouth 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 at 100 ppm. 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 was 30 ppm (2 mg/kg/day) and the NOAEL could not be determined.

   Seacat et al. (2003) also performed an interim sacrifice for five Sprague-Dawley Crl:CD
(SD) IGS BR rats/sex/dose at the end of 14 weeks as part of the long-term cancer bioassay. The
animals were administered PFOS in the diet at concentrations of 0, 0.5, 2.0, 5.0, or 20 ppm.
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 for those sacrificed at 14 weeks. Animals were observed
twice daily for mortality and moribundity 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/dose  during week 14 for clinical chemistry,
hematology, and urinalysis evaluation. A thorough necropsy was performed at the end of 14
weeks of treatment for 5 animal s/sex/dose, and liver samples were collected for PCoAO activity,
cell 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 3-12.

   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).
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-59

-------
    Table 3-12. Mean (± SD) Values for Select Parameters in Rats Treated for 14 Weeks
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
Source: Data from Table 1 in Seacat et al. 2003
Note: 'Statistically-significant from controls, p < 0.
05
3.2.4   Neurotoxicity

    Available in vivo and in vitro studies focused on mechanistic endpoints to a greater extent
neurobehavioral indications of neurotoxicity. Effects observed included altered levels of
excitatory amino acids in the brain, changes in neurotransmitter levels and increases in miniature
post-synaptic currents along with inward calcium currents. One study found effects on learning
and memory in mice at approximately 2 mg/kg/day.

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. Adult male Wistar rats (5/group)
were administered a single 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, glycine, and GAB A.

    Rats in the 12.5, 25, and 50 mg/kg groups had significantly  (p < 0.05) decreased body
weights, by 15%, 22%, and 27%, respectively, compared to controls. Among the EAAs, the Glu
content was significantly decreased in the  hippocampus at the high dose (decrease of 77%
compared to controls; p <0.05); no other significant differences were recorded. In the cortex,  Glu
was the only excitatory amino acid (EAA) affected with significant decreases at 25 (decrease of
33% compared to controls) and 50 (decrease of 47 compared to controls) mg/kg. GS activity was
Perfluorooctane sulfonate (PFOS) - May 2016
                                                 3-60

-------
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.

Mouse. Groups of 15 adult C57BL6 mice (8 weeks old; number of each sex not specified) were
administered PFOS at doses of 0, 0.43, 2.15, or 10.75 mg/kg/day by gavage for three months
(Long et al. 2013). Learning and memory were assessed in the Morris water maze. The apoptosis
profile of hippocampal cells, as well as the levels of glutamate, GAB A, dopamine, 3,4-
dihydrophenylacetic acid (DOPAC), and homovanillic acid (HVA) were evaluated. In the water
maze trial, animals in the mid- and high-dose groups exhibited a significantly longer latency to
escape and spent significantly less time in the target quadrant. A significant increase in the
percentage of apoptotic cells was observed in the hippocampus of the mid- and high-dose
animals. Neurotransmitter levels were affected only in the high-dose group as based on
decreased dopamine and DOPAC levels plus increased glutamate levels. HVA and GABA levels
were unchanged by PFOS treatment.

   Differential protein expression at the high dose included down-regulation of Mibl protein (an
E3 ubiquitin-protein ligase), HercS (hect domain and RLD 5 isoform 2), and  Tyro3  (TYRO3
protein tyrosine kinase 3). Succinate dehydrogenase flavoprotein subunit (SDHA), Gzrna
(Isoform HF1 of Granzyme A precursor), Plau (Urokinase-type plasminogen activator
precursor), and Lig4 (DNA ligase 4) were upregulated. The 0.43 mg/kg/day dose group was the
NOAEL, and the 2.15 mg/kg/day dose group the LOAEL based on water maze performance.

In vitro. Slotkin et al. (2008) evaluated 10-250 |imol PFOS, PFOA, perfluorooctane sulfamide
(PFOSA),  and perfluorobutane sulfonate (PFBS) in vitro in differentiated and undifferentiated
PC12 cells, a neurotypic cell line. The study evaluated the following endpoints  as indications of
effects:

   •  Inhibition of DNA synthesis.
   •  Deficits in cell numbers and growth.
   •  Oxidative stress.
   •  Cell viability.
   •  Shifts in differentiation toward or away from the dopamine and acetylcholine (ACh)
       neurotransmitter phenotypes.

No effects on cell size, cell number, or neurocyte outgrowth were observed. PFOS decreased cell
viability at 250 jimol and promoted differentiation into the ACh phenotype at the expense of the
DA phenotype. The study suggests that the mechanisms for the observed effects in the
neurotypic cell lines are not the same for the individual perfluoroalkyl acids tested. The rank
order for the adverse effects measured in vitro was as follows: PFOSA > PFOS > PFBS = PFOA.

   Liao et al. (2009) assessed the 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 were tested at 100 jimol and included a variety of perfluorinated
compounds including PFOS. Testing showed the frequency of mPSCs increased in proportion to
the increase in carbon chain length. PFOS had a statistically-significant (p < 0.001) increase in
the mPSCs when compared to the four carbon PFBS. Inward calcium currents (lea) were
recorded in the presence or absence of the individual compounds with a ramp depolarization
pulse. Voltage values were recorded and plotted versus the corresponding lea every 5 mV and the
resulting current-voltage relationship curve established. All three sulfonic compounds increased


Perfluorooctane sulfonate (PFOS) - May 2016                                             3-61

-------
the lea. The longer the chain length the greater was the effect. PFOS caused the greatest increase
in ICa (% increase not provided).

   In the same study, the chronic effects of perfluorinated compounds (50 jimol) on neuronal
development were evaluated by measuring neurite outgrowth and branching. Among the sulfonic
compounds, only PFOS statistically suppressed the length of neurites (p < 0.001; 25% below that
of controls) and sum length of neurites per neuron (p < 0.001; 31% below that of controls). The
study suggested that the effects of perfluorinated sulfonates on neurons were greater than the
perfluorinated carboxylates. The study authors hypothesized that this reflects the fact that PFOS
was more likely to be incorporated into the lipid bilayer of cell membranes. This is consistent
with the results from a study by Matyszewska et al. (2008) who found that PFOS incorporation
into a model biological membrane was superior to PFOA and that it caused a change in
membrane fluidity and thickness depending on the amount incorporated.

3.2.5  Developmental/Reproductive Toxicity

   Rats and mice were found to be affected in developmental/reproductive studies with orally-
administered PFOS. Prenatal exposure of rats to PFOS caused an increase in neonatal mortality
when dams were given doses > 1 mg/kg/day and lowered pup body weight occurred at maternal
doses of 0.4 mg/kg/day. Neonatal death was shown to be a direct effect of PFOS on the lung
surfactant. Other developmental and reproductive toxicity effects included decreased gestation
length and developmental delays. Higher doses resulted in fetal sternal defects and cleft palate in
both rats and mice.

   Many specialized developmental studies have also been conducted with PFOS to assess long-
term effects in offspring (see section 3.2.6). Postnatal effects of gestational and lactational
exposure included evidence of developmental neurotoxicity, changes in thyroid and reproductive
hormones, altered lipid and glucose metabolism, and decreased immune function.

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 6 weeks prior to and during mating (Luebker et al. 2005b). Treatment in males continued
through the cohabitation interval,  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-sections) performed on GD 10; others delivered naturally and
were killed on 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, mortality, clinical signs, and mating/fertility parameters were
unaffected. During pre-mating, decreases in 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

Perfluorooctane sulfonate (PFOS) - May 2016                                              3-62

-------
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, therefore 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 until sacrifice. One
rat/sex/litter was tested in a passive  avoidance paradigm at 24 days of age and one rat/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 housed individually. The Fl generation male rats were sacrificed after mating, necropsied,
and evaluated as described in the FO generation. All Fl generation females were allowed to
deliver 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 and 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 stillborn pups 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 and 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 effects in pups at 3.2 mg/kg/day
included impacts on lactation (i.e., high number [~ 75%] of pups not  nursing and not having milk
present in the stomach), an increased incidence of stillborn pups, and a high incidence of
maternal cannibalization of the pups.

   In the Fl generation offspring, pups administered 3.2 mg/kg/day could only be evaluated on
LD 1 due to the high mortality. All viable pups from the 1.6 mg/kg/day group had significantly
(p < 0.05 or 0.01) delayed  eye opening, pinna unfolding, surface righting, and air righting during
lactation. No  delays were observed in rats administered doses < 0.4 mg/kg/day. Sexual
maturation was not affected in the 0.1 and 0.4 mg/kg/day groups after weaning. The results from
the passive avoidance (beginning at 24 days of age) and water maze tests (beginning at 70 days
of age) for neurobehavioral effects showed no dose-related effects on learning and memory.

F2 Generation: Fl parental animals displayed no clinical signs or mortality. Food consumption
was transiently decreased in Fl males, but it was not affected in Fl females. Reproductive
performance was unaffected in the Fl dams.

   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 were observed at 0.1  mg/kg/day on LDs 4 and 7, but mean pup body weights were
similar to controls by LD 14.  The pups in the 0.4 mg/kg/day group displayed 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-63

-------
   Based on the decreases in body weight gain and food consumption, the LOAEL for both the
FO male and female rats was 0.4 mg/kg/day and the NOAEL was 0.1 mg/kg/day. For the Fl rats,
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 with
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 were 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. Significant
reductions in gestation length, the average number of implantation sites, total litter size (live and
dead), and live litter size were observed for treated dams.

   Live litter sizes were comparable between treated and control groups following cross-
fostering. However, on LDs 2-4, approximately 19% of the pups in the group exposed
gestationally and lactationally were either found dead or presumed cannibalized compared to
1.6% for the negative control. For pups only exposed prenatally, mortality was 9% compared to
1.1% for those exposed during lactation only. Reductions in pup body weights on LD 1 were
observed in groups exposed both gestationally and lactationally and in those with gestational
exposure only. On LDs 4-21, pup body weights were reduced in all exposed groups when
compared to the negative control (p < 0.05 or 0.01). The greatest deficit in body weight
compared to controls was the group exposed during both gestation and lactation.

   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 treated
dams. No significant differences in pup lung histopathology were observed between the negative
control group and the treated animals.

   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


Perfluorooctane sulfonate (PFOS) - May 2016                                            3-64

-------
were below the limit of detection. Serum PFOS concentrations in the pups from treated dams,
fostered with untreated dams (in utero exposures) ranged 47.6-59.2 jig/mL. Serum PFOS
concentrations of treated dams ranged 59.2-157 |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 79.5-96.9 |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
(lactational 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 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, when the pups
were nursed by dams that had been exposed there was no significant effect on pup viability even
though the dams continued to receive PFOS during the period of lactation.

    The dose-response curve for neonatal mortality in rat pups born to PFOS exposed dams and
the associated biochemical and pharmacokinetic parameters were investigated in a companion
study (Luebker et al. 2005a). At 6 weeks prior to mating, female Crl:CD(SD)IGS VAF/Plus rats
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 GD 20 for dams assigned to C-section which
included eight dams in the control, 1.6, and 2.0 mg/kg/day groups, but none from the other dose
groups. Another group (~ 20 dams per dose group) was allowed to deliver and nurse their pups
through LD 4. These dams and their pups were sacrificed 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, and early/late resorptions. Maternal liver weights were
determined and the maternal organs examined by gross necropsy. Fetuses were pooled by litter
and mean weight recorded. For the dams that were allowed to deliver, reproductive and fetal
parameters (Table 3-13) were measured and recorded. Biochemical parameters investigated in
the dams and litters included: serum lipids, glucose, mevalonic acid, thyroid hormones (TT4 and
FT4, TT3, and FT3, and TSH), milk cholesterol, and liver lipids. Mevalonic acid was included as it
is abiomarker 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, fetuses/litter) in those dams receiving C-sections. Overall absolute body
weights of the dams were reduced slightly (5%-7% of that for the controls) in the 1.6 and
2.0 mg/kg/day  group dams during gestation; the changes, although slight, were statistically-
significant. Body weight change was  significantly reduced (p < 0.05 or 0.01) during premating at
2 mg/kg/day and during lactation at > 0.8 mg/kg/day. Food consumption showed a decreasing
trend with increasing dose during pre-mating, gestation and lactation. For dams allowed to
deliver, 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-65

-------
      Table 3-13. Fertility and Litter Observations in Dams Administered 0 to 2.0 mg
                                      PFOS/kg/day

Fertility index3 (%)
Implantations per
delivered litter
Gestation length
(days)
Gestation indexb (%)
Delivered pups/litter
Live births (%)
Dams with all pups
dying on LDs 1-5
Viability indexc (%)
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"
Source: Data from Luebker et al. 2005a
Notes: a Number of dams pregnant/number of dams mated x 100
b Number of dams with live offspring/number of pregnant dams x 100
c Number of live pups on day 5 postpartum/number of live births x 100
'Statistically-significant at p < 0.05
" Statistically-significant at p < 0.01

   In the group sacrificed on LD 5, a significant decrease in gestation length was observed at
doses > 0.8 mg/kg. Offspring viability was decreased starting at 0.8 mg/kg and was statistically-
significant at 1.6 and 2.0 mg/kg. The viability indices were 97.3%, 97.6%, 93.1%, 88.8%,
81.7%, 49.3%, and 17.1% at 0, 0.4, 0.8, 1.0, 1.2, 1.6, and 2.0 mg/kg, respectively (Table 3-13).
Lipids, glucose utilization, and thyroid hormones were similar or slightly different for treated
animals compared to controls. In all treated groups, pup body weight at birth on PND 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 based on
decreased body weight and a NOAEL was not identified.

   Several benchmark dose (BMD) estimates (BMDs and benchmark dose for the lower 95th
percentile confidence bound  [BMDLs]) were presented in the study.  They were as follows:

   •   Effect on gestation length: BMDs = 0.45  mg/kg/day, BMDLs = 0.31 mg/kg/day.
   •   Birth weight effect: BMDs = 0.63 mg/kg/day, BMDLs = 0.39 mg/kg/day.
   •   Decreased pup weight (day 5): BMDs = 0.39 mg/kg/day, BMDLs = 0.27 mg/kg/day.
   •   Pup weight gain (day 5): BMDs = 0.41 mg/kg/day, BMDLs = 0.28 mg/kg/day.
   •   Decreased survival of pups to  day 6: BMDs = 1.06 mg/kg/day, BMDLs = 0.89
       mg/kg/day.

   The impact of PFOS exposure on  the hypothalamic-pituitary-testicular axis in groups of 19
adult  male rats was studied by Lopez-Doval et al. (2014) following dosing at levels of 0, 0.5, 1,
3, or 6 mg/kg/day by gavage for 28 days. Serum  LH, FSH, and testosterone were measured in all
animals. The histology of the hypothalamus, pituitary gland, and testes were examined by light
microscopy and by electron microscopy (two animals/dose group using each method).
Noradrenaline concentration in the anterior and medial hypothalamus and median eminence and
GnRH in the whole hypothalamus were also determined in five animals/dose group each. For the
Perfluorooctane sulfonate (PFOS) - May 2016
3-66

-------
remaining five animals/dose group, GnRH gene expression in the hypothalamus and LH and
FSH gene expression in the pituitary gland were assayed.

    The pituitary gonadotrophic cells examined using an electron microscope showed structural
abnormalities in all exposed animals, although under light microscopy, the cells at the lowest
exposure levels appeared normal. At doses > 3 mg/kg/day the most active gonadotrophic cells
were classified as inactive based on the lack of homogeneous endoplasmic reticulum and a well-
developed Golgi complex. Many cells in the process of degeneration were observed. The
hypothalamus appeared to be normal at the two lowest doses, but not for doses > 3 mg/kg/day at
which basophilia, vacuolation, and irregular nuclear borders were seen. Histological
abnormalities (edema around seminiferous tubules and malformed spermatids) in the testes were
seen at doses > 1 mg/kg/day. Gene expression for LH and FSH were increased compared to
controls at the two lowest doses, with subsequent decreases at the higher doses. Serum LH and
testosterone were significantly decreased and FSH was significantly increased at all doses. Gene
expression for GnRH was significantly decreased compared to controls at all doses, while GnRH
levels in the hypothalamus were increased at the high dose. The results are consistent with
inhibition of the reproductive hypothalamus-pituitary-testicular axis at doses of 0.5 mg/kg/day
and above. The 0.5 mg/kg/day was the LOAEL based on significantly decreased LH and
testosterone concentration and increased FSH concentration.  The authors stated that the various
biochemical changes observed are linked and could be due to PFOS antiandrogenic and/or
estrogenic properties as has been proposed by other researchers.

Developmental Studies

Rat. Thibodeaux et al. (2003) administered 0, 1, 2, 3, 5, or 10 mg/kg PFOS in 0.5% Tween-20
daily by gavage 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 significantly
decreased (p < 0.0001) in a dose-dependent manner at > 2 mg/kg; these data were presented
graphically. A dose-dependent increase in the serum PFOS concentration was observed with
liver concentrations approximately four times higher than serum at each dose. Liver weight was
not affected in the treated rats. Serum chemistry showed significant decreases in cholesterol
(decrease of 14% compared to controls) and triglycerides (decrease of 34% compared to
controls) at 10 mg/kg. Serum thyroxine (T4) and T3 were significantly decreased in all treated
rats when compared to controls, however, a feedback response on 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 provided for different parameters were as follows:

    •   Maternal weight reduction BMDs = 0.22 mg/kg and BMDLs = 0.15 mg/kg (polynomial
       model).
    •   T4 effects on GD 7 BMDs = 0.23 mg/kg and BMDLs = 0.05 mg/kg (Hill model).
    •   Fetal sternal defects BMDs =  0.31 mg/kg and BMDLs = 0.12 mg/kg (logistic model).
    •   Fetal cleft palate BMDs = 8.85 mg/kg and BMDLs =  3.33 mg/kg (logistic model).
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-67

-------
   Lau et al. (2003) conducted a companion study to the one by Thibodeaux et al. (2003) in
order to examine the post-natal impact of in utero exposure to PFOS. Sprague-Dawley rats were
administered 0, 1, 2, 3, 5, or 10 mg/kg/day 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
PND 1. The number of pups per litter, number of live pups in the litter 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/day 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 maintained in the study and
used for serum collection and thyroid hormone analysis and as the subjects for the
neurobehavioral tests.

   In dams administered 10 mg/kg/day, the neonates became pale, inactive, and moribund
within 30-60 minutes of birth and all died. In 5 mg/kg/day 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/day. Pups from dams treated with 2 mg/kg/day still had
significant increases in mortality, but those from dams administered 1 mg/kg/day were similar to
controls (these data were presented graphically). No differences were observed in liver weight in
the neonates. Pup body weight was significantly decreased in dams administered > 2 mg/kg/day.
A significant (p < 0.05) delay in eye opening was observed at the same dose in the pups, but no
differences in onset of puberty were observed at that dose.  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 T3 or TSH.
The thyroid hormone data were presented graphically. 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 deficiencies.
Based on the findings, the developmental LOAEL is 2 mg/kg/day PFOS for mortality, decreased
body weight, and a significant 1-day delay in eye opening; the NOAEL is 1 mg/kg/day. The
authors calculated a BMDLs for a 6 day survival of 7.02 mg/kg/day.

   Because of the high number of fetal deaths, a sub-study was performed with newborns from
the 5 mg/kg/day PFOS group wherein they were cross-fostered with control dams immediately
after parturition. Survival was monitored for 3 days. Cross-fostering the pups from PFOS-treated
rats (5 mg/kg/day) with control dams did not increase their survival. Conversely, all control pups
fostered by PFOS treated dams survived, supporting the Luebker et al. (2005a) observations.

   Grasty et al. (2003) exposed pregnant rats to 25 mg/kg/day by gavage for four consecutive
days during critical windows of development (GDs 2-5,  6-9, 10-13, 14-17, or 17-20) or at 25
or 50 mg/kg/day on GDs 19-20. Litter size at birth was unaffected, but pup weight was
decreased in dams exposed for each of the 4 day intervals. Neonates died after dosing in all the
gestation time periods tested and the number of deaths 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; all pups had died by PND 4.

   In the dams treated only on GDs 19-20,  survival of the pups was 98%, 66%, and 3% in the
control, 25, and 50 mg/kg/day  groups on PND 5, respectively. Histological examination of the
lungs showed differences in the level of maturation between the control and treated pups.

   Grasty et al. (2005) performed a study with a comparable design to their 2003 study in order
to determine whether delayed lung surfactant maturation was responsible for neonatal deaths.

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-68

-------
Dams were given 25 or 50 mg/kg/day on GDs 19-20 and offspring evaluated on GD 21 or PND
0 immediately after birth. The newborns had normal pulmonary surfactant profiles.
Morphometric measurements of the histological lung sections of newborns showed significantly
(p < 0.05) increased proportion of solid tissue and decreased proportion of small airway space at
both doses. Co-treatment of dams with dexamethasone or trans-retinol palmitate as rescue agents
did not improve survival of newborns. These agents are used therapeutically to promote lung
maturation and surfactant production.

    While lung surfactant maturation did not appear to be the cause of death in the Grasty et al.
(2003) study, some data support effects of PFOS on lung surfactants. Xie et al. (2007, 2010a,
201 Ob) found that PFOS interacts with dipalmitoylphosphatidylcholine, a major lung surfactant.
As discussed in the distribution  section, Borg et al. (2010) found that radiolabeled PFOS was
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 what was in the maternal
blood on PND 1.

    Chen et al. (2012) administered 0, 0.1, or 2.0 mg/kg/day PFOS in 0.05% Tween 80 in
deionized water by gavage to 10 pregnant Sprague-Dawley rats/group on GDs 1-21. After
parturition (PND 0), pups were counted and 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 assessed for
markers of oxidative stress  and cytoplasmic protein and examined histologically.  The serum and
lungs were also analyzed for PFOS concentration. Three additional groups of  10 rats/dose were
treated as described above and the number of deaths/litter recorded until PND 4.

    Body weight of the pups was decreased and postnatal pup mortality (by PND  3) was
increased significantly (p < 0.05 and 0.01, respectively) 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 pup
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 pup serum (|ig/mL) were
approximately 2 times greater than that found in the pup 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 they
were still greater compared to serum.  PFOS was not detected in control pups at either timepoint.

    Histopathological changes observed in pup lungs at 2.0 mg/kg/day on PND 0  included
marked alveolar hemorrhage, thickened interalveolar septum, and focal lung consolidation. On
PND 21, the lungs also had alveolar hemorrhage, thickened septum, and inflammatory cell
infiltration. Numerous apoptotic cells were observed. No abnormalities were observed on
examination of the control rats or the pups from dams receiving 0.1 mg/kg/day.

    An increase in biomarkers associated with oxidative stress was found in pups from the
2.0mg/kg/day dams. The levels of malondialdehyde (MDA) were 473% and 305% of controls on
PND 0 and 21, respectively, and glutathione levels and superoxide dismutase (SOD) activity
decreased at both time-points compared to controls. Cytochrome c release from the inner
mitochondrial membrane and increased caspase -3, -8, and -9 are biomarkers for apoptotic cell
death. Each of these factors was significantly increased above that for controls at 2.0 mg/kg/day
on both PNDs 0 and 21. No changes were observed in the pups from dams receiving
0.1  mg/kg/day. The NOAEL for histopathological lesions in the lung, oxidative stress, and
apoptosis was 0.1 mg/kg/day with a LOAEL of 2 mg/kg/day.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-69

-------
   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 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 upregulated 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 upregulated genes were involved in significant
cytoskeletal, extracellular matrix remodeling, and transport and secretion of proteins.

   Lv et al. (2013) investigated the impact of gestational and lactational exposure to PFOS on
glucose and lipid homeostasis in offspring. Groups of 6 pregnant SPF Wistar rats were given
doses of 0, 0.5, or 1.5 mg/kg/day dissolved in 0.5% Tween 20 from GD 0 to PND 20. After birth,
pups were sexed, randomly selected and cross-fostered to insure there were equal pups per litter
(5 male and 5 female). Pup weights were determined on PNDs 0, 5, 10, 15, and 21. Serum and
liver samples were also collected at PND 0 and 21 from an unspecified number of pups.  The
remaining pups were maintained for 19 weeks after weaning before final sacrifice. Blood
samples were collected at 10 and 15 weeks after weaning and examined for fasting serum
triglycerides, total cholesterol, and fasting blood glucose. A glucose tolerance test was
administered after a 16-hour overnight fast. The adult pups were sacrificed at 22 weeks of age
for collection of total liver RNA with analysis for hepatic transcription factor SREBP-lc (sterol
regulatory element binding protein Ic) as a reflection of lipogenesis linked to glucose. Other
parameters  evaluated included serum insulin, leptin, and adiponectin, and gonadal fat weight,
pancreatic beta cell area, fat accumulation in the liver as monitored through oil red and
hematoxylin and eosin staining.

   Body weight of pups from treated dams was significantly reduced (p < 0.05) at birth,
throughout lactation, and persisted until week 8 post-weaning. A dose-related increase in glucose
intolerance was observed at 10 weeks post-weaning in pups from treated dams with statistical
significance attained at 1.5  mg/kg/day. At 15 weeks, pups from the 0.5 mg/kg/day dams had
significantly increased glucose intolerance, while that for high-dose pups was increased  but did
not attain statistical significance. Fasting glucose levels and serum glycosylated serum protein
concentrations were similar between pups from treated and control dams at 10 and 15 weeks
post-weaning. At 18 weeks after weaning,  pups from dams given 1.5 mg/kg/day had significant
increases in serum insulin, insulin resistance  index, and serum leptin. Serum adiponectin was
significantly decreased in pups from both treated groups compared with that of controls. At
sacrifice, pups from both treated groups had a significant increase in epigonadal fat pad weight,
and fat accumulation was observed in the liver of high-dose animals. The lowest dose tested
(0.5 mg/kg/day) was a LOAEL for a significant decrease in birth weight that persisted until week
8 of the post-lactation period, a significant increase of the  epigonadal fat pad weight at 19 weeks
after weaning, impaired glucose tolerance at  15 weeks after weaning, and decreased serum
adiponectin.

Mouse. As described for rats, a two-part developmental study with PFOS was performed 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/day 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 also measured in the mice.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-70

-------
   Maternal body weight gain was significantly decreased at 20 mg/kg/day. Food and water
consumption were not affected by treatment. Increases in serum PFOS were comparable to the
rat. PFOS treatment increased (p < 0.05) the liver weight in a dose-dependent manner in the
mice. T4 was decreased in a dose-dependent manner on GD 6 with statistical significance
(p < 0.05) attained for the 20 mg/kg/day group; levels  of T3 and TSH were not affected by
treatment. A significant increase in post-implantation loss was observed in animals administered
20 mg/kg/day, and reduced fetal weight (p < 0.05) was 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 doses > 10 mg/kg.

   In the second part of the developmental study, the  post-natal effects of in utero exposure to
PFOS were evaluated in the mouse (Lau et al. 2003). CD-I mice were administered 0, 1,  5, 10,
15 or 20 mg/kg/day of PFOS in 0.5% Tween-20 by gavage on GDs  1-17.

   Most mouse pups from dams administered 15 or 20 mg/kg/day did not survive for 24 hours
after birth. Fifty percent mortality was observed at 10  mg/kg/day. Survival of pups in the 1 and
5 mg/kg/day treated dams was similar to controls. A significant (p < 0.0001) increase in absolute
liver weight was observed at > 5 mg/kg/day. A  significant delay in eye opening was observed at
> 5 mg/kg/day. No dose- or treatment-related effects were observed on T4, T3, and TSH levels
in the pups. The LOAEL for this study in mice was 5 mg/kg/day and the NOAEL was 1
mg/kg/day. The authors calculated a BLDLs for survival at 6 days of 3.88 mg/kg/day.

   Ten pregnant ICR mice/group were administered 0, 1, 10, or 20  mg/kg of PFOS daily by
gavage from GD 1 to GD 17 or 18 (Yahia et al. 2008). Five dams/group were sacrificed on GD
18 for fetal external and skeletal effects 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 sacrificed on GD 18, the gravid
uterus was removed and the number of live/dead fetuses, fetal body weight, and number of
resorptions were recorded. Four pups/litter were sacrificed immediately after birth for
examination of their lungs.

   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 (59%) and 20 (60%) mg/kg/day
groups.

   All neonates in the 20 mg/kg/day dose group were born pale, weak, and 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. Results from this study are summarized in Table 3-14.

   Histopathological exam showed that all fetuses examined on GD 18 from dams treated with
20 mg/kg were alive and 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


Perfluorooctane sulfonate  (PFOS) - May 2016                                             3-71

-------
intracranial blood vessel dilatation when examined histopathologically. Three neonates from
each of the five dams treated with 10 mg/kg were examined, and 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.

   Table 3-14. Effects Observed in the Mice Administered PFOS from GD 0 to GD  17/18
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 PND 4 (%)
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"
0"
Source: Data from Tables 2-3 in Yahia et al. 2008
Notes: 'Statistically-significant difference between control and treated groups, p < 0.05
** Statistically-significant difference between control and treated groups, p < 0.01

    The effects of developmental PFOS exposure during gestation and lactation on glucose
metabolism in adult CD-I mice were studied by Wan et al. (2014b). The effects observed are
consistent with those in Wistar rats (Lv et al. 2013) discussed above. The dams were exposed to
doses of 0, 0.3, or 3 mg/kg/day dissolved in dimethyl sulfoxide (DMSO) and then in corn oil
from GD 3 to sacrifice on PND 21. The final concentration of DMSO was < 0.05% throughout
gestation and lactation. At PND 21, all dams and 2 pups per litter were sacrificed. The remaining
pups were randomly divided into two groups that were fed with either a standard diet or a high
fat diet until PND 63. Dams had increased liver weight at 3 mg/kg/day but no differences in
fasting serum glucose or insulin levels.

    There were no significant differences in pup weights at PND 21 although liver weights were
increased significantly (p < 0.05) at the highest dose for both the male and female pups. Both
sexes also had significant changes in genes regulating lipids and glucose at the highest dose.
Expression of CYP4A14, lipoprotein lipase, fatty acids translocase, the hepatic insulin receptor,
and insulin-like growth factor-1 receptor were significantly increased (p < 0.05) in males and
females from high-dose dams. The genes for prolactin receptor and insulin-like growth factor-1
were significantly decreased (p < 0.05) in males and females at 3 mg/kg/day.

    When evaluated at PND 63, liver weight in the pups was significantly increased at the high
dose in males, but not females. In the animals on the standard diet, fasting serum  glucose was
significantly (p < 0.05) higher for males and females at both doses, but fasting serum insulin
attained statistical significance only for the animals in the highest dose group. There were no
significant differences in oral glucose tolerance. The HOMA-IR index was increased
significantly for the high-dose group receiving the standard diet.
Perfluorooctane sulfonate (PFOS) - May 2016
3-72

-------
   The results from the glucose tolerance test (fasting blood glucose levels and blood glucose
levels over 2 hours following oral glucose challenge) became statistically-significant (p < 0.05)
at the high dose in both sexes fed high fat diets on PND 63. Fasting serum insulin was
significantly increased (p < 0.05) at 3 mg/kg/day in males and females on both diets, with the
effects more pronounced in mice on the high fat diet than in mice on the standard diet. The
HOMA-IR index was significantly increased (p < 0.01) at both doses for males and females on
the high fat diet.

3.2.6   Specialized Developmental/Reproductive Studies

Hormonal Disruption

Rat. Yu et al. (2009a) fed pregnant adult Wistar rats (n = 20/group) a control diet or a diet
containing 3.2 mg PFOS/kg feed. Doses to the dams were not calculated, and body weight and
feed consumption data were not presented. Treatment continued for both groups throughout
gestation and lactation. Dams were allowed to deliver, 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:

   •   Litters from control dams fostered by control dams (CC, unexposed control; n = 8).
   •   Litters from treated dams fostered by control dams  (TC, prenatal exposure; n =  8).
   •   Litters from control dams fostered by treated dams  (CT, post-natal  exposure; n = 8).
   •   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 same diet as the foster dam. Pups were
weighed and sacrificed 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 1A1 and 1A6 (UGT1A1 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. Body weight in offspring from
PFOS treated groups did not differ significantly from controls. Liver weights in pups from the
pre- and postnatal exposure (group TT) were significantly increased on PNDs 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 with time in those  exposed only
prenatally (group TC). The levels increased in those in the  TT group. These results indicate that
PFOS can be transferred by the placenta and through lactation.

   The total T3 and rT3 were not affected by PFOS treatment of the pups. Compared to
controls, pups in all treated groups 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 larger than that of the TC  group. On
PNDs 21 and 35, T4 levels were 71%-75% and 63%-64% of controls for the CT and TT groups,
respectively, compared with 80%-81% of control for the TC group  on both days. Pups in the TT
group (exposed pre- and postnatally) had T4 levels that were significantly  lower than the controls
at PND 14. For  gene expression, no statistically-significant differences were observed between

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-73

-------
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 interval. Lactational exposure appears to be an important contributor to the
observed thyroid effects given that the serum PFOS levels were higher, and T4 levels lower, in
the CT group than in the TC group.

   The effects of PFOS on testosterone production by fetal Ley dig cells were investigated
following prenatal exposures (Zhao et al.  2014). Pregnant Sprague-Dawley rats (n = 4) were
administered PFOS by gavage at doses of 0, 5, or 20 mg/kg/day on GDs 11-19; controls received
the 0.05% Tween 20 vehicle. Dams were  killed on GD 20 and the male pups removed, weighed,
and measured for length and anogenital distance. The fetal testes were removed for analysis of
testosterone  production, fetal Leydig cell  numbers, ultrastructure, and gene and protein
expression levels. Dams given 20 mg/kg/day had significantly lower body weight and serum
cholesterol levels on GD 20. Male fetuses had significantly lower body weight at 5 and
20 mg/kg/day. At 20 mg/kg/day there were significant differences in body length, anogenital
distance, and testes weight; all measures were lower than those for controls.

   Testicular mRNA levels of growth factors (Kitl), cholesterol transporters (Scarbl and Star).,
steroidogenic enzymes (Cypllal, Cypl7a, and Hsd3b Injunction protein (Trmp2\ and LH
receptor (Lhcgr) were significantly reduced in fetuses from dams given 20 mg/kg/day. Fetuses
from high-dose  dams also had significantly lower testicular testosterone levels, enzyme activity,
and protein levels for 3p-hydrosteroid dehydrogenase and 17a-hydroxylase/20-lyase.  Liver
cholesterol and testes HDL-cholesterol levels were reduced in fetuses from high dose dams.
Histologically, the number of fetal Leydig cells was reduced and  showed a decreased number of
lipid droplets and features of apoptosis at 20 mg/kg/day. The 5 mg/kg/day dose was a LOAEL
for effects on male fetal body weight.

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 by gavage from GD 0 through 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, sex ratio, birth to PND 4 survival, PND 4-21 survival,
pup body weights through PND 72, and gross internal findings. Maternal body weight and  body
weight gain during gestation were comparable between the treated and control groups. On LDs
1-4, dams in the 1.0 mg/kg/day group had slightly, but not significantly, lower weight gain and
food consumption than  those of controls resulting in significantly lower (p < 0.05 or 0.01)
absolute body weight throughout lactation. Food consumption was transiently decreased
(p < 0.05 or  0.01) on GDs 6-9 for the 0.3  mg/kg/day group  and on GDs 6-12 for the
1.0 mg/kg/day group. These findings in the treated dams are not considered to be treatment-
related or adverse. 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 assessments
performed on PNDs 4, 11, 21, 35, 45, and 60. Male offspring  from dams administered 0.3 and

Perfluorooctane sulfonate (PFOS) - May 2016                                              3-74

-------
1.0 mg/kg/day had statistically-significant (p < 0.05) increases in motor activity on PND 17, but
this was not observed on PND 13,21 or 61. No effect on habituation was observed in the 0.1 and
0.3 mg/kg/day males or in the 1.0 mg/kg/day females. On PND 17, males at 1.0 mg/kg/day
showed a lack of habituation as evidenced by significantly (p < 0.05) increased activity counts
for the sequential time intervals of 16-30, 31-45, and 46-60 minutes. The normal habituation
response is for motor activity to be highest when the animals are first exposed to a new
environment and to decline during later exposures to the same  environment as they have learned
what to expect. 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 reflecting 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.

   Y. Wang et al. (2015) examined the effects of PFOS on spatial learning and memory
following pre- and post-natal exposure. Pregnant Wistar rats were administered PFOS in the
drinking water at 0, 5,  or 15 mg/L beginning on GD 1 and continuing through lactation. Doses to
the animals were not calculated, and body weight and water consumption data were not
presented. Doses were estimated as 0, 0.8, or 2.4 mg/kg/day using subchronic values for female
Wistar rats from USEPA (1988). Maternal serum levels in the treated groups were 25.7 and
99.3 |ig/mL, respectively, on PND 7 and 64.3 and 207.7 |ig/mL, respectively, on PND 35.  On
PND 1 pups were cross-fostered to establish groups for unexposed controls, only prenatal
exposure, only post-natal exposure, and continuous exposure. After weaning, pups were given
the same treated or control water as their foster dam. Three pups per group were sacrificed on
PNDs 7 and 35 for measurement of protein  and RNA levels in the hippocampus. On PND 35,
8-10 pups per group were tested in the  Morris Water Maze which consisted of one day of visible
platform tests, seven days of hidden platform tests, and a probe trial 24 hours after the last
hidden platform test.

   Offspring survival  on PND 1 was significantly reduced  from high-dose dams before cross-
fostering; survival on PND 5 was not given. On water maze testing day 1, swimming speed and
the time to reach the visible platform were similar between  all treated and control groups.
Thereafter, escape latency was significantly increased for all treated groups on one or more
testing days. The most pronounced and significant effect was in pups exposed prenatally from
dams given 15 mg/L and cross-fostered to control  dams. Similar trends were observed for escape
distance. During the probe trial for memory testing, pups continuously exposed pre- and post-
natally to 15 mg/L spent less time in the target quadrant than the unexposed controls but
statistical significance  was not achieved as consistently as that for the group exposed only during
gestation. Protein levels of growth-associated protein-43, neural cell adhesion molecule 1, nerve
growth factor, and brain-derived neurotrophic factor were significantly decreased in the
hippocampus on PND  35, especially in pups exposed prenatally to 15 mg/L and cross-fostered to
control dams.

   Ten pregnant Sprague-Dawley rats/group were administered 0, 0.1, 0.6, or 2.0 mg/kg/day of
PFOS in 0.5% Tween 80 by oral gavage 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. Astrocyte activation markers,  glial fibrillary acidic protein (GFAP) and SI00
calcium binding protein B, which are associated with morphological changes inside the cell,
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-75

-------
were evaluated with immunohistochemistry. 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
than on PND 0.

   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 1 beta (IL-1P) and tumor necrosis factor-a(TNF) 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 at > 0.6 mg/kg and TNF-
a was increased in the high dose 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 greatest increase was observed in the hippocampus on with a significant increase
in activation protein-1 (AP-1) in all dose groups and an increase in nuclear factor-xB (NF-icB)
and cAMP response element-binding protein at > 0.6 mg/kg groups at PND 0.  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.

Mouse. Fuentes et al. (2007) treated 8-10 pregnant Charles River CD-I mice/group to 0 or
6 mg/kg/day 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 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 tests to further assess development. The PFOS treatment had no effect on maternal
body weight or food/water consumption. On PNDs 4 and 8, pups from dams treated with
6 mg/kg of PFOS had reduced body weight, as well as delayed (p < 0.05) eye opening, pinna
detachment, and surface righting reflex. Female pups from dams exposed to 6 mg/kg of PFOS
and stressed by immobilization exhibited reduced open-field activity. No differences in activity
were observed for male pups and rotarod performance was not affected in any group by PFOS
alone  or combined with maternal stress.

   Ten-day old male neonatal Naval Medical Research Institute (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


Perfluorooctane sulfonate (PFOS) - May 2016                                            3-76

-------
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). The LOAEL was 0.75 mg/kg based on
decreased locomotion, rearing, and total activity in 2 month old mice.

   Johansson et al. (2009) administered a single oral dose of 0 (3 litters) or 11.3 mg/kg (four
litters) to NMRI male mice (10 days old). The exact number of male mice in each litter was not
provided. 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.

   There were no clinical signs of acute toxicity, and no treatment-related body weight
differences. The CaMKII and GAP-43 protein levels in the hippocampus were both increased in
the PFOS treated males; 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 hippocampus and (p < 0.01; 59%) in the cerebral cortex of the treated mice. The tau
protein levels in the cerebral cortex were increased significantly (p < 0.05; 80%) in treated
animals compared to controls. Overall, the study indicates that a one-time treatment with
11.3 mg//kg PFOS had a significant effect on the neuronal proteins evaluated.

Tissue and Metabolic effects

   Zeng et al. (2014) examined cardiac mitochondria mediated apoptosis in weaned rats
exposed by way of their dams (10 per dose group) to 0, 0.1, 0.6, or 2 mg/kg/day in 0.05% Tween
80 by gavage on GDs 2-21. The pups were sacrificed at the end of the lactation period. Trunk
blood and the heart were recovered. Apoptotic  cells in the heart tissue from six animals per dose
group were measured using a Terminal deoxynucleotidyl transferase dUTP nick end labeling
(TUNEL) staining assay by an individual pathologist blinded to the exposure group. The
apoptosis index was recorded as percent apoptotic cells per 1,000 cells in the same section.
PFOS exposure was associated with a dose dependent increase in the percentage of TUNEL
positive nuclei (p < 0.05). The 0.6 mg/kg/day dose was the LOAEL and the 0.1  mg/kg/day dose
the NOAEL. The researchers found that biomarkers for apoptosis were supportive of the TUNEL
results. The expression of BCL2-associated X protein and cytochrome c were upregulated and
bcl-2 downregulated.  The concentration of caspase 9 was significantly increased above the
control levels at all doses and caspase 3 levels were significantly increased for all but the lowest
dose level.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-77

-------
3.2.7   Chronic Toxicity

   Only a single chronic exposure study in animals is available (Thomford 2002/Butenhoff et al.
2012). It is the long term component of the Seacat et al. (2002) subchronic study reported in
section 3.2.3. Sprague-Dawley Crl:CD (SD)IGS BR, rats (n = 40-70) were dosed using a PFOS
containing diet for up to 105 weeks. Five per sex per dose group were sacrificed at 4 and 14
weeks as described earlier. Treatment resulted in decreased body weight, with increased liver
weight with hepatocellular hypertrophy. A satellite group of animals received 20 ppm of the
PFOS containing diet for 52 weeks, followed by the control diet until sacrifice at week 106.

   The animals received dietary levels of 0, 0.5, 2, 5, or 20 ppm PFOS as the potassium salt.
Corresponding PFOS doses were 0, 0.024, 0.098, 0.24, and 0.984 mg/kg/day, respectively, for
males and 0, 0.029, 0.120, 0.299, 1.251 mg/kg/day, respectively, for females. Five animals/sex in
the treated groups were sacrificed during week 53 and liver samples were obtained for
mitochondrial activity, hepatocellular proliferation rate, and determination of palmitoyl-CoA
oxidase activity; liver weight was recorded. The results from the 4-week and 14-week sacrifices
(Seacat et al. 2002) from this study are provided in sections 3.2.2 and 3.2.3, respectively. Serum
samples were collected at weeks 27 and 53  from 10 rats/sex/dose group and were examined for
clinical effects associated with systemic toxicity; liver samples were obtained during and at the end
of the study for determination of PFOS concentration. Data on chronic effects were not reported
for the recovery group. The concentration of PFOS in serum was measured at weeks 4, 14, and
105. In males the serum levels decreased between week 14 and 105 by 50% for all but the 0.5 ppm
group where the decrease in serum concentration was larger. A serum measurement was available
at 53 weeks for the high dose males and was comparable to the value at 14 weeks. In females
serum levels serum levels remained relatively constant at 14 and 105 weeks. In both males and
females the concentrations in the liver were lower at 105 weeks than they were at 14 weeks.

   The clinical serum observations for ALT at 53 weeks were consistent with those at 14 weeks
in demonstrating significant (p < 0.05) increases for the high dose males but not females. At
week 27, ALT was increased for high-dose males, but did not attain statistical significance. For
males at 53 weeks in the 0, 0.5, 2, 5, and 20 ppm groups, ALT values were 54 ± 66, 62 ± 52,
40 ± 7.5, 44 ± 8.3, and 83 ± 84 IU/L, respectively. The large SDs were the result of high values
in one animal in each of the control and 0.5 ppm groups and two animals in the 20 ppm  group.
Thus, some animals may be more sensitive to liver damage as a result of exposure than others.
AST levels were not increased for either sex. Serum blood urea nitrogen (BUN) was
significantly (p < 0.05) increased at 20 ppm for males and females at weeks 14, 27,  and  53 and in
5 ppm males and females at 27 and 53 weeks. The males in the 2 ppm group also had a
significant (p < 0.05) increase in BUN at 53 weeks. These data were presented graphically in
Butenhoffetal. (2012).

   Nonneoplastic lesions in the liver are shown in Table 3-15. 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 (females only), and centrilobular hepatocytic
vacuolation (males only). At the high dose, there was a significant increase in the number of
animals with single cell hepatic necrosis in both males and females at 53 weeks. Necrosis in the
recovery animals was comparable to the controls.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-78

-------
               Table 3-15. Incidence of Nonneoplastic Liver Lesions in Rats
                            (Number Affected/Total Number)
Lesion
0 ppm
0 mg/kg/day (d)
0.5 ppm
0.024 mg/kg/d
2.0 ppm
0.098 mg/kg/d
5.0 ppm
0.242 mg/kg/d
20 ppm
0.984 mg/kg/d
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
0 mg/kg/d
2/65
0/65
7/65
0.029 mg/kg/d
1/55
0/55
6/55
0.120 mg/kg/d
4/55
0/55
6/55
0.299 mg/kg/d
16/55"
7/55"
6/55
1.251 mg/kg/d
52/65"
36/65"
15/65*
Source: Data from Thomford 2002/Butenhoff et al. 2012
Notes: 'Significantly increased over control: p < 0.05
** Significantly increased over control: p < 0.01.

   No effects were observed on hepatic palmitoyl-CoA oxidase activity or increases in
proliferative cell nuclear antigen (PCNA) at weeks 4 and 14 or bromodeoxyuridine 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. The LOAEL at termination for male rats was 2 ppm
(0.098 mg/kg/day) and for female rats was 5 ppm (0.299 mg/kg/day) based on the liver
histopathology.  The NOAEL for the males was 0.5 ppm (0.024 mg/kg/day) and 2 ppm
(0.120 mg/kg/day) for females. Additional details from the study in regard to carcinogenicity are
provided in section 3.2.8.

   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 week 14 and 53 sacrifices, absolute
and relative liver weights were  significantly increased at 20 ppm in males and relative liver
weight was increased at 20 ppm in females. At week 53, liver weight data were given only for
the control and 20 ppm groups such that a dose-response could not be evaluated.

3.2.8   Carcinogenicity

Rat. Tumor data were collected as part of the chronic study (Thomford 2002/Butenhoff et al. 2012)
described above. The tumor results are provided in Table 3-16. 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 < 0.0456) in the high-dose group (7/60, 11.7%) over the
control (0/60, 0%). No hepatocellular tumors were observed in the recovery group exposed for 52
weeks and sacrificed at 106 weeks. Liver tumors were observed in males at all  doses (0%, 6%, 6%,
2%, and 11.7%). In 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 sacrifice. Here too, the response was not linear to dose with sequential values of
0%, 2%, 2%, 2%, and 8.3%. These cases were associated with  significant increases in the high-
dose group 5/60 (p = 0.0386; 8.3%) for adenomas and 6/60 (p = 0.0204; 10%) for combined
adenomas and carcinomas compared to the control. The female recovery group had 2/20 liver
adenomas (5%)  and no carcinomas. The presence of increased levels of ALT in the males of the
Perfluorooctane sulfonate (PFOS) - May 2016
3-79

-------
    high dose group at 14, 27, and 53 weeks supports hepatic tissue damage with compensatory repair
    as a probable a possible mode of action (MO A) for the liver tumors. In all cases the SDs about the
    means are broad suggesting that some animals could be less resilient than others to the liver effects.

                                 Table 3-16. Tumor Incidence (%)
Tumors
0 ppm
0 mg/kg/d
0.5 ppm
0.024 mg/kg/d
2.0 ppm
0.098 mg/kg/d
5.0 ppm
0.242 mg/kg/d
20 ppm
0.984 mg/kg/d
20 ppm
recovery
1.144 mg/kg/day
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* (9/3 9)
2.6 (1/39)
25.6 (10/39)
Females

Liver
hepatocellular adenoma+
hepatocellular carcinoma
combined+
Thyroid
follicular cell adenoma
follicular cell carcinoma
follicular cell combined
C-cell adenomas
C-cell Carcinomas
C-cell combined
Mammary
Fibroma/ Adenoma
Adenoma
Combined adenomas
carcinoma
0 mg/kg/d
0 (0/60)
0 (0/60)
0 (0/60)
0 (0/60)
0 (0/60)
0 (0/60)
20.0 (12/60)
0 (0/60)
20.0 (12/60)
33.3 (20/60)
11.7(7/60)
38.3 (23/60)
18.3(11/60)
0.029 mg/kg/d
2.0 (1/50)
0 (0/50)
2.0 (1/50)
0 (0/50)
0 (0/50)
0 (0/50)
12.0 (6/50)
2.0 (1/50)
14.0 (7/50)
54.0* (27/50)
12.0 (6/50)
60.0* (30/50)
24.0 (12/50)
0.120 mg/kg/d
2.0 (1/49)
0 (0/49)
2.0 (1/49)
0 (0/49)
0 (0/49)
0 (0/49)
12.2 (6/49)
0 (0/49)
12.2 (6/49)
39.6 (19/48)
10.4 (5/48)
45.8 (22/48)
31.3(15/48)
0.299 mg/kg/d
2.0 (1/50)
0 (0/50)
2.0 (1/50)
4.0 (2/50)
2.0 (1/50)
6.0* (3/50)
16.0 (8/50)
0 (0/50)
16.0 (8/50)
48.0 (24/50)
14.0 (7/50)
52.0 (26/50)
22.0(11/50)
1.251 mg/kg/d
8.3* (5/60)
1.7 (1/60)
10.0* (6/60)
1.7 (1/60)
0 (0/60)
1.7 (1/60)
8.3* (5/60)
0 (0/60)
8.3* (5/60)
18* (11/60)
6.7 (4/60)
25.0* (15/60)
23.3 (14/60)
1.385 mg/kg/d
5.0 (2/40)
0 (0/40)
5.0 (2/40)
2.5 (1/40)
0 (0/40)
2.5 (1/40)
15.0 (6/40)
2.5 (1/40)
17.5 (7/40)
37.5 (15/40)
10.0 (4/40)
40.0 (16/40)
10.0 (4/40)
Source: Data from Thomford 2002/Butenhoff et al. 2012.
Notes: +Significant positive trend.
* Significantly increased over the control: p < 0.05
** Significantly increased over the control: p < 0.01.

        There were cases of thyroid follicular cell adenomas and carcinomas in both the male and
    female rats but no pattern of dose-response or significant increases compared to controls. The
    incidence of thyroid follicular cell adenomas in the male recovery group was increased
    significantly (p = 0.028) over controls (23.1% vs 5%). The incidence of combined thyroid
    follicular cell adenoma and carcinoma in the recovery group males (10/39, 25.6%) did not attain
    statistical significance compared to that of the control group (6/60, 10%). The males that were
    continually dosed for 105 weeks had a much lower adenoma incidence than the recovery group
    (6.8% versus 23.1%). In no case  were thyroid tumors determined to be a cause of death.

        In females, there was a significant increase (p = 0.0471) for combined thyroid follicular cell
    adenoma and carcinoma in the mid-high (5.0 ppm) group (3/50, 6%) compared to the control
    group (0/60, 0%). The incidence data for thyroid follicular tumors lacked dose-response. C-cell
    thyroid adenomas had a higher incidence than the follicular cell tumors in female rats. The
    Perfluorooctane sulfonate (PFOS) - May 2016
3-80

-------
highest incidence was in the control group (20%); there was a lack of dose-response across
groups (8%-18%). As was the case with the combined adenomas and carcinomas, the C-cell
tumors were not identified as a cause of death.

    There was a high background incidence in mammary gland tumors in the female rats,
primarily combined fibroma adenoma and adenoma (25%-60%), but the incidence lacked dose-
response for all tumor classifications. Significant (p = 0.0318) increases combined mammary
fibroadenoma/adenoma (30/50, 60%; p = 0.0318) were observed in the low-dose (0.5 ppm)
group compared to the respective controls but there was a lack of dose response with the high
dose group having a lower incidence (25%) than the controls (38%). Mammary gland
carcinomas also lacked dose-response and had a relatively comparable incidence across dose
groups including the controls.

Mouse. The mouse model C57BL/6J-M/W/+ for intestinal neoplasia was used to study the
obesogenic and tumorigenesis effects of PFOS following in utero exposure (Ngo et al. 2014).
The C57BL/6J-y4/»cMm/+mouse has a heterozygote mutation in the tumor suppressor gene
adenomatous polyposis coli (Ape), and is therefore a sensitive model in which to test whether
chemicals can affect intestinal tumorigenesis. Wild-type females (Apc+/+), mated to heterozygous
males (ApcMm/+), were given  0, 0.01, 0.1, or 3 mg/kg/day by gavage on GDs  1-17 and allowed to
litter naturally. Offspring with ApcMin/+ genotype were terminated at 11 weeks of age for study
of intestinal tumorigenesis and obesogenic effect while wild-type (Apc+/+) offspring were
sacrificed at 20 weeks to assess any obesogenic effect at an older age. In the treated groups,
whole litter loss occurred in 6/16, 10/28, and 7/14 dams, respectively, compared with 2/22
controls; the timing of loss, late,  or early gestation, was not stated.  No clinical signs of toxicity
were observed during dosing and maternal body weight was similar between treated and control
groups. For offspring of either genotype, terminal body weight, liver and spleen weights, and
plasma glucose were not affected by in utero exposure. PFOS did not increase intestinal
tumorigenesis in susceptible,  ApcMm/+, offspring.

3.3  Other Key Data

3.3.1   Mutagenicity and Genotoxicity

    Results of genotoxicity testing with PFOS are summarized in Tables 3-17 and 3-18. PFOS
was tested for mutation 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 co//'/Mammalian-microsome reverse mutation assay with and without metabolic
activation (Mecchi 1999), in an in vitro assay for chromosomal aberrations in human whole
blood lymphocytes with and without metabolic activation (Murli 1999), and in 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 whether
induction of gene mutation in five strains of S. typhimurium and in S. cerevisiae strain D3 would
take place with and without metabolic activation (Simmon 1978). The results were negative.

    Governini et al. (2015) collected semen samples from 59 healthy-nonsmoking patients
attending a Center for Couple Sterility at the University in Siena, Italy.  The subjects were
divided into those that were normozoospermic (13) and those that were oligoasthenoterato-

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-81

-------
zoospermic (46). PFOS was present in 25% of the seminal plasma samples and 84% of the serum
samples. Conversely PFOA was present in 75% of the seminal plasma samples and only 16% of
the blood samples. Sperm were evaluated for the presence of aneuploidy and diploidy, and sperm
DNA was evaluated for fragmentation using the TUNEL assay. The frequencies of aneuploidy
and diploidy were significantly greater in the PFAS positive samples than in the PFC negative
samples (p < 0.001 and p < 0.05, respectively) suggesting the possibility for errors in cell
division. The levels of fragmented chromatin were significantly increased (p < 0.001) for the
PFC positive group compared with the PFAS negative group.

                       Table 3-17. Genotoxicity of PFOS in vitro
Species
(test system)
Salmonella strains and D4
strain of Saccharomyces
cerevisiae
Salmonella strains and
Escherichia coli WP2wvr
5 strains of S. typhimurium
and S. cerevisiae strain D 3
Human lymphocytes
Hepatocytes from Fisher
344 male rats
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 3-18. Genotoxicity of PFOS in vivo
Species
(test system)
Crl:CD-lBRmice
End-point
Presence of micronuclei in bone marrow
Results
negative
Reference
Murli 1996
3.3.2   Protein binding

   The ability of PFOS to bind to serum proteins for distribution is discussed in section 2.2. PFC
protein binding can also impact cellular function in cases where the proteins in question are
transporters (serum albumin and fatty acid binding protein), enzymes (lysine decarboxylase), or
membrane receptors such as members of the PPAR family and thyroid hormone receptors. The
mechanistic studies of the membrane receptors are described in section 3.3.4.

   Ren et al. (2015) examined the relative binding affinities of 16 perfluoroalkyl compounds for
the human thyroid hormone receptor's a ligand binding domain (TRa-LBD) using a fluorescence
competitive binding assay. Solutions of Ijimol TRa-LBD were prepared in DMSO. Changes in
TRa-LBD tryptophan fluorescence after binding to 10 (imol T3 in the absence or presence of the
PFAS was used to determine the binding properties of the PFAS. ICso values were calculated by
linear extrapolation between two responses located in the vicinity of a 50% inhibition level. All
the PFAS had a lower affinity for the receptor than T3, but the binding affinity of PFOS was
greater than that for PFOA and the other sulfonates tested. The ICso value for PFOS was
16 jimol, compared with 0.3 jimol for T3.
Perfluorooctane sulfonate (PFOS) - May 2016
3-82

-------
   Lysine decarboxylase is a key enzyme involved in the production of cadaverine from the
amino acid lysine. S. Wang et al. (2014) studied the impact of a series of 16 PFAS on the activity
and conformation of this enzyme because of its involvement in growth and development. The
interaction assays were carried out in vitro using a fluorescent probe to measure enzyme activity.
The impact of a PFAS on enzyme activity caused a decrease in fluorescence that represented
enzyme inhibition. Varying the PFAS concentrations provided the data for determining
inhibition constants for each compound tested. Members of the sulfonate family were stronger
inhibitors than the carboxylic acids, and enzyme inhibition increased as did the length of the
carbon chain. Only the 4, 6, and 8 carbon members of the sulfonate family were tested.

   Circular dichroism was used as a tool for determining changes in enzyme conformation in the
presence  of the tested PFAS (S. Wang et al. 2014). PFOS caused a greater change in  enzyme
conformation than PFOA. Cellular cadaverine production was decreased indicating the potential
for PFOS to alter metabolism by way of enzyme inhibition as a consequence of its protein
binding properties. To date there has been scant investigation of PFOS or other PFASs as
enzyme inhibitors.

   An in vitro study of the impact of PFOS (and other PFASs) on the conformation of several
proteins (BSA, ovalbumin, and p-galactosidase) in solution found that the denaturing effect of
the PFAS depended on the amino  acid composition and conformation  of the protein as well as
the individual PFAS (Ospinal-Jimenez and Pozzo 2012). The PFOS concentration (1  millimole
[mmol]) was higher than one would expect in vivo because the study was designed to examine
denaturing potential.

   Enzymes targeted by PFOS can vary. Molecular docking analysis of PFOS's potential to bind
with and  change the activity of enzymes along metabolic pathways associated with its critical
effects could provide important insights related to toxicity. The importance of the S. Wang et al.
(2014) and Ospinal-Jimenez and Pozzo (2012) studies are the evidence they produced showing
that the protein binding properties of a PFAS can impact the conformation, thereby possibly
changing activity.

3.3.3  Immunotoxicity

Human-in vitro. In a pilot study, Brieger et al. (2011) examined the effects of PFOS on human
leukocytes. Peripheral blood mononuclear cells (PBMC) were obtained from 11 voluntary
donors (n = 6 females, 5 males). The mean plasma concentrations of PFOS were 0.004, 0.0028,
and 0.0055 |ig/mL for all, female, and male volunteers, respectively. PBMCs were incubated
with varying concentrations of PFOS followed by assays for cell viability, proliferation,  and
natural killer (NK) cell activity. The human promyelocytic leukemia cell line,  HL-60, was also
used in cell viability and monocyte differentiation assays. The various components of the assays
employed and the results are identified as follows:
   1.  In the cell viability assay, the PBMCs and HL-60 cells were incubated with 0-125 ug/mL
       of PFOS for 24 hours. 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.
   2.  In the proliferation assay, the PBMCs were incubated with 0, 1, 10, or  100 ug/mL of
       PFOS for 24 hours; labeled with 6-carboxyfluorescein succinimidyl ester (CFSE);
       stimulated with concanavalinA,  a T-cell mitogen (ConA,  5 ug/mL to half of all samples);

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-83

-------
       and incubated for an additional 72 hours. Proliferation was slightly increased at
       100 ug/mL and slightly reduced with the presence of ConA, but neither effect was
       statistically-significant.
   3.  For the NK cell assays, PBMCs were incubated with 0, 1,  10, or 100 ug/mL of PFOS for
       24 hours followed by incubation for 3 hours with K562 target cells (12.5:1 ratio) labeled
       with CFSE. K562 cells are a chronic myelogenous leukemia cell line known to be
       susceptible to NK cell induced cytotoxicity. PFOS significantly (p < 0.001) reduced NK
       cell cytotoxicity to K562 cells by 32% at 100 ug/mL.
   4.  In the monocyte differentiation assay, HL-60 cells were incubated with 0, 1, 10, or
       100 ug/mL of PFOS for 72 hours. Half of each sample was stimulated with 25 nmol
       calcitrol, la,25-dehydroxyvitamin D3 (1,25D3) 24 hours into the incubation period.
       Expression of CD1 Ib and CD14 were measured as markers of differentiation. In the
       presence of 1,25D3, PFOS had no significant effect on the percentage of HL-60 cells
       expressing CD1 Ib and CD14. No differences in monocyte differentiation were observed
       in the absence of 1,25D3.
   5.  Whole blood was incubated with 0-100 ug/mL of PFOS in the presence or absence of
       25 ug/mL phytohemagglutinin (PHA), a T-cell cytokine secretion  stimulator, for
       48 hours. Lipopolysaccharide (LPS, 0 or 250 ng/mL), a monocyte  stimulator, was added
       to whole blood incubated with 0.1-100 ug/mL of PFOS either 4 or 24 hours prior to the
       end of the 48 hour incubation period. Release of the cytokines TNF-a and IL-6 from
       T-cells or monocytes was quantified. Cytokine release from T-cells was not affected by
       PFOS. PFOS significantly (p < 0.001) reduced the release of the pro-inflammatory
       cytokine TNF-a after monocyte LPS stimulation. The authors also looked at the
       correlation between basal PFOS concentration of the blood donor and cytokine release. A
       significant association was observed between PFOS concentration and the release of
       LPS-induced IL-6 by peripheral monocytes.

   This study suggests some effects on immunity in humans; however the sample size used is
small and the concentrations at which effects were observed are much higher than the levels of
PFOS in human blood samples.

   Midgett et al. (2014) examined the effects on IL-2 production using stimulated cultured
human Jurkat cells and CD4+ T cells recovered from 11 healthy volunteers. Both cell types were
stimulated with PHA/phorbal myristate acetate (PMA) or anti-CD3 to produce IL-2 and
incubated with 0-100 ug PFOS/mL; separate experiments were conducted with human Jurkat
cells in the presence  or absence of a PPAR antagonist.  Cell viability was not affected in either
cell type up to and including the highest concentration of PFOS. In the human Jurkat cells
stimulated with PHA/PMA a concentration of 10 ug/mL was a NOEL and 50 ug/mL a LOEL for
inhibition of IL-2 production in the absence and presence of a PPARa inhibitor. In the presence
of anti-CD3, the NOEL was 1 ug/mL and the LOEL 5  ug/mL. In primary  human CD4+  T cells
stimulated with PHA/PMA, the NOEL was the 10 ug/mL concentration and the LOEL
100 ug/mL for inhibition of IL-2 production. A decrease in T cell IL-2 production is a
characteristic associated with autoimmune disorders, suggesting that this population could be
sensitive  to PFOS exposures. However, the authors caution that the results from the in vitro
studies do not reflect any potential decrease in circulating PFOS as the result of protein binding
to albumin or other serum proteins. In this study the observed IL-2 effects in the Jurkat cells
were demonstrated to be independent of PPARa activation as the inhibition was similar with and
without the PPAR antagonist.

Perfluorooctane sulfonate (PFOS) - May 2016                                            3-84

-------
Mouse. 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 4-6 six male (6-8 weeks old) C57B1/6 mice/group. Doses for all dietary levels were
not presented by the study authors. 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 of 10 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 measured and checked for viability. Histology was also
performed on the thymus and spleen.

   The 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 since it is also observed when PFOS is 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), and they had thymocytes and splenocytes that were 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.

   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 PFAS exposure influenced innate immune response to
bacterial LPS. Mice were exposed to PFOS as described above. On day 10, some mice were
injected intravenously with 0.1 mL sterile saline containing 300  jig LPS (E. coli), while others
received 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 administration of LPS. The spleen, thymus,
epididymal fat, liver, and peritoneal and bone marrow cells were collected.

   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
(0.6 mg/kg/day), and the total concentration of PFOS in the serum was 340 ±16 |ig/mL (ppm).
The overall total number  of white blood cells and lymphocytes were decreased while the
neutrophil counts were similar to controls. The number of macrophages in the bone marrow was
increased but not those of the peritoneum and spleen. 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
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-85

-------
an effect on the innate immune responses in mice following a 10-day exposure to about
0.6 mg/kg/day.

   In Qazi et al. (2010), 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. 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, and
hypertrophic hepatocytes surrounded the central vein. No necrosis was noted. Total serum
cholesterol was decreased and there was a moderate increase in serum ALP. At the end of the
study, the total mean serum PFOS concentration for four mice was 125.8 |ig/mL. PFOS
increased only one type of intrahepatic immune cells (TER119+). 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. The IgM response of the intrahepatic B and T
cells was normal.

   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 PFOS
doses to the seven dose groups 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, NK cell activity, lysozyme activity,
antigen specific IgM production, lymphocyte immunophenotypes, and serum PFOS
concentrations were determined after exposure.

   Survival, behavior, body weight, spleen, thymus, kidney, gonad and liver weights, and
lymphocytic proliferation were not affected by treatment. Lysozyme activity increased
significantly in females, but not males, at 0.0033 and 0.166 mg/kg/day, respectively compared to
the control group; however, the response as not dose-related. NK 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 they 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.

   Because IgM suppression can result from effects on both T- and B-cells, antibody production
was measured in response to sheep red blood cells (SRBC) (T-dependent) and a trinitrophenyl
(TNP) LPS conjugate (T-independent). The SRBC plaque-forming response 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%. For
evaluation of T-independent (TI) responses, an additional group of female mice was treated with
0 or 0.334 mg/kg/day of PFOS orally for 21 days and challenged with a TI antigen TNP-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.
Perfluorooctane sulfonate (PFOS) - May 2016                                            3-86

-------
   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 weeks
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 sacrificed. The blood was analyzed for corticosterone and PFOS
concentration.  Spleen, thymus, liver, and kidneys were collected and weighed, and the spleen
and thymus were processed into suspensions to look at functional immune endpoints and T-cell
immunophenotype determinations.

   Starting on about day 3, mean body weights were significantly decreased compared to the
controls for the 20 and 40 mg/kg bw/day doses. However, food consumption decreased with
treatment. 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. Liver weight was 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; levels in the controls were
below the limit of detection. Serum corticosterone levels increased significantly in mice treated
with doses > 20 mg/kg/day. Splenic and thymic cellularity were 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 was 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 release assay was performed to determine NK cell activity. The
average NK-cell activity was decreased at 20 and 40 mg/kg/day compared to control mice,
18.04± 1.42 and 13.08± 1.11, respectively compared to 50.33 ± 4.08 in controls. No numeric
data were provided for the 5 mg/kg/day group. 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, 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/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 0, 0.008, 0.083, 0.417, 0.833, and 2.083 mg/kg/day PFOS with 2% Tween 80
in de-ionized water daily by gavage for 60 days (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 dose groups compared to controls, 5.17 ± 0.12 g
(control), 5.21  ± 0.17 g, 5.78 ± 0.13 g, 6.67 ± 0.11 g, 8.17 ± 0.21 g,  and 11.47 ± 0.12 g,
respectively. Serum corticosterone was decreased in mice 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/kg/day (45.43 ± 4.74%) with significant marked decreases at 0.833 mg/kg/day

Perfluorooctane sulfonate (PFOS) - May 2016                                            3-87

-------
(20.28 ± 2.51%) and 2.083 mg/kg/day (15.67 ± 1.52%). The SRBC-specific IgM plaque forming
cell response showed a dose-related decrease with statistical significance at 0.083 mg/kg/day and
higher. 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.

   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 of PFOS in 0.5% Tween-20 by gavage
daily on gestation days (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 for
dams during the study and pups after delivery. Organ weights (spleen, liver, thymus and uterus)
from the pups were recorded at sacrifice. Only litters with 6 to 9 pups were retained for the
immunotoxicity studies. One male and one female were selected from the retained litters (total of
6 male and 6 female pups) for testing of the immunotoxicity parameters; positive  controls were
included for each assay.

   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 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 parameters evaluated, 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.

   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) 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.

   The four studies in mice discussed above examined NK cell activity and SRBC response.
The results from those studies are summarized in Table 3-19. Three of the studies showed effects
on SRBC response, NK cell activity, or both at the  same dose that caused increased liver weight.
Based on the limited evidence, neither response appeared more sensitive than the  other. The NK
cell activity was enhanced at very low PFOS doses, while it was depressed at higher doses. The
animal studies indicate that females are less susceptible to impacts on NK cell activity and the
SRBC response than males.
Perfluorooctane sulfonate (PFOS) - May 2016                                            3-8

-------
    Table 3-19. Summary of SRBC and NK Cell Findings in Mice after PFOS Exposure
Study
Dong et
al. (2009)
Keil et al.
(2008)
Peden-
Adams et
al. (2008)
Zheng et
al. (2009)
Strain
C57BL/6
(M)
B6C3FJ
(M, F pups)
B6C3F1
(M,F)
C56BL/6
(M)
Duration
Days
60
CDs 1-17
Dams
only*
28
7
SRBC
NOAEL
mg/kg/day
0.008
1(M)
5(F)
0.00017 (M)
0.0033 (F)

LOAEL
mg/kg/day
0.083 (|)
SUM)
0.0017(|M)
0.017 (|M)
5(4)
NK Cell activity
NOAEL
mg/kg/day
0.008
1(F)
0.0033 (M)
0.166(F)
5
LOAEL
mg/kg/day
0.083 (t)
0.833 (4)
1(|M)
5 (IF)
0.017 (|M)
20(4)
Increased
Liver wt.
LOAEL
mg/kg/day
0.083
5 (M at 4
wks only)
None
5
Notes: Direct dosing of the dams did not continue during the lactation period. The immune system response was evaluated in
pups at 4 and 8 weeks. Effects were seen at 8 weeks but not at 4 weeks.
The direct ion of the arrow indicates if the change from control was an increase or a decrease.
M = male; F = female

3.3.4   Physiological or Mechanistic Studies of 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 (~ decrease of 47-80%) and free T4 (~ decrease of 60-82%). The total T3 was only
significantly deceased after day 5 (decrease of- 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 caused hepatomegaly, hepatocellular hypertrophy, and
macrovesicular steatosis. Genes associated with the thyroid hormone release and synthesis
pathway included type 3 deiodinase DIO3, which  catalyzes the inactivation of T3 and type  1
deiodinase DIO1, which deiodinates prohormone T4 to bioactivate T3. Treatment with PFOS
caused significant (p < 0.05) DIO1 repression and Dio3 induction only on day 5.

   Chang et al. (2007) investigated whether the decrease of 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 jimol 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. However,  FT4 was significantly
decreased in the animals when measured with two analog methods, chemiluminescence
immunoassay and simple 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.
Perfluorooctane sulfonate (PFOS) - May 2016
3-89

-------
   Chang et al. (2008) investigated whether PFOS competed with thyroxine for serum binding
proteins in rats. Three different experimental designs were employed. 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 were
obtained. Serum FT4, total thyroxine (TT4), triiodothyronine (TT3), reverse triiodothyronine
(rT3), and thyrotropin were measured at each timepoint. TSH was measured only at the 6 and 24
hour timepoints. PFOS concentrations in the blood and liver were also measured along with
hepatic transcripts for UDP-glucuronosyltransferase 1A (UGT1 A) (involved in glucuronidation
and T4 turnover) and malic enzyme (ME). ME activity is an indicator for tissue response to
thyroid hormone.

   Serum TT4 decreased significantly (p < 0.05) compared to controls after 2 hours (decrease of
24%), 6 hours (decrease of 38%), and 24 hours (decrease of 53%). The TT3 and rT3 only
decreased significantly at the 24-hour 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 (control: < LOQ; treated: 37.28, 66.90, and 61.58 ng/mL at 2, 6, and 24 hours,
respectively). A similar trend was observed with the concentration of PFOS in the liver (control:
< LOQ;  treated: 30.60, 44.84, 45.00 |ig/g at 2, 6, and 24 hours, respectively). 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,  but 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 PFOS
administration  does not induce a typical hypothyroid state or alter the hypothalamic-pituitary-
thyroid axis.
Perfluorooctane sulfonate (PFOS) - May 2016                                              3-90

-------
   In the study by Curran et al. (2008) (see section 3.2.2 of this document) 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.

   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 whether 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), FT4, total triiodothyronine (T3), and TSH. Liver and thyroid
organ weights were obtained as well. Also measured were messenger RNA (mRNA) levels for
two isoforms  of uridine diphosphoglucuronosyl transferase (UGT1A6 and UGT1A1) and 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 weight) 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 3-20 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 3-20. Thyroid Hormone Levels in PFOS Treated Rats
Dose administered
mg/L
0
1.7
5.0
15.0
Total T3
(Mg/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.
Perfluorooctane sulfonate (PFOS) - May 2016
3-91

-------
    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), 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 (section 2.2.1), Yu et al. (2011) found that liver OATp2 was
increased significantly (143% compared to controls) in rats at 3.0 mg/kg, indicating that this
transporter may be involved in hepatic T4 uptake and could potentially lead to the decrease
observed in serum TT3 and TT4. Relative liver weight and 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.

    Ren et al. (2015) examined the comparative agonist and antagonist properties of the PFCs as
revealed using a T3 cell proliferation assay in GH2 cancer cells. Antagonist activity was
measured using cell proliferation response in the presence of 0.2 nmol T3 and the PFAS. PFOS
had the strongest potency  as an agonist among the PFAS compounds tested but was still less
potent than T3. PFOS also upregulated three thyroid hormone response genes and downregulated
another three, one of those being the fatty acid binding protein gene in tadpoles. Molecular
docking analysis was used to examine the mode of interaction between the PFOS and the TRa-
ligand binding domain protein. PFOS and T3 both hydrogen bonded with Arg-228, with the
PFOS sulfonate functional group facing into the pocket and the perfluorinated carbon chain
oriented towards the exterior of the pocket.

    Kjeldsen and Bonefeld-J0rgensen (2013) conducted an in vitro study in an attempt to
elucidate the mechanisms by which  PFAS, including PFOS, affect the estrogen receptor (ER)
and androgen receptor (AR) transactivity, as well as aromatase activity. Estrogenic and
antiestrogenic activities were assessed using the stably transfected  MVLN cell line  carrying an
estrogen response element luciferase reporter vector. Androgenic and antiandrogenic activities
were assessed using the Chinese hamster ovary cell line  CHO-K1 transiently co-transfected with
an MMTV-LUC  reporter vector and an AR expression plasmid pSVARO. Effects on aromatase
activity were assessed using the human choriocarcinoma JEG-3 cell line. PFOS had no effect on
aromatase activity, but it was cytotoxic at > 1 * 10"4 M.

    In the ER transactivation assay, PFOS was cytotoxic to MVLN cells at concentrations
> 6 x  1Q-5M. The half maximal effective concentration (ECso) for PFOS was 2.9 x  10'5 M
compared  with 4.8 x 10'11 M for l?p-estradiol (E2). Co-exposure of cells with E2 and PFOS
enhanced the E2-induced ER response at the highest non-cytotoxic PFOS concentration. No
evidence of antagonism was observed.

    In the AR transactivation assay,  PFOS was cytotoxic to CHO-K1 cells at concentrations
> 1  x  10"4M. PFOS did not act as an agonist, however, it elicited a significant (p < 0.05)
inhibiting  effect (76%) on AR function at a relative high test concentration of 5 x 10"5 M.
Co-exposure of cells with dihydrotestosterone and PFOS elicited a significant (p < 0.05)
concentration-dependent antagonistic effect on DHT-induced AR transactivity; the  ICso was
4.7  x  1Q-6M.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-92

-------
PPAR activity

    Studies have been conducted in order to determine if PFOS activates 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 whether it activated human or mouse
PPARa in a COS-1 cell-based luciferase reporter rram'-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  1000 jimol were tested. The COS-1 cells were transfected with either a 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 cotransfected 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- to 6-fold and was similar to that
obtained with the positive control. The average ECso was 13 jimol in the mouse and 15 jimol in
the human PPARa.

    Both PFOS and PFOA were tested to determine whether they could activate PPARs in 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 (imol), 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-14,643 and MK-886, respectively, for PPARa, and
troglitazone and GW9662, respectively, for PPARy. Only the agonist LI65,041 was used for
PPAR.p/5. After treatment for 24 hours, activity was measured using the Luciferase reporter
assay. WY-14,643 was used for mouse and human PPARa, and it exhibited 15- and 1-fold
increase,  respectively over the luciferase response of the negative controls.  L165,041 was the
agonist for mouse and human PPAR.p/5. It 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.

    In this study, PFOS activated the mouse PPARa with a significant (p <  0.01) 1.5-fold
increase in activity at 120 (imol PFOS, compared to the negative control. PFOS did not
significantly increase activity in the human PPARa construct. PFOS activated the mouse
PPAR.p/5 but not the human PPAR.p/5 construct. It did not activate the mouse or human PPARy
construct. Table 3-21 shows summary data. The authors concluded that PFOA activated PPARa
more than PFOS and that 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                            3-93

-------
  Table 3-21. Summary of PFAS Transactivation of Mouse and Human PPARa, P/5, and y
PPAR isoform
a
P/5
Y
PFAS
PFOA
PFOS
PFOA
PFOS
PFOA
PFOS
Mouse LOEC3
10 jimol
120 nmol
40 jimol
20 nmol
NA
NA
Human LOECa
30 jimol
NAb
NA
NA
NA
NA
Source: Data from Table 1 in Takacs and Abbott 2007
Notes: a LOEC = lowest observed effect concentration; lowest concentration (|imol) at which there was a significant difference
compared to the negative control (p < 0.05)
°NA = not activated

   Wolf et al. (2008) tested PFAS, including PFOS, to determine whether 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% DMSO), a positive control
(WY-14,643), or PFOS at 1-250 |imol. At the end of 24-hours of exposure, the luciferase
activity was measured. The no observed effect concentration (NOEC) for PFOS was 60 jimol in
the mouse; the LOEC was 90 jimol (48.4 |ig/mL), and the C20max was 94 jimol. The
corresponding values for humans were: NOEC = 20 jimol, LOEC = 30 jimol (16.2 jig/mL), and
C20max = 262 jimol.

   Wolf et al. (2012) incubated transfected cells with PFAS at concentrations of 0.5 to
100 jimol, vehicle (water or 0.1% DMSO as negative control), or with 10 jimol WY-14,643
(positive control). Assays were performed with three identical plates per compound per species,
with nine concentrations per plate and eight wells per 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  PFAS significantly induced human and mouse PPARa. The
study also provided the C20max, which was the concentration at which a PFAS produced 20% of
the maximal response elicited by the most active PFAS. For PFOS, this was 94 jimol in mouse
PPARa and 262 jimol in human PPARa. For comparison, PFOA was 6 jimol and 7 jimol,
respectively.

    Several studies have suggested that PFOS may target PPARy and influence metabolism via
pathways under its control. L. Zhang  et al. (2014) examined the direct binding properties of
PFOS and other PFASs using the ligand binding domain of human PPARy. Interactions between
transfected B 1.21 (DE3) E. coli supported derivation of ICso values for the different PFAS
examined. The ICso values were derived using a fluorescence displacement method and
comparing the results from the tested  chemicals with those of decanoic and octanoic acid. The
PFAS binding increased with carbon  chain length (C4 to Cl 1). The authors also examined the
PFAS binding to the PPARy ligand binding domain. For compounds with fewer than 11 carbons
there was a correlation between binding and chain length. The authors interpreted this as an
indication  that hydrophobic interactions between the amino acids of the binding domain and the
PFAS are responsible for the stability of the  complex. PFOA and PFOS induced receptor
activation to a similar extent, 2.8 and  2.5 times greater than the control, respectively. The authors
concluded that PFASs induce disruption of lipid homeostasis and inflammation by the PPARy
pathway as well as the PPARa pathway. Among the three members of the sulfonate family tested
(4, 6, and 8 carbons), PFOS displayed the strongest activation potency.
Perfluorooctane sulfonate (PFOS) - May 2016
3-94

-------
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 1, 3, or 5 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. 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 DIO1 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. DIO1 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 constitutive androstane receptor (CAR).

   Wang et al. (2010) dosed albino Wistar female rats with 3.2 mg PFOS/kg diet from GD 1  to
weaning (PND 21). Pups were allowed access to the treated feed until PND 35. To determine if
prenatal or lactational exposure had more effect on altering gene expression, pups were divided
into cross fostering groups on PND 2. These groups are listed below:

   •   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.
   •   Pups born to treated dams fostered by other treated dams.

   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, WY-14,648 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 WY-14,648, or 500 ppm PB  in the
diet 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 WY-14,643 and PB, behaving as a combined peroxisome proliferator
andphenobarbital-like enzyme inducer. The data suggested that PFOS may activate not only
PPARa, but also CAR and PXR.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-95

-------
Mice. To assess PPAR involvement in developmental effects of PFOS, male and female
12981/Svlm wild-type 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 wild-type 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 PNDs 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.

   Reproductive parameters measured included maternal body weight, maternal body weight
gain, 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 wild-type mice. PFOS exposure had no effect on absolute or
relative liver weight in any of the dams. In both strains of pups, PFOS exposure at 10.5
mg/kg/day caused a significant increase in relative liver weight (sexes were combined). Survival
of the pups was affected with treatment. Most post-natal deaths occurred between PNDs 1 and 2.
Survival of the wild-type 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 wild-type and KO pups. On PND 13, 44% of
the control pups and none in the 8.5 mg/kg/day wild-type group had experienced their eye
opening. In the KO mice, open eyes were reported in 23% of the 10.5 mg/kg pups and 59% of
the controls on PND 14. 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
wild-type 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 mechanistic 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 for 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 two more fetuses/litter for histological examination.

   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. At 5
mg/kg/day, 753 fully annotated genes were altered in the fetal liver. In the fetal liver, PFOS
upregulated a number of markers for fatty acid metabolism, xenobiotic metabolism, peroxisome
biogenesis, cholesterol biosynthesis, bile acid biosynthesis, and glucose and glycogen metabolism.
In the fetal lungs, up-regulation only occurred in a limited group of genes including: Cyp4al4,
enoyl-Coenzyme A hydratase (Ehhadh), and fatty acid binding protein 1 (FABP1).

   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 weights were both decreased in the wild-type mice

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-96

-------
only. The wild-type mice administered 0.02% PFOS in the diet had a pronounced decrease in the
total number of thymocytes (by 84%) and splenocytes (by 43%), as well as a decrease in the size
of all subpopulations of thymocytes and splenocytes. In the knock-out mice, the reduction in the
total number of thymocytes and subpopulations 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 activation.

    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 7 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 across the dose groups. This finding suggests that there are PPARa-independent effects
in null mice that also occur in wild-type mice. Thus, some of the liver effects in the wild-type
animals are not necessarily a reflection of PPARa activation.

    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, Cyp2bJO, 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 3-1). Unique in null mice, PFOS upregulated
Cyp7al, an important gene related to bile acid/cholesterol homeostasis. The results support those
from other studies that indicate PFOS exposure results in metabolic changes both linked to, and
independent of, PPAR-a.

    The variability in the serum lipid profiles in humans suggests that response to PFOS
exposure could be impacted by individual physiological differences and that environmental
factors such as diet could contribute to intraspecies differences in response. L. Wang et al. (2014)
reported on the differences in response of male BALB/c mice (4-5 weeks old) administered
PFOS (0, 5, or 20 mg/kg) for 14 days while concurrently given diets that varied in dietary  fat
[regular fat (RF)  versus high fat (HF)] content. Following PFOS exposure, there was an increase
in liver fat content in both groups and a decrease in liver glycogen. However, the increase  in fat
content was more pronounced with the RF mice than in the mice on the HF diet. This study is
described in more detail in section 3.2.2.
    The fat content of the diet alone was associated with significantly higher serum levels  of
glucose, HDL cholesterol, LDL cholesterol, and total cholesterol. For glucose, cholesterol, HDL,
and LDL the levels declined as the dose of PFOS  increased. In the case of triglycerides, the
levels increased with 5 mg/kg/day PFOS and decreased at 20 mg/kg/day. PPARa expression at
the end of the 14-day PFOS treatment increased for the RF group but decreased for the HF
groups (significant for the high dose).
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-97

-------
                             2-  —
                              E   E
c  s.
       tu
       OH
       DC
   O   I
E
s
M(
la
E
                                             8
                                             I
                                             »
                                                Fatty acid metabolism


                                                Proteasome activation

                                                Peroxisome biogenesis


                                                Inflammatory response




                                                Xenobiotic metabolism


                                                Cholesterol biosynthesis


                                                Oxidative phosphorylation
                                                Ribosome biogenesis
                                         Fold change
                          2  1.5  I  -1.5 -2

Figure 3-1. Functional Categories of Genes Modified by PFOS in Wild-Type and Null Mice

    The high fat diet alone increased the expression of CPT1A and CYP7A1 genes involved with
lipid metabolism. On the RF diet, the exposure to PFOS was associated with a significant dose-
related increase in CTP1A expression, whereas for the high fat diet plus PFOS there was a
significant decrease in expression. For CYP7A1 expression there was no significant impact of
PFOS with the RF diet, whereas with the high fat diet there was a highly significant decrease in
expression with PFOS.

    The L. Wang et al. (2014) study demonstrates a clear influence of diet alone on the liver and
lipid profile that was combined with some dose-related differences in the responses to PFOS
exposure. The data support a possible role for PFOS in inhibiting pathways for metabolism and
export of liver lipids and identify some PFOS associated liver responses that are independent of
PPARa activation.

    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 interperitoneal (ip) injection for 7 days. Twenty-four hours after the last dose,
Perfluorooctane sulfonate (PFOS) - May 2016
                                                     3-98

-------
the animals were 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 1,502 unique proteins; 71 showed a greater than 1.5-fold change in expression.
Sixty-two proteins showed increased expression, and nine showed decreased expression. Figure
3-2 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 3-2 shows that the majority of
enriched proteins were involved with lipid metabolism, transport, biosynthetic processes,
catabolic processes, and carbohydrate metabolic processes.
                                                       l-'vpxl
                               MacmrreilccuLir cttmplex

                                          l>   5   III  15   2(1  25  VI
                                    rX.'U'L
-------
   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), mRNA transcript data for the control
and 1.0 mg/kg/day dose group (GD 20 dams and fetuses and PND 21  male pups) was obtained
by quantitative reverse transcription polymerase chain reaction. Results for this part of the study
were reported by Chang et al. (2009). Statistically-significant changes included:

   •   Increased Cyp2b2 levels in dams and male pups (2.8-fold and 1.8-fold, respectively) than
       in controls on GD 20 and PND 21.
   •   Higher mean acyl CoA (ACoA) and Cyp4al levels in male pups (1.5-fold and 2.1-fold,
       respectively) than those of controls.
   •   Lower mean Cyp7al (3.5-fold) than that for controls.

These results suggest induction of PPARa as well as hepatic CAR. Transcripts possibly related
to thyroid status were all similar between the treated dams and pups and the controls.

Oxidative Damage

   Liu et al. (2009) conducted a study of KM mice in which 3-6 pups/sex/group were
administered one subcutaneous injection of 0 or 50 mg PFOS/kg bw on 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 MDA content, SOD activity, and total antioxidation
capability (T-AOC). Animals were sacrificed 24 hours post-treatment, blood was collected, and
liver and brain were removed and weighed.

   No treatment-related effects were observed on body weight. Relative 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 (decrease of 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 (decrease of 15%). Male liver T-AOC was
decreased significantly at PND 7 (decrease of 25%) and 14 (decrease of 27%). Female brain
T-AOC had no significant differences from controls and the liver T-AOC was decreased only at
PND 21 (decrease of 15%). The study also demonstrated that distribution increased in the liver
and lessened in the blood and brain with postnatal 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. PFOS brain levels were about
5-fold higher on PND 7 than they were on PND 28. 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.
Perfluorooctane sulfonate (PFOS) - May 2016                                           3-100

-------
Gap Junctional Intercellular Communication (GJIC)

    Gap junctions are found in the cell's plasma membrane and formed by proteins that connect
to 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 whether 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) and a dolphin kidney epithelial cell line (CDK) to
PFOS at concentrations of 0, 3.1, 6.25, 12.5, 50, 100, or 160 |imol for 30 minutes. GJIC effects
were measured using the scrape loading dye technique.  PFOS inhibited GJIC rapidly in a dose-
dependent method starting at 12.5 jimol, but it was reversible once exposure ended.  Additionally,
4 to 6 Sprague-Dawley 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, but it was comparable to that seen after 3 days. No
differences were observed between the male and female rats.

    Wan et al.  (2014a) isolated  and cultured Sertoli cells from testes of 20-day old rats to
examine PFOS's effects on blood testes barrier function. By day 3 the cultures had established a
functional tight junction barrier. Gap junction communication was assayed by means of
fluorescence recovery using a photo bleaching assay that measured the ability of a fluorescent
dye to move from one cell to another in the  presence or  absence of PFOS (20 jimol; a 3-hour
exposure) in a 120 second period. Cells treated with PFOS displayed significantly lower
fluorescence recovery than the  control cells in the absence of cytotoxicity. The assays were
performed in triplicate. The authors identified this as a matter of concern because it represents
diminished function of the blood testes barrier in coordinating an intercellular junction necessary
for intercellular communication during spermatogenesis. The authors also examined other
characteristics of the blood testes barrier finding effects of PFOS on the cytoskeleton manifest in
the form of shortened F-actin filaments.

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.

    Wallace et al. (2013) also examined the impact of 16 different PFASs on mitochondrial
respiration rate and oxidative phosphorylation as measured in vitro using isolated rat liver
mitochondria.  Inhibition was determined through the reduction in oxygen consumption in
response to the addition of ADP to isolated  mitochondria. PFOS displayed a 20-30 times more
potent inhibitory effect among the other sulfonates evaluated (PFBS and PFHxS). PFOS was
three times more potent than PFOA. The inhibition mode of action seemed to vary across
different PFAS families. In the  case of PFOS, its impact on membrane fluidity appeared to be


Perfluorooctane sulfonate (PFOS) - May 2016                                             3-101

-------
responsible for the observed respiratory inhibition. The authors' proposed mode of action for this
effect from PFOS is consistent with the findings of Matyszewska et al. (2008) that PFOS
increased the membrane fluidity and thickness of a model biological membrane in a manner
similar to that resulting from cholesterol insertion into a lipid by-layer.

3.3.5   Structure-Activity Relationship

In vitro. Bjork and Wallace (2009) performed a study to see whether 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 jimol PFAS for 24
hours to determine the structure-activity relationships across various chain lengths. The
concentration used was the maximum concentration that did not lead to cell injury in any of the
cell lines. The PFAS 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 PFAS 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 PFASs in HepG2/C3A cells.
However, this gene expression was stimulated by PFAS 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 PFAS, 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 PFAS 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 was no 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.

3.3.6   ToxCast Assays

   The Toxicity Forecaster (ToxCast) database is a large high throughput screening compilation
of public in vitro and in vivo assays on over 9,000 chemicals. PFOS was tested in  1,087 assays
and was active in 175. Assay activation defines the occurrence of the molecular event within the
assay (cytotoxicity, induction, binding, and so forth.) with the concentration resulting in 50%
activity, ACso, used  for comparison to other assays. Assays with < 50% reported efficacy  or
over-fitting issues are not included in the results discussed. Some of the data from the ToxCast
assays such as the interactions with PPAR and CAR support the experimental data for PFOS and
PFOA. In cases where effects were only observed at concentrations greater than those causing
cytotoxicity, attributing the outcome to PFOS rather than the cytotoxicity is less certain.

Cytotoxicity. Of the active assays, 20 were examining endpoints based on cytotoxicity. Most
cell types offered at  least one cytotoxicity ACso for comparison to other in vitro assays. If no
cytotoxicity ACso was reported for a specific cell type, the minimum in vitro cytotoxicity

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-102

-------
endpoint for PFOS was used for comparison. The lowest PFOS induced ACso recorded in the
ToxCast database is 5.34 uM in the assay for induction of tumor protein 53 using liver cells and
the highest ACso recorded is 172.02 uM for measuring tumor protein in intestinal cells
(SD = 45.15; standard error = 9.9).

Endocrine Disruption. Four different estrogen receptor (ESR) assays reported activation
following PFOS treatment, all of which were Estrogen Receptor 1- (ESR1-) related. Estrogen
and its receptors are essential for sexual development and reproductive function, but they also
play a role in other tissues, such as bone. Estrogen receptors are also involved in pathological
processes including breast cancer, endometrial cancer, and osteoporosis (NCBI2016). Two
assays of the same cell type were related to ESR1 induction with the lowest ACso, 18.06 uM.
This is lower than the cytotoxicity ACso reported for the cell type used, 23.76 uM, and is
indicative of ESR1 induction. In a different ESR1 assay, antagonism was recorded at an AC50 of
83.57 uM, a value higher than the cytotoxicity AC50 for that cell type, 66.31 uM. PFOS induced
the estrogen DNA binding site with an ACso of 87.42 uM. There was no cell-specific reference
cytotoxicity value for comparison. The ToxCast assays suggest that PFOS has the ability to
induce ESR1.

   PFOS antagonized the androgen receptor (AR) at 12.57 uM in human cells and 4.27 uM in
rats. Although there is no direct cellular cytotoxicity value to compare, PFOS rat AR antagonism
occurred at lower concentrations than the minimum cytotoxicity value (5.34 uM). This implies
that PFOS reacts with the AR receptor in the rat and perhaps the human. The progesterone
receptor (PR) was also antagonized within the same human cell type, and had a higher ACso,
22.2 uM, than the minimum cytotoxicity ACso. Thyroid receptor (TR) antagonism ACso,
91.24 uM, was also higher than its respective cell specific cytotoxicity ACso (33.323 uM). This
signifies that assay activation (i.e., positive results) might have occurred due to cytotoxicity
rather than PR, TR, or human AR antagonism. However, PFOS-induced ESR1 and antagonized
rat AR was observed.

Immunotoxicity. PFOS activated a variety of genes related to immunotoxicity in the  ToxCast
database. These genes include: chemokine ligand (CXCL) 10, CXCL8, collagen type  II alpha
(COL3A), interleukin-1  alpha (IL-la), plasminogen activator (PLA), plasminogen activator
urokinase (PLAUR), vascular cell adhesion molecule (VCAM1), and the TNF receptor
subfamily gene CD40 (CD40). All of the immunological assays were performed by the vendor
BioSeek. The vendor did not have a cytotoxicity ACso for every cell type utilized and used only
two cytotoxicity ACso values for comparison. Genes that had lower ACso values than cell or
BioSeek specific cytotoxicity ACso were: CD40, PLAUR, PLA, VCAM1, and CXCL8. Given
the limited cytotoxicity reference values it is difficult to determine if all gene activity  can be
attributed to PFOS. For PLAUR and VCAM1, ACso results were lower than their cell specific
cytotoxicity ACso and have stronger translational potential. VCAM1 and PLAUR play a role in
inducing chronic inflammation and vascularization in vivo (Kleinstreuer et al. 2014). This
implies PFOS may play a role in inducing chronic inflammation and/or vascularization, both of
which are important for the development of rheumatoid arthritis (Khansari et al. 2009).

Neurotoxicity (in vitro). PFOS activated five different neurological receptor families with seven
different receptor types in cell-based assays. The receptors activated were: 5-hydroxytryptamine
receptor (5HT) 5 a, 6, and 7, the adenosine A2a receptor (ADORA2), the adrenoceptor alpha 2C
(ADRA2C), and beta 1 (ADRB1), as well as the dopamine receptor D4 (DRD4). Cell-specific
cytotoxicity ACso values for reference were lacking for all of the in vitro assays; only ADORA2
had an activation ACso lower than the lowest PFOS cytotoxicity ACso of 5.34 uM. Therefore, it

Perfluorooctane sulfonate (PFOS) - May 2016                                           3-103

-------
is difficult to draw any conclusions on the potential neurotoxicity of PFOS using the ToxCast
data.

Fish Toxicity (in vivo). Oregon State University conducted a large number of toxicity studies
using a zebrafish model. PFOS gave positive results in nine assays. Positive effects were
recorded for limb malformations 5 days after a 1-time exposure during embryonic development.
Other assays with positive results were those for Axis Malformation, Jaw Malformation,
Pericaradiac Edema, Snout Malformation, Touch Response, Trunk Malformation, Yolk Sac
Edema, and Mortality. Mortality had the lowest reported ACso at 0.54 jimol. The results provide
strong evidence for PFOS developmental toxicity in fish and suggest a potential for human
developmental human toxicity.

PPAR/PXR/RAR Receptors. PFOS activated PPARs, PXR, constitutive adrostane receptor
(CAR), and retinoic acid receptor (RAR) in assays conducted under the ToxCast program. PFOS
induced the DNA sequences for PPAR alpha (PPARa), peroxisome proliferator hormone
response elements (PPRE), and PPAR gamma (PPARy) and antagonized the PPARy receptor.
The only PPAR assay ACso that was above the cell-specific cytotoxicity ACso was PPARy
antagonism at 5.91  uM. However, it is possible that cytotoxicity occurs due to PPAR induction
or that the PPAR antagonism contributes to cytotoxicity. PFOS induced DNA sequences for
PXR, ACso 9.42 uM, at concentrations lower than the cell-specific cytotoxicity ACso. CAR and
RAR alpha antagonism were also observed but not at levels below the cell specific cytotoxicity
values of 17.57 uM and 28.45 uM respectively. PPAR, PXR, CAR, and RAR pathways are all
nuclear receptors that can form heterodimers with one another to induce translation of linked
genes. Some of these genes are important for development, reproduction, waste degradation,  and
even induction of cytotoxicity. Therefore, PFOS induction of these assays are consistent with the
experimental data on PPAR and CAR receptor activation.

Cytochrome P450s Activation. Cytochrome P450 (CYP) enzyme bindings were also examined
within the ToxCast database in order to understand any metabolic potential for  a chemical.
Though PFOS is not known to be metabolically active, it showed activation in four acceptable
CYP assays: CYP2C18, CYP2C19, and CYP2C9 in human cells, and CYP2C11 in rat. All of the
CYP assays had activation at concentrations lower than the lowest cytotoxic ACso. The CYP2C
class is known to be involved in the metabolism of xenobiotics including drugs, such as the anti-
seizure medication diazepam, the beta blocker propranolol, and the selective serotonergic
reuptake inhibitor citalopram. Though there is no evidence of metabolism of PFOS by these
CYPs, it is possibly acting as a competitive or allosteric inhibitor for other substrates. This,
coupled with PFOS's high affinity for albumin, could significantly alter the pharmacokinetics of
various necessary and habitual pharmaceuticals.

3.4   Hazard Characterization

3.4.1   Synthesis and Evaluation of Major Noncancer Effects

3.4.1.1  Liver Effects, Cholesterol, and Uric Acid

   Good correlation between serum and  hepatic levels of PFOS has been shown for human
samples (Karrman et al. 2010; Olsen et al. 2003a). However, no consistent adverse effects on the
liver were found in epidemiology studies. Biomonitoring studies performed at the 3M Decatur,
Alabama plant (Olsen et al. 2001a, 2001b, 2003b) identified occasional differences in hepatic


Perfluorooctane sulfonate (PFOS) - May 2016                                           3-104

-------
clinical chemistry values but no associated hepatic disease and or hepatic carcinogenicity was
reported.

   Multiple epidemiologic studies have evaluated serum lipid status in association with PFOS
concentration. These studies provide support for an association between PFOS and small
increases in total cholesterol. Hypercholesterolemia, which is clinically defined as cholesterol
greater than 240 mg/dL, was associated with PFOS exposure in a Canadian cohort (Fisher et al.
2013) and in the C8 cohort (Steenland et al. 2009). PFOS levels in these studies were
0.0084 |ig/mL and 0.022 |ig/mL, respectively. Cross-sectional occupational studies demonstrated
an association between PFOS and total cholesterol (Olsen et al. 2001a, 2001b, 2003b), with
much higher PFOS serum levels of up to 1.40 jig/mL. Evidence for associations between other
serum lipids and PFOS is mixed, with some studies showing an association with measurements
of concurrent HDL and/or LDL and others failing to measure the serum lipoprotien complexes.
The studies on serum lipids in association with PFOS serum concentrations are largely cross-
sectional in nature and were largely conducted in adults, but some studies exist on children and
pregnant females. The location of these cohorts varied from the U.S. population including
NHANES volunteers, to the Avon cohort in the UK, to Scandanivian countries. Limitations to
these studies include the frequently high correlation between PFOA and PFOS exposure; not all
studies control for PFOA in  study design.

   There are several characteristics of HDLs that explain the association of increased serum
concentration of HDL with PFOS or PFOA levels. HDLs accept cholesterol from other serum
lipoprotein complexes and bring it to the liver for degradation and conversion to bile salts
(Montgomery et al. 1990). Competition between PFOS and bile salts for biliary transport could
result in impeded  removal of HDL lipids from serum and increase both HDL  cholesterol and
total cholesterol. The liver is the only tissue that can rid the body of excess cholesterol by
secreting it in bile for removal with the feces (Montgomery et al.  1990). In addition, HDLs have
the highest ratio of protein to lipid (50:50)  among the serum lipoprotein complexes
(Montgomery et al. 1990). Binding of PFOS to HDL protein could impede the HDL interaction
with liver tissue receptors resulting in increased serum levels  of HDL. LDLs contain 21%
protein. LDL uptake by tissues is mediated by binding of the LDL apo-B-100 protein and by a
receptor independent route (Montgomery et al. 1990). Thus, conformational changes in the
lipoprotein proteins as a result of PFOS binding can also impact serum LDL levels.

   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 (Curran et al. 2008; Goldenthal et al.  1978a; Liu et al. 2009; Seacat et al. 2002;
Thomford 2002/Butenhoff et al. 2012). Chang et al. (2009) identified PFOS levels in the liver of
rat offspring as early as GD 20, and Stein et al. (2012) measured PFOS in human 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/Butenhoff et al. 2012). Microscopic 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
0.098-0.299 mg/kg/day after 104 weeks (Thomford 2002/Butenhoff et al.  2012), 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. Single cell necrosis

Perfluorooctane sulfonate (PFOS) - May 2016                                             3-105

-------
was also found in rats at 0.984-1.251 mg/kg/day after 104 weeks (Thomford 2002/Butenhoff et
al. 2012; Table 3-15). In these studies, no evidence of peroxisome proliferation was found in
either species.

   Hepatomegaly and increased liver weight alone are not considered adverse in cases where a
chemical such as PFOA causes stimulation of PPAR-a, CAR, and/or PXR cellular receptors
(Hall et al. 2012). However, when accompanied by necrosis (Thomford 2002/Butenhoff et al.
2012) and/or fatty acid steatosis (Bijland et al. 2011; Wan et al. 2012), liver weight increases are
considered adverse.

   In contrast with humans, rats, mice, and monkeys displayed a decrease in cholesterol levels
and high density lipoprotein cholesterol following PFOS administration in short and long term
studies (Curran et al. 2008; Seacat et al. 2003; L. Wang et al. 2014) when compared to the
controls. Male rats had decreased serum cholesterol after 14 weeks at a dose of about
1.4 mg/kg/day (Curran et al. 2008). Wan et al. (2012) found evidence for hepatic macrovesicular
steatosis in mice given > 5 mg/kg/day that was not entirely due to PPARa activation. Steatosis
was exacerbated when PFOS exposure was combined with a high fat diet.

   As discussed above in section 3.3.4, mice administered PFOS showed differential expression
of genes or proteins mainly involved in lipid metabolism, transport, biosynthetic processes, and
response to stimulus (Tan et al. 2012; L. Wang et al.  2014) 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 Bijland et al. (2011) as described
below. In addition, the nuclear hormone receptors CAR and PXR have been shown to be
activated in mice (Bijland 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, Bijland 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 3.2.2. 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; and VLDL synthesis and secretion.
Genes involved with HDL synthesis, maturation, clearance, and bile acid formation were
downregulated. 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.

   No animal studies were identified that examined  serum uric acid levels following PFOS
exposures.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-106

-------
3.4.1.2  Developmental/Reproductive Toxicity

   PFOS has been detected in amniotic fluid samples indicating that the chemical crosses the
placenta. The median ratio of maternal serum:amniotic fluid concentration was 25.5:1, and PFOS
was rarely detected in amniotic fluid until the serum concentration reached at least 0.0055 |ig/mL
(Stein et al. 2012). Studies evaluating the reproductive and developmental health in humans
exposed to PFOS have been performed in both occupational settings and in the general
population.

   Although not entirely consistent, the set of studies evaluating fetal growth retardation suggest
an association of prenatal serum PFOS with deficits in mean birth weight and with LEW.
Although three studies were null (Fei et al. 2008a; Hamm et al. 2010; Monroy et al. 2008), birth
weight deficits ranging from 29 to 149 grams were detected in 5 studies (Apelberg et al. 2007;
Chen et al. 2015; Darrow et al. 2013; Maisonet et al. 2012; Washino et al. 2009). In these
studies, PFOS serum levels ranged from 0.005 to 0.0132 |ig/mL. Three (Chen et al. 2012; Fei et
al. 2007; Stein et al. 2009) out of four (Darrow et al. 2014) studies of LEW showed increased
risks (OR range: 1.5-4.8). Studies have questioned whether low maternal GFR is a positive
confounder in epidemiology studies of birth weight and PFAS (Morken et al. 2014; Verner et al.
2015).  The Verner et al. (2015) comparison between a meta-analysis and PBPK simulations
suggests that the some but not all of the association reported between PFOS and birth weight
could be attributable to low GFR.

   A small set of studies observed an association with gestational diabetes (Zhang et al. 2015,
preconception serum PFOS), pre-eclampsia (Stein et al. 2009), and pregnancy-induced
hypertension (Darrow et al. 2013) in populations with serum PFOS concentrations of 0.012-
0.017 ug/mL. Zhang et al. (2015), and Darrow et al. (2013) used a prospective assessment of
adverse pregnancy outcomes in relation to PFAS which addresses some of the limitations the
available cross-sectional studies. Associations with these outcomes and  serum PFOA also were
observed.

   There also is generally consistent evidence of associations of serum  PFOS with decreased
fertility and fecundity (Bach et al. 2015; Fei et al. 2009; J0rgensen et al. 2014; Velez et al. 2015);
there was one null study (Vestergaard et al. 2012). While a concern over the possibility of
reverse causation explaining observed associations has been raised (Whitworth et al. 2012), the
collective findings, particularly from a more recent study (Bach et al. 2015), support a consistent
association with fertility and fecundity measures and PFOS exposures. Although there was some
suggestion of an association between PFOS exposures and semen quality parameters in a few
studies (Joensen et al. 2009; Toft et  al.  2012), most studies were largely null (Buck Louis et al.
2015; Ding et al. 2013; Joensen et al. 2013; Raymer et al. 2012; Specht  et al. 2012; Vested et al.
2013).

   No animal studies were identified that suggested effects on fertility in males or females.
However, Lopez-Doval et al. (2014) found structural effects on the hypothalamic-pituitary axis
in adult male rats after exposure to PFOS. There were histopathological lesions of the testes at
doses > 1 mg/kg/day and only a few active gonadotrophic cells at doses > 3 mg/kg/day. The
lowest  dose tested, 0.5 mg/kg/day, was accompanied by decreased LH and testosterone levels
and increased FSH levels.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-107

-------
   Increased pup mortality was observed when rat dams were treated only during gestation as
part of developmental toxicity studies (Chen et al. 2012; Lau et al. 2003; Thibodeaux 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; Lau et al. 2003; Thibodeaux 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, 2005b). 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
is 0.1 mg/kg/day in the two-generation study (Luebker et al. 2005b); this dose was not tested in
the one-generation  study (Luebker et al. 2005a) where the LOAEL was 0.4 mg/kg/day for
decreased pup body weight, decreased maternal body weight, and decreased gestation length. A
0.4 mg/kg/day  dose was a LOAEL in the both the one and two generation studies. The dose
associated with a significant decrease in pup survival for the two generation study was
1.6 mg/kg/day  and the dose for a decreased viability index was 0.8 mg/kg/day (BMDLs =
0.89 mg/kg/day) in the one-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 4-day period during various stages of pregnancy. Mortality was highest when treatment
occurred on GDs 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 GDs 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-exposed newborn pups,
suggesting that lung maturation and pulmonary surfactant interactions are potential MO As.
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, 2010b). 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
chain fatty acids and may be responsible for interfering with the normal physiological function of
pulmonary surfactant.

   Prenatal PFOS exposures appear to influence hormones during gestation, as well as in
neonate and adult animals. Zhao et al. (2014) examined the testes from male Sprague-Dawley rat
fetuses on GD 20 following maternal exposure to 0, 5, or 20 mg/kg/day on GDs 11-19.  Fetal

Perfluorooctane sulfonate (PFOS) - May 2016                                            3-108

-------
Ley dig cells were found to be reduced in number with evidence of apoptosis. Levels of
testosterone were reduced along with the levels of key enzymes or mRNA for proteins involved
with testosterone production.

   Studies have examined the impact of gestational and lactational exposures on the pups as
adults (Lv et al. 2013 rats; Wan et al. 2014b, mice). In both cases early life exposure through
maternal treatment with PFOS was associated with abnormal glucose control in the mature
offspring. In both cases, serum glucose was significantly higher in the adult animals exposed
during gestation and lactation than in controls and there was evidence of insulin resistance. The
LOAEL was 0.5 mg/kg/day  in the Lv et al. (2013) study and 3 mg/kg/day for pups fed a diet
with normal fat levels (Wan et al. 2014b). In the Wan et al. (2014b) study, the NOAEL was
0.3 mg/kg/day. When accompanied by a high fat diet, 0.3 mg/kg/day was a LOAEL for increased
insulin resistance.

3.4.1.3  Immunotoxicity

   A few studies have evaluated associations with measures indicating immunosuppression.
Two studies reported decreases in response to one or more vaccines (diphtheria, rubella) in
children aged 3, 5, and 7 years (e.g., measured by antibody titer) in relation to increasing
maternal serum PFOS levels (maternal levels ranging from 0.0056 to 0.027 |ig/mL) during
pregnancy or in the children at 5 years of age (mean child 0.0167 jig/mL) (Grandjean et al. 2012;
Granum et al. 2013). Decreased rubella and mumps antibody concentrations in relation to serum
PFOS concentration were found among 12-19 year old children in the NHANES, particularly
among seropositive children (Stein et al. 2015). A third study of adults found no associations
with antibody response to influenza vaccine (Looker et al. 2014). In the three studies examining
exposures in the background range among children (i.e., general population exposures, geometric
means < 0.02 jig/mL), the associations with PFOS were also seen with other correlated PFAS,
complicating conclusions specifically for PFOS.

   No clear associations were reported between prenatal PFOS exposure and incidence of
infectious disease among children (Fei et al. 2010b; Okada et al. 2012), although there might be
effect modification by sex. With regard to other immune dysfunction, serum PFOS levels were
not associated with risk of ever having had asthma among children in the NHANES with median
levels of 0.017 |ig/mL (Humblet et al. 2014). A study among children in Taiwan with higher
serum PFOS concentrations  (median with and without asthma 0.0339 and 0.0289  jig/mL,
respectively) found higher odds ratios for physician-diagnosed asthma with increasing serum
PFOS quartile (Dong et al. 2013). Associations also were found for other PFASs.  Among
asthmatics, serum PFOS was also associated with higher severity scores, serum total IgE,
absolute eosinophil counts, and eosinophilic cationic protein levels.

   Other data on the immunotoxicity of PFOS in humans are limited to in vitro studies using
cells recovered from human  blood (PBMCs; Brieger  et al.  2011 or CD4+ T cells; Midgett et al.
2014). In both cases the concentration of PFOS with  a demonstrated significant effect was
100 |ig/mL, and the concentration that lacked  any effects was 10 |ig/mL. A significant
(p < 0.001) decrease in TNFa and a nonsignificant trend towards increasing IL-6 release from
stimulated monocytes were seen, but no effects were measured on stimulated T cells (Brieger et
al. 2011). T cell IL-2 production was decreased in the Midgett et al. (2014) study.
Perfluorooctane sulfonate (PFOS) - May 2016                                            3-109

-------
    Studies in mice examined NK cell activity and SRBC response following oral dosing with
PFOS. Three of four studies showed effects on SRBC response and/or NK cell activity at the
same dose that caused increased liver weight (Dong et al. 2009; Keil et al. 2008; Zheng et al.
2009). Based on the limited evidence, neither response appeared more sensitive than the other.
The animal studies indicate that females are less susceptible to impacts on NK cell activity and
the SRBC response than males.

    The NK cell activity was enhanced at very low PFOS doses, while it was depressed at higher
doses. Peden-Adams et al. (2008) and Dong et al. (2009) showed increased NK cell activity in
male mice following exposure to 0.0017 mg/kg/day and 0.083 mg/kg/day, respectively. The
increased activity in Dong et al. (2009) correlated with a PFOS serum level of approximately
7.1 |ig/mL. In the Dong et al. (2009) study, the NK cell activity was significantly decreased at a
higher dose of 0.833 mg/kg/day, demonstrating a U-shaped response to dose. Doses
> 1 mg/kg/day resulted in decreased NK cell  activity in offspring of dams treated during
gestation (Keil et al. 2008)  and in adult male  mice (Zheng et al. 2009).

    In the Peden-Adams et  al. (2008)  study, IgM suppression occurred after 28 days of treatment
with 0.0017 mg/kg/day 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. Other studies also showed a suppression
of the SRBC response at higher doses of PFOS (Dong et al. 2009; Keil et al. 2008; Zheng et al.
2009). Guruge et al. (2009) found a decrease in survival in mice exposed to 0.025 mg/kg of
PFOS after exposure to influenza A virus.

    Qazi et al. (2009a) reported that approximately 40 mg/kg/day 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 size of the those present in wild-type
mice. Knock-out mice had  a reduction in the  total number of thymocytes that was 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. At the
same dose, Qazi et al. (2009b) did not find elevated levels in serum or spleen of TNF-a and IL-6
in  response to stimulation by LPS in C57B1/6 mice, but levels were increased in the cells from
the peritoneal cavity and bone marrow

3.4.1.4  Neurotoxicity

    Developmental neurotoxicity and adult neurotoxicity studies have been conducted in  rats and
mice. Mechanistic studies have examined effects on excitatory amino acids and gene profiles
following PFOS exposures.

    Butenhoff et al. (2009)  found significantly increased motor activity and decreased
habituation of male offspring at one time point (PND 17) following gestational and lactational
dosing of dams with 1.0 mg/kg/day. No effects were found on learning and memory with the
Biel swimming maze. Luebker et al. (2005b) found no effects on passive avoidance behavior or
water maze learning and memory in Fl offspring at a daily dose of 0.4 mg/kg/day. Y. Wang et
al. (2015) used water maze testing on offspring from treated dams who were cross-fostered with
either control or treated dams, and continued on the treatment of their lactational dam. Escape
latency was significantly increased for all treated groups on one or more testing days with the
most pronounced effect in pups exposed prenatally from dams given 15 mg/L drinking water and

Perfluorooctane sulfonate (PFOS) - May 2016                                            3-110

-------
cross-fostered to control dams. Y. Wang et al. (2015) did not provide data on water intake or
body weight data. A drinking water concentration of 5 mg/L was a NOAEL, and a concentration
of 15 mg/L was a LOAEL for offspring. Estimated adult doses are 0.8 and 2.4 mg/kg/day,
respectively, using the subchronic USEPA (1988) conventions for water intake and body weight.
Long et al. (2013) found a significantly longer latency to escape, with significantly less time
spent in the target quadrant in the Morris water maze test for learning and memory at a dose of
2.5 mg/kg/day in 8-week-old C57BL6 mice. The NOAEL for these effects was 0.43 mg/kg/day.

   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). Wang et al. (2010) 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. Zeng et al. (2011) also found PFOS administered to pregnant rats as low as 0.1 mg/kg
from GD 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 of PFOS when they were 10 days old (Johansson et al. 2008)
displayed abnormal habituation responses in motor activity testing. Cultured hippocampal
neurite growth and branching were suppressed by exposure to 50 jimol PFOS. The authors
hypothesized that this was a consequence of PFOS incorporation into the neuronal lipid  bilayer
membrane (Liao et al. 2009). The effect of PFOS was greater than that of PFOA. PFOS  was the
only member of the sulfonate family to exhibit this effect.

3.4.1.5  Thyroid Effects

   Numerous epidemiologic studies have evaluated thyroid hormone levels and/or thyroid
disease in association with serum PFOS concentrations. These epidemiologic studies provide
limited support for an association between PFOS exposure and incidence or prevalence of
thyroid disease, and include large studies of representative samples of the general U.S. adult
population (Melzer et al. 2010; Wen et al. 2013). These highly powered studies reported
associations between PFOS exposure (serum PFOS concentrations) and thyroid disease but not
thyroid hormone status. Melzer et al. (2010) studied thyroid disease with medication (PFOS level
of 0.025 |ig/mL in males and 0.019 |ig/mL in females) and Wen et al.  (2013) studied subclinical
thyroid disease (mean serum 0.0142 jig/mL). Thyroid function can be affected by iodide
sufficiency  and by autoimmune disease. People testing positive for the anti-TPO biomarker for
autoimmune thyroid disease showed associations with PFOS (0.0048 jig/mL) and TSH or T4
(Webster et al. 2014); this association was stronger in people with both low iodide status and
positive anti-TPO antibodies, with a PFOS level of 0.014 |ig/mL (Webster et al. 2015). These
studies used anti-TPO antibody levels  as an indication of stress to the thyroid system, not a
disease state. Thus, the association between PFOS and altered thyroid hormone levels is stronger
in people at risk for iodine deficiency than those receiving  adequate dietary iodine. In people
without diagnosed thyroid disease or without biomarkers of thyroid disease, thyroid hormones
(TSH, T3, or T4) show mixed effects across cohorts.

   Several animal models have described changes in thyroid hormone levels after administration
of PFOS. In contrast to the human epidemiology studies, the most consistent finding in animals
treated with PFOS was a decrease in T4 with slight, or no,  changes in T3. Any changes found in
T3 and T4 levels failed to activate the hypothalamic-pituitary-thyroid (HPT) feedback
mechanism to produce significant elevations of serum TSH.
Perfluorooctane sulfonate (PFOS) - May 2016                                            3-111

-------
   Rats treated orally with PFOS for 1-5 days had significant decreases in total T4 at doses of
10 and 15 mg/kg, but not at 5 mg/kg (Chang et al. 2007, 2008; Martin et al. 2007). With
treatment for 7 days, total T4 was decreased at 1 and 3 mg/kg (Yu et al. 2011).

   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 neither a dose-response nor evidence of
hypothyroidism was observed. TSH levels were variable during the study, but increased 2-fold in
the high-dose males at the end of exposure.

   Exposure of pregnant rats to PFOS at 1  mg/kg/day, which corresponded to maternal serum
concentrations of 14-26 ug/mL, resulted in decreases in T4 and T3 in dams (Thibodeaux et al.
2003) and decreased T4 in pups (Lau et al. 2003). No effect was observed on serum TSH. In
contrast, no effects were found on thyroid hormones in either dams or pups when females were
treated prior to mating and through LD 4 (Luebker et al. 2005a). 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 (Chang et
al. 2009).

   In addition to the evaluation of PFOS's effects on serum TT4, several studies have examined
the levels of circulating FT4 (Lau et al. 2003; Luebker et al. 2005a; Thibodeaux et al. 2003; 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 alterations in thyroid hormones 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. Similarly, Ren et al. (2015) demonstrated that PFOS bound to the ligand
binding domain of the human thyroid hormone receptor, although with a much lower affinity
thanT3.

   Several possibilities might account for the differential  findings of thyroid hormone disruption
between animal models and human biomonitoring data. First, decreased T3 or T4 was observed
in adult monkeys and rodents only when serum PFOS reached the 70-90 ug/mL 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 general
populations of humans is estimated to be 0.018-0.037 |ig/mL,  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


Perfluorooctane sulfonate (PFOS) - May 2016                                            3-112

-------
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. PFOS has been shown to have much a
lower binding affinity for both TTR and the thyroid hormone receptor than do T4 and T3,
respectively (Ren et al. 2015; Weiss et al. 2009).

3.4.2   Synthesis and Evaluation of Carcinogenic Effects

   The small set of epidemiology studies of PFOS exposure do not suggest that there is an
association with cancer, but the breadth and scope of the studies are not adequate to make
definitive conclusions. While an elevated risk of bladder cancer mortality was associated with
PFOS exposure in an occupational study (Alexander et al. 2003),  a subsequent study to ascertain
cancer incidence in the cohort observed elevated but statistically insignificant incidence ratios
that were 1.7- to 2-fold higher among exposed workers (Alexander and Olsen 2007). Mean
PFOS serum levels were 0.941 jig/mL. No elevated bladder cancer risk was observed  in a nested
case-control study in a Danish cohort with plasma PFOS concentrations at enrollment ranging
0.001-0.1305 |ig/mL (Eriksen et al. 2009).

   Elevated odds ratios for prostate cancer were reported for the  occupational cohort  examined
by Alexander and Olsen (2007) and the Danish population-based  cohort examined by  Eriksen et
al. (2009), however the confidence intervals included the null, and no association was reported
by another case-control study in Denmark (Hardell et al. 2014). A case-control study of breast
cancer among Inuit females in Greenland with similar serum PFOS levels to those of the Danish
population (0.0015-0.172 |ig/mL) reported an association of low  magnitude that could not be
separated from other perfluorsulfonated acids, and the association was not confirmed in a Danish
population (Bonefeld-J0rgensen et al. 2011, 2014). Some studies  evaluated associations with
serum PFOS concentration at the time of cancer diagnosis and the impact of this potential
exposure misclassification on the estimated risks is unknown (Bonefeld-J0rgensen et al. 2011;
Hardell et al. 2014). No associations were adjusted for other perfluorinated chemicals in serum in
any of the occupational and population-based studies.

   The only chronic toxicity/carcinogenicity study in animals was a rat study (Thomford
2002/Butenhoff et al. 2012). Increased incidence of hepatocellular adenomas in the male (12% at
the high dose) and female rats (8% at the high dose) and combined adenomas/carcinomas in the
females (10% at the high dose) were observed, but they did not display a clear dose-related
response. In males but not females the serum ALT levels were increased at 14, 27, and 53 weeks.
At 105 weeks there was an increase in eosinophilic clear cell foci, and cystic hepatocellular
degeneration in males given 2, 5, and 20 ppm PFOS. Low levels of single cell necrosis in all
dose groups (males and females) were identified; the increase compared to controls was
significant at the high dose in males and females (Table 3-15).

   Thyroid tumors (adenomas and carcinomas) were seen in males receiving 0, 0.5, 2, 5, or
20 ppm and in females receiving 5 or 20 ppm in their diet.  The tumor (adenomas + carcinomas)
prevalence for males was consistent across dose groups.  In males the incidence of thyroid tumors
was significantly elevated only in the high-dose, recovery group males exposed for 52 weeks
(10/39) but not in the animals receiving the same dose at 105 weeks. There were very  few
follicular cell acenomas/carcinomas in the females (5 total) with no dose-response. The most
frequent thyroid tumor type in the females was C-cell adenomas, but the highest incidence was
that for the  controls and there was a lack of dose response among the exposed groups. C-cell
adenomas were not observed in males (Thomford 2002/Butenhoff et al. 2012).


Perfluorooctane sulfonate (PFOS) - May 2016                                            3-113

-------
   There was a high background incidence in mammary gland tumors in the female rats,
primarily combined fibroma adenoma and adenoma, but the incidence lacked dose-response for
all tumor classifications. Mammary gland carcinomas also lacked dose-response and had a
relatively comparable incidence across dose groups including the controls (Thomford
2002/Butenhoff et al. 2012).

   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.

3.4.3   Mode of Action and Implications in Cancer Assessment

   Short-term genotoxicity assays suggested that PFOS is not a DNA-reactive compound. The
results from five in vitro studies (Cifone 1999; Litton Bionetics, Inc. 1979; Mecchi 1999; Murli
1999; Simmon 1978) were negative, as was the result from an in vivo bone marrow micronucleus
assay (Murli 1996).

   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, mice, or both (Ashby et al. 1994; Rao and Reddy 1996). The liver-
expressed peroxisome PPARa regulates the transcription of genes involved in peroxisome
proliferation, cell cycle control, apoptosis, and lipid metabolism. The data for PFOS illustrate the
ability of PFOS to activate PPARa (Martin et al. 2007; Shipley et al. 2004; Wolf et al. 2008,
2012). However, data are generally lacking for increased cell proliferation. No increase in
hepatic cell proliferation was detected in the subchronic study (Seacat et al. 2003) or the cancer
bioassay (Thomford 2002/Butenhoff et al. 2012); limited necrosis was observed across all doses
and significantly (p < 0.05) increased for the 20 ppm males and females. In addition, no
subchronic or longer term studies revealed evidence of preneoplastic foci in the liver. Liu et al.
(2009) studied biomarkers for oxidative stress in the liver and brain in KD mice. 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 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). These are not clearly defined MO As, and
their importance relative to PFOS exposure is not certain. Ngo et al. (2014) used the mouse
model C57BL/6J -Min/+ for intestinal neoplasia to determine effects following in utero
exposure. Maternal treatment with PFOS at doses up to 0.3 mg/kg/day during gestation did not
result in an increase of intestinal tumors in either wildtype or susceptible offspring up to 20
weeks of age.

3.4.4   Weight of Evidence Evaluation for Carcinogenicity

   Under the EPA Guidelines for Carcinogen Risk Assessment (USEPA 2005a) there is
suggestive evidence of carcinogenic potential of PFOS in humans. A single chronic cancer
bioassay in animals is available for PFOS. Although liver adenomas were  significantly increased
in males and females at the highest dose and a positive trend was observed (p = 0.03), a dose-
response pattern was not observed. In males the incidence of thyroid follicular tumors was


Perfluorooctane sulfonate (PFOS) - May 2016                                            3-114

-------
elevated only in the high-dose, recovery group 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. Based on the
available evidence, the data are inadequate to support a PPARa-linked MOA for the liver and
thyroid adenomas observed by Thomford (2002)/Butenhoff et al. (2012) in the chronic 2 year
bioassay in Sprague-Dawley Crl:CD(SD)IGS BR rats.

3.4.5   Potentially Sensitive Populations

   In humans, single blood samplings of different populations within the United States do not
support major gender differences in half-life or sensitivity to PFOS. Gender differences could not
be determined by those  exposed by occupational exposure, as the majority of those tested were
males. Serum monitoring among the NHANES populations (2004-2008) found  significantly
(p < 0.05) higher PFOS  levels in males  (0.020 ug/L) than females (0.014 ug/L).  However, this
difference is more likely to be related to exposures than to sensitivity.

   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. In animal studies of immunological effects, the response to NK cell suppression
occurred at a lower dose in males than in females (Peden-Adams et al. 2008).

   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.  2005a, 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 epidemiology 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.

   The animal data on LEW receive support from the epidemiolgy (Apelberg et al. 2007; Chen
et al. 2015; Darrow et al. 2013;  Maisonet  et al. 2012; Washino et al. 2009). For humans with low
GFR (females with pregnancy-induced hypertension or preeclampsia in late pregnancy), the
impact on body weight is likely due to a combination of the low GFR and the serum PFOS
(Verner et al. 2015). Low GFR  in pregnant females will tend to cause an increase in serum PFOS
compared to individuals with a  normal GFR. Females with hypertension during  pregnancy could
have an increased risk for having a LEW baby.

   The fat content of the diet appears to be an important variable that influences the effects from
PFOS exposures. Elevated total cholesterol, HDL, and sometimes triglycerides are effects seen
in a number of the human epidemiology studies. However, none of the studies evaluated
appeared to control for fat content in the typical diet of the subjects. Martin et al. (2007), Bijland
et al. (2011), and Wan et al. (2012) found hepatic steatosis in PFOS-treated animals. Liver fat
increased with both a high fat diet alone and with a high fat diet  plus PFOS (Wan et al. 2012). In
the same  study, significant increases in the expression of fatty acid translocase and lipoprotein
lipase was observed at the 10 mg/kg/day PFOS dose. Mobilization of liver lipids appeared to

Perfluorooctane sulfonate  (PFOS)  - May 2016                                            3-115

-------
decrease following the PFOS exposure leading to lower serum LDL/VLDL levels; VLDLs are
carriers of liver triglycerides and other lipids from liver to serum.

   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 4-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 (GDs) 17-20,
identifying late gestation as a critical exposure window for increasing the risk of neonatal
survival. The effects of PFOS at this stage of development could be related to an impact of PFOS
on lung surfactants leading to respiratory  distress syndrome. Both Luebker et al. (2005a) and Lau
et al. (2003) identified pup mortality as adverse effects of gestational PFOS exposures.
Perfluorooctane sulfonate (PFOS) - May 2016                                             3-116

-------
4.  DOSE-RESPONSE ASSESSMENT

   A Reference Dose (RfD) or Reference Concentration (RfC) is used as a benchmark for the
prevention of long-term toxic effects other than carcinogenicity. RfD/RfC determination assumes
that thresholds exist for toxic effects, such as cellular necrosis, significant body or organ weight
changes, blood disorders, and so on. 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.

4.1   Dose-Response for Noncancer Effects

4.1.1   RfD Determination

Human Data. In humans, data have been obtained from studies evaluating both occupational
and general population exposure scenarios. Some studies monitored similar populations over
time to determine whether or not a trend was present. Pathways of exposure in the general
population appear to be from drinking water, food (especially fish/seafood), and some
environmental exposures (e.g., carpets, house dust). In general, PFOS levels in the serum of the
general population have decreased since production was stopped in the United States.

   Multiple epidemiology  studies evaluated serum lipid status in association with PFOS
concentration. These  studies provide support for an association between PFOS and small
increases in total cholesterol. Hypercholesterolemia, (clinically defined as cholesterol
> 240 mg/dL) was associated with PFOS exposure in a Canadian cohort (Fisher et al. 2013) and
in the C8 cohort (Steenland et al. 2009); PFOS levels in these studies were 0.0084 ng/mL and
0.022 |ig/mL, respectively. Cross-sectional occupational studies demonstrated an association
between PFOS and total cholesterol (Olsen et al. 2001a, 2001b, 2003b), with much higher PFOS
serum levels of up to  1.40 jig/mL. Evidence for associations between PFOS and other serum
lipids including HDL cholesterol, LDL, VLDL, non-HDL cholesterol, and triglycerides is mixed.

   The studies on serum lipids in association with PFOS serum concentrations are largely cross-
sectional in nature and were largely conducted in adults. Some studies exist on children and
pregnant females. The location of these cohorts varied from the U.S. population including
NHANES volunteers, to the Avon cohort in the UK, to and Scandanivian countries. Limitations
to these studies include the frequently high correlation between PFOA and PFOS exposure; not
all studies control for PFOA in study design.

   Studies that evaluated thyroid hormone levels and/or thyroid disease in association with
serum PFOS concentrations include large representative samples of the general U.S. adult
population and provide limited support for an association between PFOS exposure and the
incidence or prevalence of thyroid disease. PFOS levels in Melzer  et al. (2010) were
0.025 |ig/mL in males and 0.019 |ig/mL in females, and in Wen et al. (2013) they were
0.0142 |ig/mL. Pregnant females testing positive for the anti-TPO biomarker for autoimmune
thyroid disease showed a positive association with PFOS (0.0048 jig/mL) and TSH (Webster et
al. 2014). In a second study, Webster et al. (2015) found an association with PFOS
(0.014 |ig/mL) and TSH and T3 in a subset of the NHANES population with both low iodide
status and  positive anti-TPO antibodies. Anti-TPO antibody levels are an indication of stress to

Perfluorooctane sulfonate (PFOS) - May 2016                                              4-1

-------
the thyroid system, not a disease state. Thus, the association between PFOS and altered thyroid
hormone levels is stronger in people at risk for thyroid insufficiency or disease. In people
without diagnosed thyroid disease or without biomarkers of thyroid disease, thyroid hormones
(TSH, T3, or T4) show mixed effects across cohorts.

   A few studies evaluated associations with measures of immunosuppression. Two studies
reported decreases in response to one or more vaccines (diphtheria, rubella) in children aged 3, 5,
and 7 years (e.g., measured by antibody titer) in relation to increasing maternal serum PFOS
levels (ranging 0.0056-0.027 jig/mL) during pregnancy or at 5 years of age (Grandjean et al.
2012; Granum et al. 2013). Decreased rubella and mumps antibody concentrations in relation to
serum PFOS concentration were found among 12-19 year old children in the NHANES,
particularly among seropositive children (Stein et al. 2015). A study of adults found no
associations with antibody response to influenza vaccine (Looker et al. 2014). In the three studies
examining exposures in the background range among  children (i.e., general population
exposures, geometric means < 0.02 |ig/ml), the associations with PFOS were also correlated with
other PFASs, complicating conclusions as they applied to PFOS.

   No clear associations were reported between prenatal PFOS exposure and incidence of
infectious disease among children (Fei et al. 2010b; Okada et al. 2012), although an elevation in
risk of hospitalizations for infectious disease was found among girls, suggesting effect
modification by sex. PFOS levels were not associated with risk of ever having had asthma
among children in the NHANES with median levels of 0.017 |ig/mL (Humblet et al. 2014). A
study among children in Taiwan with higher serum PFOS concentrations (median with and
without asthma 0.0339 and 0.0289 |ig/mL, respectively) found higher odds ratios for physician-
diagnosed asthma with increasing serum PFOS quartile (Dong et al. 2013). Associations  with
other PFASs were also positive. Among asthmatics, serum PFOS was associated with higher
severity scores, serum total IgE, absolute eosinophil counts, and eosinophilic cationic protein
levels.

   The set of studies evaluating fetal growth retardation suggest an association of prenatal
serum PFOS with deficits in mean birth weight and with LEW, however it is not entirely
consistent. Birth weight deficits ranging from 29 to 149 grams were detected in five studies
(Apelberg et al. 2007; Chen et al. 2015; Darrow et al.  2013; Maisonet et al. 2012; Washino et al.
2009). In these studies, PFOS serum levels ranged from 0.005 to 0.0132 |ig/mL. Three (Chen et
al. 2012; Fei et al. 2007; Stein et al. 2009) out of four  (Darrow et al. 2014) studies of LEW
showed increased risks (OR range: 1.5-4.8). Studies have questioned whether low maternal GFR
is a confounder in epidemiology studies of birth weight and PFOS (Morken et al. 2014; Verner et
al. 2015). The Verner et al. (2015) study compared the results from a meta-analysis of the
epidemiology data with PBPK simulations and concluded that the some, but not all, of the
association reported between PFOS and birth weight is attributable to low GFR. Thus, the
interpretation of the observed associations is unclear.

   A small set of studies observed an association with gestational diabetes (Zhang et al.  2015,
preconception serum PFOS), pre-eclampsia (Stein et al.  2009), and pregnancy-induced
hypertension (Darrow et al. 2013) in populations with serum PFOS concentrations of 0.012 -
0.017 ug/mL. Zhang et al. (2015) and Darrow et al. (2013) used a prospective assessment of
adverse pregnancy outcomes in relation to serum PFAS thereby avoiding some of the limitations
of the available cross-sectional studies. Associations with serum PFOA and adverse pregnancy
outcome were identified.
Perfluorooctane sulfonate (PFOS) - May 2016                                              4-2

-------
   There is consistent evidence of associations of serum PFOS with decreased fertility and
fecundity (Bach et al. 2015; Fei et al. 2009; J0rgensen et al. 2014; Velez et al. 2015). While a
concern over the possibility of reverse causation explaining observed associations has been
raised (Whitworth et al. 2012), the collective findings, particularly from a more recent study
(Bach et al. 2015), support a consistent association with fertility and fecundity measures and
PFOS exposures. Although there was some suggestion of an association between PFOS
exposures and semen quality parameters in a few studies (Joensen et al. 2009; Toft et al. 2012),
most studies were largely null (Buck Louis et al. 2015; Ding et al. 2013; Joensen et al. 2013;
Raymer et al. 2012;  Specht et al. 2012; Vested et al. 2013).

Animal Data. Adequate studies were available for short-term, subchronic, chronic,
developmental, and reproductive parameters in rats, mice, and primates. Subchronic, chronic,
and reproductive toxicity animal studies, all with exposure duration greater than 60 days, have
been summarized in Table 4-1. Shorter duration studies that focused on immunotoxicity
endpoints and developmental toxicity studies are summarized in Table 4-2. Although the
exposure durations are shorter in developmental studies, they are important in quantification of
dose-response because the exposures occur during critical windows of development and are often
symptomatic of effects that can occur later in life. It is noted, however, that in some of these
studies, steady states of PFOS might not have been achieved due to the long half-life of PFOS in
animal models (see discussion of steady state in section 4.1.1.1)

   Seacat et al. (2002) treated monkeys with PFOS 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
PCoAO activity, was increased significantly in the females at 0.75 mg/kg/day; however, the
magnitude was less than the 2-fold increase typically indicating biological significance and
PPARa activation. There were no treatment-related effects on cell proliferation in the liver,
pancreas, or testes; survival was decreased among the males. At the dose with no effects
observed (0.15 mg/kg/day), the serum concentration was 83 |ig/mL in males and 67 |ig/mL in
females. At the effect level (0.75 mg/kg/day), the serum concentrations were 173 |ig/mL in
males and 171 |ig/mL in females, about twice those for the no-effect serum level despite a 5-fold
increase in dose.

   Microscopic 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 0.098 mg/kg/day in males
and 0.299 mg/kg/day in females after 105 weeks (Thomford 2002/Butenhoff et al. 2012). Liver
lesions included centrilobular hypertrophy and vacuolation after the subchronic and chronic
exposures with eosinophilic granules observed after 104 weeks. No evidence of peroxisome
proliferation was found during either phase of the study. Mean no effect levels in males and
females were 0.34 mg/kg/day and 0.40 mg/kg/day, respectively, after 14 weeks and 0.024 mg/
kg/day and 0.120 mg/kg/day, respectively, after 104 weeks.

   Rat dams were treated with PFOS for 63 or 84 days in one- and two-generation reproductive
studies, respectively (Luebker et al. 2005a, 2005b). 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 study. A NOAEL for pup body weight effects
was 0.1 mg/kg/day in the two-generation study; the one-generation study (Luebker et al. 2005a)
lacked a NOAEL, as pup body weight was impacted at the lowest dose tested (0.4 mg/kg/day).
Perfluorooctane sulfonate (PFOS) - May 2016                                               4-3

-------
    Table 4-1. NOAEL/LOAEL and Effects for Longer-Term Duration Studies of PFOS
Species
Monkey
Monkey
Rat
Rat
Rat
Rat
Rat
Rat
Mouse
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
90 days
NOAEL
(mg/kg/day)
ND
0.15
ND
0.40 (F)
0.34 (M)
0.1
0.4
ND
0.120(F)
0.024 (M)
0.008
0.43
LOAEL
(mg/kg/day)
0.5
0.75
2.0
1.56 (F)
1.33 (M)
0.4
0.8
0.4
0.299 (F)
0.098 (M)
0.083
2.15
Critical Effect(s)
diarrhea, anorexia
I survival, body wt gain
t liver wt; hepatocyte
hypertrophy,
|T3 and |TSH
t liver wt
hepatocyte hypertrophy
t liver wt
1 cholesterol (M)
t ALT (M), |BUN (M/F)
t liver hypertrophy
hepatic centrilobular
vacuolization
1 adult body wt gain
1 pup body wt
1 maternal wt gain
I gestation length
I pup survival
J, pup body weight
Cystic degeneration,
centrilobular vacuolation (M)
and centrilobular eosinophilic
granules (F); | hepatic necrosis
centrilobular vacuolation at
higher doses
t liver wt
t splenic NK cell activity; J,
SRBC response
Impaired spatial learning and
memory
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/Butenhoff
etal. 2012
Dong et al.
2009
Long etal. 2013
Notes: ND = not determined
BUN = blood urea nitrogen
M = male; F = female

    Offspring survival was affected in a dose-related manner in the one-generation study, with a
biologically important decrease in viability index attained at 0.8 mg/kg/day and statistical
significance reached at 1.6 mg/kg/day (Luebker et al. 2005a). In the two generation study
(Luebker et al. 2005b), Fl 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

    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), trends for reduced total triiodothyronine (T3) and increased TSH (males only) were
observed and reached statistical significance for T3 in males and females. In the case of TSH, the
decrease was significant only for males. The trend in females  lacked clear dose-response. There
was no evidence of hypothyroidism. PFOS-induced alterations of thyroid hormones were also
seen studies on adult rats (Martin et al. 2007; Thibodeaux et al. 2003; Yu et al. 2009b, 2011);
Perfluorooctane sulfonate (PFOS) - May 2016
4-4

-------
however, most reductions involved circulating TT4, instead of T3. In most animal studies,
however, the changes in T3 and TT4 failed to activate the HPT feedback mechanism to produce
significant elevations of serum TSH.

   Across the range of longer-term studies, the lowest LOAEL is 0.098 mg/kg/day for
histopathological changes in the liver of male Sprague-Dawley rats following a 104-week
(2-year) exposure (Thomford 2002/Butenhoff et al. 2012). Histological changes observed
included centrilobular eosinophilic granules, centrilobular vacuolation, and centrilobular
hypertrophy with single cell necrosis at a higher dose. Significant increases in absolute and
relative liver weights were not observed. The LOAEL for comparable effects in females was
about 3 times higher. 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. An increase in serum ALT at
the same dose is suggestive of liver damage, but these data were highly variable and did not
notably progress in the Thomford 2002/Butenhoff et al. 2012 study at 27 and 53 weeks. In
monkeys, decreased survival, increased relative liver weight, and decreased cholesterol were
seen at a LOAEL of 0.75 mg/kg/day administered for 6 months (Seacat et al. 2002).

   In the Dong et al. (2009) study, an increase in splenic NK cell activity, a decrease in the
SRBC response, 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. At a
10-fold higher dose, NK response was decreased and indicative of a U-shaped response to dose.
No other studies of an immunological endpoint with a comparable exposure duration were
identified.

   The most severe of the effects observed  in the longer-term studies was the decrease pup
survival in the one-generation study by Luebker et al. (2005a) in rats at a LOAEL of 0.8
mg/kg/day, a dose not evaluated in the two-generation study. The LOAEL for the less  serious
effect of decreased pup  body weight was 0.4 mg/kg/day in the one- and two-generation studies.

   The short-term and developmental exposure studies compiled in Table 4-2 below support the
concern for low dose-effects on pup body weight and survival. The majority of the short-term,
dose-response studies of PFOS were designed to examine developmental end-points.

   Similar to the decreased offspring survival described in the one-generation reproductive
toxicity study (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; Lau et
al. 2003; Thibodeaux 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. Data from Borg et al. (2010) demonstrated
significantly increased levels of fetal and neonatal PFOS concentrations in the lung between GD
18 and PND 1 compared with their dams, providing a possible link to the changes observed by
Chen et al. (2012). Thibodeaux et al. (2003) and Lau et al. (2003) both found decreased maternal
and pup weight, but no effects on maternal liver weight, when dams were  dosed at 2 mg/kg/day
from GD 2 to 20.
Perfluorooctane sulfonate (PFOS) - May 2016                                              4-5

-------
         Table 4-2. NOAEL/LOAEL Data for Short-Term Oral Studies of PFOS
Species
Rat
Rat
Rat
Rat
Rat
Rat
Rat
Rat
Mouse
Mouse
Mouse
Mouse
Mouse
Study Duration
28 days
CDs 1-21
GD 0 to PND 20
CDs 11-19
CDs 2-20
GD 0-PND 20
CDs 0-20
GDO-LD21
CDs 1-17
GD3-PND21
(dams)
(offspring
evaluated on PND
63)
21 days
28 days
CDs 1-17
NOAEL
(mg/kg/day)
ND(F)
0.14(M)
0.1
-
-
1.0
0.3
0.8
0.8
1.0
0.3
1
0.00017 (M)
0.0033 (F)
(M)
1(F)
LOAEL
(mg/kg/day)
0.15(F)
1.33 (M)
2.0
0.5
5
2.0
1.0
2.5
2.5
5.0
3.0
5
0.0017 (M)
0.017 (F)
1(M)
5(F)
Critical Effect(s)
t relative liver wt (M/F), |T4
(M/F)
J, pup survival histopathological
changes to lungs (pups)
J, body weight impaired glucose
tolerance
jbody weight, J, fetal Leydig
cells, and J, testosterone
J, dam and pup body weight
J, pup survival
t motor activity and decreased
habituation in male pups
fwater maze escape distance
and escapre latency
fwater maze escape distance
and escapre latency
t liver wt, dams and pups;
delayed eye opening
t liver weight, increased insulin
resistance
t liver weight
hepatic steatosis
1 SRBC plaque-forming cell
response
I NK cell activity at postnatal
week 8
Reference
Curran et al.
2008
Chen et al.
2012
Lvetal. 2013
Zhao et al.
2014
Thibodeaux et
al. 2003 ;Lau
et al. 2003
Butenhoffet
al. 2009
Y. Wang et
al. 2015
Y. Wang et
al. 2015
Thibodeaux et
al. 2003 ;Lau
et al. 2003
Wan et al.
2014a
Wan et al.
2012
Peden-Adams
et al. 2008
Keil et al.
2008
Note: M = male; F = female


    In the standard developmental neurotoxicity study by Butenhoffet 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. In Y. Wang et al. (2015), the
NOAEL for learning and memory as reflected in Morris water maze results for rats exposed
during gestation and gestation/lactation was 0.8 mg/kg/day and the LOAEL was 2.4 mg/kg/day.
In the longer-term 90-day study by Long et al. (2013), the NOAEL for effects on learning and
memory was 0.43 mg/kg/day with a LOAEL of 2.12 in  mice first exposed at 8 weeks. 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.

    The studies by Lv et al. (2013) in rats and Wan et al. (2014b) in mice provide evidence for
long lasting impacts on blood glucose control in adult animals exposed to PFOS gestationally
and lactationally. In both studies, dams were exposed throughout gestation and lactation, but the
offspring were not directly treated. In the Lv et al. (2013) study,  the animals were evaluated at
22 weeks of age and in the Wan et al. (2014b) study animals were evaluated at 63 days of age
Perfluorooctane sulfonate (PFOS) - May 2016
4-6

-------
(9 weeks). In both cases, the rats exposed during gestation had signs of insulin resistance,
resulting in elevated serum glucose levels.

   Peden-Adams et al. (2008) identified immunotoxicity in male mice exposed to 0.0017
mg/kg/day. IgM production was suppressed after 28 days of treatment although no overt signs of
toxicity were observed at any dose. In the Keil et al. (2008) study, crossbred mice exposed
during gestation had decreased NK cell activity in males and females at postnatal week 8. The
SRBC IgM response was suppressed in males at a higher dose (5 mg/kg/day), but not in females.
The 52%-78% decrease 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 is the only effect at a LOAEL less than that in male rats (0.072 mg/kg/day)
from the Thomford (2002)/Butenhoff et al.  (2012) chronic study. The number of animals per
dose group utilized by Peden-Adams et al. (2008) was small (n = 5). The SRBC response
suppression in male pups (n = 6) from the Keil et al. (2008) developmental exposure was higher
at 5 mg/kg/d; females showed no response.  The longer duration study by Dong et al. (2009) also
had a higher LOAEL at 0.083 mg/kg/day for SRBC suppression and increased liver weight.

   Decreased NK cell activity occurred at a lower dose than the  SRBC response in the Keil et al.
(2008) study, at a higher dose in the Peden-Adams et al. (2008) study, and at the same dose in
the Dong et al.  (2009) study. The NK cell activity was enhanced at very low PFOS doses, while
it was depressed at higher doses. These differences highlight the need for additional research to
confirm the NOAEL and LOAEL for the immunological endpoints. In all three studies with the
low dose responses, males responded at lower doses than females.

   Studies in mice examined NK cell activity and SRBC response. Three of four studies showed
effects on  SRBC response, NK cell activity, or both at the same dose that caused increased liver
weight (0.083 mg/kg/day,  Dong et al. 2009; 5 mg/kg/day, Keil et al. 2008; Zheng et al. 2009).
The extremely low-dose effects found in Peden-Adams et al. (2008) with a LOAEL for SRBC
response of 0.0017 mg/kg/day after 28 days are not supported by the LOAEL of 0.083
mg/kg/day for a dosing duration of 60 days from Dong et al. (2009).

   Taken together, the lower antibody liters associated with PFOS levels in humans and the
consistent suppression of SRBC response in animals indicates  a concern for adverse effects on
the immune system. However, lack of human dosing information and lack of low-dose
confirmation of effects in animals for the short-duration study precludes the use of these
immunotoxicity data in setting the RfD.

4.1.1.1  Pharmacokinetic Model

   Among the studies summarized in Tables 4-1 and 4-2, a number reported low-dose adverse
effects and had data on measured serum concentrations that made them suitable for
pharmacokinetic  modeling in order predict  a time-integrated average serum concentration for the
exposure duration and experimental doses. Because of the complexities of the pharmacokinetic
differences between animals and humans and across animal species, the average serum values
are a superior point of departure (POD) for  RfD derivation, rather than the external doses in the
studies. Generally, it was assumed that animals were observed at the end of dosing. The
published Wambaugh et al. (2013) model described in section 2.5.1 was applied to the selected
studies. The use of the animal data and the available PK model allows for the incorporation of
species differences in saturable renal resorption, dosing duration, and serum measurements for
Perfluorooctane sulfonate (PFOS) - May 2016                                             4-7

-------
doses administered to determine human equivalent doses based on average serum concentration
and clearance.

   The results for studies in the rats are summarized in Table 4-3. For the Butenhoff et al.
(2009) study two different AUCs were calculated—gestational only (for the male offspring
endpoint) and gestational plus 20 days postnatal (for the maternal endpoint). This separation of
the two exposures neglects lactational transfer of compound, which was not modeled.

   The predicted results from studies in mice and the monkey are provided in Tables 4-4 and
4-5, respectively. The Lau et al. (2003) data on mice are representative of the impact of PFOS on
developmental endpoints. Although the duration of this study is relatively short at 19 days, the
average serum levels associated with the observed effects on pup body weight and
developmental milestones merit consideration.  The Seacat et al. (2002) study on monkeys is a
long term (6 month), multiple dose study of systemic toxicity in which the LOAEL for effects on
liver weight, liver histopathology, cholesterol, body weight gain,  T3, and TSH was accompanied
by early death in two of six monkeys.

   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. The average serum concentration given in Table 4-6 for the LOAEL or
NOAEL was determined through numeric simulation. Averaging the serum concentrations for
the duration of exposure is important because of the variability in the times of exposure across
the studies (17-182 days).

   Average serum concentration has the advantage of normalizing 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 4-6 provides dosing duration and the predicted average serum concentration from each
of the modeled studies. Internal doses associated with developmental toxicity were 19.9-25
|ig/mL for reduced pup body weight (Luebker et al. 2005a, 2005b), 34.6 |ig/mL for changes in
motor activity (Butenhoff et al. 2009), and 35.1-39.7 |ig/mL for pup survival (Lau et al. 2003;
Luebker et al. 2005a). In comparison, internal doses associated with increased liver weight were
64.6-157 |ig/mL (Seacat et al. 2002, 2003). Thus, the internal doses associated with the
developmental and liver effect levels (LOAELs) differ by less than an order of magnitude
(19.9-157 |ig/mL), while the  corresponding AUC values (Tables 4-3 through 4-5) differ by more
than an order of magnitude (30,100 |ig/mL*h-684,000 |ig/mL*h).
Perfluorooctane sulfonate (PFOS) - May 2016                                               4-8

-------
          Table 4-3. Predicted Final Serum Concentration and Time Integrated Serum
                      Concentration (AUC) for Different Treatments of Rat
Study
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/
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
98 Days
98 Days
Gestation (22
Days)
Gestation (21
Days) +
Postnatal (20
Days)
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
0.03
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
fig/mL
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
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Female
Rat/Sprague-
Dawley
Predicted
Final Serum
Concentration
fig/mL
2.29 (0.0888)
9.94 (0.386)
25.9 (0.976)
101 (3.94)
4.86 (0.0978)
18.2(0.364)
48.3(1.07)
187(7.98)
3.7(0.121)
11.1(0.367)
37.1 (1.2)
6.36(0.167)
19.1(0.512)
63.5(1.67)
32.4(1.05)
64.8 (2.23)
97(3.26)
162(5.61)
321 (15)
11(0.226)
43.8(0.882)
174 (5.73)
342 (24.5)
35.7(0.765)
71.3(1.65)
88.9 (2.25)
107(2.91)
142(4.13)
177(6.38)
Predicted AUC
jig/mL*h
3,430 (108)
14,900 (480)
38,900(1,230)
152,000(4,860)
6,620 (143)
24,800(561)
65,800(1,500)
256,000 (7,500)
1,060(37.7)
3,180(114)
10,600 (376)
3,410(105)
10,300 (323)
34,100(1,040)
8,020 (279)
16,000 (594)
24,000 (866)
40,100(1,430)
79,800 (3,070)
12,600(312)
50,400(1,180)
201,000(5,250)
398,000 (17,700)
30,100 (794)
60,100(1,640)
75,000 (2,060)
90,000 (2,600)
120,000 (3,400)
150,000(4,530)
Notes: Numbers in parentheses indicate SD
GD = gestation day; NT = not tested
a Thibodeaux et al. (2003) data available only in a graph in the published paper; the values for the model obtained from author.

          Table 4-4. Predicted Final Serum Concentration and Time Integrated Serum
                               Concentration (AUC) for the Mouse
Study
Lauet
al. 2003
Species /
Strain
Female
Mouse/CD-I
Study
Duration
And Type
GDs 1-17
(17 days)
Administere
d Doses
mg/kg/day
1
5
10
15
20
Measured Final
Serum
Concentration
fig/mL
NT
NT
NT
NT
NT
Species /
Strain Used
for
Prediction
Female
Mouse /
CD1
Predicted Final
Serum
Concentration
fig/mL
54.8 (1.78)
195 (38.4)
259(103)
289 (158)
312(217)
Predicted AUC
fig/mL*h
13,500 (460)
57,700 (5,220)
88,900 (19,700)
106,000 (35,000)
118,000(50,300)
Notes: Numbers in parentheses indicate SD
GD = gestation day; NT = not tested
   Perfluorooctane sulfonate (PFOS) - May 2016
4-9

-------
           Table 4-5. Predicted Final Serum Concentration and Time Integrated Serum
                                Concentration (AUC) for the Monkey
Study
Seacat et
al. 2002
Species /
Strain
Monkey /
Cynomol-
gus
Study
Duration
And Type
182 days
Administered
Doses
mg/kg/day
0.03
0.15
0.75
Measured Final
Serum
Concentration
fig/mL
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
Concentration
fig/mL
14.3 (0.228)
68.8 (0.978)
225 (6.28)
Predicted AUC
jig/mL*h
33,800 (547)
166,000 (2460)
684,000 (10,700)
Notes: Numbers in parentheses indicate SD
M = male; F = female
               Table 4-6. Average Serum Concentrations for the Duration of Dosing
Study
Seacat et al. 2002
monkey: | liver weight +
histopathology; jbody
weight; |T3; |TSH
Seacat et al. 2003 male
rat: t liver weight,
centrilobular
vacuolization,! ALT,
|BUN
Luebkeretal. 2005b: |
rat pup body weightb
Luebker et al. 2005a: |
rat pup body weightb
Luebker et al. 2005a rat:
J, maternal body weight,
gestation length and pup
survival13
Butenhoffetal. 2009 rat
developmental
neurotoxictiy: t increased
motor activity!
habituation
Lauetal. 2003: jratpup
survival; jmaternal and
pup body weight
Dosing
duration
days
182
98
84
63
63
41
19
NOAEL
mg/kg/day
0.15
0.34
0.1
None
0.4
0.3
1.0
NOAEL
(Av serum
fig/mL)a
38
(0.564)
16.5
(0.522)
6.26
(0.155)
None
19.9
(0.525)
10.4
(0.328)
17.5
(0.609)
LOAEL
mg/kg/day
0.75
1.33
0.4
0.4
0.8
1.0
2.0
LOAEL
(Av serum
fig/mL)a
157
(2.45)
64.6
(2.06)
25
(0.583)
19.9
(0.525)
39.7
(1.09)
34.6
(1.05)
35.1
(1.3)
   Notes: a Average serum concentrations predicted from PK simulations of dose regimens were performed using species-specific
   parameter distributions. The number in parentheses is the SD.
   b Multiple effects are included for the Luebker et al. (2005a, 2005b) studies to distinguish between the effects quantified for dose-
   response.
   Perfluorooctane sulfonate (PFOS) - May 2016
4-10

-------
   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
6.26-19.9 |ig/mL for developmental/neurodevelopmental endpoints (Butenhoff et al. 2009; Lau
et al. 2003; Luebker et al. 2005a, 2005b) and 16.5-38 |ig/mL for liver weight changes and
accompanying liver pathology and changes in serum biochemistry (Seacat et al. 2002, 2003).
Despite the similarity in average serum concentrations, the AUC values differ by an order of
magnitude (12,600 |ig/mL*h-l66,000 |ig/mL*h). Given the differences in external doses, the
projected serum levels are proportionally quite similar. Table 4-6 identifies 6.26 and 10.4 |ig/mL
as the lowest average serum concentrations associated with a NOAEL for offspring effects; the
associated LOAELs were based on 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, liver histopathology, changes in body weight, and serum biochemistry
in monkeys (38 jig/mL; Seacat et  al. 2002) and male rats (16.5 jig/mL; Seacat et al. 2003) are
very similar to the average no effect serum value in Lau et al. (2003) for decreased pup survival
with a shorter averaging time (17.5 |ig/mL). Thus, it appears that the NOAELs are consistent
across gender, species, and treatment with respect to average serum concentration. Assuming
that mode of action and susceptibility to toxicity do not vary and that pharmacokinetics alone
explain variation, it is reasonable to expect similar concentrations to cause similar effects in
humans.

   The Wambaugh et al.  (2013) model employed here to generate the average serum
concentrations shown in Table 4-6 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 34.6 |ig/mL as seen in Table 4-6, which is very similar to the approximately 25 |ig/mL for the
dams that can be estimated from the graph (Loccisano et al. 2012b). The slightly higher value
calculated from the Wambaugh et al. (2013) model might 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 Wambaugh et al. (2013) model predicts maternal values,  a direct
comparison to the fetal 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 Wambaugh et al.
(2013) model can be accepted with reasonable confidence that the predicted AUC values
accurately represent maternal levels during gestational and lactational exposures.

   The Andersen et al. (2006) model, used to make the predictions in Tables 4-3 through 4-6,
calls for numerical simulation in order to make predictions for serum concentrations resulting
from a regimen of discrete doses. However, one can predict the steady-state concentration (Css)
resulting from a fixed infusion dose rate (DR, in units of |imol/h):

                                    DR
                            bb   free * Qfil V    Qfil * kT + DR/

   The Css depends non-linearly on DR. The PFOS studies in Tables 4-1 and 4-2, used discrete,
daily doses that can be converted to DR by dividing the daily dose (mg/kg/day) by 24 hours to
Perfluorooctane sulfonate (PFOS) - May 2016                                             4-11

-------
give and approximate measure of DR. For each DR and species a range of Css values can be
calculated by using species-specific combinations of parameters from the Bayesian analysis of
the available PK data. In Table 4-7, the Css is compared with the average serum concentration
predicted for each of the studies in Table 4-6. The average serum concentration fraction of the
Css for the 182-day Seacat et al. (2002) study in monkeys is approximately 69% of the steady-
state concentration. The 19-day average serum concentration from Thibodeaux et al. (2003) is
only approximately 9% of Css, while the average serum concentration for the rest of the modeled
studies ranges 17%-50%  of Css.

    The shortest duration  study in Table 4-7 had a higher administered LOAEL dose than the
longest studies (0.75  mg/kg/day for 182 days versus 2.0 mg/kg/day for 19 days). Despite the
higher administered dose, the short 19-day study resulted in effects at a lower serum
concentration than that for the longest duration of exposure, the one closest to steady state. In
fact, the average serum values from the studies that do not approach steady state have lower
average serum LOAELs for endpoints of toxicological concern. Thus, the data do not appear to
indicate increasing sensitivity as steady-state is approached. If anything, the average serum
values appear to be more  protective than serum concentrations at steady state.

 Table 4-7. Comparison of Average Serum Concentration and Steady-State Concentration
Study
Seacat et al. 2002: monkey:
t liver weight +
histopathology; jbody
weight; |T3; |TSH
Seacat et al. 2003 : male rat:
t liver weight, centrilobular
vacuolization,! ALT, |BUN
Luebker et al. 2005b: | rat
pup body weight
Luebker et al. 2005a: |rat
pup body weight
Luebker et al. 2005a: rat pup
survival and J, maternal body
weight
Butenhoff etal. 2009: |rat
pup body weight
Lau et al. 2003rat: pup
survival; J, maternal and pup
body weight
Dosing
duration
days
182
98
84
63
63
41
19
LOAEL
mg/kg/day
0.75
1.33
0.4
0.4
0.8
1.0
2.0
Css (mg/L) for
constant infusion
of LOAEL
227 (6.95)
128 (7.9)
83.4 (6.96)
83.3 (7.08)
163 (15.9)
203 (22.5)
397 (57.6)
Average Serum
Cone, for Study
(mg/L)
157 (2.45)
64.6 (2.06)
25 (0.583)
19.9 (0.525)
39.7(1.09)
34.6(1.05)
35.1(1.3)
Fraction of Css
(Average / Css)
0.689(0.0131)
0.504(0.0211)
0.302 (0.027)
0.24 (0.0232)
0.246 (0.0273)
0.173 (0.0245)
0.0911(0.0202)
Notes: Average serum concentrations from PK simulations of toxicity study treatment regimens and Css were both predicted
using species-specific parameter distributions. The number in parentheses is the SD.

    For human exposure to PFOS one needs to rely on average serum calculations since there is a
lack of both the sufficient PK and exposure knowledge to make more complicated estimates. The
average serum concentrations of the LOAEL in Table 4-7 range from 19.9 to 157 |ig/mL; all are
within one order of magnitude. The predicted toxic 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.
Perfluorooctane sulfonate (PFOS) - May 2016
4-12

-------
   A reliable measure of half-life in humans is available from a retired worker population
followed for 5 years. Olsen et al. (2007) calculated the PFOS half-life in this former worker
population as 5.4 years (see section 2.5.2). Thompson et al. (2010) give a volume of distribution
of 0.23 L/kg bw (see section 2.5.3). 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 x (0.693 H-  1,971 days) = 0.000081 L/kg bw/day

   Where:
       Vd    =      0.23  L/kg
       In 2   =      0.693
       ti/2    =      1,971 days (5.4 years x 365 days/year = 1,971 days)
       These values combined give a clearance of 8.1 x 10"5 L/kg bw/day.

   Scaling the derived average concentrations (in |ig/mL) for the  NOAELs and LOAELs in
Table 4-6 gives predicted oral HEDs in mg/kg bw/day  for each corresponding serum
measurement. The HED 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, HED = average serum concentration (in |ig/mL) x CL

   Where:
       Average serum is from model output in Table 4-6
       CL = 0.000081 L/kg bw/day

   The resulting HED values are shown in Table 4-8. Endpoints considered as critical effects  in
multiple studies include offspring growth and survival, liver weight changes, liver
histopathology, and changes in serum biochemistry indicative of systemic effects. Each study
selected for modeling was of high quality and show effects at low doses. In all cases but one
(Luebker et al. 2005a) the POD for the analysis was a NOAEL rather than a LOAEL. The
developmental effects of reduced pup body weight and survival occurred in the absence of
changes in maternal liver weight, indicating that maternal toxicity and PPARa were not
confounding variables.

   The external dose NOAELs and LOAELs from other studies summarized in Tables 4-1 and
4-2 that lacked serum information are comparable to those in the modeled studies. For example,
the NOAEL in the Long et  al. (2013) 90-day mouse study for effects on learning and memory  is
0.43 mg/kg/day (Table 4-1) compared to the 0.3 mg/kg/day for Butenhoff et al. (2009) in rats
and the LOAEL for mice is 2.15 mg/kg/day compared to the value of 1 mg/kg/day for rats. The
LOAEL from Luebker et al. (2005a) of 0.4 mg/kg/day  for decreased pup body weight is not
unlike the 0.5 mg/kg/day observed by Lv et al. (2013) for decreased pup body weight and
increased insulin resistance (Table 4-2). The 1.0 mg/kg/day NOAEL and 2.0 mg/kg/day LOAEL
for decreased body weight in rat dams and pups combined with decreased pup survival (Lau et
al. 2003; Thibodeaux et al.  2003) are quite similar to the corresponding values of 1  and
5 mg/kg/day, respectively, in the study of mice conducted by the same authors (increased
maternal liver weight and delayed pup eye opening).
Perfluorooctane sulfonate (PFOS) - May 2016                                            4-13

-------
     Table 4-8. Human Equivalent Doses Derived from the Modeled Animal Average
                                    Serum Values


Study
Seacat et al. 2002 monkey:
t liver weight +
histopathology; jbody
weight; |T3; |TSH
Seacat et al. 2003 male rat:
t liver weight, centrilobular
valorization,! ALT, |BUN
Luebker et al. 2005b rat:
|pup body weight
Luebker et al. 2005a rat: |
pup body weight
Luebker et al. 2005a rat:
I maternal body weight,
gestation length and pup
survival
Butenhoffetal. 2009 rat
developmental neurotoxictiy:
tmotor activity, jhabituation
Lau et al. 2003 rat: | pup
survival; maternal and pup
body weight
Dosing
duration
days

182


98

84
63

63


41


19


NOAEL
mg/kg/d

0.15


0.34

0.1
None

0.4


0.3


1.0

NOAEL
Av serum
fig/mL

38


16.5

6.26
None

19.9


10.4


17.5


RED
mg/kg/d

0.0031


0.0013

0.00051
None

0.0016


0.00084


0.0014


LOAEL
mg/kg/d

0.75


1.33

0.4
0.4

0.8


1.0


2.0

LOAEL
Av serum
fig/mL

157


64.6

25
19.9

39.7


34.6


35.1


RED
mg/kg/d

0.013


0.0052

0.002
0.0016

0.0032


0.0028


0.0028

4.1.1.2  RfD Quantification

   Several acceptable PODs can be used in the process of RfD development based on the
modeled human equivalent doses (Table 4-9).

   All modeled 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) with a
LOAEL of 0.4 mg/kg/day. The same external dose was also a LOAEL for the same effect in the
two-generation study by Luebker et al. (2005b), with a NOAEL of 0.1 mg/kg/day, a dose not
tested in the one-generation study. The calculated HED values associated with no adverse effects
on developmental and liver endpoints (NOAELs) were very similar with a range of 0.00051-
0.0031 mg/kg/day.

   Two effect-level doses were modeled from the Luebker et al. (2005a) one-generation rat
study: (1) the NOAEL for the effects on pup survival (0.4 mg/kg/day), which was the LOAEL
for the body weight effect, and (2) the LOAEL (0.8 mg/kg/day) for the pup  survival effect to
illustrate the  importance of the body weight LOAEL in both the one- and two-generation
Luebker et al. (2005a, 2005b) studies. In the two-generation study, 1.6 mg/kg/day resulted in the
death of > 26% of the pups between LD 2 and 4. Support for the pup survival  serum level
LOAEL is provided by the Lau et al. (2003) rat study, with a HED for the same end point that is
comparable to that in the Luebker et al. (2005b) study (0.0028 mg/L and 0.0032 mg/L,
respectively).
Perfluorooctane sulfonate (PFOS) - May 2016
4-14

-------
         Table 4-9. POD Outcomes for the HEDs from the Pharmacokinetic Model
                                 Average Serum Values
POD
PK-HED (Seacat et al.
2003;rat,NOAEL,t ALT,
|BUN
PK-HED (Lauetal. 2003;
rat, NOAEL |pup
survival)
PK-HED (Butenhoff et al.
2009; rat, NOAEL tmotor
activity jhabituation)
PK-HED (Luebkeretal.
2005b; rat, NOAELjpup
body wt)
PK-HED LOAEL
(Luebker et al. 2005a; rat,
LOAELJ,pup body wt)
PK-HED (Luebker et al.
2005a; rat, NOAEL |pup
survival)
POD Value
mg/kg/day
0.0013
0.0014
0.00084
0.00051
0.0016
0.0016
UFH
10
10
10
10
10
10
UFA
3
3
o
J
o
J
o
J
3
UFL
1
1
1
1
3
1
UFs
1
1
1
1
1
1
UFD
1
1
1
1
1
1
UFtotal
30
30
30
30
100
30
Candidate
RfD
mg/kg/day
0.00004
0.00005
0.00003
0.00002
0.00002
0.00005
Notes: UFn: Intra-individual uncertainty factor, UFA: Interspecies uncertainty factor, UFs: Subchronic to chronic uncertainty
factor, UFL: LOAEL to NOAEL uncertainty factor, UFo: incomplete database uncertainty factor, UFtotai: Total (multiplied)
uncertainty factor

    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.
However, the modeled average serum values from animals are several orders of magnitude
greater than measured values in humans. Thus, extrapolation to humans adds a layer of
uncertainty that needs to be accommodated in deriving the RfD.

HED PODs. The PK HEDs derived from Seacat et al. (2003), Lau et al. (2003), Butenhoff et al.
(2009),  and Luebker et al. (2005a, 2005b) were each examined as the potential basis for the RfD
(ph). The Seacat et al. (2002) results for male monkeys were not utilized in the derivation of the
RfD because of the premature deaths in two of the six males at the LOAEL. Each of these
studies, except one,  contained a NOAEL from which the HED could be derived. The outcomes
for potential RfD values are similar demonstrating the ability of the model to normalize the
animal data across species, gender, and exposure duration.

Uncertainty Factors

    An uncertainty factor for intraspecies variability (UFn) of 10 is assigned to account for
variability in the responses within the human populations because of both intrinsic (genetic, life
stage, health status) and extrinsic (life style) factors that can influence the response to exposure.
No information was available relative to variability in the human population that supports a
factor other than 10.

    An uncertainty factor for interspecies variability (UFA) of three was applied to account for
uncertainty in extrapolating from laboratory animals to humans (i.e., interspecies variability).
The 3-fold factor is applied to account for toxicodynamic differences between the animals and
Perfluorooctane sulfonate (PFOS) - May 2016
4-15

-------
humans. The HEDs were derived using average serum values from a model to account for
pharmacokinetic differences between animals and humans.
   An uncertainty factor for LOAEL to NOAEL extrapolation (UFL) of one was applied to all
PODs, except the LOAEL of 0.4 mg/kg/day for effects on pup body weight in the one-generation
Luebker et al. (2005a) study. A value of three is assigned for this study based on the fact that the
NOAEL for this effect was 0.1 mg/kg/day in the two-generation (Luebker et al. 2005b) study, a
dose that was not used in the one-generation study. The LOAEL in the two-generation study was
0.4 mg/kg/day, demonstrating that the difference between a NOAEL and LOAEL for the  body
weight is not a factor of 10, the default value for NOAEL/LOAEL extrapolation.

   An uncertainty factor for extrapolation from a subchronic to a chronic exposure duration
(UFs) of one was applied because the PODs are based on average serum concentrations for all
studies except Seacat et al. (2013). The studies for developmental endpoints are not adjusted for
lifetime exposures because they cover a critical window of exposure with lifetime consequences.
The average serum value associated with the developmental (Luebker et al. 2005b) POD is lower
than that for any of the other modeled studies including those with systemic effects after longer
exposures. It is accordingly more protective of adverse effects than the POD for any of the
longer-term studies despite the limited exposure duration. The serum from the Seacat et al.
(2013) study was collected at 14 weeks. Some of the animals in the study continued to be dosed
for a total of 105 weeks, but the effects observed at the LOAEL did not increase in magnitude.
Serum measurements taken before sacrifice were 2-fold higher at 14 weeks in males than they
were at 105 weeks. Concentrations of PFOS in the liver were lower at 105 weeks than they were
at 14 weeks. The PFOS concentrations in the diet were constant. SDs about the monitored ALT
and BUN were broad indicating higher sensitivity is some animals than others. The serum and
effects data for the male rats justify the subchronic to chronic adjustment to the study NOAEL
for this study.

   A database uncertainty factor (UFo) of one was applied to account for deficiencies in  the
database for PFOS. The epidemiology data provide strong support for the identification of
hazards observed following exposure to PFOS in the laboratory animal studies and human
relevance. However, uncertainties in the use of the available epidemiology data precluded their
use at this time in the quantification of the effect level for derivation of the drinking water health
advisory. In animals,  comprehensive oral short term, subchronic, and chronic studies in three
species and several strains of laboratory animals have been conducted and published in the peer
reviewed literature. Additionally, there are several neurotoxicity studies (including
developmental neurotoxicity) and several reproductive (including one- and two-generation
reproductive toxicity  studies) and developmental toxicity studies including assessment of
immune effects following developmental exposure.

RfD Selection

   Based  on the consistency of the response and of the use of the most sensitive endpoint,
developmental toxicity, as the critical effect, the RfD of 0.00002 mg/kg/day from Luebker et al.
(2005a) is selected as the RfD for PFOS. This RfD is derived from reduced pup body weight in
the two-generation study in rats.  The POD for the derivation of the RfD for PFOS is the HED of
0.0005 1 mg/kg/day that corresponds to a NOAEL that represents approximately 30% of steady-
state concentration. An UF of 30 (10 UFn and 3 UFA) was applied to the HED NOAEL to derive
an RfD of 0.00002 mg/kg/day. This is supported by the 0.00002 mg/kg/day value derived from
the LOAEL for the same effect in the one-generation Luebker et al. (2005a) study and the

Perfluorooctane sulfonate (PFOS) - May 2016                                             4-16

-------
0.00003 mg/kg/day value for neonatal neurodevelopmental effects in the Butenhoff et al. (2009)
study.

    Low body weights in neonates are a biomarker for developmental deficits and linked to
problems often manifest later in life. A study by Lv et al. (2013) that lacked serum data for
pharmacokinetic modeling identified 0.5 mg/kg/day as a LOAEL for effects on body weight in
Wistar rat pups exposed during gestation, an observation that was accompanied by increased
insulin resistance, problems with glucose homeostasis, and hepatic fat accumulation in the pups
as adults. A similar effect on glucose homeostasis was observed in CD-I mice at PND 63 in a
study by Wan et al. (2014b), with a  dose of 3 mg/kg/day for animals receiving a diet with regular
fat content. For animals receiving a  high fat diet, the LOAEL was 0.3 mg/kg/day. Support for the
neurodevelopmental effects in Butenhoff et al. (2009) at a dose 1 mg/kg/day kg/day is provided
by the NOAEL (0.43 mg/kg/day) in the Long et al. (2013) 90-day mouse study for effects on
learning and memory.

    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 (Karrman et al. 2010), and in serum
samples from children aged 5-15 years (Dong et al. 2013; Grandjean et al. 2012). A human
epidemiology study found no association with maternal PFOS levels and motor or mental
development of their children; the mean maternal serum concentration was approximately
0.035 |ig/mL (Fei et al. 2008b).

4.1.2  RfC Determination

    The only inhalation study available is an acute lethality inhalation study in rats (Rusch et al.
1979); no inhalation data are available in humans. Thus, data are insufficient for the development
of an RfC  for PFOS.

4.2  Dose-Response for Cancer Effects

    Under the EPA (2005a) Guidelines for Carcinogen Risk Assessment, when the evidence from
the epidemiology studies and the cancer bioassays is suggestive for carcinogenicity, a
quantitative estimate of risk is generally not performed unless there is a well-conducted study
that could serve a useful purpose by providing a sense of the magnitude and uncertainty of
potential risks, ranking potential hazards, or setting research priorities. In the case of PFOS, the
existing evidence does not support a strong correlation between the tumor incidence and dose to
justify a quantitative assessment.
Perfluorooctane sulfonate (PFOS) - May 2016                                             4-17

-------
5.   REFERENCES

3M Company. 2000. Determination of Serum Half-Lives of Several Fluor ochemicals. Interim
     Report #1, June 8, 2000. 3M Company, St. Paul, MN.

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. 3M Company,
     St. Paul, MN.

Abbott, B.D., CJ. Wolf, K.P. Das, R.D. Zehr, I.E. Schmid, A.B. Lindstrom, MJ. Strynar, and
     C. Lau. 2009. Developmental toxicity of perfluorooctane sulfonate (PFOS) is not
     dependent on expression of peroxisome proliferator activated receptor-alpha (PPARa) in
     the mouse. Reproductive Toxicology 27:258-265.

Alexander, B.H., G.W. Olsen, J.M. Bums, J.H. Mandel, and J.S. Mandel. 2003. Mortality of
     employees of a perfluorooctanesulfonyl fluoride manufacturing facility. Occupational
     and Environmental Medicine 60:722-729.

Alexander, B.H., and G.W. Olsen. 2007. Bladder cancer in perfluorooctanesulfonyl fluoride
     manufacturing workers. Annals of Epidemiology 17:471-478.

Andersen, M.E., 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. Toxicological Sciences 102:3-14.

Andersen, C.S., C. Fei, M. Gamborg, E.A. Nohr, T.I. A.  S0rensen, and J. Olsen. 2010. Prenatal
     exposures to perfluorinated chemicals and anthropometric measures in infancy. American
     Journal of Epidemiology 172:123 0-123 7.

Andersen, C.S., C. Fei, M. Gamborg, E.A. Nohr, T.I.A.  S0rensen, and J. Olsen. 2013. Prenatal
     exposures to perfluorinated chemicals and anthropometry at 7 years of age. American
     Journal of Epidemiology 178:921-927.

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. Environmental Science &
     Technology 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. Human & Experimental
     Toxicology 13(Suppl. 2):S1-S117.

Ashford, R.D. 1994. Ashford's Dictionary of Industrial Chemicals: Properties, Production,
     Uses. Wavelength Publications, Ltd.

Perfluorooctane sulfonate (PFOS) - May 2016                                              5-1

-------
ATSDR (Agency for Toxic Substances and Disease Registry). 2015. Draft Toxicological
     Profile for Perfluoroalkyls. United States Department of Health and Human Services,
     Agency for Toxic Substances and Disease Registry, Public Health Service, Atlanta, GA.
     Accessed May 2016. http://www.atsdr.cdc.gov/ToxProfiles/tp200.pdf

Augustine, L.M., RJ. Markelewicz, K. Boekelheide, andNJ. Cherrington. 2005. Xenobiotic
     and endobiotic transporter mRNA expression in the blood-testis barrier. Drug
     Metabolism and Disposition 33(1): 182-189.

Bach, C.C., Z. Liew, B.H. Bech, E.A. Nohr, C. Fei, B.C. Bonefeld-J0rgensen, T.B. Henriksen,
     and J. Olsen. 2015. Perfluoroalkyl acids and time to pregnancy revisited: An update from
     the Danish National Birth Cohort. Environmental Health  14:59.

Barrett, E.S., C. Chen, S.W. Thurston, L.S. Haug, A. Sabaredzovic, F.N. Fjeldheim, H.
     Frydenberg, S.F. Lipson, P.T. Ellison, and I.  Thune. 2015. Perfluoroalkyl substances and
     ovarian hormone concentrations in naturally  cycling women. Fertility and Sterility
     103(5):1261-1270.

Beesoon, S., G.M. Webster, M. Shoeib, T. Harner, J.P. Benskin, and J.W. Martin. 2011.
     Isomer profiles of perfluorochemicals in matched maternal, cord, and house dust
     samples: Manufacturing sources and transplacental transfer. Environmental Health
     Perspectives 119:1659-1664.

Beesoon, S., and J.W. Martin. 2015. Isomer-specific binding affinity of
     perfluorooctanesulfonate (PFOS) and perfluorooctanoate  (PFOA) to serum proteins.
     Environmental Science & Technology 49:5722-5731.

Berg, V., T.H. N0st, S. Hansen, A. Elverland, A.-S. Veyhe, R. Jorde, J.0. Odland, and T.M.
     Sandanger. 2015. Assessing the relationship between perfluoroalkyl substances, thyroid
     hormones and binding proteins in pregnant women; a longitudinal mixed effects
     approach. Environment International 77:63-69.

Biesemeier, J.A., and D.L. Harris. 1974. Eye and Skin Irritation Report on Sample T-1117.
     Project No. 4102871. WARF Institute, Inc.

Bijland, S., P.C.N. Rensen, EJ. Pieterman, A.C.E. Maas, J.W. van der Hoorn, M.J. van Erk,
     L.M. Havekes, K.W. van Dijk, S.-C. Chang,  E.J. 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. Toxicological Sciences 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. Toxicological Sciences 111:89-99.

Bjork, J.A., J.L. Butenhoff, and K.B. Wallace. 2011. Multiplicity of nuclear receptor activation
     by PFOA and PFOS in primary human and rat hepatocytes. Toxicology 228:8-17.

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.
     Physiology & Behavior 99:240-245.
Perfluorooctane sulfonate (PFOS) - May 2016                                              5-2

-------
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.

Bonefeld-J0rgensen, E.G., M. Long, R. Bossi, P. Ayotte, G. Asmund, T. Kriiger, M. Ghisari,
     G. Mulvad, P. Kern, P. Nzulumiki, and E. Dewaily. 2011. Perfluorinated compounds are
     related to breast cancer risk in Greenlandic Inuit: A case control study. Environmental
     Health 10:88.

Bonefeld-J0rgensen, E.G., M. Long, S.O. Fredslund, R. Bossi, and J. Olsen. 2014. Breast
     cancer risk after exposure to perfluorinated compounds in Danish women: A case-control
     study nested in the Danish National Birth Cohort. Cancer Causes Control 25(11): 1439-
     1448.

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
     sulfonate (PFOS) in C57B1/6 mice following late gestational exposure. Reproductive
     Toxicology 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.

Buck Louis, G.M., Z. Chen, E.F. Schisterman, S. Kim, A.M. Sweeney, R. Sundaram, C.D.
     Lynch, R.E.  Gore-Langton, and D.B. Barr. 2015. Perfluorochemicals and human semen
     quality: The LIFE study. Environmental Health Perspectives 123(1): 57-63.

Burris, J.M., J.K. Lundberg, G.W. Olsen, C. Simpson, and J. Mandel. 2002. Determination of
     Serum Half-Lives of Several Fluor ochemicals. Interim Report #2. 3M Company, St. Paul,
     MN.

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 cynomolgus monkeys after oral dosing for 6 months. Toxicological Sciences
     69:244-257.

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 cynomolgus monkeys. Toxicological Sciences 82:394-406.

Butenhoff, J., and S. Chang. 2007. ADME Study of Perfluorooctane Sulfonate  in Rats. ST-179
     Final Report. 3M Strategic Toxicology Laboratory, 3M Center, St. Paul,  MN.

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. Reproductive Toxicology 27:319-330.

Butenhoff, J. L., S.C. Chang, G.W. Olsen, and P.J. Thomford. 2012. Chronic dietary toxicity
     and carcinogenicity study with potassium perfluorooctane sulfonate in Sprague Dawley
     rats. Toxicology 293:1-15.
Perfluorooctane sulfonate (PFOS) - May 2016                                              5-3

-------
Chan, E., I. Burstyn, N. Cherry, F. Bamforth, and J.W. Martin. 2011. Perfluorinated acids and
     hypothyroxinemia in pregnant women. Environmental Research  111:559-564.

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. 2007. Negative bias from analog
     methods used in the analysis of free thyroxine in rat serum containing
     perfluorooctanesulfonate (PFOS). Toxicology 234:21-33.

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. 2008. Thyroid hormone status and
     pituitary function in adult rats given oral doses of perfluorooctanesulfonate (PFOS).
     Toxicology 243:3 3 0-3 3 9.

Chang, S.-C., D.J. 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. Reproductive Toxicology 27:387-399.

Chang, S.-C., P.E. Noker, G.S. Gorman, S.J. Gibson, J.A. Hart, D.J. Ehresman, and J.L.
     Butenhoff. 2012. Comparative pharmacokinetics of perfluorooctanesulfonate (PFOS) in
     rats, mice and monkeys. Reproductive Toxicology 33:428-440.

Chang, E.T., H. Adami, P. Boffetta, C. Cole, T.B. Starr, and J.S. Mandel. 2014. A critical
     review of perfluorooctanoate and prefluorooctanesulfonate exposure and cancer risk in
     humans. Critical Reviews in Toxicology 44(51): 1-81.

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). Environmental Research 110:710-717.

Chen, Y.-M., and L.-H. Guo. 2009. Fluorescence study on site-specific binding of
     perfluoroalkyl acids to human serum albumin. Archives of Toxicology 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. Reproductive Toxicology 33:538-545.

Chen, H., P. He, H. Rao,  F. Wang, H. Liu, and J. Yao.  2015. Systematic investigation of the
     toxic mechanism of PFOA and PFOS on bovine serum albumin by spectroscopic and
     molecular modeling. Chemosphere 129:217-224.

Christensen, K.Y., M. Maisonet, C. Rubin, A. Holmes, A.M. Calafat, K. Kato, W.D. Flanders,
     J. Heron,  M.A. McGeehin, and M. Marcus. 2011. Exposure to polyfluoroalkyl chemicals
     during pregnancy is not associated with offspring age at menarche in a contemporary
     British cohort. Environment International 37:129-135.

Cifone, M.A. 1999. UnscheduledDNA Synthesis in Rat Liver Primary Cell Cultures with PFOS.
     Final Report, Covance Study No. 207840447. Covance Laboratories Inc., Vienna, VA.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-4

-------
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. Toxicology and Applied Pharmacology 103:512-
     527.

Corsini, E., A. Avogadro, V. Galbiati, M. Dell'Agli, M. Marinovich, C.L. Galli, and D.R.
     Germolec. 2011. In vitro evaluation of the immunotoxic potential of perfluorinated
     compounds (PFCs). Toxicology and Applied Pharmacology 250(2): 108-116.

Corsini, E., R.W. Luebke, D.R. Germolec, and J.C. DeWitt. 2014. Perfluorinated compounds:
     Emerging POPs with potential immunotoxicity. Toxicology Letters 230(2):263-270.

Corsini, E., E. Sangiovanni, A. Avogadro, V. Galbiati, B. Viviani, M. Marinovich, C.L. Galli,
     M. Dell'Agli, and D.R. Germolec. 2012. In vitro characterization of the immunotoxic
     potential of several perfluorinated compounds (PFCs). Toxicology and Applied
     Pharmacology 258(2):248-255.

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.
     Archives of Environmental Contamination and Toxicology 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. Archives of Environmental
     Contamination and Toxicology 58:205-213.

Curran, L, 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). Journal of Toxicology and Environmental Health, Part A 71:1526-1541.

D'Alessandro, M.L., D.A. Ellis, J.A. Carter, N.L. Stock, and R.E. March. 2013. Competitive
     binding of aqueous perfluoroctanesulfonic acid and ibuprofen with bovine serum albumin
     studied by electrospray ionization mass spectrometry. InternationalJournal of Mass
     Spectrometry 345-347:28-36.

Dallaire, R., E. Dewailly, D. Pereg, S. Dery, and P. Ayotte. 2009. Thyroid function and plasma
     concentrations of polyhalogenated compounds in Inuit adults. Environmental Health
     Perspectives 117:1380-1386.

Dankers, A.C., M.J.E. Roelofs, A.H. Piersma, F.C.G.J. Sweep, F.G.M. Russel, M. van den
     Berg, M.B.M. van Duursen, and R. Masereeuw. 2013. Endocrine disrupters differentially
     target ATP-binding cassette transporters in the blood-testis barrier and affect Leydig cell
     testosterone secretion in vitro. Toxicological Sciences 136(2):382-391.

Darrow, L.A., C.R. Stein, and K. Steenland. 2013. Serum perfluorooctanoic acid and
     perfluorooctane sulfonate concentrations in relation to birth outcomes in the Mid-Ohio
     Valley, 2005-2010. Environmental Health Perspectives 121:1207-1213.

Darrow, L.A., P.P. Howards,  A. Winquist, and K. Steenland. 2014. PFOA and PFOS serum
     levels and miscarriage risk. Epidemiology 25:el-8.
Perfluorooctane sulfonate (PFOS) - May 2016                                              5-5

-------
Davies, B., and T. Morriss. 1993. Physiology parameters in laboratory animals and human.
     Pharmaceutical Research 10:1093-1095.

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.

DeWitt. J.C., W.C. Williams, J. Creech, andR.W. Luebke. 2015. Supression of antigen-
     specific antibody responses in mice exposed to perfluorooctanoic acid: Role of PPARa
     and T- and B-cell targeting. Journal of Immunotoxicology 13(l):38-45.

Ding, G., J. Zhang, Y. Chen, L. Wang, M. Wang, D. Xiong, and Y. Sun. 2013. Combined
     effects of PFOS and PFOA on zebrafish (Danio rerio) embryos. Archives of
     Environmental Contamination and Toxicology 64(4):668-675.

Donauer, S., A. Chen, Y. Xu, A.M. Calafat, A. Sjodin, andK. Yolton. 2015. Prenatal exposure
     to polybrominated diphenyl ethers and polyfluoroalkyl chemicals and infant
     neurobehavior. The Journal of Pediatrics 166(3):736-742.

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.
     Archives of Toxicology 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.
     Environmental Health Perspectives 121:507-513.

EFSA (European Food Safety Authority).  2008. Opinion of the scientific panel on
     contaminants in the food  chain on perfluorooctane sulfonate (PFOS), perfluorooctanoic
     acid (PFOA) and their salts.  The EFSA Journal 653:1-131.

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. Journal of Exposure Science and Environmental Epidemiology 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.

Environment Canada. 2006. Ecological Screening Assessment Report on Perfluorooctane
     Sulfonate, Its Salts and Its Precursors that Contain the C8Fi?SO2 or  CsFnSOs, or
     C8Fi?SO2N Moiety. Accessed May 2016. http://www.ec.gc.ca/lcpe-
     cepa/documents/substances/spfo-pfos/ecological_sar_pfos_eng.pdf.

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. Journal of the National Cancer
     Institute 101:605-609.
Perfluorooctane sulfonate (PFOS) - May 2016                                              5-6

-------
Eriksen, K.T., O. Raaschou-Nielsen, J.K. McLaughlin, L. Lipworth, A. Tj0nneland, K.
     Overvad, and M. S0rensen. 2013. Association between plasma PFOA and PFOS levels
     and total cholesterol in a middle-aged Danish population. PLOS ONE 8:e56969.

Fabrega, F., V. Kumar, M. Schuhmacher, J.L. Domingo, and M. Nadal. 2014. PBPK modeling
     for PFOS and PFOA: Validation with human experimental data. Toxicology Letters
     230:244-251.

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. Environmental Health
     Perspectives 115:1677-1682.

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. American Journal
     of Epidemiology 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. Environmental Health Perspectives
     116:1391-1395.

Fei, C., J.K. McLaughlin, L. Lipworth, and J. Olsen. 2009. Maternal levels of perfluorinated
     chemicals and subfecundity. Human Reproduction 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. Scandinavian Journal of Work, Environment & 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.
     Environmental Research 110:773-777.

Fei, C., and J. Olsen. 2011. Prenatal exposure to perfluorinated chemicals and behavioral or
     coordination problems at age  7. Environmental Health Perspectives 119:573-578.

Fenton, S.E. 2015. A special issue dedicated to a complex tissue. Reproductive Toxicology
     54:1-5.

Filgo, A.J., E.M. Quist, M.J. Hoenerhoff, A.E. Brix, G.E. Kissling, and S.E. Fenton. 2015.
     Perfluorooctanoic acid (PFOA)-induced liver lesions in two strains of mice following
     developmental exposures: PPARa is not required. Toxicologic Pathology 45:558-568.

Fisher, M., I.E. Arbuckle, M. Wade,  and D.A. Haines. 2013. Do perfluoroalkyl substances
     affect metabolic function and plasma lipids? - Analysis of the 2007-2009, Canadian
     Health Measures Survey (CFJJVIS) Cycle 1. Environmental Research 121:95-103.

Fitz-Simon, N., T. Fletcher, M.I. Luster, K. Steenland, A.M. Calafat,  K. Kato, and B.
     Armstrong. 2013. Reductions in serum lipids with a 4-year decline in serum
     perfluorooctanoic acid and perfluorooctanesulfonic acid. Epidemiology 24:569-576.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-7

-------
Fletcher, T., T.S. Galloway, D. Melzer, P. Holcroft, R. Cipelli, L.C. Pilling, D. Mondal, M.
     Luster, and L.W. Harries. 2013. Associations between PFOA, PFOS and changes in the
     expression of genes involved in cholesterol metabolism in humans. Environment
     International 57-58:2-10.

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. Archives of Pediatrics & Adolescent
     Medicine 164:860-869.

Fu, Y., T. Wang, Q. Fu, P. Wang, and Y. Lu. 2014. Associations between serum
     concentrations of perfluoroalkyl acids and serum lipid levels in a Chinese population.
     Ecotoxicololgy and Environmental Safety 106:246-52.

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.  Toxicological Sciences
     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. Environmental Health Perspectives 120:655-660.

Geiger, S.D., J. Xiao, A. Ducatman, S. Frisbee, K. Innes, and A. Shankar. 2014a. The
     association between PFOA, PFOS and serum lipid levels in adolescents. Chemosphere
     98:78-83.

Geiger, S.D., J. Xiao, and A.  Shankar. 2014b. No association between perfluoroalkyl
     chemicals and hypertension in children. Integrated Blood Pressure Control 7:1-7.

Genuis, S.J., D. Birkholz, M. Ralitsch, andN. Thibault. 2010. Review paper: Human
     detoxification of perfluorinated compounds. Public Health 124:367-375.

Ghisari, M., H. Eiberg, M. Long, and E.G. Bonefeld-J0rgensen. 2014. Polymorphisms in phase
     I and phase II genes and breast cancer risk and relations  to persistent organic pollutant
     exposure: A case-control study in Inuit women. Environmental Health 13(1): 19.

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.

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.

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.
Perfluorooctane sulfonate (PFOS) - May 2016

-------
Governini, L., C. Guerranti, V. De Leo, L. Boschi, A. Luddi, M. Gori, R. Orvieto, and P.
     Piomboni. 2014. Chromosomal aneuploides and DNA fragmentation of human
     spermatozoa from patients exposed to perfluorinated compounds. Andrologia 47:1012-
     1019.

Governini, L., C. Guerranti, V. De Leo, L. Boschi, A. Luddi, M. Gori, R. Orvieto, and P.
     Piomboni. 2015. Chromosomal aneuploides and DNA fragmentation of human
     spermatozoa from patients exposed to perfluorinated compounds. Andrologia 47:1012-
     1019.

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. Journal of the American Medical Association 307:391-397.

Grandjean, P., and R. Clapp. 2015. Perfluorinated alkyl substances emerging insights into
     health risks. New Solutions: A Journal of Environmental and Occupational Health Policy
     15(2): 147-163

Granum, B., L.S. Haug, E. Namork,  S.B. St01evik, C. Thomsen, IS. Aaberge, H. van Loveren,
     M. L0vik, and U.C. Nygaard. 2013. Pre-natal exposure to perfluoroalkyl substances may
     be associated with altered vaccine antibody levels and immune-related health outcomes
     in early childhood. Journal of Immunotoxicology 10:373-379.

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: Developmental and Reproductive Toxicology 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: Developmental and Reproductive Toxicology 74:405-416.

Grice, M., B. Alexander, R. Hoffbeck, and D. Kampa. 2007. Self-reported medical conditions
     in perfluorooctanesulfonyl fluoride manufacturing workers. Journal of Occupational and
     Environmental Medicine 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 Journal of
     ToxicologicalSciences 34:687-691.

Hall, A.P., C.R. Elcombe, J.R. Foster, T.  Harada, W. Kaufmann, A. Knippel, K. Kuttler, D.E.
     Malarkey, R.R. Maronpot, A. Nishikawa, T. Nolte, A. Schulte, V. Strauss, and MJ.
     York. 2012. Liver hypertrophy: A review of adaptive (adverse and non-adverse) changes
     - conclusions from the 3rd International ESTP Expert Workshop. Toxicologic Pathology
     40:971-994.

Halldorsson, T.I., D. Rytter, L.S. Haug, B.H. Bech, I. Danielsen, G. Becher, T.B. Henriksen,
     and S.F. Olsen. 2012. Prenatal  exposure to perfluorooctanoate and risk of overweight at
     20 years of age: A prospective  cohort study. Environmental Health Perspectives
     120:668-673.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-9

-------
Hamm, M., N.M. Cherry, E. Chan, J. Martin, and I. Burstyn. 2010. Maternal exposure to
     perfluorinated acids and fetal growth. Journal of Exposure Science and Environmental
     Epidemiology 20(7):589-97.

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 perfluorooctanoate
     and perfluorooctane sulfonate in humans. Environmental Toxicology and Pharmacology
     24:134-139.

Hardell, E., A. Karrman, B. van Bavel, J. Boa, M. Carlberg, and L. Hardell. 2014. Case-control
     study on perfluorinated alkyl acids (PFAAs) and the risk of prostate cancer. Environment
     International 63:35-39.

Higgins, C.P. and R.G. Luthy. 2006. Sorption of perfluorinated surfactants on  sediments.
     Environmental Science &  Technology 40:7251-7256.

Hoffman, K., T.F. Webster, M.G. Weisskopf, J. Weinberg, and V.M. Vieira. 2010. Exposure to
     polyfluoroakyl chemicals and attention deficit hyperactivity disorder in U.S. children
     ages 12-15 years. Environmental Health Perspectives 118:1732-1767.

H0yer, B.B., C.H. Ramlau-Hansen, M. Vrijheid, D. Valvi, H.S. Pedersen, V. Zviezdai, B.A.G.
     Jonsson, C.H. Lindh, J.P. Bonde, and G. Toft. 2015a. Anthropometry in 5- to 9-year old
     Greenlandic and Ukrainian children in relation to prenatal exposure to perfluorinated
     alkyl substances. Environmental Health Perspectives 123:841-846.

H0yer, B.B., C.H. Ramlau-Hansen, C. Obel, H.S. Pedersen, A. Hernik, V. Ogniev, B.A.G.
     Jonsson, C.H. Lindh, L. Rylander, A. Rignell-Hydbom, J.P.  Bonde, and G. Toft. 2015b.
     Pregnancy serum concentrations of perfluorinated alkyl substances and offspring
     behavior and motor development at age 5-9 years - a prospective study.  Environmental
     Health 14:2.

HSDB (Hazardous Substances Data Bank). 2012. Hazardous  Substances Data  Bank. National
     Institutes of Health, Health & Human Services, U.S. National Library of Medicine,
     Bethesda, MD. Accessed May 2016.
     http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen7HSDB.

Hu, W., P.O. Jones, B.L. Upham, J.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.
     ToxicologicalSciences 68:429-436.

Humblet, O., L.G. Diaz-Ramirez, J.R. Balmes, S.M. Pinney, and R.A. Hiatt. 2014.
     Perfluoroalkyl chemicals and asthma among children 12-19 years of age: NHANES
     (1999-2008). Environmental Health Perspectives 122:1129-1133.

Innes, K.E., J.H. Wimsatt, S. Frisbee, and A.M. Ducatman. 2014.  Inverse association of
     colorectal cancer prevalence to serum levels of perfluorooctane sulfonate (PFOS) and
     perfluorooctanoate (PFOA) in a large Appalachian population. BMC Cancer 14:45.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-10

-------
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.
     Environmental Health Perspectives 112:1204-1207.

Jain, R.B. 2014. Contribution of diet and other factors to the levels of selected perfluorinated
     compounds: Data from NHANES 2003-2008. InternationalJournal of Hygiene and
     Environmental Health 217:52-61.

Jain, R.B., 2015. Estimation of the total concentration of perfluoroalkyl acids (PFAA) in
     human serum: Data from NHANES 2005-2012. Chemosphere, 134: 387-394.

Joensen, U.N., R. Bossi, H. Leffers, A.A. Jensen, N. Skakkebaek, and N. J0rgensen. 2009. Do
     perfluoroalkyl compounds impair human semen quality? Environmental Health
     Perspectives 117:923-927.

Joensen, U.N., B. Veyrand, J.-P. Antignac, M.B. Jensen, J.H. Petersen, P. Marchand, N.E.
     Skakkebaek, A.-M. Andersson, B. Le Bizec, andN. J0rgensen. 2013. PFOS
     (perfluorooctanesulfonate) in serum is negatively associated with testosterone levels, but
     not with semen quality, in healthy men. Human Reproduction 28:599-608.

Johansson, N., A. Fredriksson, and P.  Eriksson. 2008. Neonatal exposure to perfluorooctane
     sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural defects in
     adult mice. NeuroToxicology 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. ToxicologicalSciences 108:412-418.

J0rgensen, K.T., I.O. Specht, V. Lenters, C.C. Bach, L. Rylander, B.A. Jonsson, C.H. Lindh,
     A. Giwercman, D. Heederik, G.  Toft, and J.P. Bonde. 2014. Perfluoroalkyl substances
     and time to pregnancy in couples from Greenland, Poland and Ukraine. Environmental
     Health 13(1): 116.

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. Environmental Science and Pollution Research
     17:750-758.

Keil, D., T. Mehlmann, L. Butterworth, and M. Peden-Adams. 2008. Gestational exposure to
     perfluorooctane sulfonate suppresses immune function in B6C3F1 mice. Toxicological
     Sciences 103:77-85.

Kerger, B.D., T.L. Copeland, and A.P. DeCaprio. 2011. Tenuous dose-response correlations
     for common disease states: Case study of cholesterol and perfluorooctanoate/sulfonate
     (PFOA/PFOS) in the C8 Health  Project. Drug and Chemical Toxicology 34:396-404.

Kerstner-Wood, C., L. Coward, and G. Gorman. 2003. Protein Binding ofPerfluorohexane
     Sulfonate,  Perfluorooctane Sulfonate and Perfluorooctanoate to Plasma (Human, Rat
     and Monkey) and Various Human-Derived Plasma Protein Fractions. Study ID 9921.7.
     Southern Research Institute.

Perfluorooctane sulfonate (PFOS) - May 2016                                             5-11

-------
Khansari, N., Y. Shakiba, and M. Mahmoudi. 2009. Chronic inflammation and oxidative stress
     as a major cause of age-related diseases and cancer. Recent Patents on Inflammation &
     Allergy Drug Discovery 3(1):73-80.

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, and J.Y. Ha. 2011. Distribution of perfluorochemicals between sera and milk
     from the same mothers and implications for prenatal and postnatal exposures.
     Environmental Pollution 15 9( 1): 169-174.

Kjeldsen, L.S., and E.G. Bonefeld-J0rgensen. 2013. Perfluorinated compounds affect the
     function of sex hormone receptors. Environmental Science and Pollution Research
     20(11):8031-8044.

Klaassen, C.D., and L.M. Aleksunes. 2010. Xenobiotic, bile acid, and cholesterol transporters:
     Function and regulation. Pharmacological Reviews 62:1-96.

Kleinstreuer, N.C., J. Yang, E.L. Berg, T.B. Knudsen, A.M. Richard, M.T. Martin, D.M. Reif,
     R.S. Judson, M. Polokoff, DJ. Dix, and R.J. Kavlock. 2014. Phenotypic screening of the
     ToxCast chemical library to classify toxic and therapeutic mechanisms. Nature
     Biotechnology 32(6):583-591.

Knox, S.S., T. Jackson, B. Javins, S.J. Frisbee, A. Shankar, and A.M. Ducatman. 2011.
     Implications of early menopause in women exposed to perfluorocarbons. The Journal of
     Clinical Endocrinology & Metabolism 96:1747-1753.

Koustas, E.,  J. Lam, P. Sutton, P.I. Johnson, D.S. Atchley, S. Sen, K.A. Robinson, D.A.
     Axelrad, and T.J. Woodruff. 2014. The Navigation Guide - evidence-based medicine
     meets environmental health: systematic review of nonhuman evidence for PFOA effects
     on fetal growth. Environmental Health Perspectives 122:1015-1027.

Kristensen, S.L., C.H. Ramlau-Hansen,  E. Ernst, S.F. Olsen, J.P. Bonde, A. Vested, T.I.
     Halldorsson, G.  Becher, L.S. Haug, and G. Toft. 2013. Long-term effects of prenatal
     exposure to perfluoroalkyl substances on female reproduction. Human Reproduction
     28(12):3337-48.

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. Environmental Science & Technology 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. ToxicologicalSciences 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. Toxicological Sciences
     99:366-394.

Lewis, R.J.,  Sr., ed. 2004. Sax's Dangerous Properties of Industrial Materials, llth ed. Wiley-
     Interscience, Wiley & Sons, Inc., Hoboken, NJ.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-12

-------
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. Environmental Science &
     Technology 43:2099-2104.

Liew, Z., B. Ritz, B.C. Bonefeld-J0rgensen, T.B. Henriksen, E.A. Nohr, B.H. Bech, C. Fei, R.
     Bossi, O.S. von Ehrenstein, E. Streja, P. Uldall, and J. Olsen. 2014. Prenatal exposure to
     perfluoroalkyl substances and the risk of congenital cerebral palsy in children. American
     Journal of Epidemiology 180:574-581.

Lin, C.-Y.,  Y.-C. Lin, P.-C. Chen, and L.-Y. Lin. 2009. Association among serum
     perfluoroalkyl chemicals, glucose homeostasis, and metabolic syndrome in adolescents
     and adults. Diabetes Care 32:702-707.

Lin, C.-Y.,  L.-Y. Lin, C.-K. Chiang, W.-J. Wang, Y.-N. Su, K.-Y. Hung, and P.-C. Chen.
     2010. Investigation of the associations between low-dose serum perfluorinated chemicals
     and liver enzymes in US  adults. The American Journal of Gastroenterology 105:1354-
     1363.

Litton Bionetics, Inc. 1979. Mutagenicity Evaluation ofT-2014 CoC in the Ames
     Salmonella/Microsome Plate Test. Final Report. LBI Project No. 20838. Litton
     Bionetics, Inc.

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 Journal of Toxicological
     Sciences 34:245-254.

Liu, W., B. Yang, L. Wu, W. Zou, X. Pan, T. Zou, F. Liu, L. Xia, X. Wan, and D. Zhang.
     2015. Involvement of NRF2 in perfluorooctanoic acid-induced testicular damage in male
     mice. Biology of Reproduction 93(2): 1-7.

Loccisano,  A.E., J.L.  Campbell, M.E. Andersen, andH.J. Clewell. 2011. Evaluation and
     prediction of pharmacokinetics of PFOA and PFOS in the monkey and human using a
     PBPK model. Regulatory Toxicology and Pharmacology 59:157-175.

Loccisano,  A.E., J.L.  Campbell, J.L. Butenhoff, M.E. Andersen, andH.J. Clewell. 2012a.
     Comparison and evaluation of pharmacokinetics of PFOA and PFOS in the adult rat
     using a physiologically based pharmacokinetic model. Reproductive Toxicology 33:452-
     467.

Loccisano,  A.E., J.L.  Campbell, J.L. Butenhoff, M.E. Andersen, andH.J. 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. Reproductive Toxicology 33:468-490.

Loccisano,  A.E., M.P. Longnecker, J.L. Campbell, Jr., M.E. Andersen, andH.J. Clewell. 2013.
     Development of PBPK models for PFOA and PFOS for human pregnancy and lactation
     life stages. Journal of Toxicology and Environmental Health, Part A 76:25-57'.

Long, Y., Y. Wang, G. Ji, L. Yan, F. Hu, and A. Gu. 2013. Neurotoxicity of perfluorooctane
     sulfonate to hippocampal cells in adult mice. PLOS ONE 8(l):e54176.

Perfluorooctane sulfonate (PFOS) - May 2016                                            5-13

-------
Looker, C., M.I. Luster, A.M. Calafat, VJ. Johnson, G.R. Burleson, F.G. Burleson, and T.
     Fletcher. 2014. Influenza vaccine response in adults exposed to perfluorooctanoate and
     perfluorooctanesulfonate.  ToxicologicalSciences 138:76-88.

Lopez-Doval, S., R. Salgado, N. Pereiro, R. Moyano, and A. Lafuente. 2014. Perfluorooctane
     sulfonate effects on the reproductive axis in adult male rats. Environmental Research
     134:158-168.

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. Environmental Science & Technology 45:8160-8166.

Lu, Y., B. Luo, J. Li, and J. Dai. 2015. Perfluorooctanoic acid disrupts the blood-testes barrier
     and activates TNFa/p38 MAPK signaling pathway in vivo and in vitro. Archives of
     Toxicology 90(4):971-983.

Luebker, D.J., K. J. 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.

Lv, Z., G. Li, Y. Li, C. Ying, J.  Chen, T. Chen, J. Wei,  Y. Lin, Y. Jiang, Y. Wang, B. Shu, B.
     Xu, and S. Xu. 2013. Glucose and lipid homeostasis in adult rat  is impaired by early-life
     exposure to perfluorooctane sulfonate. Environmental Toxicology 28:532-542.

Maisonet, M., M.L. Terrell, M.A. McGeehin, K.Y. Christensen, A. Holmes, A.M. Calafat, and
     M. Marcus. 2012. Maternal concentrations of polyfluoroalkyl compounds during
     pregnancy and fetal and postnatal growth in British girls. Environmental Health
     Perspectives 120(10): 143 2.

Maisonet, M., S. Nayha, D.A. Lawlor, and M. Marcus.  2015. Prenatal exposures to
     perfluoroalkyl acids and serum lipids at ages 7 and 15 in females. Environment
     International 82:49-60.

Maloney, E.K., and D.J. Waxman. 1999. trans-Activation of PPARa and PPARy by
     structurally diverse environmental chemicals. Toxicology and Applied Pharmacology
     161:209-218.

Mandel, J., and R. Johnson.  1995. Mortality Study of Employees at 3M Plant in Decatur,
     Alabama. University of Minnesota, School of Public Health, Division of Environmental
     and Occupational Health,  Minneapolis, MN.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-14

-------
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. Toxicological Sciences 97:595-613.

Matyszewska, D., J. Leitch, R. Bilewic, and J. Lipkowski. 2008. Polarization modulation
     infrared reflection - absorption spectroscopy studies of the influence of perfluorinated
     compounds on the properties of a model biological membrane. Langmuir 24:7408-7412.

Mecchi, M.S. 1999. Salmonella-Escherichia co\i-Mammalian-Microsome Reverse Mutation
     Assay with PFOS. Final Report. Covance Study No. 20784-0-409. Covance Laboratories,
     Vienna, VA.

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.
     Environmental Health Perspectives 118:686-692.

Midgett, K., M.M. Peden-Adams, G.S. Gilkeson, and D.L. Kamen. 2014. In vitro evaluation of
     the effects of perfluorooctanesulfonic acid (PFOS) and perfluorooctanoic acid (PFOA) on
     OL-2 production in human T-cells. Journal of Applied Toxicology 35(5):459-465.

Miller, R.T., L.A. Scappino, S.M. Long, and J.C. Gorton. 2001. Role of thyroid hormones in
     hepatic effects of peroxisome proliferators. Toxicologic Pathology 29:149-155.

Mondal, D., MJ. Lopez-Espinosa, B. Armstrong, C.R. Stein, and T. Fletcher. 2012.
     Relationships of perfluorooctanoate and perfluorooctane sulfonate serum concentrations
     between mother-child pairs in a population with perfluorooctanoate exposure from
     drinking water. Environmental Health Perspectives 120(5):752.

Mondal, D., R.H. Weldon, E.G. Armstrong, LJ. Gibson, MJ. Lopez-Espinosa, H.M. Shin, and
     T. Fletcher. 2014. Breastfeeding: A potential excretion route for mothers and
     implications for infant exposure to perfluoroalkyl acids. Environmental Health
     Perspectives 122(2): 187-192.

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. Environmental Research 108:56-62.

Montgomery, R., T.W. Conway, and A.A. Spector. 1990. Biochemistry: A Case-Oriented
     Approach. 5th ed. The C.V. Mosby Company,  St. Louis, MO.

M0rck, T.A., F. Nielsen, J.K.  Nielsen, V.D. Siersma, P. Grandjean, and L.E. Knudsen. 2015.
     PFAS concentrations in plasma  samples from Danish school children and their mothers.
     Chemosphere 129:203-209.

Morken, N.-H.,  G.S. Travlos, R.E. Wilson, M. Eggesb0, and M.P. Longnecker. 2014. Maternal
     glomerular filtration rate in pregnancy and fetal size. PLOS ONE 9:el01897.

Murli, H. \996.Mutagenicity Test on T-6295 in an In-vivoMouse Micronucleus Assay. Final
     Report. CHV Study No. 17403-0-455. Corning Hazelton Inc., Vienna, VA.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-15

-------
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.

NCBI (National Center for Biotechnology Information). 2016. ESR1 estrogen receptor 1
     [Homo sapiens (human)]. National Institutes of Health, National Center for
     Biotechnology Information, Washington, DC. Accessed May 2016.
     http://www.ncbi.nlm.nih.gov/gene/2099.

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.
     Environmental Health Perspectives 118:197-202.

Ngo, H.T., R.B. Hetland, A. Sabaredzovic, L.S. Haug, and I.-L.  Steffensen. 2014. In utero
     exposure to perfluorooctanoate (PFOA) or perfluorooctane sulfonate (PFOS) did not
     increase body weight or intestinal tumorigenesis in multiple intestinal neoplasia (Minl+)
     mice. Environmental Research 132:251-263.

Nohr, E.A., M. Frydenberg, T.B. Henriksen, and J. Olsen. 2006. Does low participation in
     cohort studies induce bias? Epidemiology 17:413-418.

Noker, P., and G. Gorman. 2003. A Pharmacokinetic Study of Potassium
     Perfluorooctane sulfonate in the Cynomolgus Monkey. U.S. Environmental Protection
     Agency, Washington, DC. U.S. EPA docket AR-226-1356.

NRC (National Research Council). 1983. Risk Assessment in the Federal Government:
     Managing the Process. National Research Council, Committee on the Institutional Means
     for Assessment of Risks to Public Health, Commission on Life Sciences.  National
     Academy Press, Washington, DC.

OECD (Organisation for Economic Co-operation and Development). 2002. Hazard Assessment
     of Perfluorooctane Sulfonate (PFOS) and its Salts. ENV/JM/RD(2002)17/FINAL.
     Report of the Environment 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. Environmental Research 112:118-125.

Okada, E., S. Sasaki, I. Kashino, H. Matsuura,  C. Miyashita, S. Kobayashi, K. Itoh, T. Ikeno,
     A. Tamakoshi, and R. Kishi. 2014. Prenatal exposure to perfluoroalkyl acids and allergic
     diseases in early childhood. Environment International 65:127-134.

Olsen, G.W., M.M. Burlew, J.M. Burris, and J.H. Mandel. 2001a. 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 andDecatur Fluorochemical
     Medical Surveillance Program. Final Report. 3M Company,  St. Paul, MN.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-16

-------
Olsen, G.W., M.M. Burlew, J.M. Burris, and J.H. Mandel. 2001b. A Longitudinal Analysis of
     Serum Perfluorooctane Sulfonate (PFOS) andPerfluorooctanoate (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. Final Report. 3M Company, St. Paul, MN.

Olsen, G., K. Hansen, L.  Stevenson, J. Burris, and J. Mandel. 2003a. Human donor liver and
     serum concentrations of perfluorooctanesulfonate and other perfluorochemicals.
     Environmental Science & Technology 37:888-891.

Olsen, G., J.M. Burris, M.M. Burlew, and J.H. Mandel. 2003b. Epidemiologic assessment of
     worker serum perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA)
     concentrations and medical surveillance examinations. Journal of Occupational and
     Environmental Medicine 45:260-270.

Olsen, G.W., J.M. Burris, D.J.  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. Environmental Health Perspectives 115:1298-1305.

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.
     Reproductive Toxicology 27:212-230.

Ospinal-Jimenez, M., and D.C. Pozzo. 2012. Structural analysis of protein denaturation with
     alkyl perfluorinated sulfonates. Langmuir 28:17749-17760.

Osuna, C.E., P. Grandjean, P. Weihe, and H.A. El-Fawal. 2014. Autoantibodies associated
     with prenatal and childhood exposure to environmental chemicals in Faroese children.
     Toxicological Sciences 142(1): 158-66.

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. Toxicological Sciences 104:144-154.

Perez, F., M. Nadal, A. Navarro, F. Fabrega, J. Domingo, J.L. Barcelo,  D. Barcelo, and M.
     Farre. 2013. Accumulation of perfluoroalkyl  substances in human tissues. Environment
     International 59:354-362.

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.

Post, G.B., P.D. Cohn, and K.R. Cooper. 2012. Perfluorooctanoic acid (PFOA), an emerging
     drinking water contaminant: a critical review of recent literature. Environmental research
     116:93-117.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-17

-------
Qazi, M.R., Z. Xia, J. Bogdanska, S.-C. Chang, DJ. 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. International
     Immunopharmacology 10:1420-1427.

Quist, E.M., A.J. Filgo, C.A. Cummings, G.E. Kissling, andM.J. Hoenerhoff  2015. Hepatic
     mitochondrial alteration in CD-I mice  associated with prenatal exposures to low doses of
     perfluorooctanoic acid (PFOA). Toxicologic Pathology 41:546-557.

Rao, M.S., and J.K. Reddy. 1996. Hepatocarcinogenesis of the peroxisome proliferators.
     Annals of the New York Academy of Sciences 804:573.

Raymer, J.H., L.C.  Michael, W.B. Studabaker, G.W. Olsen, C.S. Sloan, T. Wilcosky, and O.K.
     Walmer. 2012. Concentrations of perfluorooctane sulfonate (PFOS) and
     perfluorooctanoate (PFOA) and their associations with human semen quality
     measurements. Reproductive Toxicology 33(4):419-427.

Ren, X.-M., Y.-F. Zhang, L.-H. Guo, Z.-F. Qin, Q.-Y. Lv, andL.-Y. Zhang. 2015. Structure-
     activity relations in binding of perfluoroalkyl compounds to human thyroid hormone T3
     receptor. Archives of Toxicology 89:233-242.

Rigden, M., G. Pelletier, R. Poon, J.  Zhu, C. Auray-Blais, R. Gagnon, C. Kubwabo, I. Kosarac,
     K. Lalonde, S. Cakmak, and B. Xiao. 2015. Assessment of urinary metabolite excretion
     after rat acute exposure  to perfluorooctanoic acid and other peroxisomal  proliferators.
     Archives of Environmental Contamination and Toxicology 68(1): 148-158.

Rosen, M.B., J.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.
     Reproductive Toxicology 27:278-288.

Rosen, M.B., J.R. Schmid, J.C. Corton, 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 Research Volume 2010, Article ID
     794739, 23 pp.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-18

-------
Roth, N., and M.F. Wilks. 2014. Neurodevelopmental and neurobehavioural effects of
     polybrominated and perfluorinated chemicals: A systematic review of the
     epidemiological literature using a quality assessment scheme. Toxicology Letters
     230(2):271-281.

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.

Salvalaglio, M., I. Muscionico, and C. Cavallotti. 2010. Determination of energies and sites of
     binding of PFOA and PFOS to human serum. The Journal of Physical Chemistry B
     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 Journal of Toxicological Sciences 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.  Toxicological Sciences 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.

Shabalina, I.G.,  A.V. Kalinovich, B. Cannon, and J. Nedergaard. 2015 (epub). Metabolically
     inert perfluorinated fatty acids directly activate uncoupling protein 1 in brown-fat
     mitochondria. Archives of Toxicology 90(5): 1117-28.

Shankar, A., J. Xiao, and A. Ducatman. 2011. Perfluoroalkyl chemicals and chronic kidney
     disease in  US adults. American Journal of Epidemiology 174(8):893-900.

Sheng, N., J. Li, H. Liu, A. Zhang, and J. Dai. 2016. Interaction of perfluoroalkyl acids with
     human liver fatty acid-binding protein. Archives of Toxicology 90(l):217-227.

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. Toxicological Sciences 80:151-160.

Shrestha, S., M.S. Bloom, R. Yucel, R.F. Seegal, Q. Wu, K. Kannan, R. Rej, and E.F.
     Fitzgerald. 2015. Perfluoroalkyl substances and thyroid function in older adults.
     Environment International 75:206-214.

Simmon, V.F. 1978. In-vhro MicrobiologicalMutagenicity Assays of 3M Company
     Compounds T-2247 CoC and T-2248 CoC. Final Report. SRI Project LSC-4442-016.
     SRI International,  Menlo Park, CA.

Slitt, A.L., K. Allen, J. Morrone, L.M. Aleksunes, C. Chen, J.M. Maher, J.E. Manautou, N.J.
     Cherrington, and C.D. Klaassen. 2007. Regulation of transporter expression in mouse
     liver, kidney, and intestine during extrahepatic cholestasis. Biochimica etBiophysica
     Acta (BBA)-Biomembranes 1768(3):637-647.
Perfluorooctane sulfonate (PFOS) - May 2016                                             5-19

-------
Slotkin, T., E. MacKillop, R. Melnick, K. Thayer, and F. Seidler. 2008. Developmental
     neurotoxicity of perfluorinated chemicals modeled in vitro. Environmental Health
     Perspectives 116:716-722.

Specht, I.O., K.S. Hougaard, M. Spano, D. Bizzaro, G.C. Manicardi, C.H. Lindh, G. Toft, B.A.
     Jonsson, A. Giwercman, and J.P. Bonde. 2012. Sperm DNA integrity in relation to
     exposure to environmental perfluoroalkyl substances—a study of spouses of pregnant
     women in three geographical regions. Reproductive  Toxicology 33:577-583.

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. Environmental Science &
     Technology 42:5361-5367.

SRC (Syracuse Research Corporation). 2016. PHYSPROP Database. SRC, Inc., North
     Syracuse, NY. Accessed May 2016.
     http://www.srcinc.com/what-we-do/environmental/scientific-databases.html.

Starkov, A.A., and K.B. Wallace. 2002. Structural determinations of fluorochemical-induced
     mitochondrial dysfunction. ToxicologicalSciences 66:244-252.

Starling, A.P., S.M. Engel, K.W. Whitworth, D.B. Richardson, A.M. Stuebe, J.L. Daniels, L.S.
     Haug, M. Eggesb0, G. Becher, A. Sabaredzovic, C. Thomsen, R.E. Wilson, G.S. Travlos,
     J.A. Hoppin, D.D. Baird, and M.P. Longnecker. 2014. Perfluoroalkyl substances and
     lipid concentrations in plasma during pregnancy among women in the Norwegian Mother
     and Child Cohort Study. Environment International 62:104-112.

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. American Journal of Epidemiology 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. Environmental Health
     Perspectives 118:229-23 3.

Steenland, K., L. Zhao, and A.  Winquist. 2015. A cohort incidence study of workers exposed
     to perfluoroctanoic acid (PFOA). Occupational & Environmental Medicine 0:1-8.

Stein, C.R., D.A.  Savitz, and M. Dougan. 2009. Serum levels of perfluorooctanoic  acid and
     perfluorooctane sulfonate and pregnancy outcome. American Journal of Epidemiology
     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. Reproductive Toxicology 34:312-316.

Stein, C.R., K.J. McGovern, A.M. Pajak, PJ.  Maglione, and M.S. Wolff. 2015 (epub).
     Perfluoroalkyl and polyfluoroalkyl substances and indicators of immune function in
     children aged 12-19 y: National Health and Nutrition Examination Survey. Pediatric
     Research 79:348-357.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-20

-------
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. lexicological Sciences 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. Toxicology
     Letters 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). Environmental Science & Technology
     46:12170-12177.

Tan, X., G. Xie, X. Sun, Q. Li, W. Zhong, P. Oiao, X. Sun, W. Jai, and Z. Zhou. 2013. High fat
     diet feeding exaggerates perfluorooctanoic acid-induced liver injury in mice via
     modulating multiple metabolic pathways. PLOS One  8(4):e61409.

Taylor, K.W., K. Hoffman, K.A. Thayer, and J.L. Daniels.  2014. Polyfluoroalkyl chemicals
     and menopause among women 20-65 years of age (NHANES). Environmental Health
     Perspectives 122:145-150.

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. Toxicological
     Sciences 74:369-381.

Thomford, P.J. 2002. 104-Week Dietary Chronic Toxicity and Carcinogenicity Study with
     Perfluorooctane Sulfonic Acid Potassium Salt (PFOS; T-6295) in Rats. Final Report.
     Volumes I-IX.  Covance Study No. 6329-183. 3M Company, St. Paul, MN.

Thompson, J., 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. Environment International 36:392-397.

Timmermann, C.A.,  L.I. Rossing, A. Gr0ntved, M. Ried-Larsen, C. Dalgard, L.B. Andersen, P.
     Grandjean, F. Nielsen,  K.D. Svendsen, T.  Scheike, and T.K.  Jensen. 2014. Adiposity and
     glycemic control in children exposed to perfluorinated compounds. The Journal of
     Clinical Endocrinology & Metabolism 99(4):E608-14.

Toft, G., B.A.G. Jonsson, C.H. Lindh, A. Giwercman, M. Spano, D. Heederik, V. Lenters, R.
     Vermeulen, L.  Rylander, H.S. Pedersen, and J.K. Ludwicki. 2012. Exposure to
     perfluorinated compounds and human semen quality in Arctic and European populations.
     Human Reproduction 27(8):2532-2540.

Tucker, D.E., M.B. Macon, M.J. Strynar, S. Dragnino, E. Andersen, and S.E. Fenton. 2015.
     The mammary  gland is a pensitive pubertal target in CD-I and C57BL/6 mice following
     perinatal perfluorooctamoic acid (PFOA) exposure. Reproductive Toxicology 54:26-36.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-21

-------
UNEP (United Nations Environment Programme). 2006. Report of the Persistent Organic
     Pollutants Review Committee on the Work of its Second Meeting. Addendum: Risk
     Profile on Perfluorooctane Sulfonate. UNEP/POPS/POPRC.2/17/Add.5. United Nations
     Environment Programme. Accessed May 2016.
     http://chm.pops.int/Default.aspx?tabid=2301.

USEPA (U.S. Environmental Protection Agency).  1986a. Guidelines for the Health Risk
     Assessment of Chemical Mixtures. U.S. Environmental Protection Agency. Federal
     Register 51(185):34014-34025.

USEPA (U.S. Environmental Protection Agency).  1986b. Guidelines for Mutagenicity Risk
     Assessment. U.S. Environmental Protection Agency. Federal Register 51(185):34006-
     34012.

USEPA (U.S. Environmental Protection Agency).  1988. Recommendations for and
     Documentation of Biological Values for Use in Risk Assessment. EPA 600/6-87/008.
     U.S. Environmental Protection Agency, Washington, DC.

USEPA (U.S. Environmental Protection Agency).  1991. Guidelines for Developmental
     Toxicity Risk Assessment. U.S. Environmental Protection Agency. Federal Register
     56(234):63798-63826.

USEPA (U.S. Environmental Protection Agency).  1994a. Interim Policy for Particle Size and
     Limit Concentration Issues in Inhalation Toxicity Studies. U.S. Environmental Protection
     Agency. Federal Register 59(206):53799.

USEPA (U.S. Environmental Protection Agency).  1994b. Methods for Derivation of Inhalation
     Reference Concentrations and Application of Inhalation Dosimetry. EPA/600/8-90/066F.
     U.S. Environmental Protection Agency, Washington, DC. Accessed May 2016.
     https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=71993&CFID=57984601&CFTO
     KEN=763 87071.

USEPA (U.S. Environmental Protection Agency).  1995. Use of the Benchmark Dose Approach
     in Health Risk Assessment. EPA/630/R-94/007. U.S. Environmental Protection Agency,
     Washington, DC. Accessed May 2016.
     hero.epa.gov/index.cfm/reference/download/reference id/5992.

USEPA (U.S. Environmental Protection Agency).  1996. Guidelines for Reproductive Toxicity
     Risk Assessment. U.S. Environmental Protection Agency. Federal Register
     61(212):56274-56322. Accessed May 2016.
     https://www.gpo.gov/fdsvs/pkg/FR-1996-10-31/pdf/96-27473.pdf.

USEPA (U.S. Environmental Protection Agency).  1998. Guidelines for Neurotoxicity Risk
     Assessment. U.S. Environmental Protection Agency. Federal Register 63(93):26926-
     26954. Accessed May 2016.
     https://www.epa.gov/sites/production/files/2014-11/documents/neuro  tox.pdf.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-22

-------
USEPA (U.S. Environmental Protection Agency). 2000a. Science Policy Council Handbook:
     Peer Review. 2nd ed. EPA 100-B-OO-OOl. U.S. Environmental Protection Agency,
     Office of Science Policy, Office of Research and Development, Washington, DC.
     Accessed May 2016.
     http://nepis.epa.gov/EP A/html/DLwait.htm?url=/Exe/ZvPDF.cgi/50000UAG.PDF?Doc
     kev=5 OOOOUAG.PDF.

USEPA (U.S. Environmental Protection Agency). 2000b. Supplemental Guidance for
     Conducting for Health Risk Assessment of Chemical Mixtures. EPA/630/R-00/002.
     U.S. Environmental Protection Agency, Washington, DC. Accessed May 2016.
     https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=20533.

USEPA (U.S. Environmental Protection Agency). 2002. A Review of the Reference Dose and
     Reference Concentration Processes. EPA/630/P-02/0002F. U.S. Environmental
     Protection Agency, Washington, DC. Accessed May 2016.
     https://www. epa.gov/osa/review-reference-dose-and-reference-concentration-processes.

USEPA (U.S. Environmental Protection Agency). 2005a. Guidelines for Carcinogen Risk
     Assessment. EPA/630/P-03/001B. U.S. Environmental Protection Agency, Washington,
     DC. Accessed May 2016.
     https://www.epa.gov/risk/guidelines-carcinogen-risk-assessment.

USEPA (U.S. Environmental Protection Agency). 2005b. Supplemental Guidance for
     Assessing Susceptibility from Early-Life Exposure to Carcinogens. EPA/630/R-03/003F.
     U.S. Environmental Protection Agency, Washington, DC. Accessed May 2016.
     https://www.epa.gov/risk/supplemental-guidance-assessing-susceptibility-early-life-
     exposure-carcinogens.

USEPA (U.S. Environmental Protection Agency). 2006a. Peer Review Handbook: 3rdEdition.
     EPA/100/B-06/002. U.S. Environmental Protection Agency, Science Policy Council,
     Washington, DC. Accessed May 2016. https://www.epa.gov/sites/production/files/2015-
     09/documents/peer  review handbook 2006  3rd  edition.pdf.

USEPA (U.S. Environmental Protection Agency). 2006b. A Framework for Assessing Health
     Risks of Environmental Exposures to Children. EPA/600/R-05/093F. U.S. Environmental
     Protection Agency, National Center for Environmental Assessment, Washington, DC.
     Accessed May 2016. https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=l58363.

USEPA (U.S. Environmental Protection Agency). 2011. Exposure Factors Handbook: 2011
     Edition. EPA/600/R-090/052F. U.S. Environmental Protection Agency, National Center
     for Environmental Assessment, Washington, DC. Accessed May 2016.
     https ://cfpub. epa.gov/ncea/risk/recordisplay. cfm?deid=236252.

USEPA (U.S. Environmental Protection Agency). 2012. Benchmark Dose Technical Guidance.
     EP A/100/R-12/001. U.S. Environmental Protection Agency, Washington, DC.

USEPA (U.S. Environmental Protection Agency). 2014a. Child-Specific Exposure Scenarios
     Examples (FinalReport). EPA/600/R-14-217F. U.S. Environmental Protection Agency,
     Washington, DC.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-23

-------
USEPA (U.S. Environmental Protection Agency). 2014b. Emerging Contaminants Fact Sheet
     - Perfluorooctane Sulfonate (PFOS) andPerfluorooctanoic Acid (PFOA). EPA 505-F-
     14-001. U.S. Environmental Protection Agency, Washington, DC. Accessed May 2016.
     http://nepis.epa.gov/Exe/ZvPDF.cgi/P100LTG6.PDF?Dockev=P100LTG6.PDF.

USEPA (U.S. Environmental Protection Agency). 2014c. Framework for Human Health Risk
     Assessment to Inform Decision Making. EPA/100/R-14/001. U.S. Environmental
     Protection Agency, Washington, DC. Accessed May 2016.
     https://www.epa.gov/risk/framework-human-health-risk-assessment-inform-decision-
     making.

USEPA (U.S. Environmental Protection Agency). 2015. ToxCast & Tox21 Summary Files
     from invitrodb_v2. U.S. Environmental Protection Agency, Washington, DC. Accessed
     May 2016. http://www2.epa.gov/chemical-research/toxicity-forecaster-toxcasttm-data.

Vanden Heuvel, J.P. 2013. Comment on "Associations between PFOA, PFOS and  changes in
     the expression of genes involved in cholesterol metabolism in humans" by Fletcher et al.,
     Environment International 57-58 (2013) 2-10. Environment International 61:150-153.

Vassiliadou, I, D. Costopoulou, A. Ferderigou, and L. Leondiadis. 2010. Levels of
     perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) in blood samples from
     different groups of adults living in Greece. Chemosphere 80:1199-1206.

Velez, M.P., I.E. Arbuckle, and W.D. Fraser. 2015. Maternal exposure to perfluorinated
     chemicals and reduced fecundity: The MIREC study. Human Reproduction 30:701-709.

Verner, M.A., and M.P. Longnecker. 2015. Comment on "Enhanced elimination of
     perfluorooctanesulfonic acid by menstruating women: Evidence from population-based
     pharmacokinetic modeling. Environmental Science &  Technology 49(9):5836-5837.

Verner, M.-A., A.E. Loccisano, N.-H. Morken, M. Yoon, H. Wu, R. McDougall, M. Maisonet,
     M. Marcus, R. Kishi, C. Miyashita, M.-H.  Chen, W.-S. Hsieh, M.E. Andersen, H.J.
     Clewell, and M.P. Longnecker. 2015. Associations of perfluoroalkyl substances (PFASs)
     with lower birth weight: An evaluation of potential confounding by glomerular filtration
     rate using a physiologically based pharmacokinetic model (PBPK). Environmental
     Health Perspectives 123:1317-13 24.

Vested, A., C.H. Ramlau-Hansen, S.F. Olsen, J.P. Bonde, S.L. Kristensen, T.I. Halldorsson, G.
     Becher, L.S. Haug, E.H. Ernst, and G. Toft. 2013. Associations of in utero exposure to
     perfluorinated alkyl acids with human semen quality and reproductive hormones in adult
     men. Environmental Health Perspectives 121:453-458.

Vestergaard, S., F. Nielsen, A.M. Andersson, N.H. Hj011und, P. Grandjean, H.R. Andersen,
     and T.K. Jensen. 2012. Association between perfluorinated compounds and time to
     pregnancy in a prospective cohort of Danish couples attempting to conceive. Human
     Reproduction 27(3):873-880.

Vesterinen, H.M., P.I. Johnson, D.S. Atchley, P. Sutton, J. Lam, M.G. Zlatnik, S. Sen, and T.J.
     Woodruff. 2014. Fetal growth and maternal glomerular filtration rate: A systematic
     review. Journal of Maternal-Fetal and Neonatal Medicine 28:2176-2181.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-24

-------
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.
     InternationalJournal of Hygiene and Environmental 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. Reproductive
     Toxicology 27:239-245.

Wallace, K.B., G.E.  Kissling, R.L. Melnick, and C.R. Blystone. 2013. Structure-activity
     relationships for perfluoroalkane-induced in vitro interference with rat liver
     mitochondrial  respiration. Toxicology Letters 222:251'-264.

Wambaugh, J.F., R.W. Setzer, A.M. Pitruzzello, J. Liu, D.M. Reif, N.C. Kleinstreuer, N.
     Ching, Y. Wang, N. Sipes, M. Martin, K. Das, J.C. DeWitt, M.  Strynar, R. Judson, K.A.
     Houck, and C.  Lau. 2013. Dosimetric anchoring of in vivo and in vitro studies for
     perfluorooctanoate and perfluorooctanesulfonate. Toxicological Sciences 136:308-327.

Wan, H.T., Y.G. Zhao, X. Wei, K.Y. Hui, J.P. Giesy,  and C.K.C. Wong.  2012. PFOS-induced
     hepatic steatosis, the mechanistic actions on p-oxidation and lipid transport. Biochimica
     et Biophysica Acta 1820:1092-1101.

Wan, H.T., D.D. Mruk, K.C. Wong, and C.Y. Cheng.  2014a. Perfluorooctanesulfomate
     (PFOS) perturbs male rat sertoli cell blood-testes barrier function affecting F-actin
     organization via p-FAK-Tyr407: An in vitro study. Endocrinology 155(l):249-262.

Wan, H.T., Y.G. Zhao, P.Y. Leung, and C.K.C. Wong. 2014b.  Perinatal exposure to
     perfluorooctane sulfonate affects glucose metabolism in adult offspring. PLOS ONE
     9:e87137.

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. Environmental Science
     &  Technology  44:1847-1853.

Wang, Y., A.P. Starling, L.S. Haug, M. Eggesbo, G. Becher, C. Thomsen, G. Travlos, D. King,
     J.A. Hoppin, WJ. Rogan, and M.P. Longnecker. 2013. Association between
     perfluoroalkyl  substances and thyroid stimulating hormone among pregnant women: A
     cross-sectional study. Environmental Health 12:76.

Wang, L., Y. Wang, Y. Liang, J. Li, Y. Liu, J. Zhang, A. Zhang, J. Fu, and G. Jiang. 2014.
     PFOS induced lipid metabolism disturbances in  BALB/c  mice through inhibition of low
     density lipoproteins  excretion. Scientific Reports 4:4582.

Wang, S., Q. Lv, Y.  Yang, L.-H. Guo, B. Wan, and L. Zhao. 2014. Cellular target recognition
     of perfluoroalkyl acids: In vitro evaluation of inhibitory effects  on lysine decarboxylase.
     Science of the  Total Environment 496:381-388.

Wang, F., W. Liu, Y. Jin, F. Wang, and J. Ma. 2015. Prenatal and neonatal exposure to
     perfluorooctane sulfonic acid results in aberrant changes in miRNA expression  profile
     and levels in developing rat livers. Environmental Toxicology 30(6):712-723.
Perfluorooctane sulfonate (PFOS) - May 2016                                              5-25

-------
Wang, Y., W. Liu, Q. Zhang, H. Zhao, and X. Quan. 2015. Effects of developmental
     perfluorooctane sulfonate exposure on spatial learning and memory ability of rats and
     mechanism associated with synaptic plasticity. Food and Chemical Toxicology 76:70-76.

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. Environmental Health Perspectives
     117:660-667.

Watkins, D.J., J. Josson, B. Elston, S.M. Bartell, H.-M. Shin, V.M. Vieira, D.A.  Savitz, T.
     Fletcher, and G.A. Wellenius. 2013. Exposure to perfluoroalkyl acids and markers of
     kidney function among children and adolescents living near a chemical plant.
     Environmental Health Perspectives 121:625-630.

Webster, G.M., S.A. Venners,  A. Mattman, and J.W. Martin. 2014. Associations between
     perfluoroalkyl acids (PFASs) and maternal thyroid hormones in early pregnancy: A
     population-based cohort study. Environmental Research  133:338-347.

Webster, G.M., S.A. Rauch, N. Ste Marie, A. Mattman, B.P. Lanphear, and S.A. Venners.
     2015. Cross-sectional associations of serum perfluoroalkyl acids and thyroid hormones in
     U.S.  adults: Variation according to TPOAb and Iodine status (NHANES 2007-2008).
     Environmental Health Perspectives.

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. Toxicological Sciences 109:206-216.

Wen, L.-L., L.-Y. Lin, T.-C. Su, P.-C. Chen, and C.-Y. Lin. 2013. Association between serum
     perfluorinated chemicals and thyroid function in U.S. adults: The National Health and
     Nutrition Examination survey 2007-2010. The Journal of Clinical Endocrinology &
     Metabolism 98(9):E1456-E1464.

Whitney, E.N., Cataldo, C.B. and  S.R. Rolfes. 1987. Understanding Normal and Clinical
     Nutrition. West Publishing Company,  St. Paul, MN.

Whitworth, K.W., L.S. Haug, D.D. Baird, G. Becher, J.A. Hoppin, R. Skjaerven, C. Thomsen,
     M. Eggesbo, G. Travlos,  R. Wilson, and M.P. Longnecker. 2012. Perfluorinated
     compounds and subfecundity in pregnant women. Epidemiology 23(2):257.

WHO (World Health Organization).  2012. Guidance for Immunotoxicity Risk Assessmentfor
     Chemicals.  Harmonization Project Document No. 10. WHO Document Production
     Services, Geneva, Switzerland.

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
     7(3):340-350.

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. Toxicological Sciences 106:162-171.
Perfluorooctane sulfonate (PFOS) - May 2016                                            5-26

-------
Wolf, C.J., I.E. Schmid, C. Lau, and B.D Abbott. 2012. Activation of mouse and human
     peroxisome proliferator- activated receptor-alpha (PPARa) by perfluoroalkyl acids
     (PFAAs); further investigation of C4-C12 compounds. Reproductive Toxicology 33:546-
     551.

Wong, F., M. MacLeod, J.F. Mueller, and IT. Cousins. 2014. Enhanced elimination of
     perfluorooctane sulfonic acid by menstruating women: Evidence from population-based
     pharmacokinetic modeling. Environmental Science & Technology 48:8807-8814.

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. Biochimica et Biophysica Acta
     1768:1299-1808.

Xie, W., G.D. Bothun, and H.-J. Lehmler. 2010a. Partitioning of perfluorooctanoate into
     phosphatidylcholine bilayers is chain length-independent. Chemistry and Physics of
     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 Journal of Toxicological Sciences 33:219-
     226.

Yan, S., H. Zhang, F. Zheng, N. Sheng, X. Guo,  and J. Dai. 2015. Perfluorooctanoic acid
     exposure for 28 days affects glucose homeostasis and induces insulin hypersensitivity in
     mice. Scientific Reports 5:11029.

Yang, X., L.  Wang, W. Sun, and Z. Xue. 2009. Effects of perfluorooctane sulfonate on amino
     acid neurotransmitters and glutamine synthetase in rats. Wei Sheng Yan Jiu 38:19-21.

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. Toxicology Letters 209:270-276.

Yeung, L.W.Y., K.S. Guruge, S.  Taniyasu, N. Yamashita, P.W Angus, and C.B. Herath. 2013.
     Profiles of perfluoroalkyl substances in the liver and serum of patients with liver cancer
     and cirrhosis in Australia. Ecotoxicology and Environmental Safety 96:139-146.

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. Environmental
     Science & Technology 43:8416-8422.

Yu, W.-G., W. Liu, and Y.-H. Jin. 2009b. Effects of perfluorooctane sulfonate on rat thyroid
     hormone biosynthesis and metabolism. Environmental Toxicology and Chemistry
     28:990-996.

Yu, W.-G., W. Lu, L. Liu, and Y.-H. Jin. 2011. Perfluorooctane sulfonate increased hepatic
     expression of OAPT2 and MJAP2 in rats. Archives of Toxicology 85:613-621.

Perfluorooctane sulfonate (PFOS) - May 2016                                             5-27

-------
Yu, N., X. Wang, B. Zhang, J. Yang, J. Li, W. Shi, S. Wei, and Y. Yu. 2015. Distribution of
     perfluorooctane sulfonate isomers and predicted risk of thyroid hormonal perturbation in
     drinking water. Water Research 76:171-180.

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.
     NeuroToxicology 32:130-139.

Zeng, H.-X., Q.-Z. He, Y.-Y.  Li, C.-Q. Wu, Y.-M. Wu, and S.-q. Xu. 2014. Prenatal exposure
     to PFOS caused mitochondria-mediated apoptosis in heart of weaned rat. Environmental
     Toxicology 30:1082-1090.

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.

Zhang, T., H. Sun, Y. Lin, Y.  Qin, X. Geng, and L. Kannan. 2013. Distribution of poly- and
     perfluoroalkyl substances in matched samples from pregnant women and carbon chain
     length related maternal transfer. Environmental Science & Technology 47:7974-7981.

Zhang, L., X.-M. Ren, and L.-H. Guo. 2013. Structure-based investigation on the interaction of
     perfluorinated compounds with human liver fatty acid binding protein. Environmental
     Science & Technology 47:11293-11301.

Zhang, Y., S. Beesoon, L. Zhu, and J.W. Martin. 2013. Biomonitoring of perfluoroalkyl acids
     in human urine and estimates of biological half-life. Environmental Science  &
     Technology 47(18): 10619-10627.

Zhang, T., H. Sun, X. Qin, Z.  Gan, and K. Kannan. 2014. PFOS and PFOA in paired urine and
     blood from general adults  and pregnant women: Assessment of urinary elimination.
     Environmental Science and Pollution Research 22(7): 5 572-5 579.

Zhang, L., X.-M. Ren, B. Wan, and L.-H.  Guo. 2014. Structure-dependent binding and
     activation of perfluorinated compounds on human peroxisome proliferator-actovated
     receptor y. Toxicology and Applied Pharmacology 279:275-283.

Zhang, W., F. Wang, P. Xu, C. Miao, X. Zeng, X. Cui, C. Lu, H. Xie, H. Yin, F. Chen, and J.
     Ma. 2014. Perfluorooctanoic acid stimulates breast cancer cells invasion and up-regulates
     matrix metalloproteinase-2/-9 expression mediated by activating NF-KB. Toxicology
     Letters 229(1): 118-125.

Zhang, C., R. Sundaram, J. Maisog, A.M. Calafat, D. Boyd Barr, and G.M. Buck Louis. 2015.
     A prospective study of prepregnancy serum concentrations of perfluorochemicals and the
     risk of gestational diabetes. Fertility and Sterility 103:184-189.

Zhao, B., L. Li, J. Liu, H. Li,  C. Zhang, P. Han, Y. Zhang, X. Yuan, R.S. Ge, and  Y. Chu.
     2014. Exposure to perfluorooctane sulfonate in utero reduces testosterone production in
     rat fetal Leydig cells. PLOS ONE 9:e78888.

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.
     Archives of Toxicology 83:679-689.

Perfluorooctane sulfonate (PFOS) - May 2016                                            5-28

-------
Appendix A: Literature Search Strategy Developing the Search

The literature search strategy was planned with input from EPA library services staff. CAS
numbers served as the basis for identification of relevant search terms. Trial searches were
conducted and results were evaluated to refine the search strategy (e.g., to prevent retrieval of
citations unrelated to health and occurrence). The search string was refined to improve the
relevancy of the results. All searches were conducted in the PubMed database, which contains
peer-reviewed journal abstracts and articles in various biological, medical, public health, and
chemical topics. The first search string (as well as future iterations) is presented below.

Every two weeks, a search was run in PubMed and a bibliography of the search results was
compiled.

In 2012, the State of New Jersey Department of Environmental Protection (NJDEP) initiated a
monthly search in PubMed for emerging literature on perfluorinated chemicals primarily from
the carboxylic acid and sulfonate families. These searches were provided to the EPA on a
monthly basis. There was a high degree of overlap with the results from the EPA search, thus
increasing the confidence in the search strategy.

In 2013, the EPA search strategy was expanded to cover other members of the
perfluorocarboxylic acids (C-4 to C-12) and sulfonate families (C-4, C-6, C-8). The search string
was altered in June of 2013 to rely more on the search features offered by PubMed.

A change in the PubMed database structure in 2015 required some modification to the search
strategy. A search in August 2015 returned more than 4,000 records, a number that was
inconsistent with prior searches. The cause was PubMed's lack of recognition of the  search term,
"Heptadecafluorooctane-1-sulphonic acid" and interpreting the term as "ACID." The
resolution is highlighted in the search strings below.

All search iterations are noted below.

Search Strategy Examples: (Arranged from most recent to oldest).

2015

Search: perfluorooctanoate OR "perfluorooctanoic acid" OR "perfluoroctanoic acid" OR pfoa
OR "perfluorinated chemicals" OR "perfluorinated compounds" OR "perfluorinated  homologue
groups" OR "perfluorinated contaminants" OR "perfluorinated surfactants" OR perfluoroalkyl
acids OR "perfluorinated alkylated substances" OR "perfluoroalkylated substances" OR pfba OR
"perfluorobutanoic acid" OR perfluorochemicals OR "telomer alcohol" OR "telomer alcohols"
OR "fluorotelomer alcohols" OR "polyfluoroalkyl compounds" OR "perfluorooctane sulfonate"
OR pfos OR "perfluorooctanesulfonic acid" OR "perfluorooctane sulfonic acid" OR
"perfluorooctane sulphonate" OR perfluorooctane sulfonate OR "perfluorooctanyl sulfonate" OR
"Heptadecafluorooctane-1-sulphonic" OR"Heptadecafluoro-l-octanesulfonic acid" OR
perfluorononanoate OR pfhxa OR "perfluorohexanoic acid" OR "fluorinated surfactants"

Filters: English.
Frequency: Every 2 weeks

September 2013
Perfluorooctane sulfonate (PFOS) - May 2016                                              A-1

-------
Search: perfluorooctanoate OR "perfluorooctanoic acid" OR "perfluoroctanoic acid" OR pfoa
OR "perfluorinated chemicals" OR "perfluorinated compounds" OR "perfluorinated homologue
groups" OR "perfluorinated contaminants" OR "perfluorinated surfactants" OR perfluoroalkyl
acids OR "perfluorinated alkylated substances" OR "perfluoroalkylated substances" OR pfba OR
"perfluorobutanoic acid" OR perfluorochemicals OR "telomer alcohol" OR "telomer alcohols"
OR "fluorotelomer alcohols" OR "polyfluoroalkyl compounds" OR "perfluorooctane sulfonate"
OR pfos OR "perfluorooctanesulfonic acid" OR "perfluorooctane sulfonic acid" OR
"perfluorooctane sulphonate" OR perfluorooctane sulfonate OR "perfluorooctanyl sulfonate" OR
"Heptadecafluorooctane-1-sulphonic acid" OR"Heptadecafluoro-l-octanesulfonic acid" OR
perfluorononanoate OR pfhxa  OR "perfluorohexanoic acid" OR "fluorinated surfactants"

Filters: English.
Frequency: Every 2 weeks

June 2013

Search: (PFOA[tw] OR perfluorooctanoic acid[tw] OR335-67-l[tw] ORPFBA[tw] OR
perfluorobutanoate[tw] OR 3794-64-7[tw] ORPFDA[tw] OR perflurordecanoic acid[tw] OR
335-76-2[tw] OR PFHpA[tw]  OR perfluoroheptanoic acid[tw] OR 375-85-9[tw] OR PFHxA[tw]
OR perfluorohexanoic acid[tw] OR 307-24-4[tw] OR PFNA[tw] OR perfluorononanoic acid[tw]
OR 375-95-1 [tw] OR PFPtA[tw] OR perfluoropentanoic acid[tw] OR 2706-90-3[tw] OR
PFPA[tw] OR pentafluoropropionic acid[tw] OR 422-64-0[tw]) AND (human* [tw] OR
mammal*[tw]) NOT (environment* OR ecolog*)

Filters: English.
Frequency: Every 2 weeks

February 2013

Search: perfluorooctanoate OR "perfluorooctanoic acid" OR "perfluoroctanoic acid" OR pfoa
OR "perfluorinated chemicals" OR "perfluorinated compounds" OR "perfluorinated homologue
groups" OR "perfluorinated contaminants" OR "perfluorinated surfactants" OR perfluoroalkyl
acids OR "perfluorinated alkylated substances" OR "perfluoroalkylated substances" OR pfba OR
"perfluorobutanoic acid" OR perfluorochemicals OR "telomer alcohol" OR "telomer alcohols"
OR "fluorotelomer alcohols" OR "polyfluoroalkyl compounds" OR "perfluorooctane sulfonate"
OR pfos OR "perfluorooctanesulfonic acid" OR "perfluorooctane sulfonic acid" OR
"perfluorooctane sulphonate" OR perfluorooctane sulfonate OR "perfluorooctanyl sulfonate" OR
"Heptadecafluorooctane-1-sulphonic acid" OR"Heptadecafluoro-l-octanesulfonic acid" OR
perfluorononanoate OR pfhxa  OR "perfluorohexanoic acid" OR "fluorinated surfactants"

Filters: English.
Frequency: Every 2 weeks

June 2011

Search (perfluorooctanoate OR "perfluorooctanoic acid" OR "perfluoroctanoic acid" OR pfoa
OR "perfluorinated chemicals" OR "perfluorinated compounds" OR "perfluorinated homologue
groups" OR "perfluorinated contaminants" OR "perfluorinated surfactants" OR
perfluoroalkylacids OR "perfluorinated alkylated  substances"  OR "perfluoroalkylated
substances" OR pfba OR "perfluorobutanoic acid" OR perfluorochemicals OR "telomer alcohol"
OR "telomer alcohols" OR "fluorotelomer alcohols" OR "polyfluoroalkyl compounds" OR

Perfluorooctane sulfonate (PFOS) - May 2016                                             A-2

-------
"perfluorooctane sulfonate" OR pfos OR "perfluorooctanesulfonic acid" OR "perfluorooctane
sulfonic acid" OR "perfluorooctane sulphonate" OR perfluorooctanesulfonate OR
"perfluorooctanyl sulfonate" OR"Heptadecafluorooctane-l-sulphonic acid" OR
"Heptadecafluoro-1-octanesulfonic acid" OR perfluorononanoate OR pfhxa OR
"perfluorohexanoic acid" OR "fluorinated surfactants" OR 335-67-1 [rn])

Limits: Publication Date [Dates will change for each search], English Language only.

June 2009

Search (perfluorooctanoate OR "perfluorooctanoic acid" OR "perfluoroctanoic  acid" OR pfoa
OR "perfluorinated chemicals" OR "perfluorinated compounds" OR "perfluorinated homologue
groups" OR "perfluorinated contaminants" OR "perfluorinated surfactants" OR
perfluoroalkylacids OR "perfluorinated alkylated substances" OR "perfluoroalkylated
substances" OR pfba OR "perfluorobutanoic acid" OR perfluorochemicals OR "telomer alcohol"
OR "telomer alcohols" OR "fluorotelomer alcohols" OR "polyfluoroalkyl compounds" OR
"perfluorooctane sulfonate" OR pfos OR "perfluorooctanesulfonic acid" OR "perfluorooctane
sulfonic acid" OR "perfluorooctane sulphonate" OR perfluorooctanesulfonate OR
perfluorononanoate OR pfhxa OR "perfluorohexanoic acid"  OR "fluorinated surfactants" OR
335-67-1 [rn] OR 1763-23-1 [rn])

Limits: Entrez Date from 2009/04/07 to 2009/04/12.

New Jersey Search Terms

Search: perfluorinated OR perfluorooctanoate OR perfluorononanoate OR
perfluorooctanesulfonate OR perfluorooctanesulphonate OR perfluoroalkylated OR
perfluoroalkyl OR polyfluoroalkyl OR polyfluorinated OR PFBA OR PFBS OR PFDA OR
PFHA OR PFFffA OR PFHXA OR PFHXS OR PFNA OR PFOA OR PFOAs OR PFOS OR
PFUNDA OR "perfluorooctanoic acid" OR "perfluoro octanoic acid" OR "perfluorooctane
sulfonate" OR "perfluorooctane sulfonic acid" OR "perfluorooctanesulfonic acid" OR
"perfluorooctane sulphonate" OR "perfluorooctanyl sulfonate" OR "perfluorobutanoic acid" OR
"perfluoroalkyl acids" OR "perfluorononanoic acid" OR "perfluorohexanoic acid"  OR
"perfluorohexane sulfonate" OR "perfluorohexane sulphonate" OR perfluorobutanoate OR
"perfluoro butanoate" OR perfluorohexanoate OR "perfluoro hexanoate"

Filters: 1
Perfluorooctane sulfonate (PFOS) - May 2016                                              A-3

-------
Appendix B: Studies Evaluated Since August 2014

   The tables that follow identify the papers that were retrieved and reviewed for inclusion
following the August 2014 peer review for the draft PFOS Health Effects Support Document.
The papers listed include those recommended by the peer reviewers or public commenters, as
well as those identified from the literature searches between the completion of the peer review
draft and December 2015. Papers included in the final Health Effects Support Document (HESD)
are noted and reasons provided for those that were not included in the final document.

   The tables for document retrieval and review are followed by updated versions of the
summaries of the epidemiology summary tables from the peer reviewed draft as recommended
by the peer reviewers. They are a useful tool to facilitate a high level comparison of the study
outcomes for each of the epidemiological study groupings.

   The criteria utilized in determining the  papers that were included in the HESD after the peer
review and presented in the Background were the following:
   1.  The study examines a toxicity endpoint or population that had not been examined by
       studies already present in the draft assessment.
   2.  Aspects of the study design, such as the size of the population exposed or quantification
       approach, make it superior to key studies already included in the draft document.
   3.  The data contribute substantially  to the weight of evidence for any of the toxicity
       endpoints covered by the draft document.
   4.  There are elements of the study design that merit its inclusion in the draft assessment
       based on its contribution to the mode of action or the  quantification approach.
   5.  The study elucidates the mode of action for any toxicity endpoint or toxicokinetic
       property associated with PFOS exposure.
   6.  The effects observed differ from those in other studies with comparable protocols.
        Table B-l. PFOS Epi Papers—Post Peer Review (Retrieved and Reviewed)
Authors and year
Andersen et al. 2013
Back etal. 2015
Barrett etal. 2015
Berg etal. 2015
Bonefeld-Jorgenson et al. 2014
Bonefeld-Jergenson et al. 201 1
Brieger etal. 2011
Buck Louis et al. 2015
Chang etal. 2014
Chen etal. 2015
Dankers etal. 2013
Darrow etal. 2013
Darrow etal. 2014
Topic — key words
Postnatal growth
Time to pregnancy
Ovarian hormone
Thyroid
Breast cancer
Breast cancer
Immune effects
Semen quality
Analysis of human cancer
studies
Birth weight
Blood-testis barrier
Reproductive outcome
Miscarriage
Status/Notes
Added PFOA/PFOS
Added PFOA
Not Added — No association observed for
PFOA; PFOS was not included in the
assessment
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Already presented in PFOS/PFOA
Added PFOA/PFOS
Added PFOA in the cancer weight of
evidence section
Added PFOS
Reviewed, — not added; Study of an assay
that used PFOA as one chemical in the test
battery
Added PFOA/PFOS
Added PFOA/PFOS
Perfluorooctane sulfonate (PFOS) - May 2016
B-l

-------
Authors and year
Donaueretal. 2015
Eriksenetal. 2013
Fitz-Simonetal. 2013
Fisher etal. 2013
Fletcher etal. 2013
Fu etal. 2014
Geigeretal. 2014a
Geigeretal. 2014b
Ghisari etal. 2014
Governini et al. 2015
Grandjean and Clapp 2015
Granumetal. 2013
Hardell etal. 2014
Hey er etal. 201 5a
Hey er etal. 201 5b
Humblet etal. 2014
Jain 20 14
Innesetal. 2014
Joensenetal. 2013
Kerger etal. 2011
Kjeldsen and Bonefeld-Jergensen
2013
Kristensen et al. 2013
Liew etal. 2014
Looker etal. 2014
Lopez-Doval et al. 2014
Maisonet et al. 2015
Maisonet etal. 2012
Merck etal. 2015
Okada etal. 2014
Osuna etal. 2014
RothandWilks2014
Shrestha etal. 2015
Starling etal. 2014
Steenland etal. 2015
Topic — key words
Infant Neurobehavior
Total cholesterol — Danish
Serum lipids
Plasma lipids
Cholesterol-genes
Serum lipids in Chinese
subjects
Lipids/children
Hypertension/children
Breast cancer — Inuit
DNA effects in sperm
Health Risks
Immune children
Prostate cancer
Human weight
Behavior motor development
Asthma
NHANES
Colorectal cancer
Sperm
Cholesterol C8
Sex hormones
Prenatal female repro
Cerebral palsy children
Immune
Male repro
Gestational diabetes
Birth weight
PFAS levels in children
Allergy children
Antibodies PFOS PFOA
Neurodevelopmental
Thyroid
Plasma lipids
Workers
Status/Notes
Not added — negative for PFOS; No
statistical differences in PFOA levels
during pregnancy and any neuro endpoint.
Better studies.
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Not added: Chinese population, dataset
available on U.S. population. More
branched chain isomers found among the
people in China.
Added PFOA/PFOS
Added PFOA/PFOS
Not added; same population as Bonefeld-
Jergensen et al. 2014; this study focuses on
gene polymorphisms
Added PFOA/PFOS
Not added; the primary studies are already
included in the documents.
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added; demographics for cholesterol and
PFOS in summary section of epi studies
Covered multiple PFAS in vitro no impact
on weight of evidence
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Not added; No significant impact
Added PFOS
Not added; focus more on methylHg and
PCB than PFAS; only n = 38 as
preliminary study
. Not added; no significant impact
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA
Perfluorooctane sulfonate (PFOS) - May 2016
B-2

-------
Authors and year
Stein et al. 2009
Taylor etal. 2014
VandenHeuvel2013
Vassiliadou etal. 2010
Velez etal. 2015
Verner etal. 2015
Verner and Longnecker 2015
Vested etal. 2013
Vesterinen et al. 2014
Wang etal. 2013
Watkins etal. 2013
Webster etal. 2014
Webster etal. 2015
Wen etal. 2013
Yeung etal. 2013
Zhang etal. 2015
Topic — key words
Pregnancy
Menopause
Serum lipids
PFOS in cancer vs non-cancer
patients
Fertility
Fetal growth GFR
Menstruation/excretion
Semen quality and hormones
Fetal Growth GFR
Thyroid
Kidney function
Maternal thyroid
Thyroid — iodine statue
Thyroid
Liver cancer
Gestational diabetes
Status/Notes
Added PFOA
Added PFOA/PFOS
Not added; is a rebuttal of Fletcher et al.
2013 conclusions. No significant impact
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOS
Added PFOS/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
               Table B-2. PFOA Post Peer Review Animal Toxicity Studies
Authors and year
Bjork etal. 2011
Corsinietal. 2014
Corsinietal. 2012
Dewitt etal. 2015
Fenton2015
Filgo etal. 2015
Hall etal. 2012
Koustas etal. 2014
Liu etal. 2015
Long etal. 2013
Lu etal. 2015
Ngo etal. 2014
Post etal. 2012
Quist etal. 2015
Rigden etal. 2015
Shabalina etal. 2015
Sheng etal. 2016
Topic
Nuclear receptor activation
Immune data review
Immune in vitro data review
Immunotoxicity
Repro editorial
Liver tumors in females
developmentally exposed
PPARa and cancer
Fetal growth (animal studies)
navigation guide
Testes
Neurotoxicity adult PFOS
Testes
Tumors mice Min/+ PFOS
Review paper
Liver histopathology/high fat
diet post weaning exposure
Acute liver effects
Brown fat uncoupling protein 1
Binding to liver fatty acid
binding protein
Action notes
In vitro, mechanistic findings comparable to
studies already included
Not added; no significant impact
Not added; no significant impact
Added PFOA
Not added
Added PFOA
Cited in synthesis. Paper on adversity of liver
hypertrophy PFOA/ PFOS
Added PFOA
Added PFOA
Added PFOS
Added PFOA
Added PFOS
Not added. Key studies included in the document;
no significant impact
Added PFOA
Added PFOA
Not added. Mechanistic; no significant impact
Not added; no significant impact, topic covered by
other papers
Perfluorooctane sulfonate (PFOS) - May 2016
B-3

-------
Authors and year
Tan etal. 2012
Tan etal. 2013
Tucker etal. 2015
Wallace etal. 2013
Wan etal. 20 14b
Wan etal. 2012
Wan etal. 20 14a
F.Wang etal. 2015
S.Wang etal. 2014
L.Wang etal. 2014
Y.Wang etal. 2015
Yan etal. 2015
Yu etal. 2015
Zeng etal. 2014
L.Zhang etal. 2013
Y.Zhang etal. 2013
W.Zhang etal. 2014
Zhao etal. 2014
Topic
Gene activation
Gene activation dietary fat
Mammary gland
Mitochodrial respiration
Glucose metabolism
Hepatic steatosis
Sertoli cells
MiRNA liver PFOS early life
Lysine decarboxylase
Inhibition of LDL
Special learning and memory
Glucose homeostasis
Thyroid PFOS isomers
Mitochondria! mediated
apoptosis of the heart
Fatty acid binding protein
Biological half-life
Breast cancer cell invasion —
mechanistic
Testosterone reduction in
Leydig cells PFOS
Action notes
Added PFOA/PFOS
Added PFOA
Added PFOA
Not added. No significant impact, topic covered
by other papers
Added PFOS
Added PFOS
Added PFOS
Not added; no significant impact
Added PFOA/PFOS
Added PFOS
Added PFOS
Not added. Dose-response in Wan (2014b)
presented (more robust). Single dose for whole
animal
Added PFOS
Added PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Not added; in vitro, no significant impact
Added PFOS
Perfluorooctane sulfonate (PFOS) - May 2016
B-4

-------
                       Table B-3. Toxicokinetics: Post Peer Review
Authors and year
D'Alessandro et al. 2013
Augustine et al. 2005
Beesoonetal. 2011
Beesoon and Martin 2015
Cuietal. 2010
Fabrega etal. 2014
Kerstner-Wood et al. 2003
Klaassen and Aleksunes
2010
Loccisano et al. 2013
Mondal et al. 2014
Ospinal-Jimenez and Pozzo
2012
Perez etal. 2013
Ren etal. 2015
Rigden etal. 2015
Shabalina etal. 2015
Slitt et al. 2007
Tucker etal. 2015
Verner and Longnecker
2015
Wambaugh etal. 2013
Wong etal. 2014
T.Zhang etal. 2014
L.Zhang etal. 2014
Y.Zhang etal. 2013
T.Zhang etal. 2013
Topic
Serum albumin
Transporter expression testes
Isomer profile
Albumin binding
Excretion subchronic
PK model
Plasma protein binding
Transporter paper — Provided
diagram of kidney transporters
PK model — Human
Breast milk
Protein denaturation
Human tissue levels
Thyroid hormone receptor
binding (in vitro)
Liver and excretion
Brown fat
Transporter expression PFOA
Menstruation-excretory route
Excretion PFOS
PK model
Menstrual blood as excretory
route
Excretion general population
and pregnancy
PPAR gamma
Excretion, half-life
Maternal transfer
Action Notes
Added PFOS
Not added background paper on testes transporters
-no relationship to PFOA PFOA or any PFAS
Added PFOA
Added PFOA
Added PFOA/PFOS
Added PFOA/PFOS
Added— PFOA/PFOS
Added PFOA
Added PFOA/PFOS
Added PFOS/PFOA
Added PFOS
New PFOA/PFOS
Added PFOA/PFOS
Added PFOA
Not added; No information on MOA for body
weight effects in the animal or human studies
Not added. Reported on transporters during
extrahepatic cholestasis. No data on PFOA and
PFOS. No significant impact.
Added PFOA
Added PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Added PFOS
Added PFOA/PFOS
Added PFOA/PFOS
Perfluorooctane sulfonate (PFOS) - May 2016
B-5

-------
                         Table B-4. Association of Serum PFOS with Serum Lipids and Uric Acid
Reference
Study type
n
Mean or median
serum PFOS (jig/mL)
TC
VLDL
LDL
HDL
Non-HDL
TG
UA
Occupational Populations
Olsen et al.
200 la, 2003b
Olsen et al.
200 Ib, 2003b
Cross-sectional
Longitudinal;
~5 years
263 (Decatur)
255 (Antwerp)
175
(Decatur and
Antwerp
combined for
analysis)
1.40
0.96
2.62
(baseline)
1.67 (follow-up)
(Decatur)
1.87 (baseline)
1.16
(follow-up)
(Antwerp)
t
4-^>
NM
NM
NM
NM
•
4-^>
NM
NM
t
4-^>
NM
NM
General Populations
Steenland et
al. 2009
Steenland et
al. 2010
Frisbee et al.
2010
Fitz-Simon et
al. 2013
Nelson et al.
2010
Lin et al.
2009
Maisonet et
al. 2015
Timmermann
etal. 2014
Chateau-
Begat et al.
2010
Eriksen et al.
2013
Cross-sectional
(C8)
Cross-sectional
(C8)
Cross-sectional
(C8, children)
Longitudinal;
4.4 years (C8)
Cross-sectional
(NHANES)
Cross-sectional
(NHANES)
Longitudinal;
prenatal and
aged 7 and 15
years
Cross-sectional
(children 8-10
years)
Cross-sectional
Cross-sectional
46,294
54,951
12,476
521
860
3,685
111 (age 7 years)
88 (age 15 years)
499
723
663
0.022
0.023
0.023
0.023 (baseline)
0.0 11 (follow-up)
0.025
0.003 1(12-< 20 yrs)
0.0032 (> 20 yrs)
0.022
0.0412
0.019
0.036
t
NM
t
<— >
t
NM
4-^>
4-^>
•
t
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
t
NM
t
•
4-^>
NM
<— >
<—>•
<— >
NM
4-^>
NM
t
•
4-^>
t
4-^>
4-^>
t
NM
NM
NM
NM
NM
t
NM
NM
NM
NM
NM
t
NM
<— >
<— >
NM
<— >
<—>•
<-> (normal wt)
t (overweight)
<— >
NM
NM
t
NM
NM
NM
NM
NM
NM
NM
NM
Perfluorooctane sulfonate (PFOS) - May 2016
B-6

-------
Reference
Starling et al.
2014
Fisher et al.
2013
Study type
Cross-sectional
(maternal at
14-26 weeks
gestation)
Cross-sectional
n
891
2,700
Mean or median
serum PFOS (fig/mL)
0.013
0.0084
TC
t
4-^>
VLDL
NM
NM
LDL
t
4-^>
HDL
t
4-^>
Non-HDL
NM
NM
TG
<— >
NM
UA
NM
NM
Notes: | = 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 measured
     Perfluorooctane sulfonate (PFOS) - May 2016
B-7

-------
                 Table B-5. Association of Serum PFOS with Reproductive and Developmental Outcomes
Study
Study type
n
Mean or median
serum PFOS
(jig/mL)
Outcome
Measures at
birth
Growth/
Development
Fecundity/
Fertility
Occupational Populations
Grice et al. 2007
Survey
263
Not measured;
exposure
categorized by job
<-> (any adverse)
NM
NM
NM
General Populations — Measures at Birth
Fei et al. 2007,
2008a, 2010a
Monroy et al. 2008
Washino et al.
2009
Hamm et al. 2009
Stein et al. 2009
Darrow et al.
2013,2014
Apelberg et al.
2007
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
(C8)
Cross-sectional
(C8)
Cross-sectional
1,400
101
428
252
5,262
1330
293
0.033-0.039 (first
trimester)
0.018 (maternal at
24-28 weeks)
0.016
(maternal at
delivery)
0.0072 (umbilical
cord blood)
0.0056 (maternal)
0.009 (maternal)
0.014
0.016-0.017
0.005 (cord blood)
<-> (gestation
length)
I (length of
breastfeeding)
<-> (gestation
length)
NM
<-> (gestation
length)
<-> (miscarriage)
<-> (preterm)
t (miscarriage)
<-> (gestational
age)
<-> (weight)
<-> (size)
<-> (Apgar score)
<->• (weight)
t (low weight
females only)
<-> (weight)
t (low weight)
<-> (low weight)
t (birth weight
decreased)
I (weight, head
circumference,
ponderal index)
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
t (hypertension)
NM
General Populations — Measures of Postnatal Growth
Fei et al. 2008b
Liewetal. 2014
Cross-sectional
Cross-sectional
1,400
156 cases
550 controls
0.033-0.039 (first
trimester)
0.026-0.029 (first
trimester)
NM
NM
NM
NM
<-> (at 6 months)
t (at 18 months;
sitting up later)
t (cerebral palsy
in boys)
MN
NM
Perfluorooctane sulfonate (PFOS) - May 2016
B-8

-------
Study
Andersen et al.
2010
Andersen et al.
2013
Fei and Olsen
2011
Hey er etal. 201 5b
Hoffman et al.
2010
Hey er etal. 201 5a
Lopez-Espinosa et
al. 2011
Kristensen et al.
20 13; Vested etal.
2013
Christensen et al.
2011
Halldorsson et al.
2012
Study type
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
(NHANES)
Cross-sectional
Cross-sectional
(C8)
Cross-sectional
Cross-sectional
Cross-sectional
n
1,010
811 (children at
age 7 years)
787 (behavior)
537
(coordination)
1,106
571 (children)
1,022 (children)
3, 076 boys
2,931 girls
343 women
169 men
(~ 20 years)
448 girls
665
Mean or median
serum PFOS
(jig/mL)
0.0334 (first
trimester)
0.033-0.039 (first
trimester)
0.036 (first
trimester)
0.01 (maternal)
0.023
0.005-0.0202
(maternal)
0.0098-0.036
0.0211-0.0212
(maternal)
0.019-0.02
(maternal)
0.0285 (maternal)
Outcome
NM
NM
NM
NM
NM
NM
NM
NM
NM
NM
Measures at
birth
I (birth weight in
girls)
NM
NM
NM
NM
NM
NM
NM
NM
NM
Growth/
Development
I (weight and
BMI at 12
months in boys)
<-> (height,
weight, waist
measurement,
risk of
overweight)
<-> (behavior and
coordination at 7
years)
<-> (motor skills,
hyperactivity)
t (ADHD)
<-> (overweight)
t (waist-to-height
ratio)
t (delayed
puberty)
NM
<-> (age at
menarche)
t (overweight in
females at 20
years)
Fecundity/
Fertility
NM
NM
NM
NM
NM
NM
NM
<-> (measures of
reproductive
function)
NM
NM
General Populations — Male and Female Fertility
Zhang etal. 2015
Velez etal. 2015
Fei et al. 2009
Cross-sectional
Cross-sectional
Cross-sectional
258
1,743
1,400
0.012-0.0131
(preconception)
0.005
0.033-0.039 (first
trimester)
t (gestational
diabetes)
NM
NM
NM
NM
NM
NM
NM
NM
NM
<-> (time to
pregnancy)
<-> (infertility)
t (time to
pregnancy)
t (infertility)
Perfluorooctane sulfonate (PFOS) - May 2016
B-9

-------
Study
Knoxetal. 2011
Joensen et al. 2009
(PFOA/PFOS
combined)
Joensen etal. 2013
Buck Louis et al.
2014
Study type
Cross-sectional
(C8)
Cross-sectional
Cross-sectional
Cross-sectional
n
25,957
105
247
462
Mean or median
serum PFOS
(jig/mL)
0.018
0.025
0.0085
0.017-0.021
Outcome
NM
NM
NM
NM
Measures at
birth
NM
NM
NM
NM
Growth/
Development
NM
NM
NM
NM
Fecundity/
Fertility
t (early
menopause)
t (lower number
normal sperm)
<-> (testosterone)
<-> (semen
parameters)
I (testosterone)
t (lower % sperm
with coiled tail)
(total of six PFAS
associated with
changes in sperm
quality)
Notes: | = positive association; j= negative association; <-> = no association; NM = Not Measured
Perfluorooctane sulfonate (PFOS) - May 2016
B-10

-------
          Table B-6. Association of PFOS Level with the Prevalence of Thyroid Disease and Thyroid Hormone Levels
Study
Olsenetal. 200 la
Dallaire et al. 2009
Bloom etal. 2010
Melzeretal. 2010
Shrestha etal. 2015
Pirali et al. 2009
Wang etal. 2013
Berg etal. 2015
Inoue et al. 2004
Chan etal. 2011
Webster etal. 2014
Study type
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Cross-sectional
Population (n)
Adult workers
(263 Decatur)
(255 Antwerp)
Adults (623)
Adults (31)
Adult (NHANES;
3,966)
Adults (51 men, 36
women)
Adults (28)
Women at gestation
week 18 (Norwegian
Mother/Child
Cohort; 903)
Women at gestation
week 18, day 3 and
week 6 after delivery
(Norwegian
Mother/Child
Cohort; 375)
Newborns (15)
Women at gestation
week 15-20 (974)
152 women at
gestation week 15-
18
Mean serum PFOS
(Mg/niL)
1.4
0.96
0.018
0.0196
0.025 (men)
0.019 (women)
0.036
5.3 ng/g thyroid tissue
0.0128
0.00803
0.0016-0.0053 (cord
blood)
0.0074
0.0017
Thyroid
Disease
NM
<— >
<—>•
<-> (women)
t (men)
<— >
<— >
NM
NM
NM
4-^>
NM
TSH
<— >
I
4-^>
NM
4-^>
NM
t
t
4-^>
4-^>
4-^>
T3
<— >
4
NM
NM
4-^>
NM
NM
<— >
NM
NM
4-^>
T4
<— >
t
4-^>
NM
t
NM
NM
<— >
<—>•
<—>•
<—>•
Notes: f= positive association; |= negative association; ^= no association; NM = Not Monitored
Perfluorooctane sulfonate (PFOS) - May 2016
B-ll

-------
          Table B-7. Association of Serum PFOS with Markers of Immunotoxicity
Study
Okadaetal. 2012
Feietal. 2010b
Grandjean et al.
2012
Grandjean et al.
2012
Granum et al.
2013
Humblet et al.
2014
Dong etal. 2013
Looker etal. 2014
Study type
Prospective
cohort
Cross-sectional
Prospective
cohort
Prospective
cohort
Prospective
cohort
Cross-sectional
Cross-sectional
Cross-sectional
Population (n)
Maternal, third
trimester (343)
Maternal, first
trimester (1,400)
Maternal at
gestation week
32 (587)
Children age 5
years (587)
Women at
delivery (56)
Children at 12-
19 years (1,877)
Children age
10-15 years
(23 1 asthmatics
and 225
controls)
Adults (411)
Mean or median
serum PFOS
(jig/mL)
0.0056
0.0353
0.0273 (maternal)
0.0 167 (child at
age 5 years)
0.0056 (maternal)
0.017
(asthmatics)
0.0168 (non-
asthmatics)
0.0455
(asthmatics)
0.0334 (non-
asthmatics)
0.0083
Disease
prevalence in
children
<— >
<— >
NM
NM
<— >
<->• for asthma
t for asthma
NM
Vaccine
response
NM
NM
1 (antibody liter
in child at age 5
yrs)
1 (antibody liter
in child at age 7
yrs)
J, (antibody tiler
in child al age 3
years)
NM
NM
4-^>
Notes: f= positive association; |= negative association; <-»= no association; NM = Not Measured
Perfluorooctane sulfonate (PFOS) - May 2016
B-12

-------
Appendix C: Summary of Data
Perfluorooctane sulfonate (PFOS) - May 2016                                            C-1

-------
                                      Table C-l. PFOS Toxicokinetic Information
Species
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Dose
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Route of
exposure
Unknown
Unknown
Drinking water
Drinking water
Drinking water
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown/
drinking water
Unknown/
drinking water
Unknown
Unknown
Effects observed
t TC; t TG
None observed
on cholesterol
t TC; t TG; t
LDL; t UA
t TC; t LDL; |
HDL
None observed
on cholesterol
t TC; t non-
HDL
|HDL
TTC
t TC; t LDL; |
HDL
None observed
on cholesterol
Developmental
delays
LEW
None on birth
outcome; birth
weight and
length; growth to
7 years
t time to
pregnancy
PFOS in liver
(Mg/g)
M
F
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
PFOS in blood
(jig/mL)
M
F
0.96-1.40
1.16-2.62
0.022-0.023
0.023
0.011-0.023
0.025
0.019
0.036
0.013
0.0084
NS
NS
NS
NS
0.0098-0.039
0.0056-0.016
0.009-0.039
0.033-0.039
Reference
Olsen et al.
200 Ib, 2003b
Olsen et al.
200 Ib, 2003b
Steenland et al.
2009,2010
Frisbee et al.
2010
Fitz-Simon et al.
2013
Nelson et al.
2010
Chateau-Degat et
al. 2010
Eriksen et al.
2013
Starling et al.
2014
Fisher etal. 2013
Fei et al. 2008b;
Lopez-Espinosa
etal. 2011
Washino et al.
2009; Stein et al.
2009; Darrow et
al. 2013
Fei et al. 2007,
2008a; Monroy
et al. 2008;
Hamm et al.
2009; Andersen
etal. 2013
Fei et al. 2009
Perfluorooctane sulfonate (PFOS) - May 2016
C-2

-------
Species
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Human
Dose
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
Route of
exposure
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Effects observed
Effects on sperm
numbers and
morphology
None on semen
parameters
None on thyroid
hormones
|TSH, T3; |T4
None on thyroid
hormones
t incidence of
thyroid disease
(men only)
|T4
tTSH
None on thyroid
hormones
None on diseases
in children
None on diseases
in children
I antibody titer
in children
1 antibody titer
in children
t asthma
PFOS in liver
(Mg/g)
M
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
F
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
PFOS in blood
(jig/mL)
M
0.017-0.025
0.0085
F
NS
NS
0.96-1.4
0.018
0.0196
0.025
0.019
0.036
NS
NS
NS
NS
NS
0.008-0.0128
(gestation wk 18)
0.0074
(gestation wk
15-20)
0.0056
(maternal, third
trimester)
0.0353
(maternal, first
trimester)
0.0273
(maternal,
gestation wk 32)
0.0167 (child age 5 years)
NS
0.0056
(maternal at
delivery)
0.0455 (asthmatic children)
Reference
Joensen et al.
2009; Buck
Louis etal. 2014
Joensen et al.
2013
Olsen et al.
2001b;2003b
Dallaire et al.
2009
Bloom et al.
2010
Melzer et al.
2010
Shrestha et al.
2015
Wang et al.
20 13; Berg etal.
2015
Chan etal. 2011
Okada etal. 2012
Feietal. 2010b
Grandjean et al.
2012
Granum et al.
2013
Dong etal. 2013
Perfluorooctane sulfonate (PFOS) - May 2016
C-3

-------
Species
Human
Dose
NA
Route of
exposure
Unknown
Effects observed
None on vaccine
response
PFOS in liver
(Mg/g)
M
NS
F
NS
PFOS in blood
(jig/mL)
M
F
0.0083
Reference
Looker et al.
2014

Monkey
Monkey
0.15 mg/kg/day
for 26 weeks
with 52 week
recovery
0.75 mg/kg/day
for 26 weeks
with 52 week
recovery
capsule
capsule
None observed
t liver wt
I cholesterol and
body wt
NS
NS
NS
NS
(serum)
wk 1:4.60
wk 26: 82.6
wk 35: 84.5
wk79: 19.1
(serum)
wk 1:21.0
wk26: 173
wk35: 181
wk 79: 41.1
(serum)
wk 1:3.71
wk 26: 66.8
wk 35: 74.7
wk 79: 21.4
(serum)
wk 1:20.4
wk26: 171
wk35: 171
wk 79: 41.4
Seacat et al. 2002
Seacat et al. 2002

Rat
Rat
Rat
Rat
(male only)
Rat
(male only)
0.018-0.023
mg/kg/day for
104 weeks
0.184-0.247
mg/kg/day for
104 weeks
0.765-1.10
mg/kg/day for
104 weeks
5 mg/kg for 28
days
20 mg/kg for 28
days
diet
diet
diet
oral gavage
oral gavage
None observed
t liver histopath.
lesions
t body and liver
wt
t hepatocellular
adenoma
1 body wt
10/10 died (day
26)
hepatic
hypertrophy
wkO: 11.0
wk 10: 23.8
wk 105: 7.83
wk 0:47.6
wk 10: 358
wk 105: 70.5
wk 0:282
wk 10: 568
wk 105: 189
345
648
wk 0:8.71
wk 10: 19.2
wk 105: 12.9
wk 0:83.0
wk 10: 370
wk!05: 131
wk 0:373
wk 10: 635
wk 105: 381
NS
NS
(serum)
wk 0:0.91
wk 14: 4.04
wk!05: 1.31
(serum)
wk 0:7.57
wk 14: 43.9
wk 105: 22.5
(serum)
wk 0:41.8
wk 14: 148
wk 105: 69.3
(whole blood)
72.0
(whole blood)
NS
(serum)
wkO: 1.61
wk 14: 6.96
wk 105: 4.35
(serum)
wkO: 12.6
wk 14: 64.4
wk 105: 75.0
(serum)
wk 0:54.0
wk 14: 223
wk 105: 233
NS
NS
Thomford 2002
Thomford 2002
Thomford 2002
Cui et al. 2009
Cui et al. 2009
Perfluorooctane sulfonate (PFOS) - May 2016
C-4

-------


Species
Rat





Rat





Rat





Rat













Dose
0.4 mg/kg
42 days prior to
cohabitation
through GD 21


1.6 mg/kg
42 days prior to
cohabitation
through GD 21


3.2 mg/kg
42 days prior to
cohabitation
through GD 21


0.1 mg/kg GDO
to PND 20











Route of
exposure
oral
gavage




oral
gavage




oral
gavage




oral gavage













Effects observed
I maternal and
pup body wt




J, maternal and
pup body wt
J, pup survival



100% pup
mortality by
PND 2



None observed in
dams or
offspring









PFOS in liver
(Mg/g)
M
NS





NS





NS








PND 21:
Offspring =
5.98
PND 72:
Offspring =
0.98



F
GD21:
dams =107
fetuses = 30.6



GD21:
dams = 388
fetuses = 86.5



GD21:
dams = 610
fetuses = 230



GD20:
Dams = 8.35
Offspring =
3.21
PND 21:
Dams = NS
Offspring =
5.28
PND 72:
Dams = NS
Offspring =
0.80
PFOS in blood
(jig/mL)
M
NS





NS





NS








(serum)
PND 21:
Offspring =
1.73
PND 72:
Offspring =
0.04


F
(serum)
GD 1:40.7
GD 7: 40.9
GD21:
dams = 26.2
fetuses = 34.3
(serum)
GD 1: 160
GD 7: 154
GD21:
dams = 136
fetuses = 101
(serum)
GD 1:318
GD 7: 306
GD21:
dams = 155
fetuses =164
(serum)
GD20:
Dams= 1.72
Off spring = 3. 91
PND 21:
Dams = 3.16
Offspring = 1.77
PND 72:
Dams = NS
Offspring = 0.21




Reference
Luebker et al.
2005b




Luebker et al.
2005b




Luebker et al.
2005b




Butenhoff et al.
2009; Chang et
al. 2009









Perfluorooctane sulfonate (PFOS) - May 2016
C-5

-------


Species
Rat











Rat


Rat




Dose
l.Omg/kgGDO
to PND 20










1.0 mg/kg/day
CDs 2-20

2.0 mg/kg/day
CDs 2-20


Route of
exposure
oral gavage











oral gavage


oral gavage




Effects observed
t motor activity
and I habituation
in male offspring









none


I dam and pup
wt; |pup survival

PFOS in liver
(Mg/g)
M



PND 21:
Offspring =
44.89
PND 72:
Offspring =
7.17



NS


NS


F
GD20:
Dams = 48.88
Offspring =
20.03
PND 21:
Dams = NS
Offspring =
41.23
PND 72:
Dams = NS
Offspring = 7.2

NS


NS


PFOS in blood
(jig/mL)
M



(serum)
PND 21:
Offspring =
18.61
PND 72:
Offspring =
0.56


NS


NS


F
(serum)
GD20:
Dams = 26.63
Offspring =
31.46
PND 21:
Dams = 30.48
Offspring =
18.01
PND 72:
Dams = NS
Offspring =1.99
19.69


44.33




Reference
Butenhoff et al.
2009; Chang et
al. 2009









Thibodeaux et al.
2003;Lauetal.
2003
Thibodeaux et al.
2003;Lauetal.
2003
Note: NS = no sample obtained or recorded
Perfluorooctane sulfonate (PFOS) - May 2016
C-6

-------
             Table C-2. 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)
Length of study
20 days
90 days
182 days
single dose
single dose
single dose
thyroid hormone
activity
1 hour
28 days
28 days
Species
monkey
monkey
monkey
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
1,000 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
1.77 mg/kg/day —
females
(0,0.5,2, 5, or 20
ppm)
5 rats/sex/dose
0.14, 1.33,3.21,6.34
mg/kg/day — males
0.15, 1.43,3.73,7.58
mg/kg/day — females
(0, 2, 20, 50, or 100
mg/kg diet)
15 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 J, body weight
Total thyroxine (TT4)-
significant J, at 2, 6 and
24hrs
Triiodo thy ro nine (TT3)
and reverse
triiodothyronine (rT3)-
significant J, at 24 hrs
Free thyroxine-
significant t 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 | food
consumption (F)
NOAEL = 0.14mg
/kg/day in males and NA
in females
LOAEL = 1.33 mg
/kg/day in males and 0.15
mg/kg/day in females
based on f relative liver
wt
Reference
Goldenthal et al.
1978a
Goldenthal et al.
1979
Seacat et al. 2002
Deanetal. 1978
Yang et al. 2009
Chang et al. 2008
Ruschetal. 1979
Seacat et al. 2003
Curran et al. 2008
Perfluorooctane sulfonate (PFOS) - May 2016
C-7

-------
Method of
exposure
oral gavage



oral gavage



oral (in diet)



oral (in diet)









oral gavage






oral gavage











oral gavage










Length of study
28 days



28 days



90 days



98 days









GD 0 to PND 20a
developmental
neurotoxicity
study



CDs 2-20











CDs 2-21










Species
rat



rat



rat



rat









rat






rat











rat










Concentration
0, 5, or 20 mg/kg/day
10 males/dose


0,0.5, 1,3, or 6
mg/kg/day
19 males/dose

0,2,6, 18, 60, or 200
mg/kg/day
5 rats/sex/dose

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, 2, 3, 5, or 10
mg/kg










0,0.1,0.6, or 2.0
mg/kg









Results
NOAEL= NA
LOAEL= 5 mg/kg/day
based on J, body wt and
lung congestion
NOAEL = NA
LOAEL = 0.5 mg/kg/day
based on J, LH and
testosterone and t FSH
NOAEL= NA
LOAEL= 2 mg/kg/day,
from t liver wt, J, food
consumption
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 f motor activity
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 = cannot be
determined
LOAEL=0.1 mg/kg
based on changes in the
cortex and hippocampus
(astrocyte activation
markers, pro-
inflammatory
transcription factors)

Reference
Cui et al. 2009



Lopez-Doval et al.
2014


Goldenthal et al.
1978b


Seacat et al. 2003









Butenhoff et al.
2009





Thibodeaux et al.
2003 and Lau et
al. 2003









Zengetal. 2011









Perfluorooctane sulfonate (PFOS) - May 2016
C-8

-------
Method of
exposure
oral gavage




oral gavage






oral gavage




oral gavage



oral gavage




















Length of study
CDs 2-21




CDs 1-21






GD 0-PND 20




CDs 11-19



6 wks prior to
mating and
Males — 22 days
Females —
through
gestation,
parturition and
lactation

reproductive
study










Species
rat




rat






rat




rat



rat




















Concentration
0,0.1,0.6, or 2
mg/kg/day



0,0.1, or 2.0
mg/kg/day





0,0.5, or 1.5
mg/kg/day
6 dams/dose


0, 5, or 20 mg/kg/day
4 dams/dose


0,0.1,0.4, 1.6, or 3.2
mg/kg/day
35 rats/sex/dose


















Results
Offspring on PND 21
NOAEL = 0.1 mg/kg/day
LOAEL = 0.6 mg/kg/day
based on increased
apoptosis in heart cells
Offspring
NOAEL= 0.1 mg/kg/day
LOAEL = 2.0 mg/kg/day
based on
histopathological
changes in lungs, J, body
wt and t mortality
NOAEL = NA
LOAEL = 0.5 mg/kg/day
based on J, offspring
body wt, impaired
glucose tolerance
NOAEL = NA
LOAEL = 5 mg/kg/day
based on J, offspring
body wt
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 = NA, higher
dose not tested

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

Reference
Zengetal. 2014




Chen etal. 2012






Lvetal. 2013




Zhao etal. 2014



Luebker et al.
2005b


















Perfluorooctane sulfonate (PFOS) - May 2016
C-9

-------
Method of
exposure
oral gavage














oral (diet)












oral gavage




oral gavage




oral gavage










Length of study
6 wks prior to
mating and
continued
through mating,
gestation and
LD4

reproductive
study






104 weeks












1 time on PND
10

developmental
neurotoxicity
7 days

immunotoxicity
study

CDs 1-17

developmental
immunotoxicity







Species
rat














rat












mouse




mouse




mouse










Concentration
0,0.4,0.8. 1.0, 1.2,
1.6 and 2.0
mg/kg/day
20-28 dams/dose











0, 0.024, 0.098,
0.242, or 0.984
mg/kg/day — males
0,0.029,0.120,
0.299, or 1.251
mg/kg/day — females

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



0, 5, 20, or 40 mg/kg
12 male mice/dose



0,0.1, 1, or 5 mg/kg
10-12 female
mice/dose








Results
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= 0.024
mg/kg/day
Males LOAEL= 0.098
mg/kg/day based on liver
histopathology

Females NOAEL = 0.120
mg/kg/day
Females LOAEL = 0.299
mg/kg/day based on liver
histopathology
Suggestive of
carcinogenicity
Mice at both
concentrations showed J,
activity and t
neuroprotein levels in the
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

Reference
Luebker et al.
2005a













Thomford 2002/
Butenhoff et al.
2012










Johansson et al.
2008, 2009



Zheng et al. 2009




Keil et al. 2008









Perfluorooctane sulfonate (PFOS) - May 2016
C-10

-------
Method of
exposure
oral gavage












oral gavage






oral gavage










oral gavage



oral gavage




oral gavage






Length of study
CDs 1-17












CDs 12-18

developmental




CDs 1-17/18

developmental








14 days

With regular or
high fat diet
3-21 days




GDO-PND21






Species
mouse












mouse






mouse










mouse



mouse




mouse






Concentration
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, 5, or 20 mg/kg/day
16 males/dose/diet


0, 1, 5, or 10
mg/kg/day
4 males/dose


0, 0.3, 3 mg/kg/day
6 dams/dose





Results
Maternal
NOAEL= 1 mg/kg
LOAEL= 5 mg/kg based
on 1 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
NOAEL = NA
LOAEL = 5 mg/kg/day
based on wt loss on high
fat diet
NOAEL = 1 mg/kg/day
LOAEL = 5 mg/kg/day
based on f liver wt,
changes in oxidation
biochemical parameters
NOAEL = 0.3 mg/kg/day
LOAEL = 3 mg/kg/day
based on f liver wt in
dams and male offspring,
t fasting serum insulin in
males

Reference
Thibodeaux et al.
2003;Lauetal.
2003










Fuentes et al. 2007






Yahia et al. 2008










L. Wang et al.
2014


Wan etal. 2012




Wanetal. 2014b





Perfluorooctane sulfonate (PFOS) - May 2016
C-ll

-------
Method of
exposure
oral gavage










oral gavage





oral gavage






dermal

ocular



Length of study
28 days

immunotoxicity








60 days

immunotoxicity



90 days

neurotoxicity




single dose

single dose



Species
mouse










mouse





mouse






rabbit

rabbit



Concentration
0, 0.00017, 0.0017,
0.0033,0.017,0.033,
or 0.166 mg/kg
5 mice/dose







0, 0.008, 0.083,
0.417, 0.833, or 2.083
mg/kg
10 male mice/group


0,0.43,2.15, or 10.75
mg/kg/day
15/group, sex not
specified



0.5g*
(no data on gender)
0.5 g*
(no data on gender)


Results
Males NOAEL= 0.00017
mg/kg
Males LOAEL= 0.00 17
mg/kg based on J, plaque
forming cell response
Females NOAEL=
0.0033 mg/kg
Females LOAEL= 0.0 17
mg/kg
based on J, plaque
forming cell response
NOAEL = 0.008
mg/kg/day
LOAEL = 0.083 mg/kg
based on f splenic NK
cell activity and t liver
weight
NOAEL = 0.43
mg/kg/day
LOAEL = 2. 15
mg/kg/day based on
changes in water maze
and histopath. in
hippocampus
No irritation

Exact score not provided
except maximal score at
1 and 24 hrs

Reference
Peden-Adams et
al. 2008









Dong et al. 2009





Long etal. 2013






Biesemeier and
Harris 1974
Biesemeier and
Harris 1974

Notes: *Exact dose not provided; NA= not applicable; could not be determined
a GD = gestation day and PND = post natal day
M = male; F = female
Perfluorooctane sulfonate (PFOS) - May 2016
C-12

-------