DEQ-CFW_00004207

Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
�1?
2 Evaluation and prediction pharmacokinetics of PFOA and PFOS
using3 and human a PBPKmodel
4 Anne E. Loccisano *, Jerry L. Campbell Jr., Melvin E. Andersen, Harvey J. Clewell III
5 Center for Human Health Assessment, The Hamner Institutes for Health Sciences, 6 Davis Drive, P.O. Box 12137, Research Triangle Park; NC 27709, United States
6
2§
ARTICLE I NFO
9
Article history:
10
Received 18 August 2010
11
Available online xxxx
12
Keywords:
13
Physiologically -based pharmacokinetic
14
modeling
15
PBPK
16
Toxicokmetic modeling
17
Perfluoroalkyls
18
Pharmacokinetics
19
PFOA
20
PFOS
21
V 1.Introduct1on
A B S T R A C T
Perfluoroalkyl acid carboxylates and sulfonates;,(PFAAs) have many consumer and industrial applications. 23
The persistence and widespread distribution of these compounds in humans have brought them under 24
intense scrutiny. Limited pharmacokinetic data is available in humans; however, human data exists for 25
two communities with drinking water contaminatedby PFAAs. Also, there is toxicological and pharma- 26
cokinetic data for monkeys, which can be quite useful for cross -species extrapolation to humans. The goal 27
of this research was to develop a physiologically -based pharmacokinetic (PBPK) model for PFOA and PFOS 28
for monkeys and then scale this model to humans in order to describe available human drinking water 29
data. The monkey model simulations were consistent with available PK data for monkeys. The monkey 30
model was then extrapolated to the human and then used to successfully simulate the data collected 31
from residents of two communities exposed to PFOA in drinking water. Human PFOS data is minimal; 32
however, using the half-life estimated from occupational exposure, our model exhibits reasonable agree- 33
ment with the available human serum PFOS data. It is envisioned that our PBPK model will be useful in 34
supporting human health risk assessments for PFOA and PFOS by aiding in understanding of human 35
pharmacokinetics. 36
© 2010 Published by Elsevier Inc. 37
38
39
42 Perfluorooctanoate (PFOA) and perfluorooctane sulfonate
43 (PFOS) are fully fluorinated man-made chemicals that have wide
44 spectrum of industrial and consumer uses. These compounds are
45 used as surfactants in situations requiring stability towards heat
46 and chemical degradation or other unique ,properties imparted
47 Ql by the perfluorinated chain (Lau et al., 2007). They can also be
48 formed from the degradation (environmental' or metabolic) of cer-
49 tain fluorinated materials (Olsen et al., 2007, 2003a). Due to the
50 high electronegativity of the fluorine atom, the C-F bond has a sig-
51 nificant dipole moment, imparting ionic character to the bond
52 through partial charges; the partial charges are attractive, making
53 the C-F bond one of the strongest in organic chemistry (Lau et al.,
54 2007). The stability imparted to these compounds by the C-F bond
55 is very desirable industrially; they are stable in air at high temper-
56 ature, nonflammable, and not readily degraded by strong acids,
57 bases, or oxidizers (Lau et al., 007). The C-F bond also provides
58 these compounds with low surface tension making them ideal
Abbreviations: PFAAs, collectively refers to perfluoroalkyl acid carboxylates and
sulfonates; PBPK model, physiologically based pharmacokinetic model; PFOA,
perfluorooctanoate; PFOS, perfluorooctane sulfonate.
* Corresponding author. Fax: [phone].
E-mail address: [e-mail] (A.E. Loccisano).
0273-2300,'S - see front matter 2 2010 Published by Elsevier Inc.
doi:10.1016(j.yrtph.2010.12.004
surfactants (Lau et al., 2004). However, the same properties that 59
make these compounds so useful make them resistant to break- 60
down by metabolism, hydrolysis, photolysis, or biodegradation 61
and thus they can be persistent in the environment indefinitely. 62
Industrial use of PFOA began in the late 1940s, and since then it 63
has been estimated that cumulative global emissions are between 64
2400 and 5200 metric tons (Prevedouros et al., 2006). Production 65
and use of PFOS and materials which can form PFOS were initially 66
greater than that of PFOA, but since the phaseout of PFOS by its 67
major manufacturer between 2000 and 2002, global production 68
has dropped (175 metric tons in 2003) (3M, 2003). 69
Although exposure sources and routes for humans are not well 70
understood, perfluoroallyl carboxylates and sulfonates (collec- 71
tively perfluoroalkyl acids or PFAAs) escape to the environment 72
through production, manufacture, disposal, and degradation of 73
other fluoropolymers (Andersen et al., 2008). These compounds 74
are now widespread throughout the global ecosystem. They have 75
been found in wildlife samples from remote ocean sites and the 76
Arctic, although concentrations are usually greater in animals liv- 77
ing in more populated and industrial areas (Lau et al., 2007). These 78
compounds have also been identified in parts -per -billion levels in 79
human serum from the general population and in parts -per -mil- 80
lion levels for occupationally exposed workers and other popula- 81
tions with high exposure to these compounds (Calafat et al., 82
2007b; Emmett et al., 2006; Olsen et al., 2007). The long plasma 83
DEQ-CFW 00004207

2 A.E. Loccisano el al./Regulatory Toxicology and Phonnocology xxx (2010) xxx-xxx
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
ill
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
half-lives of PFOA and PFOS (3-5 years) observed in humans is of
particular concern because this indicates that they can accumulate,
which may result in higher body burdens and potential adverse
health effects. Possible exposure sources for humans include drink-
ing water, dust in homes, food and food packaging, fabrics, carpet-
ing, and cookware (Fromme et al., 2009; Ostertag et al., 2009;
Sinclair et al., 2007; Skutlarek et al., 2006; Strynar and Lindstrom,
2008; Vestergren and Cousins, 2009; Washburn et al., 2005; Wil-
helm et al., 2008).
The pharmacokinetic properties of PFOA and PFOS have been
studied. In summary, animal studies have shown that these com-
pounds are well absorbed orally, poorly eliminated, and not
metabolized (Johnson et al., 1984; Kuslikis et al., 1992; Ophaug
and Singer, 1980; Vanden Heuvel et al., 1991). They are distributed
mainly to the serum, liver, and kidney, usually with liver concen-
trations being 3-5 times higher than serum concentrations in the
rat, while liver concentrations in humans and non -human primates
are only -1.3-2 times higher than in serum (DePierre, 2009;
Hundley et al., 2006; Johnson et al., 1979; Kemper, 2003; Kudo
et al., 2007; Olsen et al., 2003b; Seacat et al., 2002). Distribution
at steady state is mainly extracellular (Butenhoff et al., 2004a; No-
ker and Gorman, 2003). Both compounds have a high affinity for
binding to albumin and are thus highly bound in plasma, and to
a lesser extent, P-lipoproteins and liver fatty acid binding protein
(Han et al., 2003; Kerstner-Wood et al., 2003; Luebker et al.,
2002). Both compounds have been found in rodent fetuses as well
in as umbilical cord blood of humans and in breast milk indicating
that they can cross the placenta and partition into milk, exposing
the fetus and neonate (Apelberg et al., 2007; Hinderliter et al.,
2005; Karrman et al., 2007; Kuklenyik et al., 2004; Thibodeaux
et al., 2003). The most notable aspect of PFOA and PFOS PK is that
they exhibit major differences in plasma half-lives across species,
and for PFOA, a sex difference in elimination is observed in rats,
The elimination half-life for PFOA in female rats is 2-4 h; it is 4-
6 days in male rats (Hanhijarvi et al., 1987; Kemper, 2003; Kudo
Both compounds are agonists for peroxisome proliferator activated
receptor alpha (PPAR-a) and the tumor induction has been attrib-
uted to this activation (Elcombe et al., 2010). However, PPAR-OE
activation may not be of human relevance because humans express
much lower levels of these receptors than do rodents and human
liver cells have been shown to be much less responsive to the ef-
fects of ammonium PFOA than rat liver cells (13jork and Wallace,
2009; Lake, 2009).
Due to their prevalence and stability in the environment, toxic
effects observed in animal studies, and long half-life in humans,
PFAAs have drawn considerable attention from public and regula-
tory agencies with regard to health risks that they may present.
Risk assessments and interpretation of available human data are
slowed due to lack of framework to understand and estimate hu-
man pharmacokinetics. Previously, a biologically based compart-
mental model was developed by our group that was able to
describe the complex PK of PFOA and PFOS in rats and monkeys
(Andersen et al., 2006; Tan et al,, 2008). The key process required
to describe kinetics was resorption by renal transporters in the fil-
trate compartment, The goal of the present work was to develop a
complete PBPK model for the monkey for PFOA and PFOS and then
scale up to the human, The consistency of renal resorption to de-
scribe the kinetics in both the rat and monkey suggests the exis-
tence of a saturable, high -affinity resorption process that governs
the kinetics of these compounds in other species, including hu-
mans (Tan et al., 2008). If confirmed, once the transporter protein
that is responsible for renal resorption of PFOA and PFOS is identi-
fied, information on the activity of that transporter can then be
used to identify susceptible sub -populations due to polymorphic
variation. The human model was used to successfully simulate
available data and also to predict and thus examine variability in
human half-lives of PFOA. The result of these efforts is a human
PBPK model with reasonable predictive ability that can be used
to aid in risk assessment for perfluorinated compounds.
et al., 2002). The gender difference in elimination is age -dependent 2. Methods
and the slower elimination in males is observed starting around 3-
5 weeks in age (Hinderliter et al., 2006). This gender difference in
11 11 F_ Dune ; 11 IF I�F_ ; '.� 1) 2. 1. Monkey model development
elimination s no o serve r n rats (half S - -
50 days, depending on the isomer) or mice (half-life is -40 days)
(Benskin et al., 2009; Butenhoff, 2010). Monkeys have half-lives
of 21-30 days for PFOA and 4-6 months for PFOS while human
half-lives are even longer for both compounds (3-5 years) (Bartell
et al., 2010; Butenhoff et al., 2004a; Noker and Gorman, 2003; Ol-
sen et al., 2007; Seacat et al., 2002). Neither the monkey nor human
appears to have any major sex differences in elimination of either
compound, and this appears to be true for the mouse also (Buten-
hoff et al., 2004a; Noker and Gorman, 2003; Olsen et al., 2007;
Rodriguez et al., 2009; USEPA, 2005). The cause of the differential
elimination across sexes and species has not been conclusively
demonstrated, but there is evidence from studies performed both
in rats and in vitro with rat and human transporters that transport-
ers in the proximal tubule of the kidney may be responsible
(Andersen et al., 2006; Kudo et al., 2002; Tan et al., 2008; Weaver
et al., 2009; Yang et al., 2010, 2009).
Potential toxicities in animals have been characterized. Sub -
chronic exposure to PFOS led to significant weight loss, reductions
in serum cholesterol and thyroid hormones, and hepatotoxicity.
Developmental toxicity studies with PFOS in rats, mice, and rabbits
reveal reduction of fetal weight, cleft palate, and reduced neonatal
survival (Lau et al., 2004). Repeated -dose studies with PFOA in ro-
dents showed induction of liver tumors, Leydig cell tumors, and
pancreatic acinar cell tumors (Andersen et al., 2008; Biegel et al.,
200 1; USEPA, 2005). Reproductive toxicity studies with PFOA in ro-
dents showed increased post weaning mortality, decreased body
weight, and delayed sexual maturation (Butenhoff et al., 2004b).
The structure for the monkey model was developed from a bio-
logically based compartmental model for PFOA and PFOS in rats
and monkeys (Fig. 1) (Andersen et al., 2006; Tan et al., 2008).
The compartmental model contained compartments for plasma, li-
ver, tissues, and filtrate, and it was developed to examine the role
of renal resorption through a saturable transport process that is
thought to be responsible for the long half-life of PFOA and PFOS
in rats and monkeys. Free chemical in the central (plasma) com-
partment can move to the liver compartment, tissue compartment,
or be filtered by the filtrate (i.e., kidney) compartment. Once in the
filtrate compartment, the chemical can be resorbed back to the
plasma compartment by a saturable process with transporter max-
imum Tm and affinity constant Kt. This model was able to success-
fully describe PFOA and PFOS kinetics from oral and IV dosing in
rats and monkeys, but in order to extrapolate the model to the hu-
man to aid in risk assessment, a physiologically based model with
realistic tissue volumes and blood flows and physicochemical and
biochemical properties of the chemical under study were needed.
Thus, a physiologically -based pharmacokinetic (PBPK) model was
developed for PFOA and PFOS in the monkey for use in extrapola-
tion to the human.
2.1.1. Study designs and data - PFOA in monkeys
Pharmacokinetic data for PFOA in the monkey have been re-
ported by Butenhoff et al. (2004a) for oral and IV dose routes. Both
data sets were used to develop the current model. In the IV dosing
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
DEQ-CFW-00004208

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
Oral dose
Liver compartment F---> feces
gof m
Central Compartment
IV- ___ (vol. of distr., free fraction of chemical in
plasma)
k1 3, k31 QFill Tin, Kt
Tissue Filtrate Compartment
compartment (vol. of renal filtrate, renal filtrate rate,
saturable resorption)
storage > urine
to develop the current monkey model. The IV data has been re-
ported by Noker and Gorman (2003). Male and female monkeys
(3(sex) were administered a single IV dose of 2 mg/kg PFOS and
then monitored for 161 days after dosing. PFOS concentrations in
plasma and urine of the monkeys were measured 0.5, 2, 4, 8, and
24 h post dose and on days 2, 7, 14, 21, 28, 42, 56, 70, 105, and
161 post dose. The other data used for model development was
from a subchronic toxicity study of PFOS reported by Seacat et a].
(2002). Male and female monkeys received daily oral doses (by
capsule) of 0.03, 0.15, or 0.75 mg/kg PFOS for 26 weeks. Two mon-
keys per sex in the 0.15 and 0.75 mg/kg dose groups were allowed
to recover and monitored for a year. Plasma concentrations of PFOS
were measured throughout the dosing period for all 3 dose groups
and during the recovery period for the mid- and high -dose groups.
Significant adverse effects were observed only in the high -dose
group, which included death of 2 of 6 male monkeys for which
PFOS exposure could not be ruled out as a contributing cause.
The monkey model was used to simulate the data from both the
IV and oral studies with PFOS in the cynornoIgus monkey.
Fig. 1. Structure of biologically motivated model for PFOA and PFOS in rats and
monkeys. Figure adapted from Tan et al. (2008). Free chemical in the central
2.2. Model structure - monkeys
compartment can move to the liver compartment based on flow rate OL and
partition coefficient PL, to the tissue compartment based on the rate constants k13
The PBPK model structure for PFOA and PFOS in the monkey is
and k31, or be filtered by the filtrate compartment with flow rate Qfil. Chemical can
resorbed back to the plasma compartment with transporter maximum Trn and
the same (Fig, 2). The model contains compartments for plasma,
affinity constant Kt.
gut, liver, kidney, filtrate, fat, skin, and a lumped compartment
for remaining body tissues. The plasma and liver are the primary
target tissues for PFOA and PFOS with possible involvement of
211
study, male and female cynornoIgus monkeys (3/sex) received a
enterohepatic circulation (Butenhoff et al., 2004a; Johnson et al.,
212
single IV dose of 10 mg/kg PFOA and were then monitored for
1979; Kemper, 2003; USEPA, 2005). These compounds also activate
213
123 days after dosing. PFOA concentration in plasma was mea-
nuclear receptors PPAR-ot and CAR/PXR known to be associated
214
sured at 0.5, 2, 4, 8, and 24 h post dose and on days 2, 4, 7, 11,
with a hypertrophic response in liver tissue (Elcombe et al.,
215
14, 21, 28, 57, 79, 87, and 123. No adverse effects related to PFOA
2010; Rosen et al., 2010). The liver in some cases is the site ofaccu-
216
were reported (Butenhoff et al., 2004a). In the oral dosing study,
mulation (Andersen et al., 2008; Kennedy et al., 2004; Lau et al.,
217
male cynornolgus monkeys were administered a daily oral dose
218
by capsule of 0 (n = 6),3 (n = 4), 10 (n = 6), or 20 mg/kg (n = 6) PFOA
219
for 6 months. Several of the animals were then allowed to recover
k Oral dose., drinking water
220
and were monitored for 90 or more days after dosing cessation
Gut
221
(n = 2 for the 10 mg/kg dose group and n = 2 in the 20 mg/kg dose
QGUt
222
group). Monkeys in the 20 mg/kg dose group were originally
223
started on 30 mg/kg PFOA; however, dosing had to be stopped after
giver j
224
12 days due to toxicity. Dosing began again at day 21, and the dose
QLiv
225
was lowered to 20 mg/kg. Dosing again had to be discontinued for
Fat
226
three of the monkeys due to toxicity on day 43, 66, and 81. Only
Plasma U-al - --------------------------------------
227
two monkeys in the high -dose group received treatment for the
IV
Skin
228
entire 6 months. During the dosing period, PFOA concentration in
Free QS!kn -------------------------------------
229
plasma and urine of the monkeys still receiving doses were taken
fraction
230
at weeks 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, and 26. During
Best of body
231
recovery, plasma and urine PFOA concentrations were measured
QR
232
for only the 10 and 20 mg/kg dose groups; these were measured
Kidney
K,'-'ne
233
once every two weeks from weeks 28 through 40 of the study. Be-
QIK i d T rn, M
234
cause there was no recovery data for the 3 mg/kg dose group, and
235
the monkeys in the 30/20 mg/kg dose group exhibited signs of tox-
Filtrate
QH Filtraa te
236
icity and had an irregular dosing regimen, only the plasma and ur-
------------ ----
237
ine data from the 10 mg/kg dose group was used for development
238
of the monkey model for PFOA by the oral route. Also, although the
storarle
239
10 mg/kg dose would be a much higher concentration than hu-
kurine
240
mans would be exposed to environmentally, the lower dose and
urine
241
the effects and PK associated with it would be more relevant for
242
development of a human model. The monkey model was used to
Fig. 2. Structure of PBPK model for PFOA and PFOS in monkeys and humans.
243
simulate the data from both the IV and oral studies with PFOA in
Chemical is taken up into the plasma (iv) or into the gut (oral). From the gut,
244
the cynomolgus monkey.
chemical is transported to the liver by the portal blood. Only the free fraction of
chemical in plasma is assumed to be available for partitioning into tissues. Chemical
is eliminated through the filtrate compartment to storage into urine. While in the
245
2.1.2. Study designs and data - PFOS in monkeys
filtrate compartment, chemical can be reabsorbed back into the plasma through a
246
Similar to PFOA, there is existing PK data for PFOS in cynornol-
.saturable process with a transporter maximum Tm and affinity constant Kt. The Qs
indicate blood flows into and out of tissues. Qfil is not a blood flow; it is a clearance
247
gus monkeys for IV and oral dose routes. Both data sets were used
(L/h) from the plasma to the filtrate compartment.
DEQ-CFW-00004209

A.E. Loccisano el al./Regulatory Toxicology and Phonnocology xxx (2010) xxx-xxx
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
2004; Luebker et al., 2002; USEPA, 2005). The PBPK model retains
the same features that were important for describing kinetics of
PFOA and PFOS in the biologically based model of Andersen and
Tan, namely, a free fraction of chemical in the plasma and renal
resorption in the filtrate compartment (Andersen et al., 2006;
Tan et al., 2008). The fat and skin compartments were built into
the monkey model because these compartments will be of interest
in the human. PFOA and PFOS are lipophobic, thus fat would serve
as an exclusionary volume for these chemicals, and dermal absorp-
tion is a possible exposure route for the human (USEPA, 2005). An
IV dose to the monkey enters the plasma, and oral doses enter the
gut and are then passed to the liver by the portal blood. Only the
free fraction of chemical in plasma is assumed to be available to
partition into tissues. Chemical is eliminated through the filtrate
compartment to storage and subsequently to urine. The clearance
from plasma to the filtrate compartment is described by a flow
rate, QFil. In the filtrate compartment, chemical can be reabsorbed
by a saturable process with a transporter maximum (Tm) and
affinity constant (Kt) into the kidney and back into systemic circu-
lation. Coding and simulations for both the PFOA and PFOS monkey
models was performed in the Berkeley -Madonna program (Macey
et al., 2000). Model code can be found in the Supplementary
Material.
from tissue concentration data from male C57131/6 mice (mice were
fed PFOS in their diet from one to five days) (DePierre, 2009). PFOA
and PFOS have been shown to be >97% bound to plasma proteins
such as albumin in the rat, monkey, and human (Han et al.,
2003; Jones et al., 2003; Kerstner-Wood et al., 2003); a constant
free fraction of each chemical (<0.03) was used. Because of the lim-
ited evidence of extensive binding for PFOA and PFOS, it was nec-
essary to estimate the free fraction of chemical in plasma (Free) for
each compound to provide the best fit the plasma concentration
data from the IV and oral studies while adhering to the aforemen-
tioned constraint.
It is hypothesized that the long plasma half-lives of PFOA and
PFOS are due to renal resorption by transporters in the proximal
tubule of the kidney (Andersen et aL, 2006; Tan et al., 2008). In fact,
more recent research has suggested that renal tubular resorption
transporters are capable of transporting PFOA (Nakagawa et al.,
2009; Weaver et al., 2009; Yang et al., 2010, 2009). In order to de-
scribe the kinetics of these chemicals, a description of saturable re-
nal resorption was implemented in the filtrate compartment in the
model. The parameters used to describe renal resorption are the
transporter maximum (Tm) and the transporter affinity constant
(Kt). In the rat, a significant sex difference is observed in the clear-
ance of PFOA but not PFOS (Andersen et al.. 2008: Kemper. 2003:
Kennedy et al., 2004; Kudo et al., 2002). It has been shown that
302
303
304
305
306
307
308
309
2.3. Model parameterization
2.3.1. Chemical -specific parameters for PFOA and PFOS
Chemical -specific parameters used for the monkey PFOA and
PFOS models are presented in Table 1. Tissue:plasma partition
coefficients for PFOA for all tissues in the model were estimated
from tissue concentration data reported by Kudo et al. (2007) for
a single IV dose of PFOA to male Wistar rats. Tissue:blood partition
coefficients for PFOS for all tissues in the model were estimated
Table I
Chemical parameters for PBPK models for PFOA and PFOS in monkeys and humans.
male rats exhibit higher expression of a particular organic anion
transporter, Oatplal, that is thought to play a role in renal resorp-
tion, thus extending the half-life of PFOA in the male rat (Weaver
et al., 2009; Yang et al., 2009). However, no significant sex differ-
ences are observed in elimination of either PFOA or PFOS in the
monkey (Butenhoff et al., 2002, 2004a; Seacat et al., 2002), both
chemicals have long half-lives in both sexes of the monkey
(20-30 days for PFOA; 5-6 months for PFOS). Therefore, in the
monkey model, the male and female monkeys have the same
PFOA Definition and units Value
Source
Tmc Resorption maximum 0.15 (monkey); 10 (human; 3.8 yrs); 6
Fit to plasma concn. in monkey; estimated to give desired half-life in
(mg/h(kg075)a (human (2.3 yrs)
human (3.8 or 2.3 yrs)
Kt Affinity constant (mg/L) 0.055
Fit plasma and urine concn. in monkey
Free Free fraction of chemical in 0.02
Fit to plasma concn. in monkey
plasma
PL Liver:plasma partition 2.2
Rat tissue data (Kudo et al., 2007)
coefficient
PF Fat:plasma partition coefficient 0.04
Rat tissue data (Kudo et al., 2007)
PK Kidney:plasma partition 1.05
Rat tissue data (Kudo et al., 2007)
coefficient
PSk Skin:plasma partition coefficient 0A
Rat tissue data (Kudo et al., 2007)
PR Rest of body:plasma partition 012
Rat tissue data (Kudo et al., 2007)
coefficient
PG Gut:plasma partition coefficient 0.05
Rat tissue data (Kudo et al., 2007)
Kurine Urinary elimination rate (/h/ 50 (monkey); 3e-' (human)
Fit to urine concn. in monkey and human
kg -0.25)
PFOS
Tmc Resorption maximum (mglhl 1.3 (monkey); 3.5 (human)
Fit to plasma concn. data in monkey; estimated to give desired half-life in
kg 0.75r
human (5.4 yrs)
Kt Affinity constant (mg(L) 0.023
Fit to plasma concn. data in monkey
Free Free fraction of chemical in 0.025
Fit to plasma concn. in monkey
plasma
PL Liver:plasma partition 3.72
Mouse tissue data (DePierre)
coefficient
PF Fat:plasma partition coefficient 0.14
Mouse tissue data (DePierre)
PK l(idney:plasma partition 0.8
Mouse tissue data (DePierre)
coefficient
PSk Skin:plasma partition coefficient 0.29
Mouse tissue data (DePierre)
PR Rest of body:plasma partition 0.2
Mouse tissue data (DePierre)
coefficient
PG Gut:plasma partition coefficient 0.57
Mouse tissue data (DePierre)
Kurine Urinary elimination rate (/h/ 0.003 (monkey); 0.001 (human)
Fit to urine concn. in monkey and human
kg --0.25)
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
DEQ-CFW-0000421 0

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
61
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
values for Tm and Kt; i.e., we assume that the transporter(s)
responsible for resorption of PFOA would be the same in both
sexes. The same goes for PFOS; however, we assume that either
the transporter(s) responsible for resorption of PFOS is different
from that of PFOA or simply that the activity of each compound
with the transporter(s) is different. PFOS has a longer half-life;
the capacity and affinity of the transporter are thus higher than
those for PFOA. The transporter maxima and affinity constants
for each chemical were estimated by fitting them to the plasma
and urine concentration data for PFOA and PFOS in both sexes.
The transporter maximum was scaled as BW°.75. The urinary elim-
ination rate constant (Kurine) was estimated to fit the urine con-
centration data from the IV doses of chemicals.
2.3.2. Physiological parameters for the monkey model - PFOA and
PFOS
The physiological parameters for the monkey can be found in
Table 2. Body weights for the monkeys were obtained from Buten-
hoff et al. (2004a) and Seacat et al. (2002). Fractional tissue vol-
umes were the values reported from Brown et al. (1997) and
Forsyth et al. (1968) for rhesus monkeys. Tissue volumes were
scaled linearly to body weight. Cardiac output and fractional blood
flows to tissues were obtained from Fisher et al. (2000) and Forsyth
et al. (1968) for rhesus monkeys. Blood flows were scaled as
BW0.75. Because PFOA and PFOS do not partition into red blood
cells, (Ehresman et al., 2007) blood flows were adjusted for plasma
flows to tissues by multiplying the blood flow by 1-hematocrit.
sure that the simulations were consistent with both the plasma
and urine concentrations. The shapes of the simulated curves were
assessed to make sure that they were in line with the trends ob-
served in the experimental data. Agreement between the experi-
mental data and the model simulations was good overall. Further
efforts to refine the model parameters will require more detailed
studies (i.e., plasma protein binding and transporter elucidation
in the monkey).
A normalized local sensitivity analysis was performed on the
model to examine the influence of each model parameter on the
model output. Sensitivity coefficients were calculated for the pre-
dicted plasma area under the curve (AUC; total concentration) with
the original parameters and for that resulting from a 1% change in
each parameter value. The normalized sensitivity coefficients were
calculated by the following equation:
A B
Sensitivity coefficient - rB
A is the AUC resulting from the 1 % increase in parameter value, B is
the AUC resulting from the original parameter value, C is the param-
eter value increased by 1%,'and D is the original parameter value.
Plasma AUC predictions for the monkeys were run at 3 and
10 mg1kg doses for PFOA and 0.03 and 0.75 mg/kg doses for PFOS.
2.5. Human model development
2.5.1. Study designs and data - human
PFOA and PFOS health and pharmacokinetic data in humans is
2.4. Model evaluation available from occupational and environmental and epidemiologi-
cal investigations and worker medical surveillance. Developmental
For PFOA, we used the monkey model to simulate plasma and and immunotoxic effects have been reported in studies with rats
urine concentrations of the chemical resulting from IV and oral and mice exposed to PFOA and PFOS. However, fetal effects in
dose routes. The simulated concentrations were compared to the the offspring of treated animals were observed only at high doses
experimental data of Butenhoff et al. (2002, 2004a). For PFOS, the that caused maternal toxicity (Lau et al., 2004). Similarly, the inter -
monkey model was used to simulate plasma concentrations of pretation of decreased spleen and thymus weights and altered nat-
PFOS from a single IV dose and daily oral doses; the simulated con ural killer cell activity in the animals in the immunotoxicity studies
centrations were then compared to the experimental data of Noker were complicated by the high doses used in the studies and also re -
and Gorman (2003) and Seacat et al. (2002). duced food intake (DeWitt et al., 2009; Dong et al., 2009; Qazi
The adequacy of the monkey models and the corresponding et al., 2010; Yang et al., 2000). Repeat -dose studies in rodents
parameters were assessed based on whether the model was able and monkeys with PFOS have shown reduced body weight,
to consistently describe the available pharmacokinetic data on increased liver weight, reduced cholesterol, and a steep dose-
PFOA and PFOS in cynomolgus monkeys. To evaluate the current response curve for mortality (Goldenthal, 1978; Seacat et al.,
monkey models, the correspondence between the data points for 2002). A chronic bioassay with PFOS in SD rats showed an increase
tissue concentrations of the chemicals and the model -predicted in hepatocellular adenomas at only the highest dose of 20 ppm in
values at different time points were examined visually to make the diet (Lau et al., 2007). Two separate chronic bioassays have
Table 2
Physiological parameters for monkey and human PBPI{ models
Physiological parameters
Monkey
Human
Source
Cardiac output (L(h/kg075), QCC
15
12.5
Forsyth et al. (1968), Brown et al. (1997)
Blood flow, fraction of CO
Fat, QFC
0.02
0.052
Fisher et al. (2000), Brown et al. (1997)
Liver, QLC
0.2
0.25
Fisher et al. (2000), Brown et al. (1997)
Kidney, QKC
0.123
0.175
Forsyth et al. (1968), Brown et al. (1997)
Filtrate, QFiIC
0.0246
0.035
20% of kidney blood flow; GFR
Gut, QGC
0.096
0.181
Forsyth et al. (1968), Brown et al. (1997)
Skin, QSkC
0.048
0.058
Forsyth et al. (1968), Brown et al. (1997)
Hematocrit, Htc
0.41
0.44
Davies and Morris (1993), Brown et al. (1997)
BW (kg, average)
3.5-5.5
70
Butenhoff et al. (2004a,b), Seacat et al. (2002), Brown et al. (1997)
Tissue volume, fraction of BW
Liver, VLC
0.026
0.026
Brown et al. (1997)
Fat, VFC
0.05
0.214
Fisher et al. (2000), Brown et al. (1997)
Kidney, VKC
0.004
0.004
Brown et al. (1997)
Filtrate, VFilC
0.0004
0.0004
10 0 of kidney volume
Gut, VGC
0.017
0.017
Brown et al. (1997)
Plasma, VP1asC
0.0448
0.0428
Davies and Morris (1993), Brown et al. (1997)
DEQ-CFW 00004211

6 A.E. Loccisano el al./Regulatory Toxicology and Phonnocology xxx (2010) xxx-xxx
428
examined the turnorigenecity of PFOA. In the study by Biegel et al.,
measured from a cross-section (divided into public water users
494
429
PFOA was found to increase the incidence of benign hepatocellular,
and bottled water users) of the Little Hocking and Lubeck popula-
495
430
Leydig cell, and pancreatic acinar cell tumors at only a high dose of
tions by Bartell et al. (2010). Pre -filtration concentrations were
496
431
300 ppm in the diet (Biegel et al., 2001). In a Riker Pharmaceuticals
measured in May/June 2007; filtration systems were installed in
497
432
study, PFOA increased Leydig cell hyperplasia only at a high dietary
Lubeck in June 2007 and in Little Hocking in November 2007.
498
433
dose of 300 ppm (Riker, 1987).
Post -filtration concentrations were measured in May/June 2008.
499
434
However, the PPAR-a mode of action that is responsible for tu-
Although the population size for this study was small (N = 132
500
435
mor development in rodents is most likely not relevant for humans
for Lubeck; N = 40 for LH), the model was used to simulate the
501
436
because PPAR-a activation in humans does not lead to a hyperplas-
pre- and post -filtration serum PFOA concentrations. When simu-
502
437
tic response (Gonzalez and Shah, 2008; Kennedy et al., 2004; Lake,
lating the LH population, the pre -filtration drinking water concen-
503
438
2009). PPAR-ot activators are non-genotoxic carcinogens, and they
tration was set at 3.55 ppb; for the Lubeck population the drinking
504
439
include hypolipidemic drugs and cholesterol -lowering drugs. In
water PFOA concentration ranged from 0.41 to 1 ppb (Bartell et al.,
505
440
contrast to the liver tumors observed when these agents are chron-
2010), and for the simulations, 0.7 ppb (middle of the range) was
506
441
ically administered to rats and mice, there is no evidence that
used.
507
442
administration over an extended period of time in humans are
Another study performed by Holzer et al. examined the serum
508
443
associated with liver or any other types of tumors in humans,
PFOA concentrations in a population in Arnsberg, Germany in July
509
444
which indicates a species difference in turnorigenic response (Ash-
2006 (Holzer et al., 2008). As in Little Hocking, the drinking water
510
445
by et al., 1994; Gonzalez and Shah, 2008). Epidemiology studies of
had been contaminated with PFOA, and this is thought to have aris-
511
446
exposure to PFOA and PFOS and adverse health events are incon-
en from farmers dumping soil conditioner mixed with industrial
512
447
elusive; in fact, no causal associations of adverse health effects to
waste into nearby rivers. The average concentration of PFOA in
513
448
PFOA or PFOS exposure have been observed (Lau et al., 2007; USEP-
the German public water was found to be approximately seven
514
449
A, 2009). Most of the human data available for PFOA is for a popu-
times lower than that measured in Little Hocking public water
515
450
lation in Little Hocking, Ohio. The residents of this community get
(0.519 ppb versus 3.55 ppb) (Holzer et al., 2008). The data in the
516
451
their drinking water from the Ohio River, which is adjacent to a
HbIzer et al. study was from approximately 355 people (n = 101
517
452
plant in Washington, West Virginia that has manufactured fluoro-
for men, n = 164 for women, n = 90 for kids); as in the Emmett
518
453
polymers since 1951. The highest levels of PFOA in public water
et al. study, only a mean and standard deviation for the population
519
454
supplies in the US have been found in the Little Hocking water sys-
surveyed was available, and it is not known how long the people
520
455
tem (Emmett et al., 2006). Emmett et al. measured the serum PFOA
had been exposed to PFOA; however, it is thought to be only a
521
456
levels in residents of the Little Hocking population as well as the
few years in contrast to the Little Hocking population that has been
522
457
concentration of PFOA in the public drinking water (mean concen-
exposed for over 30 years (Holzer et al., 2008). After the study was
523
458
tration = 3.55 ppb), which is the primary route of exposure for
conducted in July 2006, a filtration system was installed to remove
524
459
these residents (Emmett et al., 2006). The available data from this
the PFOA in the drinking water; the PFOA concentrations in the
525
460
study for Little Hocking was from 291 individuals whose primary
drinking water were reduced to concentrations below the detec-
526
461
water source was the public water system. Only a mean, median,
tion limit (LOD was 10 ng/Q. Hblzer et al. measured the serum
527
462
and interquartile range is available for the population's serum
PFOA concentrations again in July 2007 to quantify the changes
528
463
PFOA concentration; we were not able to obtain individual data.
in serum concentration (Holzer et al., 2009). 288 participants from
529
464
Also, it is unknown for how long the residents were exposed to
the original 2006 study were included in the 2007 study. In addi-
530
465
the contaminated water. No data was collected on serum PFOS
tion to Little Hocking, we used the human model to simulate the
531
466
concentrations in this study. Aside from Emmett's data, more data
serum PFOA concentrations in Arnsberg in both 2006 and 2007.
532
467
for the Little Hocking population is available on the Little Hocking
The half-lives of PFOA and PFOS are much longer in the human
533
468
Water Association (LHWA) website (LHWA, 2005), This data con-
than in laboratory animals; the half-lives in humans are reported
534
469
tains serum PFOA, PFOS, and several other perfluorinated chemical
to be on the order of several years (Andersen et al., 2008; Olsen
535
470
concentrations for some residents (N = 25) as of May 2005 in addi-
et al., 2007). Two studies have been performed to investigate the
536
471
tion to gender, age, and exposure duration. Both data sets were
half-life of PFOA, and the gender difference in elimination of PFOA
537
472
used to evaluate the predictive ability of the human model.
in the rat has not been observed in humans (Olsen et al., 2007). The
538
473
Another data set from the Little Hocking population used to test
first, reported by Olsen et al. (2007) examined the half-lives of sev-
539
474
the model was from the CS Health Project, which was conducted as
eral perfluorinated chemicals, including PFOA and PFOS, in 26 re-
540
475
part of a settlement of a class-action lawsuit (Leach v. E.I. du Pont
tired fluorochernical workers. The average serum half-life
541
476
de Nemours & Co., No. 01C-608 (W. Va., Wood County Cir. Ct. filed
estimated for PFOA was 3.8 years and that for PFOS was 5.4 years.
542
477
Apr. 10, 2002)) from August 2005-July 2006. The participants in
The second study was performed by Bartell et al. (2010) using the
543
478
the study were from Little Hocking and surrounding communities
pre- and post -filtration serum PFOA concentration of 200 residents
544
479
that were served with public drinking water, and the purpose of
of Little Hocking and the nearby town of Lubeck. The average ser-
545
480
the study was to assess the serum PFOA levels in the residents,
um half-life of PFOA estimated from their study was 2.3 years. We
546
481
The results of the CS Health Project were published by Steenland
considered both half-life estimates when running the model
547
482
et al. (2009), and the average serum PFOA concentration was
simulations.
548
483
-227 ppb, which is approximately half of what was measured by
PFOA and PFOS have been detected in the serum of the US gen-
549
484
Emmett et al. (2006) (mean -448 ppb). The Steenland et al. study
eral population; however, the chemicals are present in much lower
550
485
did not give a drinking water concentration of PFOA; thus, the
levels than what has been found in workers and people drinking
551
486
range from the Emmett study (1.5-7.2 ppb) was tested for simulat-
the contaminated water. Exposure pathways for the general popu-
552
487
ing the serum PFOA concentrations. The concentration chosen from
lation are not well characterized (Andersen et al., 2008; Calafat
553
488
the Emmett study to simulate the Steenland et al. data was 1.5 ppb
et al., 2007a; Paustenbach et al., 2007). Calafat et al. reported and
554
489
as this concentration gave the best fit to the data when a 3.8 year
compared the serum concentrations of several perfluorinated
555
490
half-life was assumed.
chemicals of the general population from 1999-2000 and 2003-
556
491
To reduce exposure to PFOA, activated carbon filters were in-
2004 National Health and Nutrition Examination Survey (NHANES)
557
492
stalled in Little Hocking and the surrounding communities in
(Calafat et al., 2007b). Due to reduction in production and manu-
558
493
2007. Pre- and post -filtration serum PFOA concentrations were
facturing emissions of these chemicals by 2002, the serum concen-
559
DEQ-CFW-00004212

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx—xxx
7
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
trations of PFOA and PFOS were found to have decreased in the
general population (Calafat et al., 2007b). We used the human
model to estimate the exposures of the general population to PFOA
and PFOS from both 1999-2000 and 2003-2004.
Olsen et al. also examined the decline in serum PFOA and PFOS
concentrations (Olsen et al., 2003a, 2008). In these studies, the ser-
um of Red Cross adult blood donors in six geographic locations
(100 donors from each location) around the US was analyzed for
PFOA, PFOS, and five other fluorochemicals. The first study was
conducted in 2000-2001; the follow-up study was performed in
2006. As with the NHANES participants, a decrease in the levels
of PFOA and PFOS was observed in the serum of the Red Cross do-
nors. The human model was used to estimate the exposure of all
Red Cross donors to PFOA and PFOS in 2000-2001 and in 2006.
2.6. Model structure - human
The PBPK model structure for PFOA and PFOS in the human is
the same as that for the monkey (Fig. 2); extrapolation to the hu-
man was the principal reason for development of the monkey
model. As mentioned in the methods section for the monkey mod-
el, the skin compartment was built in because dermal uptake is of
interest for the human, and the fat compartment serves as a
restriction volume; a larger fat compartment would result in a
smaller volume of distribution for these chemicals. Coding and
simulations for both the PFOA and PFOS human models was per-
formed in the Berkeley -Madonna program (Macey et al., 2000).
Model code can be found in the Supplementary Material.
resorption process that is hypothesized to be responsible for the
kinetics and long half-life in the monkey is also responsible for
the persistence in humans (Andersen et al., 2006). If this is con-
firmed and the transporter protein(s) that is/are responsible for re-
nal resorption of these chemicals is/are identified, information on
the activity of that transporter can then be used to identify suscep-
tible sub -populations due to polymorphic variation. The human
PBPK model thus includes the description of renal resorption that
we included in the monkey to allow for eventual incorporation of
data describing human population variability in renal transporter
activities. We assumed that the transporter protein would be the
same one in the monkey and the human; studies on the cloning
and function of renal transporters in the monkey suggest that the
monkey is a good predictor of renal uptake of organic anions in
the human (Tahara et al., 2005). Therefore, the transporter affinity
constant (Kt) is the same in the human as was estimated for the
monkey. We then adjusted the transporter capacity (Tm) for each
chemical in the human to get the correct plasma half-lives. Since
the clearance of both chemicals is so much slower in the human,
the transporter capacity is greater than in the monkeys.
2.7.2. Physiological parameters for the human model - PFOA and PFOS
The physiological parameters used for the human models can
be found in"rable 2. Body weight, cardiac output, fractional tissue
volumes, and fractional blood flows to tissue were the values re-
ported by Brown et al. (1997). Tissue volumes were scaled linearly
to body weight and blood flows were scaled as BW1_75. To adjust for
plasma flows to tissues tissue blood flows were multiplied by I-
hematocrit.
586
2.7. Model parameterization - human
587
2.7.1. Chemical -specific parameters for PFOA and PFOS
3. Results
588
Chemical -specific parameters used for the human PFOA model
589
are presented in Table 1. The tissue -plasma partition coefficients
3.1. Monkey model for PFOA
590
used for both chemicals were the same as those used for the mon-
591
key. Fiserova-Bergerova (1975) examined partition coefficients for
3.1.1. IV data
592
two compounds (Forane and methylene chloride) in several spe-
The simulated plasma and urine PFOA concentrations for a sin-
593
cies (man, monkey, dog, and rat). Tissue:blood partition coeffi-
gle IV dose of 10 mg/kg PFOA to the monkey were in good agree-
594
cients were very similar for the tissues examined across species;
ment with the experimental data (Fig. 3). There appears to be
595
thus, it is reasonable to use the same partition coefficients in both
some variability in the plasma clearance of PFOA for the monkeys;
596
the human and monkey models. For PFOA, the PCs were estimated
at about 60 days post dose, some monkeys have eliminated most of
597
from tissue concentration data reported by Kudo et al. (2007) for a
the chemical while some still have higher plasma concentrations.
598
single IV dose of PFOA to male Wistar rats. For PFOS, PCs were esti-
Adjustment of the transporter maximum (Tm) will affect how
599
mated from tissue concentration data from male C57131/6 mice
slowly or quickly the chemical is eliminated in the model. The
600
(DePierre, 2009). Only the free concentration of chemical is as-
I'm estimated to the average plasma concentration provided the
601
sumed to be available to partition into tissues. As in the monkey
best fit to the data.
602
and rat, PFOA and PFOS are highly bound to plasma proteins in
603
the human. The species differences in half-lives of these com-
3.1.2. Oral data
604
pounds does not appear to be due to differences in protein binding,
As mentioned in Section 2, only the data from the 10 mg/kg dai-
605
(Han et al., 2003) so we used the same free fraction of chemical
ly oral dose of PFOA to the monkeys was used for model develop-
606
(Free) estimated for the monkeys for PFOA and PFOS in the human
ment. The simulated plasma and urine concentrations are in good
607
models. Urinary elimination rate constants were scaled as BW-0-25
agreement with the observed data from the 10 mg/kg dose (Fig. 4).
608
and estimated by fitting the model to the data of Harada et al.
The model simulates the fast rise to steady state in the plasma and
609
(2005), where serum and urine PFOA and PFOS concentrations
the slower terminal half-life. The model was used to simulate the
610
were measured in residents in Kyoto City, Japan.
30/20 mg/kg daily dose, and the irregular dosing regimen reported
611
Similar to the monkey, humans have long half-lives for these
in the study was used. However, the observed experimental data
612
chemicals; however, instead of months as for the monkey, the hu-
for the 30/20 mg/kg dose produced different shapes of plasma
613
man plasma half-life is on the order of several years. Based on a
and urine concentration curves from the 10 mg/kg dose during
614
study of retired fluorochemical workers, the estimated arithmetic
the elimination phase. Tm could be increased to fit the high dose
615
mean serum elimination half-life in the human is approximately
data (data not shown); however, because the monkeys in the
616
3.8 years for PFOA (geometric mean was 3.5 years) and 5.4 years
high -dose group exhibited toxicity, the data observed may be
617
for PFOS (geometric mean was 4.8 years) (Olsen et al., 2007). A
due other effects, such as impaired renal function (decreasing the
618
more recent study conducted on a population that had previous
rate of flow to the filtrate compartment also allowed for fitting of
619
exposure to PFOA in drinking water estimated a plasma half-life
the high dose data). The lower dose is also more relevant for devel-
620
for PFOA of2.3 years (Bartell et al., 2010). The long half-lives across
opment of the human model, so we chose to use the parameters
621
species (monkey to human) suggests that the same saturable renal
that best fit the 10 mg/kg dose.
DEQ-CFW-00004213

8 A.E. Loccisano el al./Regulatory Toxicology and Phonnocology xxx (2010) xxx-xxx
100
0.01
0
20 40 60 80
days after dosing
100
0.01 4 --------------------------------------- 7 ------------------- I ------------------- I ------------------- 7 ------------------- 7 ------------------- I
0 5 0 15 20 25 30 35
days after dosing
Fig. 3. Comparison of model simulations (lines) with experimental data (points)
from Butenhoff et al. (2004). The plots contain PFOA plasma (top) and urine
(bottom) concentrations in monkeys administered a single 10 mg/kg iv dose of
PFOA and then allowed to recover. Females in triangles; males in squares.
100
0.01
0
10
0.01
0,001
50 '100 150 200 250 7300 350 400
thne (days of study)
0 50 100 150 200 250 300 350 400
time (days of study)
Fig. 4. Comparison of model simulations (lines) with experimental data (points)
from Butenhoff et al. (2004). The plots contain PFOA plasma (top) and urine
(bottom) concentrations in male monkeys administered daily gavage doses of
10 mg/kg PFOA for 26 weeks and were then allowed to recover.
682 3.2. Monkey model for PFOS
683 3.2.1. IV data
684 The simulated plasma and urine PFOS concentrations for a
685 single IV dose of 2 mg/kg PFOS to monkeys exhibits good
correspondence with the observed data of Noker and Gorman
(Fig. 5). Up to about 70 days post dose the monkeys exhibit some
variability in plasma elimination of PFOS, but at the last two time
points, the animals have similar plasma PFOS levels. Because PFOS
appears in the urine at a slower rate than it disappears from the
plasma, a storage compartment was added to the model. This "de-
lay" compartment was also used in the compartmental model of
Andersen et al. (2006) and Tan et al. (2008) and its use allows for
a better fit to the urine data. The biologically based model by Tan
et al. (2008) for PFOS in the monkey incorporated a time -depen-
dent description for the free fraction of chemical in the plasma
(this was necessary to describe the kinetics resulting from the IV
dose). This description was not necessary to implement in the
PBPK model in order to describe PFOS pharmacokinetics; thus,
the free fraction (Free) was kept constant.
3.2.2. Oral data
The simulated PFOS plasma concentrations resulting from an
oral dose of 0.03, 0.15, or 0.75 mg/kg daily to the monkeys are
shown in Fig. 6. Overall, the simulated concentrations are in good
agreement with the experimental data of Seacat et al. (2002). The
0.15 mg/kg dose lines up well with the experimental data while
the highest dose is slightly overestimated by the model, especially
during the dosing period. The model simulated a sharper rise in
plasma concentration and faster approach to steady state during
the dosing period that what was observed in the experimental
data. Adjustment of Tm and Kt allowed for better fitting of the high
dose curve; however, the best overall fit for all three doses was
with the Tm and Kt for the 0.15 mg/kg dose.
3.3. Sensitivity analysis
The normalized sensitivity coefficients for the model parame-
ters for PFOA and PFOS in the monkey with respect to serum PFOA
_J
E
-a
Z
W
0
LL
Q.
W
1 4
50 100 150 200
days after dosing
0 o'c"'...
0 20 do
60 80 100
days after dosing
Fig. 5. Comparison of model simulations (lines) with experimental data of Noker
and Gorman (2003). The plots contain PFOS plasma (top) and urine (bottom)
concentrations after a single iv dose of 2 mg/kg PFOS. Male and female monkeys
were administered an iv dose of PFOA and allowed to recover for 5 months.
Squares -males, triangles -females.
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
DEQ-CFW-00004214

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
9
350
300
250
_J
E
200
0
150
E
E
0
100
50
0
0 100 200 300 400 600 700
time (days)
Fig. G. PFOS in monkey plasma during and after oral closing for 6 months. Male and female monkeys were administered 0.03, 0.15, or 0.75 mg/kg PFOS orally for 6 months and
monitored for a vear after dosing stopped.
717 or PFOS AUC are shown in Fig. 7. Only parameters with sensitivity
718 coefficients greater than 0.1 are shown. For PFOA, sensitivity anal-
719 ysis was performed at 3 and 10 mg/kg daily doses. There do not ap-
720 pear to be any dose -dependent differences in model sensitivity to
0.5
.2
0
0
U
_0.5
IL
0.5
.2
0 0
W
] 0 * CC QRIC BW VUvC Htc Tmc Kt Free PLiv PTis
-2
Fig. 7. Calculated sensitivity coefficients for PBPK model parameters for PFOA (top)
and PFOS (bottom) in monkeys with respect to serum PFOA or PFOS AUC. Only
parameters with SCs > 0.1 are shown on the plots.
any of the parameters shown. Serum PFOA is primarily dependent
on the free fraction of chemical in the plasma and the parameters
governing elimination, which are the flow to the filtrate compart-
ment and the transporter maximum. Cardiac output and hemato-
crit will both influence the plasma flow to the filtrate
compartment since blood flows are dependent on cardiac output
and also must be adjusted for plasma flows because PFOA does
not partition into red blood cells. For PFOS, sensitivity analysis
was performed at 0.03 and 0.75 mg/kg daily doses. Unlike PFOA,
there appear to be some dose -dependent differences in model sen-
sitivity at the low and high doses of PFOS. At 0.03 mg/kg/day, ser-
um PFOS appears to be more dependent on liver volume,
liver:plasma partitioning, and rest of body:plasma partitioning
than at the higher dose. At 0.75 mg/kg, the free fraction of PFOS
in plasma and the elimination parameters (QFiIC, Tmc) appear to
be more important than at 0.03 mg/kg. At the higher dose, uptake
into tissues may be limited and clearance parameters show a
greater influence on serum PFOS.
3.4. Human model
3.4.1. Comparison with available data
For PFOA, we used the human model to simulate plasma con-
centrations of the chemical resulting from exposure to contami-
nated drinking water in both the Little Hocking, Ohio and
Arnsberg, Germany populations. The adequacy of the human PFOA
model and the corresponding parameters were assessed based on
whether the model was able to describe the available plasma con-
centration data on PFOA in humans, which exhibits high variabil-
ity. Overall, agreement between the experimental data and the
model simulations was good. Further efforts to refine the model
parameters will require more detailed kinetic studies in the
human. The exposure conditions (drinking water rate, concentra-
tion of PFOA on drinking water) used to simulate the different
populations are in Table 3.
3.4.1.1. Little Hocking, Ohio. Since it is not known how long the peo- 754
ple had been exposed (i.e., drinking and using the water), the mod- 755
el was run until the plasma PFOA concentration reached steady 756
state and until the model predictions were within the measured 757
DEQ-CFW-00004215

10 A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
Table 3
Exposure estimates used in simulations with human PBPK model for PFOA.
Parameters Values Sources
PFOA in water Mean: 3.55 ppb (Little Hocking)Mean: 0.519 ppb(Arnsberg) Emmett et al. (2006) and Holzer et al. (2008)
Intake Mean: 11 mL/kg/day(range:1-95) EPA-822-R-00-008
PFOA in serum (Little Hocking)(N = 291) Mean: 448 µg/LMedian: 374 µg/Llnterquartile range:221-576 µg(L Emmett et al. (2006)
PFOA in serum Mean: 28.5 µg/LSD: 12.9 Holzer et al. (2008)
(Arnsberg, Germany)(N = 101: men only)
mean and standard deviation. For the Little Hocking population,
the model had to be run for approximately 30 years before steady
state and the mean PFOA plasma concentration was reached. In
addition to malting sure the model reached steady state and the
mean/SD, we made sure that we had adjusted the transporter max-
imum (Tm) to yield the proper half-life. When simulating the Little
Hocking population, we assumed that their only exposure to PFOA
was through the drinking water as this was their primary exposure
source. Because there are two different plasma half-life estimates
that we found in the literature, (Bartell et al., 2010; Olsen et al.,
2007) the model was run assuming both half-lives. Fig. 8 shows
the plot of simulated plasma concentrations with the measured
data from Emmett's study. From the plot, it can be seen that either
half-life (3.8 versus 2.3 years) is possible; both yield simulated
concentrations that are within the mean and SD. The Little Hocking
Water Association (LHWA) website contains plasma PFOA concen-
tration data for 25 residents. The average and standard deviation
calculated from this data were slightly different that reported in
Emmett's 2006 study, so for comparison with the simulations, both
data sets (mean and standard deviations for LHWA website data
(triangles) and mean and interquartile range for Emmett data
(squares)) are shown in Fig. 8.
In addition to the data from Emmett et al. and the LHWA web
site, data collected from the participants in the C8 Health Project
was also simulated with the human model. The purpose of the
C8 Health Project was to assess the serum PFOA levels of the LH
and surrounding area residents that were served with water con-
taining PFOA. This data was collected from August -2005 July
2006 and was published by Steenland et al. (2009). As with the Em-
mett data, the model was run for 30 years of exposure to order to
r�tiir
600
500
o)400
Lo 300
a-
200
100
0
reach steady state, and both available estimates of the plasma half-
life for PFOA were tested (2.3 versus 3.8 years). Fig. 9 shows the
plot of the simulated plasma concentrations along with the mea-
sured data from both Emmett's study and the C8 Health Project
study. For this plot, a drinking water concentration of 3.55 ppb
was used. The mean serum PFOA concentration measured from
the LH residents participating, in the C8 Health Project is
—227 ppb, which is approximately half that measured in the Em-
mett et al. study (-448 ppb); however, the mean value from the
C8 Health Project overlaps with the lower end of the interquartile
range (221-576 ppb) measured in Emmett's study.
The serum half-life of PFOA is not the only factor that probably
has variability for this population. No drinking water concentra-
tions of PFOA were listed in the Steenland et al. study; however,
in the Emmett et al. study, the measured drinking water concentra-
tions ranged from 1.5 to 7.2 ppb (mean = 3.55 ppb). In Fig. 9, the
3.8 year half-life appeared to overestimate the serum PFOA con-
centrations measured from the C8 Health Project participants, so
the model was then run with a reduced drinking water concentra-
tion (1.5 ppb) but the half-life was kept at 3.8 years. Fig. 10 shows
the results of this (bottom) and also the model predictions plotted
with the Emmett et al. data (top) and the 3.55 ppb intake. Lowering
the drinking water concentration allowed for good predictions
with the 3.8 year half-life. The people in the Emmett cohort were
from LH only, which was the area with the highest water concen-
tration of PFOA, while the participants in the C8 Health Project
were from LH and surrounding communities where the water con-
centration of PFOA was lower; thus, it is reasonable that the people
from the C8 Health Project have lower serum PFOA concentrations.
In addition to variability in serum half-life and the concentration of
0 5 10 15 20 25 30 35 40 45 50
time exposed (years)
Fig. 8. Comparison of model simulations (lines) with experimental data from Emmett et al. (2006) (squares) and the Little Hocking Water Association website in 2005
(triangles). The plot contains PFOA plasma concentrations for the Little Hocking, OH population that was exposed to drinking water contaminated with PFOA (3.55 ppb). The
simulations were run for an exposure period of 30 years assuming two different half-lives: 3.8 years (red curve) from a study of occupational workers by Olsen et al. (2007),
and 2.3 years (blue curve) from a study by Bartell et al. (2010). Both half-lives appear to be probable.
788
789
790
791
792
793
794
795
796
797
798
799
Soo
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
DEQ-CFW 00004216

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
191
818
819
820
700
600
500
a 400
d
0 300
LL
200
100
0
0 ;> 10 15 20 25 30 35 40 45 50
time exposed (years)
Fig. 9. Comparison of model simulations (lines) with experimental data from Emmett et al. (2006) (squares) and Steenland et al. (2009) (triangles). The plot contains PFOA
plasma concentrations for the Little Hocking, OH population that was exposed to drinking water contaminated with PFOA. Both data sets were collected in 2005-2006. The
simulations were run for an exposure period of 30 years assuming a half-life of 3.8 years (top curve) and 2.3 years (bottom curve). The data from Steenland et at. (2009) is
from the participants in the C8 Health Panel. The mean serum concentration is —227 ppb, about half of that measured by Emmett et al. in 2006 (mean —448 ppb). A drinking
water concentration of 3.55 ppb PFOA was assumed.
700
600
500
400
0 300
u_
o.
200
100
0
350
300
250
u 200
CO 1;>0
o<
100
50
0
0 10 20 30 40 50
time exposed (years)
0 10 20 30 40 50
time exposed (years)
Fig. 10. Comparison of model simulations (tines) with experimental data from
Emmett et al. (2006) (top, squares) and Steenland et at. (2009) (bottom, triangles).
The plots contain PFOA plasma concentrations for the Little Hocking, OH population
that was exposed to drinking water contaminated with PFOA. Both data sets were
collected in 2005-2006. The simulations were run for an exposure period o
30 years and a serum half-life of 3.8 years was assumed. The water concentration of
PFOA for simulating the Emmett data sets was set at 3.55 ppb; however, for the
Steenland data, the drinking water concentration had to be reduced to 1.5 ppb if a
3.8 year half-life is used. The water concentrations in LH between 2002 and 2005
were found to range from 1.5 to 7.2 ppb (Emmett et at., 2006), so the people studied
by Steenland et al. simply may have been in a lower exposure region. Variable water
concentrations and/or half-lives of PFOA in individuals is likely.
PFOA in drinking water in various regions, drinking water rates and
how much exposure comes from drinking water versus other
sources most likely differ among the populations examined as well.
To reduce exposure to PFOA in the drinking water, an activated
carbon filtration system was installed in Little Hocking and the
nearby community of Lubeck in 2007 (Bartell et al., 2010). Pre -
and post filtration serum PFOA serum concentrations were mea-
suredby Bartell et al. in a cross-section of the populations, and
the participants were divided into public water or bottled water
users (Bartell et al., 2010). The model was used to simulate the ser-
um PFOA concentrations of the public water users in both Lubeck
and LH (Fig. 11). For both Lubeck and LH, both available serum
half-lives of PFOA were taken into consideration when running
the simulations. For LH public water users (top), exposure was as-
sumed to be 30 years (to reach steady-state) plus 6 months, as fil-
tration systems were installed 6 months after the pre -filtration
serum concentrations were measured. The pre -filtration water
concentration was assumed to be 3.55 ppb; when filtration started,
the exposure was simply shut off, as the PFOA water concentration
after the filtration system was installed was nonquantifiable or
nondetectable (Bartell et al., 2010). For Lubeck public water users
(bottom), exposure was assumed to be 30 years and then it was
shut off; filtration systems were installed in Lubeck right after
the pre -filtration serum concentrations were measured. PFOA in
drinking water samples in Lubeck ranged from 0.41 to 1 ppb be-
tween 2001 and 2006; 0.7 ppb (middle of this range) was used in
running the simulations. Due to the small population sizes used
in the Bartell et al. study, the ± SD from the mean for the post -fil-
tration serum PFOA concentrations are quite small. However, from
the pre -filtration concentrations, it appears from Fig. 11 that either
the 2.3 or 3.8 year half-life is possible. A larger population size
would be helpful in examining the extent of the variability in ser-
f um PFOA concentration.
3.4.1.2. Arnsberg, Germany. As with the people of Little Hocking, it
is not known how long the people of Arnsberg were exposed to
PFOA in their drinking water. It is believed that the contamination
occurred in the time period between 2000 and 2006; this popula-
tion was therefore not exposed as long as the people of Little Hock-
ing (Holzer et al., 2008). The concentration of PFOA in the drinking
water was approximately 7 times lower than that measured in Lit-
tle Hocking (0.519 ppb versus 3.55 ppb). For the Arnsberg popula-
tion, the plasma PFOA concentration reached measured mean
DEQ-CFW 00004217

12
A.E. Loccisano et al./Regulatory Toxicology and Phonuacology xxx (2010) xxx-xxx
860
861
862
863
800
700
600
< 500 -
0
400
E 300
�y 200 -
100
0
25
3.8 year haif-life
2.3 year nalf-life
spring 2007
n spring 2008 LH
.................................................................:::..>.
27
250
3.8 year half-life
200
2.3 year half-life
spring2007
spring2008
0 150
uv
Ei
E 100
-
50
0 1-
0
0
0
25 27 29 31 33 35 0 10 20 30 40 50
time exposed (years) time (years)
Fig. 11. Comparison of model simulations (lines) with experimental data from Fig.13. Comparison of model simulations (lines) with serum PFOA concentration of
Bartell et al. (2010) (points). The plots contain PFOA plasma concentrations for the general population (points). The plots contain the simulated serum PFOA concen-
Little Hocking (top) and Lubeck (bottom), OH populations that were exposed to trations using a half-life of 2.3 years (top) and 3.8 years (bottom). Estimated general
drinking water contaminated with PFOA. Serum PFOA was measured in spring 2007 population exposure was simulated as a direct input to blood. For a 2.3 year half -
(before filtration systems were installed; diamonds) and in spring 2008 (-6 months life, the estimated exposure range from 1999 to 2000 was 4.25-5.2e-4 ug(kg/day
after filtration started; squares). The simulations were run for an exposure period of and 3.3-3.9e-4 ug(kgjday for 2003-2004. For a 3.8 year half-life, the estimated
30 years assuming a half-life of 3.8 years (top curves) and 2.3 years (bottom curves). exposure range was 2.6-3.2e 4 ug(kgjday for 1999-2000 and 2.0-2.4e -4 ug(kg/day
Average drinking water concentration of PFOA in LH was 3.55 ppb and 0.7 ppb in for 2003-2004.
Lubeck The population size for this study was small (N = 132 for Lubeck; N = 40 for
LH); however, both half-lives appear to be likely.
installed by the waterworks. The drinking water concentration
within a few years (Fig. 12); however, the plasma concentration was lowered considerably by filtration (mostly below the LOD,
does not reach steady state in this time. Two studies were per which was 10 ng/L). The follow-up study was then performed in
formed on this population. The initial study was performed in July' July 2007 to examine the effects of the filtration on serum PFOA
2006, (Holzer et al., 2008) after which a filtration system was concentration (Holzer et al., 2009). The authors of the studies
10 20 30 40 50
time (years)
qr
30
25
ra
20
0
tL
a- 15
E
as
10
5
0
0 0.5 1 1,5 2 2.5 3 3.5 4 4.5 5
time Gears exposed)
Fig. 12. Comparison of model simulations (tines) with experimental data from Holzer et al. (2008) and Holzer et al. (2009) (points). The plot contains PFOA plasma
concentrations for a population in Arnsberg, Germany that was exposed to drinking water contaminated with PFOA (0.519 ppb). A filtration system was installed right after
the initial serum concentrations were measured (July 2006), and the serum PFOA concentrations were measured a year after this (July 2007). The exposure duration is
unknown, but it appears to be on order of a few years. The simulations pictured were run for two years assuming two different half-lives: 3.8 years (blue curve) from Olsen
et al. (2007) and 2.3 years (red curve) from Bartell et al. (2010). As for the Little Hocking population, both half-lives appear to be likely.
864
865
866
867
868
DEQ-CFW 00004218

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
13
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
measured the plasma concentration of PFOA in men, women, and
F5 10
children; however, since the means were similar for all groups
70
(28.5 ppb for men, 26.7 ppb for women, and 24.6 ppb for kids in
60
the initial study; 23.4 ppb for men, 18.8 ppb for women, and 17.4
for kids in the follow-up study), only the simulations for the men
are shown in Fig. 12. As with the Little Hocking simulations, we as-
40
0
surned that drinking water was the only source of exposure to
LL 30
CL
PFOA and the simulations were run using both a 3.8 and 2.3 year
20
half-life. From Fig. 12, it can be seen that either half-life fits with
10
the data (although the 3.8 year half-life yields a slightly better fit).
3.4.1.3. General population. PFOA and PFOS have also been mea-
0 10 20 30 40 50
time exposed (years)
sured in the blood of the general population (Calafat et al.,
2007a; Calafat et al., 2007b). These levels are much lower than
40
b
those found in the serum of the residents of Little Hocking and
35
Arnsberg; however, the exposure sources are unknown and could
2 31)
1999-2000 ....................
potentially be from a number of sources, including drinking water,
25
food and food packaging, consumer products, and in the air. Serum
0
20
LL
2003-2004
concentrations of PFOA and PFOS from the general population are
available from 1999-2000 to 2003-2004 from the National Health
E 5
and Nutrition Examination Survey (NHANES). We used the model
10
to simulate the serum concentrations and to estimate how much
5
PFOA and PFOS the general population might be exposed to daily.
0
----------------------- T ----------------------------------- I ----------------------------------- I ----------------------------------- 1
Figs. 13 and 14b show the comparison of the model simulations
0 20 40 60 80
with the measured concentrations reported by Calafat et al.
time exposed (years)
(2007b). The top plot in Fig. 13 shows the simulation for PFOA
assuming a 2.3 year half-life and the bottom plot assumes a
Fig. 14. (a) Comparison of model simulations (lines) of serum PFOS concentration
3.8 year half-life. Fig. 14b shows the simulation for PFOS assuming
with Little Hocking Water Association website data (squares). The data from the
webSiLe is the average and SO of the plasma concentrations of residents in Little
a 5.4 year average serum elimination half-life (Olsen et al., 2007).
Hocking (N = 25). The concentration of PFOS in the drinking water was not reported
Because the exposure sources of PFOA and PFOS are unknown,
(or unknown) at this time. As for PFOA, the simulation for PFOS was run to reach
the simulations were performed with chemical exposure being a
steady state (30 years) and to reach the average concentration measured in plasma
direct input to blood (effectively an IV dose). The major US manu-
(the concentration of PFOS in the drinking water was estimated). A half-life of
facturer of PFOS and compounds that can degrade to PFOS ceased
5.4 years (from occupational data of Olsen et al.) was assumed for the simulation.
(b) Comparison of model simulations (lines) with serum PFOS concentration of
all production by the end of 2002, and, while PFOA continues to
general population (points). The plots contain the simulated serum PFOS concen-
be produced and used in the US, emissions have been reduced. This
trations using a half-life of 5.4 years. Estimated general population exposure was
is apparently reflected in the serum concentrations of both PFOA
simulated as a direct input to blood. The estimated exposure range was 2.3-
3.Oe -' ug/kg/day for 1999-2000 and 1.6- 1.9e -' ug/kg/day for 2003-2004.
and PFOS: the geometric mean of the serum PFOA concentration
decreased from 5.2 ppb in 1999-2000 to 3.1 ppb in 2003-2004
and from 30.4 to 20.7 for PFOS. As far as exposure estimates go,
if a 2.3 year half-life is assumed for PFOA the exposure range is
chemicals was assumed to be constant (the simulation was run
4.25-5.2e 4 ug/kg/day in 1999-2000 and 3.0-3.9e -A ug/kg/day in
so that concentrations reached steady state, but exposure was
2003-2004. If a 3.8 year half-life is assumed, the exposure range
not stopped; Figs. 15 and 16, top panels). For PFOS, the half-life
is 2.6-3.2e --4 ug/kg/day in 1999-2000 and 2.0-2.15e -4 ug/k,g/day
was assumed to be that estimated from retired fluorochernical
in 2003-2004. For PFOS, the exposure range is 2.3-3.0e-ug/kg/
workers (arithmetic mean of 5.4 years) by Olsen et al., for PFOA,
day in 1999-2000 and 1.6-1.9e --3 ug/kg/day in 2003-2004. These
both the 3.8 year arithmetic mean half-life estimated by Olsen et
ranges were estimated using the 95% confidence intervals for the
al., and 2.3 year half-life estimated by Bartell et al. were considered
geometric means reported by Calafat et al. (2007b). For PFOA, if
(Bartell et al., 2010; Olsen et al., 2007). The estimated daily expo -
we compare these estimates to what the average person in Little
sure to PFOS in 2000-2001 was 2.2-4.5e -3 ug/kg BW; in 2006, it
Hocking ingested daily in 2006 (assuming consumption of 11 mL
was estimated between 8.5e --4 and 1.8e --3 ug/kg BW. The esti-
water/kg BW/day and 3.55 ppb PFOA in drinking water), the gen-
mated daily exposures for PFOA in 2000-2001 using the model
eral population exposure is approximately 195 times less than that
are 3.2-6.Oe -4 u,u/kg BW (2.3yr half-life) and 1.9-3.7e -4 u,u/kg
of a person in Little Hocking. For PFOS (if the estimate of 0.34 ppb
BW (3.8 yr half-life); the estimated daily exposures in 2006 are
PFOS in drinking water in LH is used; see section on PBPK human
1.3-2.65e -4 (2.3 year half-life) and 2.2-4.4e -4 (3.8 year half-life).
model for PFOS), if the general population estimates are compared
These estimates were made using the 95% confidence intervals
to that of LH (assuming consumption of 11 mL/kg BW/day of
for the geometric means and also the interquartile ranges reported
water), the general population exposure is approximately 2.3 times
by Olsen et al. (2003a, 2008) and are close to the values estimated
less than for the people in LH.
for the participants in the NHANES study, which ranged from
In addition to the general population data from the NHANES, Ol-
1.6e --3 to 3.Oe --3 ug/kg/day for PFOS and 3.0-5.2e --4 ug/kg/day for
sen et al. reported the serum PFOA and PFOS concentrations in
PFOA. In the second set of simulations, the concentrations were al -
adult Red Cross blood donors from 2000-2001 to 2006 in six geo-
lowed to reach steady state and then exposure was shut off two
graphic locations around the US (N = -600 in both studies; -100
years after the initial measurements were taken. We assumed that
samples from each clinic) (Olsen et al., 2003a, 2008). We used
steady state was reached when the initial measurements were ta-
the model to simulate the serum concentrations of PFOA and PFOS
ken (2000-2001); 3 M completely ceased production of PFOS-pro-
at both time periods in the donors and also the decline in serum
ducing chemicals in 2002. For both PFOA and PFOS, the exposure
concentrations of these chemicals after the phase -out of PFOS-pro-
was estimated for all donors in the study in both 2000 and 2006.
ducing chemicals. In the first set of simulations, exposure to both
The estimated exposure in 2000 was then used to examine the de-
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
DEQ-CFW-00004219

14 A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
60
50
ra
CL
a. 40
93
0 -0
n.
20
U1
Ij
0
M
70
� 60
en
50
u- 40
CL
E 30
En
20
10
0
10 20 30 40 50 60 70 80
years
80
70®rn0dEll
de 60 <,
Ilage.r c; en
-' SC
0 Boston
u. 40
CL
E 30 <;
2
20
10
0 ---------------------------------------------
0 20 40 60 80 0 20 40 60 80
years years
Fig. 15. Comparison of model simulations (lines) with serum PFOS concentration of Red Cross adult blood donors in 2000 and 2006 (points). A serum half-life of 5.4 years for
PFOS was assumed. The top panel shows the serum PFOS concentrations from donors in all six locations examined, assuming constant exposure in both 2000 (estimated
exposure was 0.0022-0.0045 ug/kg) and 2006 (estimated) exposure was 0.00085-0.0118 ug/kg). The bottom panel shows the serum PFOS concentrations with exposure shut
off 2 years after the initial measurements were taken to simulate the phase -out of PFOS-producing chemicals. The points are the serum PFOS concentrations measured in each
of the six cities. The exposure estimated for all donors and the 5.4-year half-life yield good agreement for the individual cities.
r 7
2000-2001
5 2flflfl-2001
4 4
LL
9L 2006
a -
::::.
2006
J E
2 2
" 3.8 yr half-life � 2.3 yr half-life
n C
0 19 20 39 40 59 0 10 20 30 40 50
years years
6 12
5 10 - model 3.8 yr. half-life
Charlotte
e
� 4 Portland
c( Boston
—rru i-3.8 yr half- �y
av 3 ife u. 0 rnodel 2.3 yr. half-life
CL LA M �
E E
2 4 .
Portland ro
1 hf Eapolis 2
-----------
0 1ti 20 39 40 59 0 10 20 30 40 50
years years
Fig. 16. Comparison of model simulations (lines) with serum PFOA concentration of Red Cross adult blood donors in 2000 and 2006 (points). A serum half-life of 3.8 and
2.3 years for PFOA was assumed. The top panel shows the serum PFOA concentrations from donors in all six locations examined, assuming constant exposure in both 2000
and 2006. The bottom panel shows the serum PFOS concentrations with exposure shut off 2 years after the initial measurements were taken to simulate the phase -out of
perfluoroalkyl compounds. The points are the serum PFOA concentrations measured in each of the six cities. The exposure estimates for all donors and both half-lives yield
good agreement for the individual cities.
961 cline in serum concentration that was observed in 2006 for each of Although the concentration ranges for the donors in each city 963
962 the six cities in the study (Figs. 15 and 16, bottom panels). may vary, the estimated exposure for all donors in 2000 appears 964
DEQ-CFW 00004220

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx 15
965
to yield good model predictions for each city and for the decrease
3.4.1.5. Predicted serum PFOA half-lives. Only two studies have been
966
in concentration observed in 2006.
conducted that have estimated the half-life of PFOA in humans
(Bartell et al., 2010; Olsen et al., 2007). The average half-life esti-
mated from the 26 retired fluorochernical workers (3.8 years)
967
3.4.1.4. PBPK human model for PFOS. Data in the human is extremely
was longer compared to the environmentally exposed populations
968
limited for PFOS. The largest US manufacturer of PFOS completed its
in Lubeck and Little Hocking (2.3 years). From running the human
969
phaseout of the chemical between 2000 and 2002 (Butenhoff et al.,
model for the Little Hocking and Arnsberg populations and taking
970
2006). As for PFOA, no definite association between exposure to
both half-lives into consideration, it was observed that both esti-
971
PFOS and adverse health effects in humans has been established.
mated half-lives appear likely; both allowed for the predicted PFOA
972
Serum PFOS concentrations have been measured in residents from
serum concentrations to fall within the measured means and stan-
973
Little Hocking; however, PFOS was present in levels similar to those
dard deviations (Figs. 8-12). In addition to running the model to
974
measured from the general US population (Frisbee et al., 2009;
compare the simulations with available data, we used the model
975
LHWA, 2005). The study conducted by Olsen et al. (2007) with re-
to predict the serum half-lives of the 25 individuals for whom data
976
tired fluorochernical workers estimated an average half-life of
was provided on the Little Hocking Water Association website. The
977
5.4 years for PFOS, and the LHWA website included serum PFOS con-
data on the individuals and their serum PFOA concentration is gi-
978
centrations of 25 individuals. Using this data, we were able to esti-
ven in Table 4. For simplicity, the drinking water concentration
979
mate a transporter maximum for the human (Tmc = 3.5) that
was assumed to be 3.55 ppb and the drinking water rate was set
980
would yield the estimated 5.4 year half-life and run the model to
at 11 mL/kg BW/day for each person unless noted in the table
981
reach the measured mean serum PFOS concentration in the Little
(USEPA, 1997). For men, a standard BW of 70 kg was used; for wo-
982
Hocking people. Fig. 14a shows the comparison of the simulated ser-
men, a BW of 58 kg was used in the model. To predict the serum
983
um PFOS concentration compared to the measured average and
half-life for each person in the table, the exposure period was mod-
984
standard deviations from the LHWA website data. No PFOS concen-
ified to reflect how long they had been drinking the water (the ser-
985
tration in the drinking water could be found for Little Hocking;
um concentration given in the table was assumed to be the
986
therefore, we had to estimate this to achieve the correct serum con-
maximum serum concentration and the end -of -exposure concen-
987
centration. The mean estimated concentration was 0.34 ppb, which
tration). To get the correct serum PFOA concentration for each indi-
988
is ten times less than the PFOA concentration in the drinking water.
vidual, the transporter maximum Jrn) was adjusted. The predicted
989
The water levels are most likely lower because the nearby produc-
half-lives can be found in Table 4. The mean half-life of the 25
990
tion plant did not produce PFOS. As for PFOA, it took about 30 years
people is approximately 2.6 years, and they range from 0.8 to
991
of exposure to reach steady state and it was assumed that drinking
8 years. There were two individuals in the table for whom the max-
992
water was the only source of PFOS exposure. Although the existing
imum serum PFOA concentration could not be reached simply by
993
data for human PK for PFOS is scarce, we were able to make a human
adjusting Tm. In order to reach their serum concentrations in the
994
model for this chemical that works reasonably well. Further refine-
time period that they were exposed, their drinking water intake
995
ment of the model will require that more data for PFOS in humans
rates had to be increased from 11 mL/kg BW/day to 29 and
996
becomes available.
22 mL/kg BW/day. These increased drinking water rates are not
Table 4
Model prediction of serum PFOA half-lives in the Little Hocking population.
LHWA
Sex
May-05
Age
ppb
PFOA
Years drinking
H2O
Tmc (PFOA)
PFOA
Half-life (years)
F
46-55
281
28
5.5
2.04
M
46-55
395
51
7.5
2.89
F
46-55
248
55
4.9
1.82
F
26-35
228
27
4.5
1.67
M
<15
629
3
10
3.7 a (29 mL/kg BW/day)
F
>65
442
35
8.7
3.23
M
36-45
220
35
4.1
1.61
F
36-45
112
9
2.2
0.82
M
56-65
532
34
10
3.87
F
26-35
116
4
2.3
0.86
F
26-35
211
6
4.3
1.61
M
<15
268
4
7.1
2.93
F
<15
436
3
10
3.81' (22 mL,'kg BW/day)
M
26-35
216
4
4.7
1.86
M
36-45
325
25
6.2
2.4
F
46-55
475
33
9.25
3.44
M
46-55
432
33
8.2
3.15
F
>65
358
25
7
2.64
M
>65
346
20
6.5
2.52
M
36-45
1040
37
20
7.71
F
36-45
240
35
4.7
1.75
M
56-65
204
16
3.8
1.46
M
26-35
176
24
3.3
1.28
F
16-25
488
16
9.7
3.59
M
>65
238
34
4.5
1.74
Mean
346.24
23.84
2.576
SD
196.350087
Median
281
a Indicates that serum PFOA concentration could not be attained by adjusting only Tmc. Drinking water rate had to be increased from 11 mL/
kg BW/day to the amount indicated. Drinking water concentration was set at 3.55 ppb for each individual.
DEQ-CFW-00004221

16 A.E. Loccisano el al./Regulatory Toxicology and Phonnocology xxx (2010) xxx-xxx
unreasonable; these individuals may have had higher activity lev-
els (USEPA, 1997), simply drank more water, or other physiological
factors could be involved.
The estimated half-lives from retired workers and the environ-
mentally exposed population obviously fit into the predicted range
of half-lives from the people on the LHWA website (3.8 and
2.3 years). The retired workers in the study were no longer occupa-
tionally exposed, but their serum concentrations were higher than
the general population (Calafat et al., 2007b; Olsen et al., 2007).
Fig. 17 compares the distributions of the predicted half-lives from
Little Hocking with those calculated for the retired workers. The
workers' average initial serum concentration was higher than that
of the average concentration in Little Hocking (691 ppb for workers
versus 346 ppb for LH), but the half-lives for both groups are com-
parable. Little Hocking has a higher frequency of shorter half-lives
(in the 1-2 year range) while the workers have a higher frequency
of half-lives in the 3-4 year range. Neither group has many individ-
uals with a half-life longer than 4 years.
Prediction of individual half-lives gives an idea of the variability
in how long different people retain or eliminate PFOA. From the
half-lives predicted by the model, the half-life does not appear to
be dependent on how long the people were exposed to PFOA in
the drinking water (Fig. 18, top). There is no correlation between
length of exposure and half-life; the person with the longest
half-life (8 years) was drinking the water for 37 years, but people
who drank the water for over 50 years had shorter half-lives (2-
3 years). The same was found for the fluorochernical workers;
there does not appear to be a correlation between how long they
worked in the plant and their half-life of PFOA (Fig. 18, bottom).
If renal resorption is the main process responsible for the long
half-lives observed in the human, the variability in elimination of
PFOA could be due to polymorphisms in renal transporters or var-
iability in expression of the transporter(s). Once the transporter
responsible for retention of PFOA in human is identified, informa-
tion on the activity and expression of the transporter can then be
used to determine if there are potentially sensitive individuals.
4. Discussion
Previously, a biologically motivated compartmental model was
developed that was able to describe the PK of PFOA and PFOS in the
cynomolgus monkey. However, to extrapolate the model to a hu-
man to aid in risk assessment, a physiologically, based model with
realistic tissue volumes, blood flows, and biochemical properties of
PFOA and PFOS were needed. We have therefore developed a PBPK
model for PFOA and PFOS in the cynomolgus monkey, and the
model is able to describe the available pharmacokinetic data in
monkeys exposed to these chemicals via IV and oral dosing.
12 -
workers
10
LHWA
B-
ea .....
..... .....
...... ..... ..... .....
...... ..... ..... .....
...... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
...... ..... ..... ..... ..... .....
..... ..... ..... .....
...... ..... ..... ..... .....
...... ..... ..... ..... .....
...... ..... ..... ..... .....
...... ..... ..... ..... .....
...... ..... ..... ..... .....
...... ..... .....
..... .....
...... .....
2 . ..... ......
...... ..... .....
..... ..... .....
..... .....
..... ..... .....
..... ..... .....
..... ..... .....
..... ..... .....
..... .....
...... ..... .....
..... .....
...... ..... .....
..... .....
...... ..... .....
..... .....
...... ..... .....
..... .....
...... ..... .....
..... .....
...... ..... .....
..... .....
...... ..... .....
..... .....
..... .....
..... .....
0
1 2 3 4 5 7 8 10
half-life (years)
Fig. 17. Distribution of half-lives for LH residents and retired fluorochemical
workers.
41
7 -1
6
Z' 5 Al
41
,5 3
2
0 ------------------------------------------------ 7 ------------------------ r ----------------------- - ------------------------------------------------ 1
0 10 20 750 40 50 60
years drinking H2O
10
5
4
3
0 41
2 41
a.
J
0
19 24 29 734 39
years worked
Fig. 18. Half-lives of residents of LH vs. exposure period (top) and retired workers
(bottom). There is no correlation between their time drinking the water or working
with fluoropolymers and their half-lives of PFOA
The monkey model was extrapolated to the human based on the
hypothesis that the long -half lives in humans for PFOA and PFOS,
like the monkey, can be attributed to saturable renal resorption
in the proximal tubule of the kidney. We used the human model
to successfully describe the available PFOA data from two commu-
nities exposed to PFOA in their tap water and also the general pop-
ulation. In addition, we used the model to predict the serum half-
lives of 25 residents of the Little Hocking community.
Obviously, just because models that incorporate saturable renal
resorption are able to describe the existing body of kinetic infor-
mation does not mean that this process controls the serum half-
lives of these chemicals in the animal or that the same process is
responsible in the human. However, existing data for renal trans-
porters in the rat suggest a tubular resorptive role for at least
one transporter (Oatplal), and we were able to achieve good
agreement between the observed data experimental data and the
model predictions for both monkeys and humans. In addition, we
have developed a PBPK model for these compounds in the rat using
the renal resorption description, and this model successfully de-
scribes kinetics in the rat (manuscript in preparation). This is evi-
dence that the saturable resorption process is responsible for the
kinetics of these chemicals across species. If this process is con-
firmed to the be the process responsible for the long half life in hu-
mans and the transporter proteins are identified, information on
the activity of the transporters can be used to identify susceptible
individuals due to polymorphic variation.
A recent study by Yang et al. showed that two human renal or-
ganic anion transporters on the apical membrane side of proximal
tubule cells are capable of transporting PFCs and thus may play a
key role in tubular resorption of these compounds (Yang et al.,
2010). This further suggests that renal resorption is at least partly
responsible for the long plasma half-life of these chemicals. Two
other human OATs, OAT1 and OAT3, which are located on the baso-
lateral membrane of proximal tubule cells, can also transport PFCs
(tubular secretion) (Weaver et al., 2009); OAT1 and OAT3 thus ex-
crete the chemicals from the blood to the urine side, while OAT4
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
DEQ-CFW-00004222

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx—xxx
17
1113
1114
1115
1116
1117
ills
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
and URAT1 bring the chemicals back into the cell. In the rat, OAT1
and OAT3 have been shown to secrete PFCs; the higher expression
of OATplal (apical side) in the male rat is related to the longer
half-life observed in the male rat. The human orthologs of OAT1
and OAT3 have similar affinities for PFCs and due to the high
expression levels of these transporters in the human, tubular
secretion is expected to be similar to that in the rat (Nakagawa
et al., 2008) (Hilgenclorf et al., 2007; Nakagawa et al., 2009). How-
ever, because PFCs do not accumulate in the kidney, Yang, et al.
hypothesize that there may be an efflux transporter localized to
the basolateral membrane that could aid in transporting PFCs back
into systemic ci rcu lation. (Yang et al., 2010) If identified, this efflux
transporter(s) would, in addition to the apical transporters respon-
sible for tubular resorption, probably contribute to extending the
half-life of PFCs in humans.
4.1. Monkey PBPK models
In the oral dosing study with PFOA, it was observed that there is
a rapid rise to steady state plasma concentrations of PFOA with a
much slower terminal half-life (Butenhoff et al., 2002, 2004a). In
the studies with PFOS, a similar trend is observed, and the terminal
half-life is even slower (Seacat et al., 2002). In addition, the differ-
ences observed in elimination with increasing dose are indicative
that a capacity -limited, saturable process must be involved in the
pharmacokinetic behavior of these compounds (Andersen et al.,
2006; "ran et al., 2008). It is hypothesized that saturable renal
resorption is the process responsible for the kinetics and long
half-lives of these chemicals; thus, this description was incorpo-
rated into both the compartmental model of Andersen and Tan
and the current PBPK model for PFOA and PFOS. This process was
described with a transporter maximum (Tm) and transporter affin-
ity constant (Kt) where chemical was reabsorbed from the filtrate
compartment back to the kidney into the plasma. The longer half-
life of PFOS compared to PFOA was accounted for by using a higher
capacity and higher affinity for PFOS in the model. The biologically
motivated model required a time -dependent feature to better
characterize the kinetics of these chemicals, which was a changing
free fraction of chemical in the plasma. This feature was not neces-
sary for description of kinetics in the PBPK model; a constant free
fraction (Free) was sufficient. Because both PFOA and PFOS are
highly bound to plasma proteins (>97%), a free fraction of <0.03
(0.02 for PFOA and 0.025 for PFOS) was estimated to fit the plasma
time course data for the PBPK model. The monkey models success-
fully describe the available PK data, and further improvement to
the monkey PBPK model will rely on additional data. information
on plasma binding proteins in the monkey (dissociation constants,
number of binding sites for PFOA and PFOS) and elucidation and
information (any sex differences in expression, transporter capac-
ities and affinity constants) on renal transporters responsible for
tubular resorption in the monkey would aid in refinement of the
model.
The goal of developing a PBPK model for the monkey was to
extrapolate to the human to in order to enhance the scientific basis
for human health risk assessment of PFOA and PFOS. Due to the
physiological structure of the monkey model, it was possible and
easy to extrapolate the model to the human. Epidemiological data
suggest that the half-life of PFOA and PFOS is on the order of years
in the human in contrast to months in the monkey. If renal resorp-
tion is the main process responsible for the long half-life of these
compounds in both species, the slower clearance in the human
can now be easily described in the model by a higher transporter
capacity. In addition to extrapolation to the human, the PBPK mod-
el for the monkey could be extended to describe the kinetic data
available for other perfluoroalkyl acids.
4.2. Human PBPK models
Available human data for PFOA and PFOS suggest that humans
are slow eliminators of these chemicals. It is hypothesized that
these long half-lives are due to saturable renal resorption as they
are in the monkey and rodent. The human PBPK model thus incor-
porated saturable renal resorption in the kidney and exhibits good
agreement with available data for PFOA in Little Hocking and Arns-
berg residents and also in the general population. Human data on
PFOS is even sparser than that for PFOA, but using the half-life esti-
mate from retired workers and the data available on serum PFOS
concentrations from Little Hocking, the model shows reasonable
agreement with the data. Plasma protein binding is another possi-
ble cause of the long plasma half-lives in humans. PFOA and PFOS
have been found to be >90% bound to plasma proteins in rats, mon-
keys, and humans; however, binding is not well -characterized
(Han et al., 2003; Jones et al., 2003; Kerstner-Wood et al., 2003).
For example, it may be difficult to distinguish between 90% and
99% bound, but if humans have a higher bound fraction, this would
mean a 10-fold difference in the free fraction and therefore a
slower clearance and longer half-life would be possible. Further
refinement will depend on future studies of both renal transporters
and plasma protein binding.
Even though the model exhibits good predictive ability, we had
only means, medians, and standard deviations of human serum
measurements at our disposal. We were not able to obtain the data
on the individuals in the populations in the studies, so we were not
able to incorporate factors that may be important to determining
variability in serum concentration, plasma half-life, or susceptibil-
ity to toxic effects. These factors would include, but may not be
limited to: gender, age, how long each person had been drinking
the water, how much tap water each person drinks daily, the PFOA
concentration measured in different wells or water supplies, and
any preexisting medical conditions, such as compromised renal
function. The more data we can obtain on individuals in exposed
populations, the better we can characterize population variability
in a risk assessment context. Several years ago, residents from Lit-
tle Hocking and the surrounding communities of the fluoropoly-
mer production plant filed a class-action lawsuit alleging adverse
health effects due to consumption of water containing PFOA. As
part of the settlement, the CS Health Project, which is an indepen-
dent panel of researchers investigating the potential link between
PFOA exposure and adverse health effects, was created. We believe
that using this information when it becomes available to us would
greatly improve the value of the model for describing the variabil-
ity in the human p harm acoki netics of PFOA.
When running simulations, we took into consideration both
serum elimination half-life estimates available in the literature
for PFOA. From the model predictions, both half-lives appear to
be equally likely. Both populations used to estimate the half-lives
were small (26 for workers and 200 for environmental exposure);
however, the 26 workers exhibited half-lives from 1.5 to 9 years
(the environmental study did not give the individuals' half-lives).
From the half-lives that the model predicted from the 25 residents
in Little Hocking, the half-lives range from 0.9 to 8 years. Even in
small populations a wide spectrum is observed, indicating that
there is probably a lot of variability in how individuals retain, han-
dle, and eliminate PFOA. The average half-life from the model pre-
dictions was 2.7 years, which is between that measured from the
workers (3.8 years) and the environmentally exposed people
(2.3 years).
Future refinement of the human model will depend on addi-
tional human studies and information on activity and variation in
renal transporters. Use of the human PBPK model along with indi-
vidual data and information on exposure routes and sources of
PFAAs will enable a better understanding of pharmacokinetics of
DEQ-CFW-00004223

is A.E. Loccisano el al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
1241 PFAAs in humans, link possible exposure routes and external doses
1242 to blood concentrations, aid in understanding interindividual vari-
1243 ability, and ultimately aid in human health risk assessment for
1244 PFAAs.
1245 5. Conflict of interest statement
1246 The authors declare that there are no conflicts of interest.
1247
1248
1249
1250
1251
1252
1253
1254
1255
Acknowledgments
We would like to thank Dr. John Butenhoff at 3M Company and
Mr. Gerald Kennedy, Jr. at DuPont Company for helpful discussion
and comments. Financial support was supported by an EPA STAR
Grant (No. R833450) and the 3M and DuPont Companies.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.yrtph.2010.12.004.
References
3M, 2003. Environmental and Health Assessment of Perfluorooctane Sulfonate and
its Salts. USEPA Administrative Record AR 226-1486.
Andersen, M.E., Butenhoff, J.L., Chang, 3­C_ Farrar, D.G., Kennedy Jr., ".L., Lau, C.,
Olsen, G.W., Seed, J., Wallace, K.B., 2008. Perfluoroalkyl acids and related
chemistries-toxicokinetics and modes of action. Toxicol. Sci. 102, 3-14.
Andersen, M.E., Clewell, H.J., Tan, Y.-M., Butenhoff, J.L., Olsen, G., 2006.
Pliarmacokinetic modeling of saturable, renal absorption of
perfluoroalkylacids in monkeys -probing the determinants of long plasma
half-lives. Toxicology 227, 156-164.
Apelberg, B.J., Witter, F.R., Herbstman, J.B., Calafat, A.M., Halclen, R.U., Needham, L.L.,
Goldman, L.R., 2007. Cord serum concentrations of perfluorooctane sulfonate
(pfos) and perfluorooctanoate (pfoa) in relation to weight and size at birth.
Environ. Health Perspect. 115, 1670-1676.
Ashby, J., Brady, A., Elcombe, C.R., Elliott, B.M., Ishmael, J., Odum, J., Tugwood, J.D_
Kettle, S., Purchase, I.F.H., 1994. Mechanistically -based human hazard
assessment of peroxisome proliferator-induced hepatocarcinogenesis, Hum.
Exp. Toxicol. 13 (Suppl. 2), S1 -Sl 17.
Bartell, S.M., Calafat, A.M., Lyu, C., Kato, K., Ryan, P.B., Steenland, K, 2010, Rate of
decline in serum pfoa concentrations after granular activated carbon filtration
at two public water systems in Ohio and West Virginia. Eaviron- Health
PerspeCL. 118, 222-228.
Benskin, J.P., De Silva, A.C., Martin, L.J., Arsenault, G., McCrindle, R., Riddell, N.,
Mabury, S.A., Martin, J.W., 2009. Disposition of perfluorinated acid isomers in
sprague-dawley rats: part 1: single dose. Environ. Toxicol. Chem. 28, 542-554.
Biegel, L.B., Hurtt, M.E., Frame, S.R., O'Connor, J.C., Cook, J.C., 2001, Mechanisms of
extrahepatic tumor induction by peroxisome proliferators in male cd rats.
Toxicol. Sci. 60, 44-55.
Bjork,J.A., Wallace, K.B., 2009. Structure -activity relationships and human relevance
for perfluoroalkyl acid -induced transcriptional activation of peroxisome
proliferation in liver cell cultures. Toxicol. Sci. 111, 89-9.9.
Brown, R.P., Delp, M.D., Lindstedt, M.D., Rhomberg, LP,, Beliles, R.P., 1997.
Physiological parameter values for physiologically based pharmacokmetic
models. Toxicol. Ind. Health 13, 407-484.
Butenhoff, J.L, 2010. Personal communication.
Butenhoff, J.L., Costa, G., Elcombe, C., Farrar, E., Hansen, K., Iwai, H., Jung, R., Kennedy
Jr., G., Lieder, P., Olsen, G., Thomford, P., 2002. Toxicity of ammonium
perfluorooctanoate in male cynomolgus monkeys after oral dosing for 6
months. Toxicol. Sci. 69, 244-257.
Butenhoff, J.L., Kennedy Jr., G., Hinderliter, P.M., Lieder, P., Jung, R., Hansen, K,
Gorman, G.S., Noker, P.E., Thornford, P., 2004a. Pharmacokinetics of
perfluorooctanoate in cynomolgus monkeys. Toxicol. Sci. 82, 394-406.
Butenhoff, J.L., Kennedy Jr., G.L., Frame, S.R., O'Conner, J.C., York, R.G., 2004b. The
reproductive toxicology of ammonium perfluorooctanoate (apfo) in the rat.
Toxicology 196, 95-116.
Butenhoff, J.L, Olsen, G.W., Pfahles-Hutchens, A., 2006. The applicability of
biomonitoring data for perflurooctanesulfonate to the environmental public
health continuum. Environ. Health Perspect. 114, 1776-1782.
Calafat, A.M., Kuklenyik, Z., Reidy, J.A, Caudill, S.P., Tully, J.S., Needham, LL., 2007a.
Serum concentrations of 11 polyfluoroalkyl compounds in the US population:
data from the national health and nutrition examination survey (nhanes) 1999-
2000. Environ. Sci. Technol. 41, 2237-2242.
Calafat, A.M., Wong, L.-Y., Kuklenyik, Z., Reidy, J.A., Needham, L.L., 2007b.
Polyfluoroalkyl chemicals in the US population: data from the national health
and nutrition examination survey (nhanes) 2003-2004 and comparisons with
nhanes 1999-2000. Environ. Health Perspect. 115, 1596-1602.
DePierre, J.W., 2009. Personal communication.
DeWitt, J., Shnyra, A., Badr, M., Loveless, S., Hoban, D., Frame, S.R., Cunard, R.,
Andersen, S., Meade, BJ., Peden -Adams, M., Luebke, R., Luster, M., 2009.
Immunotoxicity of perfluorooctanoic acid and perfluorooctane sulfonate and
the role of peroxisome proliferator-activated receptor alpha. Crit. Rev. Toxicol.
39,76-94.
Dong, G.-H., Zhang, Y.H., Zheng, L., Liu, W., Jin, Y.-H., He, Q.-C., 2009. Chronic effects
of perfluorooctanesulfonate exposue on immunotoxicity in adult male c57b1/6
mice. Arch. Toxicol. 83, 805-815.
Eliresman, D.J., Froehlich, J.W., Olsen, G.W., Chang, S.-C., Butenhoff, J.L, 2007.
Comparison of human whole blood, plasma, and serum matrices for the
determination of perfiLiOrOOCLaiiesulfonate (pfos), perfluorooctanoate (pfoa),
and other fluorochemicals. Environ. Res. 103, 176-184.
Elcombe, C.R., Elcombe, B.M., Foster, J.R, Farrar, D.G., Jung, R., Chang, S.-C.,
Kennedy Jr., G.L., Butenhoff, J.L., 2010. Hepatocellular hypertrophy and cell
proliferation in sprague-dawley rats following dietary exposure to
ammonium perfluorooctanoate occurs through increased activation of the
xenosensor nuclear receptors ppar (alpha) and car/pxr. Arch. Toxicol. 84,
787-798.
Emmett, E.A., Shofer, F.S., Zhang, H_ Freeman, D., Desai, C., Shaw, L.M., 2006.
Community exposure to perfluomoctanoate: relationships between serum
concentrations and exposure sources. J. Occup. Environ. Med. 48, 759-770.
Fiserova-Bergerova, V., 1975. Biological -mathematical modeling of chronic toxicity.
Wright -Patterson Air Force Base, OH: Aerospace Medical Research Laboratory,
AMRL-TR-75-5.
Fisher, J.W., Dorman, D.C., Medinsky, M.A., Welsch, F., Conolly, R.B., 2000. Analysis
of respiratory exchange of methanol in the lung Of the monkey using a
physiological model, Toxicol. Sci. 53, 185-193.
Forsyth, R.P., Nies, AS_ Wyler, F., Neutze, J., Melmon, K.L., 1968. Normal distribution
of cardiac output in the unanesthetized, restrained rhesus monkey. J. Appl.
Physiol. 25, 736-741.
Frisbee, S., Brooks, A.P_ Maher, A., Flensborg, P., Arnold, S., Fletcher, T., Steenland, K.,
Shankar, A., Knox, S,S., Pollard, C., Halverson, J.A., Vieira, V.M., Jin, C., Leyden,
K.M., Ducatman, A. 2009. The c8 health project: design, methods, and
participants, Environ. Health Perspect. 117, 1783-1882.
Fromme, H., Tittlemier, S.A., Volkel, W., Wilhelm, M., Twardella, D., 2009.
Perfluorinated compounds - exposure assessment for the general population
in western countries. hit. J. Hyg. Environ. Health 212, 239-270.
Goldenthal, E.I., 1978. Final report, ninety day subacute rat toxicity Study on fluorad
fluorochemical fc-143. International Research and Development Corporation,
Study No. 137-089, 3M Reference No. T-3141.
Gonzalez, 'F.J., Shah, Y.M., 2008. Ppar-o,.: mechanism of species differences and
hepatocarcinogenesis of peroxisome proliferators. Toxicology 246, 2-8.
Han, X, Snow, T.A., Kemper, RA, Jepson, G.W., 2003. Binding of perfluorooctanoic
acid to rat and human plasma proteins. Chem. Res. Toxicol. 16, 775-781.
Hanhijarvi, H., Ylinen, M., Kojo, A., Kosma, V.-M., 1987. Elimination and toxicity of
perfluorooctanoic acid during subchronic administration in the Wistar rat.
Pharm. Toxicol. 61, 66-68.
Harada, K., Inoue, K., Morikawa, A., Yoshinaga, T., Saito, N., Koizurni, A., 2005. Renal
clearance of perfluorooctane sulfonate and perfluorooctanoate in humans and
their species -specific c excretion. Environ. Res. 99, 253-261.
Hilgenclorf, C., Ahlin, G., Seithel, A., Artursson, P., Ungell, A.L., Karlsson, J., 2007.
Expression of thirty-six drug transporter genes in human intestine, liver,
kidney, and organotypic cell lines. Drug Metab. Dispos. 35, 1333-1340.
Hinderliter, P.M., Han, X., Kennedy Jr., G.L., Butenhoff, J.L., 2006. Age effect on
perfluorooctanoate (pfoa) plasma concentration in post -weaning rats following
oral gavage with ammonium perfluorooctanoate (apfo). Toxicology 225, 195-
203.
Hinderliter, P.M., Mylchreest, E., Gannon, S.A., Butenhoff, J.L., Kennedy Jr., G.L., 2005.
Perfluorooctanoate: placental and lactational transport pharmacokinetics in
rats. Toxicology 211, 139-148.
H61zer, J., Goen, T., Rauchfuss, K., Kraft, M., Angerer, J., Kleeschulte, P., Wilhelm, M.,
2009. One-year follow-up of perfluorinated compounds in plasma of german
residents from arnsberg formerly exposed to pfoa-contaminated drinking
water. Int. J. Hyg. Environ. Health 212, 499-504.
H61zer, J., Miclasch, 0., Rauclifuss, K., Kraft, M., Reupert, P_ Angerer, J., Kleeschulte,
P., Marschall, N., Wilhelm, M., 2008. Bionionitoring of perfluorinated
compounds in children and adults exposed to perfluorooctanoate-
contaminated drinking water. Environ. Health PerspeCL. 116, 651-657.
Hundley, S.G., Sari -if, A.M., Kennedy Jr., G.L., 2006. Absorption, distribution, and
excretion of ammonium perfluorooctanoate (apfo) after oral administration to
various species. Drug Chem. Toxicol. 29, 137-145.
Johnson, J.D., Gibson, S.J., Ober, R.E., 1979. Extent and Route of Excretion and Tissue
Distribution of Total Carbon-14 in Rats after a Single Intravenous Dose of fc-95-
14c. Riker Laboratories, Inc., St. Paul, MN, USEPA Docket No. 8(e)HQ-1180-
00374.
Johnson, J.D., Gibson, SJ., Ober, R.E., 1984. Cholestyramine-enhanced fecal
elimination of carbon-14 in rats after administration of ammonium
[ 14C jperfluorooctanoate or potassium [ 14C jperfluorooctanesulfonate. Fundam.
Appl. Toxicol. 4, 972-976.
Jones, P.D., Hu, W., De Coen, W., Newsted, J.L., Giesy, J.P., 2003. Binding of
perfluorinated fatty acids to serum Proteins. Environ. Toxicol. Chem. 22, 2639-
2649.
Karrman, A., Ericson, L, Van Bavel, B., Darnerud, P.O., Aune, M., Glynn, A., Lignell, S.,
Lindstrom, G., 2007. Exposure of perfluorinated chemicals through lactation:
Levels of matched human milk and serum and a temporal trend, 1996-2004, in
Sweden. Environ. Health Perspect. 115, 226-230.
DEQ-CFW-00004224

A.E. Loccisano et al./Regulatory Toxicology and Pharmacology xxx (2010) xxx-xxx
19
Kemper, R., 2003. Perfluooctanoic acid: toxicokinetics in the rat. DuPont Haskell
Laboratories, Project ID: DuPont-7473. USEPA Administrative Record AR-226.
Kennedy Jr., G.L., Butenhoff, J.L, Olsen, G.W., O'Conner, J.C., Seacat, A.M., Perkins,
R.G., Biegel, L.B., Murphy, S.R., Farrar, D.G., 2004. The toxicology of
perfluorooctanoate. Crit. Rev. Toxicol. 34, 351-384.
Kerstner-Wood, C., Coward, L., Gorman, G., 2003. Protein binding of perfluorobutane
sulfonate, perfluorohexane sulfonate, perfluorooctane sulfonate, and
perfluorooctanoate to plasma (human, rat, and monkey) and various human -
derived plasma Protein fractions. Southern Research Corporation, Study 9921.7.
USEPA Docket AR-226-1354.
Kudo, N., Katakura, M., Sato, Y., Kawashima, Y., 2002. Sex hormone -regulated renal
transport of perfluorooctanoic acid. Chem.-Biol. Int. 139, 310-316.
Kudo, N., Sakai, A., Mitsurnoto, A., Hibino, Y., Tsuda, T., Kawashima, Y., 2007. Tissue
distribution and hepatic subcellular distribution of perfluorooctanoic acid at
low doses are different from those at high doses. Biol. Pharm. Bull. 30, 1535-
1540.
Kuklenyik, Z., Reich, J.A., Tully, J.S., Needham, L.L, Calafat, AM., 2004. Automated
solid -phase extraction and measurement of perfluorinated organic acids and
amides in human serum and milk. Environ. Sci. Technol. 38, 3698-3704.
Kuslikis, B.I., Vanden Heuvel, J.P., Peterson, R.E., 1992. Lack of evidence for
perfluorodecanoyl- or perfluorooctanoyl-coenzyme a formation in male and
female rats. J. Biochem. Toxicol. 7, 25-29.
Lake, B.G., 2009. Species differences in the hepatic effects of inducers of cyp2b and
cyp4a subfamily forms: relationship to rodent liver turnout formation.
Xenobiotica 39, 582-596.
Lau, C., Anitole, K, Hodes, C., Lai, D., Pfahles-Hutchens, A., Seed, J., 2007.
Perfluoroalkyl acids: a review of monitoring and toxicological findings.
Toxicol. Sci. 99, 366-394.
Lau, C., Butenhoff, J.L, Rogers, J.M., 2004. The developmental toxicity of
perfluoroalkyl acids and their derivatives. Toxicol. Appl. Pharmacol. 198, 231-
241.
Lhwa Member Blood Testing Results (May 2005), 2005. Little Hocking Water Board,
<http://www.littlebockingwater.org,, 2005 (accessed May 2009).
Luebker, D.J., Hansen, K.J., Bass, N.M., Butenhoff, J.L, Seacat, A.M., 2002. Interactions
of fluorochemicals with rat liver fatty acid -binding protein. Toxicology 176,
175-185.
Macey, R., Oster, G., Zalmley, T., 2000. Berkeley-Madomia. University of California,
Department of Cellular and Molecular Biology. Berkeley, CA.
Nakagawa, H., Hirata, T., Terada, T., JULabha, P., Miura, D., Harada,J.H., Inoue, K.,
Anzai, N., Endou, H., Inui, K, 2008. Roles of organic anion transporters in the
renal excretion of perfluorooctanoic acid. Basic Clin. Pharmacol. Toxicol. 103, 1-
8.
Nakagawa, H., Terada, T., Harada, K.H., Hitorm, T., Inoue, K, Inui, K., Koizumi, A_
2009. Human organic anion transporter hoat4 is a transporter of
perfluooctanoic acid. Basic Clin. Pharmacol. Toxicol. 105,136-138.
Noker, P., Gorman, G., 2003. A pharmacokinetic study of potassium
perfluorooctanesulfonate in the cynomolgus monkey. Southern Research
Institute, Study ID: 9921.6, USEPA Public Docket, Administrative Record AR-
226-1356.
Olsen, G.W., Burris, J.M., Ehresman, D.J., Froehlich, J.W., Seacat, AM., Butenhoff, J,L,
Zobel, I-R., 2007. Half-life of serum elimination of perfluorooctanesulfonate,
perflLiorohexanesulfonate, and perfluorooctanoate in retired fluorochemical
production workers. Environ. Health Perspect. 115, 1298-1305.
Olsen, G.W., Church, T.R., Miller, J.P., Burris, J.M., Hansen, K.J., Lundberg, JX,
Armitage, J.B., Herron, R.M., Medlidizadehkashi, Z., Nobiletti, J.B., O'Neill, E.M.,
Mandel, J.H., Zabel, L.R., 2003a. Perfluorooctanesulfonate and other
fluorochernicals in the serum of American red cross adult blood donors.
Environ. Health Perspect. 111, 1892-1901.
Olsen, G.W., Hansen, K.J., Stevenson, LA., Burris, J.M., Mandel, J.H., 2003b. Human
donor liver and serum concentrations of perfluooctane sulfonate and other
perfluorochemicals. Environ. Sci. Technol. 37, 888-891.
Olsen, G.W., Mair, D.C., Church, T.R., Ellefson, M.E., Reagen, W.K., Boyd, T.M., Herron,
R.M., Medhdizadehkashi, Z., Nobiletti, J.B., Rios, J.A., Butenhoff, J.L., Zobel, L.R.,
2008. Decline in perfluorooctanesulfonate and other polyfluoroalkyl chemicals
in American Red Cross adult blood donors, 2000-2006. Environ. Sci. Technol. 42,
4989-4995.
Ophaug, R.H., Singer, L., 1980. Metabolic handling of perfluorooctanoic acid in rats.
Proc. Soc. Exp. Biol. Med. 163, 19-23.
Ostertag, SX, Chan, H.M., Moisey, J., Dabeka, R., Tittlernier, S.A., 2009. Historic
dietary exposure to perfluorooctane sulfonate, perfluorinated carboxylates, and
fluorotelemer unsaturated carboxylates from the consumption of store-bought
and restaurant foods for the Canadian population. J. Agric. Food Chem. 57,
8534-8544.
Paustenbach, D.J., Panko, J.M., Scott, P.K., Unice, K.M., 2007. A methodology for
estimating human exposure to petfluorooctanoic acid (pfoa): a retrospective
exposure assessment of a community (1951-2003). J. Toxicol. Environ. Health,
A 70, 28-57.
Prevedouros, K., Cousins, I.T., Buck, R.C., Korzeniowski, S.H., 2006. Sources, fate, and
transport of perfluorocarboxylates. Environ. Sci. Technol. 40, 32-44.
Qazi, M.R., Nelson, B.D., Depierre, J.W., Abedi-Valugerdi, M., 2010. 28-Day dietary
exposure of mice to a low total dose (7 mg/kg) of perfluorooctane sulfonate
(pfos) alters neither the cellular compositions of the thymus and spleen nor
Immoral immune responses: does the route of administration play a pivotal role
in pfos-inducted immunotoxicity? Toxicology 267, 132-139.
Riker, 1987. Two year oral (diet) toxicity/carcinogenicity study offluorochemical fc-
143 in rats. Riker Laboratories, Inc.. Experiment No. 0281CR0012..
Rodriguez, C.E., Setzer, R.W., Barton, H.A., 2009. Pharmacokinetic modeling of
perfluorooctanoic acid during gestation and lactation in the mouse. Reprod.
Toxicol. 27, 373-386.
Rosen, M.B., Schmid, j.R., Carton, J.C., Zehr, R.D., Das, K.P., Abbott, B.D., Lau, C., in
press. Gene expression profiling in wild -type and ppara-mill mice exposed to
perfluorooctane sulfonate reveals pparcx-independent effects. PPAR Res.
Seacat, AM., Thomford, P.J., Hansen, K.J., Olsen, G.W., Case, M.T., Butenhoff, J.L.,
2002. Subchronic toxicity studies on perfluorooctanesulfonate potassium salt in
cynomolgus monkeys. Toxicol. Sci. 68, 249-264.
Sinclair, E., Kim, SX, Akinleye, H,B., Kannan, K, 2007. Quantitation of gas -phase
perfluoroalkyl surfactants and fluorotelomer alcohols released from nonstick
cookware and microwave popcorn bags. Toxicol. Sci. 41, 1180-1185.
Skutlarek, D., Exner, M., Farber, H., 2006. Perfluorinated surfactants in surface and
drinking waters. Environ. SCI. PCIILIL. Res. 13, 299-307.
Steenland, K., Jin, C., MacNeil, J., Lally, C., Ducatman, A., Vieira, V., Fletcher, T., 2009.
Predictors of pfoa levels in a community surrounding a chemical plant. Environ.
Health Perspect. 117, 1083-1088.
Strynar, M.J., Lindstrom, A.B., 2008. Perfluorinated compounds in house dust from
Ohio and North Carolina, USA Environ. Sci. Technol. 42, 3751-3756.
Tahara, H., Shona, M., Kusuhara, H., Kinoshita, H., Fuse, E., Takadate, A., Otagiri, M.,
Sugiyama, Y., 2005. Molecular cloning and functional analyses of oatl and oat3
from cynomolgus monkey kidney. Pharm. Res. 22, 647-660.
Tan, Y.-M., Clewell, H.J., Andersen, M.E., 2008. Time dependencies in
perfluorooctylacids disposition in rat and monkeys: a kinetic analysis. Toxicol.
Lett. 177, 38-47.
Thibodeaux, J.R., Hanson, R.G., Rogers, J.M., Grey, B.E., Barbee, B.D., Richards, J.H.,
Butenhoff, J.L., Stevenson, LA., Lau, C., 2003. Exposure to perflU01`00CLane
sulfonate during pregnancy in rat and mouse. 1: maternal and prenatal
evaluations. Toxicol. Sci. 74, 369-381.
USEPA, 1997. Exposure Factors Handbook. <http://www.epa.gov/ncea/efh/pdfs/eth-
front-gloss.pdf>.
USEPA, 2005. Draft Risk Assessment of the Potential Human Health Effects
Associated with Exposure to Perfluorooctanoic Acid and its Salts. <http:/'/
www.epa.gov/opptintr/pfoa/pubs/pfoarisk.pdf>.
USEPA, 2009. Provisional health advisories for perfluorooctanoic acid (pfoa) and
perfluorooctanesulfonate (pfos). http://www.epa.gov/waterscience/criteria/
drinking/pba-PFOA PFOS.pdf.
Vanden Heuvel, J.P., Kuslikis, B.I., Rafelghem, M.J., Peterson, R.E., 1991. Tissue
distribution, metabolism, and elimination of perfluorooctanoic acid in male and
female rats. J. Biochem. Toxicol. 6, 83-92.
Vestergren, R., Cousins, I.T., 2009. Tracking the pathways of human exposure to
perfluorocarboxylates. Environ. Sci. Technol. 43, 5565-5575.
Washburn, S.T., Bingman, T.S., Braithwaite, S.K., Buck, R.C., Buxton, LW., Clewell 111,
H.J., Haman, L.A., Kester, J.E., Rickard, R.W., Shipp, A.M., 2005. Exposure
assessment and risk characterization for perfluorooctanoate in selected
consumer articles. Environ. Sci. Technol. 39, 3904-3910.
Weaver, Y.M., Ehresman, D.J., Butenhoff, J.L., Hagenbuch, B., 2009. Roles of rat renal
organic anion transporters in transporting perfluorinated carboxylates with
different chain lengths. Toxicol. Sci. 113, 305-314.
Wilhelm, M., Kraft, M., Rauchfuss, K, Holzer, J., 2008. Assessment and management
of the first german case of a contamination with perfluorinated compounds
(pfc) in the region Sauerland, North Rhine-Westphalia. J. Toxicol. Environ.
Health, A 71, 725-733.
Yang, C.-H., Glover, K.P., Hall, X., 2010. Characterization of cellular uptake of
perfluoroctanoate via organic anion -transporting polypeptide 1a2, organic
anion transporter 4, and urate transporter I for their potential roles in
mediating human ren], reabsorption of perfluorocarboxylates. Toxicol. Sci.
117,294-302.
Yang, C., Glover, K.P., Han, X., 2009. Organic anion transporting polypeptide (oatp)
lal-mediated perfluorooctanoate transport and evidence for a renal
reabsorption mechanism of oatplal in renal elimination of
perfluorocarboxylates in rats. Toxicol. Lett. 190, 163-171.
Yang, Q., Xie, Y., Depierre, J.W., 2000. Effects of peroxisome proliferators on the
thymus and spleen of mice. Clin. Exp. Immunol. 122, 219-226.
DEQ-CFW-00004225