<>EPA
www.epa.gov
Evaluating Per- and Polyfluoroalkyl Substances (PFAS) by In Vitro
Toxicokinetic Data Generation with In Vitro-ln Vivo Extrapolation (IVIVE)
Smeltz, M.a, Crizer, D. b, McMillan, L.a, Patlewicz, G.a, DeVito, M.a, Wetmore, B.A.a	f '
a United States Environmental Protection Agency, Office of Research and Development, Research Triangle Park, North Carolina; b Division of the National Toxicology Program, National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina
Background
New approach methodologies (NAMs) make use of in vitro high-throughput screening (HTS) data and in
silico approaches to inform chemical safety assessment through the translation of bioactive in vitro
assay concentrations to administered dose equivalents. This approach utilizes in vitro point of departure
information from HTS studies where in vitro toxicokinetic (TK) data is applied to assigned administered
dose equivalents, and to compare exposure estimates to define ad hoc margin of exposure.
in vitro TK data and modeling -¥ in vivo concentrations ad hoc MOE
Per- and polyfluoroalkyl substances (PFAS) have become chemicals of concern for human health as
more is learned about their widespread presence and persistence in the environment. Given the
inclusion of 1,220 PFAS on the Toxic Substances Control Act (TSCA) inventory, the availability of in vivo
toxicologic data and exposure information on only a subset is inadequate to provide an understanding of
the potential exposures, TK, and toxicities across this structurally diverse domain. To address this
deficiency, we are applying NAMs to a panel of PFAS, containing carboxylic acid, sulfonate, and ether
functionalities, for use in IVIVE models to predict systemic concentrations and eventual application in
read-across approaches.
In Vitro Toxicokinetic Assays
•	Human plasma (10-donor pool, mixed sex) centrifuged at
850,000xg to separate aqueous fraction from albumin,
lipoproteins, and fatty acids
•	Analyte quantitation (multiple reaction monitoring, MRM) for
aqueous fraction performed on Waters Xevo TQ-S (ultra-high-
performance liquid chromatography-tandem mass
spectrometry, UPLC-MS/MS)
FU=AF/(T5)
•	Substrate depletion approach using
primary human hepatocytes (50-donor
pool, mixed sex) at 1 jjM PFAS
concentration
•	Time course: 0, 15, 30, 60, 90, 120, and
240 min with non-linear regression fit to
determine half-life (T1/2)
•	Single Ion Monitoring mode on Thermo *
Vanquish with Q Exactive Plus /
U.S. Environmental Protection Agency
Office of Research and Development
Determining Fraction Unbound by LC-MS/MS
Fraction unbound (Fu) was determined by measuring PFAS concentration in aqueous fraction and
plasma by Waters Xevo TQ-S micro (UPLC-MS/MS) in 6.5 min run with low ppb (pg/pL) sensitivity



DTXSID3031864
PFOS
DTXSID7027831
NMeFOSE
Category-based Grouping and Distribution Analyses of PFAS Plasma Protein Binding Data

cO.001 0.005 0.01 0.05 0.10 0.15
Fu Distribution
Marci Smeltz I smelt2.marci@epa.gov I 919-541-1064
IVIVE Modeling Results
IVIVE modeling was performed using collected in vitro Clhepatic and Fu data for a representative set of
PFAS to predict steady-state blood concentrations (Css) for an adult population:
Compound Name
Fu
Clrenal (L/hr)
Clhepatic (L/hr)
Css (nM)
Potassium perfluorohexanesulfonate
0.0011
0.0075
3.82E-07
894.5132
Ammonium perfluorooctanoate
0.0014
0.0094
1.16E-04
713.7360
Perfluorononanoic acid
0.0013
0.0088
8.33E-03
368.6974
Perfluorohexanoic acid
0.0076
0.0507
2.33E-04
183.6569
Potassium perfluorobutanesulfonate
0.0087
0.0581
2.75E-02
101.5252
Perfluorooctanesulfonic acid
0.0073
0.0490
5.38E-02
57.1902
Perfluoro(4-methoxybutanoic) acid
0.0142
0.0950
2.97E-01
26.7545
Perfluorobutanoic acid
0.1032
0.6927
1.68E-05
19.8299
2H,2H,3H,3H-Perfluorooctanoic acid
0.0072
0.0483
5.15E-01
15.2577
Perfluoro-3,6,9-trioxatridecanoic acid
0.0026
0.0176
6.38E-01
7.9748
4:2 Fluorotelomer sulfonic acid
0.0142
0.0951
5.55E+00
1.5874
N-Ethylperfluorooctanesulfonamide
0.0464
0.3110
5.57E+00
0.9485
N-Methylperfluorooctanesulfonamide
0.0113
0.0757
7.43E+00
0.7633
Perfluorooctanesulfonamide
0.0229
0.1536
3.60E+01
0.1630
Assumptions: 1 pg/kg/day exposure, linear kinetics, 100% bioavailability, and nonmetabolic renal clearance only
Summary
.) are being measured on over 100 PFAS for
Experimental in vitro toxicokinetic data (Fu and Clhe
use in IVIVE modeling.
Plasma protein binding data on 49 PFAS measured to date using UPLC-MS indicate high binding
rates, with 75% exhibiting Fu values from 0.001 - 0.05.
Preliminary category-based analyses of Fu values show perfluoroalkyl carboxylates have the greatest
range, while other functional groups exhibit tighter distributions.
Assuming an external exposure of 1 pg/kg/day, Css predictions ranged from 0.16-895 pM, with a
median value of 23.29 pM. Half-life (T!4) estimates tracked similarly, with 50% of tested PFAS having
T1/4 s < 100 hr and ~20% with VA values exceeding 1000 hr.
Data generation across additional PFAS and toxicokinetic assays for bioavailability and renal
reuptake are currently underway.
Disclaimer: The views expressed in this poster are those of the authors and do not necessarily
represent the views or policies of the U.S. Environmental Protection Agency.

-------