Data Requirements for Developing IVIVE Models Nynke I. Kramer, PhD Institute for Risk Assessment Sciences Utrecht University N.I.Kramer@uu.nl

Conflict of Interest Statement

I disclose that I have no conflicts of interest

Overview

What data is needed for IVIVE models?

In vitro assays in risk assessment QIVIVE: an integrated approach Parameters in PBPK models Toxicokinetics and toxicodynamics assays Challenges translating in vitro effect concentrations

In Vitro Assays

Toxicodynamics

Toxicokinetics concentration % o

f liv

ing

cells

EC50

In Vitro Assays in Risk Assessment

Toxicity Testing in 21st Century – National Research Council (2007), Toxicity Testing: A Vision and a

Strategy, DOI: 10.17226/11970

From hazard identification to hazard characterization

Quantitative In Vitro-In Vivo Extrapolation (QIVIVE) – Estimating environmental chemical exposures producing

target tissue exposures in humans equivalent to those associated with effects in in vitro toxicity tests

QIVIVE Louisse et al. (2010) Tox. Sci. 118, 470

In vitro toxicodynamics

Mouse embryonic stem cell test

In vitro toxicokinetics Biotransformation kinetics in hepatocytes

alkoxyacetic acid formation

Blood

Fat

Liver

Slowly perfused tissue

Rapidly perfused tissue

GI tract

Blood

Fat

Liver

Slowly perfused tissue

Rapidly perfused tissue

Input: BMC values meta-bolite alkoxyacetic acid

Output: predicted oral embryotoxic dose parent glycol ethers Evaluation of IVIVE model with in

vivo embryotoxic doses

Integrated Approach

PBPK, reverse dosimetry Exposure assessment Sensitivity analysis In silico tools (e.g. read-across, QSAR, QSPR) AOP In vitro tools, KE / TK Margin of exposure Tiered/iterative process

Steps in IVIVE Model Development

Potential Target Tissue

PBPK Model

In vivo Human Toxicty Estimate

In vitro Dynamics

In Vitro Kinetics

QSAR QSPR

Metabolite ID

Target tissue response

In vivo exposure profiles

Nature of Toxicity

• Hepatic clearance • Intestinal uptake /

metabolism • Renal clearance • Partitioning

Reverse Dosimetry

From Miyoung Yoon Yoon et al. 2012, Crit Rev Toxicol. 42, 633

Bottom-up approach

Parameterization PBPK for IVIVE

– Physiological/Anatomical Tissue blood flow or

respiration rate (Q, e.g. L/hr) Tissue volumes (V) Glomerular filtration rate

Well characterized: Brown

et al. (1997) Toxicol. Indust. Health 13, 407. Tissues

Liver

Veno

us b

lood

Arte

rial b

lood

QtxCin

QtxCin

QtxCout

QtxCout

QtxCin

QtxCin

QtxCin

QtxCout

Fat QtxCout QtxCin

metabolism

dAt/dt = Vt·dC/dt = Qt · (Cin – C/P) Cout = Cfree = Ctissue/P dAm/dt = R = Cfree·Vt·K = (Vmax·Cfree)/(KM + Cfree)

Gut/lung/skin

Parameterization PBPK for IVIVE

– Physicochemical Tissue-blood partition

coefficients (P = Ctissue/Cfree) Protein binding constants Tissue clearance/elimination

rates (K, Vmax, KM) Absorption (bioavailability)

In silico tools – QSPR estimating

partitioning/protein binding – Challenge: metabolic

clearance parameter (QSAR limited, mostly qualitative, METEOR, OECD Toolbox, Multicase) In vitro tools

Tissues

Liver

Veno

us b

lood

Arte

rial b

lood

QtxCin

QtxCin

QtxCout

QtxCout

QtxCin

QtxCin

QtxCin

QtxCout

Fat QtxCout QtxCin

metabolism

dAt/dt = Vt·dC/dt = Qt · (Cin – C/P) Cout = Cfree = Ctissue/P dAm/dt = R = Cfree·Vt·K = (Vmax·Cfree)/(KM + Cfree)

Gut/lung/skin

Toxicokinetics Input

In silico tools Tier 1

Passive diffusion down C gradient

Diffusion into tissue Non-saturable protein binding Well-stirred tissue Whole-body D perfusion-limited

Bessems et al. (2014) Reg. Toxicol. Pharmacol. 68, 119 Peyret and Krishnan (2011) SAR QSAR Environ. Res. 22, 129 Liao et al. (2007) Risk Anal. 27, 1223 Wetmore (2015) Toxicol. 332, 94

J = Papp*SA*C

Toxicokinetics Input

In vitro tools Bessems et al. (2014) Reg. Toxicol. Pharmacol. 68, 199

Absorption Oral (Caco-2, PAMPA) Dermal (OECD TG 428, Skin PAMPA) Inhalation (Head space model for VOS, Calu-3)

Distribution Partition coefficients Protein binding Permeation coefficients

Metabolism Human microsomes,

(cryopreserved) primary hepatocytes, HepaRG, HepG2

Excretion See distribution & human

physiology, headspace model, e.g. cell lines RPTEC/TERT1

In Vitro Tools for Distribution Parameters

Method Advantages Disadvantages Equilibrium Dialysis Widely used Time consuming,

membrane binding

Ultracentrifugation No binding to apparatus

Expensive apparatus

Ultrafiltration

Fast and inexpensive

Filter binding, equilibrium shift

Heringa et al. (2003) TAC 22: 575-587

Solid Phase Microextraction

Measuring Kcompartment

SPME

• Measuring free conc. • Nd-SPME

Amount fiber << Amount medium

])[1(1

1

AKKV

VF

aff

mf

⋅+⋅

+=

Cf Cfree

Cb

Ka

Kf

)(

)(

eqfree

eqff C

CK =

)(

)(

infree

eqff C

CK =

well cells

medium fiber

serum protein

Increasing serum concentration

In Vitro Tools for Metabolism Parameters

• Tissue slices • Primary hepatocytes • Hepatic cell lines • Microsomes, cytosol • Recombinant CYPs

• Parent chemical disappearance vs. metabolite generation • Dose-response (Vmax, Km) vs. activity

Toxicodynamics Input

Concentration-effect relationships

AOP ‘Omics’

Interaction with target

Effect tissue/ pathway

Effect organism

AOP framework MIE > KE > AO

Toxicodynamics (TD) Chemical agnostic (biological pathway)

Parameterization Challenges

Variability in in vitro readout Poorly defined / Uncertainties

– Chemical applicability domain – Biological applicability domain/link effect with adversity – In vitro effective dose

Good Modeling Practice – Formal verification process useful

Good Cell Culture Practice – maximise reproducibility, reliability, credibility, acceptance and

proper application of in vitro results – minimum standards in cell and tissue culture

Price and Coecke (2011) in Cell Culture Techniques, Springer, DOI 10.1007/978-1-61779-077-5_1 Loizou et al. (2008) Reg Toxicol. Pharmacol. 50, 400

In Vitro Dose Metrics

plastic binding

free in medium

protein binding

evaporation

cell binding

target free in cell metabolism CELL

MEDIUM

PLASTIC (well)

plastic binding

free in medium

protein binding

cell binding

target free in cell metabolism CELL

MEDIUM

PLASTIC (well)

Cell Assay Components …

… determining target concentration in vitro Medium components (e.g. serum, 3D matrix) Headspace/air flow Well plate dimensions Cell density Exposure time Temperature pH Metabolic capacity Transporter expression

Physicochemical Properties …

…determining target concentration in vitro Kd /LogP/LogD7.4/ KOW

H pKa Solubility Reactivity Transporter affinity Metabolic enzyme affinity

0.91 ± 0.08%

27.50 ± 5.21%

51.49 ± 6.00 %20.10 ± 2.43%

0.96 ± 0.04%

18.99 ± 5.82%

14.80 ± 4.17%

65.75 ± 7.45 %

5.61 ± 1.82% 0.96 ± 0.10%

10.55 ± 3.21%

82.88 ± 3.96 %

Medium Plastic Loss Cells

1.25%

2.5%

5%

Kramer et al. (2012) Chem. Res. Toxicol. 25, 436

Medium Constituents

Kramer et al. (2012) Chem. Res. Toxicol. 25, 436; Kramer (2010) PhD thesis, Utrecht University

3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.502.50

2.75

3.00

3.25

3.50

3.75

4.00

Log KOW

Log

K s (m

3 /mol

)Nominal EC50: 73 µM 105 µM 187 µM

Serum: 1.25% 2.5% 5%

Free EC50: 5 µM 5 µM 6 µM Free: 7% 5% 3%

Polycyclic aromatic hydrocarbons

Medium Components

Kramer et al. (2012) Chem. Res. Toxicol. 25, 436; Kramer (2010) PhD thesis, Utrecht University

3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.502.50

2.75

3.00

3.25

3.50

3.75

4.00

Log KOW

Log

K s (m

3 /mol

)Nominal EC50: 73 µM 105 µM 187 µM

Serum: 1.25% 2.5% 5%

Free EC50: 5 µM 5 µM 6 µM Free: 7% 5% 3%

Polycyclic aromatic hydrocarbons

Medium Components

• Stronger partitioning than Endo and Goss (2011) Chem. Res. Toxicol. 24, 2293, DeBruyn and Gobas (2007) Environ. Toxicol. Chem. 26, 1803 (neutral chemicals)

• Proteins vs. lipids • Neutral vs. ionized

Well Plate Dimensions

5% 2% 0% FBS

Kramer et al. (2012) Chem. Res. Toxicol. 25, 436 Kramer (2010) PhD Thesis, Utrecht University

3.00 4.00 5.00 6.00 7.00-4.00

-3.50

-3.00

-2.50

-2.00

-1.50

-1.00

-0.50

Log KOW

Log

K p (m

) Polycyclic aromatic hydrocarbons

Well Plate Dimensions

5% 2% 0% FBS

Kramer et al. (2012) Chem. Res. Toxicol. 25, 436 Kramer (2010) PhD Thesis, Utrecht University

3.00 4.00 5.00 6.00 7.00-4.00

-3.50

-3.00

-2.50

-2.00

-1.50

-1.00

-0.50

Log KOW

Log

K p (m

) Polycyclic aromatic hydrocarbons

Dependent on: • Chemical (ionization) • Concentration • Time EU FP7 Predict-IV • Kramer et al. 2015 Toxicol. In Vitro 30, 217

Headspace

0.0 0.5 1.0 1.5 2.0 2.5 3.0

-20

0

20

40

60

80

100

120

Alamar BlueCFDA-AMNeutral Red

Log 1,2,4-trichlorobenzene (concentration in µM)

% o

f con

trol

Recovery from medium after 48h

Conventional dosing: 11% Continuous dosing: 105%

EC50 (µM) Conventional: 135 Continuous total: 38 Continuous free: 11 Fathead Minnow: 16

Bound to Serum Constituents

70%

Nominal Cont. total

Cont. free

Kramer et al. (2010) Chem. Res. Toxicol. 23, 1806

Exposure Time

Gülden et al. (2010) Free Radical Bio Med 49, 1298 1800 uM*min = EC50 (see also Gulden et al. (2015) Toxicology 335, 35, time to incipient EC50 is chemical

dependent)

Cell Density

1 10 100

0

20

40

60

80

100

120

Dieldrin (µM)

% o

f via

ble

cells

15 million cells/mL31 million cells/mL62 million cells/mL123 million cells/mL

4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0

4.04.55.05.56.06.57.07.58.0

Log KOW

Log

Klip

id

Gülden et al (2001) Toxicol. In Vitro 15: 233

Jonker et al (2007) Environ. Sci. Technol. 41: 7363

Cell-Associated Concentration Groothuis et al., manuscript in preparation

Cell-Associated Concentration Groothuis et al., manuscript in preparation

• Chemicals acting by narcosis (basal cytotoxicity, accumulation in cell membranes)

• Passive vs. active uptake/efflux • Time

Repeated Dosing

Area under the curve or Cumulative dose (irreversible mechanism?)

Wilmes et al. (2013) J Proteomics 79: 180 Kramer et al. (2015) Toxicol. In Vitro 30: 217

Modeling Kinetics In Vitro

Austin et al (2002) Drug Metab. Disp. 30: 1497-1503 & Austin et al (2005) Drug Metab. Disp. 33: 419-425

Gülden et al (2003) Tox. 189: 211-222 & Gülden et al (2006) Tox. in Vitro 20: 1114-1124

Kramer et al (in preparation) & Armitage et al. (2014) Environ. Sci. Technol.: 48, 9770−9779

Zaldivar Comenges et al (in press) Toxicol. In Vitro & Stadnicka et al. (2014) PloS One 9(3): e92303

Modeling Kinetics In Vitro

F = 1 1 + Ks [S] + Kp [P] + Kc [C] + Ka [A]

HEADSPACE

evaporation (Ka)

PLASTIC (well)

free in cellmetabolismtarget

cell binding (Kc)

free in medium

plastic binding (Kp) protein binding (Ks)

MEDIUM

CELL

HEADSPACE

evaporation (Ka)

PLASTIC (well)

free in cellmetabolismtarget

cell binding (Kc)

free in medium

plastic binding (Kp) protein binding (Ks)

MEDIUM

CELL

m

aKowKowKow

VV

THCPS

F⋅++++

=−−−

3144.8][10][10][101

170.3log25.194.6log97.029.0log37.0

Modeling Kinetics In Vitro

Serum 0% 2% 5% Measured free

21% 8% 5%

Modeled free

32% 9% 5%

Measured in cells

10% 4% 2%

Modeled in cells

14% 5% 3%

RTgill-W1 exposed to phenanthrene

RTgill-W1

Conclusions

Data needed for IVIVE model dependent on – Required precision – Chemical – Exposure

Physiological vs. chemical-specific parameters Tiered, integrated approach Partitioning parameters from in silico tools Clearance parameters from in vitro tools Define applicability domains tools for parameterization Consider in vitro kinetics as well as in vivo kinetics

Thank you

Questions? – At this time, please limit questions to those specific to this

presentation, and save general questions for the Roundtable Discussion with all the presenters.

References

Armitage et al. (2014) Environ. Sci. Technol. 48, 9770−9779 Bessems et al. (2014) Reg. Toxicol. Pharmacol. 68, 119–139 Brown et al. (1997) Toxicol. Indust. Health 13, 407-484 Groothuis et al. (2015) Toxicol. 332, 30-40 Louisse et al. (2010) Toxicol. Sci. 118, 470–484 National Research Council (2007), Toxicity Testing: A Vision and

a Strategy, The National Academies Press, DOI: 10.17226/11970 Yoon et al. (2012) Crit. Rev. Toxicol. 42, 633-652 Peyret and Krishnan (2011) SAR QSAR Environ. Res. 22, 129-

169

Acknowledgements

Research funding: EU FP6 AcuteTox, EU FP7 Predict-IV, Unilever, CEFIC-LRI

Members of the Toxicology Division of UU IRAS incl. Bas Blaauboer, Joop Hermens, Floris Groothuis, Steven Droge, and Chiel Jonker

Miyoung Yoon, ScitoVation