Computational Toxicology and High-Throughput Risk …...All of these target-cancer links are backed...

Office of Research and Development

Computational Toxicology and High-Throughput Risk AssessmentRichard Judson U.S. EPA, National Center for Computational ToxicologyOffice of Research and Development

The views expressed in this presentation are those of the author and do not necessarily reflect the views or policies of the U.S. EPA

ToxForum July 2011

Office of Research and DevelopmentNational Center for Computational Toxicology

ToxCast / Tox21 ContextEDSP as case study

• Congress has mandated that EPA test 4000-8000 chemicals for their potential to be endocrine disruptors –Through interaction with estrogen and other pathways–US testing labs and EPA reviewers have a capacity to test ~100

chemicals per year in guideline studies (Tier 1)–40-80 year backlog

• Prioritization is a logical approach–Test chemicals for ability to interact with E, A, T, S in vitro–“Validate” in vitro assays against expert-derived reference chemicals –Estimate dose at which activity can occur–Compare with estimated human exposure levels–Suggest testing those with an overlap first in science-based, standard,

GLP, guideline studies

2


High-Throughput Risk Assessment (HTRA)

• Uses HTS data for initial, rough risk assessment of data poor chemicals

• Risk assessment approach–Estimate upper dose that is still protective (~ RfD, BMD)–BPAD (Biological Pathway Altering Dose)–Compare to estimated steady state exposure levels

• Contributions of high-throughput methods–Focus on molecular pathways whose perturbation can lead to adversity–Screen 100s to 1000s of chemicals in HTS assays for those pathways–Estimate oral dose using High-Throughput PK modeling (R. Thomas talk)

• Incorporate population variability and uncertainty3


High Throughput Risk Assessment (HTRA)

4BPAD Distribution

Pathway / Target / Model

CAR/PXR PathwayER / AR / Endocrine TargetsReproTox SignatureDevTox SignatureVascular Disruption SignatureThyroid Cancer Signature

Exposure / DoseLow High

Upper no effect dose

Critical Effect

No detectNo detect

No detect

HTRA Report Card For Chemical: ABC

Estimated Exposure


HTRA Outline

5

Identify biological pathways linked to adverse effects

Measure Biological Pathway Altering Concentration (BPAC) in vitro

Estimate in vivo Biological Pathway Altering Dose (BPAD) (PK modeling)

Incorporate uncertainty and population variability estimates

Calculate BPAD lower limit – Estimated health protective exposure limit


Assay targets linked to rodent cancer

6

Out of hundreds of assays and targets, only a few are statistically associated with adversity and therefore useful for toxicity testing prioritization

All of these target-cancer links are backed up by the broader biomedical literature


HTS – Pharmacodynamics (PD)

7

Sample curves for BPA in two ER assays

Full concentration-response profiles can be measured, at arbitrary spacing and to arbitrarily low concentrations

Result: AC20, AC50


HTRA HTS Chemical Space

8

…

ToxCast

EDSP21

Tox21

1,000 2,000 10,000

500 ToxCast assays ToxCast EATS

assays

Tox21 Assays

Pesticides (active and inert), industrial chemicals, consumer products,marketed and failed pharmaceuticals, food additives, water contaminants, natural human metabolites


In collaboration with Hamner Institutes / Rusty Thomas, et al.

Reverse Toxicokinetics

Human Hepatocytes

(10 donor pool)

Add Chemical(1 and 10 µM)

Remove Aliquots at 15, 30, 60, 120 min

Analytical Chemistry

-5

-4

-3

-2

-1

0

1

2

3

0 50 100 150

Ln C

onc

(uM

)

Time (min)

Hepatic Clearance

HumanPlasma

(6 donor pool)

Add Chemical(1 and 10 µM)

Analytical Chemistry

Plasma Protein Binding

EquilibriumDialysis

9

Combine experimental data with PK Model to estimate dose-to-concentration scaling

“Reverse Toxicokinetics”


Steady State Blood Concentration (Css) :

Upper 95th Percentile Among Healthy Individuals of Both

Sexes from 20 to 50 Yrs Old

Oral Doses Required to Achieve the Upper 95th

Percentile Steady State Blood Concentrations Across All In

Vitro Assays(Represented as a Box Plot)

Reverse Dosimetry

Oral Exposure

Blood Concentration

ToxCast AC50 or LEC Value

~500 In VitroToxCast Assays

Metabolic Clearance

Plasma Protein Binding

Population-Based In Vitro-to-In Vivo

Extrapolation

Oral Equivalent Distributions


Etox

azol

eEm

amec

tinBu

prof

ezin

Dibu

tyl p

htha

late

Pyra

clos

trobi

nPa

rath

ion

Isox

aben

Pryr

ithio

bac-

sodi

umBe

ntaz

one

Prop

etam

phos

2,4-

DS-

Bioa

lleth

rinM

GK

Atra

zine

Br

omac

ilFe

noxy

carb

Forc

hlor

fenu

ron

Met

hyl P

arat

hion

Tric

losa

nRo

teno

neCy

prod

inil

Isox

aflu

tole

Acet

amip

ridZo

xam

ide

Diur

onBe

nsul

ide

Vinc

lozo

linO

xyte

tracy

clin

e DH

Dicr

otop

hos

Met

ribuz

inTr

iadi

mef

onTh

iazo

pyr

Fena

mip

hos

Clot

hian

idin

Bisp

heno

l-AAl

achl

orAc

etoc

hlor

Diaz

oxon

Dich

lorv

osCh

lorp

yrip

hos-

oxon

TriclosanPyrithiobac-sodium

log

(mg/

kg/d

ay)

Rotroff, et al. Tox.Sci 2010

11

Range of in vitro AC50 values converted to humanin vivo daily dose

Actual Exposure (est. max.)

Safety margin

Combining in vitro activity and dosimetry


Chl

oron

ebPa

clob

utra

zol

Pyrip

roxy

fen

Mes

otrio

neD

icam

baO

xyflu

orfe

nC

acod

ylic

aci

dA

ceph

ate

Chl

orid

azon

Linu

ron

Imaz

amox

Met

hida

thio

nC

arbo

xin

Imaz

apic

Pirim

icar

bO

xam

ylEt

hopr

opC

yana

zine

Fena

mip

hos

Tria

dim

efon

Cyc

lani

lide

Thia

clop

ridFl

urox

ypyr

-mep

tyl

Phen

oxye

than

olB

isph

enol

-ATe

tram

ethr

inM

etal

axyl

Die

thyh

exyl

pht

hala

teFe

nhex

amid

Icar

idin

Trifl

usul

furo

nPr

opam

ocar

b H

Cl

Ben

dioc

arb

Tebu

thiu

ron

Vinc

lozo

linTr

ibuf

osB

ifena

zate

Imid

aclo

prid

Met

olac

hlor

Etho

fum

esat

eTh

ioph

anat

e m

ethy

lFl

ufen

acet

Daz

omet

Bro

mox

ynil

Sulfe

ntra

zone

Dim

ethe

nam

idS-

Bio

alle

thrin

Car

bary

lA

lach

lor

Hex

azin

one

Azi

npho

s-m

ethy

lA

ceto

chlo

rd-

cis,

tran

s- A

lleth

rinPy

met

rozi

neFo

rmet

anat

e H

Cl

Flua

zifo

p-P-

buty

lD

ibut

yl p

htha

late

Dim

ethy

l pht

hala

teD

iazo

xon

0.00001

0.0001

0.001

0.01

0.1

1

10

100

1000

10000

100000

Compound

Ora

l Equ

ival

ent D

ose

or E

stim

ated

Exp

osur

e(m

g/kg

/day

)

DicambaFenhexamidFluroxypyr-meptyl


BPAD Probability Distribution

• “Biological Pathway Altering Dose”

• BPAD = BPAC / Css / DR

• Add in uncertainty and population variability

• Take low dose end of distribution (BPADL) as health-protective estimate of allowable safe exposure level

13

BPAD Distribution

This is being used where no animal data is available!


BPAD Variability and Uncertainty

• Variability–PK – models are available (CYPs, liver mass, age,

populations)–PD – HapMap vs. chemicals vs. assays (UNC, NCGC)

• Uncertainty–PK – model uncertainty, experimental measurement –PD – assay background, experimental noise

14

BPAD Distribution


Conazoles and Liver Hypertrophy

• Conazoles are known to cause liver hypertrophy and other liver pathologies

• Believed to be due (at least in part) to interactions with the CAR/PXR pathway

• ToxCast has measured many relevant assays

• Calculate BPADL for 14 conazoles –Compare with liver hypertrophy NEL/100

15


Conazole / CAR/PXR results

16

LEL, NEL

BPAD Range

Exposure estimate

NEL/100

BPAD Distribution


Conazole Summary

• Rough quantitative agreement–Significant BPADL vs. NEL/100 rank correlation (p=0.025)–12 of 14 chemicals have BPADL within 10 of NEL/100–For only 3 is BPADL significantly less protective than

NEL/100–All BPADL > Exposure estimate

• Some apples to oranges: human BPADL, rat NEL–Rat RTK underway for some of these chemicals

17


BPADL vs. Exposure

• “Allowable exposure levels” (RfD, BPADL, etc.) are only relevant compared to exposure

• HTRA looks at many chemicals simultaneously in a prioritization context–Prioritize further testing if BPADL < estimated exposure

• Drives ExpoCast program (exposure parallel to ToxCast)–Hard to measure exposure in HT–Need to model

18


HTRA Summary

1. Select toxicity-related pathways2. Develop assays to probe them3. Estimate concentration at which pathway is “altered” (PD)4. Estimate in vitro to in vivo PK scaling5. Estimate PK and PD uncertainty and variability6. Combine to get BPAD distribution and health protective

exposure limit estimate (BPADL)

• Many (better) variants can be developed for each step (1-6)• Use for analysis and prioritization of data-poor chemicals

19


Select doses for toxicity testing

Relate real-world exposures with toxicity pathway perturbations

Translate in vitro results for risk assessment

Exposure

HumanEnvironment

Biomonitoring

Population

N

N N

NH NH

Cl

Products

Sources

Chemicals

HostSusceptibility

Biotransformation

Rapid Prioritization

ExposomeInformatics Approaches

Network Models

Knowledge Systems

Mechanistic Models

Data Repositories

N

N N

NH NH

Cl

Toxicity endpoints

In vivo bioassays

HTS assays

Distribution/Fate

ExpoCast


High-throughput Exposure Estimates

• Hard, because most chemicals are not BPA, dioxin-like in their data-richness

• Data–Use class, products–Measured data

• Pesticide residues• Water / air / monitoring

–Captured in ACToR (http://actor.epa.gov)• Models

–USEtox–RAIDAR –MENTOR –GxE FRAMES

21


Exposure-Based Prioritization Model Challenge

• Improve understanding of existing prioritization schemes or models for evaluating potential exposures.

• Participating teams/models–USEtox–RAIDAR –MENTOR –GxE FRAMES

• Models are being applied to evaluate common list of 52 chemicals• High interest to EPA• Relatively data rich • Do not span the full range of potential exposure

22


http://www.usetox.org/

“The USEtoxTM model is an environmental model for characterization of human and ecotoxic impacts in Life Cycle Impact Assessment and for comparative assessment and ranking of chemicals according to their inherent hazard characteristics.”

Rosenbaum, R.K. et al., (2008) Int J Life Cycle Assess DOI 10.1007/s11367-008-0038-4

Predicts increase in eleven compartments (five global, five continental, and urban air) due to additional 1 kg/day emitted

Steady-state model

Far Field Exposure (Fate and Tranport) Models: USEtox

http://www.usetox.org/�


Far Field Exposure (Fate and Transport) Models:RAIDAR

Risk Assessment, IDentification, And Ranking (RAIDAR) model

Steady-state model

Arnot, J.A., Mackay, D., and Webster, E. Environ. Sci. Technol. (2006) 40 (7), pp 2316–2323


LRI RfP: Developing Exposure Indices for Rapid Prioritization of Chemicals in Consumer Products:

• Leverage the best existing exposure models–Derive exposure classification indices

• Consider multiple metrics to cover important aspects of exposure space and product lifecycle:

– Physical -Chemical properties– Product characteristics (manufacture, formulation, use, lifecycle)– Emission characteristics (indoor/outdoor, media of release, amount available for

release/contact)– Pathways (media, routes)– Scale (far-field, near-field)– Target characteristics (individual, population, lifestage, lifestyle, susceptibility)– Dosimetry (ADME)

• Demonstrate application ~100-1000 chemicals

25


HTRA Prioritization Scheme

26

Project-relevant Chemicals

Adversity-related HTS Assays

Positive(few?)

Negative(many?)

Exposure EstimatesRTK

Overlap?

Exposure Estimates (?)

Targeted TestingYes or close

Highexposurepotential (?)

Computational Toxicology and High-Throughput Risk …...All of these target-cancer links are backed...

Documents

Transcript of Computational Toxicology and High-Throughput Risk …...All of these target-cancer links are backed...