ToxCast and Tox21: High Throughput Screening for Hazard & Risk ...
ToxCast and Tox21: High Throughput Screening for Hazard & Risk ...
ToxCast and Tox21: High Throughput Screening for Hazard & Risk ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>ToxCast</strong> <strong>and</strong> <strong>Tox21</strong>: <strong>High</strong> <strong>Throughput</strong> <strong>Screening</strong><br />
<strong>for</strong> <strong>Hazard</strong> & <strong>Risk</strong> of Environmental Chemicals<br />
Office of Research <strong>and</strong> Development<br />
National Center <strong>for</strong> Computational Toxicology<br />
David Dix<br />
NLSOT, 07oct2010<br />
This work was reviewed by EPA <strong>and</strong> approved <strong>for</strong> presentation but does not necessarily reflect official Agency policy.
Future of Toxicity Testing <strong>and</strong> Ultimately,<br />
Environmental <strong>Risk</strong> Assessments<br />
Science, February 2008<br />
EPA strategy, March 2009<br />
CTRP 2 nd gen plan, Sept 2009<br />
Science, , August g 21, , 2009<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program 1
Exposure p<br />
Gr<strong>and</strong> Challenge <strong>for</strong> Computational Toxicology:<br />
PPredicting di ti HHuman TToxicity i it<br />
Tissue<br />
Dose<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Complex<br />
Cellular <strong>and</strong><br />
HCS HTS<br />
�<br />
Biochemical<br />
HTS<br />
Tissues<br />
Cellular Systems<br />
Molecular<br />
Targets<br />
Cell<br />
Changes<br />
Cellular<br />
Networks<br />
�<br />
Molecular<br />
Pathways<br />
��<br />
Cell-Based<br />
HTS<br />
��<br />
Model<br />
Organism<br />
MTS<br />
Virtual Tissues<br />
ToxRefDB<br />
�<br />
2<br />
Toxicity
10000<br />
1000<br />
Too Many Chemicals Too Little Data (%)<br />
100<br />
10<br />
1<br />
9912<br />
IRIS TRI Pesticides<br />
Inerts<br />
MPV<br />
CCL 1 & 2 HPV<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
Acute Cancer Gentox<br />
Dev Tox Repro Tox<br />
3
4<br />
Robotic Plat<strong>for</strong>m<br />
Pathway<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
<strong>High</strong> <strong>Throughput</strong> <strong>Screening</strong> 101<br />
(HTS)<br />
96-, 96 ,38 384-, , 1536 536 Well e Plates ates<br />
Target Biology (e.g.,<br />
Estrogen Receptor)<br />
Chemical Exposure<br />
Cell Population<br />
4<br />
HTS: <strong>High</strong> <strong>Throughput</strong> <strong>Screening</strong>
HTS in Drug Development<br />
HTS tests<br />
Make<br />
Id Identify tif CCreate t >100,000 100 000 modifications difi ti<br />
target, testing chemicals with to most active<br />
pathway,<br />
or cellular<br />
phenotype h t<br />
system<br />
(aka,<br />
“ “assay”) ”)<br />
no known<br />
activity<br />
<strong>for</strong> effect on<br />
target<br />
chemicals to<br />
make suitable<br />
f<strong>for</strong> in i vivo i<br />
testing<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Test in<br />
animals <strong>for</strong><br />
safety,<br />
effectiveness<br />
Test in<br />
humans <strong>for</strong><br />
safety,<br />
effectiveness<br />
effectiveness effectiveness<br />
5
HTS iin TToxicology i l<br />
Test prioritized<br />
chemicals in<br />
animals<br />
Categorize as<br />
inactive subject<br />
to further testing<br />
Computational<br />
analysis<br />
&<br />
Synthesis<br />
of HTS results<br />
HTS tests<br />
chemicals<br />
<strong>for</strong> effect<br />
on assays<br />
Obt Obtain i or<br />
create<br />
testing<br />
systems<br />
(“ (“assays”) ”)<br />
Id Identify tif<br />
toxicity<br />
pathways,<br />
cellular<br />
phenotypes h t<br />
Chemicals C e cas<br />
with known<br />
or<br />
suspected<br />
toxicity/ y<br />
activity<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
6
<strong>ToxCast</strong> <strong>and</strong> <strong>Tox21</strong> <strong>High</strong> <strong>Throughput</strong> <strong>Screening</strong><br />
of Chemical Bioactivity<br />
• Addresses chemical screening <strong>and</strong> prioritization<br />
needs <strong>for</strong> chemicals regulated by EPA<br />
• Comprehensive use of HTS technologies to<br />
generate biological fingerprints <strong>and</strong> predictive<br />
signatures<br />
• Committed to stakeholder involvement <strong>and</strong><br />
transparency p y<br />
• Communities of Practice- Chemical Prioritization;<br />
Exposure<br />
• Release of all data upon peer review publication<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
7
GGoals l of f T<strong>ToxCast</strong> C t <strong>and</strong> dT <strong>Tox21</strong> 21<br />
• Identify y toxicity yppathways y<br />
• Obtain HTS assays <strong>for</strong> pathways<br />
• Screen a large chemical library<br />
• Initially link HTS results to adverse effects<br />
- Toxicity signatures<br />
• Ultimately identify points of departure from HTS data<br />
- Toxicity pathways<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
8
96-well plate<br />
<strong>Tox21</strong> <strong>Screening</strong> g <strong>Throughput</strong> g p<br />
� 8 rows x 12 columns<br />
� 88 test samples<br />
384-well plate<br />
4 x 96-well plates<br />
If @ 100 microtiter plates per day:<br />
samples Time to screen<br />
1 MM samples<br />
§ Plate <strong>for</strong>mat samples /day Time to screen<br />
(wells/day) 1 MM samples<br />
§ Plate <strong>for</strong>mat<br />
/day<br />
(wells/day)<br />
96-well<br />
384-well<br />
1,536-well<br />
8,800 (9,600)<br />
35,200 (38,400)<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
140,800 (153,600)<br />
4 months<br />
4 weeks<br />
7 days<br />
� 16 rows x 32 columns<br />
� 352 test samples<br />
1536-well plate<br />
16 x 96-well plates<br />
� 32 rows x 48 columns<br />
� 1,408 test samples<br />
§ wells remaining after subtraction of control wells;<br />
NCGC uses left 4 columns of a 1536-well plate <strong>for</strong><br />
control wells<br />
9
Comparison p of Volumes<br />
<strong>for</strong> <strong>Screening</strong> Compounds in 7 Concentrations<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Total<br />
VVolume l<br />
96-well plate: 100 �l x 7 pts = 700 �l<br />
384-well plate: 40 �l x 7 pts = 280 �l<br />
1536-well plate: 5 �l x 7 pts = 35 �l<br />
10
• 1536 well HTS<br />
• 10,000 chemicals<br />
• 25 assays per year<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
11
<strong>ToxCast</strong><br />
Phase I<br />
<strong>ToxCast</strong> Phase I <strong>and</strong> II, <strong>Tox21</strong><br />
Number of Chemicals<br />
<strong>ToxCast</strong><br />
Phase II <strong>Tox21</strong><br />
Actives 272 120 700<br />
Inerts 24 100 1000<br />
Antimicrobials 33 100 500<br />
HPV 35 170 1300<br />
MPV 7 60 1500<br />
Green 4 60 500<br />
PCCL 73 150 500<br />
Pharmaceuticals 0 100 2500<br />
Consumer<br />
Products /Food<br />
additives 0 0 1500<br />
Total 309 ~700 ~10000<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program 12
Biochemical Assays<br />
• Protein families<br />
<strong>ToxCast</strong> HTS Assays<br />
~500 Total Endpoints<br />
Cellular Assays<br />
• Cell lines<br />
– HepG2 human hepatoblastoma<br />
– GPCR<br />
– A549 human lung carcinoma<br />
– HEK 293 human embryonic y kidney y<br />
– NR<br />
– Kinase<br />
• Primary cells<br />
– Human endothelial cells<br />
– Phosphatase<br />
– Human monocytes<br />
– Protease<br />
– HHuman kkeratinocytes ti t<br />
– Other enzyme<br />
– Ion channel<br />
– Transporter<br />
– Human fibroblasts<br />
– Human proximal tubule kidney cells<br />
– Human small airway epithelial cells<br />
– Rat hepatocytes p y<br />
– Mouse embryonic stem cells (Sid Hunter)<br />
• Assay <strong>for</strong>mats<br />
– Radiolig<strong>and</strong> binding<br />
– Enzyme activity<br />
– Co-activator recruitment<br />
Primarily Human / Rat<br />
EException: ti ZZebrafish b fi h ddevelopment l t (St (Stephanie h i Padilla)<br />
P dill )<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
• Biotrans<strong>for</strong>mation competent cells<br />
– Primary rat hepatocytes<br />
– Primary human hepatocytes<br />
• Assay <strong>for</strong>mats<br />
– Cytotoxicity<br />
– Reporter gene<br />
– Gene expression<br />
13<br />
– Biomarker production<br />
– <strong>High</strong>-content imaging <strong>for</strong> cellular phenotype
Assays<br />
Chemicals<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
The <strong>ToxCast</strong> In Vitro Data Set<br />
Many hits – median value=50<br />
Fewer in cell-free HTS<br />
(Novascreen, red)<br />
Many hits are at or near top of<br />
tested concentration range<br />
A few are in nanomolar range<br />
14
…<br />
4<br />
Emax<br />
3 2 1 0<br />
Concentration<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
<strong>ToxCast</strong> Data Analysis<br />
AC50<br />
Chemicalls<br />
Assays<br />
828 Assay-Chemical Pairs<br />
had AC50s of less than 1µM<br />
: Assay-Chemical Hit<br />
500 Assays X 320 Chemicals X {5-18} Concentrations X<br />
{1-3} Replicates X {1-3} Time Points ≈<br />
3.2 Million Data Points<br />
Emax: Maximal efficacy/response/activity<br />
15<br />
AC50: Concentration whereby 50% of maximal response was achieved<br />
Conc (uM): Micromolar concentrations of chemical
micals<br />
Chem<br />
Digitizing Legacy in Vivo Data in ToxRefDB<br />
Chronic/Cancer<br />
Multigenation<br />
Developmental<br />
Martin et al 2009a,b<br />
Knudsen et al 2009<br />
30 years <strong>and</strong> more than $2B worth of data 16<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program 16<br />
SOT 2010
Predicting Toxicity <strong>and</strong> Disease<br />
from HTS Data<br />
ToxRefDB <strong>ToxCast</strong> Human Disease<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
17
<strong>ToxCast</strong>: Multiple Assays <strong>and</strong> Technologies per Target<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
18
Biologically Multiplexed<br />
Activity Profiling (BioMAP)<br />
Multiplex Transcription<br />
Reporter Assay<br />
Cell-based HTS Assays<br />
Cell-free HTS Assays<br />
<strong>High</strong> Content Cell Imaging<br />
Assays<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
<strong>ToxCast</strong>: Multiple Targets per Pathway<br />
19
No Pathology<br />
Proliferative Lesions<br />
Pre-neoplastic Lesions<br />
Neoplastic Lesions<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Building Toxicity Signatures <strong>for</strong><br />
RRat t Li Liver Histopathology<br />
Hi t th l<br />
from Chronic Bioassays<br />
37<br />
68<br />
21<br />
N = 248 Chemicals in ToxRefDB<br />
20
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Judson et al, EHP (2010)<br />
<strong>ToxCast</strong> Identied Genes Associated with<br />
Progression of Rat Liver Disease<br />
Any Lesion Pre-Neoplastic Neoplastic<br />
132 60 21<br />
Liver Disease Progression<br />
21
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Toxicity Pathways<br />
Chemical<br />
Receptors p / Enzymes y / etc.<br />
Direct Molecular Interaction<br />
Pathway Regulation /<br />
Genomics<br />
Cellular Processes<br />
Tissue / Organ / Organism Tox Endpoint<br />
22
Predictive Signatures from <strong>ToxCast</strong> <strong>for</strong> Chronic,<br />
Developmental p <strong>and</strong> Reproductive p Toxicityy<br />
• Chronic/Cancer endpoints from rat, mouse <strong>and</strong> dog<br />
• Developmental endpoints from rat <strong>and</strong> rabbit<br />
• Reproductive endpoints from rat<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
23
ToxPi: Prioritizing Chemicals <strong>for</strong><br />
PPotential t ti l Endocrine E d i Activity A ti it<br />
Prioritization Index = ToxPi = f(HTS ( assays y + Chemical properties p p + Pathways) y )<br />
Bisphenol A Tebuthiuron<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Other<br />
XME/ADME<br />
Other NR<br />
LogP_TPSA<br />
Predicted<br />
CaCO-2<br />
Reif et al, EHP, 2010<br />
TR<br />
KEGG<br />
pathways<br />
ER<br />
AR<br />
Disease<br />
classes<br />
Ingenuity<br />
pathways
ToxPi score<br />
309 3 Chemicaals<br />
sorted by<br />
lowest<br />
Linuron<br />
Tebuthiuron<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
ToxPi Ranking of Chemicals<br />
ToxPi<br />
Pyrimethanil<br />
ToxScore<br />
hi highest h<br />
Other<br />
XME/ADME<br />
Other<br />
NR<br />
LogP_TPSA<br />
CaCO-2<br />
HPTE<br />
TR<br />
KEGG<br />
path<br />
ER<br />
AR<br />
Disease<br />
classes<br />
Ingenuity<br />
path<br />
Reif et al, 2010
Exp<strong>and</strong>ing the ToxPi Approach <strong>for</strong> Prioritization of Toxicity<br />
Testing Based on <strong>ToxCast</strong> Chemical Profiling<br />
DEVEL<br />
CANCER<br />
PPathways th<br />
associated with<br />
CANCER in vivo<br />
endpoints<br />
In vitro assays<br />
associated with<br />
CANCER in vivo<br />
endpoints<br />
REPRO SYSTEMIC<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Identify in vitro assays,<br />
targets, g ,ggenes<br />
<strong>and</strong><br />
pathways associated<br />
with multiple sectors<br />
of in vivo toxicity toxicity.<br />
26
ToxPi Scores <strong>for</strong> Antimicrobial Pesticide Actives<br />
One of the methods be<br />
developed to use <strong>ToxCast</strong><br />
data <strong>for</strong> classifying <strong>and</strong><br />
prioritizing antimicrobials<br />
<strong>and</strong> inerts.<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
27
Future ToxPi (Toxicological Prioritization Index)<br />
ToxPi = f(Exposure + Chemical properties + In vitro assays + Pathways)<br />
Incorporate additional components<br />
(slices) ( ) from other domains:<br />
- Exposure<br />
- Chemical properties<br />
- QSAR<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
28
log (mg/kgg/day)<br />
Pyrithiobac-sodium<br />
Triclosan<br />
Range of in vitro AC50<br />
values l converted t d tto s<br />
human in vivo daily dose<br />
Margin<br />
Estimated Exposure<br />
Etoxxazole<br />
Emammectin<br />
Buproofezin<br />
Dibutyl phthhalate<br />
Pyraclosttrobin<br />
Paraathion<br />
Isoxxaben<br />
Pryrithiobac-soodium<br />
Bentaazone<br />
Propetammphos<br />
2,4-D<br />
S-Bioalleethrin<br />
MGK<br />
Atraazine<br />
Broomacil<br />
Fenoxxycarb<br />
Forchlorfennuron<br />
Methyl Paraathion<br />
Tricclosan<br />
Roteenone<br />
Cyprrodinil<br />
Isoxafflutole<br />
Acetammiprid<br />
Zoxaamide<br />
Diuron D<br />
Benssulide<br />
Vincloozolin<br />
Oxytetracyclinne<br />
DH<br />
Dicrotoophos<br />
Metribuzin<br />
Triadimmefon<br />
Thiaazopyr<br />
Fenamiphos<br />
Clothiaanidin<br />
Bispheenol-A<br />
Alaachlor<br />
Acetoochlor<br />
Diazzoxon<br />
Dichlorvos<br />
Chlorpyriphos--oxon<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Rotroff, et al. Tox.Sci. In Press<br />
29<br />
Reverse<br />
Tooxicokkinetics
Systems Approaches to Modeling Toxicity:<br />
From Pathways to Virtual Tissues<br />
chemicals pathways networks cell states tissue function<br />
Moving beyond<br />
empirical models, to<br />
multi-scale models of<br />
complex biological<br />
systems: vLiver,<br />
vEmbryo…<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
Identify Key Targets <strong>and</strong> Pathways<br />
For Prioritization<br />
Quantitative<br />
Dose-Response<br />
Models<br />
Next Generation<br />
<strong>Risk</strong> assessments<br />
30<br />
SOT 2010
Exposure<br />
HIGH THROUGHPUT SCREENING <strong>and</strong><br />
CHARACTERIZATION OF HAZARD & RISK<br />
Tissue<br />
Dose<br />
Molecular<br />
Targets<br />
Molecular<br />
Pathways<br />
Tissues<br />
Cellular Systems<br />
Cell<br />
Changes<br />
Cellular<br />
Networks<br />
IN VITRO: concentration response<br />
Toxicity<br />
Human IN SILICO: exposure dose response<br />
Animal IN VIVO: exposure dose response<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program 31<br />
SOT 2010
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program<br />
<strong>ToxCast</strong> Phase I Publications & Data Release<br />
32<br />
http://epa.gov/ncct/toxcast/<br />
SOT 2010
Robert Kavlock<br />
Keith Houck<br />
Matt Martin<br />
Richard Judson<br />
Ann Richard<br />
David Reif<br />
Daniel Rotroff<br />
Woody Setzer<br />
Holly Mortenson<br />
AAndrew d BBeam<br />
Acknowledgements<br />
Tom Knudsen<br />
Imran Shah<br />
JJohn h WWambaugh b h<br />
Ray Tice<br />
Chris Austin<br />
http://www.epa.gov/ncct/<br />
Office of Research <strong>and</strong> Development<br />
Computational Toxicology Research Program 33<br />
SOT 2010