Computational Toxicology: New Approaches for the 21st Century

September 9th, 2009  Session IV: ToxCast and the Comparative ToxicogenomicsDatabase (CTD)

David Dix, Acting Deputy Director of EPA/ORD’s National Center for Computational Toxicology

Carolyn Mattingly, Mount Desert Island Biological Laboratory

1

September 9, 2009

1

David DixActing Deputy Director, EPAs National Center for Computational Toxicology

ToxCast- Screening and Prioritization of Environmental Chemicals Based on Bioactivity Profiling and Predictions of Toxicity

NIEHS Risk e Learning Sept 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

This work was reviewed by EPA and approved for presentation but does not necessarily reflect official Agency policy.

2

September 9, 2009

2

Office of Research and DevelopmentNational Center for Computational Toxicology

Predicting Toxicity Will Not Be Easy

Chemical

September 9, 2009 3

3

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

Key Challenges Of Pathway Profiling

•Find the Toxicity Pathways•Hepato vs developmental nuerotoxicity

•Obtain HTS Assays for Them• Including metabolic capability

•Screen Chemical Libraries• Coverage of p-chem properties

•Link Results to in vivo Effects• Gold standard and dosimetry

September 9, 2009 4

4

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

ToxCastTM Background

• Research program of EPA’s National Center for Computational Toxicology

• Addresses chemical screening and prioritization needs for pesticidal inerts, anti-microbials, CCLs, HPVs and MPVs

• Comprehensive use of HTS technologies

• Coordinated with NTP and NHGRI/NCGC via Tox21

• Committed to stakeholder involvement and public release of data

• Chemical Prioritization Community of Practice

• NCCT website- http://www.epa.gov/ncct/

September 9, 2009 5

5

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

The ToxCast_320

Classes with > 3 chemicals

Misc MOA classes with3 or fewer representatives

Acetylcholine esterase inhibitorsconazole fungicidesSodium channel modulatorspyrethroid ester insecticidesorganothiophosphate acaricidesdinitroaniline herbicidespyridine herbicidesthiocarbamate herbicidesimidazolinone herbicidesorganophosphate insecticidesphenyl organothiophosphate insecticidesaliphatic organothiophosphate insecticidesamide herbicidesaromatic fungicideschloroacetanilide herbicideschlorotriazine herbicidesgrowth inhibitorsorganophosphate acaricidesoxime carbamate insecticidesphenylurea herbicidespyrethroid ester acaricidesstrobilurin fungicidesunclassified acaricidesunclassified herbicides

Misc

309 Unique Structures

Replicates for QC

291 Pesticide Actives9 Industrial Chemicals8 Metabolites

56/73 Proposed Tier 1 EDSP

53 of 80 with DNTs

122 in IRIS

14 HPV11 HPV Challenge

6

6

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

ToxRefDB

• Relational phenotypic/toxicity database

• Provides in vivo anchor for ToxCast predictions

• Three study types• Chronic/Cancer Rat and Mouse (Martin, et al, EHP 2008)

• Rat multigenerational Reproduction (Martin, et al, 2009)

• Rat & Rabbit Developmental Toxicity (Knudsen, et al, 2009)

• Two types of synthesis• Supervised (common individual phenotypes)

• Unsupervised (machine based clustering of phenotype patterns)

September 9, 2009 7

7

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology SOURCE: Matt Martin, NCCT, 2009

A = Rat B = Mouse C = Rabbit

CHRONIC/CANCER (CHR)Martin et al. (2008) Environ Hlth Perspdoi:10.1289/ehp.0800074

PRENATAL DEVELOPMENTAL (DEV)Knudsen et al. (2009) Reprod Toxicoldoi: 10.1016/j.reprotox.2009.03.016

MULTIGENERATION REPRODUCTIVE (MGR)Martin et al. (2009) Toxicol Scidoi: 10.1093/toxsci/kfp080

ToxRefDB Endpoint Coveragedata evaluation records ToxRefDB

September 9, 2009 8

8

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

Rat Chronic Bioassay Results

Martin, et al EHP, 2009September 9, 2009 9

September 9, 2009

9

Office of Research and DevelopmentNational Center for Computational Toxicology

Digitizing Legacy in Vivo Data in ToxRefDB

Chronic/CancerMultigenationDevelopmental

Che

mic

als

30 years and more than $2B worth of data

Martin et al 2009a,bKnudsen et al 2009

September 9, 2009 10

10

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

ToxRefDB in Predictive Modeling

STRENGTHS– Source data from >2,000 guideline studies– Puts >$2B worth of legacy data into a computable form– in vivo database anchoring HTS in vitro assays– Enables comparison of endpoint incidence between species– Searchable database will be public (www.epa.gov/ncct/toxrefdb/)

LIMITATIONS– Endpoints aggregated as independent features– Data largely qualitative (LELs, LOAELS)– Not all ToxCast™ chemicals represented in ToxRefDB– Not all ToxRefDB chemicals represented in ToxCast™– Species dimorphism may link to biology or study design–Limited mode of action information available in source DERs

September 9, 2009 11

11

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

ToxCast Assays

• Cell lines– HepG2 human hepatoblastoma– A549 human lung carcinoma– HEK 293 human embryonic kidney

• Primary cells– Human endothelial cells– Human monocytes– Human keratinocytes– Human fibroblasts– Human proximal tubule kidney cells– Human small airway epithelial cells

• Biotransformation competent cells– Primary rat hepatocytes– Primary human hepatocytes

• Assay formats– Cytotoxicity– Reporter gene – Gene expression– Biomarker production– High-content imaging for cellular

phenotype

• Protein families– GPCR– NR– Kinase– Phosphatase– Protease– Other enzyme– Ion channel– Transporter

• Assay formats– Radioligand binding– Enzyme activity– Co-activator recruitment

Cellular AssaysBiochemical Assays

467 Endpoints

12

12

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

Confidence Builders:Some Expected Results…

• Estrogen receptor (ER)–Bisphenol A, Methoxychlor, HPTE

• Androgen Receptor (AR)–Vinclozolin, Linuron, Prochloraz

• PPAR –PFOA, PFOS, Diethylhexyl Phthalate, Lactofen

• Mitochondrial Poisons–Azoxystrobin, Fluoxastrobin, Pyraclostrobin

• Acetylcholinesterase Inhibition–Multiple organophosphorus pesticides

September 9, 2009 13

13

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

Confidence Builders (2):Multiple Assays and Technologies per Target

September 9, 2009 14

September 9, 2009

14

Office of Research and DevelopmentNational Center for Computational Toxicology

Biologically Multiplexed Activity Profiling (BioMAP)

Cell-free HTS Assays

Cell-based HTS Assays

Multiplex Transcription Reporter Assay

High Content Cell Imaging Assays

Confidence Builders (3):Pathway Based Analysis

September 9, 2009 15

September 9, 2009

Office of Research and DevelopmentNational Center for Computational Toxicology

Data Analysis:What is a hit?

16

September 9, 2009

16

Office of Research and DevelopmentNational Center for Computational Toxicology

Biochemical HTS from Novascreen

hCYP 2C9 hERαrAdrRa2B

hLynAActivator

hM1hKATPase

17

September 9, 2009

17

Office of Research and DevelopmentNational Center for Computational Toxicology

qHTS from the NCGC on NRs

ERα PPARγ

September 9, 2009 18

18

Office of Research and DevelopmentNational Center for Computational Toxicology

trans: ERa

cis: ERE

Bisphenol A HPTE

Attagene: cis and trans Assays

September 9, 2009 19

September 9, 2009

19

Office of Research and DevelopmentNational Center for Computational Toxicology

CellzDirect: Data ExamplesCYP1A1-AhR CYP2B6-CARHMGCS2-PPARα

September 9, 2009 20

20

Office of Research and DevelopmentNational Center for Computational Toxicology

ToxCast Phase I Assay Hits (n=624 measurements)

Judson et al, submitted

Cell Free HTSMultiplexed TFHuman BioMapHCSqNPAsXMEsImpedanceGenotoxicity

828 Assay-Chemical Pairshad AC50s of less than 1µM

September 9, 2009 21

September 9, 2009

21

Office of Research and DevelopmentNational Center for Computational Toxicology

Minimum Human Pathways in ToxCast Phase I

Judson et al, submitted

September 9, 2009 22

22

Office of Research and DevelopmentNational Center for Computational Toxicology

“Hits” per ChemicalAs a Function of AC50/LEC Cutoff

Judson et al, submitted

9 Chemicals have at least 20 hits at an AC50 of <30µM

September 9, 2009 23

September 9, 2009

23

Office of Research and DevelopmentNational Center for Computational Toxicology

“Hit” Distribution for Chemical Classes Against 33 Minimal Pathways

(at least 10 chemicals per class)

Judson et al, submitted

September 9, 2009 24

24

Office of Research and DevelopmentNational Center for Computational Toxicology

Judson et al, submitted

Activity of ConazolesAgainst Minimal Pathway Set

September 9, 2009 25

September 9, 2009

25

Office of Research and DevelopmentNational Center for Computational Toxicology

Rat Liver Histopathologyfrom Chronic Bioassays

68

37

21

No PathologyProliferative LesionsPre-neoplastic LesionsNeoplastic Lesions

N = 248 Chemicals

Martin et al, 2009 26

26

Office of Research and DevelopmentNational Center for Computational Toxicology

Rat Liver Tumor Correlations

Fisher’s Exact test, p<0.01 Judson, et al, submittedSeptember 9, 2009 27

27

Office of Research and DevelopmentNational Center for Computational Toxicology

Gene Networks Associated with Progression of Rat Liver Tumor Endpoints

Judson et al, submittedSEVERITYSeptember 9, 2009 28

September 9, 2009

28

Office of Research and DevelopmentNational Center for Computational Toxicology

Some Challenges Faced or to be Faced

• Organizing the chemical library• Quality control of the chemical library

– Acceptable purity, stability• Defining concentration response ranges to the assayed• Definition/Calculation of a hit

– Minimum fold change; minimum r-squared; limit on Hill function• Assay performance

– Replicates, artifacts• Sufficient coverage of biological pathways

– Including those that represent tissue level processes• Incorporation of metabolic competency• Establishment of target prediction

– Pathway perturbation– Rodent bioassay data– Rodent mechanistic studies– Human effects

• Sufficient representation of positives to predict against29

29

Office of Research and DevelopmentNational Center for Computational Toxicology

FY08 FY09 FY10 FY11 FY12

Proof of Concept: ToxCastVerification/Extension

Reduce to Practice

Tox21

Prioritization Product Timeline

FY07

FY09 -10~$15 -20K>200PMN Nanomaterials>12IId

FY09~$20 -25k>400ExtrapolationKnown Human Toxicants>100IIb

FY09$10K166PilotNanomaterials15Ib

FY09~$20 -25k>400ValidationData Rich Chemicals>300IIa

>300

>400

552

Number of Assays

Data poor

Expanded Structure and Use Diversity

Data Rich(pesticides)

Chemical Criteria

FY11 -12

FY10

FY08

TargetDate

~$15 -20k

~$20 -25k

$20k

Cost per Chemical

Prediction and Prioritization

Extension

Signature Development

PurposeNumber of ChemicalsPhase

ThousandsIII

>300IIc

320I

FY -~$15 -20K>200PMN Nanomaterials>12IId

FY09-11~$20 -25k>400ExtrapolationKnown Human Toxicants>100IIb

FY09$10K166PilotNanomaterials15Ib

FY09-11~$20 -25k>400ValidationData Rich Chemicals>300IIa

>300

>400

552

Number of Assays

Data poor

Expanded Structure and Use Diversity

Data Rich(pesticides)

Chemical Criteria

FY11 -12

FY09-11

FY07-09

TargetDate

~$15 -20k

~$20 -25k

$20k

Cost per Chemical

Prediction and Prioritization

Extension

Signature Development

PurposeNumber of ChemicalsPhase

ThousandsIII

>300IIc

320Ia

FY09-11

FY10-11

30

September 9, 2009

30

Office of Research and DevelopmentNational Center for Computational Toxicology

Phase II Plans• Done in conjunction with Tox21 10k Library

–Subset of 700 will seed Phase II• Chemical Diversity

–More food use pesticides –Failed pharmaceuticals (preclinical and clinical)– “Green” chemicals–HPV Categories–Liver toxicants –OECD Molecular Screening Group nominations

• Evaluation of Phase I Assays• Addition of new assays via competitive procurements• Timing

–Chemical procurement completed 4thQ FY09–Launch of Assays, 1st Q FY10–Results Available early FY11

31

31

September 9, 2009

32

Carolyn J. MattinglyThe Mount Desert Island Biological Laboratory

Salisbury Cove, Maine

The Comparative Toxicogenomics Database (CTD)

32

September 9, 2009

33

Helping scientists explore the etiologies of environmental diseases

• What diseases are associated with arsenic?

• Arsenic affects which genes and proteins?

• What biological processes are affected by arsenic?

• Which molecular pathways are affected by arsenic?

• Which other chemicals affect the same molecular pathways?

• Which diseases are implicated with other chemicals?

33

September 9, 2009

34

DiseasesGenes

Chemicals

Curated Data

34

September 9, 2009

35

DiseasesGenes

Chemicals

chemical-diseaserelationships

chemical-geneinteractions

gene-diseaserelationships

Curated Data

35

September 9, 2009

36

DiseasesGenes

Chemicals

Curated Data

MeSH (modified)

Entrez-Gene MeSH/OMIM

CTD interactions

chemical-diseaserelationships

chemical-geneinteractions

gene-diseaserelationships

36

September 9, 2009

37

DiseasesGenes

Chemicals

chemical-diseaserelationships

chemical-geneinteractions

gene-diseaserelationships

181,150

9,179

5,757

Curated Data

37

September 9, 2009

38

38

September 9, 2009

39

39

September 9, 2009

40

40

September 9, 2009

41

41

September 9, 2009

42

42

September 9, 2009

43

43

September 9, 2009

44

44

September 9, 2009

45

45

September 9, 2009

46

46

September 9, 2009

47

47

September 9, 2009

48

48

September 9, 2009

49

Arsenates

HMOX1 AlzheimerDisease

HMOX1

Inferred chemical-disease relationships

49

September 9, 2009

50

Arsenates

HMOX1 AlzheimerDisease

HMOX1

Inferred chemical-disease relationships

50

September 9, 2009

51

51

September 9, 2009

52

52

September 9, 2009

53

53

September 9, 2009

54

54

September 9, 2009

55

55

September 9, 2009

56

56

September 9, 2009

57

Chem/Gene Comps

57

September 9, 2009

58

Tools

58

September 9, 2009

59

Tools

59

September 9, 2009

60

Tools

60

September 9, 2009

61

Tools: VennViewer

61

September 9, 2009

62

Tools: VennViewerInteracting Genes/Proteins

Pathways

127 1357118

Folic acid Arsenicals

4 7621

Folic acid Arsenicals

62

September 9, 2009

63

Tools

63

September 9, 2009

64

Tools: MyGeneVenn

64

September 9, 2009

65

0 10 100

65

September 9, 2009

66

0 10 100

64 168920

Array data CTD data

66

September 9, 2009

67

0 10 100

64 168920

Array data CTD data

Mattingly, C. J., T. Hampton, K. Brothers, N. E. Griffin and A. J. Planchart (2009). Perturbation of defense pathways by low-dose arsenic exposure in zebrafish embryos. Environ Health Perspect doi:10.1289/ehp.0900555.

67

September 9, 2009

68

Gohlke, J., R. Thomas, Y. Zhang, M. D. Rosenstein, A. P. Davis, C. Murphy, C. J. Mattingly, K. G. Becker and C. J. Portier (2009). The Genetic And Environmental Pathways to Complex Diseases. BMC Syst Biol.May 5 3:46.

68

September 9, 2009

69

2096Chemicals

213Genes

213Genes Autism

Environmental etiology of autistic disorders

69

September 9, 2009

70

2096Chemicals

213Genes

213Genes Autism

Environmental etiology of autistic disorders

• What do these chemicals have in common?

– Structure

– Regulatory features(e.g., High production, Carcinogen)

– Function(e.g., Associated pathways)

– Other associated diseases(e.g., Neurological)

70

September 9, 2009

71

Next

• Analysis tools and visualization capabilities

• Integration of additional data sets

• Text mining

71

September 9, 2009

72

• Gohlke, J., R. Thomas, Y. Zhang, M. D. Rosenstein, A. P. Davis, C. Murphy, C. J. Mattingly, K. G. Becker and C. J. Portier (2009). The Genetic And Environmental Pathways to Complex Diseases. BMC Syst Biol.May 5 3:46.

• Mattingly, C. J., T. Hampton, K. Brothers, N. E. Griffin and A. J. Planchart (2009). Perturbation of defense pathways by low-dose arsenic exposure in zebrafish embryos. Environ Health Perspect doi:10.1289/ehp.0900555.

• Davis, A. P., C. G. Murphy, C. A. Saraceni-Richards, M. C. Rosenstein, T. C. Wiegersand C. J. Mattingly (2009). Comparative Toxicogenomics Database: a knowledgebase and discovery tool for chemical-gene-disease networks. Nucleic Acids Res 37(Database issue): D786-92.

• Mattingly, C. J. (2009). Chemical databases for environmental health and clinical research. Toxicol Lett. 186(1):62-5.

• Davis, A. P., C. G. Murphy, M. C. Rosenstein, T. C. Wiegers and C. J. Mattingly (2008). The Comparative Toxicogenomics Database facilitates identification and understanding of chemical-gene-disease associations: arsenic as a case study. BMC Med Genomics 1: 48.

Recent CTD references

72

September 9, 2009

AcknowledgementsScientific Curators

Allan Peter Davis, PhDCindy Murphy, PhDCynthia Richards, PhD

Scientific Software Engineers

Michael C Rosenstein, JDThomas Wiegers

System Administrator

Roy McMorran

James L. Boyer, MD (Yale)

Thomas Hampton (Dartmouth)

http://ctd.mdibl.org/

Zebrafish workAntonio Planchart, PhD

Funding

NIEHS and NLM (ES014065 and ES003828)NCRR (RR016463)Contact Us!

cmattin@mdibl.orgctd@mdibl.org 73

74

After viewing the links to additional resources, please complete our online feedback form.

Thank You

Links to Additional Resources

Feedback Form

74

74

September 9, 2009