adverse outcome pathways outcome pathways what are they? what are they good f?for? how can they be...

Adverse Outcome PathwaysAdverse Outcome PathwaysAdverse Outcome PathwaysAdverse Outcome PathwaysWhat are they?What are they?

Wh h d f ?Wh h d f ?What are they good for?What are they good for?How can they be used in risk assessment?How can they be used in risk assessment?

Daniel L. Villeneuve, Ph.D.US EPA, National Health and Environmental Effects Laboratory (NHEERL)*

Mid-continent Ecology Division (MED)

SOT Risk Assessment Specialty Section, Monthly WebinarWednesday April 8 2015

6201 Congdon Blvd, Duluth, MN

Wednesday April 8, 2015.

*The contents of this presentation neither constitute nor, necessarily, reflect US EPA views or policies.

Disclaimers

*The contents of this presentation neither constitute nor, necessarily reflect US EPA views or policiesnecessarily, reflect US EPA views or policies.

*I’m an EcotoxicologistI m an Ecotoxicologist

• I’ll be presenting “eco” examples

• I’m not an expert on human health risk assessment

• I’m interested in learning from youg f y

Introduction

p,p’‐DDE

Contribution of DDT to population declines in sensitive bird populations

Perhaps the most well known incident in wildlife ecotoxicology

Helped spark the environmental movement, and in part the mission of EPA

Introduction

A central challenge for regulatory toxicology

How do we identify the other chemicals that may cause similar adverse effectsHow do we identify the other chemicals that may cause similar adverse effects

Before we see impacts on human health or wildlife populations

Traditional Approach

Avian reproduction study(OPPTS 850.2300; OECD 206)

$>250 000$>250,000 >30 weeks to perform

Introduction

If we understand HOW chemicals cause adverse outcomes

And…biological activities that lead to/are associated with progression toward those AOs

Prostaglandin Ca2+ and HCOProstaglandin

Creates opportunities to use new types of data for hazard id and/or risk‐based decision‐making

gsynthaseinhibition(cox‐1 or 2 and peroxidase)

Ca2+ and HCO3 transport to shell‐gland

lumen impaired

Eggshell thickness reduced

Crushed eggs, breeding failure

E2 concentrations in eggshell

gland mucosa, reduced

gap

High throughput toxicology• > 600 in vitro assays• daysy• ≈ $ 20,000

1

Introduction

“Transform toxicity testing from a system based on whole‐animal i f d d i il i i h d h ltesting to one founded primarily on in vitro methods that evaluate

changes in biologic processes using cells, cell lines, or cellular components, preferably of human origin”

“The vision emphasizes the development of suites of predictive, high‐throughput assays …..”p , g g p y

“The mix of tests in the vision include tests that assess critical mechanistic endpoints involved in the induction of overt toxic effects rather than the effects themselves.”

6

t e se es

Introduction

ToxCastToxCast> 600 assays, >2000 chemicals,

Toxicity Testing in the 21st Century

21st Century Toxicity Testing is here….

We can rapidly and cost effectively generate pathway‐based data

•Activity of 1000s of chemicals in•Activity of 1000s of chemicals in 100s of pathways.

Conceivable that majority of chemicals in commerce could be “tested” within the decade.

1

Example

So What?

•Can we expect this perturbation lead to an adverse outcome?adverse outcome?

Introduction

An Adverse Outcome Pathway (AOP) is a conceptual framework that portrays existing knowledge concerning the An Adverse Outcome Pathway (AOP) is a conceptual framework that portrays existing knowledge concerning the

linkage between a direct linkage between a direct molecular initiating eventmolecular initiating event and an and an adverse outcomeadverse outcome, at a level of biological organization , at a level of biological organization

relevant to risk assessment.relevant to risk assessment.

((AnkleyAnkley et al. 2010, Environ. et al. 2010, Environ. ToxicolToxicol. Chem., 29(3): 730. Chem., 29(3): 730‐‐741.)741.)

ToxicantCellular Responses

OrganResponses

OrganismResponses

PopulationResponses

Macro‐MolecularInteractions

ChemicalProperties

Receptor/LigandInteraction

DNA Binding

Gene activation

Protein production

Altered physiology

Disrupted h i

Lethality

Impaired Development

Toxicant Responses Responses Responses

Structure

Recruitment

ResponsesInteractions

Protein Oxidation Altered signaling

homeostasis

Altered tissue development/

function

Impaired Reproduction

Extinction

• Helps us organize what we know

A d tili th t k l d t t i k b d d i i ki• And utilize that knowledge to support risk‐based decision‐making

Example

Prostaglandin synthaseinhibition(cox‐1 or 2 and peroxidase)


lumen impaired



Prostaglandin E2

concentrations in eggshell

gland mucosa, gap

peroxidase) reduced

Ah-ha

Introduction

Prostaglandin synthaseinhibition(cox‐1 or 2 and

)


lumen impaired



Prostaglandin E2


gland mucosagap

peroxidase) pgland mucosa,

reduced

A set of chemicals for which there may be reason expect egg‐shell thinning

Introduction

Prostaglandin synthaseinhibition(cox‐1 or 2 and

)


lumen impaired



Prostaglandin E2


gland mucosagap

peroxidase) pgland mucosa,

reduced

•Alternative tests: Suggests informative endpoints we may be able to measure more rapidly and cost‐ff ti l i l b t t i it t teffectively in laboratory toxicity tests.

•Biomarkers: Suggests biomarkers we could measure in animals from the environment – early warning; diagnostic

• Particularly if we can translate into a quantitative prediction of probability or severity of AO.p p y y

•Diagnostic potential: Suggests an etiology for observed patterns of biological response – may help trace back to causative agents and/or sourcestrace back to causative agents and/or sources.

Introduction

•HTS is not the only relevant source of mechanistic data

•Toxicology research community has been studying mechanismsToxicology research community has been studying mechanisms underlying toxicity for decades.

•Considerable mechanistic data exists in the literature currently•Considerable mechanistic data exists in the literature – currently under‐utilized for regulatory applications.

• Biomarkers•Omics

Molecular Cellular Tissue Organ Organ System Individual Population Ecosystem

•Survival•Growth/development

•QSAR and Read‐Across•High content imaging• In vitro•Alternative model organisms (e g zebrafish drosophila C elegans)

Growth/development•Reproduction

Human Health Ecosystem Health

Alternative model organisms (e.g., zebrafish, drosophila, C. elegans)

Introduction

AOPs and environmental surveillance and monitoring.

Connecting chemicals to potential hazards

21st Century Regulatory Decision‐Making

The data are there….

We have entire libraries of scientificWe have entire libraries of scientific publications available at our fingertips…..

Scientific challenge of the 21st C….

How do we make effective use of those data and our wealth of existing scientific knowledge to support regulatory decision making?decision‐making?

AOPs are part of the solution.

AOP Concept in Ecotoxicology

SETAC Pellston workshop on biomarkers – identified need for linkages across levels of organization to support use of biomarkers in ERA.1992

Mid 90 Schmieder Bradbury Veith others in ecotox community – concept toMid 90s Schmieder, Bradbury, Veith, others in ecotox community concept to support application of QSARs and biomarkers in ERA.

2004 Bradbury et al. (ES&T Dec. 1, 2004) – publication of the concept as a means to support greater use of in silico and in vitro approaches in risk assessment – termed “Toxicity Pathway”

2004 Schmieder et al. (ES&T 38:6333‐6342) – published a “toxicity pathway” linking ER binding to potential population‐level consequences.

2007 NRC report on Toxicity Testing in the 21st Century – advocated paradigm similar to Bradbury et al – defined “toxicity pathway” as “cellular response pathways that, when sufficiently perturbed, are expected to result in adverse health effects”

20092009, 2010

Use of AOP term at McKim conferences – introduction of AOP terminology into OECD QSAR tool‐box discussions

2010MED working group published definition of “Adverse O h ” d ib li i i d2010 Outcome Pathways”, describe application in Ecotox and Ecological Risk Assessment. ET&C 2010.

Regulatory Application

Prostaglandin Prostaglandin

Until recently, AOP development was an ad hoc process.

Prostaglandin synthaseinhibition(cox‐1 or 2 and peroxidase)


lumen impaired



gE2


gland mucosa, reduced

gap

•When can you trust that the relationships depicted provide a sound/defensible foundation for regulatory decision‐support?/ g y pp

•What is the biological/toxicological applicability domain?•Taxa life stage target organ(s) route of exposure•Taxa, life stage, target organ(s), route of exposure

•How do we present in a systematic and transparent manner?

Mode of Action (MOA) Concept in HH Toxicology

•In many respects synonymous with AOP conceptAOP concept.

•MOA generally applied in a chemical‐specific context (RA)

•Intended to address a specific question (animal to human relevance)

•Formal incorporation of BH considerations

Formalism of the AOP Framework

•2012 launch of OECD AOP development programme•2013 OECD Guidance on Developing and Assessing AOPsp g g

•Conventions and terminology•Information content of an AOP description•Weight of evidence evaluation (adopted from MOA)•Weight of evidence evaluation (adopted from MOA)•Introduce standardization and rigor to AOP development

AOP Development

and Description

Users’ handbook supplement to

dDescription Case Studies OECD guidance

document for developing and assessing AOPs.g


•March 2014 – Advancing AOPs for Integrated Toxicology and Regulatory Applications Workshop

Workshops


Principles of AOP

Development

Weight of evidence evaluationDevelopment evaluation

•Supports for a systematic approach to describing AOPs.

•Evaluates the relative level of confidence/uncertainty in the relationships depicted.

•Provides a rationale for why they’re described that way.

p p

• Informs suitability for different applications.

Principles of AOP Development

1. AOPs are not chemical‐specific•Not trying to describe what a single chemical does

• Trying to describe what ANY chemical that perturbs the MIE with sufficient potency and duration is likely to do‐ Biological motifs of failure

•Describing AOP does not require chemical‐specific information.•Applying those motifs in a predictive context requires understanding chemical‐specific properties (e.g., potency, ADME) that dictate the magnitude and duration of perturbation

h MIEat the MIE.

Volz et al. 2011. Toxicol. Sci. 123: 349‐358Russom et al. 2014. Environ. Toxicol. Chem. 33: 2157‐2169


2. AOPs are ModularTwo Primary Building Blocks

HepatocyteReduced VTG production

OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn

PopulationDeclining Trajectory

AromataseInhibition

GranulosaReduced E2 synthesis

PlasmaReduced

circulating E2

Key EventsKey Events (KEs)

•Functional unit of observation/verification

•Observable ∆ biological state (measurable)•Essential (but not necessarily sufficient)( y )

•Description•Methods for observing/measuring•Methods for observing/measuring

•Taxonomic applicability

Building an AOP

•Molecular initiating event (MIE) – A specialized type of KE that represents the initial point of chemical interaction onthat represents the initial point of chemical interaction, on the molecular level, within an organism, that results in a perturbation that starts the AOP.

•Adverse Outcome (AO) – A specialized type of KE that is generally accepted as being of regulatory significance on the basis of correspondence to an established protection goal or p p gequivalence to an apical endpoint in an accepted regulatory guideline toxicity test.


2. AOPs are ModularTwo Primary Building Blocks


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn


AromataseInhibition


PlasmaReduced

circulating E2

Key Event Relationships (KERs)

Key event down

Key event up

y p ( )

•Functional unit of inference/extrapolation

•Define a directed relationship•Define a directed relationship•State of KEup provides some ability to predict or infer state of KEdown

•Supported by plausibility and evidence•Quantitative understanding


3. AOPs are a pragmatic functional unit of development and evaluationp


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn


AromataseInhibition


PlasmaReduced

circulating E2

•By convention AOP consists of a single sequence of key events connecting an q y gMIE to AO (no branches)

•AOP is a pragmatic simplification of complex biology

•For a “pure ligand” – functional unit of di i One set of directions from point A to pointprediction One set of directions from point A to point

B, not the map of all possible routes


4. For most real‐world applications, AOP networks are the functional unit of predictionp

Diuron

Chemicals with multiple biological activities Exposure to multiple chemicals

ImidaclopridLinuronTerbuthylazine2,4-DMCPPPropachlor ESAButalbitalDiclofenacFurosemideGemfibrozilIbuprofenNaproxenPhenobarbitalPhenytoinySulfamethoxazoleTriclosanAcebutololAlbuterolAmitriptyline

AOPs are not triggered in isolation. They interact.


4. For most real‐world applications, AOP networks are the functional unit of predictionp

Key events shared by multiple AOPs

KERs shared by multiple AOPs

AR

Agonism

Theca/Granulosa

Reduced T & E2 synthesis

Hypothalamic Neurons

(‐) Feedback


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawning


PlasmaReduced

circulating E2


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawning


Aromatase

Inhibition


PlasmaReduced

circulating E2

ER Hepatocyte Ovary Female Population

CYP17, CYP11A

Inhibition

Theca/GranulosaReduced T & E2

Synthesis


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawning


PlasmaReduced

circulating E2

AntagonismReduced VTG production

Impaired Oocyte Dev.

Decreased ovulation/spawning

Declining Trajectory


4. For most real‐world applications, AOP networks are the functional unit of prediction

AR

Agonism

Hypothalamic Neurons

(‐) Feedback

p


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawning


Aromatase

Inhibition


PlasmaReduced

circulating E2

CYP17, CYP11A

Inhibition

•By building modular AOPs, we gradually describeER

Antagonism

By building modular AOPs, we gradually describe the complexity of potential interactions.

•AOPs meet systems biology


•Each AOP is one sequential sequence or path through a broader network of AOPs.

•If properly represented in KB, can be analyzed as a network.


5. AOPs are living documents

•AOPs are a way of organizing existing knowledge•AOPs are a way of organizing existing knowledge

•As methods for observing biology evolve:•New possibilities for KEs•Ability to measure KEs with greater precision/accuracy

•As new experiments are published:•Weight of evidence supporting (or rejecting) KERs grows•New AOPs and new branches in AOP networks discovered•New AOPs and new branches in AOP networks discovered

•There is no objective “complete AOP”


5. AOPs are living documentsOperationally‐defined “stages” of AOP development

Stages of AOP Development Characteristics

Putative AOPs: Hypothesized set of KEs and KERs primarily supported b bi l i l l ibili d/ i i l i f

Increasing

• Depth of Putative AOPs: by biological plausibility and/or statistical inference

Formal AOPs:

Include assembly and evaluation of the supporting weight of evidence – developed in AOP knowledgebase in accordance with internationally harmonized OECD

evidence /understanding

• Transparency /defensibilityin accordance with internationally‐harmonized OECD

guidance

QuantitativeSupported by quantitative relationships and/or computational models that allow quantitative

/de e s b y

• Quantitative precision

C tQuantitative AOPs:

computational models that allow quantitative translation of key event measurements into predicted probability or severity of adverse outcome

• Cost• Data needs

• Time

All t h t ti l tilit•All stages have potential utility•Level of development desired/required depends on the application

AOP ApplicationTh i bj ti “ f l” “ it bl ” AOP•There is no objective “useful” or “suitable” AOP

•AOPs are developed independent of the regulatory applications for which they may be applied.pp

• The suitability of an AOP for any risk assessment or regulatory decision is dependent on the problem formulation or regulatory context.

Data needed for development

Prediction/extrapolation uncertainties

Cost and time required for development

Quantitative risk assessment (high tier,

Quantitative risk assessment (low

Informed approaches to

Screening and prioritization

high impact)tier, low impact)testing and assessment

prioritization

Elements of WOE

Elements of WOE Evaluation Considerations

Biological Plausibility of KERs Are the KERs consistent with current biological understanding

Essentiality of KEs Are the effects are reversible if the stressor is removed? Are the downstream KEs prevented if one or more upstream KEs are blocked? (e.g., knock‐out, over‐expression, pharmacological agents)

Empirical Support for KERs Dose response ‒ Upstream KEs observed at doses below or similar toEmpirical Support for KERs Dose response Upstream KEs observed at doses below or similar to those associated with downstream KEs and the AO?

Temporality ‒ The KEs are observed in hypothesized order?Incidence ‒ The frequency of occurrence of the apical effect less than

that for the key events?Consistency – relationship consistently observed within the relevant

domain of applicability, inconsistencies are explainable.

Quantitative Understanding of the KERs

Extent to which the state of KE downstream and can be precisely inferred from the measured or predicted state of KE upstream.Extent to which relevant modulating factors are understood.

Example


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn


AromataseInhibition


PlasmaReduced

circulating E2

Biological Plausibility

•Aromatase is rate‐limiting for 17β‐estradiol synthesisAromatase Granulosa

CYP17CYP11A

CYP17

Cholesterol

Inhibition Reduced E2 synthesis

Pregnenolone 17‐OH‐Pregnenolone

CYP17 (hydroxylase)

3‐HSD 3‐HSD

DHEA

CYP17 (lyase)

3‐HSDCYP17 CYP17

Progesterone 17‐OH‐Progesterone Androstenedione

17‐HSDCYP19

20‐HSD

CYP17 (hydroxylase)

CYP17 (lyase)

estroneCYP19

17‐HSDCYP21 CYP21

Testosterone 17‐estradiol17α20β‐dihydroxy‐4‐

pregnen‐3‐one

11β‐OH‐Testosterone

CYP11B1

11‐deoxycorticosterone

corticosterone

CYP11B1 11‐deoxycortisol

CYP11B2 Testosterone

11‐Ketotestosterone

11βHSDaldosterone

CYP11B2

cortisol

Biological Plausibility

A bit of fish reproductive biologyEstrogenReceptor

Hepatocyte Ovary

l i

Female

Stable or

Population

Agonism Vtgproduction

Oocytedevelopment

Ovulation& spawning

Stable orincreasing trajectory

OH

OH

ERE‐Vtg

•17β‐estradiol regulates vitellogenin (Vtg) synthesisVt iti l lk•Vtg critical egg yolk precursor

• accounts for up to 95% of egg mass/volume•Reproduction required for stable/increasing pop trajectoryReproduction required for stable/increasing pop. trajectory

1

Forecast Population Trajectories

11

Forecast Population Trajectories

Empirical Support

N y

150

Male Female

a

b

y

150

Male Female

a

bPerturbed fish reproductive biology

0.6

0.8

1

e Po

pula

tion

Size

of C

arry

ing

Cap

acity

)

A

0.6

0.8

1

ulat

ion

Size

rryin

g C

apac

ity)

A 0%

0.6

0.8

1

e Po

pula

tion

Size

of C

arry

ing

Cap

acity

)

0.6

0.8

1

e Po

pula

tion

Size

of C

arry

ing

Cap

acity

)

A

0.6

0.8

1

ulat

ion

Size

rryin

g C

apac

ity)

A 0%

N

Fadrozole

Aro

mat

ase

Act

ivity

(fmol

/mg-1

hr-1

)

75

Aro

mat

ase

Act

ivity

(fmol

/mg-1

hr-1

)

75

0

0.2

0.4

0 5 10 15 20

Ave

rage

(Pro

port

ion

B

C

DE0

0.2

0.4

Aver

age

Pop

u(P

ropo

rtion

of C

ar

B

CDE

25%

50%

75%>95%0

0.2

0.4

0 5 10 15 20

Ave

rage

(Pro

port

ion

0

0.2

0.4

0 5 10 15 20

Ave

rage

(Pro

port

ion

B

C

DE0

0.2

0.4

Aver

age

Pop

u(P

ropo

rtion

of C

ar

B

CDE

25%

50%

75%>95%

CN

Fadrozole (µg / L)0 50

0

cc

Fadrozole (µg / L)0 50

0

cc

10

Fadrozole (ug/L)

10

Fadrozole (ug/L)

Time (Years)0 5 10 15 20Time (Years)Time (Years)Time (Years)0 5 10 15 20Time (Years)

8

4

6

E2

(ng/

ml)

8

4

6

E2

(ng/

ml)

4

6

8

(Tho

usan

ds)

lativ

e N

umbe

r of E

ggs

Control21050

Fadrozole (ug/L)

***

4

6

8

(Tho

usan

ds)

lativ

e N

umbe

r of E

ggs

Control21050

Fadrozole (ug/L)

***

0

2 *

*

20g/

ml)

0

2 *

*

20g/

ml)

-20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20Exposure (d)

0

2

Cum

u

-20 -18 -16 -14 -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20Exposure (d)

0

2

Cum

u

0

10Vtg

(mg

*

* *Control 2 10 50

Fadrozole (µg/l)

0

10Vtg

(mg

*

* *Control 2 10 50

Fadrozole (µg/l)Ankley et al., 2002, Toxicol. Sci. 67:121‐130

Empirical Support


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn


AromataseInhibition


PlasmaReduced

circulating E2

NN

Aromatase inhibition

8

6

8

6 8

10

gs

Control2

Fadrozole (ug/L)

8

10

gs

Control2

Fadrozole (ug/L)

Reduced E2, Vtg synthesis

Impaired vitellogenesis Reduced fecundity

CN0

2

4E2

(ng/

ml)

*

*

10

20

Vtg

(mg/

ml)

*

0

2

4E2

(ng/

ml)

*

*

10

20

Vtg

(mg/

ml)

*

2

4

6

(Tho

usan

ds)

Cum

ulat

ive

Num

ber o

f Egg 2

1050

***

2

4

6

(Tho

usan

ds)

Cum

ulat

ive

Num

ber o

f Egg 2

1050

***

0* *

Control 2 10 50Fadrozole (µg/l)

0* *

Control 2 10 50Fadrozole (µg/l)

-20 -18 -16-14-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20Exposure (d)

0-20 -18 -16-14-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20

Exposure (d)

0

Consistent profile of effects have been observed with other cyp19 inhibitors and in other species :

Prochloraz, fathead minnow: Toxicol. Sci. 2005. 86: 300‐308

Propiconazole, fathead minnow: Toxicol. Sci. 2013. 132: 284‐297.Propiconazole, fathead minnow: Toxicol. Sci. 2013. 132: 284 297.

Letrozole, Japanes medaka: Compar. Biochem. Physiol. Pt. C, 2007, 145: 533‐541

Empirical Support


OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn


AromataseInhibition


PlasmaReduced

circulating E2

3

4

ng/m

l)

<6h 12 h 24 h days weeks years

Temporal concordance

0

1

2

**

Ex vivo E2

(

**

6

8

(ng/ml)

)

6 Hour 12 Hour 24 Hour

0

2

4

***

Plasma E2

(

6 Hour 12 Hour 24 Hour 30

40

50

*

(c)

TG (m

g/ml)

6h 12h 24hPRO (200 μg/L)

P 6 Hour 12 Hour 24 Hour

0

10

20

Plasma VT

6 Hour 12 Hour 24 HourPRO (200 μg/L)

Skolness et al. 2011 Aquat. Toxicol. 103:170‐178

Empirical Support


AromataseInhibition


PlasmaReduced

circulating E2

PlasmaReduced

circulating VTG

•KE1 – declining through d1 then compensatingTemporal concordance

KE1 declining through d1 then compensating•KE2 – declining through d2 then compensating•KE4 – declining through d4

P t i t l d•Post‐exposure recovery in same temporal order

Essentiality of the KEs

Recovery Experiments Provide Evidence for Essentiality

HepatocyteAromatase Granulosa Plasma Plasmap yReduced VTG production

Inhibition Reduced E2 synthesis

Reduced circulating E2

Reduced circulating VTG

Quantitative Understanding


AromataseInhibition


PlasmaReduced

circulating E2

PlasmaReduced

circulating VTG

Input parameter – in vitro measure of aromatase inhibition

200 Simulate…

0

50

100

150

Veno

us VTG

(μM)

0‐5 5 15 25 35

150

200 Simulated …

(μM)

HPG axis model that accounts for 0

50

100

5 0 5 10 15 20 25 30 35

Veno

us VTG

Breen et al. 2013 Toxicol. Sci. 133: 234‐237

feedback and compensatory responses developed.

‐5 0 5 10 15 20 25 30 35Time (d)


PlasmaReduced

circulating VTG

OvaryImpaired

Oocyte Dev.

FemaleDecreased

ovulation/spawn

50

100

150

200 Simulated …

0

50

‐5 0 5 10 15 20 25 30 35

Statistically based model predicts cumulative fecundity and spawning from VTG dataStatistically‐based model predicts cumulative fecundity and spawning from VTG data.


Ad t tit ti d t di t f MIE t di t it f AO ( d

HPG axis model

A t Granulosa Plasma Hepatocyte Plasma

Adequate quantitative understanding to use measure of MIE to predict severity of AO (and dose‐response time‐course behaviors of intermediate KEs)

AromataseInhibition


PlasmaReduced circulating

E2

HepatocyteReduced VTG

production

PlasmaReduced circulating

VTG

Oocyte growth dynamics modelOvary

Impaired Female

Decreased l ti /

l d d l

Oocyte Dev. ovulation/spawning

Population dynamics modelPopulationDeclining Trajectory


Model accuracy still being tested

Elements of WOE

Elements of WOE Evaluation Considerations

Biological Plausibility of KERs Are the KERs consistent with current biological understanding

Essentiality of KEs Are the effects are reversible if the stressor is removed? Are the downstream KEs prevented if one or more upstream KEs are blocked? (e.g., knock‐out, over‐expression, pharmacological agents)

Empirical Support for KERs Dose response ‒ Upstream KEs observed at doses below or similar toEmpirical Support for KERs Dose response Upstream KEs observed at doses below or similar to those associated with downstream KEs and the AO?

Temporality ‒ The KEs are observed in hypothesized order?Incidence ‒ The frequency of occurrence of the apical effect less than

that for the key events?Consistency – relationship consistently observed within the relevant

domain of applicability, inconsistencies are explainable.

Quantitative Understanding of the KERs

Extent to which the state of KE downstream and can be precisely inferred from the measured or predicted state of KE upstream.Extent to which relevant modulating factors are understood.

WOE Calls

•WOE calls and analysis are summarized in the AOP knowledgebase

•Data from which the calls are derived, are captured on AOP, KE, and KER pages –transparency.

•Allows user to evaluate fit‐for‐purpose suitability for varying regulatory contexts

Summary

•AOPs are a framework for organizing existing knowledge concerning biological motifs of failure.

•Intended to support the use of mechanistic/pathway perturbation data in risk‐based regulatory decision‐making.data in risk based regulatory decision making.

•AOPs are NOT risk assessments

•Application in chemical risk assessment requires additional info:•Exposure (concentration, duration, route)ADME/T i ki ti ( t ti d ti f f t b ti•ADME/Toxicokinetics (concentration, duration, frequency of perturbation at target site)

Th d l d i d d t f th l t li ti f•They are developed independent of the regulatory applications for which they may be applied.

The Role for AOPs in 21st Century Toxicology

Cart

Cvfat

Cven

Cexp Cinsp

Gills

Cart

QfatFat Tissue Group

Effective Respiratory Volume

Cardiac Output Qc

Qw

Qc

Qw

Cvcarc Cart

QcarcCarcass Tissue Group

Cvliv Cart

Qliv

Liver

Km, Vmax

Source EnvironmentalFate Exposure

Molecular InitiatingEvent

Pathway Perturbation / Key Events

Apical Outcomes

Population/Community Impacts

ADME, Toxicokinetics

Cart

Qgut

Cvliv Cart

Qdigesta

CdigestaCdiet

Qdiet

Cvgut

Gut

kgut

Qbile Cbile

Gut Lumen

Gut Tissue

Exposure Assessmentff

AOP

Effects Assessment

•An AOP is not a risk assessment

•AOP describes a generalizable/predictable biological motif of failure that can be expected when a particular biological pathway or process is perturbed.

Summary

•Ad hoc process of AOP development/description has evolved•Guided by set of core principles•Conventions•Emerging best practicesEmerging best practices

•Process is aimed at assembling knowledge in a manner that facilitates evaluation and applicationfacilitates evaluation and application.

•Evaluation of AOPs and the level of confidence and quantitative understanding of the relationships they depict dictates suitable applications (including for MOA analyses).

Summary

•AOP facilitates translation of mechanistic data to outcomes that meaningful to decision‐making.

•Facilitates MOA‐type analysis of cross‐species relevance.

•Application in chemical risk assessment requires additional info:•Exposure (concentration, duration, route)•ADME/Toxicokinetics (concentration, duration, frequency of perturbation at target site)

Using AOP KnowledgeHow certain

Needs from a regulatory standpoint:•Systematically organized

How certain are we?

•Systematically organized•Transparent, well documented•Scientifically‐defensible, credibleScientifically defensible, credible

Pathway‐based bioactivity profile

Potential/probable AO

OECD Guidance on Developing and Assessingand Assessing

AOPsReview and evaluation

Conceptually, AOP description documents could meet those needs.

process

Need for an AOP‐KB

Documents are not ideal for:•Facilitating collaboration and crowd‐sourcingFacilitating collaboration and crowd sourcing•Avoiding duplicative effort•Integration and analysis•Accessible and searchable

CSSCSSNeed for an AOP Knowledgebase

AOPs are modular•KEs and KERs are shared by multiple AOPsy p•No need to re‐write the same descriptions over and over•Reusability (best practices)

AOPs are living documentsAOPs are living documents•KE and KER descriptions can be expected to evolve over time•As descriptions are updated and expanded – all AOP descriptions they link to update automaticallyy p y

•AOP networks for prediction•Entry of structured information in KB allows for de‐facto assembly of AOP networksAOP networks.

How you can access AOP Knowledge

Released and in use

Integration with KB in 2015

On‐going development

https://aopkb.org/

KB in 2015development, case studies

In development

InIn development

Components of the AOP‐KBModular organization.Allows independent KE, KER pages to be linked to multiple AOPs.

https://aopkb.org/aopwiki/index.php/Main_Page

Contents of the AOP‐KB

•Modular structure implemented via a common knowledgebase provides a tractable approach to tackle “real‐world” toxicological complexity and supports more integrated toxicology.

OECD AOP Development Programme– 29 development projects & 9 case studies

Only one reviewed and approved AOP Skin Sensitization– Only one reviewed and approved AOP ‐ Skin Sensitization– 14 AOPs (10 projects) being reviewed by EAGMST– 2 AOPs (2 projects) are scheduled for EAGMST review

AOP Wiki AOPs– 17 AOPs Open for Commenting

bl d• Public comments are encouraged• Includes the 14 under EAGMST review

– 40+ AOPs Under Development• Available for viewing, but not comments• Note: Not all are under active development

Acknowledgements

AOP Training Team:Mirjam Luijten, Bette Meek, Carole Yauk, Stephen Edwards, Kristie Sullivan, Dan Villeneuve

EAGMSTEAGMST

Users’ Handbook Development Team:Bette Meek, Carole Yauk, Grace Tier, Joop DeKnecht, Stephen Edwards, Julia Filipovska, Rick Becker, Anne Gourmelon, Nathalie Delrue, Hristo Aladjov, Dan VilleneuveEAGMSTEAGMST

Somma Lombardo AOP Workshop WG2Dan Villeneuve, Doug Crump, Natalia Garcia‐Reyero,

EAGMSTEAGMST

, g p, y ,Markus Hecker, Tom Hutchinson, Carlie LaLone, Brigitte Landesmann, Teresa Lettieri, Sharon Munn, Malgorzata Nepelska, Mary Ann Ottinger, Lucia Ver a en Ma ri e WhelanVergauwen, Maurice Whelan.

Acknowledgements

• Clemens Wittwehr• Brigitte Landesmann

• Stephen Edwards• David Lyons

• Marina Goumenou• Sharon Munns• Maurice Whelan

• Ryan Durden• Max Felsher• Harriet Ashcroft

K l P i t• Ed Perkins• Natalia Garcia Reyero• Tanwir Habib

• Kyle Painter• Dan Villeneuve• Kevin Crofton• Gary Ankley• Tanwir Habib

• Hristo AladjovJoop DeKnecht

• Gary Ankley• Lyle Burgoon• Robert Kavlock

• Joop DeKnecht• Collaborative Partners

– OECD External Advisory Group on Molecular Screening & Toxicogenomicsg g

– IPCS/WHO Mode of Action Steering Committee

SMALL FISH

• USEPA, NHEERL – Duluth, MN, and Grosse Isle, MI

• G. Ankley, J. Berninger, B. Blackwell, K Jensen, M Kahl, C. LaLone, D. Miller, A Schroeder D VilleneuveA. Schroeder, D. Villeneuve,

• USEPA, NERL – Athens, GA

• T. Collette, D. Ekman, J. Davis, D. Skelton, Q. Teng

• USEPA, NERL – Cincinnati, OH

• D. Bencic, M. Kostich, A. Biales, J. Lazorchak, R. Wang,

• USEPA, NHEERL, ‐RTP, NC

• M. Breen, R. Conolly, S. Edwards, L. Burgoon, S. Bell, D. Lyons

• USACE, ERDC – Vicksburg, MS

• E. Perkins, T. Habib, M. Mayo

• Other partners• Other partners• Mississippi State University, N. Garcia‐Reyero, • University of St. Thomas, D. Martinovic• Oregon Health and Science University, K. Watanabe

• Current Contractors and FellowsCurrent Contractors and Fellows• J. Cavallin, K. Nelson, R. Milsk, E. Randolph, T. Saari

Retired: E. Durhan, E. Makynen,

Former Contractors, Fellows, Students, Interns:M. Breen, M. Hughes, L. Wehmas, L. Thomas, K. Stevens, S. Seidl, E. Eid, M. Weberg, M. Lee, L. Blake, N. Mueller, K. Green, J. Brodin, S. Skolness, O. Adedeji, S. Robinson