NexGen Risk Assessment - toxicology.org

NexGenRisk Assessment

Daniel Krewski, PhD, MHAProfessor and DirectorMcLaughlin Centre for

Population Heath Risk Assessment&

Risk Sciences International

January 12, 2011

McLaughlin Centre for Population Health Risk Assessment

Cornerstone #1:Toxicity Testing in the 21st Century


Toxicity Testing in the 21st Century:A Vision and A Strategy

Committee on Toxicity Testing and Assessment ofEnvironmental Agents

Board on Environmental Studies and Toxicology

Institute for Laboratory Animal Research

Division on Earth and Life Studies

National Research Council

Components of the Vision

Chemical Characterization

Toxicity Testing

Endorsement by the Scientific Community

Collins, F.S., Gray, G.M. & Bucher, J.R. (2008), Science (Policy Forum). Vol. 319. pp. 906 - 907

US Environmental Protection AgencyStrategic Plan and Strategic Goals

http://www.epa.gov/osa/spc/toxicitytesting/docs/toxtest_strategy_032309.pdf

Reaction from Experts in Risk Assessment

“Suresh Moolgavkar, our Area Editor for Health Risk Assessment,asked six experts with different perspectives to comment on thepaper. Each praises the vision and offers suggestions for making itmore useful.”

Michael Greenberg & Karen Lowrie, Editors

Reaction from the Legal Community

Reaction from the Animal Law Community

International Symposia onChallenges and Opportunities in Implementation

June 29-30, 2009 September 12, 2009 November 5, 2009 June 21-23, 2010

“There is widespread support for the NAS vision. There are alsoreal but surmountable challenges in moving the vision into routineregulatory practice. Progress is being made in producing thenecessary science and knowledge base — we need to redouble ourefforts to see that these insights carry over into the worlds of lawand policy.”

Paul Locke, Johns Hopkins UniversityCenter for Alternatives to Animal Testing

Cornerstone #2:Population HealthRisk Assessment



Canadian Institute for Advanced Research (1989)

Health&

FunctionDisease Health Care

Interventions

SocialEnvironment

PhysicalEnvironment Genetic

Endowment

Well-being Prosperity

Individual Response•biology•behaviour

What is “Population Health Risk Assessment”?

Risk assessment is a process to characterize risk using scientific methods.

Population health risk assessment is the comprehensive assessment of health risks in the general population based on genetic, environmental, social & behavioural determinants of health.

This forms the basis for evidence-based population health risk policy analysis, and, ultimately, cost-effective population health risk management decisions.


Health Risk ScienceDeterminants and Interactions

Health Risk Policy AnalysisEvidence Based Policy

Biologyand

Genetics

Socialand

Behavioural

Biology-social interactions

Environmentand

Occupation

Biology-environmentinteractions

Population Health

Environment-socialinteractions

Multiple Interventions

AdvisoryRegulatory Economic Community Technological

Cornerstone #3:Science and Decisions:

Advancing Risk Assessment


KEY MESSAGES

Enhanced framework Formative focus Four steps still core Matching analysis to decisions Clearer estimates of population risk Advancing cumulative assessments People and capacity building

Phase IFormulating and Scoping ProblemFor environmental condition:• What’s theproblem?

• What are theoptions for altering?

• What assessmentsare needed toevaluate options?

Phase IIIRisk Management

• Relative benefits ofproposed options?

• How are other factors(e.g., costs) affected byoptions?

• Which option is chosen?What’s the uncertaintyand justification?

• How to communicate it?• Should decision

effectiveness beevaluated? If so, how?

Phase IIPlanning and

Risk AssessingStage 1: Planning for:

• Options Assessment• Uncertainty and

Variability AnalysisStage 2: Assessing

Stage 3: Confirming Utility of Assessment

Not

OK

OK

Stakeholder involvement at each phase

“Risk-Based Decision-Making” Framework


Background Paper #1


NRC Vision for the Future of Toxicity Testing



Interim and Final Reports

www.nas.edu


Original Editorial

8 +1 Invited Commentaries 2009-2010


• Definition of adversity

• Predicting in vivo results from in vitro toxicity pathway assay results

• Setting standards from results of in vitro assays

• How can the change from current practices to a new paradigm occur?

Recurring Themes in the Commentaries

.



• Part A: NRC Report on Toxicity Testing in the 21st Century (reprint with permission)

• Part B: U.S. EPA Strategic Plan for Toxicity Testing (reprint)

JTEH Special Issue onFuture Directions in Toxicity Testing

• Part C: Individual contributions on future directions in toxicity testing


Building the Scientific Toolbox(Andersen et al., 2010)

Case Study Prototypes



Tool Lung Injury and Ozone

Functional genomics • Microarray analysis of both in vivo and in vitro airway epithelial cells exposed to clean air and ozone • Proteomic changes will be monitored and compared with mRNA changes • Comparison of altered toxicity pathways between in vivo and in vitro cell systems

Bioinformatics • Mining of known toxicity pathways, key biological events, biomarkers and relevant endpoints related to ozone toxicity

Systems biology • Systems biology approach to integrate information for modeling • NFκB inflammatory pathway due to ozone exposure • Goals are to understand network mechanisms behind disease process

Computational systems biology

• Development of BBDR models • Framework to illustrate key events and MOA due to ozone exposure

Structure-activity relationships

• Ozone structure linked to perturbation pathways

Molecular epidemiology

• Examination of a number of inflammatory markers to assess inflammatory response in lungs such as neutrophils, cytokines, prostaglandins, LDH, etc.

Genetic epidemiology • Comparison of toxicity pathways from in vitro and in vivo systems come from the same subjects therefore the variability in response across all the subjects will be taken into account



Tool Developmental Impairment and Thyroid Hormone Disruptors

High throughput screens

• ToxCastTM : screened for endpoints relevant to thyroid hormone disruption • 90 HTS assays were identified as thyroid hormone pathway relevant :

1) endpoints had a direct impact on thyroid hormone pathways, 2) 85 had impacts related to indirect thyroid hormone homeostasis and neurodevelopment

Functional genomics • HPT axis is well mapped and points of possible disruption due to chemical exposure are shown in Figure 1 • Figure 2 shows thyroid hormone control pathways and points of disruption • 90 assays were identified that measured endpoints and had dose- response curves related to HPT pathway • The identification of a ‘signature’ profile could involve the combined analysis of a diverse set of assays endpoints

Bioinformatics: • ToxCast database, developed by NCCT and NIEHS • Mining of this database identified relevant endpoints found in thyroid hormone disruption • Quality analysis and summarization of data entered into a database is needed for effective predictive modeling • Simplification of data into binary formats are highly dependant on appropriate criteria and may be inconsistent across

different assays

Systems biology • 90 assays identified with endpoints related to thyroid hormone disruptions; pathways and cellular responses include • hepatic stimulation and metabolism, thyroid hormone clearance and thyroid hormone pathway


• ToxCast database: analysis of dose-response curves by visual inspection and modeling 1) species concordance, 2) effect identification, and 3) relative potency.

• Chemical classification is highly dependant on quality data analysis and summarization for predictive modeling


• Table 1 identifies IRIS chemicals that have direct effects on thyroid hormone pathway

Biomarkers • Table 2 identifies several assay targets to measure thyroid hormone disruption and homeostasis • Table 3 identifies possible endpoints that still need to be evaluated and measured in assays • CAR, PXR nuclear receptor activation: activation linked directly to reduced thyroid hormone • Other possible nuclear receptors related to reduced thyroid hormone: PPAR, LXR



Tool Cancer and Polycyclic Aromatic Hydrocarbons


• ToxCastTM: HTS assays for genotoxicity and detecting genes in the p53 signaling pathway • Three HTS assays were evaluated:

1) the GreenScreen HC assay using a (TK6 human cell line with a GADD45a-GFP (green fluorescent protein) gene reporter)

2) the p53 endpoint in the CellCiphr Cytotox Profiling Panel uses (HepG2 cells and an anti-p53 antibody) 3) the Invitrogen CellSensor p53RE-bla HCT-116 assay (HCT-116 colon cancer cells with a beta-lactamase reporter gene

controlled by p53 response elements)

Functional genomics • Table 3.2 evaluates data for key mechanistic events from ‘omics’ literature: 1) High carcinogenic potency of dibenzo[a,l]pyrene potentially related to genotoxic key events (i.e., DNA adduct

formation, ras-oncogene mutation, and tumor initiation) 2) Carcinogenicity of benzo[a]pyrene could be due to direct interaction of metabolites with DNA (i.e., adduct formation,

oncogene and tumor suppressor gene mutation, tumor initiation) and possible epigenetic key events (AhR-mediated effects, inhibition of gap junctional intracellular communication [GJIC])

3) dibenzo[a,l]pyrene and dibenzo[a,h]anthracene induced changes in genes involved in apoptosis • Pathway analysis demonstrated that the potent PAHs modulated several pathways whereas less potent PAHs

(fluoranthene and 1-methylphenanthrene) did not significantly affect any pathways. • Proteomic changes were limited to evaluation of p53 accumulation • Venn diagrams showing overlap in gene expression changes

Bioinformatics • Micro array data from many studies were analyzed and summarized in several tables, for example Table 3-4 list various genes affected by PAHs and significantly correlated with cancer potency, DNA adduct formation, or Ah induction in HepG2 cells

Systems biology • PAHs have the ability to bind to AhR and this is suspected to suppress estrogen response element-controlled gene expression through crosstalk between the AhR and ERα.



Tool Cancer and Polycyclic Aromatic Hydrocarbons (Part 2)


• QSAR models capable of predicting mechanisms and/or cancer potency for PAHs based on structural features: 1) PAHs having four or more benzene rings exhibit greater carcinogenic potency 2) PAHs with two or three benzene rings have less carcinogenic potency 3) carcinogenicity is related to the specific arrangement of the benzene rings 4) PAH metabolites resistant to detoxification due to stereochemical effects are more likely to be mutagenic and cause

cancer 5) Dihydrodiol epoxides formed at certain positions on the PAH molecule are more accessible to glutathione transferase

detoxification and are less potent mutagens and carcinogens • Table 3-1 tabulates chemical with three in vivo pathways linked to differing potencies of carcinogenesis • Table 3-4 shows the genes affected by PAHs and significantly correlated with cancer potency, DNA adduct formation, or

Ah induction in HepG2 cells

Biomarkers • Possible biomarkers could be found from microarray data and Venn diagrams • Biomarkers of AhR, estrogen response, and p53 pathway

Exposure assessment

• Extracts from soils contaminated with PAHs were exposed to HepG2 cells and various proteins levels related to different pathways were measured; p53, Cdk2, cyclin D1, p21, Chk2 Thr68, Mdm2 Ser166, pErk Tyr204,53BP1, γ-H2AX, and Mdm2 2A10



Tool Cancer and Benzene

Functional genomics • Genome-wide expression profiles of human promyelocytic leukemia HL-60 cells exposed to VOCs was analyzed using a microarray

• Yeast screening approach identified highly conserved genes involved in susceptibility to HQ, CAT, and BT. NF1 (neurofibromin) was identified as a candidate susceptibility gene to the benzene metabolite, hydroquinone (HQ)

• Microarray platforms to identify global gene expression changes associated with occupational benzene exposure in the peripheral blood mononuclear cells (PBMC) of a population of shoefactory workers. Four genes (CXCL16, ZNF331, JUN and PF4were among the top 100 genes identified by both platforms

• Carcinogenesis epigenomics data showed effects of benzene on the DNA methylation of specific genes and on miRNA expression

Systems biology • The roles of the human homologs of selected genes in benzene toxicity were examined through mechanistic studies in human cell lines. WRN protein using siRNA in HeLa cells and examined sensitivity to toxicity following exposure to the benzene metabolite, HQ

Biomarkers • Biomarkers to assess benzene exposure: Urinary trans,trans-muconic acid (t,t-MA), S-phenylmercapturic acid, and benzene (U-benzene)

Molecular epidemiology • 125 workers exposed to a range of benzene exposures : microarray mRNA transcriptome analysis of peripheral blood mononuclear cells identified genes and pathways (apoptosis, immune response, and inflammatory response)

• 239 workers recruited among traffic policemen, taxi drivers and gasoline pump attendants of the city of Parma (Italy) to study benzene exposures at low levels: Study investigates nucleic acid oxidation and biomarkers associated with exposure.

• Gas station attendants, urban policemen, bus drivers, and two groups of controls (415 subjects): Biomarkers such as urinary trans, trans-muconic acid (t,t-MA), S-phenylmercapturic acid, and benzene (U-benzene) were measured to assess benzene exposure

Genetic epidemiology • Genotyping studies of 250 benzene-exposed workers and 140 unexposed controls in China: analyzed 1395 SNPs in 411 potential carcinogenesis-related genes and associations between DNA repair, genomic maintenance and development of benzene hematotoxicity were found


Tool

Lung Injury and

Ozone

Developmental Impairment and

Thyroid Hormone Disruptors

Cancer and Polycyclic Aromatic

Hydrocarbons

Cancer and Benzene


Stem cell biology

Functional genomics

Bioinformatics

Systems biology


Physiologically-based pharmacokinetic models


Biomarkers

Molecular and genetic epidemiology .

Exposure assessment


Background Paper #2

Cornerstone #1:Toxicity Testing in the 21st Century


Toxicity Testing and Risk Assessment(from Krewski et al., 2010, Annual Review of Public Health, in press)






Cornerstone #2:Population HealthRisk Assessment


Health Risk ScienceDeterminants and Interactions

Health Risk Policy AnalysisEvidence Based Policy

Biologyand

Genetics

Socialand

Behavioural

Biology-social interactions

Environmentand

Occupation

Biology-environmentinteractions

Population Health

Environment-socialinteractions

Multiple Interventions

AdvisoryRegulatory Economic Community Technological

Health Determinants

Biology & Genetics

Genetic Polymorphisms

Gender

Immune status

Metabolism

Health Determinants

Environment & Occupational

Physical environment: air, food, water, soil, built environment

Working conditions

Health Determinants

Social & Behavioural

Personal health practices

Coping skills

Employment/ working conditions

Education and literacy

Social environment

Healthy child development

Culture

Customs, traditions, and beliefs

Social stability

Intervention Examples

Regulatory Governmental laws and regulations, policies, guidelines, audits, inspections and fines

Economic Levies or other cost structures, insurance, investments in future technology, government grants to invest in new technology

Advisory Health promotion campaigns, safety awareness workshops

Community Action Volunteer public monitoring groups, increases in capacity building

Technological Updating equipment to newer technology or repairing old technology, financial investment in research and development of new emerging technologies

Multiple Interventions: REACT

Cornerstone #3:Science and Decisions:

Advancing Risk Assesment


Table 4. Potential Modifications of Risk Assessment Approaches in a NexGen Context

Risk Issue Current Approach NexGen Approach

Defining Adversity Adversity is presently defined in terms of observation of apical endpoints in mammalian systems.

Adversity will be defined in terms of critical perturbations of toxicity pathways, ultimately in the absence of information on apical outcomes. Defining adversity will require knowledge of dose response for various pathway assays and in vitro models that assess conditions leading to excessive pathway perturbations in relevant assays.

Point of departure (PoD)

NOAEL or BMD, based on apical endpoints.

Dose response curves are presently based on measurement of apical endpoints in animal models, with inter-species extrapolation of the health outcome from animal to human.

NexGen risk assessments will still use empirical statistical models to establish dose response relationships for pathway perturbations, and select established points of departure such as the BMD for in vitro responses deemed ‘adverse’. Alternative approaches for PoDs may include new methods, such as SNCD (signal to noise crossover dose) or statistical methods to assess likelihood of threshold behaviours.

Inter-species extrapolations will no longer be required when human cells are used in HTS in vitro assays.





Default assumptions

Current default assumptions used in risk assessment (such as a 10-fold variation in sensitivity within the human population) are usually based on limited empirical evidence.

Understanding toxicity pathways in more depth will permit a move away from default assumptions, towards a more mechanistic approach guided by scientific evidence and knowledge of the behaviour of the toxicity pathway in shifting from basal levels of activity to enhanced function with excessive perturbation. (It will likely be possible to characterize phenotypic variation with some precision using suites of human cell lines representing inherent differences in sensitivity and differences among life stages and through knowledge of pathway components and polymorphisms in these components that affect function).



Analysis of various life stages

In vivo generational studies using mammalian systems can be conducted at different life stages, but are expensive, time consuming and labour intensive.

In the absence of evidence to the contrary, it is assumed that individuals at all life stages are equally vulnerable.

NexGen risk assessments can make use of systems biology and virtual tissue modeling that simulates human organs in 3 dimensions (e.g., 3D liver model), to determine the effects that could be expected from the exposures of humans of various ages to toxic substances.

High throughput in vitro assays using human cells and human cell lines representing different life stages could facilitate a more complete assessment of vulnerability at different life stages.



Analysis of multiple exposure doses

Present risk assessments are based on dose-response curves generated by high dose acute exposures of animal models, and then low dose extrapolation is performed.

Laboratory exposures of animals to environmentally-relevant and sublethal doses are not generally performed for the purpose of conducting a risk assessment but may be stated in the risk assessment report.

It will be feasible in NexGen risk assessments to conduct exposures to a wide range of doses for risk assessment purposes because of the rapid and cost efficient use of HTS. Using medium throughput, multi-dimensional tissue surrogates should allow some conditions where in vitro responses will be assessed over days to weeks of exposure.

Selected Risk Issues to be Addressed

• Chemical mixtures

• Joint effects of multiple stressors

• Assessment of delayed effects

• Reversible or transient effects

• Analysis of various life stages

• Analysis of multiple exposure doses

• Assessment of exposures of different durations (e.g., acute, chronic, and intermittent exposures)


NexGen Risk Assessment - toxicology.org

Documents