Progress in the Toxicity Testing in the 21 st Century (TT21C) Proposal AXLR8 Workshop 2011 Berlin-Dahlem, Germany May 24-26, 2011 Harvey Clewell Director, Center for Human Health Assessment The Hamner Institutes for Health Sciences hclewell@thehamner.org
Topics Various approaches for implementing the 2007 report Pilot Project Focused efforts Humane Society Human Toxicology Project and an ongoing program at The Hamner Progress & Directions with the Hamner pilot projects
From the report Get all the research, assay development completed and make the change
Incrementally bring tools to bear -Tox21 Profiling and Prioritization High Throughput Screening and Computational Toxicology - 2008 Predict results of animal studies Prioritize for in vivo testing Assist in risk assessment In the last year has added a high-throughput risk assessment piece
Catalyze & Accelerate Change Pilot Project Approaches Humane Society of the United States 1. Select a group of well-studied prototype compounds/pathways 2. Design human/rodent cell or tissue surrogate based toxicity pathway assays 3. Examine results from the panel of assays to assess adversity
Catalyze & Accelerate Change Pilot Project Approaches Humane Society of the United States 4. Refine quantitative risk assessment tools, i.e., computational systems biology pathway models and in vitro-in vivo extrapolation. 5. Integrate results into proposed health safety/risk assessments. Show an entireprocess in vitro safety assessment in practice.
Catalyze & Accelerate Change To date: Humane Society of the United States Consortium is now operational http://htpconsortium.wordpress.com/about-2/ Workshop held last fall in Washington, DC all talks available General endorsement of pilot project approach no direct financial research support to date
A Hamner Initiative: A parallel pilot project effort Similar components as noted with the HTP Support for key pathway research from several sponsors in place Endorsed by Humane Society of the United States Hamner teams involved with pathway mapping and modeling, assay design, refining in vitro assay systems Risk/safety assessment focus
The process Assessing adversity in vitro Computational Systems Biology Pathway (CSBP) Modeling Panel of pathway assays Point of Departure (concentration) Boekelheide and Campion (2010). Tox. Sci., Boekelheide and Andersen (2010). ALTEX,4, 243-252. Bhattacharya et al. (2011). PLoS One, in press. Acceptable concentration in vitro (ug/l) in vivo human exposure standard mg/kg/day in vitro-in vivo dosimetry PK Modeling
Current Support Dow Chemical Unilever Exxon Mobil Foundation Dow Corning ACC-LRI Sumitomo visiting scientist
Chemical Exposure DNA-damage response pathway DNA-reactivity Chemical level Chemical reacts with DNA Functional consequence at cellular/gene level with dose response Genomic responses to exposure Genomic level DNA-damage pathway upregulation Cellular Response level Mutation Adversity Coordinated cellular response through pathway cascades with dose-response and pathway mapping Assess Likelihood of high dose carcinogenicity Boekelheide and Campion. Tox. Sci., 2010 Functional consequence: assessing degree of pathway perturbation necessaryfor adverse organ system outcome Mutated cell
Unilever/Hamner Project DNA damage, µ nuclei, pathway mapping HT-1080 cells with flow cytometry or HCA assays measures of damage Next step select treatment concentrations (grey arrows, box A ) for genomic studies to do pathway mapping (components) and modeling (dose response) A B C D
Develop computational cell biology model of pathway function to evaluate dose responses already started Basal Response Exogenous DNA damage Cell 142, 89 100, July 9, 2010. Elsevier Inc. Coherent feed forward loop Examine individual cells
DNA-Damage Pathway Model Simple model that damage Batchelor et al., Nature Rev. Cancer, 9, 371-377, 2009.
Visualizing data collection and modeling for pathway responses Concentration γor X-ray exposure Dosimetry Pathway dynamics Mapping and Modeling Genomics, Proteomics, Metabolomics Phenotypic Response Basal Function Biochemical effects, feedback control Adaptive Changes Mutation Bhattacharya, S., Zhang, Q., Carmichael, P.C., Boekelheide, K. and Andersen, M.E. Toxicity Testing In the 21 st Century: Defining New Risk AssessmentApproaches Based on Perturbation of Intracellular Toxicity Pathways Using the p53 DNA-Damage and Repair Pathway as a Prototype. PLoS One. in press. Endpoint assays, e.g. gene expression, biomarker changes, mutagenesis, apoptosis, etc.
The safety assessment process mdm2 mdm2 Assays p53 EC XX** a. conc 1 b. conc 2 c. conc 3 d. conc 4 promoter DNA Adversity xx = f(conc i ) Systems model for DNA repair in vitro ug/ml in vitro-in vivo extrapolation Reverse dosimetry Exp c mg/kg/day
Mapping and modeling PPAR-α toxicity pathway PPARα Are figures of this kind useful for enhancing safety assessment?
Extracting Circuit Structure and Dose Resposne Modeling PPARα Coactivator 1 Coactivator 2 Coactivator 3 Coactivator 4 PPARα Kinase 1 Kinase 2 Kinase1 3 TF1 Kinase 4 TF2 TF3 TF4 TF5 TF6 TF7 TF8 TF9 TF10 TF13 TF12 TF11 Adapted from Bromberg et al Science (2008).
Our Experimental System Primary Hepatocytes (3-D sandwich culture) PPARα agonist: GW 7647 Dose (µm) 0.001 0.01 0.1 1 10 Time (hr) 2 6 12 24 72 RNA and Nuclear Extracts
Our Approach Measure whole genome transcriptional expression Confirm expression of select genes by RTPCR Use TRANSFAC to predict TF regulators Identify active response elements using high-content Protein-DNA binding arrays Confirm active response elements by Gel-shift Use Transfac Database to predict TFs binding to these elements Knockdown TFs in primary hepatocytes and evaluate 1) temporal gene expression changes and 2) cellular phenotype(fatty acid oxidation)
Mapping Inferred Pathways in Humans GW7647 gene expression for 1uM at 2, 6, 24 & 72 hr 2hr 6hr Gene input Predicted regulatory TF 24 hr 72 hr
Mapping Inferred Pathways in Humans Using expression data at of GW7647 24hr 0.1, 1 or 10uM 0.1 um 1uM Gene input Predicted regulatory TF 10 um
Next Steps: Evaluate network structure using gene knockdowns, kinase inhibitors and their effect on key nodes. The hypothetical network to the left consists of key nodes ( Y and Z factors) and feed forward loops (FFLs). Define the FFL2 operative for PPAR-α Create computational systems biology (CSB) model for the pathway Use the CSB model for dose-response assessment and for validating the assay i.e., show that assay output is understood based on network biology
Advantages with a Prototype Approach 1. Design assays for purpose i.e., to collect information for defining adversity and for use in risk/safety assessment 2. Establish (or at least discuss) from the start the optimal use of in vitro information so we avoid developing institutionalized default methods for in vitro safety assessments 3. Create a risk/safety based process for prototypes that can be applied as other toxicity pathways are enumerated 4. Discuss operational application of the methods even lacking information on majority of pathways.
Reverse Dosimetry Exposure Modeling for Interpreting In VitroAssay Results Ln Conc (um) 3 2 1 0-1 -2-3 -4-5 0 50 100 150 Time (min) Hepatocellular Clearance Reverse Dosimetry Equivalent Exposure Plasma Protein Binding Estimated Renal Clearance Prediction of in vivo clearance Concentration from in vitro assay
Results From Reverse Dosimetry Analysis The Same EC50 Does Not Imply the Same Exposure! Est Oral Minimum EC50 or Equivalent Chemical ToxCast Endpoint LEL (um) (mg/kg/day) Acetamiprid BSK_BE3C_uPAR 1.481 0.384 Atrazine BSK_KF3CT_IP10 1.481 1.215 Bromacil BSK_BE3C_IP10 1.481 0.888 Forchlorfenuron BSK_BE3C_uPAR 1.481 1.277 Metribuzin BSK_hDFCGF_MMP1 1.481 6.577 Isoxaflutole BSK_hDFCGF_EGFR 1.481 1.209 Dicrotophos BSK_hDFCGF_PAI1 1.481 2.632 Clothianidin BSK_hDFCGF_EGFR 1.481 7.580 Diazoxon BSK_KF3CT_IP10 1.481 0.266 Oxytetracycline BSK_BE3C_IL1a 1.481 0.567 2,4-D BSK_BE3C_IL1a 1.481 1.389 Similar LEL Values Different Oral Equ
Office of Research and Development National Center for Computational Toxicology 27
Comparison of in vitro-to-in vivo extrapolation results with estimates based on in vivo PK Chemical PK-or PBPK- Derived C ss (µm) IVIVE C ss a,b (um) IVIVE Caco-2 c C ss a,b um) IVIVE, F ub =0.99 Css (um) IVIVE F ub =0.99, Caco-2 c Css (um) 2,4-dichlorophenoxyacetic acid 9.05-90.05 39.25 40.34 39.25 40.34 Bisphenol-A < 0.13 d 0.35 0.40 0.06 0.07 Cacodylic acid 1.80 3.06 tbv e 3.06 tbv e Carbaryl 0.03 0.04 0.04 0.03 0.03 Fenitrothion 0.03 17.91 -- f 0.10 -- f Lindane 0.46 13.21 tbv e 0.07 tbv e Oxytetracycline dihydrate 0.36 2.00 0.44 2.00 0.44 Parathion 0.17 24.63 -- f 0.14 -- f Perfluorooctanoic acid 20,120 g 55.34 g 58.19 g 0.4 g 0.4 g Picloram 0.27 57.63 32.01 0.37 0.19 Thiabendazole 0.45 13.76 15.20 13.76 15.20 Triclosan 2-10 1.56 1.59 0.01 0.01
Conclusions In vitro assays for hepatocyte clearance, plasma protein binding, and Caco-2 transport can reduce uncertainty regarding the chronic in vivo doses that are equivalent to in vitro toxicity assay concentrations from over 4 orders of magnitude to less than 2 More complicated approaches are necessary for toxicity mediated by metabolites, or for acute exposures
Evaluation of IVIVE Approach for Toxicity Assessment prediction Comparison of QSAR prediction with in vivo data QSAR prediction of toxicity/target tissue In vitro toxicity assays (EC50) confirmation Comparison of QSPR prediction with in vivo data QSPR prediction of properties, metabolites In vitro metabolism assays Conduct IVIVE to estimate equivalent in vivo exposure Comparison of IVIVE prediction with in vivo results Evaluate IVIVE approach Collaborative Effort between The Hamner and Utrecht University
Considerations for IVIVE Pharmacokinetic factors that affect in vivo toxicity but are not appropriately reflected in in vitro toxicity tests: - Bioavailability - Transport - Protein binding - Clearance - Metabolic - Renal - Biliary - Exhalation
Our Experience in the 1980 s with PBPK 1. Half the work for implementing new methods on a broad scale gets completed by doing the first two-or threecompounds establish process. 2. Insure that the procedural parts are kept in mind assays for purpose, dose response and in vitro-in vivo extrapolation and do the examples as close to A to Z as possible. 3. Expand the group of prototypes and look at the broader assays to pick up models of action for new compounds.
Acknowledgements Companies currently supporting The TT21Cresearch at The Hamner and the HTP and those that are will join the research in the future. http://htpconsortium.wordpress.com/ Hamner colleagues and collaborators Mel Andersen Rebecca Clewell Courtney Woods Kim Boekelheide, Brown University Paul Carmichael, Unilever Sudin Bhattacharya Qiang Zhang Barbara Wetmore Rusty Thomas www.thehamner.org http://www.thehamner.org/institutes-centers/institute-for-chemical-safetysciences/toxicity-testing-in-the-21st-century/