You are here:
Identification of Putative Cardiovascular System Developmental Toxicants using a Classification Model based on Signaling Pathway-Adverse Outcome Pathways
Citation:
Hunter, S., K. Grode, Mitch Rosen, D. Kapraun, N. Baker, R. Judson, AND T. Knudsen. Identification of Putative Cardiovascular System Developmental Toxicants using a Classification Model based on Signaling Pathway-Adverse Outcome Pathways. Teratology Society, Denver, CO, June 25 - 29, 2017.
Impact/Purpose:
An important challenge for an integrative approach to developmental systems toxicology is associating putative molecular initiating events (MIEs), cell signaling pathways, cell function and modeled fetal exposure kinetics. We have developed a chemical classification model based on an integrated systems toxicology approach to identify putative cardiac developmental toxicants.This analysis reveals a diverse set of signaling pathway MIEs that correlated with apical endpoints of prenatal cardiovascular developmental toxicity. The classification model is built upon an integrative approach using pathway-based mechanistic prediction in defining the biological applicability domain for prenatal cardiovascular system developmental toxicity.
Description:
An important challenge for an integrative approach to developmental systems toxicology is associating putative molecular initiating events (MIEs), cell signaling pathways, cell function and modeled fetal exposure kinetics. We have developed a chemical classification model based on an integrated systems toxicology approach to identify putative cardiac developmental toxicants. The Mouse Genome Informatics (MGI) Phenotype database was mined to identify genes associated with cardiovascular system (CVS) phenotypes. This gene list was associated with curated signaling pathways using the Panther Gene Ontology database establishing a gene – pathway – phenotype linkage. Similarly, a chemical – gene – pathway linkage was established using the ToxCast in vitro high-throughput assay data consisting of 1017 chemicals over 82 pathway mapped gene-target assays. This produced a matrix of predicted chemical-pathway interactions that linked to CVS defects. A human pregnancy PBPK model was used to estimate the peak fetal serum concentration produced by administration of 1 or 100 mg/kg/day doses during the 14th week of gestation. A chemical was defined as bioactive when the peak serum concentration was greater than the AC50 for any pathway element. For bioactive chemicals, a pathway score was created by summing the AC50s for the constitutive pathway gene elements. Using a classification model based on the signaling pathways associated with CVS defects, we classified 88 chemicals as bioactive at a 1 mg/kg/day exposure and 184 chemicals at 100 mg/kg/day. Acetochlor, rotenone, tebufenpyrad, and valproic acid were classified as bioactive chemicals at both doses. Although all 4 four these chemicals produce developmental toxicity in vivo, only valproic acid is associated with heart defects. This analysis reveals a diverse set of signaling pathway MIEs that correlated with apical endpoints of prenatal cardiovascular developmental toxicity. The classification model is built upon an integrative approach using pathway-based mechanistic prediction in defining the biological applicability domain for prenatal cardiovascular system developmental toxicity. This abstract does not represent EPA policy.