Integrated Omic Analyses Identify Pathways and Transcriptomic Regulators Associated With Chemical Alterations of In Vitro Neural Network Formation

Development of in vitro new approach methodologies has been driven by the need for developmental neurotoxicity (DNT) hazard data on thousands of chemicals. The network formation assay characterizes DNT hazard based on changes in network formation but provides no mechanistic information. This study investigated nervous system signaling pathways and upstream physiological regulators underlying chemically induced neural network dysfunction. Rat primary cortical neural networks grown on microelectrode arrays were exposed for 12 days in vitro to cytosine arabinoside, 5-fluorouracil, domoic acid, cypermethrin, deltamethrin, or haloperidol as these exposures altered network formation in previous studies. RNA-seq from cells and gas chromatography/mass spectrometry analysis of media extracts collected on days in vitro 12 provided gene expression and metabolomic identification, respectively. The integration of differentially expressed genes and metabolites for each neurotoxicant was analyzed using ingenuity pathway analysis. All 6 compounds altered gene expression that linked to developmental disorders and neurological diseases. Other enriched canonical pathways overlapped among compounds of the same class; eg, genes and metabolites altered by both cytosine arabinoside and 5-fluorouracil exposures are enriched in axonal guidance pathways. Integrated analysis of upstream regulators was heterogeneous across compounds, but identified several transcriptomic regulators including CREB1, SOX2, NOTCH1, and PRODH. These results demonstrate that changes in network formation are accompanied by transcriptomic and metabolomic changes and that different classes of compounds produce differing responses. This approach can enhance information obtained from new approach methodologies and contribute to the identification and development of adverse outcome pathways associated with DNT.

Automated summary

This study investigated the signaling pathways and physiological regulators underlying chemically induced neural network dysfunction. The results showed that different classes of compounds produce different responses and changes in network formation are accompanied by transcriptomic and metabolomic changes.