Application of transcriptomics concentration-response modeling for prioritization of contaminants detected in tributaries of the North American Great Lakes.

Abstract

As part of the Great Lakes Restoration Initiative, chemical monitoring and surveillance efforts have detected ∼330 chemicals in surface water of Great Lakes tributaries. There were 140 chemicals for which no empirical toxicity data were available. The aim of the present study was to generate transcriptomic points of departure (tPODs) for 10 of these compounds and demonstrate how they could be applied in a screening-level prioritization. Organisms representing three trophic levels of the aquatic food-web (P. promelas, D. magna, and R. subcapitata) were exposed for 24 h to a ½-log dilution series of nominal exposure concentrations typically ranging from 66.7 to 0.021 µM of each chemical. In addition to observations of apical effects (e.g., survival and morphology), whole body transcriptomic responses (tPODs) to each chemical were evaluated with targeted analysis using TempO-seq for P. promelas and D. magna and non-targeted RNA-seq for R. subcapitata. The tPODs ranged from 0.18 to 10.8 µM for P. promelas and 0.32 to 29 µM for D. magna, with the most potent of the chemicals tested being fipronil carboxamide for both species. For R. subcapitata, the tPODs ranged from 0.04 to 1.77 µM, with gabapentin as the most potent chemical tested. Empirically derived tPODs from these data-poor chemicals were compared to concentrations detected in the Great Lakes basin. Environmental concentrations were less than the tPODs, except for R. subcapitata and 3,4-dichlorophenyl isocyanate. Similarly, tPODs from previously tested data-rich chemicals were compared with environmental concentrations, in which case tPODs from several chemicals overlapped environmental concentrations. This work demonstrates the potential utility of emerging ecological high-throughput transcriptomics assays to support screening and prioritization of data-poor environmental contaminants.