Category: Chemical Biology
Current initiatives in the field of environmental toxicology include integrating toxicity testing data from in vitro and alternative methods into risk assessment practices. Frameworks have been proposed that use potency estimates of bioactivity from in vitro screening models, coupled with high throughput toxicokinetic predictions, to calculate human equivalent doses that can, in turn, be used as the basis for screening-level human health risk assessments. Previously, in vitro point of departure determination has been based on panels of biochemical or functional assays that report a single endpoint or a limited number of endpoints. However, increasing efficiency and declining cost of generating whole transcriptome profiles has made high-throughput transcriptomics (HTTr) a practical option for determining bioactivity thresholds in in vitro models. We present a microfluidics-based laboratory workflow for HTTr screening of MCF7 cells in 384-well format as well as a HTTr analysis pipeline for plate-level and sample level quality control and concentration-response modeling of HTTr data. A total 2200 chemicals from the Toxicity Forecaster (i.e. ToxCast) inventory were screened in 8-point concentration-response (0.03 – 100 µM) in MCF7 cells at a single exposure duration (6 hr) using 3 independent cultures (n = 1/treatment/culture). Cell viability and apoptosis assays were run in parallel to identify treatments producing a marked loss of cell viability. For assays, chemicals were applied using an acoustic dispenser and each test plate was uniquely randomized with respect to treatment positioning. In the TempO-Seq HTTr assay, a subset of wells in each assay plate was also treated with 3 reference chemicals (single concentration) in triplicate to assess laboratory workflow performance over time. In addition, reference RNAs and bulk lysates from DMSO and Trichostatin A treated MCF7 cells were installed on each plate to assess repeatability of the TempO-Seq assay across plates and experimental blocks. Total mapped reads, percent mapped reads, correlation of like samples on an assay plate, coefficient of variation in probe level expression and estimates and repeatability of differentially expressed gene (DEG) calls were evaluated as sample level quality control metrics towards the goal of establishing a reference range for assay performance. Concentration-response modeling demonstrated a broad range of gene and pathway level potency values across chemicals, and facilitated identification of no observable transcriptional response levels (NOTELs), which are candidates for use in screening level next generation risk assessments. In addition, an existing database of whole transcriptome microarray data from MCF7 cells treated with well-characterized chemicals (i.e. Broad Institute CMAP database) was annotated in terms of known molecular targets. Connectivity mapping of TempO-Seq data to this database was able to identify putative molecular targets for reference chemicals and a subset of chemicals testing during the HTTr screening campaign. This abstract does not necessarily reflect USEPA policy.
Joshua Harrill– Cellular and Molecular Toxicologist, USEPA, Durham, NC
Cellular and Molecular Toxicologist
I am a toxicologist for the U.S. Environmental Protection Agency. My laboratory is focused on application of high-throughput, high-content screening assays for environmental chemical hazard identification and characterization. We currently use high-throughput transcriptomics (i.e. Targeted RNA-Seq) and high content imaging as a our primary screening tools.