Use of Neural Models of Proliferation and Neurite Outgrowth to Screen Environmental Chemicals in the ToxCast Phase I Library

The U.S. Environmental Protection Agency's ToxCast program aims to develop rapid and cost-effective toxicity testing approaches. As part of this effort, we screened the ToxCast Phase I library of 309 chemicals using previously developed and optimized high-content imaging and analysis assays for proliferation in human ReNcell CX neuroprogenitor cells and neurite outgrowth in PC12 cells. Cytotoxicity was determined concurrently in both models. For the initial screening, cells were exposed in triplicate to chemicals at 40 μM for 24 (proliferation) or 96 hr (neurite outgrowth). Chemicals were deemed active if effects on any endpoint were >3×the standard deviation of control means. A total of 130/309 chemicals (42.1%) altered at least one of the four endpoints examined. In ReNcell CX cells, 126 (41%) of chemicals were active, and 46 (15%) were active in PC12 cells. In ReNcell CX cells, 63 chemicals selectively inhibited proliferation. In PC12 cells, only four chemicals (methyl isothiocyanate, phosalone, prodiamine, and pyridaben) selectively increased neurite outgrowth, while none selectively decreased neurite outgrowth. All active chemicals were re-tested in a concentration–response study (1 nM–40 μM) and similar results were obtained. These results demonstrate (1) high reproducibility and concordance for both assays; (2) that the sensitivity of single-concentration screening may be limited by biological variability; (3) that, for environmental compounds, cytotoxicity must be used in the interpretation of effects on selected cellular endpoints; and (4) that proliferation in human neuroprogenitor cells responds similarly to proliferation in nonneuronal human cells, but has higher specificity when concurrent cytotoxicity data are considered.