Target promiscuity and physicochemical properties contribute to pharmacologically induced ER-stress

In vivo toxicity of drug candidates remains a major problem in the pharmaceutical industry, and is a significant cause of late stage attrition. As a consequence predictive in vitro assays are developed and put in place early in the discovery pipeline to aid compound selection. Endoplasmic reticulum stress (ER-stress) has been implicated in many disease states, as well as compound-induced organ toxicities. We explored the role of ER-stress as a general mechanism of toxicity by utilizing a high-throughput in vitro assay to screen 316 chemically diverse Pfizer proprietary compounds with known in vivo toxicity outcome for nuclear accumulation of spliced x-box binding protein 1 (XBP1s), a key transcription factor of the unfolded protein response (UPR). We examined the correlation between physicochemical properties, such as molecular weight, pKA, lipophilicity, topological polar surface area, and passive permeability, as well as target promiscuity, between XBP1s hits and non-hits and found that lipophilicity, target promiscuity and low passive permeability significantly contributed to ER-stress. In addition, we have shown that compounds which cause ER-stress in the form of XBP1s activation at concentrations below 40 μM have a more than four times greater chance of causing in vivo toxicity at 10 μM plasma exposure.

► We tested 316 in vivo toxic and clean compounds for induction of XBP1 splicing. ► Physicochemical compound properties contribute to ER-stress. ► Promiscuous compounds are more likely to cause ER-stress. ► Compounds that induce ER-stress are more likely to have a toxic finding in vivo.