The LPA1 receptor antagonist BMS-986020, although efficacious in a Phase 2 IPF trial, was withdrawn voluntarily from clinical development due to hepatobiliary toxicity. Understanding the BMS-986020 clinical findings is complicated by the fact that similar toxicity was not observed in rats or dogs. Retrospective nonclinical investigations using in vitro and in vivo models were conducted to identify the potential mechanism for these hepatobiliary changes and to help define a preclinical testing strategy for hepatobiliary toxicity. BMS-986234, a potent LPA1 receptor antagonist and clinical candidate structurally distinct from BMS-986020, was similarly extensively evaluated nonclinically for its potential to induce hepatobiliary toxicity. The principal hypotheses for the BMS-986020-induced toxicity are: 1) inhibition of bile acid (BA) and phospholipid efflux from liver leading to hepatotoxicity, 2) alteration of bile composition and flow leading to biliary toxicity, and 3) direct toxicity to hepatocytes and cholangiocytes. In human hepatocytes, BMS-986020 inhibited the BA efflux transporters BSEP (1.8 µM IC50), MRP3 (22 µM) and MRP4 (6.2 µM), and the phospholipid efflux transporter MDR3 (7.5 µM) at clinically relevant drug concentrations. BMS-986020 inhibited BA efflux from Matrigel® human hepatocyte sandwich cultures (68% at 10 µM) and inhibited mitochondrial function in human hepatocytes (61 µM IC50) and cholangiocytes (37 µM), but did not increase oxidative stress (&gt; 200 µM IC50). These results are consistent with reports that concurrent combinations of altered BSEP, MRP, MDR3, and mitochondrial function may predispose drugs to increased risk for DILI. In the rat, BMS-986020 increased bile acids secondary to Bsep inhibition (1.8 µM IC50). In monkeys, BMS-986020 produced bile ductule hyperplasia, cholangitis, and cholestasis. In contrast, BMS-986234 did not increase BA in the rat or produce hepatobiliary toxicity in monkey, and did not alter BA, phospholipid transporters (&gt;100 µM IC50) or mitochondrial function (&gt;100 µM IC50) thus indicating that the hepatobiliary tox of BMS-986020 was compound- rather than LPA1-mechanism-based. Though the mechanism of clinical toxicity for BMS-986020 has not been clearly defined, the investigations described here became the basis for an emerging BMS decision tree for drug discovery candidates when there is a question regarding the potential for hepatobiliary toxicity (e.g., target or chemical class, BSEP inhibitor, etc.).

Strategies for early safety assessment of potential drug candidates at Bristol-Myers Squibb: Preclinical evaluation of hepatobiliary toxicity

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.