Species-Specific Urothelial Toxicity With an Anti-HIV Noncatalytic Site Integrase Inhibitor (NCINI) Is Related to...

GS-9695 and GS-9822 are next-generation noncatalytic site integrase inhibitors (NCINIs) with significantly improved potency against human immunodeficiency virus compared with previous drugs such as BI-224436. Development stopped due to vacuolation of the bladder urothelium seen in cynomolgus monkey but not in rat; this lesion was absent in equivalent preclinical studies with BI-224436 (tested in dog and rat). Lesions were unlikely to be attributable to target because NCINIs specifically target viral integrase protein and no mammalian homologue is known. Secondary pharmacology studies, mitochondrial toxicity studies, immunophenotyping, and analysis of proteins implicated in cell-cell interactions and/or bladder integrity (E-cadherin, pan-cytokeratin, uroplakins) failed to offer any plausible explanation for the species specificity of the lesion. Because it was characterized by inflammation and disruption of urothelial morphology, we investigated physicochemical changes in the bladder of cynomolgus monkey (urinary pH 5.5-7.4) that might not occur in the bladder of rats (urinary pH 7.3-8.5). In measurements of surface activity, GS-9822 showed an unusual transition from a monolayer to a bilayer at the air/water interface with decreasing pH, attributed to the strong association between drug molecules in adjacent bilayer leaflets and expected to be highly disruptive to the urothelium. Structural analysis of GS-9822 and GS-9695 showed zwitterionic characteristics over the range of pH expected in cynomolgus monkey but not rat urine. This exotic surface behavior is unlikely with BI-224436 since it would transition from neutral to cationic (never zwitterionic) with decreasing pH. These data provide useful insights to guide discovery and development of NCINIs, related compounds, and zwitterions.