Off-target autophagy disruption associated with a novel liver toxicity in dogs for a highly basic heterobifunctional protein degrader.

INTRODUCTION: The observation of hepatobiliary toxicity in a repeat-dose Good Laboratory Practice-compliant dog toxicology study was a primary driver for the deprioritization of a preclinical heterobifunctional protein degrader candidate, Compound X. The pathology of large bile duct epithelial hyperplasia was novel and its pathogenesis unknown. METHODS: In this study, a thorough characterization and mechanistic investigation are presented with both short-term exploratory animal studies and in vitro recapitulation. Cholangiocytes, epithelial cells lining bile ducts, were the toxicity target, with an accumulation of Compound X in both bile and the affected cells. RESULTS: Proteome profiling and high-content imaging highlighted a significant disruption to autophagy, with a dramatic increase in autophagosomes. A whole genome CRISPR-Cas9 screen identified the lysosomal V-ATPase as a key mediator of cell sensitivity to Compound X. This was further demonstrated by a rescue of toxicity in vitro by the V-ATPase inhibitor, bafilomycin A1, directly linking the pathology to disruption of the autophagy-lysosome system. Importantly, neither the degradation target of Compound X nor the E3 ligase it recruits, CRBN, were similarly implicated. An analog degrader with differentiated physicochemical properties, most notably a reduced pKa, was identified with significantly reduced hepatobiliary toxicity despite similar bile concentration. Together, these data indicate that uptake of the large, basic, and lipophilic Compound X into cholangiocyte lysosomes drives a unique bile duct pathology. DISCUSSION: This mechanism is a further demonstration of how the physicochemical properties of bifunctional degraders may challenge preclinical development, and its elucidation provides a path forward for development of degrader compounds with improved toxicity profiles.