Antiviral Nucleotide Incorporation by Recombinant Human Mitochondrial RNA Polymerase Is Predictive of Increased In Vivo Mitochondrial Toxicity Risk.

Nucleoside or nucleotide inhibitors are a highly successful class of antivirals due to selectivity, potency, broad coverage, and high barrier to resistance. Nucleosides are the backbone of combination treatments for HIV, hepatitis B virus, and, since the FDA approval of sofosbuvir in 2013, also for hepatitis C virus (HCV). However, many promising nucleotide inhibitors have advanced to clinical trials only to be terminated due to unexpected toxicity. Here we describe the in vitro pharmacology of compound 1, a monophosphate prodrug of a 2'-ethynyluridine developed for the treatment of HCV. Compound 1 inhibits multiple HCV genotypes in vitro (50% effective concentration [EC50], 0.05 to 0.1 μM) with a selectivity index of >300 (50% cytotoxic concentration [CC50], 30 μM in MT-4 cells). The active triphosphate metabolite of compound 1, compound 2, does not inhibit human α, β, or γ DNA polymerases but was a substrate for incorporation by the human mitochondrial RNA polymerase (POLRMT). In dog, the oral administration of compound 1 resulted in elevated serum liver enzymes and microscopic changes in the liver. Transmission electron microscopy showed significant mitochondrial swelling and lipid accumulation in hepatocytes. Gene expression analysis revealed dose-proportional gene signature changes linked to loss of hepatic function and increased mitochondrial dysfunction. The potential of in vivo toxicity through mitochondrial polymerase incorporation by nucleoside analogs has been previously shown. This study shows that even moderate levels of nucleotide analog incorporation by POLRMT increase the risk of in vivo mitochondrial dysfunction. Based on these results, further development of compound 1 as an anti-HCV compound was terminated.