5 μM and 0.08 μM respectively. However, their triphosphates were equally effective against HCV NS5B polymerase (IC50 values both 0.3 μM). In the replicon system, the triphosphate of the N-Nuc (MK608) was formed more efficiently than that of the C-Nuc1, thus explaining the lower activity of the C-Nuc1. However, in primary human hepatocytes, C-Nuc1 was phosphorylated to the triphosphate more efficiently than the N-Nuc (MK608). This illustrates the importance of using primary human cells. C-Nuc1 seemed to have a benign in vitro toxicity profile, including not inhibiting the mitochondrial DNA polymerase-gamma, but it had very significant toxicity

in animals. In a collaboration between Gilead and Craig Cameron at Pennsylvania State University, the researchers sought to identify the toxicity target(s) for ribonucleotide analogues, including C-Nuc1 and selleck kinase inhibitor others that had been stopped in Phase II trials. These studies showed a correlation between C-Nuc1 and the Phase II candidates, R1626, NM283 and BMS986094/IDX184. All the latter were efficiently incorporated into RNA by the mitochondrial RNA polymerase (>70% of the corresponding natural nucleotide). The triphosphate of C-Nuc1 was also an efficient substrate (22% the rate of ATP). In contrast, the active nucleotide analogs, formed by drugs approved

for the treatment of HCV, were poor substrates. Ribavirin was poorly incorporated (about 5%) and sofosbuvir was below the limit of detection (= 0.02%). More extensive Pictilisib clinical trial in vitro and cell culture evaluation of the compounds could have saved the expense of taking them into clinical trials. Understanding

that the mitochondrial RNA polymerase is an important target for ribonucleotide toxicity, the Gilead team sought analogs that were not incorporated by this polymerase. Adding a CN group to the 1′ position of C-Nuc1 did not change its activity as an HCV NS5B polymerase inhibitor (IC50 0.3 μM) but it did reduce incorporation in the mitochondrial RNA assay (<0.02%). However, in the absence of a nucleotide prodrug to bypass the first GNA12 phosphorylation step, the resulting di-substituted nucleoside analog would not be a drug candidate because it was not efficiently activated in cells. Application of a nucleotide prodrug strategy allowed this nucleotide to be pursued further. Oral absorption, delivery of the monophosphate into hepatocytes and high hepatic extraction were criteria used as part of the prodrug optimization process. A nucleotide prodrug, GS-464335 (a mixture of diastereoisomers at phosphorous) was well absorbed in dogs (>80%). Comparing the pre-hepatic and post-hepatic plasma drug levels, about 80% of the absorbed drug was taken up by the liver. Inside cells, GS-464335 was converted to the corresponding monophosphate which was efficiently converted to the triphosphate. At 24 h, the triphosphate levels remained about 2-fold above the IC90 value. A pure stereoisomer was selected and later named GS-6620.