Nucleotide Prodrug Containing a Nonproteinogenic Amino Acid To Improve Oral Delivery of a Hepatitis C Virus Treatment

Abstract

Delivery of pharmacologically active nucleoside triphosphate analogs to sites of viral infection is challenging. In prior work we identified a 2'-C-methyl-1'-cyano-7-deaza-adenosine C-nucleotide analog with desirable selectivity and potency for the treatment of hepatitis C virus (HCV) infection. However, the prodrug selected for clinical development, GS-6620, required a high dose for meaningful efficacy and had unacceptable variability due to poor oral absorption as a result of suboptimal solubility, intestinal metabolism, and efflux transport. While obtaining clinical proof of concept for the nucleotide analog, a more effective prodrug strategy would be necessary for clinical utility. Here, we report an alternative prodrug of the same nucleoside analog identified to address liabilities of GS-6620. A phosphoramidate prodrug containing the nonproteinogenic amino acid methylalanine, an isopropyl ester and phenol in the (S) conformation at phosphorous, GS2, was found to have improved solubility, intestinal stability, and hepatic activation. GS2 is a more selective substrate for hepatically expressed carboxyl esterase 1 (CES1) and is resistant to hydrolysis by more widely expressed hydrolases, including cathepsin A (CatA) and CES2. Unlike GS-6620, GS2 was not cleaved by intestinally expressed CES2 and, as a result, was stable in intestinal extracts. Levels of liver triphosphate following oral administration of GS2 in animals were higher than those of GS-6620, even when administered under optimal conditions for GS-6620 absorption. Combined, these properties suggest that GS2 will have better oral absorption in the clinic when administered in a solid dosage form and the potential to extend the clinical proof of concept obtained with GS-6620.

Keywords: GS-6620; GS2; HCV; antiviral; antiviral agents; carboxyl esterase; carboxyl esterase 1; cathepsin A; nucleotide prodrug.

FIG 1

Chemical structures of GS2, GS-6620, and key metabolites. Metabolite X for GS2 R is CH₃ and that for GS-6620 R is H.

FIG 2

Comparison of the HCV replicon activity of GS2 to expression of the hydrolases CatA and CES1. Protein levels were visualized by Western blotting. The EC₅₀s for GS2 are 380, 25,000, and 2,200 nM for 1a, 2a, and 1b replicons, respectively. The stable replicon cell lines were individual clones selected from transfected Huh-Lunet cells. Data were from n = 3 to 5 independent experiments done in triplicate. NA, not applicable.

FIG 3

Triphosphate level of prodrugs GS-6620 and GS2 in CES1-high (Huh-1C; open bar) and -low (Huh-Lunet; filled bar) cells. Cells were incubated with 10 μM the nucleotide prodrugs for 1 h. This study was done with GS2(R+S), which showed cell loading within 2-fold of the pure isomer GS2 under these assay conditions. Representative data are shown here with averages from duplicate measurements. There is no difference between GS-6620 and GS2 in 1C cells (P = 0.82) and significantly lower levels of triphosphate for GS2 than GS-6620 in Lunet cells (P = 0.05).

FIG 4

Maximal intracellular triphosphate concentrations following 1-h pulse incubations with primary hepatocytes from hamsters, rats, dogs, or humans (isolated from 3 separate human donors) with 10 μM GS-6620 (open bars) or GS2 (filled bars). Values are the means from duplicate samples collected for each individual primary hepatocyte donor.

FIG 5

Portal and jugular plasma concentrations versus time for GS2 following oral administration to portal vein-cannulated male beagle dogs (5 mg/kg; solution) (means ± standard deviations, n = 3). (A) The portal (closed circle) and jugular (open circle) plasma concentrations are shown. GS2 showed an F_a (faction absorbed) value of 52% and an E_H (hepatic extraction) value of 75%. (B) Portal plasma concentration versus time for GS2 (5 mg/kg; solution) and GS-6620 (5 mg/kg; suspension of micronized material) following oral administration to portal vein-cannulated male beagle dogs (means ± standard deviations, n = 3).

FIG 6

Liver levels of the active triphosphate in multiple animal species. (A) Per os dosing in Golden Syrian hamster of either GS-6620-containing mixture (10.6 mg/kg; open circle) or GS2(R+S) (9.64 mg/kg; closed circle); these doses contained 5-mg/kg equivalent of parent nucleoside. (B) Per os dosing in Sprague-Dawley rat of either GS-6620-containing mixture (21.11 mg/kg; open circle) or GS2(R+S) (50 mg/kg, dose normalized to 10 mg/kg parent nucleoside; closed circle). (C) Per os dosing in beagle dog of either GS-6620 (7.7 mg/kg, dose normalized to 2.59 mg/kg parent nucleoside; open circle) or GS2 (5 mg/kg, equivalent to 2.59 mg/kg parent nucleoside; closed circle). The values represent averages from terminal liver collections from two animals per time point.