Viral genome wide association study identifies novel hepatitis C virus polymorphisms associated with sofosbuvir treatment failure

Abstract

Persistent hepatitis C virus (HCV) infection is a major cause of chronic liver disease, worldwide. With the development of direct-acting antivirals, treatment of chronically infected patients has become highly effective, although a subset of patients responds less well to therapy. Sofosbuvir is a common component of current de novo or salvage combination therapies, that targets the HCV NS5B polymerase. We use pre-treatment whole-genome sequences of HCV from 507 patients infected with HCV subtype 3a and treated with sofosbuvir containing regimens to detect viral polymorphisms associated with response to treatment. We find three common polymorphisms in non-targeted HCV NS2 and NS3 proteins are associated with reduced treatment response. These polymorphisms are enriched in post-treatment HCV sequences of patients unresponsive to treatment. They are also associated with lower reductions in viral load in the first week of therapy. Using in vitro short-term dose-response assays, these polymorphisms do not cause any reduction in sofosbuvir potency, suggesting an indirect mechanism of action in decreasing sofosbuvir efficacy. The identification of polymorphisms in NS2 and NS3 proteins associated with poor treatment outcomes emphasises the value of systematic genome-wide analyses of viruses in uncovering clinically relevant polymorphisms that impact treatment.

Fig. 1. Association between hepatitis C virus (HCV) amino acids and sofosbuvir-based treatment outcome.

A Manhattan plot of association tests between HCV amino acids and treatment outcome. At each viral site only the p-value for the most associated amino acid is plotted. The dashed line indicates 15% false discovery rate. For the three sites significantly associated with outcome and the most associated site in the non-structural 5B (NS5B) protein, the amino acid associated with the lowest cure rate at the site and its position within its respective protein is indicated. B Schematic of HCV polyprotein.

Fig. 2. Covariation of treatment outcome polymorphisms (TOPs) and their impact on sustained virological response (SVR) rate.

A SVR rate for patients with different numbers of TOPs in baseline sequences. The dots indicate the SVR rate in each group and the lines indicate its 95% confidence intervals. The numbers in each group are shown at the bottom of the figure. B Covariation between the three novel TOPs and their association with treatment outcome. Combinations tested are listed in the table on the left. The squares show the estimated effect size (log(odds ratio)) for each group and the lines show its 95% confidence interval estimated from logistic regression. The P-value (logistic regression) and the SVR rate for each group are shown on the right.

Fig. 3. Reduction in viral load during the first week of therapy and its association with treatment outcome polymorphisms (TOPs).

A Reduction in log10 viral load between baseline and week one of therapy stratified by presence and absence of the residue associated with the lowest SVR rate at each TOP (A: Alanine, non-A: any amino acid other than alanine, V: Valine, non-V: any amino acid other than Valine). n = 507 patients. The mean change in viral load is shown as a black dot for each TOP and the lines indicate its 95% confidence intervals. P-values for difference in mean calculated using one-sided Mann–Whitney test. B Reduction in log10 viral load between baseline and week one of therapy against increasing numbers of TOPs (presence or absence of residue associated with the lowest SVR rate at the target sites). n = 507 patients. The mean change in viral load is shown as a black dot and the lines indicate its 95% confidence interval. P-value for association between number of TOPS and reduction in log10 viral load calculated using linear regression.