Hepatitis C virus diversity and evolution in the full open-reading frame during antiviral therapy

Abstract

Background: Pegylated interferon plus ribavirin therapy for hepatitis C virus (HCV) fails in approximately half of genotype 1 patients. Treatment failure occurs either by nonresponse (minimal declines in viral titer) or relapse (robust initial responses followed by rebounds of viral titers during or after therapy). HCV is highly variable genetically. To determine if viral genetic differences contribute to the difference between response and relapse, we examined the inter-patient genetic diversity and mutation pattern in the full open reading frame HCV genotype 1a consensus sequences.

Methodology/principal findings: Pre- and post-therapy sequences were analyzed for 10 nonresponders and 10 relapsers from the Virahep-C clinical study. Pre-therapy interpatient diversity among the relapsers was higher than in the nonresponders in the viral NS2 and NS3 genes, and post-therapy diversity was higher in the relapsers for most of HCV's ten genes. Pre-therapy diversity among the relapsers was intermediate between that of the non-responders and responders to therapy. The average mutation rate was just 0.9% at the amino acid level and similar numbers of mutations occurred in the nonresponder and relapser sequences, but the mutations in NS2 of relapsers were less conservative than in nonresponders. Finally, the number and distribution of regions under positive selection was similar between the two groups, although the nonresponders had more foci of positive selection in E2.

Conclusions/significance: The HCV sequences were unexpectedly stable during failed antiviral therapy, both nonresponder and relapser sequences were under selective pressure during therapy, and variation in NS2 may have contributed to the difference in response between the nonresponder and relapser groups. These data support a role for viral genetic variability in determining the outcome of anti-HCV therapy, with those sequences that are more distant from an optimal sequence being less able to resist the pressures of interferon-based therapy.

Trial registration: ClinicalTrials.gov NCT00038974.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Viral titers in relapsers and nonresponders during the first 24 weeks of therapy.

Viral titers are shown as the log(titers [IU/mL]) at each time point. All relapsers (blue) dropped to the detection limit. One of the relapsers had a rebound of viral titers during the first 24 weeks. All others had rebounds later in therapy. All nonresponders (red) declined by less than 2.1 log10.

Figure 2. Unique amino acid variability among nonresponders and relapsers.

Amino acid variations found exclusively in one response class but not in the other were compared between nonresponders and relapsers. Statistical significance of the difference in the number of variations was compared using the Mann-Whitney test. A) Pre-therapy. B) Post-therapy. The line represents the median value, and the box represents the 25–75% range. Whiskers represent samples within 1.5 box lengths, and the ° and * represent outliers between 1.5 and 3 box lengths and beyond 3 box lengths respectively.

Figure 3. Protein distance among the nonresponder and relapsers sequences.

The mean protein distance among the nonresponders and relapsers was compared, and the statistical significance was evaluated using the Mann-Whitney test. A) Pre-therapy. B) Post-therapy.

Figure 4. Protein distance in pre- and post-therapy samples.

Comparison of pre- and post-therapy protein distances in A) nonresponders and B) relapsers. Statistical significance was determined using the Mann-Whitney test.

Figure 5. Protein distance in nonresponders, relapsers, and responder pre-therapy sequences.

Comparison of the protein distances in three-phenotypes of patients.

Figure 6. Mutations in nonresponder and responder sequences over the course of therapy.

A) The number of changes in nonresponders and relapsers were compared using Mann-Whitney test. One relapser does not appear on the chart (113 changes). B) The sums of the BLOSUM90 score for all of the changes occurring in a given sample were compared between nonresponders and relapsers. Totals were compared using a Kolmogorov-Smirnov test. C) Distribution of BLOSUM90 scores for E2. D) Distribution for BLOSUM90 scores for NS2.

Figure 7. The dN/dS is similar in nonresponders and relapsers.

A. Comparison of dN/dS ratios between nonresponders and relapsers. dN/dS was determined by the Nei-Gojobori method. B. dN/dS in ten codon windows as determined by SWAPSC using the Li-Kimura method. Each sample is represented by a different color and each dot indicates a specific 10 codon window. Regions of positive selection are overlapping 10 codon windows with dN/dS >1, and are denoted as a collection of points on the graph forming an upward spike.

Figure 8. Quasispecies breadth is lower in relapser than nonresponder sequences.

Nonresponder and relapser sequences were analyzed using Phred and positions at which the dominant nucleotide was present at