Background:  The relationship between HCV-specific immune responses and outcome of treatment with interferon (IFN) and ribavirin (RBV) is not well defined, but may allow us to better explain patient characteristics associated with treatment effectiveness. We hypothesized that individuals with more vigorous HCV-specific Th1-like immune responses would be more likely to clear HCV on treatment.

Methods:  A5071 compared PEG-IFN-a-2a or standard IFN, both with RBV, in 133 subjects, with a primary endpoint of virological response at 24 weeks and a secondary endpoint of sustained virological response (HCV <600 IU/mL 24 weeks after therapy end). PBMC were isolated at baseline (n = 108 representative subjects), week 24 (n = 93) and week 72 (n = 57). ELISpot tests were performed for IFN-g and IL-10 using HCV antigens core, NS3 and NS5, Candida, and PHA. Mean spot-forming cells (SFC) in control wells were subtracted from antigen wells and reported as SFC/million PBMC. Data were analyzed using nonparametric methods and recursive partitioning (CART).

Results:  The median baseline CD4 count was 486 (range 177 to 1001). Consistent with prior reports of low frequency HCV-specific responses, median values for HCV-specific responses were 0 at baseline, week 24, and week 72. Among those with responses, the range of HCV-specific responses was diminished at 24 weeks compared to baseline. Median values for IFN-g secretion to Candida at baseline, week 24 and week 72 were, respectively, 24 SFC, 13 SFC, and 16 SFC while for IL-10 they were 1.3 SFC, 0 SFC, and 4.7 SFC. Overall sustained virologic response in this trial was 19.4%. In univariate analysis, there were no differences between subjects with sustained virologic response and no sustained virologic response in terms of IFN-g or IL-10 secretion to HCV proteins or Candida at baseline, week 24, or week 72. There was also no correlation between entry CD4⁺ cell count and baseline or week 24 HCV-specific or Candida responses. In a multivariable CART model, predictors of sustained virologic response were (in hierarchical order) genotype non-1, a <2 SFC decrease in IFN-g secretion to NS5 between baseline and week 24, a < 103 SFC decrease in IFN-g secretion to Candida, HAI score for fibrosis >1, and age <46 years, with a sensitivity of 0.95 and specificity of 0.64.

Conclusions: HCV-specific and recall immune responses diminished during treatment with IFN/RBV in this cohort of persons co-infected with HIV and HCV. A relative preservation of Th1-like HCV-specific immune responses at week 24 was predictive of sustained virologic response, and immune responses augmented clinical parameters in predicting sustained virologic response.

Keywords: Hepatitis C Virus; Treatment; Immunologic responses