Infection by Hepatitis C virus (HCV) is seen in 15-30% of all HIV-infected individuals in the US, due to the shared routes of viral transmission [1, 2]. The introduction of antiretroviral viral therapy (ART) has improved clinical outcomes in patients infected with HIV. However, liver disease has become a leading cause of morbidity and mortality in this population [3, 4]. HIV-HCV co-infection is also associated with higher HCV viral levels in serum [5, 6], rapid progression of liver disease [7] and lower efficacy of pegylated interferon [peg-IFN] plus ribavirin treatment [5, 8]. Development of biomarkers that can accurately predict therapeutic responses are needed to optimize HCV therapy in this co-infected population. Previously, HIV-HCV coinfected patients who were not responsive to peg-IFN and ribavirin HCV therapy were found to have a gene activation signature present before treatment indicative of the activation of many immune-related molecules including interferon stimulated genes [9]. Quantitative and qualitative humoral responses over the course of HCV therapy among HIV co-infected subjects have never been studied. The ability to clearly predict and monitor outcomes of HCV infection in a robust and simple serological test would have obvious clinical utility. Recently, luciferase immunoprecipitation system (LIPS) assays has been used to accurately quantify antibody responses against various viral pathogens [10]. In the present study, we used LIPS antibody profiling against the whole proteome of HCV and partial proteome of HIV to evaluate its utility in predicting and monitoring response to HCV treatment in HCV-HIV coinfected individuals.

This was a prospective, open-label trial performed at the National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health at Bethesda, Maryland. All 29 patients signed informed consent approved by the NIAID institutional review board. HIV-HCV coinfected patients were treated with peg interferon alpha-2b at 1.5 μg/kg subcutaneously every week (Peg-Intron; Schering-Plough) and ribavirin daily (Rebetol, Schering-Plough, at 400mg Qam and 600mg Qpm for < 75kg, 600mg twice per day for > 75kg) for 48 weeks and followed up for 24 weeks after the end of treatment. All patients irrespective of virologic response were treated for 48 weeks. One patient discontinued Ribavirin at week 24 due to refractory anemia but continued Peg-interferon until week 48.

Antibody titers in all patients and two additional control samples were evaluated against six different recombinant HCV antigens, essentially derived from the whole proteome of HCV. A heat map, employing log₁₀-transformed antibody titers, was used to display the differing antibody responses toward the six antigens in individual samples from these subgroups (figure 1A). As shown by the heatmap, LIPS profiling against these 6 HCV antigens clearly distinguished the 29 HCV-infected sera from the 2 uninfected control sera. The most useful antibody response was directed against the HCV core, in which all but one of the 29 HCV-HIV coinfected samples was positive. The second most useful antibody response was against NS3 (figure 1A). The other 4 HCV proteins, E1, E2, NS4 and NS5 showed variable immunoreactivity with HIV-HCV-coinfected sera (figure 1A). Of interest, one NR patient was completely negative for anti-core, anti-E1 and anti-E2 antibodies, but showed strong immunoreactivity to three other non-structural HCV proteins (figure 1A). Antibody titers against the six HCV antigens correlated poorly with each other (r_s>0.60) suggesting marked heterogeneity in humoral responses (supplementary Table I).

Our study suggests that highly quantitative HCV proteome-wide antibody responses can be a valuable tool for monitoring and predicting HCV therapeutic responses among HIV co-infected patients. Few studies have examined the utility of anti-pathogen antibodies for predicting and monitoring drug therapy. LIPS provided a clearer summary of the marked patient variability in humoral responses to the whole HCV proteome than previously reported. None of the baseline antibody responses to the 6 different HCV proteins predicted response to HCV therapy. This suggests that pre-existing host humoral responses to HCV generally do not impact the response to HCV therapy. Previously, pre-treatment HCV mono-infected patients not normalized for HCV viral load showed higher anti-NS4A and anti-NS5a antibody titers in the SVR patients compared to the NR [12]. It should be noted that our study differs from these published studies in that the patient population were HCV/HIV-coinfected controlled for HCV viral loads. Nevertheless, one NR patient completely lacked anti-core, anti-E1 and anti-E2 antibodies, but had high levels of other HCV antibodies, possibly explaining the lack of responsiveness to HCV therapy. Since this NR patient (HCV-genotype I) had ample antibodies to HIV and nonstructural HCV proteins, it is likely that selective B-cell exhaustion or deletion of certain populations of plasma B cells may have occurred [13, 14]. Intriguingly, anti-p24 HIV antibodies detected in the pre-treatment samples inversely correlated with response to treatment. The highest anti-p24 antibody titers were in the NR, intermediate titers were in the RELAP and the lowest titers were in the SVR. The higher anti-p24 antibody titers in NR compared to the SVR subgroup suggests that some of the NR patients who respond poorly to IFN treatment may have an abnormal immune response to HIV.