Factors that predict post-treatment control (PTC) after stopping antiretroviral therapy remain undetermined. In animal models, HIV-specific T cell immunity has been associated with improved viral control following treatment interruption (TI). We conducted a prospective study of treatment interruption in primary HIV infection (PHI) to look for PTC and determine immune correlates of protection.
PITCH was a prospective open-label study of participants who initiated ART within an estimated 6 months of PHI with viral suppression (<50 copies HIV RNA/ml) for at least 2 years and HIV DNA levels ?3.25 log copies/million CD4 cells. Participants underwent TI and were followed weekly. The decision to re-start ART was determined using a rule-based algorithm incorporating degree and duration of viraemia. T cell immunity to HIV antigens was assessed by IFN? ELISpot; 200,000 cryopreserved PBMCs were stimulated with Gag, Pol, Acc, Env or Nef overlapping peptide pools. 50,000 cryopreserved PBMCs/well were used for positive controls PHA, and antigen-specific controls CMV and FECT. A total of 6 timepoints, before and after TI, were analysed for each participant and associations between immune responses and viraemia were determined by Spearman correlations.
11 participants were screened and 7 underwent TI. Of the 7 participants who agreed to take a TI, all rebounded and re-started ART achieving fully suppressed plasma viral loads. There were no major adverse events, including during the ATI. Time to rebound after TI ranged from 15 – 170 days (median 36 days) with 4 participants showing evidence of viral control for >30 days. 1 participant showed evidence of longer post-treatment control (170 days) with stepwise increases in Gag-specific responses after viral blips. There was heterogeneity in T cell responses following TI. With regards to immunodominance, participants tended to show dominant responses towards Gag, Env and Nef prior to TI. After ART resumption, all participants had dominant responses toward Gag and/or Pol. Only Env-specific responses were positively correlated with viral load (r=0.44, p=0.0015).
Treatment interruptions enhance and can cause changes in the immunodominance of host T cell responses. Further studies will resolve epitope-specific responses in the context of times to rebound.