Background: Our knowledge of the source of rebounding HIV after analytic treatment interruption (ATI) is limited. Understanding the origin of HIV variants during early rebound would provide insight into the composition of the HIV reservoir and has implications for the design of curative interventions. We examined HIV single-genome sequences (SGS) of ten AIDS Clinical Trials Group (ACTG) participants who underwent an ATI and compared their rebounding virus with those from pre-ART plasma, and latent and expressed HIV reservoirs as defined by on-ART PBMC cell-associated DNA and RNA (CA-DNA and CA-RNA).
Methods: Ten ACTG participants were identified who enrolled in a first-line ART initiation study and subsequently participated in an ATI study while on suppressive ART. SGS were obtained from pre-ART plasma, PBMCs collected during virologic suppression, and from early post-ATI plasma. The genetics of the HIV populations at each time point and in each compartment were compared phylogenetically and using tests for panmixia. Sequence diversity was calculated by average pairwise distance (APD).
Results: PBMCs were collected a median of 4 years after ART initiation and post-ATI plasma were obtained a median of 1 week after initial detectable viremia with a median viral load of 17,286 HIV RNA copies/mL. A median of 25 SGS were obtained from each participant time point. Despite many years of suppressive ART, the viral diversity of on-ART CA-RNA and CA-DNA were similar to that of the pre-ART plasma HIV RNA (median APD of CA-RNA vs. CA-DNA vs. pre-ART plasma RNA: 1.1% vs. 1.2% vs. 1.2%, repeated measures ANOVA P=0.70). Hypermutated HIV sequences were detected in both CA-DNA and CA-RNA. For most participants, the oligoclonal populations of rebounding HIV were found to have a population structure not significantly different than HIV from pre-ART and on-ART samples, although the rebounding virus were most closely related to either CA-RNA or CA-DNA. In one instance, we found CA-RNA that closely matched rebound viremia. The diversity of pre-ART plasma HIV RNA in this participant was low (0.23%), indicating that treatment may have been initiated early after infection.
Conclusions: HIV diversity in the latent and expressed PBMC HIV reservoirs were similar to that found in pre-ART plasma RNA despite several years of ART. In some cases, early rebound virus closely matched those found in the expressed PBMC HIV reservoir, but in other cases, rebounding HIV may have originated from sources other than circulating PBMCs.