Identifying the source of viral rebound during a monitored analytical treatment interruption (ATI) would reveal potential targets for cure strategies. Therefore, we examined the genetic composition of proviral DNA in different subsets from participants on antiretroviral therapy and compared this to rebounding virus after an ATI.
Eleven participants underwent a monitored ATI and were sampled from different anatomical sites prior to and after the ATI. From the peripheral blood, Naïve (TNA), central (TCM), transitional (TTM) and effector (TEM) memory CD4+ T cells were sorted as were CD45 cells from gut-associated lymphoid tissue (GALT). Using single-genome sequencing (SGS) the env region of HIV DNA and plasma-derived RNA was sequenced. In an ongoing study, Full-Length Individual Proviral Sequencing (FLIPS) and Integration Site Loop Amplification (ISLA) assays were performed on the T cell subsets from 2 participants.
For participant STAR10, 87 integration sites (IS) and 113 proviral genomes were sequenced while only 3 unique intact proviruses (3%) were identified. A cluster of 17 identical defective proviruses were linked to an IS (9% of all IS) in STAT5B located in TCM, TNA, TEM and TTM. When comparing the FLIPS to SGS env sequences a 100% match was found between one defective provirus and one plasma HIV RNA sequence after rebound. For participant STAR11, 37 IS and 105 proviral genomes were sequenced yielding 14 intact proviruses (13%) with the highest proportion found predominantly in the TEM subset (n=13, 45%). Four different clusters of identical sequences could be identified of which 2 (n=3 and n=9) consisted of intact TEM sequences with the smaller cluster linked to an IS in ZNF274. A 99% match between 2 env from rebounding plasma RNA and this smaller cluster of intact proviral genomes was identified.
Comparing proviral sequences and their IS to plasma-derived RNA sequences after an ATI reveals additional information in terms of the source of viral rebound. However, this comparison is complicated by multiple factors. For example, we found a plasma-derived RNA sequence obtained during viral rebound matched a defective proviral sequence which highlights the problem of using one HIV RNA subgenomic region for identifying replication-competent virus. In addition, ongoing viral replication during rebound may prevent a 100% match with genetically intact proviral sequences making it challenging to determine the absolute source of rebound.