Background
Latency-reversing agents (LRAs) have shown promise in reactivating HIV reservoir within “shock and kill”approaches. However, molecular mechanisms underlying HIV reactivation within CD4+ T cells as well as predictors of LRA efficacy remain partially unknown. Here, we investigated the transcriptomic and whole genome DNA methylation profile of CD4+ T cells cultured in vitro with or w/o two LRAs and assessed for associations with HIV viral reservoir inducibility.
Methods
Large-scale samples (2,5 – 9,4×109 PBMCs) were collected through leukapheresis from 9 study participants (10 – 27 years old, 7 female and 2 male) with perinatally acquired HIV who initiated ART within the first year of life and maintained viral suppression for several years (IQR: 13-22 years). The HIV reservoir was characterized by Flip-Seq, IPDA and QVOA. HIV inducibility was calculated as infectious units/106 cells (by QVOA) divided by the number of intact proviruses/106 cells. Purified CD4+ T cells were cultured in-vitro for 20h with two LRAs: fimepinostat (CUD) and Ingenol-3-angelate (PEP), left unstimulated (UNS) or cultured with PMA+ionomycin+IL2 as positive control (CTR). After stimulation, RNA and DNA were extracted and sequenced using the long-reads Oxford Nanopore. DNA methylation was analyzed for both 5-methylcytosine (5mC) and N6-methyladenosine (6mA). An algorithm was developed to identify genes and genomic sites potentially associated with HIV inducibility. Transcriptome and epigenome were integrated using 4 public databases by intersecting the methylated sites with the genomic position of enhancer binding sites.
Results
The analysis of 3 distinct OMIC layers revealed 2743 Features Associated with Inducibility (FAI) upon four distinct experimental conditions: UNS, CUD, PEP, and CTR vs UNS. CUD showed the highest number of FAI at RNA and 5mC level (see Fig 1). The intersection of FAI informed by RNA-Seq and 5mC showed that Vacuolar Protein Sorting 8 (VPS8) and Phosphodiesterase 8 (PDE8A) genes were associated with HIV inducibility at both OMIC layers. These proteins are involved in HIV replication mechanisms, potentially representing important in vitro correlates of the potential of HIV reactivation by LRAs.
Conclusions
MultiOMIC analysis of CD4+ T cells upon LRA stimulation revealed molecular features associated with HIV inducibility. These features may shed light on mechanisms of latency reversal and could guide the selection of new and more effective LRAs.