CROI 2018 Abstract eBook

Abstract eBook

Oral Abstracts

that active HIV-1 suppression during CD4+ T cell division could be one of the mechanisms that maintains the latent reservoir and that interfering with these cellular safeguards could contribute to accelerating latent HIV-1 clearance. 153 IDENTIFICATION OF INTEGRATION SITES OF INDUCIBLE HIV-1 USING HIV-1 RNA SORTSEQ Ales Varabyou 1 , C. Conover Talbot 1 , Hao Zhang 1 , Subul A. Beg 1 , Ross Pollack 1 , Haiping Hao 1 , Joseph B. Margolick 1 , Robert Siliciano 1 , Mihaela Pertea 1 , Ya-Chi Ho 2 1 Johns Hopkins University, Baltimore, MD, USA, 2 Yale University, New Haven, CT, USA Background: HIV-1-infected cells may undergo clonal expansion through integration into cancer-related genes. A significant proportion of the clonally expanded HIV-1-infected cells are replication competent. However, given the technical barrier to identify full-length HIV-1 proviral genome (requiring nested PCR spanning the HIV-1 genome) and integration site (requiring shearing of the HIV-1 genome) at the same time, it remains unclear whether HIV-1 proviruses which are integrated into cancer-related genes are intact or defective. Methods: To identify the integration sites of inducible HIV-1 proviruses, we developed a novel assay, HIV-1 RNA SortSeq, which can identify HIV-1-infected cells containing inducible HIV-1 proviruses for RNAseq analysis. Resting CD4+ T cells from HIV-1-infected individuals on suppressive antiretroviral therapy were activated with PMA/ionomycin in the presence of antiretroviral therapy for 18 hours to induce HIV-1 RNA expression. HIV-1 RNA expression serves as a surrogate to identify cells containing inducible HIV-1 proviruses. Cells were then fixed, permeabilized, and hybridized with HIV-1 RNA-specific fluorescent probes. Cells expressing HIV-1 RNA were isolated by flow cytometric sorting for subsequent RNAseq. We designed bioinformatic pipelines to identify the HIV-1- host genome junctions and the HIV-1 RNA genome. Results: Using HIV-1 RNA SortSeq, we identified HIV-1-host genome chimeric RNA from cells containing inducible HIV-1 from virally suppressed HIV-1-infected individuals. We found that some of the read-through transcripts contain both HIV-1 LTR and the host genomic RNA in cancer-related genes, indicating that HIV-1 proviruses which are integrated into cancer-related genes are actually inducible, producing readily detectable cell-associated HIV-1 RNA. Strikingly, we found that some of the HIV-1-host chimeric RNAs contain host exons of cancer-related genes splicing into canonical splice acceptors of HIV-1, indicating that HIV-1 integrated into cancer-related genes can produce aberrant HIV-1-host chimeric RNA encoding novel open reading frames. Conclusion: The novel HIV-1 RNA SortSeq assay allows identification of the integration sites of inducible HIV-1 proviruses and overcomes the technical barrier of simultaneous identification of the integration sites and the HIV-1 genome. We showed that HIV-1 proviruses which are integrated into cancer- related genes can not only be induced but also produce novel HIV-1-host chimeric RNAs through aberrant splicing. 154LB HERITABLE CLONE-SPECIFIC DIFFERENCES IN HIV-1 GENE EXPRESSION Edmond A. Atindaana , David Read, Kalyani Pyaram, Sarah Emery, Cheong- Background: Because host genes are regulated by a complex interplay of epigenetic and non-epigenetic mechanisms, it seems possible that integrated proviruses might be subject to host locus-specific regulation despite HIV-1’s independent promoter. Here, we studied how the genetic neighborhoods of integrated HIV-1 proviruses affected their expression patterns and whether clonal expression phenotypes were heritable. Methods: We developed a system for ensemble analysis of single-cycle replication outcomes and persistence properties of thousands of HIV-1 Env- proviruses. Proviruses marked with “zipcodes”-sequence tags within viral sequences that identify clonal progeny of individual integration events--were quantified by high-throughput sequencing. The systemwas benchmarked using selectably marked proviruses using adherent cells, which allowed informatics to be developed and optimized for a known number of integrants. These studies confirmed our ability to analyze zipcodes from both cell DNA and viral RNA, and to address the fidelity of one round of replication by analyzing the remobilization of first-round integrants on fresh cells. In a second experiment, Jurkat cells were infected with an Env-Vpr-PuroR virus harboring gfp in the nef ORF and were used to study the stability of HIV-1 gene expression. Results: Over 90% of proviruses maintained infectivity through one cycle of replication in low-restriction cells, which is on the upper end of previous fidelity Hee Chang, Jeffrey Kidd, Alice Telesnitsky University of Michigan, Ann Arbor, MI, USA

estimates. Surprisingly, based on virion-to-cell zip code ratios, many first-round integrants’ second-round success appeared limited by clonal differences in gene expression. Separately sorting pool halves into GFP-positive and GFP-negative subpopulations revealed reproducible large differences among zipcodes in GFP+ cell proportions that persisted through an additional week of separately propagating in subculture. In contrast, first sort GFP- cells largely remained silent through subsequent cell passaging, likely reflecting epigenetic proviral silencing. When analyzed by integration site, no strong correlation to any known genome features was observed. Conclusion: This new zipcoding system, shows that in transformed cells, provirus inactivation by replication-induced mutagenesis is lower than that estimated in earlier studies. However, individual proviruses’ expression phenotypes, as monitored by GFP+/total cell levels within integrant clones, differed reproducibly and fairly stably across over three logs of magnitude. 155 CD32+PD-1+ TFH CELLS ARE THE MAJOR HIV RESERVOIR IN LONG-TERM ART-TREATED INDIVIDUALS Alessandra Noto , Francesco Procopio, Jean-Marc Corpataux, Giuseppe Pantaleo Lausanne University Hospital, Lausanne, Switzerland Background: HIV-1 persists after many years of suppressive antiretroviral therapy (ART). It is well established that lymphoid tissues serve as primary anatomic sites for HIV-1 replication. We have previously shown that PD-1+/Tfh cells serve as primary cell compartment for HIV-1 replication in viremic patients and for persistent HIV transcription in long-term treated aviremic individuals. Recently, CD32 has been identified as a marker of HIV reservoir in blood memory CD4 T cells. We therefore investigated the distribution of the HIV reservoir in lymph nodes (LNs) memory CD4 T cell populations defined by the expression of CD32 and PD-1. Methods: LN biopsies were obtained from 10 HIV-1 infected viremic individuals naive to ART and 13 aviremic long-term treated individuals. Expression of CD32, PD-1 and a large panel of cell lineage, activation and migration markers was assessed in memory CD4 T cells using flow- and mass cytometry that included 30 isotope conjugated antibodies. HIV-DNA was measured in total CD32+ and PD-1+ cell populations and cell associated HIV-RNA in memory CD4 T cells isolated on the basis of PD-1 and CD32 expression (PD-1-CD32-, PD-1-CD32+, PD-1+CD32-, PD-1+CD32+). Results: Similar to what previously shown for PD-1+ CD4 T cells, the frequency of CD32+ CD4 T cells was increased in viremic as compared to treated individuals (3% vs 1.2% p<0.0001) and positively correlated with viremia while negatively with years of suppressive ART and CD4 cell count. CD32+ CD4 T cells were enriched for total HIV DNA as compared to CD32- cells in both viremic and treated individuals (average 210 fold, n=6, p=0.01, and 1.9 fold, n=7, p=0.03 respectively) but no difference was found as compared to PD-1+ cells. In both viremics and ART treated individuals CD32+ CD4 T cells were found predominantly within the PD-1+ and Tfh cell populations. Double positive CD32+ PD-1+ cells were phenotypically similar to CD32-PD-1+ cells and Tfh cells as indicated by the expression of ICOS, CD57, CD38, CXCR5, CD40L, CCR5 and CXCR4. CD32+PD-1+ CD4 T cells were enriched in cell associated HIV RNA as compared to CD32-PD-1- (average 7.6 fold), to CD32+ PD-1- (5.2 fold) and to CD32-PD-1+ cell populations (average 7.6 fold) (n=4 ART treated). Conclusion: Co-expression of CD32 and PD-1 defines a population of Tfh cells serving as the major HIV cell reservoir in HIV long-term ART treated individuals. 156 MAJORITY OF THE LATENT RESERVOIR RESIDES IN CD32A NEGATIVE CD4 T CELLS Lynn N. Bertagnolli 1 , Jennifer White 1 , Subul A. Beg 1 , Gregory Laird 1 , Francesco R. Simonetti 1 , Jun Lai 1 , Costin Tomescu 2 , Alexandra J. Murray 1 , Annukka Antar 1 , Hao Zhang 1 , Joseph B. Margolick 1 , Luis Montaner 2 , Robert Siliciano 1 , Janet Siliciano 1 1 The Johns Hopkins University, Baltimore, MD, USA, 2 Wistar Institute, Philadelphia, PA, USA Background: The persistence of HIV-1 in a stable reservoir in resting CD4+ T cells is the major barrier to curing HIV-1 infection. Cell surface biomarkers that could distinguish cells comprising this latent reservoir from uninfected cells have been lacking. If identified, these biomarkers could significantly enhance the progress towards an HIV cure. A recent study (Descours et al., 2017) identified the cell-surface protein CD32a, a low-affinity Fc receptor for IgG that

Oral Abstracts

56

CROI 2018