CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

accessibility and gene expression in autologous CD4+ T cells were analyzed by ATAC-Seq and RNA-Seq. Results: We identified paired proviral sequences and integration sites for 100 intact and 84 defective proviruses. Among these, we detected several clusters of clonally-expanded proviruses exhibiting identical viral sequences and integration sites (8 intact clusters, 6 defective clusters). Relative to defective proviruses, intact proviruses were enriched for non-genic or pseudogenic sites (18% vs 8%, p=0.03) and more frequently displayed an opposite orientation relative to host genes (74% vs 57%, p=0.02). Additionally, intact proviruses were preferentially integrated either in relative proximity (2 participants) or with increased distance (1 participant) to active transcriptional start sites and to accessible chromatin regions, suggesting an enrichment of sites that are either more susceptible to transcriptional interference or located in genomic regions with more limited access to host transcriptional machinery, respectively. Conclusion: Our results suggest that prolonged ART is associated with a selection of intact proviruses with multiple discrete features of deeper latency, likely due to immune-mediated selection pressure. The intact reservoir may thus be vulnerable to interventions aimed at accelerating the selection of proviruses with deeper latency and reduced ability to fuel rebound viremia. 340 RELATIONSHIP BETWEEN INTACT HIV-1 PROVIRUSES AND PLASMA REBOUND VIRUSES Ching-Lan Lu , Joy A. Pai, Lilian Nogueira, Pilar Mendoza, Marina Caskey, Mila Jankovic, Michel Nussenzweig The Rockefeller University, New York, NY, USA Background: Combination antiretroviral therapy controls but does not cure HIV-1 infection due to a small fraction of cells harboring latent viruses that can produce rebound viremia upon therapy interruption. The circulating intact latent reservoir has been documented by either viral outgrowth assays (VOAs), or by amplifying near full length (NFL) proviral sequences from DNA. Analysis of samples obtained in clinical studies whereby individuals underwent analytical treatment interruption (ATI), showed little overlap between latent viruses from VOAs pre-ATI, and viruses isolated from plasma during viral rebound. To determine whether intact proviruses from DNA are more closely related to rebound viruses than those obtained from VOAs, we assayed 12 individuals who underwent ATI after infusion of two broadly neutralizing anti-HIV-1 antibodies (bNAbs). Methods: NFL proviral genomes were amplified from DNA extracted from CD4+ T cells obtained from 2 leukapheresis samples (pre- and post-bNAb infusions) from 9 individuals that maintained viral suppression for >12 weeks after ATI. A single pre-infusion sample was also available for 3 additional individuals that experienced viral rebound within 12 weeks of ATI due to pre- existing bNAbs-resistant proviruses in the latent reservoir. VOA was performed on all of these samples to determine the number the inducible replication- competent proviruses. Results: The env sequences from 435 intact proviruses obtained by NFL sequencing were compared with 650 latent viruses from VOAs and 246 plasma rebound viruses. Although, intact NFL and outgrowth culture sequences showed similar levels of stability and diversity with 39% overlap, the size of the reservoir estimated from NFL sequencing did not correlate with that obtained by VOA. Although all of the rebound viruses in plasma were >96% identical to at least one sequence from the reservoir, we did not find a single instance of 100% env identity among intact NFL sequences and rebound viruses. Moreover, only 12 out of 246 rebound sequences could be accounted for by mutation of reservoir sequences during the ATI window. However, 48% of the rebound viruses could be derived from recombination between intact NFL and/or VOA proviruses. Conclusion: We find that intact proviruses obtained from DNA overlap in part with those obtained by VOA, but do not overlap with rebound viruses. However, nearly half of all rebound sequences appear to be recombinants derived from circulating latent viruses characterized by VOA or NFL sequencing. 341 LANDSCAPE OF HIV-1 INTEGRATION SITES IN LYMPHOID TISSUE FROM ART-TREATED INDIVIDUALS Vincent H. Wu 1 , Christopher L. Nobles 1 , Leticia Kuri Cervantes 1 , Perla Del Río- Estrada 2 , Gustavo Reyes-Terán 2 , Frederic Bushman 1 , Michael R. Betts 1 1 University of Pennsylvania, Philadelphia, PA, USA, 2 National Institute of Respiratory Diseases, Mexico City, Mexico Background: The integration of HIV DNA into the host genome occurs from the activity of both viral proteins and host cellular factors. Prior analyses in cell

lines and peripheral blood mononuclear cells (PBMC) samples have assessed the HIV-1 integration site distribution and characterized regions such as active transcription units that favor integration. The host encoded LEDGF/p75 protein tethers HIV-1 integrase at active transcription units, accounting for this bias. To date, no studies have compared integration site distributions within tissue resident cells – namely in lymphoid tissues where HIV-1 is known to replicate and persist even after antiretroviral treatment (ART). Methods: Tonsil samples were collected from three ART-treated and HIV-1 infected patients to sort into non-naïve CD4 + cell subsets: circulating (HLA-DR - CD69 - ), tissue-resident (HLA-DR - CD69 + ), and germinal-center Tfh (HLA-DR - CXCR5 + PD-1 hi ) cells. Genomic DNA was extracted from the different cell types, sonicated, and then uniquely labeled by cell subset and patient. Libraries were created via ligation-mediated PCR and sequenced using the MiSeq platform (Illumina). Downstream analyses were done using the INSPIIRED software pipeline and R. Results: Of the integration sites sequenced, the majority across patients were found to be enriched within transcription units. Viral-host junctions were detected in 21 of the human autosomal chromosomes as well as the X chromosome in the combined dataset. We detected an average ratio (integration sites per one thousand cells) of 0.046 ± 0.025 for circulating cells, 0.137 ± 0.106 for tissue resident cells, and 0.067 ± 0.065 for Tfh cells. Curiously, we observe little to no overlap in integration site coverage between the circulating, tissue resident, and germinal-center Tfh cells by patient. Conclusion: Our findings agree with previous studies regarding HIV-1 integration within transcription units. However, the lack of gene overlap across cell subsets may suggest unique integration signatures in lymphoid tissue. The novelty of these results demonstrates the need for further analysis on integration sites in lymphoid resident cells as well as PBMC cells at different stages of HIV-1 progression. Analyzing the integration site signature in the lymphoid resident cells will help contribute more insight to the goal of understanding and eliminating the latent reservoir. 342 GENETIC AND AGE DISTRIBUTION OF LATENT HIV SEQUENCES IN CD4+ T-CELL SUBSETS Rachel Miller 1 , Bradley R. Jones 2 , Olivia Tsai 1 , Bemuluyigza Baraki 1 , Natalie Kinloch 1 , Hanwei Sudderuddin 1 , Hawley Rigsby 3 , Art Poon 4 , Rémi Fromentin 3 , Nicolas Chomont 3 , Jeffrey Joy 2 , Zabrina Brumme 1 1 Simon Fraser University, Burnaby, BC, Canada, 2 British Columbia Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada, 3 Centre Hospitalier de l’Université de Montréal, Montreal, QC, Canada, 4 University of Western Ontario, London, ON, Canada Background: HIV latency is the main barrier to cure, but our understanding of within-host latent proviral landscapes, particularly in distinct CD4+ T-cell subsets, is incomplete. We characterized sequence diversity and estimated age distribution of latent HIV sequences in naïve, central memory (CM), transitional memory (TM) and effector memory (EM) CD4+ T-cells from HIV-infected individuals with long-term viremia suppression on cART. Methods: CD4+ T-cell subsets were sorted from PBMC from 5 participants with a median 9 [IQR 9-13] years pVL suppression on cART. Proviral DNA was sequenced from these subsets using single-genome approaches (nef region); sequence compartmentalization was assessed using the Slatkin-Maddison (SM) test following maximum likelihood phylogenetic inference. For 4 participants, single-genome HIV RNA sequences were also obtained from a median 11 [IQR 6-15] pre-cART plasma samples spanning a median 8 [IQR 3-11] years; these data were used in a novel within-host phylogenetic approach to infer proviral sequence ages. Results: 539 Nef proviral sequences were isolated; 424 (78%) were genetically intact, of which 347 (82%) were unique. Intact sequence percentage varied between hosts (68-93%) and between T-cell subsets (naïve 71%; CM 79%; EM 86%; TM 88%). EM harbored the lowest % uniqueness (56%) and CM the highest (96%). Within-host latent HIV phylogenetic diversity varied between hosts (average tip-to-tip phylogenetic distances 2.1e-2-9.6e-2 nucleotide substitutions/site), though there was no clear relationship between within-host latent HIV diversity and length of uncontrolled viremia, or length of cART suppression. In 3 participants, proviral genetic diversity differed between subsets (Kruskal-Wallis p<0.05). Two of these participants, plus one other showed evidence of genetic compartmentalization (SM p<0.01). Proviral sequence ages varied markedly between hosts (median 10.4; max 23 years) and in two cases also differed between subsets: for example, in one participant, latent HIV sequences in naive T-cells were younger than the other subsets

Poster Abstracts

CROI 2019 124