CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

suppressed HIV+ women by fluorescence in situ hybridization in combination with immunohistochemistry (n=4). Results: Cervical CD4+ TRM cells expressed a unique repertoire of clusters of differentiation on their surface compared to non-CD4+TRM that shaped them highly susceptible to HIV infection (α4β7, CXCR4, CXCR6 and CCR6) while revealed self-renewing potential (CD122, CD132, CD127) (p<0.032 for all markers). CD4+TRM preferentially sustain HIV infection ex vivo and harbored more HIV protein (p=0.003) and DNA (p=0.062) than non-TRM from the same tissues. Conclusively, cervical tissue from ART-suppressed HIV+ women contained up to two logs more molecules of viral DNA per cell (median 12,929) compared to blood (median 1,092), and the CD4+TRM fraction was the principal contributor to this reservoir (median of 98.25%). Further, persistent viral RNA was detected within CD69 positive cells in cervical samples from ART- suppressed HIV+ women. Conclusion: Here we identified first the cervical mucosa as an overlooked HIV sanctuary and second CD4+TRM as a critical cellular reservoir. Thus, the contribution of CD4+TRM to viral persistence in tissues requires major attention in order to reach a functional cure in HIV infected patients Gilles Darcis 1 , Neeltje A. Kootstra 1 , Berend Hooibrink 1 , Thijs van Monfort 1 , Kevin Groen 1 , Suzanne Jurriaans 1 , Margreet Bakker 1 , Carine Van Lint 2 , Ben Berkhout 1 , Alexander Pasternak 1 1 Academic Medical Center, Amsterdam, Netherlands, 2 Université Libre de Bruxelles, Brussels, Belgium Background: CD32 was reported to mark the HIV reservoir, but several recent reports challenged this finding. We aimed to confirm or deny the role of CD32 as a marker of the viral reservoir and to further characterize the phenotype of these CD32+CD4+ T cells. Methods: CD32 expression and co-expression of HLA-DR, PD-1, TIGIT, LAG-3 was measured by flow cytometry on PBMCs from ART-suppressed HIV-infected individuals with undetectable plasma viremia. HIV DNA was quantified in bulk PBMC samples and in CD32+ and CD32- fractions of CD4+ T cells obtained by magnetic sorting (negative selection to isolate CD4+ T cells followed by positive selection to isolate CD32+CD4+ cells). Results: The median frequency of CD32+CD4+ T cells in HIV-infected individuals (n=19) was 0.07%. We found a positive correlation between the percentage of CD32+CD4+ T cells and total HIV DNA load in PBMCs (rho=0.58; p=0.012). CD32+CD4+ T cells demonstrated increased expression of LAG-3 (p=0.016), TIGIT (p=0.016) and HLA-DR (p<0.0001) compared with CD32- CD4+ T cells. CD32+CD4+ T cells were not enriched for HIV DNA (normalized to the total cell numbers) compared with CD32-CD4+ cells. However, the CD32+ fraction was found to contain many B cells, due to the abundance of CD32+ B cells in the input sample, of which some remained after one round of CD4+ T-cell purification. Remarkably, when HIV DNA was normalized to CD3G T-cell-specific mRNA, a significant positive enrichment in the CD32+ fraction was observed (p=0.0001). Therefore, we optimized the protocol to isolate a more pure fraction of CD32+CD4+ T cells from an additional set of HIV-infected individuals (n=19). An extra round of CD4+ purification resulted both in a 19-fold decrease in B-cell contribution to the CD4+CD32+ fraction (p<0.0001) and in an 11-fold enrichment in HIV DNA in this fraction (p=0.0007), the latter observed even when HIV DNA was normalized to the total cell numbers. In a subset of these individuals (n=8), we performed two additional rounds of CD32+ positive selection and observed a very high enrichment (mean 300-fold) for HIV DNA in the CD32+CD4+ fraction. Conclusion: We confirm that CD32+CD4+ T cells are highly enriched in HIV DNA. Our results highlight the importance of obtaining a sufficiently pure CD32+CD4+ T-cell fraction for analysis, and provide a plausible explanation for the negative results recently obtained by others. Our data suggest that the CD32+CD4+ T cells are activated, and that they often co-express the immune checkpoint receptors TIGIT and LAG-3.

346 CD32+CD4+ T CELLS ARE ENRICHED IN HIV DNA

Poster Abstracts

347 DIFFERENT T-CELL SUBSETS CONTAIN INTACT PROVIRUSES IN BLOOD AND LYMPH NODE DURING ART Bonnie Hiener 1 , Katie Fisher 1 , Bethany A. Horsburgh 1 , Timothy E. Schlub 2 , Vincent Morcilla 1 , John-Sebastian Eden 1 , Susanne von Stockenstrom 3 , Jeffery Milush 4 , Teri Liegler 4 , Rebecca Hoh 4 , Rémi Fromentin 5 , Nicolas Chomont 5 , Steven G. Deeks 4 , Frederick M. Hecht 4 , Sarah Palmer 1 1 The Westmead Institute for Medical Research, Westmead, NSW, Australia, 2 University of Sydney, Camperdown, NSW, Australia, 3 Karolinska Institute, Stockholm, Sweden, 4 University of California San Francisco, San Francisco, CA, USA, 5 Université de Montréal, Montreal, QC, Canada Background: Understanding the distribution of genetically intact, and therefore potentially replication-competent, proviruses in different CD4+ T cell subsets and within different anatomical sites is important for identifying targets for future eradication strategies. Methods: Naïve (NV), central (CM), transitional (TM) and effector memory (EM) cells were isolated from both the peripheral blood (PB, 13 participants) and lymph nodes (LN, paired, 5 participants) of HIV positive individuals on long-term ART (3-17 years). HIV proviral sequences were obtained using the Full-Length Individual Proviral Sequencing assay (FLIPS) which amplifies single, near full-length (92% of genome) HIV proviruses followed by Next-Generation Sequencing. Genetically intact HIV proviruses were identified as those lacking inversions, stop codons/hypermutation, insertions, deletions or frameshifts. Results: We sequenced 1893 genomes and identified genetically intact proviruses in all cell subsets except LN EM (3 participants). The frequency of infection with any HIV genome (intact and defective) differed across the cell subsets in LN and PB (P<0.001). We observed the highest levels in TM and EM cells, followed by CM and then NV cells, with the trend significant in blood (P<0.05). The infection frequency of genetically intact proviruses also differed across the subsets in blood (P<0.001) and LN (P=0.02). In blood, intact genomes showed a trend of EM>NV>TM>CMwith evidence for EM>CM (P=0.01). In LN, the intact genomes showed a different trend: NV>CM>TM>EM. Overall, the infection frequencies of intact proviruses within the LN and PB anatomic sites were similar (P=0.67), but this result was sensitive to 2 participants with high levels of intact proviruses in LN NV cells (10% and 26%were intact of all LN NV sequences). The intact LN sequences were genetically unique. We did not identify any identical intact sequences between PB and LN.

CROI 2019 126