CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

Methods: Paired PB, TI and R samples from 20 HIV+ patients (HIV-RNA < 20 cop/ml) and 11 healthy controls were studied (median CD4+ T cells of HIV+ [cells/µl]: 696 [495-955]; time on ART [years]: 8 [5.5-10]). Expression of CXCR3 and PD-1 on sorted CD4+CD45RO+ T-cells memory subsets (T central memory (TCMCD27+CCR7+), T transitional memory (TTMCD27+CCR7-), and effector memory cells (TEMCD27-CCR7-)) was assessed. Cell associated total HIV DNA was quantified by qPCR within memory subsets (expression as cop/CD4*10^6). IP-10 in serum and tissue supernatant was measured by ultrasensitive digital ELISA (Simoa; Quanterix, pg/mL). Data were analyzed using Mann-Whitney and Spearman rank (correlation) tests. Results: HIV+ patients had significantly higher HIV-DNA levels in the TI compared to PB and R. Distribution of memory subsets in PB, TI and R was similar between HIV+ and controls. However, in HIV+ in all compartments PD1 and CXCR3 expression was significantly higher. HIV+ had increased expression of CXCR3 and PD-1 on memory subsets in the TI+R as opposed to PB: TI showed highest PD1 expression in TTM, whereas CXCR3 was highly expressed on TCM in R. A positive correlation was observed between CXCR3 expression in PB and HIV- DNA in TCM of TI (r=0.84 p=0.0021). IP-10 levels were increased in PB and TI of HIV+ [B: 32(23-54), TI: 21(8-28)], compared to controls [B: 17(7-25) p=0.04, TI: 0.75(0.75-16) p=0.04]. In HIV+ the highest IP-10 levels were found in PB compared to TI and R [B: 32(23-54)> TI: 21(8-28) = R: 25(12-42); p=0.03; p>0.05)]. Finally, serum IP-10 levels positively correlated to total HIV-DNA in TI (r=0.89 p=0.03). Conclusion: Highest levels of HIV-DNA are found in TI. Increased IP-10 levels as well as PD1 and CXCR3 expression in HIV+ TI correlate with total HIV DNA. Different anatomical compartments of the intestinal immune system contribute to maintain the HIV-reservoir.

expression of latent HIV. Viral kinetics and SGS analyses demonstrated the replication competence of reactivated virus. dQVOA reservoir measurements demonstrated a correlation with clinical markers in addition to a large reduction in the coefficient of variation, suggesting that understanding the key mechanisms of latency reversal will reduce the reliance upon stochastic HIV reactivation, which improves assay reproducibility. Conclusion: Differentiation into an effector phenotype supports more effective latency reversal of replication competent HIV in resting CD4+ T cells and offers key insights into mechanisms of HIV latency reversal that may offer potential targets for therapeutic interventions. 334 CYTOF CHARACTERIZATION OF THE IN VIVO LATENT HIV RESERVOIR IN BLOOD AND TISSUES Jason Neidleman 1 , Xiaoyu Luo 2 , Grace Xie 1 , Rebecca Hoh 1 , Peter W. Hunt 1 , Steven G. Deeks 1 , Warner C. Greene 2 , Marielle Cavrois 2 , Nadia R. Roan 1 1 University of California San Francisco, San Francisco, CA, USA, 2 Gladstone Institute of Virology and Immunology, San Francisco, CA, USA Background: One of the major hurdles for directly phenotyping the in vivo latent HIV reservoir is the need to stimulate these cells ex vivo. As the cells are stimulated, they change expression of many cell surface markers making precise phenotyping of the latent cells difficult. We recently established an approach called PP-SLIDE (Predicted Precursor as determined by SLIDE) that can effectively backtrack in time to establish the cell surface phenotype of a cell before infection (Cavrois et al 2017) using high-dimensional datasets generated by CyTOF phenotyping. This approach was implemented to chart the in vivo latent HIV reservoir and identify biomarkers of these cells. Methods: The ability of PP-SLIDE to trace the phenotype of a reactivated cell to its precursor (non-stimulated) state was first confirmed in an in vitro model of HIV latency. Then, freshly isolated blood and lymph node cells from ART- suppressed, HIV-infected individuals from the SCOPE cohort were stimulated with PMA/ionomycin, and reactivated cells (expressing Gag) were deep- phenotyped using a 39-parameter latency-focused CyTOF panel. PP-SLIDE was used to map the reactivated cells onto a T cell atlas of unstimulated cells created for each patient sample analyzed. Results: Comparison of latent to non-latent cells revealed unique signatures associated with latency. Receptors preferentially expressed on latently-infected T cells from blood include previously described markers such as PD1, TIGIT, and OX40, as well as novel ones including homing integrin α4β1. Latent cells from lymph nodes also preferentially expressed PD1, TIGIT, and OX40, and differed from latent cells in blood in that they exhibited features of resident memory cells and preferentially expressed the T follicular helper marker CXCR5 and the costimulatory molecules CD28 and ICOS. Conclusion: By simultaneously detecting the expression of 39 proteins in individual reactivated cells and mapping this information onto an atlas of unstimulated cells, we have begun to establish a high-resolution view of the types of latent cells that persist in HIV-infected individuals. Applying this method to chart the reservoir in blood and tissues of additional donors will provide a more complete picture of the nature of the persistent reservoir. Our findings thus far reveal that latent cells exhibit unique features, some of which differ between blood and tissues consistent with the notion that unique mechanisms of persistence are present in tissues. 335 TREASURE HUNT: HIV DNA IN DISTINCT SUBSETS IN THE BLOOD, TERMINAL ILEUM, AND RECTUM Max Augustin 1 , Carola Horn 1 , Meryem S. Ercanoglu 1 , Vincent Bondet 2 , Isabelle Suarez 1 , Seung-Hun Chon 1 , Dirk Nierhoff 1 , Elena Knops 1 , Eva Heger 1 , Mark Oette 3 , Florian Klein 1 , Gerd Fätkenheuer 1 , Darragh Duffy 2 , Michaela Müller-Trutwin 2 , Clara Lehmann 1 1 Cologne University Hospital, Cologne, Germany, 2 Institut Pasteur, Paris, France, 3 Academic Hospital Augustinerinnen, Cologne, Germany Background: The gut associated lymphatic tissue (GALT) shows major differences in immune cell composition along the intestine. The largest part of the GALT is located in the terminal ileum (TI), and the HIV reservoir has been associated with CXCR3 and PD1 expression on CD4+ T-cells, and the chemokine IP-10. However, in contrast to the rectum (R), studies of the TI are scarce due to the difficulty in obtaining biopsies. The aim of this study was to compare the interplay of IP-10, CXCR3, PD-1 and the size of the viral reservoir between peripheral blood (PB), R and TI in HIV+ patients on ART.

Poster Abstracts

336 PROPORTION OF INDUCIBLE AND INTACT HIV-1 IN BLOOD IS REFLECTIVE OF LYMPH NODES Alyssa R. Martin 1 , Alexandra M. Bender 2 , Kyungyoon J. Kwon 2 , Diane M. Brown 2 , Niraj Desai 2 , Dorry Segev 2 , Thomas C. Quinn 1 , Aaron Tobian 2 , Christine Durand 2 , Andrew D. Redd 1 , Robert Siliciano 2 1 National Institute of Allergy and Infectious Diseases, Bethesda, MD, USA, 2 Johns Hopkins University School of Medicine, Baltimore, MD, USA Background: The latent reservoir (LR) for HIV-1 persists in CD4 + T cells and is a barrier to cure. The LR has been well characterized in peripheral blood, but lymph nodes have been proposed as a unique sanctuary with concerns that they harbor a greater proportion of latently infected cells, and that measurements of the LR in peripheral blood would not be representative of other tissue sites. Measuring the LR is complicated by an excess of defective virus that is detected by many assays but is not of concern in cure strategies. Additionally, only a subset of intact proviruses are readily activated in vitro. To resolve these issues we characterized total intact, defective, and inducible virus in paired blood and lymph node samples. Methods: Peripheral blood and lymph node samples were collected from 8 HIV-infected, virally suppressed participants immediately prior to solid organ transplantation. Purified CD4 + T cell populations were analyzed using the intact proviral DNA assay (IPDA) to determine the number of intact and defective proviral genomes by ddPCR, and by a novel quantitative viral induction assay (QVIA) to determine the number of inducible proviruses in each sample by quantification of cell-associated RNA at limiting dilution. Results: No difference in inducible virus was detected between lymph node and blood samples (median= 4.3, 7.9 inducible proviruses per million, respectively). The median number of intact provirus per million CD4 + T cells was 104 (IQR 48-591) in lymph nodes and 61 (IQR=43-440) in blood; this difference did not reach statistical significance (P=0.109). There was also no difference

CROI 2019 122