CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

in LN with the percentage of Tfh cells proliferating (Ki-67+ ; p=0.0002) or harboring SIV-DNA (p=0.0256). We observed a selective enrichment during ART of SIV-DNA+ cells being IL-10+ in LN within the BCF (>85%; p=0.0095) as compared to untreated RMs, with the SIV-DNA+ IL-10+ cells being remarkably more stable between the pre- and on-ART time points than their IL-10 negative counterpart. Furthermore, plasma IL-10 at pre-ART predicts several key parameters of residual disease after 7 months of suppressive ART: CD4 T cell counts in PB (p=0.0305), %Ki-67+ CD4 T cells in PB (p=0.0163), %Tfh cells in LN (p=0.0002), SIV-DNA content in PB CD4 T cells (p=0.0218) and in RB (p=0.0383). Conclusion: Plasma and LN content of IL-10, which is induced by SIV infection and not fully normalized with ART, critically contributes to SIV persistence by promoting maintenance of SIV-DNA+ cells, including TFH. Modulation of IL-10 represents a novel therapeutic avenue towards an HIV cure. 377 NOVEL SHIVS ENCODING TRANSMITTED/FOUNDER ENVS FOR LATENCY AND CURE RESEARCH Anya Bauer 1 , Fang-Hua Lee 1 , Hui Li 1 , Ronald Veazey 2 , Meagan Watkins 2 , George Shaw 1 , Katharine J. Bar 1 1 University of Pennsylvania, Philadelphia, PA, USA, 2 Tulane National Primate Research Center, Covington, LA, USA Background: A novel, robust simian-human immunodeficiency virus (SHIV)-macaque model of HIV-1 latency is critical to investigate eradicative and suppressive strategies that engage Env. We have developed a novel strategy to generate designer SHIVs encoding native TF (transmitted/founder) or primary Envs with tier 2 neutralization that consistently confer productive infection, high peak viremia, and desirable early viral kinetics. Here, we evaluate two promising TF SHIVs, SHIV.D.191859 and SHIV.C.CH848, which encode TF subtype D and C HIV-1 Envs, respectively, for their viral kinetics and persistence during suppressive combination antiretroviral therapy (cART) and treatment interruption in rhesus macaques (RM). Methods: 12 Indian RM were intravenously or mucosally inoculated with SHIV.D.191859 and followed longitudinally. A second cohort of 8 RMwere intravenously inoculated with SHIV.C.CH848. Viral kinetics through the establishment of peak and setpoint viremia, 24 weeks of cART, and treatment interruption were assessed via plasma RT-PCR. Single genome sequencing of plasma virus was used to characterize the diversity of rebounding viruses. Results: Inoculation of 12 RM with SHIV.D.191859 led to productive infection with peak viral loads between 10 5 -10 8 copies/ml. In 11 of 12 animals, viremia was maintained for at least 6 months. At between 6 and 18 months of infection, 4 RMwith high setpoint viremia (viral load of 10 3 -10 7 copies/mL) were placed on cART for 24 weeks. Viral suppression was rapidly achieved and durably maintained. Viral rebound between day 7 and 17 was observed in all four rhesus macaques upon cessation of ART. Sequencing of rebound plasma vRNA revealed multiple genetically distinct virus populations at or near first detectable rebound in all four animals. Inoculation of 8 RMwith SHIV.C.CH848 produced desirable viral kinetics with peak viremia of 10 7 -10 8 and set point viremia between 10 3 -10 5 copies/ml. After 16 weeks of infection, 4 RM were placed on cART and viral suppression was rapidly achieved and maintained for 24 weeks. Viral rebound occurred at day 12-29 after treatment interruption. In both SHIV.D and SHIV.C infected RM, time to rebound correlated with setpoint viremia. Conclusion: The antigenic properties and viral kinetics before, during, and upon interruption of cART make SHIV.D.191859 and SHIV.C.CH848 promising reagents for a SHIV model of HIV-1 latency and cure. 378 CLONES WITH INTACT PROVIRUSES ARE FOUND IN MULTIPLE CD4+ T CELL MATURATION SUBSETS AndrewMusick 1 , Eli A. Boritz 2 , Jonathan Spindler 1 , Michael J. Bale 1 , Michele Sobolewski 3 , Wei Shao 4 , Stephen H. Hughes 1 , John M. Coffin 5 , John W. Mellors 3 , Frank Maldarelli 1 , Mary F. Kearney 1 1 NIH, Frederick, MD, USA, 2 NIH, Bethesda, MD, USA, 3 University of Pittsburgh, Pittsburgh, PA, USA, 4 Leidos Biomedical Research, Inc, Frederick, MD, USA, 5 Tufts University, Boston, MA, USA Background: An important mechanism for maintaining the HIV reservoir is the proliferation of cells that were infected with replication-competent (intact) proviruses prior to ART initiation. It is not known in which CD4+ T cell subsets clonally-expanded, intact proviruses persist. To address this question, we assayed naïve, central and transitional memory (CTM), and effector memory CD4+ T cell subsets (EM) for sequence matches to expanded clones with intact

proviruses, and compared their distribution to those with defective proviruses, and to those with proviruses without obvious defects but that did not produce replicating virus in VOA (non-induced). Methods: Using p6-PR-RT single genome sequencing (SGS), we identified 84 different WT proviruses that were likely in clonally expanded cells (i.e. multiple identical sequences detected) in PBMC from “Patient 1” in Simonetti, et al. (PNAS 2016). At this same timepoint, the plasma contained a mixture of diverse, replicating drug-resistant (DR) variants and a population of identical WT virus produced by a highly-expanded clone. PBMC from the same sample were sorted into naïve, CTM, and EM subsets and SGS was performed to identify sequence matches to proviruses in probable clonally-expanded cells. Results: Three of the 84 probable clones matched replicating viral sequences in VOA, 10 were defective (contained stop codons), and 71 were non-induced. About 45,500 naïve, 60,000 CTM and 57,000 EM cells were analyzed and 22, 48, and 37 SGS were obtained, respectively. A test for panmixia suggested that each cell subset had a different proviral population (probability of panmixia <10-6). DR proviruses were found primarily in naive cells (14/20 in naïve, 5/20 in CTM, 1/20 in EM (p=0.01)). Sequences matching the intact AMBI-1 provirus, the source of the WT clonal plasma virus in the patient, were found primarily in EM (14/37 proviruses) and, to a lesser degree, in CTM (3/48) (p=6x10-4). A second sequence matching infectious provirus was found only in a single naïve T cell, and a third only in two EM cells. Sequences matching defective and non- inducible proviruses were found in each of the three T cell subsets. Conclusion: We identified intact proviruses that appear to have been present in clonally-expanded cells in each of the CD4+ T cell maturation subsets. This study suggests that the HIV reservoir resides in multiple T-cell subsets, all of which will need to be targeted to eradicate HIV infection. 379 MULTI-COMPARTMENT DISSEMINATION OF GENOME-INTACT HIV-1 RESERVOIR CLONES Guinevere Q. Lee 1 , Riddhima Banga 2 , Matthias Cavassini 3 , Jean-Marc Corpataux 3 , Xu G. Yu 1 , Giuseppe Pantaleo 3 , Matthieu Perreau 3 , Mathias Lichterfeld 4 1 Ragon Institute of MGH, MIT and Harvard, Cambridge, MA, USA, 2 University of Lausanne, Lausanne, Switzerland, 3 Lausanne University Hospital, Lausanne, Switzerland, 4 Brigham and Women’s Hospital, Boston, MA, USA Background: Latently infected CD4 T cells provide a highly durable reservoir for HIV-1, but the ability to support viral persistence appears to differ considerably among different CD4 T cell populations. Here, we used single- template near full-length viral deep sequencing to map proviral sequences across nine different CD4 T cell subsets from blood and lymph nodes (LN). Methods: PBMC from three HIV-infected patients treated with suppressive ART were sorted into the following CD3+ CD4+ T cell subsets: CD45RA- cells (total memory), CXCR3+CXCR5- (Th1-enriched), CCR4+CCR6- (Th2-enriched), CCR4+CCR6+ (Th17-enriched), CXCR3-CXCR5+ (circulating Tfh-like cells) and CXCR3+CXCR5+ (enriched for Th1/Tfh-like cells). Simultaneously, the following CD3+ CD4+ subsets were isolated from autologous LN: CD45RA- (total memory), PD-1- CXCR5-, PD-1- CXCR5+, PD-1+ CXCR5+/- (enriched for Tfh), and CD45RA+ naïve T cells. Total DNA was extracted, diluted for single-template full-HIV-1 genome PCR, and Illumina deep-sequenced. HIV sequences were classified into “genome-intact” or “defective” using established criteria. Results: A total of 312 HIV proviral sequences were detected from 3 patients. Frequencies of genome-intact sequences relative to defective copies did not notably differ between blood (17/135, 12.6%) and lymph nodes (20/177, 11.3%) (p=0.73, Fisher’s). Except for lymph node naïve cells, genome-intact HIV DNA was detected in all analyzed CD4 T cell subsets from both blood and lymph nodes. In LN, proportions of intact proviruses were highest in CXCR5+ or PD-1+ CD4 T cells, while in blood, distribution of intact proviral sequences was more variable. Proviral sequences from blood and LN phylogenetically intermingled without obvious compartmentalization (Slatkin-Maddison test p=1, 0.2, 0.1). Notably, we observed a cluster of clonally-expanded intact provirus derived from cells with discrete functional polarization from blood and LN (Figure1). Clonally-expanded defective HIV-1 sequences with mixed contributions from blood and lymph nodes were also observed. Conclusion: This study suggests dynamic interchanges between “genome- intact” viral reservoir cell populations from blood and lymphoid tissues. Clusters of clonally-expanded intact proviruses encompassing sequences from cells with distinct anatomic localization and functional polarization suggest infection of common precursor cells as a driving force for viral reservoir stabilization.

Poster Abstracts

CROI 2018 132