CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

Furthermore, ART-suppressed immunological non-responders (INR) had significantly higher median rectal p24 levels than immunological responders (IR (0.024 vs 0.009 pg p24 per million rectal CD4+ T cells, P=0.009). Rectal tissue from viremic and elite controllers also had measurable p24 pre-ART that declined significantly following ART. Among all ART-suppressed participants, higher rectal p24 levels were associated with lower CD4 counts and among the controllers, rectal p24 levels are better predictors of HIV-specific CD107a+ CD8 response than HIV RNA. p24 was also detected in aviremic FNA LN samples and early data suggests associations with PD1, CXCR5+ PD1 and Tfh CD4+ cells. Conclusion: HIV gag p24 was detected in most rectal and lymph node biopsies from ART-suppressed donors and controllers suggesting a higher burden of p24 in tissues than blood. The capacity to measure and characterize HIV reservoir at the level of viral protein and assess the relationship between immune phenotype revealed p24 correlates with immune function and will be important for immune-based clearance strategies. 317 LACK OF COMPARTMENTALIZATION IN THE LATENT RESERVOIR OF BLOOD AND LYMPH NODES Charles Kirby 1 , Jada Hackman 2 , Alyssa R. Martin 2 , Alexandra M. Bender 1 , Kyungyoon J. Kwon 1 , Craig Martens 3 , Michael J. Bale 4 , Niraj Desai 1 , Sander S. Florman 5 , Dorry Segev 1 , Aaron Tobian 1 , Christine Durand 1 , Robert Siliciano 1 , Andrew D. Redd 2 Background: There are conflicting reports on the similarity of the HIV-1 latent reservoir (LR) in the lymph node (LN) compared to the peripheral blood (PB). Characterizing the composition and any possible differences in anatomical compartments remains a crucial step to understanding the barriers to HIV cure. Methods: HIV+ individuals on ART with suppressed viral loads who were undergoing solid organ transplantation consented to have LN removed at the time of transplant, and PB and LN mononuclear cells (MC) were collected and isolated (n=10). CD4+ cells frommatched PBMC and LNMC samples were plated in a novel quantitative viral induction assay (QVIA). Sequence data was obtained from positive wells using a validated site-directed next-generation sequencing based assay that amplified the gp41 region of the viral envelope to identify the prominent induced viral variants (>2.5% of amplicons; Illumina Inc.). Subjects who had sequences obtained from>75% of the positive wells in both compartments were used to examine for compartmentalization. Neighbor- joining trees of all prominent patient sequences were inferred (Geneious prime), and identical variants were classified as ‘replicates’. A Bayesian model (IUPMBayes) was used to estimate the relative size of the LR for each participant, as well as the proportion of the LR made up by each variant for their respective compartments. Compartmentalization was assessed on samples using a Hudson based test for panmixia (non-parametric) and a branch length tree correlation coefficient (parametric). Results: In four individuals with sufficient sequences, a median of 29 induced variants were identified in PB (IQR=54.8-16.0), as compared to 26 in LN (IQR=27.8-21.3). The estimated frequency of latently infected cells was 10.4 induced proviruses per million cells (IPPMC) in PB (IQR=16.5-7.5) and 8.4 IPPMC in LN (IQR=13.2-6.8). Replicant variants and variants with unique sequences were found in both compartments in all patients. The estimated proportion of the LR made up of variants that were replicated in the patient’s matched compartment varied between compartments and between patients [median % LR shared for PB=37.1% (IQR=40.1-20.5%) and LN=42.4% (IQR=57.9-29.8%); Figure]. There was no significant compartmentalization between PB and LN across all patients. Conclusion: These data provide further evidence of intermingling and limited compartmentalization between the LN and PB, and support previous data that the LR found in the blood can be a good representation of the LR in the lymph node 1 Johns Hopkins University School of Medicine, Baltimore, MD, USA, 2 NIAID, Bethesda, MD, USA, 3 NIAID, Hamilton, MT, USA, 4 National Cancer Institute, Frederick, MD, USA, 5 Mt Sinai School of Medicine, New York, NY, USA

318 CHARACTERIZATION OF CD8+ TRM TOWARD THE CONTROL OF THE HIV RESERVOIR IN CERVIX Núria Massana 1 , Jon Cantero 1 , Judith Grau-Exposito 1 , Laura Luque-Ballesteros 1 , Josep Castellví 2 , Laura Mañalich-Barrachina 2 , Cristina Centeno-Mediavilla 2 , Vicenç Falcó 1 , María J. Buzón 1 , Meritxell Genescà 1 1 Vall d'Hebron Research Institute, Barcelona, Spain, 2 Hospital Universitario de la Vall d’Hebron, Barcelona, Spain Background: In tissues, resident memory CD8+T cells (TRM) are most likely necessary to eliminate remaining cellular HIV-1 reservoirs. However, TRM signature includes expression of molecules associated to exhausted phenotypes during chronic viral infections. Here we addressed the functional capacity of CD8+TRM from the cervical mucosa of HIV-infected women on ART to determine the most effective phenotypes at limiting viral persistence. Methods: CD8+T cells from cervical tissues were phenotyped based on CD69 expression to determine TRM signature (n=6-9). Frequency and activation of CD103+/-CD8+TRM subsets were compared between healthy (n=9) and ART- suppressed HIV+ women (n=18). In a subset of these patients, we determined total vDNA in blood and cervix (n=7). A functional assay was established to determine suppression of viral reactivation by CD8+TRM in ART-suppressed HIV+ women. Results: Cervical CD69+CD8+T cells protein profile was compatible with >90% belonging to bona fide CD8+TRMs, as determined by CCR7, S1PR1, T-bet, Eomes, Hobit, α1 and PD-1 expression. Further, CD8+TRMs expressed more frequently CXCR3, CCR5, CCR2 and CD161 compared to non-CD8+TRMs, and less frequently α4, CD122 and gdTCR (p<0.05). Cervical samples from ART-suppressed patients were enriched in total CD8+T cells compared to uninfected women, including higher frequencies of non-TRMs (p<0.01) and TRMs (p<0.05), and higher expression of HLA-DR (p<0.01). Importantly, the frequency of cervical CD8+TRMs correlated with proviral HIV-1 DNA in cervix (r=-0.82; p=0.03) and tissue CD8+TRMs showed better control of the reservoir in reactivated cells than effector circulating CD8+T cells. Conclusion: Alterations of the CD8+T cell compartment within the cervical mucosa remain in HIV+ women even after several years of effective ART- suppression. The association between higher proportion of CD8+TRM in cervix and less proviral HIV-1 DNA, together with data showing higher control of virally-reactivated infected cells by CD8+TRM, indicates that these cells may be critical to control persisting virus in tissues. 319 FACTORS ASSOCIATED WITH VIRAL CONTROL AFTER STRUCTURED TREATMENT INTERRUPTION Nikolaus Jilg 1 , Behzad Etemad 2 , Ruth Dele-Oni 2 , Colline Wong 2 , Jesse Fajnzylber 2 , Abbas Mohammadi 2 , Evgenia Aga 3 , Ronald Bosch 3 , Daniel R. Kuritzkes 2 , Ian Frank 4 , Jeffrey Jacobson 5 , Jonathan Z. Li 2 1 Massachusetts General Hospital, Boston, MA, USA, 2 Brigham and Women's Hospital, Boston, MA, USA, 3 Harvard T.H. Chan School of Public Health, Boston, MA, USA, 4 Hospital of the University of Pennsylvania, Philadelphia, PA, USA, 5 Case Western Reserve University, Cleveland, OH, USA Background: ACTG A5068 was a RCT of people with chronic HIV infection (18% females) receiving continuous ART versus intermittent structured treatment interruptions (STIs) with or without administration of a therapeutic HIV vaccine. In the CHAMP study, HIV post-treatment controllers (PTCs) were clustered among

Poster Abstracts

CROI 2020 109