CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

to suppress plasma viremia in infected individuals. However, the ability of VRC01 to suppress plasma viral rebound in HIV-infected patients following cessation of antiretroviral therapy (ART) remains unclear. Methods: An exploratory, open-label clinical trial was conducted to examine the effect of passive transfer of VRC01 on plasma viral rebound following discontinuation of ART in HIV-infected individuals who initiated ART during the chronic phase of infection and who suppressed plasma viremia >3 years with CD4 + T cell count > 450 cells/mm 3 at enrollment. Subjects received VRC01 (40mg/kg) 3 days prior to and 14 and 28 days following interruption of ART, and monthly thereafter for up to 6 months. Levels of plasma viremia and VRC01 were measured at day -7, -3, 0, 3, 7, 14, 21, and 28 and biweekly thereafter. In addition, the capacity of VRC01 and other bNAbs to neutralize autologous infectious HIV prior to and following infusions of the antibody was examined. Results: 10 subjects were enrolled in the study. Mean duration of ART was 10.6 years. Mean CD4 + and CD8 + T cell counts at baseline were 796 and 768 per mm 3 , respectively. Multiple infusions of VRC01 were safe and well tolerated. 9/10 subjects experienced plasma viral rebound (>40 copies/ml) between 11-54 days (median 39) following cessation of ART; 8 subjects reinitiated ART per protocol. Plasma concentration of VRC01 ranged between 142-352 µg/ml (median 160) at time of first detectable plasma viremia. Preliminary analyses of autologous replication-competent viral isolates revealed the existence of VRC01-resistent virus prior to infusion of antibody in several subjects; further assessment of prior and post-treatment resistance is ongoing. Conclusions: While multiple infusions of VRC01 were safe and well-tolerated, the majority of patients experienced plasma viral rebound despite adequate levels of antibody in plasma. Therefore, therapeutic strategies involving passive transfer of bNAbs may require a combination (s) of Abs and/or resistance prescreening in order to achieve sustained virologic control in HIV-infected individuals upon withdrawal of ART. 312 CD8 T Cells FromHIV Controllers Recognize and Kill HIV+ Non-Activated CD4 T Cells Blandine Monel 1 ; Julie Boucau 2 ;Yovana Pacheco 2 ; Annmarie McKeon 2 ; Daniel Kavanagh 2 ; Sylvie Le Gall 2 ; Bruce D.Walker 2 1 Howard Hughes Med Inst/Ragon Inst of MGH, MIT, and Harvard, Cambridge, MA, USA; 2 Ragon Inst of MGH, MIT, and Harvard, Cambridge, MA, USA Background: Understanding the underlying immunological mechanisms of spontaneous HIV control by HIV+ controllers remains an important step in the development of efficient HIV vaccines and functional cures. Methods: To understand control mechanisms triggered by CD8+ T cells we established a system in non-activated peripheral blood mononuclear cells (PBMC) using infectious HIV (NL4.3 IRES GFP) carrying the fusion protein Vpr-Beta lactamase (to assess viral entry) and a staining strategy to study the CD8+ T cell response by flow cytometry. The ability of the CD8 T cells from HIV controllers to form immunological synapses with HIV+ non-activated CD4 T cells was studied by confocal microscopy and the killing of the target cells was assessed by chromium release assay. In the study we included 10 HIV controllers and 10 HIV chronic progressors under antiretroviral therapy for less than 4 years. The statistical differences were calculated using PRISM software and an unpaired T test. Results: We showed that non-activated CD4+ T cells are permissive for HIV entry but this leads mostly to non-productive infection (b-lac+ GFP-). CD8+ T cells from HIV controllers were significantly more efficient than CD8 T cells from progressors to recognize these HIV+ cells, became activated (IFNg+ p=0.0068, MIP1b+ p=0.0162) and underwent degranulation (CD107a+ p=0.0038) . This activation was initiated by the presentation of HIV peptides from incoming viruses since viral entry but not reverse transcription was necessary and HLA-I blockade abolished this response. This CD8 T cell response was allowed by the formation of functional immunological synapses with HIV+ non-activated CD4+ targets and blocking adhesion molecules like LFA-1 disturbed the CD8 T cell degranulation. We confirmed that the CD8 T cell response lead to the killing of the target cells by chromium release assay. Finally we showed by flow cytometry that the TCR-beta chain may play an important role on this recognition. Conclusions: In conclusion, recognition of HIV+ non-activated CD4+ T cells directly after HIV entry by CD8+ T cells before the establishment of latency or abortive infection participates to HIV control and should be further investigated for new vaccine strategies. 313 HIV-1-Specific CD8 T Cells Are Poorly Cross-Reactive During Acute Infection Yimin Du 1 ; Anju Bansal 1 ; Jonathan Carlson 2 ; Jesus Salazar-Gonzalez 1 ; Kristin Ladell 3 ; Stephanie Gras 4 ; Sonya L. Heath 1 ; David Price 5 ; Jamie Rossjohn 4 ; Paul A. Goepfert 1 1 Univ of Alabama at Birmingham, Birmingham, AL, USA; 2 Microsoft Rsr, Redmond, WA, USA; 3 Cardiff Univ, Cardiff, UK; 4 Monash Univ, Clayton, Australia; 5 NIH, Bethesda, MD, USA Background: HIV-1 has a high capacity to mutate so elucidating HIV-1 epitope-specific CD8 T cells that can efficiently target epitope variants is important for effective vaccine development. Prior studies indicating the capacity for CD8 T cells to cross-recognize epitope variants mainly used chronically infected samples, whereby presence of multiple viral quasispecies makes it difficult to ascertain the HIV-1 strain inducing the CD8 T-cell response in-vivo . To overcome this limitation, we evaluated the extent of T-cell cross-reactivity by analyzing samples from acutely HIV-1 infected patients with known identity of their transmitted founder virus (TFV) sequence. Methods: Cross-reactive CD8 T-cell responses were assessed using PBMC obtained from 11 acute clade B infected patients. Autologous epitopes were designed based on each patient’s TFV and HLA-I sequence. Commonly occurring variants of the autologous epitopes (>10% circulating clade B virus) were selected from the LANL database. CD8 T cells targeting autologous epitopes (N=18) were evaluated for cross-reactivity by stimulating PBMC with their cognate epitope variants in an IFN-γ ELISPOT assay. Cytotoxicity of cross-reactive T cells was determined by co-culturing epitope pulsed target cells with effector CD8 T-cell lines in a 7-AAD flow-based killing assay. In addition, effector/cytokine polyfunctionality, HLA-I binding affinity (predicted from NetMHC), T-cell avidity, epitope/HLA structural analysis and TCR clonotyping assays were performed. Results: Only 5 of the 18 immunogenic autologous epitopes elicited T cells that cross-recognized one or more variants. Cross-reactive CD8 T cells exhibited poor target cell killing. This impaired killing was not driven by polyfunctionality; neither was HLA-I binding affinity, which was lower only for non-immunogenic epitope variants. Structural and TCR clonotype analyses of an autologous/variant epitope pair revealed that amino-acid changes in the TCR contact site and recruitment of dominant T-cell repertoires, respectively, affect epitope variant detection as well as the cytotoxicity of a cross-reactive response. Conclusions: Cross-reactive CD8 T cells appear compromised during primary HIV-1 infection, associated with reduced HLA-I binding affinity and TCR recognition of epitope variants. These results suggest that for an efficacious CTL-based HIV vaccine, a mosaic or polyvalent approach aimed at inducing broad autologous responses would be a more promising vaccine strategy. 314 Deleterious Effect of KIR-HLA-Associated Gag Sites on Clinical Outcome in CRF01_AE Masahiko Mori 1 ; NuanjunWichukchinda 2 ; Reiko Miyahara 3 ; Archawin Rojanawiwat 2 ; Panita Pathipvanich 4 ;Toshiyuki Miura 3 ; MichioYasunami 3 ; Koya Ariyoshi 3 ; Pathom Sawanpanyalert 2 1 Univ of Oxford, Oxford, UK; 2 Ministry of PH, Nonthaburi, Thailand; 3 Inst of Trop Med, Nagasaki Univ, Nagasaki, Japan; 4 Lampang Hosp, Lampang, Thailand Background: Class I HLA molecule contributes to the immune control of HIV through antigen presentation to both CTLs and NK cells. Contribution of CTLs to HIV control through antigen presentation of Gag peptides by protective HLA alleles, such as HLA-B*57, B*58:01 and B*27, has been well studied previously. However, reports on the NK cell contribution in HIV control, with the exception of the protective effect of HLA-Bw4 and its receptor, killer immunoglobulin-like receptor (KIR) 3DL1 and 3DS1, are limited. In particular, studies about viral adaptation to NK cell-induced immune pressure remain sparse. Methods: 208 HIV-infected treatment-naïve patients were recruited from Thailand, and the effects of HLA-KIR-associated sites in Gag on clinical outcomes were analyzed. To avoid the overestimation of HLA-associated site detection derived from CTLs, HLA-KIR-associated site detection between subjects with or without KIR in the context of their ligand HLA alleles was introduced (e.g. mutation frequency difference at Gag residue 93E between HLA-C1+KIR2DL2+ and HLA-C1+KIR2DL2-). Results: 26 HLA-KIR-associated sites were identified; 4 of them showed significant viral load differences between KIR+ vs KIR- subjects (5.5 vs 5.0 log copies/ml (p=0.03) at E93X vs 93E restricted by HLA-C1+KIR2DL2+, 4.8 vs 5.4 log copies/ml (p=0.02) at T122X vs 122T by HLA-C1+KIR2DL2+, 4.7 vs 5.3 log copies/ml (p=0.02) at M31X vs 31M by HLA-

Poster Abstracts

120

CROI 2016