CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

309 SIV REBOUND KINETICS FOLLOWING TLR7-AGONIST & THERAPEUTIC VACCINE ADMINISTRATION Jeffrey Gerold 1 , So-Yon Lim 2 , Christa E. Osuna 2 , Dan Barouch 2 , James Whitney 2 , Alison L. Hill 1 1 Harvard University, Cambridge, MA, USA, 2 Beth Israel Deaconess Medical Center, Boston, MA, USA Background: Immunotherapy is a promising approach to prevent or control HIV rebound following cessation of antiretroviral therapy (ART). We previously tested a toll-like receptor 7 (TLR7) agonist, alone and combined with Ad26/MVA therapeutic vaccination, on rebound in SIV-infected rhesus macaques. A subset of animals in each study achieved remission, but the mechanism of action of these treatments and the source of differences in outcomes between individuals remains unknown. Methods: We conducted an in-depth characterization of viral load kinetics following ART interruption in each study using a novel viral dynamics model that includes reactivation of latent infection and adaptive immune responses. We developed a Bayesian inference algorithm to estimate model parameters for individual animals from viral load kinetics and determine which parameters are impacted by which therapy. Results: Our model describes the kinetics of rebound well in all animals in both studies. In the TLR7-only study, in which animals started ART during chronic infection, we found that although mean rebound time was the same in treated vs control groups, the kinetics were different. TLR7 treatment reduced inferred reservoir exit rate (log 10 -9.87 vs -5.65 cells/ml/day, p=0.01), but also increased viral growth rate (1.8 versus 0.77 per day, p=0.03). We used the model to estimate the number of latent cells contributing to viral blips, and found that between 0.1% and 1.2% of SIV DNA was reactivated, perhaps representing a large fraction of the replication-competent reservoir. These findings are supported by the absence of rebound despite CD8 depletion and lack of detectable replication-competent reservoir in 2/11 treated animals. In the combined TLR7-agonist and vaccine study, in which animals started ART during acute infection, we found that TLR7 alone did not significantly change any viral dynamic parameters. The combined TLR7/vaccine reduced both the initial viral growth rate (0.56 vs 0.85 per day, p=0.02) and increased the inferred immune proliferate rate (log 10 -0.17 vss -3.99 per day, p=0.0006). Conclusion: Analysis of viral rebound kinetics using mathematical models suggests that TLR7-agonist alone administered during chronic infection can lead to some reductions in the functional latent reservoir in most animals and complete clearance in other animals, while the post-treatment control obtained with TLR7+therapeutic vaccination is mainly due to immunologic stimulation, not reservoir reduction. 310 IPROTECT1: A VAC-3S VACCINE PHASE II STUDY IN HIV-INFECTED PATIENTS Christine Katlama 1 , Gilles Pialoux 2 , Laurent Cotte 3 , Pierre-Marie Girard 4 , Anne Simon 1 , Jacques Reynes 5 , Jürgen K. Rockstroh 6 , Jose M. Gatell 7 , Alain Devidas 8 , Yazdan Yazdanpanah 9 , Odile Launay 10 , Laurence Weiss 11 , Gerd Fätkenheuer 12 , Béhazine Combadière 13 , Vincent Vieillard 13 1 Pitié-Salpêtrière Hospital, Paris, France, 2 Tenon Hospital, Paris, France, 3 CHU de Lyon, Lyon, France, 4 Saint-Antoine Hospital, Paris, France, 5 CHU de Montpellier, Montpellier, France, 6 Bonn University Hospital, Bonn, Germany, 7 Hospital Clinic of Barcelona, Barcelona, Spain, 8 Centre Hospitalier Sud Francilien, Corbeil-Essonne, France, 9 Bichat–Claude Bernard Hospital, Paris, France, 10 Cochin Hospital, Paris, France, 11 Georges Pompidou European Hospital, Paris, France, 12 University of Cologne, Cologne, Germany, 13 INSERM, Paris, France Background: 3S is a highly conserved motif of HIV-gp41 protein. High level of anti-3S antibodies (Abs) has been associated with lower HIV disease progression. Preclinical studies in SHIV-infected macaques demonstrated that anti-3S Abs are associated with protection of CD4 loss during infection. VAC-3S is a therapeutic vaccine composed of 3S motifs-coupled with CRM197 carrier in aluminum salt adjuvant. Previous Phase 1 study demonstrated the safety of VAC-3S. Methods: IPROTECT1 was a randomized, double-blind, placebo-controlled Phase 2 Study. Eighty-six patients (85 completed) with HIV RNA < 50 cp/mL on ART, with CD4 between 200 and 500 c/mm 3 were randomized to receive VAC-3S 16 µg (group A n=24), 32 µg (group B n=25), 64 µg (group C n=23) or placebo (n=14). All individual received 6 injections (W0, W4, W8, W12, W36, W48) except for arm C (3 injections at W0, W4, W8). The primary endpoint was the proportion of vaccine responders defined as anti-3S Abs > 35 AU one month

respectively. There was a significant increase in the magnitude of PB responses between wk26 (12 SFC/10^6 PBMC) and after the late boosts at wk50 (67 SFC/10^6 PBMC, p<0.001), wk62 (75 SFC/10^6 PBMC, p<0.0001) and wk74 (73 SFC/10^6 PBMC, p<0.0001), but not when no boost was administered (12 SFC/10^6 PBMC, p=NS). A similar trend was observed for memory responses (Table 1). Late boost at wk74 increased the magnitude of memory responses significantly compared to the late boost at wk50 (3433 vs. 2457 SFC/10^6 PBMC, p=0.01). Memory B cell responses at the late boost correlated strongly with IgG geometric mean titers (GMT) at wk96 (r=0.61, p<0.0001). At wk96 the magnitude of PB and memory responses decreased for all late boost groups to frequencies similar to wk26, however still correlated with GMT at wk96 (PB: r=0.32, p=0.01, memory B cells: r=0.66, p=0.01). Conclusion: Late boosts of the ALVAC-HIV/AIDSVAX B/E prime-boost increased the number and frequency of gp120-specific PB and memory B cell responders compared to when no boost was administered, with memory responses being predictive of IgG levels at wk96. New strategies to increase antigen-specific B cell responses could contribute to increased antibody durability.

308 EPITOPES ASSOCIATED WITH HIV-1 DYNAMICS AFTER DC-BASED THERAPEUTIC VACCINATION

Mathieu Surenaud 1 , Laura Richert 2 , Monica Montes 3 , Sandy Zurawski 3 , Karolina Palucka 4 , Jacques Banchereau 4 , Rodolphe Thiébaut 2 , Jean-Daniel Lelievre 1 , Christine Lacabaratz 1 , Yves Levy 1 1 INSERM, Créteil, France, 2 INSERM, Bordeaux, France, 3 Baylor Institute for Immunology Research, Dallas, TX, USA, 4 Vaccine Research Institute, Créteil, France Background: Improvement of therapeutic vaccine strategies in the perspective of HIV cure is warranted. The DALIA-1 phase I trial evaluated a new DC-based therapeutic vaccine that consisted in ex-vivo generated IFN-α DC loaded with LIPO-5 (HIV-1 Nef 66-97, Nef 116-145, Gag 17-35, Gag 253-284 and Pol 325-355 lipopeptides). Nineteen patients on c-ART (antiretroviral treatment) received four injections followed by a 24-week c-ART interruption (ATI). Four weeks after the last injection, an increase in breadth of immune responses was detected by IFN-γ ELISpot and Luminex assays (Y. Levy et al, EJI 2014). An inverse correlation was found between the frequency of polyfunctional CD4+ T cells detected by intracellular staining (ICS) assays after vaccination and the maximum viral load (VL) post-ATI. Here, we performed epitope mapping to dissect anti-HIV-1 responses and correlations between vaccine-induced responses and maximum VL post-ATI. Methods: PBMC from 16 participants were incubated with 36 (15-mer) and 56 (9-mer) overlapping peptides encompassing the LIPO-5 sequences. Cytokine responses were detected using 48h Luminex and 7 day ICS assays. Comparison of responses with predicted peptide/MHC class I and class II binding affinities (NetMHCpan 3.0 and NetMHCpanII 3.1 servers) was performed. Results: IL-13 and IL-2 (15-mer Luminex) T-cell responses detected after four vaccinations and before ATI were inversely correlated with maximum VL post-ATI (Spearman r=-0.66; P=0.007 and r=-0.58; P=0.02, respectively for breadth; r=-0.71; P=0.003 and r=-0.66; P=0.007, respectively for magnitude). Sum of IL-13 and IL-2 responses directed to six 15-mer peptides (from Gag, Pol and Nef regions of LIPO-5) was strongly associated with lower maximum VL post-ATI (Spearman r=-0.71; P=0.003). Using ICS, IL-13 and IL-2 were detected in both CD4+ and CD8+ T cells after 9-mer and 15-mer peptide stimulations. Luminex and ICS functional assays allowed us to identify vaccine-elicited CD4+ T cell responses that were not predicted a priori by peptide binding prediction bio-informatic tools. Conclusion: We showed that ex vivo DC vaccination, as a vehicle of HIV lipopeptide delivery, was broadly immunogenic and elicited responses against a set of epitopes from Gag, Pol and Nef. IL-2 and IL-13 functional responses were associated with a better control of VL post-ATI. Moreover, new HIV-1 CD4+ T-cell immunogenic epitopes, not predicted by peptide/MHC binding algorithms, were identified.

Poster Abstracts

CROI 2018 108