CROI 2015 Program and Abstracts

Abstract Listing

Poster Abstracts

patients, we detected IFN γ + HIV-specific EM CD8 T cells. In these patients, the mean IFN γ + HIV-specific EM CD8 T cell responses were 0.80% (range 0.12-2.39%) and 0.95% (range 0.86-4.32%) at baseline for PANO- and ROMI-treated subjects, respectively. The frequency of IFN γ + HIV-specific EM CD8 T cell responses did not change during PANO (p=0.7) or ROMI (p=0.1) treatment or at follow-up (PANO p=0.7, ROMI p=0.8). Conclusions: In two clinical trials, treatment of HIV-1 infected patients with PANO or ROMI did not decrease the levels of HIV-1-specific EM CD8 T cells. These results provide

support for the combination of HDACi and immune-based therapies in HIV-1 eradication trials. 370 Vaccine Induced Follicular CD8 T Cells Enhance Control of Pathogenic SIV Infection Geetha H. Mylvaganam ; Daniel Rios; GregoryTharp; Steven Bosinger;VijayakumarVelu; Rama R. Amara Emory University, Atlanta, GA, US

Background: Recent data demonstrated that a significant fraction of latently-infected cells reside in the germinal center (GC) resident T follicular helper cells (Tfh) during HAART of HIV-infected humans/SIV-infected macaques. Therapeutic interventions aimed at achieving a functional cure must target these viral reservoirs. Induction of highly functional anti-viral CD8 T cells that can home to GC (follicular CD8) represents an attractive therapeutic strategy. However, GCs are generally thought to exclude CD8 T cells and very little is known about whether T cell based vaccines induce anti-viral CD8 T cells with the potential to migrate to the GC and influence HIV/SIV replication in Tfh. Here, we addressed this in the context of DNA/MVA vaccination and pathogenic SIV infection in rhesus macaques (RM). Methods: RM were vaccinated with our DNA/MVA SIV vaccine with and without adjuvants and challenged intrarectally with SIVmac251. Follicular CD8 T cells were identified based on the expression of chemokine receptor CXCR5 (B cell follicle homing potential) on Gag CM9 tetramer+ CD8 T cells (Tet+). Results: Following vaccination, a small fraction (<10%) of Tet+ CD8 T cells in blood expressed CXCR5. However, following SIV infection, the frequency of Tet+ CXCR5+ CD8 T cells in blood and LN increased dramatically as early as 2 weeks post infection, were present at high levels even at 24 weeks (about 40% of Tet+ in LN) and correlated directly with their frequency post vaccination. These Tet+ CXCR5+ cells were not present in unvaccinated SIV infected animals although Tet+ cells were present. Importantly, among the vaccinated animals the frequency of Tet+ CXCR5+ CD8 T cells in blood at 2 weeks post infection was strongly associated with enhanced viral control both at peak and set point. Immunofluorescence staining of both rectal and LN tissues from vaccine controllers revealed co-localization of CD8 T cells with Tfh in GC. Impressively, the CXCR5+ CD8 T cells from the controller RM restricted the anti-CD3 driven expansion of CM9 peptide pulsed Tfh cells in vitro suggesting their killing potential . Transcriptome analysis of purified CXCR5+ Tet+ CD8 T cells identified these cells as a unique multi-faceted population expressing genes associated with both CTLs and Tfh subsets. Conclusions: Our results demonstrate that it is possible to elicit follicular CD8 T cells by vaccination and these cells contribute significantly to the control of SIV infection. These findings have implications for cure strategies against HIV. Background: Blockade of the PD-1/PD-L1 pathways using monoclonal antibodies (mAb) to PD-1 has been reported to activate HIV expression from latently infected CD4 T cells ex vivo. To further evaluate this approach, we tested the ability of the human anti-PD-L1 mAb BMS-936559 to activate virion production frommononuclear cells obtained from patients on suppressive ART. Methods: PBMC, total (t) CD4, and resting (r) CD4 cells were purified by negative selection of large-volume blood draws from HIV-infected donors. Freshly isolated PBMCs were cryopreserved and immunophenotyped for PD-1/PD-L1 expression. The remaining cells were incubated (1 million cells/well in triplicate) for 1 week with 20, 5, or 1.25 m g/ml anti- PD-L1 mAb, with 20 m g/ml isotype control [Zymogen DT-1D12g-4P (hIgG4)], or with anti-CD3/CD28-coated microbeads plus either anti-PD-L1 mAb or isotype control. On Day 8, cells were assessed for viability and supernatants tested for HIV RNA using the Roche Taqman v.2.0. A virologic response was defined as a >3-fold increase in HIV RNA over isotype control. Donors whose cells responded initially to anti-PD-L1 mAb were redrawn and tested again. Results: PBMC, tCD4, and rCD4 cells purified from ten long-term (mean 8 years) suppressed donors. Cell viability was not reduced by treatment with anti-PD-L1 mAb. 9 of 10 donors responded to anti-CD3/CD28 in all cell types (mean fold-increases of 742, 1353, and 272 for PBMC, tCD4 and rCD4, respectively). Anti-PD-L1 mAb did not enhance responses to anti-CD3/CD28. PBMC from 2 of 10 donors (donor 3: 61-fold; donor 4: 7-fold) showed an initial response to anti-PD-L1 mAB that was not reproduced upon repeat blood draw. tCD4 cells from 2 of 10 donors (donor 2: 583-fold; donor 6: 84-fold) initially responded to anti-PD-L1. However, cells from donor 2 did not respond after a repeat draw. Cells from donor 6 responded on the first repeat draw but not the second. rCD4 from 0 of 10 donors responded to anti-PD-L1. PD-1/PD-L1 expression on CD4 and CD8T cells was evident on the day of cell isolation in all donors, but expression levels did not differ between responders, non-responders, and a healthy control. Conclusions: Despite detectable PD-1/PD-L1 expression, increased HIV production from PBMC, total CD4 T cells or resting CD4 T cells after treatment with anti-PD-L1 antibody was infrequent and not reproduced longitudinally. Alternate strategies will be needed to activate proviral expression from latently infected CD4 T cells. 372 Impact of HIV Latency Reversing Agents on Natural Killer Cells Carolina Garrido ; Julia Sung; Swati Gupta; Katherine Sholtis; Nancie Archin; David Margolis Medicine, University of North Carolina University of North Carolina, Chapel Hill, NC, US Background: Eradication of HIV infection requires unmasking of the HIV latent reservoir using latency reactivating agents (LRAs) and the implementation of an immune response to clear infected cells. Thus, an assessment of the impact of LRAs immune function is required. We evaluated the effect of LRAs on natural killer cells (NKs). Methods: Five LRAs were evaluated: 3 histone deacetylase inhibitors, SAHA, Romidepsin (RMD) and Panobinostat (PNB); and 2 protein kinase C activators, Prostratin (PRT) and Ingenol (ING). Each drug was tested at 3 concentrations (cc): physiological, lower and higher. NKs from 3 HIV-negative donors were isolated and exposed 24h to the LRAs and washed out. NKs viability was assessed by AnnexinV+7AAD staining. Cytotoxicity was evaluated through degranulation upon culture with K562 cells, measuring %CD56 + CD107a + cells. Antiviral activity was measured in a viral inhibition assay using autologous CD4 + T cells infected with JR-CSF and quantifying p24 at day 7 of culture. Results: Table 1 shows results for LRAs treatment at physiological concentration. RMD, PNB and PRT showed significant toxicity on NKs. Of note, SAHA and ING did not increase cell death at any cc, while PNB produced significant cell death even at lower cc (5nM) (261.9% (63.8)). RMD, PNB and ING at physiological cc reduced significantly degranulation, and we observed an increase of both viability and degranulation impairment at higher cc (100, 100 and 1000 nM, respectively) of these drugs. Viral inhibition assays showed that any cc of SAHA impair antiviral activity, while RMD impaired it significantly at physiological and lower cc (68.8%(10.6)). Interestingly, ING had a bimodal effect, impairing viral inhibition (28.6% (9.5)) at low cc but improving it at higher cc (183.2% (91.1)). All other conditions did not produce statistically significant differences compared to untreated NKs, but we observed that PRT did not impair viral inhibition while PNB at physiological cc reduced NK antiviral effect to 37.46% (40.72). 371 Blockade of PD-L1 Does Not Reverse HIV Latency in CD4+ T Cells Ex Vivo Elizabeth Fyne 1 ; Shalyn Campellone 2 ; Huilin Qi 2 ; Amy Sheaffer 2 ; Stephen Mason 2 ; JohnW. Mellors 1 1 University of Pittsburgh, Pittsburgh, PA, US; 2 Bristol-Myers Squibb Co., Wallingford, CT, US

Poster Abstracts

277

CROI 2015