CROI 2015 Program and Abstracts

Abstract Listing

Poster Abstracts

422 Evaluation of HIV-1 Latency-Reversing Agents by a Modified Virus Growth Assay (VOA) Riddhima Banga; Francesco Procopio; Matthias Cavassini; Alessandra Noto; Song Ding; Giuseppe Pantaleo; Matthieu Perreau Lausanne University Hospital, Epalinges, Switzerland

Background: Inhibitors of histone deacetylases (HDACis) reactivate HIV replication in in vitro models of HIV latency. However, the recent study by Bullen et al. has underscored the difficulty to reactivate HIV-1 latency in primary resting memory CD4 T cells isolated from aviremic long-term treated HIV-1 infected patients (Bullen et al ., Nature Medicine 2014). We hypothesized that mimicking the modalities of administration of HDACis in patients in vivo, and in particular, the continuous exposure of primary resting memory CD4 T cells to HDACis in culture would result in efficient reactivation of HIV-1 replication. Methods: Resting memory CD4 T cells were isolated from 10 aviremic (<20 HIV-1 RNA copies/mL) long-term treated (>2 years) HIV-1 infected subjects and cultured either in the presence of CD3/CD28 (positive control) or continuous exposure to clinically relevant concentrations of HDACis i.e. SAHA, romidepsin, belinostat, panobinostat and givinostat throughout the culture period of 14 days. HIV-1 RNA and p24 protein were measured at the end of the culture period using validated diagnostic tests. Results: Using the modified VOA, all the HDACis tested (SAHA, romidepsin, panobinostat, belinostat and givinostat) induced reactivation of HIV-1 replication in primary resting memory CD4 T cells isolated from aviremic long-term treated HIV-1 infected patients as indicated by the high levels of HIV RNA and p24 production measured in culture supernatants. Of note, HIV-1 reactivated in the cell cultures was also infectious as demonstrated by the transmission of HIV-1 infection in vitro to CD8-depleted blood mononuclear cells from healthy subjects. Interestingly, givinostat was more potent than the other HDACis in reversing HIV-1 latency in vitro on the basis of the percentage of patients reactivating HIV and the proportion of HIV positive wells. Finally, the combined treatment of anti-CD3/anti-CD28 MAbs plus givinostat did not result in a significant increase of HIV-1 replication, suggesting that both HDACis and TCR signals may target similar populations of latently HIV-1 infected CD4 T cells. Conclusions: These results support further investigation and development of HDACis in the armamentarium of the therapeutic agents to achieve HIV functional cure. 423 “Kick and Kill”of Latent HIV-1 Infection in Naïve and Central Memory CD4 + T cells Background: A “kick and kill” strategy, which involves the administration of a latency reversing agent (LRA) to induce HIV-1 out of latency and promote death of the HIV-1- infected cells by cytolytic or viral cytopathic effects, may eradicate the latent HIV-1 reservoir in resting CD4 + T cells. However, the resting CD4 + T cell population is heterogeneous and is composed of distinct subsets that differ in lifespan, proliferative capacity, antigen response time, and in CCR5 and CXCR4 expression levels. It is not known whether the “kick and kill” strategy will be effective in each of the T cell subsets. Therefore, we investigated latency reversal in purified CD4 + naïve (T N ) and central memory (T CM ) T cells, which together constitute ~ 70-80% of the resting CD4 + T cell population. Methods: We developed primary cell models of both R5- and X4-tropic HIV-1 latency in T N and T CM CD4 + T cells, using a direct infection protocol as described previously (Saleh et al . Blood 2007;110:4161). Latency reversal (assessed by extracellular HIV-1 RNA) and concomitant virus-induced cell death was evaluated after exposure to α CD3/CD28 antibodies, PMA/PHA, prostratin or SAHA. Results: Consistent with previously published data, both R5- and X4-tropic viruses infected T N cells less efficiently than T CM cells. Consequently, the frequency of latently infected T N cells was significantly lower than in the T CM cells. Stimulation with α CD3/CD28 antibodies, PMA/PHA or prostratin, effectively reactivated latent HIV-1 in both T cell subsets. When the data were normalized for infection frequency more extracellular virus production was observed in X4-tropic (but not R5) infected T N cells compared to T CM cells. Interestingly, we found a strong correlation between the number of latently infected cells and extracellular virus production in T CM cells, but not in T N cells. HIV-1 DNA in both T N and T CM cells decayed with half-lives ranging from 2.5 to 5.4 days after exposure to α CD3/CD28 antibodies, PMA/PHA or prostratin. In comparison, SAHA was unable to reactivate latent HIV-1 in either T N or T CM cells, and consequently the latently infected cells persisted after exposure to drug. Conclusions: We have developed robust primary cell models of HIV-1 latency in CD4 + T N and T CM cells that can be used to better understand the establishment and reversal of latency in these key reservoirs of HIV-1 infection. Importantly, we demonstrate decay of the latent HIV-1 reservoir in both T cell subsets when the LRA effectively stimulates virus production. 424 HIV Recombination in the In Vitro T CM Latency Model – Reasons and Solutions Pawel Bonczkowski 2 ; Laura J. Martins 1 ;Ward de Spiegelaere 2 ; Alberto Bosque 1 ; Eva Malatinkova 2 ; Christopher H.Woelk 3 ; Cory H.White 4 ; AdamM. Spivak 5 ;Vicente Planelles 1 ; LinosVandekerckhove 2 1 University Of Utah School of Medicine, Salt Lake City, UT, US; 2 Ghent University and University Hospital, Ghent, Belgium; 3 University of Southampton, Southampton, United Kingdom; 4 University of California San Diego (UCSD), La Jolla, CA, US; 5 Utah University Hospital, Salt Lake City, UT, US Background: HIV latency research will benefit from in vitro latency models that closely mimic in vivo latency. The T CM model described by Bosque and Planelles generates high numbers of resting latently infected cells. The initially described model is based on a wild type env deficient virus co-transfected with an HIV env to make a single round virus. Many experiments in our laboratories led to high levels of background infection and massive cell death. These were reminiscent of a spreading infection. Methods: To assess the replication potential of the virus used in the T CM model, the original complemented virus as well as the culture supernatants of the T CM model were used to transduce Jurkat T-cells and fresh T CM . Subsequently, PCR and deep sequencing were performed to identify recombination as a likely source of replication competent virus. Finally, an adapted model using replication-competent HIV strains carrying EGFP/HSA reporter genes and including cell density manipulations in the presence of ART were assessed to provide a more standardized T CM model. Results: A spreading infection in the Jurkat and primary cells suggested the presence of a replication competent virus as a cause of background infection and cell death. Experiments with VSV-G and backbone only constructs excluded contamination as the source of this virus. PCR and deep sequencing analysis confirmed the presence of an intact env in the infected cells, indicating that the env deleted constructs recombine with the env sequence used for co-transfection. An alternative approach using replication competent NL4.3 virus with an IRES-EGFP construct in combination with antiretroviral drugs to block spreading infection proved useful for medium throughput screening of anti-latency compounds. This approach leads to up to 35% of latent infection in the presence of 15% of productive infection. The wild type NL4.3 combined with CD4 expression based cell sorting and cell crowding after infection to promote cell-to-cell HIV transmission can be used for high-throughput assays. Depletion of CD4(-) cells leaves a population of up to 20% latently infected cells with little to no active infection. Conclusions: Recombination between sequences used to generate single-round vectors in the original model introduces bias to the data and impacts data interpretation. Our study shows that the issue of recombination can be omitted. As an alternative strategy, we propose to use replication competent virus in combination with ARV to generate latently infected T CM . Jennifer M. Zerbato ; Nicolas Sluis-Cremer University of Pittsburgh, Pittsburgh, PA, US

Poster Abstracts

299

CROI 2015