CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

377 METHYLATION PROFILES OF HIV-1 PROVIRAL DNA IN ART-SUPPRESSED INDIVIDUALS Valerie F. Boltz 1 , Cristina Ceriani 1 , Wei Shao 2 , Michael J. Bale 1 , Leslie Lipkey 2 , Laura Newman 2 , Brandon F. Keele 2 , Jonathan Spindler 1 , Rebecca Hoh 3 , Jeffery Milush 3 , Steven G. Deeks 3 , Frank Maldarelli 1 , Mary F. Kearney 1 , John M. Coffin 4 1 National Cancer Institute, Frederick, MD, USA, 2 Leidos Biomedical Research, Inc, Frederick, MD, USA, 3 University of California San Francisco, San Francisco, CA, USA, 4 Tufts University, Boston, MA, USA Background: The latent HIV-1 reservoir is populated with clones of cells infected with stably integrated, intact, but transcriptionally-silent proviruses. Previously, we described the integration site of one such clone harboring a replication-competent provirus called AMBI-1. We hypothesized that the silencing of HIV-1 gene expression from this and other clones was due to DNA methylation of the 5’LTR promoter. To address this question, we investigated methylation at the single-proviral level in known CpG islands in the HIV-1 proviral genome, including one in the 5’ LTR promotor region. Methods: Using a bisulfite-based, methylation-specific single-genome- sequencing (SGS) assay, we measured the levels of methylation in CpG islands from ART-suppressed, chronically-infected individuals with samples from PBMC and lymph node mononuclear cells (LNMC) (including the donor with the AMBI-1 clone). Results: From 4 individuals an average of 30 (range: 13 to 91) bisulfite-treated SGS were obtained from each PBMC and LNMC sample. We found no significant difference in any provirus between the level of methylation of the CpG island in the 5’ LTR promoter and the assay background of cytosines not in CpG sites (averaging 3.8% and 3.3% respectively p=0.9). Furthermore, the presumed AMBI-1 provirus (matching LTR sequence) was not found to be methylated above assay background. Interestingly, we did find a significantly higher level of methylation in the CpG island in the env-tat-rev overlapping reading frame in multiple proviruses in each of the samples [averaging 21% of all CpG sites methylated vs. an average of 6% assay background (p=0.03)]. In each PBMC and LNMC sample, 78% of genomes were methylated at >1 CpG site in the env-tat- rev island and 46%were methylated at ≥3 CpG sites. Conclusion: Surprisingly, we did not find evidence that methylation of the 5’LTR promotor maintains HIV-1 latency in vivo, including LTRs of proviruses that are known to be intact and latent. Significant levels of methylation were found in a CpG island in env but its role, if any, in transcriptional silencing is unknown. Since it is well known that methylation of transcriptional enhancers located many kilobases frommRNA start sites can result in gene silencing, it is important to determine if the identified methylated CpG island in the env gene has any function in HIV-1 latency in vivo. 378 THE HIV ANTISENSE TRANSCRIPT AST INDUCES VIRAL LATENCY VIA SEVERAL SILENCING PATHWAYS Wei Zhou 1 , Kaveh Daneshvar 2 , Michelle Pleet 3 , Francesca Benedetti 1 , Gideon Wolf 1 , Davide Zella 1 , Fatah Kashanchi 3 , Balasubrahmanyam Addepalli 4 , Alan Mullen 2 , Fabio Romerio 1 1 University of Maryland, Baltimore, MD, USA, 2 Harvard University, Boston, MA, USA, 3 George Mason University, Fairfax, VA, USA, 4 University of Cincinnati, Cincinnati, OH, USA Background: We have reported that an antisense transcript (Ast) expressed from a promoter located in the HIV-1 3’LTR induces the establishment and maintenance of HIV-1 latency. We have shown that Ast recruits the Polycomb Repressor Complex 2 (PRC2) to the HIV-1 5’LTR. PRC2 catalyzes trimethylation of lysine 27 on histone H3 (H3K27me3), an epigenetic mark that leads to nucleosome assembly and transcriptional silencing. Methods: Ast mutants were tested after stable transduction in Jurkat E4 cells. To identify new binding partners, Ast was fused to a streptavidin-binding RNA aptamer, expressed in 293 cells, affinity-purified by streptavidin, and binding proteins identified by mass spectrometry (MS). To identify Ast modifications, Ast was affinity-purified via complementary biotinylated oligos and streptavidin, enzymatically digested and analyzed by MS. Results: To identify the functional domains of Ast, we divided the transcript into 5 segments and produced substitution and deletion mutants. A 376-nt segment at the 5’ end of Ast (5AST, mapping in the U3 region of the 3’LTR) mediates binding of Ast to the U3 region in the proviral 5’LTR via sequence homology. We divided the Ast sequence downstream of 5AST into four segments (A through D), and generated a substitution mutant for each of them. Mutation of segment A or B impacted significantly Ast function. Indeed,

these women initiated ART and were well suppressed for a median of 4.9 yrs. Plasma-derived virus was sequenced on average every six months from acute/ early infection to ART. These evolving sequences were compared to sequences of replication-competent reservoir viruses grown out of the latent reservoir. We used this same approach to analyse the viral reservoir in one woman for whom treatment initially failed. Illumina MiSeq with Primer ID was used to sequence partial env, gag and nef genes from pre-ART time points and PacBio sequencing was used to generate nearly full-length genome sequences of outgrowth viruses. The relatedness of reservoir and pre-ART viruses was evaluated using approximate maximum-likelihood analyses with phylogenetic placement. Results: Reservoir viruses (mean = 16; range 6-48) were sequenced from the 10 women. In the nine individuals on long-term suppressive ART, a median of 78% of reservoir viruses were most similar genetically to viruses circulating in the year before ART. We expand on this initial result by examining reservoir formation in an individual who experienced treatment failure but was virologically suppressed after ART optimization. In this individual, the reservoir contained a high percentage of variants from the time when she initiated her first and second ART regimen. Conclusion: In a cohort of nine well-suppressed women we observed that the vast majority of the persistent, replication-competent reservoir was established near the time of ART initiation. Analysis of one individual who experienced treatment failure suggests that variants may be seeded into the reservoir each time that an individual initiates therapy with a subsequent reduction in viral load. These observations suggest new strategies for reducing the size of the latent reservoir through viral clearance of variants circulating late in untreated infection. 376 TCR-ACTIVATED CD8+ T CELLS PROMOTE THE ESTABLISHMENT OF HIV LATENCY IN CD4+ T CELLS Michelle Zanoni , David J. Palesch, Claudia Pinacchio, Deanna Kulpa, Guido Silvestri Emory University, Atlanta, GA, USA Background: Virus persistence in latently-infected CD4+ T cells despite ART is the major barrier to cure HIV infection. While HIV-specific cytotoxic T lymphocytes are known to control virus replication, recent studies showed that CD8+ T cells may also suppress SIV transcription in ART-treated macaques. Identifying the mechanisms responsible for CD8+ T cell-mediated HIV silencing might reveal molecular targets to disrupt the establishment and maintenance of the HIV reservoir. Methods: CD4+ and CD8+ T cells were isolated from healthy donors, separately labeled with CellTrace Violet and CellTrace Red, and activated by TCR stimulation. CD4+ T cells were infected with an HIV reporter virus expressing eGFP under HIV-LTR control and CD8+ T cells were added for co-culture. Expression of activation markers (HLA-DR, HLA-ABC, and HLA-E), cell survival and cell proliferation were measured by flow cytometry. The non-productively- infected (eGFP-) CD4+ T-cell population frommono- or co-culture were sorted and the inducible HIV reservoir was quantified by measuring eGFP expression 24h post reactivation. Results: CD8+ T cells significantly suppressed eGFP expression in infected CD4+ T cells during co-culture as compared to CD4+ T cells cultured alone, under multiple-round and single-cycle infection conditions (mean 57%, p=0.0078, n=8, and mean 14%, p=0.0001, n=14, respectively). This observation suggests that CD8+ T cells not only inhibit virus spread but also suppress LTR-dependent viral transcription. Concomitantly, the suppressor activity of CD8+ T cells resulted in a 25% reduction of cell proliferation in the eGFP-CD4+ T cell population (mean fold change in CellTrace Violet MFI compared to CD4+ T cells alone, p=0.0009, n=14). Moreover, CD8+ T cells mitigated virus-induced cell death thus increasing CD4+ T cell survival (mean 10% increase in live CD4+T cells, p=0.0078, n=8) and down-modulated the expression of activation markers on both productively infected (eGFP+) and eGFP-CD4+T cells. Finally, CD8+ T cells increased the inducible HIV reservoir in CD4+ T cells by 62%, as shown by reactivation of sorted eGFP- CD4+ T cells from co-culture with CD8+ T cells as compared to CD4+ T cells frommono-culture (p=0.0078, n=8). Conclusion: TCR-activated CD8+ T cells from HIV uninfected donors reduce virus production by autologous in vitro HIV-infected CD4+ T cells by mitigating their level of cell activation and proliferation, and ultimately facilitate the transition of these CD4+ T cells into latent HIV infection.

Poster Abstracts

CROI 2019 135