CROI 2024 Abstract eBook

Abstract eBook

Poster Abstracts

Conclusion: Our flow-FISH approach was able to detect and differentiate between cells with transcriptionally latent or active HIV in PBMCs from PWH without the need for ex vivo stimulation. Transcriptomic profiling showed an association between diminished mitochondrial functioning and the transcriptional activity of the viral reservoir. These findings underline the relevance of altered cellular metabolism in HIV infection, and support the development of therapeutics that take this into consideration.

Methods: Using cells from 12 PLWH suppressed on ART for > 3 years cell pellets and culture supernatants from 20 million PBMC in triplicate independent tests were assessed for p24 protein by Simoa (PMID: 33796087) after 72 hours treatment with Mukungulu (1 μg/mL) or anti-CD3/CD28 positive control. Isolated CD4 T-cells (5 million) were also tested in 3 donors. Intact and defective proviral DNA were assessed by IPDA (PMID: 30700913). 12 BLT-Humanized male mice were infected with HIVSUMA, suppressed with ART for 7 weeks (PMID: 36460646), and injected i.p. with 5 μg/mL Mukungulu (N=7) or PBS vehicle (N=5) and measured after 24 hours for pVL and vRNA from human cells. Analysis was done with Prism software. Results: In PBMC from 12 ART suppressed donors, Mukungulu induced 0.30 ± 0.15 and 0.46 ± 0.24 pg/mL of p24 protein in pellets and supernatants, respectively, without cytotoxicity, compared to only 0.11 ± 0.06 and 0.17 ± 0.08 pg/mL of p24 induced by anti-CD3/CD28 (p = 0.06 and 0.07). In contrast, Mukungulu induced no detectable p24 in pellets or supernatants in isolated CD4+ cells, compared to anti-CD3/CD28 inducing 0.49 ± 0.12 and 0.38 ± 0.14 pg/mL of p24. Notably, p24 levels from PBMC pellets and supernatants induced by Mukungulu correlated well with intact provirus reservoir size (r2 = 0.60 and 0.74, respectively) but not with defective provirus reservoir size (r2 = 0.03 and 0.08). Finally, in humanized mice, Mukungulu induced 1336 ± 402 vRNA copies / million human cells plus a pVL of 391 ± 394 copies / mL, compared to no change by PBS, without obvious toxicities. Conclusion: Mukungulu is a potent LRA ex vivo and in vivo. It induces virus production from the intact viral reservoir of CD4+ cells with ~2.5 more activity than anti-CD3/CD28. However, this reactivation is dependent on PBMC and not isolated CD4+ cells, indicating that additional cell-types contribute to its mechanism of latency reversal. Transcriptomic HIV Reservoir Profiling Reveals a Role for Mitochondrial Functionality in HIV Latency Shirley Man , Stefanie Kroeze, Jade Jansen, Teunis B. Geijtenbeek, Neeltje Kootstra Academic Medical Center, Amsterdam, Netherlands Background: Improved characterization of the HIV reservoir is crucial for devising effective cure strategies. We developed a strategy for isolating and characterizing the viral reservoir in peripheral blood from people with HIV (PWH). We hypothesize that short abortive HIV transcripts are present in latently infected cells, and based on this we developed an innovative flow cytometry fluorescent in situ hybridization (flow-FISH) allowing HIV reservoir detection and cell sorting without prior activation. This method was used for direct ex vivo detection and isolation of HIV+ CD4 T cells harboring transcriptionally latent or active HIV and allowed for transcriptomic analysis of the viral reservoir. Methods: Peripheral blood mononuclear cells (PBMCs) from 21 ART-naïve PWH from the Amsterdam Cohort Studies were used for this study. Flow-FISH was performed with probes targeting either abortive (TAR+Gag-) or elongated HIV transcripts (TAR+Gag+), representing latently and productively infected cells, respectively. Flow cytometry sorting was used to isolate three distinct cell populations (i.e. TAR+Gag-, TAR+Gag+, and probe-negative) from CD4 T cells of five PWH. The transcriptomic profile was determined by 3' RNA sequencing (RNAseq). Results: Our flow-FISH method detected between 10-751 HIV+ cells per 10 4 CD4 T cells in PBMCs from PWH, of which 1-47 per 10 4 CD4 T cells harbored transcriptionally latent HIV (TAR+Gag-) and 7-727 per 10 4 CD4 T cells harbored transcriptionally active HIV (TAR+Gag+; Figure A). Supervised RNAseq analysis allowed for the identification of transcriptomic signatures that separate the isolated populations independently of person-related characteristics. Notably, we identified several differentially expressed mitochondrial genes in latently infected (TAR+Gag-) compared to productively infected (TAR+Gag+) CD4 T cells (Figure B). Interestingly, enhancing mitochondrial function increased HIV transcriptional activity in latently infected CD4 T cells from PWH.

477

Investigating the Effects of Chromatin Spatial Folding on HIV-1 Latency Nawal K Al Burtamani 1 , Jose De Las Heras 2 , Khushi Goel 1 , Alex Ward 1 , Helen Rowe 3 , Ariberto Fassati 1 1 University College London, London, United Kingdom, 2 University of Edinburgh, Edinburgh, United Kingdom, 3 Queen Mary University of London, London, United Kingdom Background: Persistence of HIV-1 in latent reservoirs is a major obstacle that prevents its eradication and cure. Understanding the mechanisms governing HIV-1 latency is crucial to eliminate this reservoir. Spatial folding of chromatin governs cellular gene expression. However how 3D genome organization affects HIV-1 latency is unknown Methods: To investigate the impact of chromatin spatial folding on HIV-1 latency, we used a comparative approach that takes advantage of the different integration profiles of wild type (WT) HIV-1 and a capsid mutant virus N74D. We mapped unique integration sites (UISs) for HIV-1 WT and N74D to DamID and HiC data. Next, to investigate how different integration site distribution may affect latency, jurkat cells were transduced with either WT or N74D single cycle HIV-1 vectors expressing GFP from the LTR. GFP+ cells were sorted after 48 hrs, and latency was established. Lastly, to identify UISs functionally relevant for latency, latently infected cells (WT and N74D) were stimulated serially by TCR engagement followed by inhibition of HDAC6 and sorted into reversible (GFP+) and deep (GFP-) latency populations at each step Results: We observed that WT HIV-1 integrated more frequently in A1 compartment, whereas N74D virus integrated more frequently into lamina associated domains (LADs). These results were confirmed in ART-treated patients' cells. We found that WT virus becomes latent more slowly and can be reactivated more efficiently than N74D virus by TCR stimulation or PMA. Epigenetic profiling showed that the pan histone deacetylase inhibitor SAHA and the G9a inhibitor BIX-0124 reversed both WT and N74D latency with equal efficiency; however, selective inhibition of HDAC6, inhibition of FOXO-1 or AzadC reversed WT more potently than N74D latency. Integration site analysis on sorted GFP+ and GFP- cell populations revealed that, in WT HIV-1 infected cells, latency was associated with significantly fewer UISs in the A1 compartment and significantly more UISs in the B1 compartment. Deep latency was associated with more UISs in intergenic regions and fewer UISs in exonic regions than reversible latency. In N74D virus infected latent cells, deep latency was associated with greater distance of UISs from transcriptional start sites Conclusion: We propose that deep latency may be linked to integration into particular 3D chromatin regions enriched for certain epigenetic marks and gene clusters, which might be specifically targeted to either induce, or prevent, virus reactivation. Modulation of IRF7-Driven Transcription as a Strategy to Control HIV-1 Latency Ifeany Ezeonwumelu 1 , Edurne Garcia-Vidal 2 , Eudald Felip 2 , Bonaventura Clotet 2 , Roger Badia 3 , Ester Ballana 2 , Eva Riveira-Muñoz 2 1 University of California San Francisco, San Francisco, CA, USA, 2 IrsiCaixa Institute for AIDS Research, Badalona, Spain, 3 Institute for Health Science Research Germans Trias i Pujol, Badalona, Spain Background: The persistence of a latent viral reservoir represents a major barrier shared by current HIV cure strategies. Emerging evidence suggests that

Poster Abstracts

476

478

CROI 2024 123