CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

FACS-sorted into 96-well plates and multiplex RT-qPCR (BiomarkHD) was used to quantify 88 human RNAs previously implicated in HIV infection/latency and 8 HIV targets (5'LTR, Gag, Pol, Nef, MS Tat-Rev, U3-PolyA, and the IPDA assays for Env&Gag). We compared HIV-unexposed, HIV-exposed but uninfected, and latently- +/- productively-infected populations from each model to identify genes with ≥2-fold difference in median expression levels and P<0.05(*) or FDR-corrected P<0.05(**). Results: As expected, multiple HIV targets(**) distinguished uninfected, latently-infected, and productively-infected cells. Each model differed in the cellular factors that distinguished populations, with some differences between donors. Compared to HIV-unexposed cells, latently-infected cells from the Verdin model showed higher expression of CXCR4(**), POL2RA(**), APOBEC3G(*), and STING(*) and lower expression of PRMT6(*), while latent cells from the Bosque model expressed higher levels of CGAS(**), and latent tonsil cells from the Greene model showed higher expression of CDK7, PBAF, RIG-I, and MDA5 (*for all). Compared to HIV-exposed but uninfected cells, latently- infected cells showed: 1) less CCR5(*), CD38(*), and NF-KBIA(*), but higher CD25(*) expression in the Verdin model; 2) less Cyclin L2(*) and more BCL6(*) in the Bosque model; and 3) no difference (except HIV targets[**]) in blood cells from the Greene model. Relative to productively-infected cells, latently- infected cells upregulated CTLA-4, BCL-11B, NFATC1, CDK7, HTATSF1, PAF-1, and PBAF expression (**for all) in the Verdin model, and exhibited lower expression of CD28, CTLA-4, PD-1, BCL-6, BCL-11B, FAS, Sp1, POLR2A, CREBBP, G9a, STAT1, and IRF9 (**for all) in the Greene tonsil model. Conclusion: Our single cell analysis reveals multiple cellular factors that distinguish latently-infected cells from uninfected and productively-infected cells, that may provide a molecular signature necessary to discriminate this population in vivo. 307 PROVIRAL/HUMAN GENOMIC CROSSTALK IN CELLULAR MODELS FOR HIV INFECTION Ulrike C. Lange 1 , Christoph P. Schwarz 2 , Julia K. Bialek 2 , Thomas Walther 2 , Roxane Verdikt 3 , Carine M. Van Lint 3 , Joachim Hauber 2 1 University Medical Center Hamburg–Eppendorf, Hamburg, Germany, 2 Heinrich Pette Institute, Hamburg, Germany, 3 Vrije Universiteit Brussel, Brussels, Belgium Background: Chronic HIV-1 infection is characterized by accumulation of proviral sequences in the genome of HIV target cells. Integration of viral-derived DNA is found at preferential loci, suggesting site-specific crosstalk between viral sequences and human genes. This crosstalk has been postulated to play a role in emergence of clonal infected cell populations. The molecular nature of this phenomenon is unclear. Paucity of HIV-infected cells in chronically infected individuals and lack of markers for HIV reservoir cells preclude functional studies in primary patient-derived cells. Methods: CRISPR/Cas9-based homologous recombination was used to target HIV-derived reporter sequences to genomic sites in T cell-derived immortalized cells. Clonal lines were generated and multiple screening steps used to verify correct targeting. Cell models were analyzed for LTR inducibility and epigenetic regulation/ transcriptomic effects of LTR activity. Results: We have established a workflow to generate cellular models for HIV infection that recapitulate proviral integration at selected genomic loci. Using this workflow, we have derived several BACH2-HIV-1 reporter models that mimic integration of proviral DNA in the BTB Domain and CNC Homolog 2 (BACH2) locus, which has been associated with recurrent integration and HIV-reservoir maintenance in chronically infected patients. We show that LTR transcriptional activity is repressed in BACH2 intronic regions associated with proviral-DNA integrations in vivo. This repression is not observed if proviral-sequences are targeted to regions that do not correlate with sites observed in patients. We demonstrate that these findings are reflected in epigenetic modifications on LTR regulatory regions. Furthermore, to study genome-wide effects of proviral/ human crosstalk at the BACH2 locus, we have undertaken transcriptome analysis in different BACH2-HIV-1 models in latent as well as LTR-activating conditions for which results will be presented. Conclusion: Our workflow is an adaptable tool for functional studies of proviral/human crosstalk. We show features of such crosstalk for the BACH2 locus, indicating that clustered BACH2 proviral integrations in vivo might be due to site-specific effects on LTR activity.

305 EPIGENOMIC CHARACTERIZATION OF A PRIMARY CELL MODEL OF HIV LATENCY Barcley T. Pace 1 , Stuart R. Jefferys 1 , Joel Parker 1 , Raghu Dronamraju 1 , Brian D. Strahl 1 , David M. Margolis 1 , Edward P. Browne 1 1 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Background: Transcriptional silencing of HIV in CD4 T cells generates a reservoir of latently infected cells that can reseed infection after interruption of therapy. As such, these cells represent the principal barrier to curing HIV infection, but little is known about the characteristics or regulation of the latent reservoir. Methods: To further our understanding of the molecular mechanisms of latency, we employed a primary cell model of HIV latency in which infected cells adopt heterogeneous transcriptional fates with a subset of infected cells establishing viral latency. We characterized this model using assay of Transposon-Accessible Chromatin sequencing (ATACseq). Results: We observed that loss of viral gene expression is a stable and heritable phenotype that is maintained through multiple rounds of stimulation and expansion, suggesting a role for epigenetic maintenance of latency. Using ATACseq we found that cells in which latency is established exhibit a significantly more closed chromatin conformation, both within the HIV genome and across the host cell genome, indicating that latency is correlated with a global process of epigenomic modification and heterochromatin expansion. We also observed that latency reversing agents (LRAs) induced distinct patterns of chromatin openning in both the HIV and host cell genomes. Furthermore, we observed that latently infected cells exhibited elevated levels of specific repressive histone modifications, including H3K27me3. Conclusion: Altogether, these data demonstrate that latency establishment in primary CD4 T cells occurs preferentially in a subset of cells that exhibit expanded H3K27me3-associated heterochromatin, and that viral silencing is connected to global cellular epigenomic reprogramming. A deeper understanding of this process will likely lead to new therapeutic strategies for blocking the initiation or maintenance of latency. 306 SINGLE-CELL ANALYSIS SHOWS MOLECULAR SIGNATURES OF HIV LATENCY IN PRIMARY CELL MODELS Sushama Telwatte 1 , Mauricio Montano 2 , Rachel S. Resop 3 , Emilie Battivelli 4 , Sara Moron-Lopez 1 , Eric Verdin 4 , Warner C. Greene 2 , Alberto Bosque 3 , Joseph K. Wong 5 , Steven A. Yukl 5 1 University of California San Francisco, San Francisco, CA, USA, 2 Gladstone Institutes, San Francisco, CA, USA, 3 George Washington University, Washington, DC, USA, 4 The Buck Institute for Research on Aging, Novato, CA, USA, 5 San Francisco VA Medical Center, San Francisco, CA, USA Background: Primary cell models have greatly advanced our understanding of HIV latency. However, it is unclear what mechanisms underlie latency in these primary cell models. We hypothesized that molecular signatures can distinguish uninfected, latently- and productively-infected populations in these models. Methods: We assessed 4 primary cell models [blood CD4T cells: models from labs of Eric Verdin, Alberto Bosque, and Warner Greene; tissue(tonsillar) CD4T cells: model fromWarner Greene]. Single cells from each model (2 donors) were

Poster Abstracts

CROI 2020 105