CROI 2024 Abstract eBook

Abstract eBook

Poster Abstracts

naïve T cells, whereas CCR5 is associated with memory T cells, viral tropism has direct implications in the composition of the persistence viral reservoir. Here, we longitudinally evaluate viral tropism using V3-loop sequences derived from an acute infection cohort. Methods: The FRESH cohort in Durban, South Africa enrolls young women at high risk for acquiring HIV infection who undergo HIV-1 RNA surveillance testing twice a week. We studied 24 women who were detected during acute HIV-1 infection and were followed up to 1102 days: 11 were subsequently treated after >1 year, and 13 were treated 1-3 days post-first-detection. Near-full-length single-genome HIV-1 DNA sequencing (HXB2 638-9632) was performed on longitudinal PBMC samples, and genomes were classified as intact or defective using HIVSeqinR. Viral subtyping was performed with RIP 3.0. Viral tropism was inferred using V3-loop sequences with a subtype-specific algorithm PhenoSeq. Results: We obtained 697 viral DNA genomes, of which 320 contained full length Env (247 intact and 73 defective) representing all 24 donors. Intact viral DNA genomes were detected in 23 out of 24 donors: V3-loop analysis revealed that 61% donors were entirely R5-tropic (n=13 subtype C, n=1 A1/K), 26% donors were entirely X4-tropic (n=6, all subtype C), and 13% donors had a mix of R5- and X4-tropic viral DNA genomes (n=3 all subtype C). Tropism remained stable in 82% (14/17) of donors who had longitudinal data, including six who initiated antiretroviral therapy within 1-2 days post-detection, as well as eight who initiated treatment during chronic infection due historic treatment guidelines and remained viremic for a median of 332 days (IQR 330-441). Mixed tropic populations were observed only in donors who were untreated during acute infection (n=3). Conclusion: We detected an unexpected high prevalence of X4-tropic archived intact HIV-1 DNA in acute infections in FRESH by PhenoSeq, confirmed by geno2pheno[coreceptor] and C-PSSM (results not shown). Given that previous studies have associated subtype C and acute infections with mostly R5-tropic variants, our results suggest phenotypic studies are needed to further assess the extent and roles of archived X4-tropic variants in intact HIV-1 DNA genomes, which may have implications in the types of cells that would fuel virologic rebound during treatment interruptions. Multiomics Profiling of HIV-Transcribing Cells in People With HIV on Suppressive ART Julie Frouard 1 , Xiaoyu Luo 1 , Sushama Telwatte 2 , Douglas Arneson 3 , Jason Neidleman 1 , Kailin Yin 1 , Antoine Chaillon 4 , Rebecca Hoh 3 , Steven G. Deeks 3 , Sara Gianella Weibel 4 , Davey M. Smith 4 , Sulggi A. Lee 3 , Phyllis Tien 3 , Steven A. Yukl 3 , Nadia R. Roan 1 1 Gladstone Institutes, San Francisco, CA, USA, 2 Peter Doherty Institute, Melbourne, Australia, 3 University of California San Francisco, San Francisco, CA, USA, 4 University of California San Diego, La Jolla, CA, USA Background: HIV reservoir cells actively transcribing HIV persist despite ART, but remain poorly characterized since standard single cell (sc)RNAseq approaches inefficiently identify these cells. In this study, we devised a comprehensive scRNAseq-based approach that increases identification of HIV- transcribing cells, enabling a deep characterization of these cells. Methods: We developed HIV-seq, an approach whereby HIV primers targeting multiple regions of the HIV genome are added during scRNAseq library preparation to increase our ability to identify HIV RNA+ cells from people with HIV (PWH) under suppressive ART. We paired HIV-seq with a new bioinformatics pipeline, named 2-Step HIV Alignment, to further increase our ability to identify HIV transcripts from scRNAseq datasets. This entailed realigning initially unmapped reads to 1) a subtype B consensus sequence under reduced stringency conditions, and 2) participant-autologous viral sequences. HIV-seq combined with 2-Step HIV Alignment were implemented in the context of multi-omic single-cell sequencing analyses where transcriptomic (scRNAseq), surface phenotypic (CITEseq), and clonal expansion (TCR analysis) data were simultaneously captured on the same cells. A total of 18 PWH under ART were recruited for this study, and 4 of them also provided viremic samples. Results: Using this combined approach, we were able to increase by up to 40% our ability to detect HIV RNA+ cells from PWH. HIV RNA+ cells were identified from all blood samples and 7 tissue sites (spleen, lymph nodes, ileum, colon, rectum, endometrium, and endocervix) of 15 donors on suppressive ART. The host transcriptome of these HIV RNA+ cells differed in a tissue site-dependent manner, and included Th17 and Trm cells. Unlike HIV RNA+ cells from viremic individuals, those from suppressed individuals did not preferentially exhibit a cytolytic phenotype characterized by higher expression levels of granzymes,

impacted the env gene. A subset of proviruses were extensively hypermutated in all subtypes. Initial sliding window analysis of deletion position and frequency suggests comparable IPDA proviral discrimination across subtypes using psi and env targets. Optimization of NFL-PCR resulted in substantially increased overall provirus recovery efficiency, and genetically intact proviruses were quantified at a higher frequency in all subtypes (including B) compared with other reports. Conclusion: The consistency in HIV-1 provirus deletion frequency and position across Group M suggests that conserved mechanisms drive formation of defective proviruses. The dominance of env deletions has important implications for antibody-mediated reservoir immune selection and for global deployment of env-targeting therapeutics. Our improvements to long-range proviral PCR reduce the bias against intact proviruses, resulting in a more accurate analysis of proviral landscape composition in PWH on ART, and will enable comparison of short- and long-range PCR-based reservoir quantification assays. Importantly, our results support the feasibility of a unified IPDA version 2 design with robust performance across Group M for global use. Solo LTR Formation Promotes Elimination of Proviruses in Persons Living With HIV Feng Li , Guanhan Li, Francesco R. Simonetti, Shawn Hill, Robert Gorelick, Chuen-Yen Lau, Frank Maldarelli National Institutes of Health, Frederick, MD, USA Background: A major obstacle for curing persons living with HIV (PLWH) is persistence of HIV-infected cells during combination antiretroviral therapy (ART). Most persistent cells harbor defective proviruses, including solo LTRs, which are the result of cellular homologous recombination during DNA replication and clonal expansion; the recombination event excises the entire coding sequence of HIV, leaving only a solo LTR provirus. Solo HIV LTRs are reported (Anderson et al., 2020; Botha et al. 2023) and may dominate proviral populations, but little is known about their frequency and persistence during ART. We combined a new screen for solo LTR with single genome sequencing (SGS) and integration site analysis (ISA) to identify and characterize solo LTR in PLWH. Methods: Peripheral blood lymphocytes (PBLs) from adults undergoing ART in clinical protocols at NIH were screened using multiplexed droplet digital PCR (ddPCR) that targets LTR and sequences downstream from the 5' LTR or upstream of the 3'LTR. A high ratio of LTR to non-LTR sequence implies an excess of LTRs and the potential presence of solo LTR proviruses. We developed specific ddPCR to characterize and quantify levels of HIV clones, and designed bridging PCR strategies that amplify solo LTR to validate provirus structure. Integration sites and single genome sequencing were performed in PBL and autopsy tissue when available. Results: Analysis of PBL from PLWH (N=48) with LTR/gag ddPCR revealed > 90% had excess LTRs (median: 832 excess LTRs/million CD4 cells; range <1 21651). Large clones of solo LTRs were identified in two PLWH; quantification of one clone integrated in HORMAD2 (Anderson et al., 2020) with bridging ddPCR demonstrated this clone now persists 10% of total proviruses >20 years after first identified during early ART. In a second PLWH, analysis of PBL and neoplastic tissue revealed a solo LTR integrated near the FAM9C gene on the X chromosome; As FAM9C is not located pseudoautosomal recombination regions (PARs) where homologous recombination on the X chromosome is restricted, these data demonstrate that cellular mechanisms responsible for homologous recombination are functional outside of PARs. Conclusion: Screening for highly deleted proviruses will be useful in identifying solo LTR proviruses. HIV solo LTR provirus form in vivo and may be prevalent among proviral populations in adults. Understanding solo LTR formation provides a new insight into requirements for eliminating functional viral sequences from host genome. Longitudinal Evaluation of Viral Tropism in Archived Intact DNA Genomes in Acute HIV-1 Infections Kavidha Reddy 1 , Elena Giorgi 2 , Kamini Gounder 1 , Nelson Sonela 3 , Krista L. Dong 4 , Bruce D. Walker 4 , Thumbi Ndung'u 1 , Guinevere Q Lee 3 1 Africa Health Research Institute, Mtubatuba, South Africa, 2 Fred Hutchinson Cancer Center, Seattle, WA, USA, 3 Weill Cornell Medicine, New York, NY, USA, 4 Ragon Institute of MGH, MIT, and Harvard, Cambridge, MA, USA Background: HIV-1 tropism refers to the ability of the virus to infect CCR5- and/ or CXCR4-expressing CD4+ cells. Since CXCR4 expression is associated with

464

Poster Abstracts

466

465

CROI 2024 119