CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

that these markers could be used as potential biomarkers predicting viral rebound.

using a real-time qPCR assay that distinguishes proviruses with deoxyuracils or thymidines, we found the proviruses in resting T cells had deoxyuracils instead of thymidines. Lastly, we revealed Vpr protected proviruses from an UNG- dependent inactivation mechanism. Conclusion: We conclude that HIV can directly infect primary resting memory CD4 T cells to establish the reservoir. HIV-infected resting CD4 T cells incorporate deoxyuracils, which is deleterious in the absence of Vpr inhibiting UNG. Finally, we believe the integrated HIV genome persists through transcription and alternatively splicing for mRNAs encoding Nef, Vif, Vpr, or Vpu-Env. Olivia D. Council 1 , Melissa-Rose Abrahams 2 , Sarah B. Joseph 1 , Nigel Garrett 3 , Matthew Moeser 1 , Shuntai Zhou 1 , Lynn Tyers 2 , David Matten 2 , Colin Anthony 2 , Sergei L. Kosakovsky Pond 4 , Nancie Archin 1 , David M. Margolis 1 , Salim S. Abdool Karim 5 , Ronald Swanstrom 1 , Carolyn Williamson 2 1 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 2 University of Cape Town, Cape Town, South Africa, 3 University of KwaZulu-Natal, Durban, South Africa, 4 Temple University, Philadelphia, PA, USA, 5CAPRISA, Durban, South Africa Background: All HIV-infected people on ART have a long-lived reservoir. We recently showed that the replication-competent portion of this reservoir originates from viruses circulating near the time of ART initiation, similar to a previous report that examined total viral DNA. Here we examine both the replication-competent reservoir and the viral DNA reservoir in the same set of participants. Methods: Plasma was collected longitudinally from 16 women in the CAPRISA- 002 cohort pre-ART, with PBMCs then collected after 4.8 years (average) of suppressive ART. MiSeq with Primer ID was used to sequence 5 genomic regions from RNA in the pre-ART plasma samples. Outgrowth virus was generated from quantitative viral outgrowth assays (QVOA) using resting CD4+ T cells collected post-ART. PCR was used to generate overlapping half genome amplicons from QVOA-derived viral RNA and from total cellular DNA from the post-ART PBMCs and sequenced using PacBio with barcodes. Phylogenetic trees were constructed using all pre-ART sequences and reservoir sequences. Reservoir entry time was estimated by the phylogenetic relationship between each reservoir sequence and the pre-ART sequences. Results: A median of 10 (range: 4 to 54) reservoir sequences were generated for each participant. In all 5 women with both QVOA and DNA sequences, we did not detect a difference in the timing of establishment of the DNA compared to the replication-competent reservoir (Fisher’s exact test, all p > 0.05). For the overall cohort (N=16; 4 DNA only, 7 QVOA only and 5 DNA and QVOA), a median of 71% of reservoir sequences were seeded in the year before ART initiation. In one individual where only late viruses had been detected using QVOA, deeper sampling of viral DNA identified a minority of early viruses, consistent with the potential for virus to enter the reservoir when active replication is ongoing. Conclusion: Viral evolution prior to ART was used to date when both replication-competent viruses and proviral DNA were seeded into the long-lived reservoir. We observed no difference in when these reservoirs formed; both formed predominantly around the time of ART initiation. Our results suggest that the probability an infected cell contributes to the long-lived HIV-1 reservoir is largely determined by the biology of the infected T cell, not the provirus that it carries. In this interpretation a larger population of cells transition to a long-lived state around the time of ART initiation, with some of these cells being nonproductively infected.

328 ESTABLISHMENT OF THE HIV-1 DNA RESERVOIR MIRRORS THE REPLICATION-COMPETENT RESERVOIR

Poster Abstracts

327 HIV DIRECTLY INFECTS RESTING MEMORY CD4 T CELLS

Rodrigo Matus-Nicodemos 1 , David R. Ambrozak 1 , Sam Darko 1 , Amy Ransier 1 , Daniel Douek 1 , Richard A. Koup 1 1 Vaccine Research Center, NIAID, Bethesda, MD, USA Background: The establishment of the latent HIV reservoir in resting memory CD4 T cells occurs early in infection. Resting CD4 T cells are more difficult to infect than activated CD4 T cells. Therefore, the HIV reservoir is thought to form when HIV infects a few activated CD4 T cells that are resting down. Furthermore, HIV encodes four proteins: Vif, Vpr, Vpu, and Nef, which play an important role for the persistence of HIV. For example, Nef is known to downregulate MHC class I molecules (pMHCs) which prevents the recognition by CD8 T cells. However, the precise timing of expression of these four HIV proteins and the downregulation of their targeted host proteins in resting memory CD4 T cells is unknown. Methods: We explored this question by direct infection and longitudinal analysis of primary resting CD4 T cells with a CCR5-tropic replication-competent reporter virus in which GFP reports the expression of Nef. We then measured pMHCs by flow cytometry and performed bulk and scRNAseq of sorted GFP+ cells to measure host and HIV mRNAs. We also performed scATACseq to identify the sites of HIV integration to determine their influence on the timing of Nef expression. Results: We detected resting memory GFP+ cells 3 to 4 days after infection. These GFP+ cells showed low surface levels of pMHCs. By scRNAseq HIV mRNAs were identified in GFP+ cells and they encoded for Nef, Vpr, Vif or Vpu-Env, but never for Gag-Pol, Tat or Rev. The analysis of scATACseq and scRNAseq revealed this differential expression of HIV mRNAs was due to HIV integration into genes that are stochastically transcribed across resting CD4 T cells. Importantly, RNAseq analysis identified the expression of a cell-cycle independent form of ribonucleotide reductase, which converts ribonucleotides to deoxynucleotides. Also, the pathway for thymidine synthesis was not active in resting T cells. Thus,

CROI 2020 113