CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

clones), or iii) identical proviruses resulted from some with identical integration sites and some with different sites (both cell and viral clones). Results: Of the 12 sets of identical P6-PR-RT sequences, 3 contained only identical integration sites (4, 2, and 2 MDA wells), 7 contained integration sites observed only once (2-5 MDA wells per set), and 2 contained a combination of identical (19 and 13 MDA wells) and singly-observed integration sites (4 and 2 MDA wells). From the 3 sets of identical env populations, one had integration sites observed only once (7 different sites) and 2 contained a combination of identical (3 and 28 MDA wells) and singly-observed integration sites (1 and 3 MDA wells). Two of the cell clones contained replication-competent proviruses, confirmed by VOA. Within the populations of defective proviruses, near-full length sequence analyses showed that sequences identical in sub-genomic regions were often identical throughout the genome, except for non- overlapping deletions, making it possible to reconstruct the sequence of the shared viral ancestor. Conclusion: The finding that identical proviruses can have different integration sites demonstrates that sub-genomic SGS is not sufficient to identify clones of infected cells and suggests that multiple infected cell clones can be established from the same viral ancestor. Such viral clones can arise from the transmission bottleneck, escape from immune pressure, or selection for drug resistant virus. 367 HIGH-THROUGHPUT SEQUENCING OF INTEGRATED HIV-1 REVEALS NOVEL PROVIRAL STRUCTURES Kevin W. Joseph 1 , Elias K. Halvas 1 , Leah D. Brandt 1 , Sean Patro 2 , Jason W. Rausch 3 , Mary F. Kearney 2 , John M. Coffin 4 , John W. Mellors 1 1 University of Pittsburgh, Pittsburgh, PA, USA, 2 National Cancer Institute, Frederick, MD, USA, 3 National Institutes of Health, Bethesda, MD, USA, 4 Tufts University, Boston, MA, USA Background: Efforts to cure HIV-1 infection will require a better understanding of the HIV-1 reservoir but characterizing individual integrated proviruses has remained difficult because of technical challenges related to the rarity of proviruses in CD4+ T-cells. Current approaches for sequencing integration sites using NGS are inefficient (most reads are off-target reads) and restricted read lengths can make it difficult to definitively identify both integration sites and proviral sequences. Methods: We have developed a new automated approach that sequences single HIV proviruses and their 5’ host integration sites by: i) amplifying the whole cellular genome at a proviral end point through multiple displacement amplification; ii) performing long-range PCR that amplifies variable and near- full length proviruses; and iii) performing nullomer-mediated PCR using a linker consisting of nullomer motifs absent in target genomes that markedly enhances specificity for integrated proviral targets. Amplicons can be sequenced by dideoxy (e.g., Sanger) and/or NGS methods. Results: Amplicons sequenced by NGS utilized >90% of reads on average during consensus generation for both proviral and integration site amplicons. The workflow sequences all but 69 bp of the 3’ LTR of the provirus. Across 5 donors, an average of 78% of HIV-positive MDA reactions yielded the 5’-host- virus junction containing 400 + 297 bp of flanking host sequence (compared to about 5 nucleotides by standard integrations site analyses) and 13.4% of proviruses were near-full length (determined by sequencing, N=33 out of 247 total proviruses). To date, the assay has been used to characterize a broad range of intact and defective integrated proviruses in blood mononuclear cells from donors on suppressive ART including replication-competent proviruses in cell clones (proven by viral outgrowth) and novel proviral structures such as asymmetrical LTR deletions (revealed by sequencing both LTRs) and genome inversions. The accuracy of the method has been confirmed by sequence identity with full-length and deleted proviruses amplified directly from blood mononuclear cells using host sequences flanking the integrated provirus. Conclusion: This novel integrated proviral sequencing assay provides an efficient and high-throughput means of characterizing HIV-1 reservoirs that need to be targeted to achieve a cure of HIV-1 infection.

368 CELL PROLIFERATION CONTRIBUTES TO THE INCREASE OF GENETICALLY INTACT HIV OVER TIME Bethany A. Horsburgh 1 , Bonnie Hiener 1 , Katie Fisher 1 , John-Sebastian Eden 1 , Eunok Lee 1 , Susanne von Stockenstrom 2 , Lina Odevall 2 , Jeffrey M. Milush 3 , Teri Liegler 3 , Rebecca Hoh 3 , Rémi Fromentin 4 , Nicolas Chomont 4 , Steven G. Deeks 3 , Frederick M. Hecht 3 , Sarah Palmer 1 1 The Westmead Institute for Medical Research, Westmead, NSW, Australia, 2 Karolinska Institute, Stockholm, Sweden, 3 University of California San Francisco, San Francisco, CA, USA, 4 Centre de Recherche du CHUM, Montreal, QC, Canada Background: Effective HIV eradication strategies require an understanding of the mechanisms maintaining persistent HIV during therapy. We examined the role of memory cell proliferation in maintaining genetically-intact proviruses over 4 years of effective therapy. Methods: Naïve (N), central (CM), transitional (TM) and effector (EM) memory CD4+ T-cells were sorted from the peripheral blood of two participants on long- term ART. Additional sequences from naïve, CM HLA-DR+/DR-, TM HLA-DR+/ DR- and EM HLA-DR+/DR- T cells were obtained 4 years later. Full-length individual proviral sequencing was used to characterise proviruses as intact or defective. Clusters of ≥2 100% genetically identical proviral sequences - indicative of host cell proliferation – were identified. Results: A total of 287 and 448 sequences were isolated from the first and second time-points, and 34 (12%) and 90 (20%) were considered intact. At both times the frequency of intact genomes differed between cell subsets, EM>TM>CM/N. In each subset, HLA-DR+ cells contained more intact provirus than HLA-DR- cells. The proportion of identical sequences was significantly higher in intact proviruses compared to defective at the second time-point (85% vs 41%, p=0.03), but not the first. There was a significant correlation at the second time-point between the proportion of identical sequences overall and the proportion of intact proviruses (R2=0.58-67, p=0.02-0.04). The majority (44/51, 86%) of sequences observed at both time-points were found in cells of the same memory phenotype. The number and size of identical sequence clusters differed depending on activation status. A greater number of identical sequence clusters were derived from HLA-DR+ cells. However, the size of clusters derived from cells of mixed activation status was larger, with 60% of all identical sequences derived from a cluster of both HLA-DR+ and HLA-DR- cells. Conclusion: Genetically intact proviruses were found most frequently in the more differentiated EM cells. However, the frequency of intact proviruses was increased in each memory cell subset when the cell expressed HLA-DR, highlighting the role of cellular activation in maintaining the reservoir. Moreover, the correlation between cellular proliferation and intact provirus highlights the importance of host cell proliferation in maintaining HIV over time. These findings demonstrate the importance of limiting cellular activation, differentiation and proliferation in strategies aimed at reducing the reservoir.

Poster Abstracts

CROI 2020 127