CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

Results: The median duration of the ATI phase was 57 days. All study participants experienced plasma viral rebound and resumed ART. HIV burden increased significantly in the CD4+ T cells during plasma viral rebound. However, the size of the HIV reservoirs, including the frequency of CD4+ T cells carrying replication-competent virus, returned to pre-ATI levels 6 to12 months after the study subjects resumed ART. Of note, the proportions of near full- length, genome-intact, and structurally defective HIV proviral DNA sequences were similar prior to ATI and following reinitiation of ART. Furthermore, no significant differences in immunologic or activation and exhaustion parameters were found between pre-ATI and post-ATI time points. Conclusion: Our data indicate that short-term ATI is not associated with permanent expansion of the persistent HIV reservoirs nor irreversible immune system abnormalities. These findings support the inclusion of ATI in future clinical trials when evaluating strategies for achieving ART-free remission. 335 SHORT-TERM ART INTERRUPTION HAS LITTLE EFFECT ON LEVELS OF INTEGRATED PROVIRAL DNA Zachary Strongin 1 , Radwa Sharaf 1 , Jake VanBelzen 2 , Jeffrey Jacobson 3 , Elizabeth Connick 4 , Paul Volberding 5 , Daniel Skiest 6 , Rajesh T. Gandhi 7 , Daniel R. Kuritzkes 1 , Una O’Doherty 2 , Jonathan Z. Li 1 1 Brigham and Women’s Hospital, Boston, MA, USA, 2 University of Pennsylvania, Philadelphia, PA, USA, 3 Temple University, Philadelphia, PA, USA, 4 University of Arizona, Tucson, AZ, USA, 5 University of California San Francisco, San Francisco, CA, USA, 6 Baystate Health, Springfield, MA, USA, 7 Massachusetts General Hospital, Boston, MA, USA Background: HIV analytic treatment interruption (ATI) represents the ultimate test to assess strategies aimed at achieving sustained ART-free remission, but questions remain about the effects of ATI on the viral reservoir. We validated a size-selection-based assay for quantifying integrated proviral HIV DNA as a measure of HIV reservoir size and assessed the impact of ATI on reservoir size after ART resumption. Methods: Cryopreserved PBMCs were obtained for 12 participants from previously completed ACTG ATI trials with available samples pre-ATI (T1) and ≥24 weeks after ART resumption (T3). Four participants also had samples available during the ATI (T2). Genomic DNA was separated from episomal DNA by size-selection using the BluePippin pulsed-field gel electrophoresis system and proviral DNA levels quantified by qPCR. Assay validation was first performed by spiking non-integrated HIV DNA into genomic DNA extracted from HIV-uninfected individuals to confirm successful elimination of the episomal HIV fragments. In addition, HIV-infected cell lines and participant samples were used to compare results to the standard Alu- gag assay. Results: The size-selection-based proviral DNA assay eliminated 99% of non-integrated HIV DNA species and correlated strongly with the established Alu- gag assay (Spearman r =0.94, P=0.02). The median ATI duration was 12 weeks (range 6-67 weeks). For the majority of individuals, integrated DNA levels increased during ATI (median +94 HIV DNA copies/10 6 PBMCs) and subsequently declined after ≥24 weeks of ART (median -109 HIV DNA copies/10 6 PBMCs). There was no significant difference in levels of HIV integrated DNA between the pre- and post-ATI time points. The median ratio of post:pre-ATI (T3:T1) HIV DNA levels was 0.95 (Q1, Q3: 0.8, 1.6). In those with higher HIV DNA levels post-ATI, the increase in proviral DNA levels was generally small, with a median change of 10.9 HIV DNA copies/10 6 PBMCs (Q1, Q3: 6.9, 19.9 HIV DNA copies/10 6 PBMCs). There was no significant correlation between duration of ATI and the ratio of post:pre-ATI reservoir size. Conclusion: We validated a size-selection-based proviral DNA assay that is less sample- and labor-intensive than the currently used assays. Levels of integrated HIV DNA showed minimal change after short-term ATI followed by ART resumption, suggesting that short-term ATI can be conducted without a significant impact on levels of integrated proviral DNA. 336 HIV REBOUND DRIVEN BY RARE SUPERSPREADING CD4+ T CELL LINEAGES Jason M. Hataye 1 , Joseph P. Casazza 1 , David R. Ambrozak 1 , Katharine Best 2 , C J. Liang 3 , Sam Darko 1 , Amy R. Henry 1 , Farida Laboune 1 , Taina T. Immonen 4 , Frank Maldarelli 4 , Takuya Yamamoto 5 , Daniel Douek 1 , Brandon Keele 4 , Alan S. Perelson 2 , Richard A. Koup 1 1 NIAID, Bethesda, MD, USA, 2 Los Alamos National Laboratory, Los Alamos, NM, USA, 3 NIAID, Rockville, MD, USA, 4 Frederick National Laboratory for Cancer Research, Frederick, MD, USA, 5 National Institutes of Biomedical Innovation, Health and Nutrition, Osaka, Japan

Background: The average number of new infections generated per individual, the basic reproductive number R0, has been estimated for within-host HIV productively infected cells, but cell variation in infectious potential remains unknown. Further, R0 was derived from the most rapid phase of exponential growth, whereas little is known of how such growth arises from a founder cell. Here we define the antecedent population dynamics of viral recrudescence from ex-vivo CD4+ T cells undergoing HIV latency disruption. Methods: Resting CD4+ T cells from 7 donors on antiretroviral therapy (ART) were stimulated and cultured in the presence of ART to quantify initial virus release by gag qRT-PCR in the absence of new infections, or in viral outgrowth cultures to determine the probability that an initial release resulted in productive infection. Single genome and deep sequencing of virus revealed clonality. The experimental results were fit to computational models to test key assumptions. Stochastic simulation allowed estimation of individual reproductive numbers for single latently infected CD4+ T cells. Results: R0 for latently infected CD4+ T cells was estimated at 4, lower than previously found for productively infected cells. The distribution giving rise to this mean was highly skewed; ~95% of single latent cells resulted in no new infections, while ~3% spawned a ‘superspreader’ lineage resulting in tens to hundreds of first round infections that were primarily responsible for establishment. Superspreading was an emergent system property resulting from at least three dynamic aspects of early virus release and growth: 1. The majority of initial latently infected CD4+ T cells and their progeny died before virus was released. 2. While the average total release from one productively infected cell was 1000 HIV RNA copies, variability in cell survival and proliferation had the potential to amplify and further disperse the total release due to one founder latent cell. 3. Establishment of productive infection was most likely when the initial release resulting from one or more latent cells exceeded a critical growth threshold of 5000 HIV RNA copies. Conclusion: Each dynamic aspect defined here is a potential viral vulnerability that could be targeted to prevent rebound. Further, the extreme variability noted among latent cell infection potentials must be accounted for to realistically predict in-vivo time to rebound following ART interruption, and thus the clinical efficacy of cure intervention candidates. 337 HIV-1 VIRAL REBOUND AND SAFETY OUTCOMES OF POSTPARTUM TREATMENT INTERRUPTION IN WOMEN Catherine N. Le 1 , Paula Britto 2 , Sean Brummel 2 , Risa M. Hoffman 1 , Patricia M. Flynn 3 , Taha E. Taha 4 , Anne Coletti 5 , Mary Glenn Fowler 6 , Karin L. Klingman 7 , James A. McIntyre 8 , Jonathan Z. Li 9 , Judith S. Currier 1 1 University of California Los Angeles, Los Angeles, CA, USA, 2 Harvard University, Cambridge, MA, USA, 3 St. Jude Children’s Research Hospital, Memphis, TN, USA, 4 The Johns Hopkins University, Baltimore, MD, USA, 5 FHI 360, Durham, NC, USA, 6 Johns Hopkins Hospital, Baltimore, MD, USA, 7 DAIDS, NIAID, Bethesda, MD, USA, 8 Chris Hani Baragwanath Hospital, Johannesburg, South Africa, 9 Brigham and Women’s Hospital, Boston, MA, USA Background: Structured, temporary treatment interruptions are useful in studies investigating aspects of HIV cure by further characterizing the size and dynamics of the latent viral reservoir. However, the associated short-term safety is not well-established, particularly in women. Methods: Participants from the 1077HS and 1077BF/FF components of the PROMISE trial were included in this analysis. Asymptomatic HIV-positive women with baseline CD4 T-cell counts ≥ 350 cells/mm 3 were randomized up to 42 days after delivery to continue or discontinue ART. All had HIV-1 RNA ≤ 400 copies/ml within 30 days of randomization. The preferred ART regimen was emtricitabine/ tenofovir disoproxil fumarate plus lopinavir/ritonavir. We evaluated time-to- detectable viral load, grade 2 or higher sign/symptom and laboratory safety events, HIV/AIDS-related events, and WHO stage 2 and 3 clinical events in the first 30 weeks of treatment interruption. HIV-1 RNA was measured at 4 weeks for 1077HS and 6 weeks for 1077BF/FF, followed by 12 week intervals. Time to virologic rebound was estimated as the time of randomization to first positive HIV-1 RNA above the limit of detection. Survival probability estimates were calculated using interval censored methods. Safety events were summarized counting the highest grade for each participant. Results: 1076 eligible women discontinued ART in the 1077HS and BF/FF trials. Median age was 28 years, CD4 count 766 cells/mm 3 (IQR 618, 957 cells/mm 3 ). Median duration on ART before discontinuation was 17 weeks. Median time to virologic rebound by interval censoring method was 2 weeks (95%CI 1.6, 2.46). The proportion who remained suppressed off ART at 8, 12 and 24 weeks was

Poster Abstracts

CROI 2018 117