CROI 2017 Abstract e-Book

Abstract eBook

Oral Abstracts

correlation of neutralization activity with trimer binding is well established we show here for the first time its efficacy of predicting neutralization breadth across HIV-1 subtypes. Based on epitope specificity (CD4bs, MPER, V1V2-loop, V3-loop) delineation analyses for 105 bnAb plasmas, we further found various binding patterns significantly associated with the presence of bnAbs directed to different antibody epitopes, highlighting that binding profiles should be included in plasma specificity delineation approaches. Conclusion: Building on the Swiss 4.5K Screen we had the unique opportunity to assess the validity of binding antibodies as predictors of neutralization activity in several thousand patients. Our data highlight that a reliable prediction of bnAb activity by binding antibodies is possible. Whether used as first-line assay or in combination with multi- clade virus neutralization, owing to their ease in standardization binding assays will be of immense value for a controlled assessment of forthcoming vaccine efficacy trials. 67 THE IMPACT OF HIV-GENETICS ON IMPRINTING ANTIBODY RESPONSES: THE SWISS 4.5K SCREEN Roger Kouyos , Peter Rusert, Claus Kadelka, Alex Marzel, Thomas Liechti, Merle Schanz, Hanna Ebner, Huldrych F. Günthard, Alexandra Trkola , for the Swiss HIV Cohort Study Univ of Zurich, Zurich, Switzerland Background: A key question on the way towards an HIV-1 vaccine is the impact of viral genetics on the neutralization response: Will a given HIV-1 envelope immunogen induce the same antibody specificities across vaccinees? Here we addressed this by studying the influence of virus genetics in potential transmission pairs identified within a recently conducted systematic screen of HIV-1 neutralization in 4,484 individuals (Swiss 4.5K Screen; Rusert, Kouyos Nat Med 2016). Methods: Potential transmission pairs were defined based on HIV-1 pol gene phylogenies and plasma neutralization activity against a 14 multi-clade pseudovirus panel and IgG1 binding activity to 13 HIV-1 Gag and Env antigens was determined. Neutralization and binding similarity within pairs was determined by Spearman correlation and a range of shuffling tests and similarity measures. Results: We identified 336 potential transmission pairs within the Swiss 4.5K Screen. For 7/14 pseudoviruses, we observed a significant (p<0.05) positive within-pair correlation of neutralization. The average Spearman correlation coefficient across all 14 viruses was weakly (ρ=0.10) but significantly (p shuf <0.001) positive, even after controlling for viral subtype, infection-length and ethnicity (p shuf <0.001). Similarly, binding to different HIV-1 antigens was significantly correlated within pairs (average ρ spear =0.18, p shuf <0.001). Notably, we found that two elite neutralizers (top 1% of neutralizers) formed a transmission pair and exhibited highly similar neutralization (ρ spear =0.73) and binding (ρ spear =0.29) fingerprints. To generalize this finding, we developed a systematic approach to identify pairs with similar neutralization and binding responses. Over a broad range of thresholds, this method identified significantly more candidate pairs than expected by chance (Figure) suggesting that the similarity of antibody responses in these transmission pairs is due to viral genetic factors. Conclusion: Our results indicate that viral genetic factors significantly affect the breadth and specificity of the antibody responses to HIV-1. A large proportion of this effect is likely due to the env gene. Not all bnAb Envs may carry the capacity to imprint identical Ab specificity, thus utilizing Envs from bnAb inducers with proven transferability of antibody reactivities as identified here may be the ultimate immunogen candidates to base vaccine design on.

Oral Abstracts

68 INTENSITY OF SUPPRESSION LINKED WITH SHIFTING COMPARTMENTALIZATION OF CNS HIV DNA Benjamin B. Gelman 1 , Joshua G. Lisinicchia 1 , Lalita D. Singh 1 , Vipulkumar N. Patel 1 , Alexander J. Gill 2 , Colleen E. Kovacsics 2 , Dennis L. Kolson 2 1 Univ of Texas Med Branch, Galveston, TX, USA, 2 Univ of Pennsylvania, Philadelphia, PA, USA Background: There is substantial information regarding the role of viral suppression on the latent pool of HIV DNA in blood cells, but analogous types of data pertaining to the deep body compartments such as the central nervous system (CNS) remain scarce. Methods: In 16 HIV-infected patients already known to have relatively low viral replication rates in the brain, we mapped total and integrated gag-pol HIV DNA and RNA in brain frontal white matter, frontal neocortex and neostriatum using PCR and the two step Alu-gag assay. Patterns of brain compartmentalization were analyzed in patients with varying levels of replication control. Results: Mean and modal averages of HIV DNA was highest in white matter versus the two gray matter sectors (p < 0.04, two-tailed t test), but CNS compartmentalization patterns deviated widely. 7 of 16 brains were reflective of the overall average and had highest concentrations in white matter; 6 brains had generally equivalent concentrations in white versus gray matter; 3 brains had higher concentrations in a gray matter compartment versus white matter. The white matter compartment contains the most mass, so the total sizes of the latent pools in the brain were nearly always highest in white matter. To determine whether the intensity of viral suppression could affect compartmentalization in the brain, patients were grouped according to those with intensely suppressed brain replication (log 1.52 +/- 0.42 copies of HIV RNA per gram, n = 8) versus less intensely suppressed (log 2.95 +/- 0.61 c/g, n = 8). The more intensely suppressed brain specimens contained significantly less HIV DNA in neocortex, but significantly more in white matter (p < 0.02), without a significant net gain or loss in total brain pool size. Precisely the same outcome was observed when the patients were grouped according to the intensity of systemic viral suppression, using spleen HIV RNA concentration as a validated postmortem surrogate for the plasma HIV replication rate (log 2.22 +/- 0.79 c/g versus 5.47 +/- 1.50 c/g). Conclusion: Results frommapping latent HIV DNA in human brain specimens imply that more intense viral suppression, both within the CNS compartment and systemically, will produce a shift in the brain burden from gray to white matter, but will not diminish the total brain pool size substantially.

26

CROI 2017