CROI 2015 Program and Abstracts

Abstract Listing

Poster Abstracts

684 Innate Immune Activation Pathways Overlap, Yet Are Distinct in HCV and HIV Infection Lenche Kostadinova ; Benigno Rodriguez; Donald Anthony Case Western Reserve University, Cleveland, OH, US

Background: While chronic innate and adaptive immune activation are well recognized during HCV and HIV infection, drivers of this process are less clear, and whether the same drivers operate in each infection is unknown. Our data led us to sCD14 and LGALS3BP (90-kDa Mac-binding protein) as potential markers/mediators of immune activation in HCV-HIV infection. Methods: We measured plasma sCD14, LGALS3BP, IL-6, LBP (LPS-binding protein) and IFABP (intestinal fatty acid binding protein) levels by ELISA in uninfected controls (n=12), HCV infected (n=30), HCV-HIV infected (n=30), and HIV infected (n=30) subjects not on therapy for HCV or HIV. HCV and HCV-HIV infected groups were stratified into low vs. hi APRI subgroups. HIV infected groups were evaluated again at 6 months on ART to determine the contribution of active HIV replication. Results: In the absence of antiviral therapy HIV infection was associated with elevated plasma sCD14 (p=0.025), IL-6 (p=0.010), IFABP (p=0.05) and LGALS3BP (p<0.0001), while HCV infection was associated with elevated IL-6 (p=0.022) and LGALS3BP (p<0.0001). The latter remained significant in the hi-APRI subgroup for IL-6, and both hi and low APRI subgroups for LGALS3BP. Additionally, sCD14 (p=0.019) and IFABP (p=0.05) levels were greater in HCV hi APRI subjects than uninfected controls. Furthermore, both sCD14 (p<0.0001 and p=0.025) and LGALS3BP (p=0.030 and p<0.0001) levels were greater in HCV-HIV co-infected subjects than both HCV and HIV mono-infected subjects, consistent with contributions from each infection. Furthermore, while LBP was not significantly elevated in either HCV or HIV mono-infection, it was elevated during HCV-HIV co-infection (p=0.005). When evaluating relationships between each marker we observed that sCD14 levels positively correlated with LGALS3BP (p=0.031), IL-6 (p= 0.05) and LBP (p=0.022) during HCV infection, while sCD14 only correlated with LBP (p=0.024) during HIV infection. During HCV-HIV co-infection relationships between sCD14, LBP, and LGALS3BP were similar to those during HCV mono-infection. After initiation of ART for HIV infection, IL-6, LGALS3BP and LBP levels in HIV infected subjects normalized to those of uninfected controls, and those in HCV-HIV infected subjects declined to levels observed in HCV mono infected subjects. Conclusions: Both HCV infection and consequent liver damage likely drive innate monocyte/macrophage activation by pathways that are overlapping but distinct from HIV infection. 685 A Novel Mechanism of Resistance to Multiple bNAbs Revealed by Natural Variation in Panel of 199 HCVpp Ramy El-Diwany ; LisaWasilewski; Madeleine Mankowski; Stuart C. Ray; Justin R. Bailey Johns Hopkins University School of Medicine, Baltimore, MD, US Background: Broadly-neutralizing monoclonal antibodies (bNAbs) are informative for vaccine development for highly variable viruses including hepatitis C virus (HCV). However, HCV resistance to bNAbs is poorly understood, in part due to a lack of neutralization testing using diverse panels of HCV variants. HC33.4 and AR4A are two of the most potent bNAbs characterized to date, targeting the amino- and carboxy-terminus of HCV envelope (E2) respectively. Methods: For each bNAb, we measured neutralization of a library of 199 genotype 1 HCV E1E2-pseudotyped particles (HCVpp). Using a novel R script, polymorphisms at each amino acid position in E1E2 were assessed for impact on neutralization resistance using Wilcoxon rank-sum test (corrected for multiple comparisons). Validation was performed using site directed mutagenesis in multiple E1E2 backgrounds. bNAb affinity for mutations in multiple E1E2 backgrounds was determined by ELISA. Results: Multiple E1E2 clones with resistance to HC33.4 and/or AR4A were identified. Surprisingly, most resistance could not be attributed to amino acid changes in known mAb binding epitopes. Subtype 1a clones were more sensitive to mAb HC33.4 than subtype 1b clones (P<1e-5) and polymorphisms at positions 359, 403, 408, 414, 473, and 641 in subtype 1a clones were associated with resistance (P corrected <1e-5 each), even though only 414 falls within the known binding epitope. Polymorphisms at 235, 293, 395, 403, and 655 were associated with AR4A resistance (P corrected <1e-5 each), even though only 655 falls near AR4A’s predicted binding epitope. Site directed mutagenesis of K408 to Methionine and L403 to Phenyalanine, in multiple E1E2 backgrounds conferred a 3.0-12.3-fold increase in resistance to HC33.4 (P<0.01). Interestingly, introduction of L403F also conferred 2.5-5.3- fold increase in resistance to AR4A (P<0.01 for each). Unlike previously characterized resistance polymorphisms, L403F did not reduce E1E2 binding affinity of either bNAb. Conclusions: Our large diverse HCVpp panel in conjunction with a comprehensive computational pipeline allows for measurement of neutralizing antibody breadth, detection of differences in HCV subtype neutralization sensitivity, and identification of naturally occurring bNAb resistance polymorphisms in E1E2. A commonly occurring polymorphism at 403 confers resistance to two potent broadly neutralizing human mAbs targeting opposite ends of E2 without affecting binding affinity, suggesting a novel mechanism of neutralization resistance in HCV. 686 Single-Variant Sequencing Revealed Rapid HCV Evolution in HIV Immune Reconstitution Background: Human immunodeficiency virus (HIV) infection accelerates the progression of liver disease in Hepatitis C virus (HCV) co-infected individuals, raising the question of how immune status in co-infected individuals affects HCV quasispecies populations. We hypothesize that the structure and evolution of HCV populations in individuals with immune reconstitution differ from that in those with immune decline in HCV-HIV co-infections. Methods: To precisely define and quantify individual HCV variants, we developed a single-variant sequencing (SVS) approach by combining random nucleotide tagging with an optimized Illumina MiSeq deep sequencing strategy. We applied this method to examine HCV quasispecies structure and evolution in envelope and NS3 genes in HCV-HIV co- infected subjects, and assess the abundance and dynamics of naturally occurring drug resistant mutations (DRMs) in NS3. Results: We sequenced 35 serum samples collected across 10 years of follow-up from 6 HCV-HIV co-infected subjects. Three subjects who received ART early during follow-up experienced immune reconstitution. The other three subjects showed a gradual decline in CD4 count with delayed ART. A total of >16 million paired-end reads were generated to build 92,226 SVS consensus sequences based on unique sequence tags. Each consensus sequence represents an individual HCV variant. We showed that the SVS approach (1) corrected PCR/sequencing errors as well as resampling and amplification bias from PCR, and (2) provided an accurate representation of HCV population structure in vivo . Rapid non- synonymous sequence evolution was observed in hypervariable region 1 (HVR1) of E2, but not NS3, in subjects with immune reconstitution compared to subjects with immune decline. Notably, using SVS consensus, a naturally occurring DRM in NS3, aaVI170T, was detected as a minor variant (<1%) in 4 of 6 subjects over multiple visits. In contrast, an even distribution of minor DRMs across all DRM sites was observed based on raw reads (P<10 -32 ). Conclusions: Single-Variant Sequencing accurately defines HCV quasispecies population structure and evolution in vivo . Immune reconstitution in HIV/HCV co-infection is associated with rapid sequence evolution in the envelope but not NS3, likely driven by increased immune pressure from enhanced humoral responses during antiretroviral therapy. Drug resistant mutations in NS3 can occur naturally at low frequencies in HCV-HIV co-infected individuals, which can be accurately identified and quantified using SVS. Lin Liu ; David Nardo; Eric Li; GaryWang University of Florida, Gainesville, FL, US

Poster Abstracts

425

CROI 2015