CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

associated with the increased inflammatory milieu we observed. Importantly, we did not observe an enrichment for Proteobacteria. Conclusion: Our findings suggest that protease inhibitors contribute modestly to microbial dysbiosis and immune dysfunction in uninfected lentiviral infections. As such, the side effects of protease-inhibitors commonly observed in HIV-infected individuals are unlikely to be attributed solely to GI tract dysbiosis or inflammation. Further research is required to determine if other ARVs interfere with intestinal stasis and whether ARVs contribute to dysbiosis in the context of ongoing lentiviral infections. 234 THE MICROBIOME MAY MODIFY HIV INFECTION RISK ASSOCIATED WITH HORMONAL CONTRACEPTIVES Laura Noël-Romas 1 , Michelle Perner 1 , Rafilwe Molatlhegi 2 , Sarah Mutch 1 , Kenzie Birse 1 , Renee Heffron 3 , Lyle McKinnon 1 , Adam Burgener 4 1 University of Manitoba, Winnipeg, MB, Canada, 2 University of KwaZulu-Natal, Durban, South Africa, 3 University of Washington, Seattle, WA, USA, 4 Public Health Agency of Canada, Winnipeg, MB, Canada Background: The injectable hormonal contraceptive depot medroxyprogesterone acetate (DMPA) has been associated with increased risk of HIV-1 acquisition in women, but these observations have been inconsistent. We examined whether the vaginal microbiome influences rates of HIV acquisition in women using different hormonal contraceptives in the CAPRISA 004 trial at study enrollment. Methods: Mass spectrometry was used to characterize the bacterial metaproteome (microbiome) from cervicovaginal lavage samples collected from study participants. Results: Among the 685 women included in this study, the majority were using hormonal contraceptives (97.7%) including DMPA (65.1%), norethisterone enanthate (NET-EN) (18.0%), and combined oral contraceptives (COC) (14.1%), and the majority did not switch contraceptives during the study (91.7%). Women belonged to two major vaginal microbiome profiles which were similarly distributed across hormonal contraceptive groups- one dominated by Lactobacillus (59.2%) and the other that was non-Lactobacillus dominant (microbial dysbiosis), where Gardnerella vaginalis predominated with other anaerobic bacteria (40.8%). Rates of HIV infection were trending higher in those using DMPA when compared to NET-EN and COC users as a single group, but this was not statistically significant (6.58 vs 4.15 infections per 100 women-years, respectively; adj. HR: 1.80, 95% CI: 0.90 to 3.59, P=0.097). In women with microbial dysbiosis, rates of HIV acquisition were similar between hormonal contraceptive types (7.13, 7.72, and 6.59 per 100 women-years in DMPA, NET-EN, and COC users, respectively), and not significantly higher in those using DMPA compared to all other hormonal contraceptives (HR: 1.16, 95% CI: 0.56 to 2.40, P=0.70). However, in Lactobacillus-dominant women, DMPA use associated with an infection rate of 6.23 per 100 women-years compared to 1.74 and 2.15 per 100 women-years with NET-EN and COC, respectively – a > 3-fold increase for DMPA users relative to women using other hormonal contraceptives (HR: 3.39; CI: 1.61 to 7.15, P=0.0152). These observations were consistent in models adjusted for study arm, study site, age, sexual behavior and other clinical variables. Conclusion: This suggests that the association between DMPA and HIV acquisition risk may depend on the composition of the microbiome, which may have important implications for safe contraceptive design and interpretation of future studies of contraceptives and HIV acquisition risk. 235 ROLE OF FREM1 IN PRO-INFLAMMATORY RESPONSES DURING VAGINAL HIV/SIV INFECTION Robert W. Omange 1 , Annelie Tjernlund 2 , Ma Luo 3 , Joshua Kimani 4 , Makobu Kimani 5 , Kristina Broliden 2 , Qingsheng Li 6 , Francis Plummer 1 , Zhe Yuan 6 , Xin Y. Yuan 1 , Mohammad Kashem 1 , Hongzhao Li 1 , Guobin Kang 6 , Lewis R. Liu 1 , Ben Liang 1 1 University of Manitoba, Winnipeg, MB, Canada, 2 Karolinska Institute, Stockholm, Sweden, 3 National Microbiology Laboratory, Winnipeg, MB, Canada, 4 University of Nairobi, Nairobi, Kenya, 5 KEMRI Wellcome Trust Rsr Prog, Kilifi, Kenya, 6 University of Nebraska–Lincoln, Lincoln, NE, USA Background: A single nucleotide polymorphism in FRAS1-related extracellular matrix 1 (FREM1) is associated with resistance to HIV. A splice variant of FREM1- Toll/Interleukin 1-like receptor regulator, (TILRR)- is an IL-1R1 co-receptor capable of potentiating inflammatory responses. This study investigated

the role for FREM1 in modulation of immune responses during vaginal HIV transmission. Methods: FREM1 protein expression was examined in human and Rhesus macaque (RM) female genital tissues, and changes in its expression measured following intravaginal SIV infection in RMs. FREM1 expression in both human and RM female genital tracts (FGTs) was similar, with high expression in the epithelium and submucosa Results: FREM1 levels increased following intravaginal SIVmac251 infection, accompanied by infiltration of SIV target cells into the genital mucosa. Different human immune cells in blood, expressed FREM1, including T cells, monocytes, and B cells to varying degrees. Notably, FREM1-expressing CD4+ and CD8+ T cells fromwomen with the protective FREM1 allele had lower cellular activation. Only Escherichia coli LPS (TLR4 agonist), and not Imiquimod (TLR7 agonist) or ssRNA40 (TLR8 agonist) alters FREM1 expression on some T cells and monocyte subsets. Co-expression analysis of FREM1 and TLR4 in PBMCs and tissues also suggests close association between these proteins. Stimulation of human monocyte populations with a TLR4 agonist or antagonist, alone and in combination with anti-FREM1 mAbs, indicates that FREM1 modulates pro- inflammatory cytokine production and co-stimulatory factor expression. Conclusion: These results suggest FREM1 potentially regulates innate immune responses, based on its association with TLR4. These findings add to the understanding of early HIV transmission in the context of cellular structural proteins being influenced vaginal microbiota driven inflammation.  236 CERVICOVAGINAL MICROBIAL STRAINS ARE ASSOCIATED WITH DISTINCT IMMUNOPHENOTYPES Matthew R. Hayward 1 , Seth M. Bloom 1 , Nomfuneko A. Mafunda 1 , Jiawu Xu 1 , Brittany A. Bowman 1 , Christina Gosmann 1 , Bjorn Corleis 1 , Mara Farcasanu 1 , Justin K. Rice 1 , Curtis Huttenhower 2 , Caroline Mitchell 1 , Douglas Kwon 1 1 Ragon Institute of MGH, MIT and Harvard, Cambridge, MA, USA, 2 Broad Institute of MIT and Harvard, Cambridge, MA, USA Background: Elevated inflammation in the female genital tract (FGT) is associated with an increased risk of HIV infection, and cervicovaginal bacteria have been shown to impact genital inflammation (Gosmann et al., 2017). These associations have been identified through bacterial 16S rRNA gene sequencing which has limited resolution and rarely achieves taxonomic assignment to the species-level. Within-species genetic differences can be vast, with some species-level pangenomes (all the unique genes observed for a species) exceeding the size of any single strain’s genome by orders of magnitude. Furthermore, 16S sequencing provides no functional information, limiting our mechanistic understanding of disease associations. Methods: To better characterize strain-level variation in the FGT microbiota we generated species-specific pangenomes from single culture genome sequences (1000 primary bacterial isolates and 2000 publicly available genomes). We produced sample specific gene profiles by mapping metagenomic sequences from 300 North America and South Africa women to the species-specific gene catalogues. Profiles were partitioned using centroid based clustering to form groups containing similar gene complements (strains). Local inflammation was measured using Luminex cytokine assays performed on cervicovaginal lavages from South African women with distinct strains. Results: We show that most species possess a small core genome (~1000 genes) with an extensive pangenome (6000 to 30,000 genes). We find that Gardnerella vaginalis comprises 4 distinct strains and that some women possess enough genes to make 4 complete genomes, suggesting some women are colonised by multiple strains. Furthermore, we show that these strain complexes are associated with higher levels of cytokines in the FGT. Conclusion: Our findings signify the importance of distinguishing microbial strains in the FGT when linking the endogenous microbiome to local inflammation influencing HIV acquisition risk.

Poster Abstracts

83

CROI 2019