CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

256 ANTIBIOTICS DISRUPT COLONIC MICROBIOTA AND MUCOSAL IMMUNE FUNTION IN RHESUS MACAQUES Alexander S. Zevin 1 , Ryan Cheu 1 , Tiffany Hensley-McBain 1 , Jacob Modesitt 1 , Jennifer A. Manuzak 1 , Charlene J. Miller 1 , Ernesto Coronado 1 , Elise Smith 1 , Brian Richardson 2 , Michael Gale 1 , Mark Cameron 2 , Jeremy Smedley 3 , Nichole Klatt 1 1 University of Washington, Seattle, WA, USA, 2 Case Western Reserve University, Cleveland, OH, USA, 3 Washington National Primate Research Center, Seattle, WA, USA Background: Antibiotics are widely used throughout the world to treat bacterial infections that occur independently or as a result of HIV infection. However, evidence suggests that antibiotic treatments disrupt the composition of the intestinal microbiome. Further, HIV infection is associated with disrupted microbiota and mucosal dysfunction independent of antibiotic use. Intestinal microbiota are critical for maintaining host immune homeostasis and protecting against the expansion of pathobionts. Additionally, microbiota-derived metabolites are key energy sources for colonic epithelial cells and regulate immunity. We hypothesized that antibiotic therapies would disrupt the GI microbiome and host mucosal immunity. Methods: We administered antibiotics to four groups of healthy female rhesus macaques and collected GI biopsies and stool (Group 1 – enrofloxacin, Group 2 – cephalexin, Group 3 – paramomycin, Group 4 – clindamycin). We tracked bacterial community composition using 16S rRNA gene sequencing and qPCR, evaluated host mucosal immunity in the colon throughout the antibiotic treatment using multicolor flow cytometry, and host gene expression using mRNA-seq. Finally, we quantified plasma biomarkers of mucosal disruption and microbial translocation using ELISA. Results: The antibiotic treatments disrupted colonic microbiota and led to expansion of facultative anaerobic and potentially pathogenic bacterial taxa as well as quantitative shifts in the bacterial abundance in the colonic mucosa and stool. There also was a significant increase in colonic mucosal neutrophils during the treatment. There also were increased frequencies of Th17 and Th22 CD4+ T-cells after the antibiotic treatment. We further found altered expression of genes involved in cell-cell junctions, cellular metabolism, and inflammatory pathways during and after the antibiotic treatment. Finally, we showed that plasma sCD14 concentrations increased during the antibiotic treatment. Conclusion: Our data demonstrate that antibiotic therapies alter colonic microbiota, and that these changes were linked to a distinct signature of mucosal inflammation and immune activation, including increased populations of HIV-1 target cells. Thus, in HIV infection, antibiotic use may exacerbate mucosal dysfunction, and resolving bacterial dysbiosis by limiting antibiotic use may be key to maintaining mucosal health. 257 EFFECT OF PROBIOTICS ON GUT MICROBIOTA AND PET/MRI ACTIVITY IN HIV-INFECTED PERSONS Caroline Arnbjerg 1 , Beate Vestad 2 , Johannes Hov 3 , Karin K. Pedersen 1 , Sofie Jespersen 1 , Helle H. Johannesen 1 , Kristian Holm 2 , Eva Fallentin 1 , Adam E. Hansen 1 , Theis Lange 4 , Andreas Kjær 1 , Marius Troseid 2 , Barbara M. Fischer 1 , Susanne D. Nielsen 1 1 Copenhagen University Hospital, Copenhagen, Denmark, 2 Oslo University Hospital, Oslo, Norway, 3 University of Oslo, Oslo, Norway, 4 University of Copenhagen, Copenhagen, Denmark Background: Despite combination antiretroviral therapy (cART) with suppressed viral replication, gut microbiota alterations, microbial translocation and low-grade inflammation persist in several HIV-infected individuals. The aim of this study was to investigate the effects of the probiotic strain Lactobacillus rhamnosus GG (LGG) on the gut microbiota composition, microbial translocation and intestinal inflammation. Methods: The gut microbiota composition (Illumina sequencing of 16s rRNA) and level of microbial translocation (lipopolysaccharide, LPS) were determined at baseline and after probiotic intervention in 45 individuals (15 cART näive and 30 cART-treated). All study participants received 6 x 109 colony-forming units of LGG twice daily for eight weeks. Gut microbial profiles were compared with measurements of intestinal inflammation by 18F-2-fluoro-2-deoxy-D-glucose positron emission tomography/magnetic resonance imaging (18FDG PET/MRI) scans in 15 individuals. Results: After probiotic intervention, no overall change in microbial translocation (LPS) or microbial alpha diversity could be detected in the study group (observed operational taxonomic units (OTUs); mean change -0.83, 95% CI -2.33 to 0.67, p=0.328, phylogenetic diversity (PD) whole tree;

22 CHI and 18 uninfected controls. A subgroup of EHI and CHI participants on ART was prospectively assessed. Plasma βDG and galactomannan levels were quantified using Fungitell® and Platelia™ assays, respectively. Plasma βDG levels were compared with age, sex, viral load, CD4 and CD8 T-cell counts, CD4/ CD8 ratio, markers of gut damage (I-FABP), microbial translocation (LPS, LBP and sCD14), inflammation (IL1-β, IL-6, IL-8, TNF-α and sCD40L) and Kynurenine/ Tryptophan ratio. Univariate and multivariate analyses were conducted at 5% α. Results: Plasma βDG levels were elevated during EHI (60.5±33.2 pg/mL, p=0.012) and CHI (135.6±48.6 pg/mL, p<0.001) in comparison to controls (30.4±5.3 pg/mL), while galactomannan levels were below level of detection in all participants. βDG levels increased over two year interval in the untreated EHI (100.6±81.2 pg/mL p=0.012) and remained stable in the ART-treated EHI. CHI on 12±4 years of ART had the highest βDG levels (190.7±68.7 pg/mL, p<0.001). A correlation of βDG was observed with viral load (r=0.430; p<0.001) in untreated participants, CD4 T-cell count (r=-0.334; p<0.001), CD4/CD8 ratio (r=-0.286; p=0.003), LBP (r=0.413; p=0.007), sCD14 (r=0.338; p=0.001), IL-6 (r=0.334; p<0.001), IL-8 (r=0.506; p<0.001), sCD40L (r=0.333; p=0.024) and Kyn/Trp (r=0.261; p=0.022). Multivariate analysis showed elevated βDG levels during HIV infection were independent of the effect of age, sex, creatinine, total cholesterol and glucose. Conclusion: Plasma βDG levels were elevated during early and chronic HIV infection and did not decrease on ART. Elevated βDG levels correlated with the validated markers of microbial translocation and inflammation and can be considered as a marker of HIV disease progression. 255 TRANSFER OF MICROBIOTA AND IMMUNOPHENOTYPES IN HIV-/+ MSM TO MICE BY FECAL TRANSPLANT Sam Li , Sharon Sen, Jennifer M. Schneider, Nichole Nusbacher, Nancy Moreno- Huizar, Catherine Lozupone, Brent E. Palmer University of Colorado, Aurora, CO, USA Background: High-risk men who have sex with men (MSM) have differences in immune activation and gut microbiome composition when compared with men who have sex with women (MSW), even in the absence of HIV infection. Gut microbiome differences that occur with HIV as assessed by 16S rRNA sequencing are relatively subtle. Understanding whether an altered gut microbiome composition drives increased immune activation in MSM and with HIV infection has important implications since immune activation has been associated with both HIV acquisition risk and disease progression. Methods: Gnotobiotic mice were gavaged with feces from HIV- MSW (n=11), HIV- MSM (n=14), and HIV+ untreated MSM (n=9). At 21 days post-gavage, CD69 and CD103 expression was evaluated on gut T cells, and fecal microbiome composition in the mice was characterized by 16S rRNA gene sequencing. HLADR and CD38 expression on blood T cells and fecal microbiome composition in the human donors were also analyzed and compared with mouse recipients. Results: We observed increased gut immune activation in gnotobiotic mouse recipients of HIV- MSM feces compared with mouse recipients of HIV- MSW feces. Additionally, several bacterial species significantly correlated with both blood HLADR+ CD38+ CD8+ T cells in human donors (which was significantly higher in MSM, p=0.01) and gut CD69+ CD8+ T cells in mouse recipients. When controlling for MSM, we found no effect of HIV status on gut immune activation or microbiome composition in mouse recipients. Gnotobiotic mice presented some challenges in interrogating MSM-associated gut microbiome differences. The MSM-associated gut microbiome is characterized in part by an increase in the relative abundance of bacteria in the Prevotella genus, but Prevotella did not colonize the mice. However, principal coordinates analysis (PCoA) revealed that the microbiome of MSM and MSWmaintained distinct clustering following transfer to mice that did replicate other differences between MSM and MSW observed in the human donors; mouse recipients of MSM feces had significantly lower abundances of the Lachnospiraceae family (p=0.01) and Blautia genus (p=0.02) compared with mouse recipients of MSW feces. Conclusion: Shared microbiome and immunological phenotypes between human donors and mouse recipients support gnotobiotic mice as a model for studying MSM- and HIV-associated microbiome changes. These data suggest that gut microbes influence immune activation in MSM, which may drive what is seen in HIV.

Poster Abstracts

89

CROI 2018