CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

Results: Median age was 36 years, CD4+ T cell count 344 cells/μL and BMI 24.5 kg/m2; 89%were male and 56% non-white. W0 median SAT and VAT density were -99 and -80 HU, respectively. Over 96 weeks, SAT and VAT HU decreased in all arms (Table). In adjusted models, female sex and higher W0 HIV-1 RNA were independently associated with greater declines in AT density (women: SAT -4.8 and VAT -4.0 HU greater than men; per log10 HIV-1 RNA copies/mL: SAT -2.3 and VAT -2.7 HU). Statistically different effects of ART type were not seen (p>0.13), though variability was high. W96 SAT and VAT HU correlated (p<0.05) positively with HDL cholesterol and adiponectin levels (r=0.19 to 0.30) and negatively with IL-6, non-HDL cholesterol, triglyceride, leptin and HOMA-IR (r=-0.23 to -0.68) even after adjusting for baseline CD4+ T cell count, HIV-1 RNA and AT area. Conclusion: VAT and SAT density decreased following ART initiation. Women and PLWH with higher HIV-1 RNA had greater decreases. Following virologic suppression, lower AT density was associated with greater systemic inflammation, lipid parameter disruption and insulin resistance independent of AT area. These findings suggest that changes in fat tissue during ART may have adverse health consequences. 682 DUAL-ENERGY X-RAY ABSORPTIOMETRY (DXA) POORLY APPROXIMATES VISCERAL FAT IN HIV Lindsay T. Fourman , Emma M. Kileel, Jane Hubbard, Tara Holmes, Ellen Anderson, Sara E. Looby, Kathleen V. Fitch, Meghan Feldpausch, Martin Torriani, Janet Lo, Takara L. Stanley, Steven K. Grinspoon Massachusetts General Hospital, Boston, MA, USA Background: People living with HIV (PLWH) are prone to visceral fat accumulation, which predisposes to comorbidities including dyslipidemia and coronary artery disease. Given the importance of visceral fat to cardiometabolic health in HIV, techniques to allow for its safe and affordable measurement are critically needed. Dual-energy x-ray absorptiometry (DXA) is an inexpensive modality that uses minimal radiation to quantify body composition. Recently, advanced software has allowed visceral fat to be ascertained from standard DXA, although this has never before been validated in HIV. Here, we investigated the accuracy of DXA in the measurement of visceral fat in comparison to computed tomography (CT) as the gold standard. Methods: We pooled data from 5 prior studies of PLWH and uninfected controls in which paired DXA and CT scans were available. For this purpose, DXA (Hologic) was re-analyzed to quantify visceral fat using APEX 6.6.0.5 software. In a cross-sectional analysis, L4-L5 visceral fat cross-sectional area (VAT) as measured by DXA and CT were compared in PLWH (n=313) and controls (n=144). In longitudinal analyses, the accuracy of DXA with respect to changes in VAT over time was assessed (1) among PLWH (n=106) and controls (n=80) on no intervention for 12 months, and (2) among PLWH on tesamorelin (n=23) – an FDA-approved medication known to reduce VAT in HIV – or placebo (n=20) for 6 months. Bland-Altman plots were used to compare DXA with CT. Results: In HIV, DXA-VAT and CT-VAT were strongly correlated (r=0.91, P<0.0001). However, the measurement bias (DXA – CT) became progressively more negative with greater VAT (P<0.0001). In this regard, whereas the bias was -9±47 cm2 overall, it was -61±58 cm^2 among those with VAT≥200 cm^2. Sex modified the inverse relationship between VAT and measurement bias (P<0.0001) such that it was particularly pronounced in men rather than women. Longitudinally, in the natural history analysis, DXA underestimated changes in VAT, irrespective of sex, with the largest bias at the extremes of VAT gain or loss (P<0.0001). DXA similarly underestimated changes in VAT among PLWH treated with either tesamorelin or placebo (P=0.004). Analogous cross-sectional and longitudinal findings were seen among uninfected controls. Conclusion: DXA underestimated VAT compared to CT in HIV-infected men with visceral fat accumulation. DXA also underestimated changes in VAT over time in both men and women with HIV. DXA-VAT should be used with caution in HIV and non-HIV alike.

683 UNIQUE MIRNA SIGNATURE IN HIV LIPODYSTROPHY WITH REDUCED ADIPOSE DICER EXPRESSION Suman Srinivasa 1 , Ruben Garcia 2 , Martin Torriani 1 , Kathleen V. Fitch 1 , Patrick Maehler 1 , Meghan Feldpausch 1 , Aaron M. Cypress 3 , C. Ronald Kahn 2 , Steven K. Grinspoon 1 1 Massachusetts General Hospital, Boston, MA, USA, 2 Joslin Diabetes Center, Boston, MA, USA, 3 NIH, Bethesda, MD, USA Background: Suppression of Dicer, an endoribonuclease that regulates microRNAs (miRNA), has evolved as a viral mechanism to enhance host HIV infectivity and may have unintended metabolic consequences. Animal knockout models of adipose-specific dicer (ADicer) acquire lipodystrophy accompanied by severe metabolic abnormalities. Data show adipose is a source of exosomal miRNAs, which function as adipokines influencing metabolic homeostasis. We hypothesized a unique miRNA profile among individuals well-phenotyped for HIV lipodystrophy and reduced ADicer expression. Methods: We evaluated >1000 miRNAs from exosomes derived from sera among the 27 male individuals [9 HIV lipodystrophy (HIV/lipo), 9 HIV without lipodystrophy (HIV/non-lipo), 9 non-HIV] whomwe previously showed variations in ADicer: most suppressed among HIV lipo, followed by HIV non-lipo and non-HIV (2.49[0.02,4.88] vs. 11.20[4.83,21.45] vs. 17.69[10.72,47.91], P=.002). To estimate miRNA abundance, data was normalized to the average expression of all measured miRNAs. Student’s T-test for 2 group comparisons and a false discovery rate analysis (FDR) was applied. Using target prediction databases (TargetScan, miRDB, Diana), we identified genes related to fat biology and lipid metabolism via a conservative approach (presence in all 3 databases + target score of >85%) with clinical relevance to lipodystrophic phenotypes. Results: HIV/lipo individuals (mean age 56±3 years, BMI 30±1 kg/m2, duration HIV 24±2 years, duration ART 20±2 years, CD4+ count 482±90 cells/μl, undetectable VL 67%) were similar to HIV/non-lipo (age 52±3 years, BMI 30±1 kg/m2) and non-HIV (age 55±3 years, BMI 30±1 kg/m2) individuals. Reduced ADicer expression was significantly related to reduced CD4+ count (r=0.55, P=.02), duration ART use (r=-0.70, P=.001) and duration PI use (r=-0.71, P=.03) and tended to be related to duration HIV (r=-0.44, P=.07) and reduced CD8+ count (r=0.42, P=.08). Accounting for the FDR, we detected miRNA- 20a-3p (P=.0026), 324-5p (P=.0059), and 186-5p (P=.0977) were expressed differentially in HIV/lipo vs. non-HIV and 324-5p(P=.0348) in HIV/lipo vs. HIV/ non-lipo. Relevant target genes per individual miRNA include: 20a-3p (EBF1, EHMT1, EZH2, NF1, PCNA, RAB4A, SPRY1, TDG), 324-5p (VDAC1), and 186-5p (MYT1L, NEGR1, NFAT5, PDE10A, PID1). Conclusion: These novel data enhance our understanding by which altered ADicer expression and specific exosomal miRNAs may affect gene expression of regulators important to fat biology and metabolic homeostasis in HIV.

Poster Abstracts

CROI 2019 261