Associations between Circulating MicroRNAs and Measures of Insulin/Glucose Dynamics in People Living with HIV with Dysglycemia

Abstract

People living with HIV (PLWH) have near normal life expectancies with advances in antiretroviral therapy, but with the increasing average age, comorbid conditions create a new health burden for this population. Insulin resistance and type II diabetes are significantly higher in PLWH. microRNAs (miRs) are small non-coding RNAs that alter protein expression, and dysregulated expression of multiple miRs are implicated in altered insulin/glucose dynamics. There is increasing interest in using miRs as non-invasive biomarkers of pathophysiological states since most miRs are found in circulation and can be traced back to the cell type where they are either actively secreted or released during cell death. In this study, we aimed to identify circulating miRs that are associated with measures of insulin/glucose dynamics in PLWH with dysglycemia. PLWH (N=98, 68 male) with fasting plasma glucose between 95-124 mg/dL were enrolled in the parent ALIVE-Ex study provided their informed consent. Anthropometric measurements were taken along with adiponectin and an oral glucose tolerance test (OGTT) where glucose, insulin, and C-peptide were measured at 0, 1, and 2 hours. Average participant age was 52 ± 10 years, and average BMI was 27.7 ± 6.8 kg/m2. Total RNA was isolated from plasma (0h) and absolute miR expression was determined. miRs were selected based on their role in skeletal muscle (miR133a, miR-206), pancreatic β-cells (miR-375), adipose tissue (miR-let-7b), and liver (miR-20a). miR expression was correlated to measures of insulin/glucose dynamics using Spearman correlations. Multiple regression analyses were completed to control for sex, age, and BMI. Additional regression analyses combining miRs that significantly added to models (p<0.05) were used to improve predictions of insulin/glucose measures. miR-20a was positively correlated with glucose-2h (p<0.01) and this association remained significant when controlling for sex, age, and BMI. miR-375 was positively correlated with adiponectin (p<0.05); however, the association was not significant when controlling for sex, age, and BMI which were all significantly associated (p<0.01) with adiponectin. miR-206, 133a, and let-7b were not significantly correlated with any insulin/glucose measures. Combining miR-20a and miR-375 improved the prediction (R=0.42, p<0.0001) of increased glucose levels (2h) compared with miR-20a alone (R=0.30, p<0.01). In this cohort of PLWH with dysglycemia, our results indicate that only miR-375 and miR-20a significantly correlated with glucose/insulin measures. miR-375 is important for maintaining β-cell mass and was positively correlated with adiponectin, but sex, age, and BMI are stronger predictors of adiponectin. miR-20a is implicated in liver function and regulates hepatic glycogen synthesis, and the increased plasma miR-20a seen with higher plasma glucose levels could indicate hepatic dysregulation. Combining miR-20a and miR-375 led to improved predictions indicating that combining expression of multiple circulating miRs into a composite score could be utilized as an indicator of altered insulin/glucose dynamics in PLWH.