Circulating microRNA signatures that predict liver fibrosis progression in HIV-1/HCV co-infected patients

The lack of available biomarkers to diagnose and predict different stages of liver disease with a non-invasive strategy is currently one of the main challenges that clinicians are facing. Recent evidence indicates that the plasma levels of specific microRNAs (miRNAs) may be significantly altered in patients with liver injury, including human immunodeficiency virus type 1 (HIV-1) infected patients. Large-scale deep sequencing analysis of small RNA expression was performed on plasma samples from 46 HIV-1/hepatitis C virus (HCV) co-infected patients that did not exhibit liver fibrosis at the time of sampling. A total of 1065 different miRNAs were identified. After a mean of 10.3 years, 26 of the former patients developed liver fibrosis (stage F2-4) and 20 remained without signs of liver fibrosis (stage F0-1). We identified a signature of seven miRNAs, 100-5p, 192-5p, 99a-5p, 122-5p, 125b-2-3p, 1246 and 194-5p, that highly correlated with patients progressing to liver fibrosis. These seven miRNAs detected liver fibrosis progression with an area under curve (AUC) of 0.910-0.806. The two miRNAs, 100-5p and 192-5p, displaying the best AUC values, yielded both a sensitivity of 88% and a specificity of 85% for liver fibrosis progression. Our results demonstrate the predictive potential of circulating levels of miRNAs to foresee liver fibrosis progression even before liver fibrosis or significant clinical differences such as liver transaminases or platelets are detectable. Thus, our study might help in predicting the progression of liver injury in HIV-1-infected patients. In total 67 plasma derived small RNA samples were analyzed: 21 from healthy reference and 46 from virally co-infected individuals with different liver fibrosis progression status. There were two patient groups based on fibrotic progression levels among co-infected individuals: 20 non-progressors and 26 progressors. Differential miRNA expression was assesed between the different groups under study. All samples were controlled for hemolysis using a qPCR based test taking only samples with miR-451/miR-23a ratios with ΔCt<7 (range −0.04 to 6.3; average 3.53). Design efforts were made to randomize all subgroups to be equally represented in the library preparation and size selection steps to be able to correct for possible batch effects. Size selection was performed in pools of 24 libraries, and 3 sets of 2 library pools were merged into 3 superpools of 48 samples each for Illumina sequencing. The experiment was performed in the context of a larger sample set including a total of 144 samples (a subset described in a previous study by Franco et al, Antiviral Res 2018 Jul;155:106-114. PMID: 29807039, GEO GSE141522 which included the same healthy control samples).The final results of this work are from a different subsetusing a different analysis pipeline.