Circulatory microRNA-365a-3p as a potential biomarker for the diagnosis of pediatric cystic fibrosis-associated liver disease

Abstract

Introduction: In Australia, cystic fibrosis (CF) affects one out of 2500 newborns, of whom 15-20% develop severe liver abnormalities and 5% die due to end-stage liver disease. CF is an autosomal recessive disorder that affects the CF transmembrane regulator channel expressed in cholangiocytes and the gall bladder. As a consequence of decreased bile flow, thickened secretions block the bile ducts and lead to hepatic and cholangiocyte injury, characteristics of CF-associated liver disease (CFLD). CFLD is one of the leading non-respiratory causes of morbidity and mortality in children with CF. Liver biopsy, an invasive procedure, remains the gold standard to assess the severity of liver damage in CFLD before the advent of cirrhosis and portal hypertension. We investigated circulatory microRNAs (miRNAs) as non-invasive biomarkers to diagnose and assess disease progression of pediatric CFLD. Methods: Our study assessed the circulating miRNA signature of 90 children allocated into three study cohorts based on clinical, biochemical, and imaging assessment, as follows: healthy controls (controls, n = 30); CF patients with no evidence of liver disease (CFnoLD, n = 30); and CFLD (n = 30; sub-divided according to liver biopsy fibrosis stages: F0, F1-2, and F3-4; n = 10 each). Serum miRNAs were analyzed using the Ion Torrent sequencing platform. Sequencing libraries were created by pooling 10 samples per group. Significant differentially expressed miRNAs were detected using normalized read counts and pair-wise comparisons. Selected miRNAs were further validated in 124 individual samples by quantitative real-time polymerase chain reaction and analyzed using ANOVA (P < 0.05). Results: We identified let-7 g-5p, miR-34a-5p, miR-122-5p, miR-365a-3p, miR-18a-5p, miR-126-5p, and miR-142-3p to be significant differentially expressed between all three study groups. Post hoc analysis showed the upregulation of miR-34a-5p (P = 0.0042), miR-122-5p (P = 0.0015), and miR-365a-3p (P = 0.0002), while let-7 g-5p (P = 0.0034) and miR- 142-3p (P = 0.0047) were downregulated in CFLD patients compared with CFnoLD patients. ROC curve analysis was used to determine the ability of these miRNAs to differentially diagnose liver disease in children with CF. miR-365a-3p showed the best predictive value, with an area under the curve (AUC) of 0.745 (P = 0.0001; sensitivity, 70.4%; specificity, 67.5%). Additionally, let-7 g-5p with an AUC of 0.708 (P = 0.0010; sensitivity, 70.4%; specificity, 65%), miR-34a-5p with an AUC of 0.707 (P = 0.0014; sensitivity, 71.4%; specificity, 64.1%), and miR-122-5p with an AUC of 0.708 (P = 0.0010; sensitivity, 70.4%; specificity, 62.5%) were also capable of differentiating between CFLD and CFnoLD. In contrast, miR-18a-5p was significantly upregulated in both CFnoLD and CFLD compared with controls (P < 0.0001), with a predictive value of AUC = 0.822 (P < 0.0001; sensitivity, 80%; specificity, 80%), suggesting an important role in CF. Conclusion: Although diagnostic panels using combinations of these miRNAs require validation in larger cohorts, our study has identified several serum miRNAs with potential to diagnose liver disease in children with CF.L6180060732017-08-31 <br />