Investigating the potential of circulating microRNAs as non-invasive biomarkers in cirrhosis and hepatocellular carcinoma

Abstract

Morbidity due to chronic liver disease (CLD) is determined by fibrosis, which is graded from F0 (no fibrosis) to F4 (cirrhosis). Progression to cirrhosis is associated with the development of portal hypertension, end-organ hepatic dysfunction and hepatocellular carcinoma (HCC). Decompensated cirrhosis is characterised by cirrhosis-related complications (e.g. ascites, varices, variceal bleeding) and prognosis can be stratified according to these features of decompensation.Liver biopsy is the gold standard, but invasive diagnostic test in CLD. Non-invasive imaging or biomarker panels have been developed to stage fibrosis and to identify patients at greatest risk of cirrhosis-related complications. The current HCC biomarker, alpha-fetoprotein (AFP), however, has poor diagnostic performance in early HCC and, in common with other non-invasive tools, provides no information about pathogenesis or therapeutic targets.MiRNAs are small, stable, quantifiable non-coding RNAs that regulate post-transcriptional events and cellular processes. Preliminary data describing differential miRNA expression in CLD and HCC exist but whether miRNAs could be developed as useful non-invasive biomarkers to assess disease severity and aid in improving patient morbidity and mortality in cirrhosis remains unknown.This thesis aimed to explore:i)nnnnnnnnnnnnnnnn differential circulating miRNA expression associated with CLD of different severities (including cirrhosis, decompensated cirrhosis and HCC);ii)nnnnnnnnnnnnnn whether miRNA panels can be used alone or in combination with current non-invasive biomarkers to discriminate CLD of different severities; andiii)nnnnnnnnnnnnn the functional roles and purported target genes of miRNAs differentially expressed during the HCC development.The most significant results of this thesis include:i)nnnnnnnnnnnnnnnn The identification of 5 miRNA candidates (miRNA-122-5p, miRNA-142-3p, miRNA-151a-5p, miRNA-409-3p and miRNA-486-5p), differentially expressed according to disease severity in serum from 60 chronic Hepatitis C patients analysed in 3 equal cohorts of mild disease (F0-2), cirrhosis (F4) and cirrhosis complicated by HCC. Stepwise logistic regression assembled a miRNA panel for cirrhosis (miRNA-409-3p + miRNA-122-5p) which, when compared with F0-2, resulted in a ROC curve with an AUC=0.80 (95% CI 0.66=0.95, pl0.001). A miRNA panel for HCC (miRNA-122-5p, miRNA-142-3p + miRNA-486-5p) was also assembled which, when compared with F4, resulted in an AUC=0.94 (95% CI 0.87n1.00, pl0.001). Combining the cirrhosis miRNA panel with currently-available non-invasive fibrosis indices, APRI and FIB-4, improved its diagnostic performance with AUCs of 0.87 (pl0.001) and 0.89 (pl0.001), respectively. Combining the HCC miRNA panel with the HCC biomarker AFP did not improve its diagnostic performance.ii)nnnnnnnnnnnnnn Following transfection of immortalised human hepatocyte (IHH) and human hepatoma (Huh7) cell lines with miRNA-151a-5p and miRNA-486-5p, pull-down assays identified:mnnnnnn For miRNA-151a-5p: 2320 (IHH) and 1041 (Huh7) putative target genesmnnnnnn For miRNA-486-5p: 2016 (IHH) and 468 (Huh7) putative target genesPutative target pathways identified via IPA included the EGF, ERK/MAPK, p53 and Molecular Mechanisms of Cancer pathways (for miRNA-151a-5p) and ERK5, Hif1a and Molecular Mechanisms of Cancer pathways (for miRNA-486-5p). Transient transfection of these miRNAs into normal hepatocyte or liver tumour cell lines had no demonstrable impact on cell proliferation or migration in vitro.iii)nnnnnnnnnnnnn Identification of a single miRNA candidate (miRNA-451a), differentially expressed (AUC=0.66, 95% CI 0.54-0.78; p=0.009) in the serum of 25 cirrhotic patients with early HCC compared with 74 cirrhotic patients without HCC. Significant differential expression remained when all 37 patients with HCC (including those with indeterminate lesions that were subsequently diagnosed as HCC) were compared with those without HCC over clinical follow-up (AUC=0.63, 95% CI 0.51n0.74; p=0.018). These miRNA performances improved when combined with AFP (AUCs of 0.68 and 0.69, respectively).iv)nnnnnnnnnnnn Identification of 5 miRNA candidates (miRNA-10a-5p, miRNA-10b-5p, miRNA101-3p, miRNA 22-3p and miRNA-27b-3p) in the serum of 136 patients with cirrhosis that are differentially expressed according to disease severity in compensated cirrhosis (n=30), cirrhosis with varices (n=30), cirrhosis with ascites p varices (n=60), and cirrhosis with variceal bleeding (n=16). MiRNA panels were assembled to distinguish patients with specific complications with the following diagnostic performance:mnnnnnn Varices panel (miRNA-10b-5p + miRNA-101-3p) with an AUC of 0.78 (95% CI 0.66n0.90, pl0.001)mnnnnnn Ascites panel (miRNA-10a-5p + miRNA-146a-5p) with an AUC of 0.83 (95% CI 0.75n0.92, pl0.001)mnnnnnn Variceal bleeding panel (miRNA-10a-5p) with an AUC of 0.80 (95% CI 0.65n0.95, pl0.001)mnnnnnn Decompensation panel, grouping ascites and variceal bleeding patients (miRNA-10a-5p, miRNA-146a-5p + miRNA-423-3p) with an AUC of 0.78 (95% CI 0.70n0.86, pl0.001).Combining platelet count l 150 (x109/L)2 with the miRNA panels improved the performance of the Varices panel and the Ascites panel (with AUCs of 0.84 and AUC of 0.87, respectively).This thesis has identified several differentially-expressed miRNA candidates across a clinically relevant spectrum of CLD and assembled promising panels of miRNAs that warrant further investigation for their diagnostic potential. Preliminary efforts to explore the possible functional roles of 2 candidates in HCC were unrewarding but further studies can build on these findings by optimising experimental conditions, validating identified miRNA candidates, and testing the prognostic potential of validated miRNAs as clinically-useful non-invasive biomarkers.n