Article Text

Download PDFPDF

MicroRNA in alcoholic hepatitis: implications for pathophysiology and treatment
  1. Felix Stickel1,
  2. Laurent Dubuquoy2
  1. 1 Department of Gastroenterology and Hepatology, University Hospital of Zurich, Zurich, Switzerland
  2. 2 LIRIC—Lille Inflammation Research International Center—U995, Univ. Lille, Inserm, CHU Lille, Lille, France
  1. Correspondence to Professor Felix Stickel, Department of Gastroenterology and Hepatology, University Hospital of Zurich, Rämistrasse 100, Zurich CH-8091, Switzerland; felix.stickel{at}, felix.stickel{at}

Statistics from

Alcoholic hepatitis alcoholic hepatitis (AH) represents a necroinflammatory complication of alcoholic liver disease, which affects a relatively small but considerable fraction of heavy drinkers with various degrees of underlying alcoholic liver damage. The true incidence of alcoholic hepatitis (AH) is unclear, but is found in around 20% of those with available liver histology, and estimated in approximately 10%–35% of alcoholics hospitalised for decompensating liver disease.1 Severe alcoholic hepatitis (AH) severe alcoholic hepatitis (sAH) has a grave prognosis with a 1-month mortality rate of up to 50%, a figure assessed for patients with a discriminant function score of ≥32 in the initial treatment trial by Maddrey and coworkers.2 Since that time, advances in intensive care medicine, response-guided treatment of selected patients with corticosteroids and improved management of infectious complications contributed to a slightly better outcome of patients with severe alcoholic hepatitis (sAH).3 Although randomised clinical trials have explored several options, effective medical treatments that improve the underlying inflammatory process are still restricted to corticosteroids to which only 50% of patients with severe alcoholic hepatitis (sAH) respond adequately.

The pathogenesis of severe alcoholic hepatitis (sAH) involves cytokine-mediated necroinflammation of the liver parenchyma, consecutive hepatocyte damage and cholestasis in the context of a systemic inflammatory response syndrome. Yet, the entire pathophysiology is still only incompletely understood and likely encompasses hitherto unknown mechanisms that could be crucial and potentially exploitable for therapeutic intervention. Hence, studies that pursue novel hypotheses to explore new therapeutic targets are of particular interest. Among the manifold potential mechanisms is that of microRNAs (miRNA) in human diseases, including liver diseases. miRNAs are highly conserved small non-coding RNA molecules that is composed of 18–25 nucleotides involved in epigenetic regulation of numerous target genes either at the posttranscriptional level or transcriptionally via targeting promoter regions of genes, while in turn target genes can also be regulated by a multitude of miRNAs. Aberrant expression of miRNAs in liver tissue was described for the pathogenesis of several liver diseases, including viral hepatitis, non-alcoholic fatty liver disease, alcohol-induced liver damage and, particularly, hepatocellular carcinoma.4 Published data on alcoholic hepatitis (AH) have been limited to experimental studies in rodents so far.5 ,6

In this issue of Gut, Blaya et al 7 add important information to the existing knowledge via a carefully executed systematic profiling analysis of miRNAs in human alcoholic hepatitis (AH) and their experimental translation in several animal models involving toxic and alcohol-mediated liver injury, and cell culture. At the outset, RNA samples obtained from liver tissues of patients with severe alcoholic hepatitis (sAH), alcoholic cirrhosis, cirrhosis from non-alcoholic liver disease and hepatitis C, and healthy patients were analysed on an Affymetrix miRNA GeneChip to identify miRNAs significantly upregulated or downregulated in severe alcoholic hepatitis (sAH) compared with the other conditions. Among the 111 miRNAs upregulated in alcoholic hepatitis (AH) and the 66 which were downregulated, miR-21, miR-155, miR-214 and miR-422 were assayed by quantitative PCR of liver tissues in an additional cohort of 35 patients with severe alcoholic hepatitis (sAH). Here, the former three miRNAs were upregulated and the latter was downregulated. Of all miRNAs on the array, miR-182 was the most highly expressed in alcoholic hepatitis (AH) and significantly more expressed in livers of patients with severe alcoholic hepatitis (sAH) compared with non-alcoholic liver diseases, and thus selected for further investigation. Interestingly, there was no difference between its hepatic expression between alcoholic hepatitis (AH) and alcoholic cirrhosis, possibly since the patients with severe alcoholic hepatitis (sAH) included into this study were not extremely ill as evidenced by a low percentage of patients with kidney failure, and model for end-stage liver disease (MELD) scores at the verge of ‘severity’. By means of a multidimensional scaling analysis of all arrayed RNA samples, authors could convincingly show that the miRNA profile in alcoholic hepatitis (AH) clearly differs from that of other liver diseases including alcoholic cirrhosis, with 17 miRNAs being upregulated and only one being downregulated. Importantly, the 17 upregulated miRNAs demonstrated only a low level of expression in non-alcoholic liver diseases indicating their specificity for alcoholic hepatitis (AH). A subsequent ingenuity pathway analysis of the 18 differentially expressed miRNAs revealed their link to pathways related to the nuclear receptors pregnan X receptor (PXR), retinoid X receptor (RXR) and farnesoid X receptor (FXR), and to cholestasis. The link to farnesoid X receptor (FXR) raises the hypothesis that farnesoid X receptor (FXR) agonists, like obeticholic acid which is currently under investigation for non-alcoholic fatty liver disease and primary biliary cholangitis (PBC),8 ,9 may deserve testing as a new therapeutic option for alcoholic hepatitis (AH). This miRNA panel appears to also target pathways essential for hepatocyte differentiation (HNF4α), liver regeneration and homeostasis (STAT3, NR5A2), which were shown to be defective in severe alcoholic hepatitis (sAH) and identified as potential target for therapeutic intervention.

Clinically, hepatic expression of miR-182 correlated with the MELD and score - age, bilirubin INR, creatinine (ABIC) score, serum bilirubin levels and histological cholestasis including immunohistochemical markers of ductular reaction (keratin 7 and epithelial cell adhesion molecule (EPCAM). Importantly, high levels of miR-182 expression correlated with higher 90-day mortality, but did not predict treatment response to corticosteroids after 7 days. miR-182 levels as measured in serum were significantly higher in patients with alcoholic hepatitis (AH) compared with control subjects without liver disease, but did not correlate with hepatic expression, disease severity or mortality supporting the assumption that serum miR-182 levels may not be ideal biomarkers for alcoholic hepatitis (AH). Apart from the abundant data in humans, authors explored miR-182 further in 4 (!) animal models of alcoholic and toxic liver injury to compensate for the lack of an ideal model of experimental alcoholic hepatitis (AH) that closely mirrors all important features of human alcoholic liver disease, and particularly of alcoholic hepatitis (AH). Here, only a combination of CCl4 plus ethanol along the Lieber-de Carli protocol produced a mild increase in hepatic miR-182 expression, while either alone had no effect. In a second model using 3,5-diethoxycarbonyl-1,4-dihydrocollidine (DDC) to induce inflammation and biliary damage, DDC resulted in a time-dependent increase of miR-182 expression in the liver, whereas the combination of DDC with alcohol had no additional effect. In an attempt to assign miR-182 expression to the cellular compartment of the liver, isolation of parenchymal liver cells (cholangiocytes, hepatocytes) and non-parenchymal cells (Kupffer cells, hepatic stellate cells/myofibroblasts) allowed to confirm that the magnitude of miR-182 stems from cells involved in ductular reactions, and to a lesser extent from hepatocytes while non-parenchymal cells revealed only a low expression. A strong case for a relevant role of miR-182 in the induction of alcohol-related liver injury, however, comes from a set of experiments in which liver injury in DDC-fed mice was significantly alleviated by administration of a neutralising decoy-182 plasmid with scavenging properties towards miR-182. Finally, transfection of a cholangiocyte cell line (H69) with miR-182 resulted in upregulation of proinflammatory genes, and a downregulation of expression of solute carrier family 1 member 1 (SLC1A1). While the former fits to the hypothesis of inflammation-triggered biliary damage, the latter observation remains somewhat unclear as it is not obvious why authors decided to assay for this gene. Another solute carrier family member, SLC38A4, has recently come into focus as a genetic variant of the SLC38A4 gene was found associated with severe alcoholic hepatitis (sAH) in a British genome-wide association study,10 and it would therefore be interesting to learn how miR-182 would affect SLC38A4 gene expression in cholangiocytes, and whether hepatic miR-182 expression correlates with SLC38A4 expression in mice and human liver of patients with severe alcoholic hepatitis (sAH).

The identified candidate miR-182 may not be so useful as a biomarker for clinical decision making as it seems to be for hepatocellular carcinoma (HCC),11 but very well suited to decipher the biliary involvement in clinical alcoholic hepatitis (AH). As convincingly shown by the authors through human, mouse and cell culture data, miR-182 mainly derives from biliary cells involved in the ductular reaction typical for the marked jaundice and histologically important bilirubinostasis encountered in patients with alcoholic hepatitis (AH).12 So, exploration of miR-182 expression in models of biliary injury, such as in bile-duct ligation, and in primary cholestatic liver diseases may serve to elucidate its role in biliary damage in general.

In conclusion, the study by Blaya et al is a fine example on how to address a novel idea with a hypothesis-generating first step (microarray) and to carry the gained leads further into clinical (human tissues) and functional analyses (animal and cell culture experiments). Such approaches are extremely helpful to expand our knowledge about a yet incompletely understood disease and to pave the way for urgently needed new treatments.


View Abstract


  • Contributors LD and myself have conceived and drafted the manuscript equally and approve its final version.

  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles