Clinical, histological and molecular profiling of different stages of alcohol-related liver disease

Abstract

Objective: Alcohol-related liver disease (ALD) ranges from never-decompensated ALD (ndALD) to the life-threatening decompensated phenotype, known as alcohol-related hepatitis (AH). A multidimensional study of the clinical, histological and molecular features of these subtypes is lacking.

Design: Two large cohorts of patients were recruited in an international, observational multicentre study: a retrospective cohort of patients with ndALD (n=110) and a prospective cohort of patients with AH (n=225). Clinical, analytical, immunohistochemistry and hepatic RNA microarray analysis of both disease phenotypes were performed.

Results: Age and mean alcohol intake were similar in both groups. AH patients had greater aspartate amino transferase/alanine amino transferase ratio and lower gamma-glutamyl transferase levels than in ndALD patients. Patients with AH demonstrated profound liver failure and increased mortality. One-year mortality was 10% in ndALD and 50% in AH. Histologically, steatosis grade, ballooning and pericellular fibrosis were similar in both groups, while advanced fibrosis, Mallory-Denk bodies, bilirubinostasis, severe neutrophil infiltration and ductular reaction were more frequent among AH patients. Transcriptome analysis revealed a profound gene dysregulation within both phenotypes when compare to controls. While ndALD was characterised by deregulated expression of genes involved in matrisome and immune response, the development of AH resulted in a marked deregulation of genes involved in hepatocyte reprogramming and bile acid metabolism.

Conclusions: Despite comparable alcohol intake, AH patients presented with worse liver function compared with ndALD patients. Bilirubinostasis, severe fibrosis and ductular reaction were prominent features of AH. AH patients exhibited a more profound deregulation of gene expression compared with ndALD patients.

Keywords: alcohol; alcohol-induced injury; alcoholic liver disease; gene expression; histopathology.

Figure 1

Histological features of AH and ndALD patients. (A) Steatosis, ndALD patient H&E viewed using a ×20 objective (x20) (B) steatosis, AH patient, H&E, ×20. (C) Pericellular and perisinusoidal fibrosis in ndALD patient, Masson-Trichrome ×20. (D) Bridging fibrosis and pericellular fibrosis on AH patient, Masson-Trichrome ×10. (E) Structural preserved parenchyma area without inflammatory infiltration, ndALD patient, H&E, ×20. (F) PMN infiltration, AH patient, H&E, ×20. (G) Normal bile duct with preserved parenchymal architecture without bilirubinostasis, ndALD, H&E, ×20. (H) Hepatocanalicular bilirrubinostasis (black arrow) on AH patient, H&E, ×20. (I) Percentage of different grade of steatosis and fibrosis, presence of PMN infiltration and types of bilirrubinostatis by cohorts (ndALD and AH patients). Scale bars indicate 50 μm. AH, alcohol-related hepatitis; MT, Masson’s trichrome; ndALD, never-decompensated alcohol-related liver disease; PMN, polymorphonuclear cell.

Figure 2

Ductular reaction analysis in control, ndALD with and without underling significant fibrosis and AH patients. Immunohistochemistry (IHC) of KT7: (A) ndALD patient without significant fibrosis (no fibrosis or only portal fibrosis without septa or with incomplete septa), KT7 IHC observed with ×20 objective, (B) ndALD patient with significant fibrosis KT7 IHC and (C) decompensated AH patients with significant fibrosis, KT7 IHC. (D) Semiquantitative assessment of KT7 within the three phenotypes (ndALD without significant fibrosis n=4; ndALD with significant fibrosis n=3, AH n=5). More details about the semiquantitative assessment can be found in online supplemental methods (E) DNA microarray analysis of the most representative genes of ductular reaction within the three phenotypes is presented as a heat map. Scale bar indicate 100 μm. AH, alcohol-related hepatitis; KT7, cytokeratin 7; ndALD, never-decompensated alcohol-related liver disease; NSF, no significant fibrosis.

Figure 3

Liver gene expression patterns of patients with ndALD and AH. (A) RNA microarray analysis is presented as a heatmap. The genes were manually selected by their role on hepatocyte function and the significant up or downregulation when comparing ndALD versus AH patients (p<0.001). (B) Gene Set Enrichment Analysis (GSEA) of differentially expressed genes comparing ndALD versus AH livers. Hallmark gene set was used and Normalised Enrichment Score (NES) is presented. Positive (red) and negative (blue) bars indicate up and downregulated gene sets in AH. AH, alcohol-related hepatitis; ndALD, never decompensated alcohol-related liver disease.

Figure 4

Overlap between molecular signature of AH versus ndALD. Cytoscape enrichment map of Gene Set Enrichment Analysis was used to visualise overlap between Molecular Signatures Database of curated gene sets (C2) that were enriched in patients with AH vs patients with ndALD. Red nodes represent upregulated while blue nodes represent downregulated pathways in AH patients. Only nodes with p value below 0.1 and edges with similarity above 0.4 are shown. AH, alcohol-related hepatitis; CREB, c-AMP responsive element binding protein; GPCR, G protein-coupled receptor; MAPK, mitogen-activated protein kinase; ndALD, never decompensated alcohol-related liver disease; SLC, solute carrier family member.