Hepatic and extrahepatic metabolic modulation in hbv-related decompensated cirrhosis and acute-on-chronic liver failure

Abstract

Acute-on-chronic liver failure (ACLF) and decompensated cirrhosis (DC) are life-threatening syndromes that can develop at the end-stage of chronic hepatitis B virus (HBV) infection. Both ACLF and DC are complicated by hepatic and extrahepatic pathogeneses. To better understand the compartment-specific metabolic modulations related to their pathogenesis, HBV-DC, HBV-ACLF patients, and controls (30 each) were analyzed by metabolomics using portal (Port), hepatic vein (Hep), and peripheral (Peri) serum. Compartment ratios of metabolites (Ratio_Hep/Port, Ratio_Peri/Hep, and Ratio_Port/Peri) were calculated. The liver tissues (10 per group) were analyzed using transcriptomics and metabolomics. An additional 75 patients with ACLF, 20 with DC, and 20 with liver cirrhosis (LC) were used to confirm oxlipid dysregulation. Both multi-omics datasets suggest suppressed energy, amino acid, and pyrimidine metabolism in the ACLF/DC liver. The serum metabolomic variations were contributed primarily by disease rather than sampling compartments, as both HBV-ACLF and HBV-DC patients demonstrated abnormal profiles of amino acids and peptides, indoles, purines, steroids, and benzimidazoles. In ACLF/DC patients, impaired hepatic metabolism resulted in a highly correlated hepatic and portal vein serum metabolome and release of inflammatory lipids and heme metabolites from the liver. HBV-ACLF showed higher Ratio_Peri/Hep of extrahepatic inflammatory oxlipids, while HBV-DC patients showed higher Ratio_Port/Peri of gut microbial metabolites. An inflammatory oxlipid outburst was confirmed in the early stages of HBV-ACLF. The inflammatory effects of the selected oxlipids were confirmed in monocytes. These findings support a synergy between liver-specific mechanisms and systemic inflammation in ACLF/DC development, and that pro-inflammatory oxlipids are metabolic signatures of early HBV-ACLF.

Keywords: Acute-on-chronic liver failure; HBV; decompensated cirrhosis; metabolomics; oxlipidomics.

Figure 1.

Overall schematic workflow of multi-omics approach designed to simultaneously investigate liver metabolomics and transcriptomics, as well as serum metabolomics from portal vein, hepatic vein and peripheral circulation of DC and ACLF patients.

Figure 2.

Characterization of liver metabolomics and transcriptomics in ACLF, DC and control patients. (a) schematic workflow of O2PLS modeling to unveil metabolomics and transcriptomics joint variation. (b) O2PLS score plot showing metabolomics and transcriptomics joint variation was related to disease progression, majorly modeled by component 1. (c) O2PLS loading plot showing correlation each metabolite and gene to component 1 (loading 1) and component 2 (loading 2), therefore reflect their contribution to the sample inter-group separation. (d) key pathway components significantly altered in liver tissue during ACLF/DC progression.

Figure 3.

Characterization of serum metabolomics at portal vein (port), hepatic vein (hep) and peripheral circulation (Peri) from ACLF, DC and control patients. (a) overall PCA plot showing overall inter-group differences of serum metabolome were majorly attributed to disease progression. (b) comparison of spearman coefficient of all metabolites between different locations showing the similarity of portal vein and hepatic vein serum metabolite profiles were higher in DC and ACLF patients. (c) summary of up- or down-regulated metabolites in each location between patient groups. (d) heatmap of significantly altered metabolites in DC and ACLF as compared to control samples. The averaged LC-MS intensity ratio of each metabolite in DC or ACLF samples versus control samples were Log2 transformed before hierarchical clustering by Euclidean distance. (e) chemical category enrichment analysis of metabolites that shown significantly alterations comparing DC or ACLF versus controls across all 3 locations, enrichment significance (p values) were visualized by heatmap. (f) number of metabolites that shown significantly alterations comparing ACLF versus DC in all 3 locations while their chemical attributes were summarized in barplot.

Figure 4.

Metabolic network connected amino acids, urea, pyrimidine, purine and polyamine metabolism (a). Key enzymes in each metabolic conversion step were encircled by box. Red and green represent up- or down-regulated metabolites or genes in ACLF patient compared to controls; while black represent no significant changes. GABA, γ-aminobutyric acid; GSH, reduced glutathione; GSSG, oxidized glutathione; MTA, 5”-deoxy-5”-(methylthio)adenosine; MTR, methylthioribose; SAM, S-adenosylmethionine; SAH, S-adenosylhomocysteine; THF, tetrahydrofolate. Alteration of hepatic metabolite and genes between patients were summarized by heatmaps grouped by amino acid metabolism pathways (b), while their corresponding levels in serum samples were summarized in (c). In all heatmaps, up- or down-regulated genes or metabolites were colored by red or green shades, comparisons reached statistical significance (student’s t-test with BH adjust FDR < 0.05) were framed by solid border.

Figure 5.

Characterization of serum metabolomics shift between locations by Ratio_Hep/Port (a), Ratio_Port/Peri (b), and Ratio_Peri/Hep (c). The left panel includes PCA plots showing overall shift of 3 pair-wise ratios attributed to disease progression. The middle panel summarizes up- or down-regulated metabolites in each pair-wise ratio between ACLF vs con or DC vs con. Right panel includes volcano plots showing significantly changed metabolites (student’s t-test with BH adjust FDR < 0.05 and fold-change >1.5) between ACLF and DC.

Figure 6.

Alteration of circulating oxlipids between patients were summarized by heatmaps (a). Average log2 transformed level within each patient group were used. Up- or down-regulation as compared to LCs were colored by red or green shades. Proposed modulation of PUFA metabolism in ACLF patients (b). Significant upregulated or down-regulated species in ACLF grade 1 patients vs DC patients were colored in red or green.