Role of MIF in coordinated expression of hepatic chemokines in patients with alcohol-associated hepatitis

Abstract

The chemokine system of ligands and receptors is implicated in the progression of alcohol-associated hepatitis (AH). Finding upstream regulators could lead to novel therapies. This study involved coordinated expression of chemokines in livers of healthy controls (HC) and patients with AH in 2 distinct cohorts of patients with various chronic liver diseases. Studies in cultured hepatocytes and in tissue-specific KO were used for mechanistic insight into a potential upstream regulator of chemokine expression in AH. Selected C-X-C chemokine members of the IL-8 chemokine family and C-C chemokine CCL20 were highly associated with AH compared with HC but not in patients with liver diseases of other etiologies (nonalcoholic fatty liver disease [NAFLD] and hepatitis C virus [HCV]). Our previous studies implicate macrophage migration inhibitory factor (MIF) as a pleiotropic cytokine/chemokine with the potential to coordinately regulate chemokine expression in AH. LPS-stimulated expression of multiple chemokines in cultured hepatocytes was dependent on MIF. Gao-binge ethanol feeding to mice induced a similar coordinated chemokine expression in livers of WT mice; this was prevented in hepatocyte-specific Mif-KO (MifΔHep) mice. This study demonstrates that patients with AH exhibit a specific, coordinately expressed chemokine signature and that hepatocyte-derived MIF might drive this inflammatory response.

Keywords: Chemokines; Hepatitis; Hepatology; Inflammation; Innate immunity.

Figure 1. WGCNA identified specific gene modules related to alcohol-associated hepatitis.

(A) The relationship of each color module to disease status. Correlation coefficients and P values are presented within each module per diagnosis. (B and C) Correlation of module membership versus gene significance was calculated in GSE28619 from green (B) and salmon (C) modules. (D) Venn diagram of the top 50% DEGs in the green module, salmon module, and chemokine ligands and receptors. (E) A total of 10 chemokine ligands and receptors were located in either the green or salmon modules.

Figure 2. Chemokine expression and correlation with diagnosis from RNA-seq in livers of HC and patients with various hepatopathies.

(A and B) Expression of C-C chemokines and CX3CL1 (A) and C-X-C chemokines (B). Data are displayed as normalized transcripts per million (tpm) in box-and-whisker plots representing the mean, interquartile range (box), and upper and lower quartiles (whiskers) for HC (Healthy, n = 10), patients with early alcohol-associated steatohepatitis (Early ASH, n = 12), nonsevere alcohol–associated hepatitis (Non-Severe AH, n = 11), Severe AH (n = 29), nonalcoholic fatty liver disease (NAFLD, n = 9), hepatitis C virus (HCV, n = 9), or hepatitis C virus with cirrhosis (HCV-Cirr, n = 9). (C) Heatmap of Pearson’s correlation coefficients for expression of chemokine ligands to patient diagnosis with values represented in the boxes.

Figure 3. Clustering of chemokine expression data by t-SNE segregated patients with AH from other diagnoses and refined the hepatic chemokine signature in patients with AH.

(A) Clustering of patients as determined by t-SNE of RNA expression for chemokines in livers of patients. (B) Heatmap of chemokine expression correlations in liver. (C) Clustering by t-SNE for expression of chemokine RNA in livers of patients from all hepatopathies.

Figure 4. MIF is required for LPS-mediated upregulation of chemokine mRNA expression in response to LPS challenge.

(A) AML-12 cells were treated with LPS at the indicated concentrations for 90 minutes, and expression of Cxcl1, Lix, Ccl2, and Ccl20 mRNA was determined by qPCR. (B) AML-12 cells were pretreated with VEH (0.1% DMSO) or MIF098 (50 μM) prior to LPS challenge for 90 minutes. Expression of Cxcl1, Lix, Ccl2, and Ccl20 mRNA was determined and normalized to Basal/VEH or LPS/VEH as indicated. (C) AML-12 cells were treated with 1 ng/mL LPS for 30 minutes, and phosphorylation of ERK and p65, as well as the abundance of IκBα, was determined by Western blot. GAPDH was used as a loading control. Values are expressed as means ± SEM. *P < 0.05 versus. BAS controls (n = 4-7). Means with different letters are significantly different, P < 0.05, by 2-way ANOVA with least square means multiple comparison tests. See complete unedited blots in the supplemental material.

Figure 5. Hepatocyte-specific Mif deletion prevents Gao-binge–induced hepatocellular injury, steatosis, and expression of the chemokine signature.

Mif^fl/fl, C57BL/6, and Mif^ΔHep mice were acclimated to a complete liquid diet and allowed free access to ethanol-containing (n = 5–6) or pair-fed (n = 4) control diets per Gao-binge feeding protocol. (A and B) Expression of Cxcl1, Lix, Ccl2, and Ccl20 chemokine mRNA (A) and expression of neutrophil markers Ly6G and Cxcr2 as well as monocyte surface marker Ccr2 mRNA (B) was determined by qPCR in mouse livers. (C) ALT (U/L) and AST (U/L) activity in circulation was determined in plasma, and hepatic triglyceride content was measured in liver homogenate. (D) Expression of ER stress–associated mRNA for spliced Xbp1 (sXbp1), Grp78, Chop, and Dr5 was determined in mouse liver by qPCR. Values are expressed as means ± SEM. P < 0.05 versus pair-fed controls within genotype, by 2-way ANOVA with least square means multiple comparison tests.