Potential Networks Regulated by MSCs in Acute-On-Chronic Liver Failure: Exosomal miRNAs and Intracellular Target Genes

Abstract

Acute-on-chronic liver failure (ACLF) is a severe syndrome associated with high mortality. Alterations in the liver microenvironment are one of the vital causes of immune damage and liver dysfunction. Human bone marrow mesenchymal stem cells (hBMSCs) have been reported to alleviate liver injury via exosome-mediated signaling; of note, miRNAs are one of the most important cargoes in exosomes. Importantly, the miRNAs within exosomes in the hepatic microenvironment may mediate the mesenchymal stem cell (MSC)-derived regulation of liver function. This study investigated the hepatocyte exosomal miRNAs which are regulated by MSCs and the target genes which have potential in the treatment of liver failure. Briefly, ACLF was induced in mice using carbon tetrachloride and primary hepatocytes were isolated and co-cultured (or not) with MSCs under serum-free conditions. Exosomes were then collected, and the expression of exosomal miRNAs was assessed using next-generation sequencing; a comparison was performed between liver cells from healthy versus ACLF animals. Additionally, to identify the intracellular targets of exosomal miRNAs in humans, we focused on previously published data, i.e., microarray data and mass spectrometry data in liver samples from ACLF patients. The biological functions and signaling pathways associated with differentially expressed genes were predicted using gene ontology and Kyoto Encyclopedia of Genes and Genomics enrichment analyses; hub genes were also screened based on pathway analysis and the prediction of protein-protein interaction networks. Finally, we constructed the hub gene-miRNA network and performed correlation analysis and qPCR validation. Importantly, our data revealed that MSCs could regulate the miRNA content within exosomes in the hepatic microenvironment. MiR-20a-5p was down-regulated in ACLF hepatocytes and their exosomes, while the levels of chemokine C-X-C Motif Chemokine Ligand 8 (CXCL8; interleukin 8) were increased in hepatocytes. Importantly, co-culture with hBMSCs resulted in up-regulated expression of miR-20a-5p in exosomes and hepatocytes, and down-regulated expression of CXCL8 in hepatocytes. Altogether, our data suggest that the exosomal miR-20a-5p/intracellular CXCL8 axis may play an important role in the reduction of liver inflammation in ACLF in the context of MSC-based therapies and highlights CXCL8 as a potential target for alleviating liver injury.

Keywords: acute-on-chronic liver failure; exosome; mesenchymal stem cells; microRNA; multi-omics.

FIGURE 1

Characterization of the ACLF model and the derived exosomes. (A) ACLF was induced in mice using carbon tetrachloride, and then primary hepatocytes were isolated and cultured in vitro, in the absence or presence of MSCs. Cells from healthy animals were collected as controls. The supernatants of the three groups of cells were then collected for exosome extraction. (B) Sirius red staining (i-ii, scale bar = 100 μm), Masson’s trichrome staining (iii–iv, scale bar = 50 μm) and H&E staining (v–vi, scale bar = 50 μm) reveal significant fibrosis, collagen deposition, and inflammatory cell infiltration and necrosis, respectively, in the livers of ACLF mice. (C) NTA and TEM-based analyses of exosomes reveal that the particles are concentrated in the 200-nm range and have a cup-like bilayer structure. (D) Western blotting reveals the expression of the exosome membrane marker CD81 and of the intra-particle marker TSG101.

FIGURE 2

Next-generation sequencing of exosomal miRNAs. The heat map shows the relative quantification of exosomal miRNAs; 50 down-regulated (A) and 50 up-regulated (B) miRNAs in ACLF hepatic exosomes. These changes are reversed upon co-culture with hBMSCs. Clustering was performed based on the expression trend. The Z value was used to normalize the RPM value. Red to blue refer to high to low.

FIGURE 3

Exosomal miRNAs targeting regulated DEGs. (A,B) Exosomal miRNAs with identical differential expression in ACLF liver samples were screened as candidate DEmiRs. The Venn diagram shows the number of up- vs. down-regulated miRNAs. The miRNA screening criteria used was |logFC| ≥ 1, p ≤ 0.05. (C,D) The number of DEmiRs targeting genes differentially expressed in the transcriptome. The transcriptome screening criteria was |logFC| ≥ 2, p ≤ 0.05. (E,F) The number of DEmiRs target genes that are differentially expressed in the proteome. DEmiRs: differentially expressed miRNAs.

FIGURE 4

GO and KEGG pathway enrichment analyses of DEGs. (A–C) The 20 most significant GO terms under the BP, CC, and MF categories. The X-axis represents the rich ratio while the Y-axis highlights the GO term names. The color of the bubbles indicates the adj. p value. The range from red to blue represents low to high. The size of the bubbles represents the number of genes enriched in the term. BP, biological processes; CC, cellular component; MF, molecular function. (D) The top 20 pathways enriched as per KEGG. (E) DEGs of the chemokine-chemokine receptor interaction pathway. Red represents up-regulation in ACLF; blue represents down-regulation.

FIGURE 5

The network of Hub genes and miRNAs. (A) Protein-protein interaction (PPI) network and 12 DEGs in the core modules. MCL parameter set to 3. (B) Regulatory network of 15 Hub genes with miRNAs. Orange diamond points represent miRNAs, and blue circle points represent Hub genes. Point and label size represent the node importance index: points, from large to small, represent high to low Neighborhood Connectivity; labels, from large to small, represent high to low Degree. (C) Heat map showing Pearson’s correlation analysis of hub genes with miRNAs. The screened miRNAs show moderate to very strong negative correlations with their target genes. The correlation coefficients of pairs of molecules with target-regulated relationships are highlighted using black borders. Pearson correlation coefficient cut-off of ≥ 0.6. All correlation analyses in the graph are adj. p ≤ 0.001.

FIGURE 6

miR-20a-5p mediates hBMSCs to regulate hepatocyte CXCL8. (A) The binding site and targeting score of miR-20a-5p and CXCL8. (B) The exosomal miR-20a-5p is down-regulated in the context of hepatocytes from ACLF mice; co-culture with hBMSCs promotes its up-regulation. (C) Acute injury by carbon tetrachloride induces the down-regulation of miR-20a-5p in human hepatocytes (L02), while hBMSCs promote its up-regulation. (D) Acute injury by carbon tetrachloride induces the up-regulation of CXCL8 in L02 cells; the phenotype is reversed in the presence of hBMSCs. (E) Enhancement of miR-20a-5p function in carbon tetrachloride-damaged hepatocytes will lead to CXCL8 downregulation.Inhibiting the function of miR-20a-5p will lead to the up-regulation of CXCL8 and block the regulation of hbMSCs on CXCL8. CCl₄, carbon tetrachloride. Data are represented as the mean ± SEM. All p values were obtained using the Student’s t-test: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. mim NC, mimic negative control; 20a mim, miR-20a-5p mimic; inh NC, inhibitor negative control; 20a inh, miR-20a-5p inhibitor.

FIGURE 7

hBMSCs inhibit liver inflammation via the miR-20a-5p/CXCL8 axis. The intracellular levels of miR-20a-5p in hepatocytes increase after the uptake of miR-20a-5p-rich exosomes. In turn, miR-20a-5p inhibits the expression of CXCL8 (together with RISC), resulting in reduced CXCL8 secretion. hBMSCs not only secrete miR-20a-5p-rich exosomes directly into the hepatic microenvironment but also may promote the endogenous synthesis of miR-20a-5p in hepatocytes and its secretion as exosomes via other pathways. The accurate detection of miR-20a-5p and its precursors is required to prove the latter hypothesis. RISC, RNA-induced silencing complex.