HLF and PPARα axis regulates metabolic-associated fatty liver disease through extracellular vesicles derived from the intestinal microbiota

Abstract

Metabolic-associated fatty liver disease (MAFLD) has become increasingly widespread. The intestine is the primary site of lipid absorption and is important for the homeostasis of lipid metabolism. However, the mechanism underlying the participation of the intestinal tract in the development of MAFLD requires additional investigation. In this study, analysis of the single-cell transcriptome of intestinal tissue from cynomolgus monkeys found that hepatic leukemia factor (HLF) participated in the genetic regulation of intestinal lipid absorption. Results obtained from normal and intestine-specific Hlf-knockout mice confirmed that HLF alleviated intestinal barrier disorders by inhibiting peroxisome proliferator-activated receptor alpha (PPARα) expression. The HLF/PPARα axis alleviated MAFLD by mediating gut microbiota-derived extracellular vesicles (fEVs), thereby inhibiting hepatocyte ferroptosis. Lipidomics and functional experiments verified that taurochenodeoxycholic acid (TCDCA), a conjugated bile acid contained in the fEVs, had a key role in the process. In conclusion, intestinal HLF activity was mediated by fEVs and identified as a novel therapeutic target for MAFLD.

Keywords: HLF; MAFLD; bile acids; extracellular vesicles; ferroptosis; gut microbiome.

Figure 1

Hepatic leukemia factor (HLF) is a novel target for regulating intestinal lipid absorption. (A) Experimental workflow diagram. (B) Single‐cell sequencing annotation map of cell types in intestinal tissues from healthy and obese cynomolgus monkeys. (C, D) Kyoto encyclopedia of genes and genomes (KEGG) bubble plot and gene set enrichment analysis (GSEA) plot of all differentially expressed genes in intestinal absorptive epithelial cells (N = 1404). (E, F) Dimensionality reduction clustering bubble plot of marker genes and annotation map for intestinal absorptive epithelial cells. (G) Bar graph showing the proportion of cell types in both groups. (H, I) Venn diagrams of upregulated and downregulated genes in long‐chain fatty acid absorptive cells, carbohydrate absorptive cells, and dual‐function absorptive cells based on transcriptome data. (J) Heatmap of differentially expressed genes. (K) Violin plot of differentially expressed genes. (L) Bar graph showing the expression of differentially expressed genes in the intestinal transcriptome of cynomolgus monkeys. (M) Immunoblot and quantification of HLF protein in cynomolgus monkeys (n = 3). (N) Expression of Hlf mRNA in cynomolgus monkeys (n = 3). (O) Immunoblot and quantification of HLF protein in mice (n = 3). (P) Expression of Hlf mRNA in mice (n = 6). (Q) Intracellular triglyceride (TG) levels in Caco‐2 cells overexpressing HLF (n = 4). (R, S) Fatty acid uptake in HLF‐overexpressing and Hlf‐silenced cells (n = 4). Primer sequences are listed in Table S1. HFD, high‐fat diet group; MAFLD, metabolic‐associated fatty liver disease; ND, control group. Data are presented as mean ± standard deviation. Repeated measures analysis of variance was used to compare trends across two curves over multiple time points. *p < 0.05, **p < 0.01.

Figure 2

Partial hepatic leukemia factor (HLF) deficiency improves metabolic‐associated fatty liver disease (MAFLD). (A) Schematic diagram of breeding for intestinal‐specific heterozygous Hlf knockout mice. (B) Genotype Identification of mice. (C, D) Western blot analysis of HLF expression in intestinal tissues (n = 3). (E) Western blot analysis of HLF expression in the heart, liver, spleen, muscle, and adipose tissues (n = 3). (F–H) Body weight and representative photographs of mouse body size at the end of the experiment (n = 6). (I) Body fat percentage of mice (n = 6). (J) Computed tomography (CT) imaging of mice. (K–M) Epididymal fat, subcutaneous fat, and liver weights along with corresponding tissue weights (n = 7). (N) H&E staining of epididymal and subcutaneous fat, and H&E, Oil Red O, and Masson staining of the liver. (O) Liver triglyceride (TG) and total cholesterol (TC) levels (n = 6). (P) Glucose tolerance test (GTT) and quantification of AUC in mice (n = 6). (Q) Insulin sensitivity test and quantification of AUC in mice (n = 6). (R, S) Energy metabolism analysis normalized to 40 g body weight (n = 3). (T–X) Liver levels of reactive oxygen species (ROS), malondialdehyde (MDA), glutathione (GSH), catalase (CAT), and Fe²⁺/Fe³⁺ (n = 6–8). Data are presented as mean ± standard deviation. The Friedman test was used for four‐group comparisons with repeated measures over time. *p < 0.05, **p < 0.01. Different letters in the figure indicate significant differences between groups.

Figure 3

Hepatic leukemia factor (HLF) regulates peroxisome proliferator‐activated receptor alpha (PPARα) expression. (A) Volcano plot of differentially expressed genes (S vs. D). D correspond to Hlf ^+/+, S correspond to Hlf ^+/−. (B) Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis of all differentially expressed genes (N = 515). (C) Gene set enrichment analysis (GSEA) of all differentially expressed genes (N = 515). (D) Heatmap of differentially expressed genes. D1–D3 correspond to Hlf ^+/+1–Hlf ^+/+3, and S1–S3 correspond to Hlf ^{+/ −}1–Hlf ^{+/ −}3. (E) Ppara promoter activity assay (n = 4). (F, G) Immunofluorescence analysis and quantification of PPARα expression in Caco‐2 (n = 3). (H–K) Western blot analysis and quantification of PPARα expression in Caco‐2 and intestinal tissues of mice (n = 3). (L) Ppara mRNA levels in Caco‐2 overexpressing HLF (n = 4). (M) Immunofluorescence analysis of PPARα in mouse intestinal tissues. (N, O) Intracellular and extracellular triglyceride (TG) levels (n = 4). (P, Q) Fatty acid uptake levels (n = 4). Data are presented as mean ± standard deviation. Repeated measures analysis of variance was used to compare trends across two curves over multiple time points. *p < 0.05, **p < 0.01. Different letters in the figure indicate significant differences between groups.

Figure 4

Inhibition of peroxisome proliferator‐activated receptor alpha (PPARα) alleviates metabolic‐associated fatty liver disease (MAFLD). (A) Schematic diagram of the mouse experimental design. (B, C) Mouse body weight and representative images of body size (n = 6). (D) Body fat percentage of mice (n = 6–7). (E) Computed tomography (CT) imaging of mice. (F) Epididymal fat, subcutaneous fat, and liver weights, along with representative images (n = 6). (G) H&E staining of subcutaneous fat, and H&E and Oil Red O staining of liver tissue. (H) Liver triglyceride (TG) and total cholesterol (TC) levels (n = 6). (I) Glucose tolerance test (GTT) and quantification of AUC (n = 6). (J) Insulin tolerance test (ITT) and quantification of AUC (n = 6). (K) Serum FITC‐dextran 4 (FD4) levels in mice (n = 4). (L–M) Serum levels of lipopolysaccharide (LPS), lipopolysaccharide‐binding protein (LBP), and cluster of differentiation 14 (sCD14) in mice (n = 6). (N) Proportional analysis of gut microbiota at the phylum level. (O) Differential species between the HCON and HGW groups. (P) Correlation analysis of gut microbial species with mouse phenotypic traits. (Q) Liver Fe²⁺/Fe³⁺ levels (n = 6–8). (R, S) Western blot analysis and quantification of hepatic solute carrier family 7 member 11 (SLC7A11), glutathione peroxidase 4 (GPX4), and acyl‐coa synthetase long chain family member 4 (ACSL4) expression (n = 3). Data are presented as mean ± standard deviation. Repeated measures analysis of variance was used to compare trends across two curves over multiple time points. *p < 0.05, **p < 0.01.

Figure 5

Gut Microbiota‐derived extracellular vesicles (fEVs) from GW6471‐treated mice improve intestinal permeability and inhibit ferroptosis. (A) Schematic diagram of the mouse experimental procedure. (B, C) Body weight curve and representative photos of mouse body morphology (n = 6). (D) Body fat percentage of mice (n = 6). (E) Computed tomography (CT) imaging of mice. (F, G) In vivo and tissue tracing of fluorescence‐labeled fEVs in mice (n = 3). (H) Weight of epididymal fat, subcutaneous fat, and liver in mice (n = 6). (I) H&E staining of adipose tissue, H&E staining, and Oil Red O staining of liver tissue. (J) Liver triglyceride (TG) and total cholesterol (TC) levels (n = 6). (K) Glucose tolerance test and area under the curve (AUC) quantification in mice (n = 6). (L) Insulin tolerance test and AUC quantification in mice (n = 6). (M, N) Energy metabolism analysis and energy expenditure normalized to 40 g in mice (n = 5). (O) Serum FD4 levels in mice (n = 4). (P, Q) Serum lipopolysaccharide (LPS), lipopolysaccharide‐binding protein (LBP), and cluster of differentiation 14 (sCD14) levels in mice (n = 6). (R) Transmission electron microscopy (TEM) images of the liver. The arrows indicate mitochondrial cristae. (S–W) Levels of reactive oxygen species (ROS), Malondialdehyde (MDA), glutathione (GSH), catalase (CAT), and Fe²⁺/Fe³⁺ in the liver (n = 6). (X) Western blot analysis and quantification of solute carrier family 7 member 11 (SLC7A11) and acyl‐coa synthetase long chain family member 4 (ACSL4) in mouse liver (n = 3). Data are presented as mean ± standard deviation. Repeated measures analysis of variance was used to compare trends across two curves over multiple time points. *p < 0.05, **p < 0.01.

Figure 6

Lipid Alterations in gut microbiota‐derived extracellular vesicles (fEVs) influence hepatic steatosis. (A) PCA plot of lipidomics data. (B) Volcano plot of differential lipid species in the lipidomics analysis (GW6471 vs CON). (C) Heatmap of differentially abundant lipids. (D) Violin plots of SPH (d18:1) and taurochenodeoxycholic acid (TCDCA) levels. (E) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis. (F) Heatmap showing the correlation between lipidomics data and in vivo phenotypes. (G) Heatmap of bacteria enriched with BSH enzymes. (H) TCDCA levels in intestinal contents of CON and GW6471‐treated mice (n = 6). (I) TCDCA levels in intestinal contents of Hlf ^+/+ and Hlf ^{+/ −} mice (n = 6). (J) Cell viability assay (CCK8) in HepG2 cells. (K) Intracellular TG levels in HepG2 cells treated with TCDCA (n = 4). (L–O) Fluorescence images and quantification of mitochondrial ROS (MitROS), Calcein‐AM (calcium green), FerroOrange, and lipid peroxidation in HepG2 cells (n = 4). Data are presented as mean ± standard deviation. *p < 0.05, **p < 0.01.