Identification of foam cell like M2 macrophages, AEBP1 biomarkers, and resveratrol as potential therapeutic in MASLD using Ecotyper and WGCNA

Abstract

The immune cell landscapes in metabolic dysfunction-associated fatty liver disease (MAFLD) and their clinical relevance have not been explored. We used Ecotyper to identify immune cell states based on gene expression and examined their roles in metabolic dysfunction-associated steatotic liver disease (MASLD) progression. Limma was applied to identify differentially expressed genes (DEGs). Weighted gene co-expression network analysis (WGCNA) was used for module identification. Bidirectional Mendelian randomization (MR) analysis was used to validate the causal effect of AEBP1 on metabolic dysfunction-associated steatohepatitis (MASH). Out of 71 immune cell states, 32 showed significant differences between MASLD and MASH. The six most significant states were Fibroblasts.3 (tumor-associated), Epithelial.cells.3 (pro-angiogenic), PMNs.3 (classically activated), Macrophages.6 (M2 foam cell-like), Mast.cells.5, and Fibroblasts.7. All six cell states belong to the CE1 ecotype. Further analysis revealed that Fibroblasts.3 had the highest discriminatory ability in distinguishing MASH from MASLD, followed by Epithelial.cells.3 and Macrophages.6. At the ecotype level, CE1 showed the strongest ability to differentiate between MASLD and MASH, with a performance score (AUC) of 0.891. CE3 followed with a slightly lower performance (AUC = 0.826). Conversely, higher CE4 effectively differentiated MASLD from MASH (AUC = 0.871). Genes up-regulated in CE1-high samples were enriched in extracellular matrix (ECM) organization and the PI3K-Akt signaling pathway, while down-regulated genes were linked to copper ion responses. These genes formed three modules associated with fibroblasts and macrophages. We identified resveratrol, a polyphenolic compound, as a potential therapeutic drug capable of modulating these immune cell states. Protein-ligand docking analysis illuminated interactions between resveratrol and the Macrophages.6 marker gene AEBP1. This study provides a comprehensive exploration of the clinical significance of immune cell states in MAFLD. It identifies potential molecular mechanisms and therapeutic candidates. Further clinical trials are needed to validate the efficacy of resveratrol and explore its structure-activity relationships to develop targeted treatments.

Keywords: Ecotyper; Fibroblasts; Immune cell landscapes; MAFLD; Resveratrol.

Fig. 1

The overall data analysis framework for our analysis. Public transcriptomic datasets (GSE167523, GSE135251, and GSE68421) were analyzed. Ecotyper was used to identify transcriptionally distinct immune cell states and ecotypes in MASLD and MASH. Differentially expressed immune cell states between MASLD and MASH were identified to reveal disease-specific signatures. WGCNA was applied to detect co-expression modules associated with disease progression. Functional enrichment analyses were performed using GO and KEGG to interpret the co-expression modules. Resveratrol was identified as a potential therapeutic candidate through statistical drug screening models. External datasets were used to validate the role of resveratrol and immune cell states in MASLD after treatment. Immune cell states’ ability to distinguish MASLD from MASH was evaluated using AUC analysis. Mendelian randomization was conducted to establish causal relationships between key genes and MAFLD progression. WGCNA: weighted gene co-expression network analysis; GO: gene ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; MASLD: metabolic dysfunction-associated steatotic liver; MASH: metabolic dysfunction-associated steatohepatitis; AUC: Area Under the Curve.

Fig. 2

The variance in cell abundance across six distinct cell states between MASLD and MASH, and marker genes across fibrosis stages in GSE167523. (A) The heatmap shows the distribution of 71 cell stats in MAFLD. (B) Box plots for the six top significant cell states between MASLD and MASH (P < 0.0001, t test). PMNs: Polymorphonuclear neutrophils, MASLD: metabolic dysfunction-associated steatotic liver disease, MASH: metabolic dysfunction-associated steatohepatitis. (C) Macrophages.6 marker gene AEBP1, Fibroblasts.3 marker gene COL10A1, Epithelial.cells.3 marker gene ITGA3, and Mast.cells.5 marker gene ARRB2 are progressively up-regulated across fibrosis stages 0 ~ 4 during MASH progression (P < 0.001, ANOVA). (D) Single-cell analysis shows that in normal liver tissue, AEBP1 is expressed in fibroblasts, smooth muscle cells, endothelial cells, and macrophages in the Human Protein Atlas. The x-axis and y-axis represent the first and second UMAP dimensions (UMAP1 and UMAP2), respectively, illustrating the two-dimensional projection of single-cell transcriptomic profiles. UMAP: Uniform Manifold Approximation and Projection. (E) AEBP1 is significantly expressed in M2 macrophages in normal and carcinoma liver (P < 0.001, ANOVA) in GEPIA2021. HCC: hepatocellular Carcinoma.

Fig. 3

Differentially expressed (A) Fibroblasts.3 and (B) Macrophages.6 were positively associated with MASH fibrosis severity in GSE135251. The comparison between disease groups was conducted by the Kruskal-Wallis Analysis of Variance (ANOVA) test with adjusted P values. MASLD: metabolic dysfunction–associated steatotic liver disease. MASH: metabolic dysfunction-associated steatohepatitis.

Fig. 4

ROC curves show that the six cell states could separate metabolic dysfunction-associated steatotic liver disease (MASLD) and metabolic dysfunction-associated steatohepatitis (MASH) patients in GSE167523. ROC: receiver operating characteristic, AUC: area under the ROC curve.

Fig. 5

The distinctive expression patterns of samples expressing high levels of CE1 in GSE167523. (A) A total of 727 genes exhibiting differential expression (DEGs) are identified. (B) The clustering heatmap of the DEGs displays the discernible expression patterns that differentiate these groups. In the color bar situated above the heatmap, the color blue denotes MASLD or men, while the color pink denotes MASH or women. (C) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enrichment analysis for the up-regulated genes. (D) Gene Ontology (GO) biological processes enrichment analysis for the down-regulated genes. Both enrichment analyses were performed in the clusterProfiler tool.

Fig. 6

Differentially expressed genes between ecotype CE1 high and low-expressing samples in GSE167523 were organized into three functional modules. (A) When a power value of 22 is set, the constructed co-expression network adheres to a power law. (B) The cluster dendrogram shows gene module assignments. The accompanying color bar denotes the respective gene module assignments. (C) Heatmap depicting the relationship between modules and traits. Each cell within the heatmap corresponds to the correlation between the expression of specific modules and distinct clinical parameters. The numerical values enclosed in brackets within each cell denote the statistical significance of the correlation. (D,E) Gene Ontology (GO) biological processes enrichment analysis is performed on the genes within modules M1 and M2 by the clusterProfiler tool.

Fig. 7

Analysis of protein-ligand docking for resveratrol and the marker genes AEBP1 in Macrophages.6. (A) The three-dimensional structure models showcase the interactions between AEBP1 and resveratrol. (B) The pose view provides a visual representation of the interaction sites, with hydrogen bonds depicted by black dashed lines between red atoms. (C) Immune-related genes IGHV3-74, IGHV3-33, IGHA1, IGKC, and IGLV1-44 were down-regulated by resveratrol treatment in human MASLD (GSE68421, n = 7). MASLD: metabolic dysfunction-associated steatotic liver disease.