Single-cell RNA sequencing reveals a novel inhibitory effect of ApoA4 on NAFL mediated by liver-specific subsets of myeloid cells

Figures

• 

  Figure 1. ScRNA-Seq profiles of liver immune cell populations and the differences between WT and ApoA4 KO mice. (A) Schematic diagram of hepatic immune cell isolation and data analysis using the 10X genomics platform. (B) UMAP visualization of hepatic immune cells showing 10 clusters. (C) Heatmap of the 6 most strongly upregulated genes in each cluster (genes were decreased by P value). (D) Violin plots showing specific marker expression in each cluster. (E) The proportion of each type of immune cell in the WT and KO groups is shown in (B). (F) Volcano plot showing differentially expressed genes (DEGs) between WT and KO mice. Average_log2 fold change > 0.2 & -log10 (P value-adjust) > 2 are red; others are grey. (G) Gene Ontology (GO) enrichment analysis of molecular function (MF), cellular component (CC), and biological process (BP) for the upregulated genes of KO mice vs. WT mice. (H) Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of upregulated genes in KO mice vs. WT mice.

• 

  Figure 2. Subclustering of macrophages and DCs. (A) UMAP plot of liver macrophages and DCs showing a six-cluster distribution. (B) Feature plots of marker gene expression in six clusters. (C) Heatmap showing the marker genes of each cluster mentioned in (A). (D) Gene enrichment for classical cell type M1 and M2 in comparisons with the macrophage and DCs subsets in NAFL. (E) GO BP terms enriched in the indicated subset. The gene list is the marker genes of each cluster, colored by -log10 (P value), scaled by row.

• 

  Figure 3. Gene expression heterogeneity and gene expression upregulation by ApoA4 knockout in macrophages and DCs. (A) Fraction of each macrophage and DC subset in WT and KO mice. (B) Volcano plot showing DEGs in macrophages and DCs between WT and KO mice. Genes upregulated in WT are shown in blue. GO (C) and KEGG (D) analyses, respectively, with the gene set in (B), cut-off = 0.5. (E) GSEA plots of enriched pathways upregulated in KO macrophage and DC subsets. (F) Violin plots showing the expression of function-related genes in (C, D) across six groups.

• 

  Figure 4. Subclustering of granulocytes revealed more effector neutrophils in the ApoA4 knockout group. (A) UMAP plot of liver granulocytes showing a three-cluster distribution; fraction of each granulocyte subset in the WT and KO. (B) Feature plots of marker gene expression in three subsets. (C) Heatmap of the 6 most upregulated genes (decreased by P value) among the three granulocyte subsets. (D) GO BP terms enriched in the indicated subsets; the gene list is the maker genes of each cluster. (E) Volcano plot showing upregulated genes in WT and KO granulocytes. Average_log2 fold change > 0.7 & -log10(P value-adjust) > 2 are red; others are grey. (F) GSEA plot of 8 MSigDB hallmark gene sets. (G) Violin plots showing the expression of genes involved in neutrophil migration, response to interferon-gamma, and NET formation pathways among granulocyte subsets.

• 

  Figure 5. ScRNA-seq data were further validated by qRT–PCR and immune staining in enlarged samples. (A) Liver qRT–PCR analysis for representative DEGs found in scRNA-seq analysis in mice fed a HFD for 16 weeks (n = 5-6). The results are expressed as the mean ± SEM. Boxes include the values from the first to third quartiles, whereas horizontal bars represent the medians. Data were analyzed using two-way ANOVA. *P < 0.05, **P < 0.01, vs. WT. (B) NE (neutrophil elastase) immunostaining experiments in liver paraffin sections for HFD sections; zoomed in areas with quantification of positive cells from multiple views in liver sections. (C) Feature plot of the inflammatory gene Il1b in all immune cell (left), granulocyte (Gran, middle) and macrophage (Mφ, right) subsets. (D) IL-1β immunostaining experiments in murine liver paraffin sections with DAPI nuclear counterstain and the quantification of positive cells from multiple views in liver sections. Data are presented as the mean ± standard error of the mean. The statistical significance of differences was assessed with Student’s t test, ***P < 0.001. KO-GFP: KO-GFP group, which were ApoA4 knockout NAFL mice subjected to AAV-GFP adenoviral…

• 

  Figure 6. Comparison of gene profile alterations during human NAFLD pathogenesis with mouse NAFL regulated by ApoA4 deficiency. (A) Venn diagram showing the overlap of upregulated genes identified in the KO group with genes identified in a cohort study by Michael et al. (2020) (36) (P = 3.7e-11), as determined by a hypergeometric test. C0 was a set of differentially expressed genes between healthy controls and patients with mild liver steatosis, while C1 included genes upregulated in advanced liver steatosis compared with healthy controls. C2 consists of differentially expressed genes between patients with mild and advanced liver steatosis. C3 was upregulated in the KO group in this study. (B) Volcano plot showing differentially expressed genes in the KO group vs. WT group. Red dots indicate genes passing our P value-adjust and fold difference thresholds. Thirty-two genes that were crossed between C1 and C3 are marked with gene names, genes crossed between C2 and C3 are colored blue, and genes crossed among C1, C2 and C3 are colored red. (C) Comparison of the expression of genes of interest between NAFLD datasets (GSE63067) (left) obtained by liver microarrays and the data from this study…

• 

  Figure 7. Immunostaining for CD48, NE, CTSS and Galtectin3 in human NAFL liver. Bars = 10 μm (×100). *P < 0.05, **P < 0.01, ***P <0.001, vs. control.

• 

  Figure 8. Cell–cell interaction plots according to receptor–ligand analysis using CellChat and the effects of ApoA4 on the immune microenvironment in NAFL livers. (A) Signal interaction between macrophages and other cell types (left) and signal interactions between granulocytes and other cell types (right). (B) Average expression values of ligands shown in (A) in WT and KO NAFL livers based on the scRNA-seq dataset. (C) Violin plots showing values of receptors in (A) in WT and KO NAFL livers based on the scRNA-seq dataset. (D) Profiling of the effects of ApoA4 deletion on the immune microenvironment in NAFL and its potential mechanisms, as described in the discussion. Wfdc17+ granulocyte: 0-Gran-Wfdc17; Cxcl9+ macrophage: 4-Mφ-Cxcl9; Cxcl2+ macrophage: 4-Mφ-Cxcl2; NET: neutrophil extracellular trap.