A Scd1-mediated metabolic alteration participates in liver responses to low-dose bavachin

Abstract

Hepatotoxicity induced by bioactive constituents in traditional Chinese medicines or herbs, such as bavachin (BV) in Fructus Psoraleae, has a prolonged latency to overt drug-induced liver injury in the clinic. Several studies have described BV-induced liver damage and underlying toxicity mechanisms, but little attention has been paid to the deciphering of organisms or cellular responses to BV at no-observed-adverse-effect level, and the underlying molecular mechanisms and specific indicators are also lacking during the asymptomatic phase, making it much harder for early recognition of hepatotoxicity. Here, we treated mice with BV for 7 days and did not detect any abnormalities in biochemical tests, but found subtle steatosis in BV-treated hepatocytes. We then profiled the gene expression of hepatocytes and non-parenchymal cells at single-cell resolution and discovered three types of hepatocyte subsets in the BV-treated liver. Among these, the hepa3 subtype suffered from a vast alteration in lipid metabolism, which was characterized by enhanced expression of apolipoproteins, carboxylesterases, and stearoyl-CoA desaturase 1 (Scd1). In particular, increased Scd1 promoted monounsaturated fatty acids (MUFAs) synthesis and was considered to be related to BV-induced steatosis and polyunsaturated fatty acids (PUFAs) generation, which participates in the initiation of ferroptosis. Additionally, we demonstrated that multiple intrinsic transcription factors, including Srebf1 and Hnf4a, and extrinsic signals from niche cells may regulate the above-mentioned molecular events in BV-treated hepatocytes. Collectively, our study deciphered the features of hepatocytes in response to BV insult, decoded the underlying molecular mechanisms, and suggested that Scd1 could be a hub molecule for the prediction of hepatotoxicity at an early stage.

Keywords: Bavachin; Hepatotoxicity; Lipid metabolism; Scd1; Single-cell RNA-Seq.

Graphical abstract

Fig. 1

Evaluation of adverse effects of different doses of bavachin (BV) on the murine liver. (A) Changes of indices in mice liver induced by different doses of BV. The liver index representing the percentage of liver weight in body weight, and two serum biochemical indices including alanine aminotransferase (ALT) and aspartate aminotransferase (AST) concentrations in the blood are presented. ^∗P < 0.05 based on Student's t-test between control and BV-administered group. U/L representing the unit of enzyme activity per liter of liquid. (B) Hematoxylin-eosin-stained histochemical images of murine liver under different doses of BV. The black arrows indicate injury areas in the liver. (C) Frequentist gamma model with benchmark response of 10% extra risk for the benchmark dose (BMD) and 95% lower confidence limit for the BMD lower confidence limit (BMDL).

Fig. 2

Decoding of lineages in healthy and bavachin (BV)-administered liver at the single-cell level. (A) Schematic of experimental design including cell dissociation and RNA sequencing. (B) Uniform manifold approximation and projection (UMAP) visualization of 15,051 cells in healthy and BV-administered livers. Cell numbers for distinct clusters are marked in parentheses. (C) Heatmap of representative marker genes for each cell type. The expressions of genes was scaled for presentation. Hepa: hepatocytes; EC: endothelial cells; SEC: sinusoidal endothelial cells; T: T cells; NK: natural killer cells; B: B cells; Mo/MoMF: monocyte or monocyte-derived macrophage; KC: Kupffer cells; DC: dendritic cells; HSC: hepatic stellate cells; scRNA-seq: single-cell RNA sequencing.

Fig. 3

Bavachin (BV)-induced feature alteration in hepatocytes. (A–C) Force-directed graph showing hepatocytes (Hepa) colored by condition (A), subpopulation (B), and pseudotime (C). The start cell and terminal states are indicated by hollow and solid arrowheads in Fig. 3C, respectively. (D) Proportion of the three subpopulations along the pseudotime. (E) Volcano plot for differentially expressed genes (DEGs) between Hepa2 and Hepa3. (F, G) Enriched terms of biological process (F) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway (G) of DEGs in Hepa2 (blue) and Hepa3 (dark blue). (H) DEGs involved in enriched functions related to lipid homeostasis in Hepa3. Orange squares: biological processes; orange diamonds: KEGG pathways. The sizes of the squares and diamonds represent the degree of enrichment. Genes are indicated by the fold changes (FC) in expression levels between Hepa3 and Hepa2. PPAR: peroxisome proliferator-activated receptors; TRP: transient receptor potential.

Fig. 4

The molecular basis of response in hepatocytes (Hepa) to low-dose bavachin (BV). (A) Albumin (ALB) and stearoyl-CoA desaturase 1 (SCD1) immunostaining in healthy and 200 mg/kg BV-induced murine livers. (B) The relative concentration of malondialdehyde (MDA), monounsaturated fatty acids (MUFA), polyunsaturated fatty acids (PUFA) and saturated fatty acids (SFA) in the livers of healthy and BV-administered mice. ^∗P < 0.05 based on the Student's t-test between the control and BV-administered group. (C) The schematic depicting the molecular changes and mechanism of steatosis in Hepa3 at the early stage of cytotoxicity induced by BV. DAPI: 40,6-diamidino-2-phenylindole.

Fig. 5

The gene expression regulation network in hepatocytes (Hepa) induced by low-dose bavachin (BV) treatment. (A) The degree of enrichment of each transcription factor (TF). The darker the blue color, the more target genes (TGs) are regulated by the TFs. (B) Gene regulation networks. TFs and TGs are squares and points colored by the fold changes (FC) of expression levels between Hepa3 and Hepa2. The sizes of the squares and points represent the connectivity degree. (C) Force-directed graphs for ferroptosis resistance-related genes colored by gene expression. Gene expressions were imputed by MAGIC algorithm.

Fig. 6

Tenascin signals received by hepatocytes were up-regulated with bavachin (BV) treatment. (A) Summary of changes of interaction strength among all clusters in BV-treated liver. The thickness of the edges indicates the strength of the corresponding interactions while the colors of the edges indicate whether the interactions were up-regulated (red) or down-regulated (blue). (B) Communication probability of significant ligand-receptor pairs where the ligands were expressed in non-parenchymal clusters and the receptors were expressed in hepatocyte clusters after BV treatment. Permutation test was applied for the significance of inter-cellular interactions predicted. (C) Mean expressions of Tnr in non-parenchymal clusters with or without BV treatment. P values were calculated based on the Wilcoxon rank-sum test between the control and BV-administered group. (D) Circle plots showing signals enriched by the ligand-receptor pairs displayed in Fig. 6B. In each circle, sources of the signals are drawn at the bottom and receptors are drawn at the top. The colors of the circles indicate distinct clusters and the colors of the links indicate individual signals. Hepa: hepatocytes; EC: endothelial cells; SEC: sinusoidal endothelial cells; T: T cells; NK: natural killer cells; B: B cells; Mo/MoMF: monocyte or monocytederived macrophage; KC: Kupffer cells; DC: dendritic cells; HSC: hepatic stellate cells.