Male C57BL/6 J mice, 8 weeks old (n = 16), were purchased from Nanjing GemPharmatech Co. Ltd., China, and housed ad libitum. A high-fat diet (HFD), comprising 60% energy from fat (Cat No. D12492 from Research Diets), was fed to all animals for 4 weeks, and then these animals were divided into two groups: HFD + vehicle (HV, n = 8) and HFD + Ashitaba (HA group, n = 8, 809/kg) groups. Our preliminary dose-ranging pre-experiments indicate that 800 mg/kg of Ashitaba can effectively improve glucose and lipid metabolism without any significant toxic side effects (data not shown). These mice were continuously fed with the HFD for another 8 weeks. Normal saline-dissolved Ashitaba dried powder, provided by Professor Zhengwu Wang, was gavaged to the animals. Mice were humanely euthanized utilizing a supra-lethal dose of sodium pentobarbital, administered at 90 mg/kg. The current animal study has been approved by the Animal Ethics Committee, Tongren Hospital, and Shanghai Jiao Tong University School of Medicine.

The serum was separated via centrifugation at a temperature of 4°C, followed by immediate storage in liquid nitrogen until analysis. Triglyceride (TG) and total cholesterol (TC) were assessed using a TG and TC detection kit (A110-2-1, A111-2-1, Nanjing Jiancheng Bioengineering Institute).

To conduct the intraperitoneal glucose tolerance test (IPGTT), the mice were intraperitoneally administered 2 g of dextrose/kg body weight after overnight fasting. Blood glucose levels were measured at various time intervals (0, 15, 30, 60, and 120 min) utilizing glucometers (Accu-Chek, Roche, Mannheim, Germany).

Liver histopathology was assessed from H&E and oil red staining slides and scanned utilizing a Pannoramic MIDI (3DHISTECH, Hungary).

RT-PCR was performed as described (18). Briefly, the total RNA from the liver was extracted using TRIzol^® Reagent (Invitrogen, Carlsbad, CA, United States) following the manufacturer’s instructions, followed by reverse transcription using Superscript III with random hexamer primers and 500 ng of total RNA. To amplify and detect the RNA, we used SYBR Premix Ex Taq Mixes on an ABI Prism 7,300 Sequence Detection System (Life Technologies, Foster City, CA, United States).

Transcriptomics was conducted by extracting total RNA from liver tissue samples (n = 4 in each group) using TRIzol^® Reagent (Invitrogen, Carlsbad, CA, United States). The concentration and quality of total RNA were evaluated using the NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States) and the 2100 Bioanalyzer system (Agilent Technologies, United States), respectively. A cDNA library was prepared using a TruSeqTM RNA sample preparation kit (Illumina, San Diego, CA, United States), and sequencing was performed using the Illumina NovaSeq 6000 system of CloudSeq Biotech Inc. (Shanghai, China).

The obtained clean data underwent differential gene expression analysis and functional enrichment analysis using the free online platform of Majorbio Cloud Platform. ¹ Differential gene expressions (DEGs) were analyzed with DESeq2 (19), based on |log₂FC| >1 and P_adjust <0.05 criteria. The identified DEGs were then subjected to Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis and Gene Ontology (GO) enrichment analysis, with a pathway deemed significantly enriched when P_adjust <0.05.

The liver samples (n = 6 in each group) were immediately placed on ice to preserve their integrity. To prepare the samples for LC-MS/MS analysis, 50 mg of each liver sample was utilized for the extraction of metabolites The LC-MS system was used to acquire all samples, following machine orders. The analytical conditions were as follows: UPLC column, Waters ACQUITY UPLC HSS T3 C18 (1.8 μm, 2.1 mm × 100 mm); column temperature, 40°C; flow rate, 0.4 mL/min; injection volume, 2 μL; solvent system, water (0.1% formic acid):acetonitrile (0.1% formic acid); and gradient program, 95:5 V/V at 0 min, 10:90 V/V at 11.0 min, 10:90 V/V at 12.0 min, 95:5 V/V at 12.1 min, and 95:5 V/V at 14.0 min.

The ProteoWizard program was used to convert the original data file to mzML format, and XCMS software was used to handle peak extraction, alignment, and retention time correction. To correct the peak area, the “SVR” method was utilized, and peaks with a deletion rate greater than 50% in each sample group were filtered. The metabolic identity information was obtained by scanning the laboratory’s self-built database and combining it with the public database and MetDNA. Finally, R software was used to conduct statistical analysis, including both univariate and multivariate statistical analyses. Univariate statistical analysis employed student’s t-test and variance multiple analysis, while multivariate statistical analysis included partial least squares discriminant analysis (PLS-DA), orthogonal partial least squares discriminant analysis (OPLS-DA), and principal component analysis (PCA).

For image analysis and data visualization, we utilized specialized software such as Zen Blue for microscopy adjustments, ImageJ for quantitative image assessments, Illustrator for high-resolution figure preparation, Photoshop for image editing, and GraphPad Prism for statistical plotting. The data points presented in the graphs reflect the mean values, accompanied by the standard error of the mean (SEM). To ensure robustness and reproducibility, all experiments were conducted with a minimum of three biological replicates and involved three to eight mice per experimental group. Statistical comparisons between groups were performed using the independent-samples t-test for normally distributed data, and non-parametric tests were applied for data sets that did not follow a normal distribution. A significance level (alpha) of 0.05 was employed to interpret statistical differences.

There was a significant decrease in body weight (12.6%) in the HA group following 8 weeks of Ashitaba treatment compared to the group without Ashitaba (p < 0.05, Figure 1A). Consistent with this, there was a reduction in liver weight in the HA group (0.92 ± 0.09 g versus 1.17 ± 0.21 g, p < 0.05, Figure 1B) following Ashitaba treatment. Notably, significantly reduced serum TG (82.9%) and TC (72.4%) levels were observed in the Ashitaba-treated group compared to those without Ashitaba (Figures 1C,D), which improved glucose tolerance in HFD mice treated with Ashitaba (Figures 1E,F). Additionally, we observed a significant reduction (32.9%) in liver fat droplets in the HA group using H&E staining (Figure 1G). Oil red staining also revealed a significant reduction in lipid levels within the liver of mice treated with Ashitaba (Figure 1H). The above data indicated that Ashitaba alleviated obesity and hepatic steatosis induced by a high-fat diet.

It was conducted as a comprehensive analysis by determining DGEs between HV and HA groups, utilizing RNA sequencing (RNA-seq) (Figure 2A), which was displayed with volcano plots. The liver gene expression profiles between HV and HA groups were markedly different, with a total of 429 DGEs identified (126 upregulated and 303 downregulated DEGs; Figure 2B). The log2 fold change hierarchical clustered heatmap illustrated the distinct transcriptional profiles between HV and HA (Figure 2C), using the top 20 DEGs (this number was utilized to enhance clarity). GO enrichment analysis of DEGs revealed several enriched pathways, including organic substance, developmental process, and anatomical structure development (Figure 3A). Furthermore, KEGG enrichment analysis of DEGs identified several enriched pathways, such as MAPK and AGE-RAGE signaling pathways in NAFLD and diabetic complications. In summary, the enrichment analysis of the liver transcriptome using 429 differentially expressed genes suggests that Ashitaba intervention in high-fat-fed mice may impact liver signaling pathways such as organic substances, MAPK, and AGE-RAGE.

To assess the discriminatory metabolites, principal component analysis (PCA) analysis was conducted, along with hierarchical clustering and correlation analyses. The quality control (QC) samples were tightly clustered within the PCA model diagram, indicating the stability of the instrument throughout the experiment. There were significant metabolic differences between the various groups (Figures 4A,B). The orthogonal partial least squares discriminant analysis (OPLS-DA) score plot demonstrated a distinct separation between the HV and HA groups (Figure 4C). Significantly upregulated metabolites were represented (red), downregulated metabolites (green), and insignificant metabolites (gray) (Figure 4D). The hierarchically clustered heatmap was presented (Figure 4E), using all 45 metabolites (25 upregulated and 20 downregulated metabolites). KEGG enrichment analysis of metabolic differences illustrated that enriched pathways encompass starch and sucrose metabolism, selenocompound metabolism, carbohydrate digestion, and absorption (Figure 4F). The 25 metabolites with the most significant differences are individually displayed using violin plots (Figure 4G). Liver non-targeted metabolomic analysis identified 45 differentially expressed metabolites, and further enrichment analysis of these metabolites suggested that Ashitaba treatment primarily influences metabolism-related signaling pathways.

The interactome network analysis of the transcriptome and metabolome was conducted to establish the connection of pathways via gene-metabolite interactions. Following Ashitaba treatment, some key pathways were altered during the development of NAFLD, such as FXR/RXR (20), NF-κB (21), AMPK (22), and PPAR (23) (Figure 5A). The interactome network illustrated that L-ornithine, L-kynurenine, and D-erythro-dihydrosphingosine contributed to the amelioration of NAFLD (Figure 5A). Thus, our data above suggest that Ashitaba ameliorates NAFLD by regulating FXR/RXR, NF-κB, AMPK, and PPAR signaling pathways. To quantify the alternation of these genes regulating NAFLD with Ashitaba, we employed qPCR to analyze these pathways. It was found that the expression of FXR in the HA group increased significantly (1.58-fold), and subsequently, the expression of CYP7A1 and SCD-1 decreased by 38.8% and 43.8%, respectively (Figure 5B). However, no differences in PPAR, AMPK, or NF-κB signaling pathways were noted between the two groups. Transcriptomic and metabolomic analyses of the liver, along with subsequent real-time quantitative PCR validation, suggest that Ashitaba may primarily ameliorate fatty liver through the FXR signaling pathway.