Hepatic transcriptome discloses the potential targets of Xuefu Zhuyu Decoction ameliorating non-alcoholic fatty liver disease induced by high-fat diet

Abstract

Background and Aim

Xuefu Zhuyu decoction (XZD), a traditional Chinese medicinal formula, was firstly recorded in the Qing dynasty of ancient China and previously demonstrated to ameliorate hepatic steatosis. In the present study, the effects of XZD on non-alcoholic fatty liver disease (NAFLD) induced by high-fat diet (HFD) were evaluated in mice and the hepatic transcriptome was detected to disclose the potential mechanisms of XZD.

Experimental procedure

The effects of XZD (low- and high-dosage) on NAFLD induced by HFD for 16 weeks were evaluated. Obeticholic acid was used as control drug. Body weight, food intake and index of homeostatic model assessment for insulin resistance (HOMA-IR) were analyzed. Hepatic histology were observed in haematoxylin and eosin stained sections and quantified with NAFLD activity score (NAS). Lipid in hepatocytes was visualized by Oil red staining. Alanine aminotransferase (ALT) and hepatic triglyceride (TG) was measured. The hepatic transcriptom was detected with RNA-sequencing and validated with real-time polymerase chain reaction, western-blotting and hepatic quantitative metabolomics.

Results

XZD ameliorated hepatic histology of NAFLD mice, accompanied with decreasing fasting insulin, HOMA-IR, NAS, ALT and hepatic TG. The hepatic transcriptom of NAFLD was significantly reversed by XZD treatment, especially the genes enriched in the pathways of arachidonic acid metabolism, fatty acid degradation, cytokine-cytokine receptor interaction and extracellular matrix (ECM) -receptor interaction. The hepatic quantitative metabolomics analysis confirmed fatty acid degradation as the key targeting pathway of XZD.

Conclusions

XZD ameliorated NAFLD induced by HFD, which probably correlated closely to the pathways of fatty acid degradation.

Graphical abstract

Highlights

• •Xuefu Zhuyu Decoction ameliorates NAFLD.
• •Xuefu Zhuyu Decoction regulates fatty acid degradation.
• •Xuefu Zhuyu Decoction regulates arachidonic acid metabolism.
• •Xuefu Zhuyu Decoction regulates cytokine-cytokine receptor interaction.
• •Xuefu Zhuyu Decoction regulates extracellular matrix - receptor interaction.

List of abbreviations

ANOVA

analysis of variance

ALT

alanine aminotransferase

ALOX12

arachidonate 12-lipoxygenase 12S type

CCL

C–C motif chemokine ligand

COL1A1

collagen type I alpha I

COL3A1

collagen type III alpha I

COL6A3

collagen type VI alpha 3 Chain

COL27A1

collagen alpha-1 (XXVII) chain

CXCL

C-X-C motif chemokine ligand

CYP

Cytochrome P450

DEGs

differentially expressed genes

ECM

extracellular matrix

EHHADH

enoyl-CoA hydratase and 3-hydroxyacyl CoA dehydrogenase

EETs

epoxyeicosatrienoic acids

FPKM

fragments per kilo-base of gene per million mapped fragments

GO

Gene Ontology

H&E

haematoxylin and eosin

HFD

high-fat diet

H-XZD

high-dose of Xuefu Zhuyu Decoction

HETEs

hydroxyeicosatetraenoic acids

HOMA-IR

homeostatic model assessment for insulin resistance

HPETEs

hydroperoxyeicosatetraenoic acids

KEGG

Kyoto Encyclopedia of Genes and Genomes

L-XZD

low-dosage of Xuefu Zhuyu Decoction

NAFLD

non-alcoholic fatty liver disease

NAS

NAFLD activity scoring

NASH

non-alcoholic steatohepatitis

OCA

obeticholic acid

OPLS-DA

Orthogonal Partial Least Square Discriminant Analysis

Real time-PCR

real-time polymerase chain reaction

RNA-Seq

RNA sequencing

TG

triglyceride

TNC

tenascin C

UHPLC-Q-Orbitrap HRMS

ultra high performance liquid chromatography-Q exactive hybrid quadrupole orbitrap high-resolution accurate mass spectrometric

UPLC-MS/MS

ultra-performance liquid chromatography coupled to tandem mass spectrometry

VIP

importance in projection

XZD

Xuefu Zhuyu decoction

1. Background

Non-alcoholic fatty liver disease (NAFLD) has been one of the most common chronic liver diseases in the world, with wide histological spectrum including non-alcoholic fatty liver, non-alcoholic steatohepatitis (NASH) and fibrosis.¹ Although many drugs for NAFLD are undergoing the clinical trials, the therapeutic approach of NAFLD is still limited. Weight loss achieved by hypocaloric diet and exercise is the primary strategy for NAFLD treatment. By now, only pioglitazone and vitamin E are recommended to treat biopsy-proven NASH.¹

As a complementary strategy, traditional Chinese medicine is commonly used to treat NAFLD in China. According to the theory of traditional Chinese medicine, stagnation of Qi and blood caused by over-nutritious diets is one of the pathogenic factors of NAFLD. Activating blood circulation and smoothing out the stagnation are recommended by the Branch of Gastrointestinal Diseases, China Association of Chinese Medicine as the primary principles of NAFLD treatment.² Xuefu Zhuyu decoction (XZD), a traditional Chinese medicinal formula, is firstly recorded in a medical classic < Yi Lin Gai Cuo>, written in the Qing dynasty of ancient China, consisting of eleven medicinal herbs (listed in Table 1). It is the representative medical formula used to activate blood circulation and smooth out the stagnation in traditional Chinese medicine and has been widely used in the treatment of coronary heart disease in China.³ In the recent decade, the effects of XZD on NAFLD were also realized and reported. XZD improved the ultrasound manifestations of liver and the serum markers of liver function in patients with NAFLD.⁴ XZD also ameliorated obesity and hepatic steatosis in metabolically stressed mice^5,6 and reduced the hepatic neutrophil chemotaxi.⁶ Previously, our group demonstrated that XZD and its fragments ameliorated hepatic steatosis and inflammatory infiltration in NAFLD mice induced by high-fat diet (HFD, 60% energy from fat) for 16 weeks.⁷ This model replicates the metabolic parameters (e.g. obesity, hyperinsulinemia and glucose intolerance) and hepatic pathological outcomes including hepatic steatosis and inflammation observed in human NAFLD.⁸ In the present study, the effects of XZD on NAFLD induced by HFD were confirmed in mouse and the transcriptome-based mechanisms of XZD were investigated to indicate the clues of the action targets of XZD.

Table 1.

Composition of XueFu ZhuYu Decoction (XZD).

Pharmaceutical name	Botanical name	Plant part use	Chinese name	Origin	w/w %
Angelicae Sinensis Radix	Angelica sinensis (Oliv.) Diels.	root, dried	Dang gui	Gan su Province, China	11.8
Rehmanniae Radix	Rehmannia glutinosa Libosch.	root, dried	Di huang	Henan, Province, China	11.8
Persicae Semen	Prunus persica (L.) Batsch.	ripe seed, dried	Tao ren	Shan dong, Province, China	15.8
Carthami flos	Carthamus tinctorius L.	flower, dried	Hong hua	Xin jiang, Province, China	11.8
Glycyrrhizae Radix Et Rhizoma	Glycyrrhiza uralensis Fisch.	root and rhizome, dried	Gan cao	Xin jiang, Province, China	3.9
Aurantii Fructus	Citrus aurantium L.	unripe fructu, dried	Zhi qiao	Jiang xi, Province, China	7.9
Paeoniae Radix Rubra	Paeonia lactiflora Pall.	root, dried	Chi shao	Nei meng gu, Province, China	7.9
Bupleuri Radix	Bupleurum chinense DC.	root, dried	Chai hu	He bei Province, China	3.9
Chuanxiong Rhizoma	Ligusticum chuanxiong Hort.	rhizome, dried	Chuan xiong	Si chuan Province, China	6.6
Platycodonis Radix	Platycodon grandiflorum (Jacq.)A. DC.	root, dried	Jie geng	An hui Province, China	6.6
Cyathulae Radix	Cyathula officinalis Kuan.	root, dried	Chuan niu xi	Si chuan Province, China	11.8

The botanical names of composition of XueFuZhuYu Decoction are from The Pharmacopoeia of the People's Republic of China (2020 Edition). The botanical names have been updated with www.theplantlist.org.

2. Methods

2.1. Preparation of XZD

The qualified medicinal materials included in the formula of XZD (Table 1) were commercially obtained from Shanghai Kangqiao Traditional Chinese Medicine Pieces Co., Ltd. (Shanghai, China). The medicinal materials were extracted with water (8 times of the total mass of raw materials) at 100 °C for 1.5h, twice. Collect and concentrate the aqueous extracts to get the high-dose of XZD (H-XZD, containing crude drug 2.02 g/ml, 20 ml/kg body weight). Low-dosage of XZD (L-XZD, containing crude drug 1.01 g/ml, 20 ml/kg body weight) was obtained by dilution of H-XZD. Based on the traditional dosage of XZD recorded in <Yi Lin Gai Cuo>, the equivalent dose on mouse was calculated as containing crude drug 1.01 g/ml, 20 ml/kg body weight (the L-XZD).⁹ A voucher specimen (TH-20200706) was deposited in the Institute of Liver Diseases, Shuguang Hospital, for the future comparison.

The quality control of XZD was described as the concentration of the main active components, including catalpol, rehmannioside D, amygdalin, hydroxysafflor yellow A, paeoniflorin, ferulic acid, liquiritin, naringin, neohesperidin, kaempferol, saikosaponin A and aikosaponin D.

2.2. Chemical profile analysis of XZD

The main active components of XZD were identified by using ultra high performance liquid chromatography-Q exactive hybrid quadrupole orbitrap high-resolution accurate mass spectrometric (UHPLC-Q-Orbitrap HRMS, Thermo Fisher Scientific Inc., Grand Island, NY, USA) as described previously.¹⁰ The reference standards including catalpol, rehmannioside D, amygdalin, hydroxysafflor yellow A, paeoniflorin, ferulic acid, liquiritin, naringin, neohesperidin, kaempferol, saikosaponin A and saikosaponin D are products of Shanghai R&D Center for Standardization of Chinese Medicines (Shanghai, China). The ions of the target compounds including catalpol (characteristic component of Rehmanniae Radix), rehmannioside D (characteristic component of Rehmanniae Radix), amygdalin (characteristic component of Persicae Semen), hydroxysafflor yellow A (characteristic component of Carthami flos), paeoniflorin (characteristic component of Paeoniae Radix Rubra), ferulic acid (characteristic component of Angelicae Sinensis Radix and Chuanxiong Rhizoma), liquiritin (characteristic component of Glycyrrhizae Radix Et Rhizoma), naringin (characteristic component of Aurantii Fructus), neohesperidin (characteristic component of Aurantii Fructus), kaempferol (characteristic component of Carthami flos), saikosaponin A (characteristic component of Bupleuri Radix) and saikosaponin D (characteristic component of Bupleuri Radix) were extracted for quantitative and qualitative analysis.

2.3. Animals and treatment

Male C57BL/6 mice, 8 weeks old, obtained from Shanghai Experimental Animal Center of Chinese Academy of Sciences (Shanghai, China), were accommodated in a pathogen-free, temperature-controlled (20–26 °C) environment under a 12h light/dark cycle in the Division of Animal Resources, Shanghai University of Traditional Chinese Medicine. The animal experiment was conducted according to the protocols approved by the animal studies ethics committee of Shanghai University of Traditional Chinese Medicine (approval number: IACUC 2020-0018).

Mice were divided into control (n = 10), HFD (n = 10), L-XZD (n = 10), H-XZD (n = 10) and obeticholic acid (OCA, n = 10) group randomly. The mice in control group were fed with control diet (D12450B, 10% kcal from fat; Research Diets Inc., NJ, USA). The others were fed with HFD (D12492, 60% kcal from fat; Research Diets Inc., NJ, USA). At the beginning of the 13th week, the mice were administrated with L-XZD (containing 1.01 g of crude drug per ml, 20 ml/kg body weight, daily), H-XZD (containing 2.02 g of crude drug per ml, 20 ml/kg body weight, daily) and OCA (10 mg/kg per day, MedChemExpress CO., Ltd. NJ, USA),¹¹ by gavage, respectively. The other mice were administrated with equal volume of double distilled water. In the end of the 16th week, the liver tissue and serum were harvested for assay.

2.4. Histology evaluation

Liver tissue was fixed with formalin and embedded in paraffin, which was ready for section. Haematoxylin and eosin (H&E) staining was performed on liver sections (4 μm thick) with commercially available kit (Nanjing Jiancheng Institute of Bio Engineering, Inc., Nanjing, China)) to visualize the histology. The NAFLD activity scoring (NAS) system including steatosis, lobular inflammation and hepatocellular ballooning (Supplementary Table 1)¹² was employed to evaluate the severity of liver disease. The histology scoring was performed by three pathologists independently.

Oil red (Sigma, MO, USA) staining was performed on the liver tissue embedded in Optimal Cutting Temperature medium (Sakura Finetek, Torrance, CA) to visualize the lipid droplets in the hepatocytes.

2.5. Measurement of glucose and insulin, and calculation of homeostatic model assessment for insulin resistance

Fasting glucose and insulin in serum were detected with QuantiChrom™ GlucoseAssay kit (BioAssay Systems, CA, USA) and Ultra Sensitive Mouse Insulin enzyme-linked immunosorbent assay (ELISA) Kit (Crystal Chem, CA, USA) respectively. Homeostatic model assessment for insulin resistance (HOMA-IR) was calculated according to the formula HOMA-IR = [insulin (m IU/L) ∗ glucose (mmol/L)]/22.5.¹³

2.6. Quantification of alanine aminotransferase and triglyceride

The activity of alanine aminotransferase (ALT) in serum and the concentration of triglyceride (TG) in liver were determined with the commercial kits (ALT assay kit, Nanjing Jiancheng Bioengineering Institute, Nanjing, China; TG commercial kit, Dongou Diagnostic Products Co., Zhejiang, China). The serum and homogenate of liver tissue was prepared as described previously.¹⁴

2.7. RNA sequencing and data analysis

Liver samples from control, HFD and H-XZD group (four samples per group) were chosen randomly for RNA sequencing (RNA-Seq), which was a contract service offered by Shanghai OE Biotech. Co., Ltd. (Shanghai, China). The protocol of extraction of RNA and strand-specific RNA library preparation was described in detail in Supplementary file S1. After validation, the RNA libraries were sequenced on the Illumina sequencing platform (Illumina NovaSeq 6000, Illumina, Inc., CA, USA) and 150bp paired-end reads were generated.

Raw reads of RNA-seq were processed with Trimmomatic,¹⁵ removing the adaptor sequences, empty reads and low quality sequences to obtain the clean reads for gene mapping by using HISAT2.¹⁶ The read counts of each gene were obtained by HTSEQ-COUNT.¹⁷ Fragments per kilo-base of gene per million mapped fragments (FPKM) values from each gene were calculated using CUFFLINKS.¹⁸

A cut off p-value <0.05 and fold-Change >2 or <0.5 were employed to determine the differentially expressed genes (DEGs) between groups by using DESeq, R package functions estimateSizeFactors and nbinomTest. The pattern of genes expression was determined by using Hierarchical cluster analysis. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment and Gene Ontology (GO) functional classification for DEGs were performed on https://cloud.oebiotech.cn/task/detail/array_enrichment.

2.8. Real-time polymerase chain reaction (real-time PCR)

Hepatic total RNA was isolated (RNA isolation kit, Sangon Biotech Inc., Shanghai, China). The reverse transcription reaction was performed by using a cDNA synthesis kit (Bio-Rad, CA, USA) to get the complementary DNA (cDNA). Real-time PCR was performed by using PCR kit (TB Green™ Premix Ex Taq™; TaKaRa Bio Inc., Japan) on Applied Biosystems ViiA7 (Thermo Fisher Scientific, CA, USA) to detect the target genes expression. The primers (synthesized by Sangon Biotech Inc., Shanghai, China) of the target genes were listed in Supplementary Table 2. The parameters of real-time PCR was as follow: pre-denaturation for 30 s at 95 °C, 95 °C for 5 s and 60 °C for 30 s, 40 cycles. The relative expression levels of the target genes were calculated by delta-delta Ct method. After being corrected by β-Actin, the data of genes expression was presented as fold changes relative to that in the control group.

2.9. Western-blot analysis

Western blot analysis was performed following the protocol described previously¹⁴ to evaluate the protein expression of arachidonate 12-lipoxygenase 12S type (ALOX12), cytochrome P450 (CYP) 4 (CYP4A), enoyl-CoA hydratase and 3-hydroxyacyl CoA dehydrogenase (EHHADH), interleukin (IL) 1 receptor type I (IL-1R1) and C–C motif chemokine ligand (CCL) 2 in the liver tissue. The protein expression of β-actin was used as internal reference. The antibody of ALOX12 (PA5-78760), EHHADH (PA5-37009) and CCL2 (MA5-17040) were the products of Thermo Fisher Scientific Inc. (MA, USA). The antibody of CYP4A (Ab140635) and IL-1R1 (Ab8154) were purchased from Abcam Inc. (MA, USA). The antibody of β-actin (66009-1-Ig) was the product of Proteintech Group, Inc. (IL, USA). The goat anti-rabbit IgG and goat anti-mouse IgG (Beijing Dingguo Changsheng Biotechnology CO.LTD, Beijing, China) were used as the secondary antibody. The blots were scanned with Fully Automatic Digital Gel/Chemiluminescence Image Analysis System (4600SF, Tanon, Shangha, China). The density of the bands was analyzed with ImageJ software.

2.10. Metabolomics analysis

The hepatic quantitative metabolomics was a contract service provided by Metabo-Profile Biotechnology (Shanghai) Co. Ltd. The metabolites in liver tissue were quantified by the ultra-performance liquid chromatography coupled to tandem mass spectrometry (UPLC-MS/MS) system (ACQUITY UPLC-Xevo TQ-S, Waters Corp., Milford, MA, USA). All of the standards were obtained from Sigma-Aldrich (St. Louis, MO, USA), Steraloids Inc. (Newport, RI, USA) and TRC Chemicals (Toronto, ON, Canada). The internal standards were used to monitor analytical variations during the entire sample preparation and analysis processes.

The raw data files generated by UPLC-MS/MS were processed using the TMBQ software (v1.0, Metabo-Profile, Shanghai, China) to perform peak integration, calibration, and quantification for each metabolite. Orthogonal Partial Least Square Discriminant Analysis (OPLS-DA), univariate analysis, and pathway enrichment analysis were performed on the iMAP platform (http://imap.metaboprofile.cloud/) developed by Metabo-Profile Biotechnology (Shanghai) Co. Ltd. The differential metabolites selection was conducted by using multi-dimensional statistics and univariate Statistics (student T-test or Mann-Whitney U test, depending on the normality of data and homogeneity of variance). For multi-dimensional statistics, the differential metabolites meeting variable importance in projection (VIP)>1 were selected. For univariate statistics, the differential metabolites meeting P < 0.05 and |log2Fold Change| >= 0 were selected. The potential biomarkers were obtained by getting intersection of the differential metabolites from univariate statistics and multi-dimensional statistics. The selected biomarkers were used in pathway Enrichment analysis.

2.11. Statistical analysis

The NAS data was presented as the medium and the individual scoring data. The others were presented as the means ± standard deviation. One-way analysis of variance (ANOVA) followed by Tukey's multiple comparisons test was used to calculate the statistical significance among groups. For the nonparametric data, the Kruskal-Wallis H-test was used to analyse the significant difference among more than two groups. Significance was accepted at the level of P < 0.05.

3. Results

3.1. Chemical components in XZD

Twelve active compounds were identified in XZD, including catalpol (3.66 μg/ml, [M + HCOO]-m/z 407.1184), rehmannioside D (12.99 μg/ml, [M + HCOO]-m/z 731.2241), amygdalin (52.43 μg/ml, [M + HCOO]-m/z 502.1555), hydroxysafflor yellow A (166.15 μg/ml, [M − H]-m/z 611.1607), paeoniflorin (81.39 μg/ml, [M + HCOO]-m/z 525.1603), ferulic acid (5.66 μg/ml, [M − H]-m/z 193.0495), liquiritin (14.84 μg/ml, [M − H]-m/z 417.118), naringin (126.89 μg/ml, [M − H]-m/z 579.1708), neohesperidin (65.45 μg/ml, [M − H]-m/z 609.1814), kaempferol (0.05 μg/ml, [M − H]-m/z 285.0394), saikosaponin A (1.59 μg/ml, [M + HCOO]-m/z 825.4631, Retention Time: 39.82 min) and saikosaponin D (6 μg/ml, [M + HCOO]-m/z 825.4631, Retention Time: 41.05 min) (Fig. 1).

Fig. 1.

Preliminary chemical analysis of Xuefu Zhuyu Decoction (XZD)

A, The total ion chromatogram (TIC) and extracted ion chromatograms (EIC) of the mixture of reference standards in negative ion mode by UHPLC-Q-Orbitrap HRMS. (1) Catalpol, (2) Rehmannioside D, (3) Amygdalin, (4) Hydroxysafflor yellow A, (5) Paeoniflorin, (6) Ferulic acid, (7) Liquiritin, (8) Naringin, (9) Neohesperidin, (10) Kaempferol, (11) Saikosaponin A, (12) Saikosaponin D. B, The total ion chromatogram (TIC) and extracted ion chromatograms (EIC) of XZD in negative ion mode by UHPLC-Q-Orbitrap HRMS. (1) Catalpol, (2) Rehmannioside D, (3) amygdalin, (4) Hydroxysafflor yellow A, (5) Paeoniflorin, (6) Ferulic acid, (7) Liquiritin, (8) Naringin, (9) Neohesperidin, (10) Kaempferol, (11) Saikosaponin A, (12) Saikosaponin D.

3.2. XZD ameliorates NAFLD and insulin resistance

With HFD-feeding, the body weight increased significantly comparing to that of control mice. With treatment of L-XZD or H-XZD, the body weight decreased obviously comparing to that in HFD group. OCA had no effect on the body weight (Fig. 2A). The food intake of HFD-fed mice with or without treatment was all lower than that of control mice, but there's no difference in food intake among the HFD, L-XZD, H-XZD and OCA groups (Fig. 2A).

Fig. 2.

Xuefu Zhuyu Decoction (XZD) ameliorates NAFLD and insulin resistance

A, Body weight and food intake. B Fasting blood glucose, insulin and homeostatic model assessment of insulin resistance (HOMA-IR). C, Haematoxylin-eosin staining for liver sections (200 times of magnification). D, Oil red staining for liver sections (200 times of magnification). E, NAFLD activity score (NAS), activity of alanine aminotransferase (ALT) and hepatic triglyceride (TG). ∗P < 0.05, ∗∗P < 0.01. HFD, high-fat diet, L-XZD, low-dosage of Xuefu Zhuyu Decoction, H-XZD, high-dosage of Xuefu Zhuyu Decoction, OCA, obeticholic acid.

The fasting blood glucose level in HFD group was significantly higher than that in control group, while XZD and OCA both had no effects on the fasting blood glucose. The fasting insulin and HOMA-IR were both increased with HFD-feeding comparing to that of control mice. With treatment of low- or high-dosage of XZD, the fasting insulin and HOMA-IR decreased comparing to that in HFD group. OCA had no effects on fasting insulin and HOMA-IR (Fig. 2B).

The images of H&E staining showed that obvious macrovesicular steatosis, ballooning degeneration of hepatocyte and inflammatory infiltration was observed in the liver of mice in HFD group. With treatment of L-XZD, H-XZD or OCA, these pathological changes in the liver tissue were ameliorated (Fig. 2C). Consistently, the NAS and serum ALT level increased obviously in HFD group comparing that in control. The medium of NAS in HFD group was more than 5, indicating the diagnosis o f NASH. With treatment of L-XZD, H-XZD or OCA, the NAS and ALT levels were all decreased significantly comparing to that in HFD group (Fig. 2E).

Lipids deposition was determined by Oil red staining and TG quantification in liver tissue. In HFD group, the hepatocytes were swelled with huge red-stained lipid droplets and the nucleus was squeezed to the side. At the same time, significatnly increased concentration of hepatic TG was detected (Fig. 2D and E). In L-XZD, H-XZD or OCA group, less and smaller red-stained lipid droplets in the hepatocytes and decreased hepatic TG contents were found (Fig. 2D and E).

3.3. Quality control of RNA-seq

For RNA-seq, there're 83.9 G of CleanData totally obtained and 6.68–7.28 G CleanData got in each sample. The average content of GC content was 50.42%. The Q30 base distribution was between 95.39 and 95.93% (Supplementary file S2). There're 585.41 M clean reads totally obtained. And on average, 98.08% of these clean reads were mapped to the reference genome (Supplementary file S3).

3.4. The profile of hepatic transcriptome

With principal component analysis, it was indicated that the profile of hepatic transcriptome of mice in control, HFD and H-XZD group was different (Fig. 3A). Simultaneously, the sample to sample distance clustering analysis disclosed that samples in H-XZD group were far from that in control and HFD group. It was indicated that treatment with XZD regulated the hepatic transcriptome of NAFLD mice remarkably. (Fig. 3B).

Fig. 3.

Profile of hepatic transcriptome

A, Principal component analysis (PCA) of transcriptome data. B, Clustering analysis of sample to sample distance. C, Volcano figures of differentially expressed genes (DEGs). D, Amount of DEGs. HFD, high-fat diet, H-XZD, high-dosage of Xuefu Zhuyu Decoction, FC, fold changes.

3.5. Identification of DEGs of HFD versus control, and XZD versus HFD

Comparing to control, there're totally 588 genes was significantly differentially expressed in the liver tissue in NAFLD induced by HFD, in which, 378 genes were up-regulated and 210 genes were down-regulated. Comparing to HFD group, there're totally 710 genes was significantly differentially expressed in the liver tissue in H-XZD group, in which, 270 genes were up-regulated and 440 genes were down-regulated (Fig. 3C and D).

3.6. XZD reversed the hepatic genes expression induced by HFD

There're totally 245 genes regulated by HFD and XZD simultaneously, in which, 233 genes were regulated by HFD and XZD in the opposite direction. (Fig. 4A and B, Supplementary Fig. 1).

Fig. 4.

The genes induced by HFD and reversed by XZD

A, Intersection of the differentially expressed genes (DEGs). B, Heatmap of top 20 genes up or down-regulated by HFD and reserved by H-XZD. C, Gene ontology (GO) analysis of total genes induced by HFD and reserved by H-XZD. D, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of total genes induced by HFD and reserved by H-XZD. E, Relative mRNA expression of genes enriched in top KEGG pathways. #p < 0.05, vs control, ##p < 0.01, vs control, ∗p < 0.05, vs HFD, ∗∗p < 0.01, vs HFD. ECM, extracellular matrix, HFD, high-fat diet, H-XZD, high-dosage of Xuefu Zhuyu Decoction.

GO enrichment analysis indicated that the genes regulated by HFD and reversed by H-XZD were enriched in the biological process of chemokine-mediated signaling pathway, chemotaxis, cell chemotaxis, acyl-CoA metabolic process, response to mechanical stimulus, inflammatory response, response to stilbenoid, cellular response to tumor necrosis factor, cellular response to IL-1 and positive regulation of extracellular signal-regulated kinase 1 and 2 cascade (Fig. 4C, Supplementary file S4).

KEGG pathway enrichment analysis suggested that the genes regulated by HFD and reversed by H-XZD were enriched in the pathways of arachidonic acid metabolism, cytokine-cytokine receptor interaction, fatty acid degradation, extracellular matrix (ECM) -receptor interaction, retinol metabolism, chemokine signaling pathway, inflammatory mediator regulation of transient receptor potential channels, peroxisome proliferator-activated receptor signaling pathway, and so on. (Fig. 4D, Supplementary file S5).

3.7. XZD regulates the genes enriched in arachidonic acid metabolism, cytokine-cytokine receptor interaction, fatty acid degradation and ECM-receptor interaction pathway

The expression of genes enriched in the top pathways in KEGG analysis, including arachidonic acid metabolism, cytokine-cytokine receptor interaction, fatty acid degradation and ECM-receptor interaction, was validated by real time PCR.

In the liver tissue of NAFLD mice induced by HFD, the expression of genes enriched in arachidonic acid metabolism and fatty acid degradation, including ALOX12, CYP4A (CYP4A12A, CYP4A12B, CYP4A31, CYP4A10, CYP4A14) and EHHADH was down-regulated comparing to that in control, which was reversed by H-XZD treatment. Meanwhile, the expression of genes enriched in cytokine-cytokine receptor interaction and ECM-receptor interaction pathway, including IL-1R1, CCL2, C-X-C motif chemokine ligand (CXCL) 1, CXCL2, collagen type I alpha I (COL1A1), collagen type III alpha I (COL3A1), collagen alpha 1 (XXVII) chain (COL27A1), collagen type VI alpha 3 chain (COL6A3) and tenascin C (TNC) was up-regulated comparing to that in control, which was reversed by H-XZD treatment (Fig. 4E).

Furthermore, the protein expression of these genes in liver tissue was detected by western-blot (Fig. 5). Consistently, the protein expression of CYP4A and EHHADH decreased in HFD group comparing to that in control and was reversed by H-XZD treatment. The protein expression of IL-1R1 and CCL2 increased in HFD group comparing to that in control and was reversed by H-XZD treatment. For ALOX12, although there's no significant difference between control and HFD group, but, consistent with the data of real-time PCR, the protein expression of ALOX12 was increased with H-XZD treatment.

Fig. 5.

The protein expression of genes enriched in the top pathways in Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis.

Western-blotting images and the relative expression of arachidonate 12-lipoxygenase 12S type (ALOX12), cytochrome P450 4A (CYP4A), enoyl-CoA hydratase and 3-hydroxyacyl CoA dehydrogenase (EHHADH), interleukin 1 receptor type I (IL1-R1), and C–C motif chemokine ligand 2 (CCL-2) in liver tissue. B-D, The hepatic quantitative metabolomic analysis. B, The Orthogonal Partial Least Square Discriminant Analysis (OPLS-DA) 2D score plot. C, Pathway Enrichment Analysis Barplot Using Pathway-associated metabolite sets. The pathways associated with fatty acid degradation are highlighted in red. D, Quantification of hepatic palmitoylcarnitine. ∗P < 0.05, ∗∗P < 0.01. HFD, high-fat diet, H-XZD, high-dosage of Xuefu Zhuyu Decoction.

3.8. Quantitative metabolomics supports fatty acid degradation pathway as the key targeting pathway of XZD

To verify the results of hepatic transcriptome, the hepatic quantitative metabolomics was conducted. There're 212 metabolites were detected, belonging to amino acids, benzenoids, benzoic acids, bile acids, carbohydrates, carnitines, fatty acids, imidazoles, indoles, nucleotides, organic acids, peptides, phenols, phenylpropanoic acids, phenylpropanoids, pyridines and short chain fatty acids (Fig. 6A). As shown in the OPLS-DA score plot, the profile of hepatic metabolomics was different between HFD and H-XZD group (Fig. 6B). There're 24 metabolites identified as the biomarkers to distinguish group HFD and H-XZD (Supplementary Fig. 6). Pathway enrichment analysis showed that these biomarkers enriched in the pathways of histidine metabolism, ketone body metabolism, β-oxidation of very long chain fatty acids, α-linolenic acid and linoleic acid metabolism, ubiquinone biosynthesis, oxidation of branched chain fatty acids, mitochondrial β-oxidation of short chain saturated fatty acids, β-alanine metabolism, aspartate metabolism, fatty acid biosynthesis, propanoate metabolism, fatty acid metabolism, pyrimidine metabolism and tryptophan metabolism. In these metabolism pathways, there're 4 pathways related to fatty acid degradation, the one of the top KEGG pathways that the genes reversed by XZD enriched in (Fig. 4, Fig. 6C). When we looked deep into the pathway of fatty acid degradation, palmitoylcarnitine was found as an important intermediate upstream of EHHADH (Supplementary Fig. 2). And in the quantitative metabolomics analysis, the content of palmitoylcarnitine was consistently increased in HFD group comparing to that in control, which was reversed by XZD treatment (Fig. 6D).

Fig. 6.

Hepatic quantitative metabolomics.

A, Class statistics of metabolites, pie plot of each group and bar plot of each sample. B, The Orthogonal Partial Least Square Discriminant Analysis (OPLS-DA) 2D score plot. C, Pathway Enrichment Analysis Bar plot Using Pathway-associated metabolite sets. The smaller the P-value, the more differential metabolites enriched in this pathway compared to other pathways. The pathways associated with fatty acid degradation are highlighted in red. D, Quantification of hepatic palmitoylcarnitine. ∗P < 0.05, ∗∗P < 0.01. HFD, high-fat diet, H-XZD, high-dosage of Xuefu Zhuyu Decoction.

4. Discussion

In the present study, the effects of XZD in different dosage were evaluated in NAFLD mice induced by HFD. Both low- and high-dosage of XZD presented significant effects on NAFLD including reducing hepatic TG content, improving the hepatic histology and insulin resistance. The data of RNA-Seq disclosed that XZD obviously regulated the hepatic transcriptome of NAFLD mice and suggested the potential pathways which related to the pharmacological effects of XZD closely.

In the present study, XZD restored the hepatic expression of CYP4 enzymes which connecting the metabolism of fatty acid and arachidonic acid. CYPs are one of the major proteins involved in oxidative biotransformation.¹⁹ Enzymes in the CYP4 family are involved in the metabolism of fatty acids, xenobiotics, therapeutic drugs, and signaling molecules, including eicosanoids, leukotrienes, and prostanoids.²⁰ In the present study, the gene expression of CYP4A12A, CYP4A12B, CYP4A31, CYP4A10 and CYP4A14 involved in fatty acid degradation was down-regulated in the liver of NAFLD mice, which was recovered with XZD treatment. The CYP4 mainly catalyzes the oxidation of fatty acids to eliminate the excess free fatty acids from the body and the subsequent produce of energy in the mitochondria (Supplementary Fig. 2). Reduction of the expression of CYP4 family was associated with fat accumulation in liver.²¹ Non-steroidal anti-inflammatory drugs caused hepatic steatosis in mice, which were associated with significant down-regulation of the gene expression of CYP4A12 in liver tissues.²¹ EHHADH is peroxisomal L-bifunctional enzyme involved in the classical peroxisomal fatty acid β-oxidation pathway²² (Supplementary Fig. 2). In the present study, similar to CYP4A enzymes, gene and protein expression of EHHADH were both down-regulated in the liver tissue of NAFLD mice and recovered with XZD treatment. It was indicated that XZD probably excited the blunt oxidation of fatty acids in NAFLD. Interestingly, it was confirmed by the quantitative hepatic metabolomics analysis. In the metabolism pathways that metabolic biomarkers enriched in, there're 4 pathways related to fatty acid degradation (Fig. 6C). And the important intermediate upstream of EHHADH, palmitoylcarnitine, was consistently increased in HFD group comparing to that in control, which was reversed by XZD treatment (Fig. 6D).

Arachidonic acid is a 20-carbon chain fatty acid and belongs to the omega-6 (n-6) polyunsaturated fatty acids, which is found to be dysregulated in NAFLD.²³ Arachidonic acid is metabolized by three different enzyme families including cyclooxygenase to produce the prostaglandins, lipoxygenase to produce mid chain hydroxyeicosatetraenoic acids (HETEs), lipoxins, leukotrienes and hydroperoxyeicosatetraenoic acids (HPETEs), and CYPs to produce epoxyeicosatrienoic acids (EETs) and 20-HETE.²⁴ 20-HETE promotes oxidative stress and inflammation in endothelial cells by activating nicotinamide adenine dinucleotide phosphate oxidase system and nuclear factor-кB pathway, while EETs have potent anti-inflammatory and protective effects.²⁵ Targeted disruption of epoxide hydrolase 2 to increase EET levels significantly attenuated hepatic steatosis, inflammation and injury induced by methionine-choline deficient diet or high fat diet.²⁵ Since CYP4 enzymes catalyze the production of to both 20-HETE and EETs from arachidonate (Supplementary Fig. 3), the effects of XZD on 20-HETE and EETs need more precise investigation. ALOX12 is a member of the lipoxygenase family and catalyzes the metabolism of arachidonic acid to 12-HPETE and the ultimate production of 12-HETE (Supplementary Fig. 3). ALOX12-12-HETE pathway was demonstrated to be enhanced in NAFLD.²⁶ In the present study, gene expression of ALOX12, CYP4A12A, CYP4A12B, CYP4A31 and CYP4A10 were all down-regulated in HFD group, which was recovered with XZD treatment. Although the protein expression of ALOX12 was not significantly changed in HFD group comparing to that in control in our study, the effect of XZD on ALOX12 protein was confirmed. The effects of XZD on arachidonic acid need to be investigated thoroughly in our future study.

XZD reduced the hepatic expression of cytokines and ECMs induced by HFD. In the present study, the hepatic expression of genes enriched in cytokine-cytokine receptor interaction pathway including IL1R1, CCL2, CXCL1 and CXCL2 and ECM-receptor interaction pathway including COL1A1, COL3A1, COL27A1, TNC and COL6A3 were all increased with HFD feeding, which was reversed with XZD treatment. And consistently, we observed the similar changes on the protein expression of IL1R1 and CCL2. But, unfortunately, we cannot detect the protein expression of CXCL1, CXCL2, collagen I and collagen Ⅲ by western-blot in the present study. It probably attributed to the sensitivity of the detection method and the pathological characteristics of the NAFLD model used (mainly steatosis and rarely fibrosis). Considering the high sensitivity of real-time PCR and the fact that steatosis is the initiating factor of inflammation and then fibrosis, we cannot exclude the possibility that the changes of inflammatory and fibrotic genes were detectable by real-time PCR in the early stages of NAFLD. Therefore, we still accepted the results of the genes expression of CXCL1, CXCL2, COL1A1, COL3A1, COL27A1, TNC and COL6A3 without protein data getting from Western blot experiments. And we will also validate the effects of XZD on these proteins in the NASH and fibrosis model in our further study.

CCL2, CXCL1 and CXCL2 are members of chemokine family (Supplementary Fig. 4). CCL2 is secreted by macrophages, endothelial cells, hepatic stellate cells, and vascular smooth muscle cells in response to inflammatory stimulus and performs biological functions through its receptor, CCR2. Consistently, hepatic CCL2 expression was demonstrated to be up-regulated in genetic or diet-induced obesity, recruiting CCR2 (+) myeloid cells that promote hepatosteatosis.²⁷ Blockade of CCL2/CCR2 signaling significantly improved hepatosteatosis, inflammation, obesity, and insulin resistance.²⁸ CCL2 can recruit myofibroblasts directly²⁹ or via recruitment of CCR2+/Ly-6C^hi monocytes into the inflammatory sites to activate hepatic stellate cells or other precursor cells to become collagen-producing myofibroblasts.³⁰ On the other hand, myofibroblasts can contribute to inflammation by releasing CCL2, CCL21 and IL1β.³¹

CXCL1 and CXCL2 are homologues of CXCL8 (only identified in human) in mice, and mainly recruit neutrophil to inflammatory sites. In NASH patients, higher level of CXCL8 was found comparing to that in patient with steatosis or healthy controls.³² Hepatic CXCL1 is induced in a toll-like receptor 4-myeloid differentiation factor 88-dependent manner and promotes NASH and fibrosis via recruitment of neutrophil.³³

IL1R1 is a receptor for IL1α, IL1β, and IL1 receptor antagonist (IL1RA). In hepatic steatosis, necrosis leads to rapid release of IL1α and Kupffer cells can up-regulate IL1α expression.³⁴ Furthermore, IL1α leads to hepatic accumulation of TG via the up-regulation of de novo lipogenesis genes in hepatocytes.³⁵ In liver injury, the denatured DNA from damaged cells stimulates Kupffer Cells to express IL1β via activation of Toll-like receptor 9. IL1β then activates hepatic stellate cells to secrete tissue inhibitor of metalloproteinase-1, COL1A1 and COL4A1.³⁶ On the other hand, IL1β inhibits the expression of peroxisome proliferator-activated receptor α in hepatocytes, leading to the down-regulation of Acyl-CoA oxidase and carnitine palmitoyltransferase 1A, which have essential roles in fat oxidation.³⁷ As the antagonist, IL1RA blocks IL1 signaling by preventing the binding of IL1α and IL1β to IL1R1 and attenuates inflammation.³⁴

Our data indicated that XZD treatment decreased IL1 and chemokine signaling in NAFLD. But the detailed mechanisms of XZD action on pathways of IL1, CCL2, CXCL1 and CXCL2 need more investigation.

The major components of ECM include collagens, proteoglycans, elastin, and cell-binding glycoproteins.³⁸ Our data showed that the ECM regulated by XZD was the members of collagen and TNC (Supplementary Fig. 5). Liver fibrosis is a dynamic progress characterized as the excess accumulation of ECM due to the activation of hepatic myofibroblasts (e.g. hepatic stellate cells) in response to inflammation. COL1A1 and COL3A1 are fibril-forming collagens, which replace the collagen IV in the space of Disse following activation of hepatic stellate cells, leading to the capillarization of sinusoids during fibrosis.³¹ COL6A3 is a biomarker of hepatic fibrosis, and its cleaved form, endotrophin, plays a critical role in adipose tissue dysfunction, insulin resistance, and NASH.³⁹ Higher expression of COL1A1, COL1A2, COL3A1 and COL6A3 were found in severe NAFLD both in patients and mice.⁴⁰ COL27A1 is a member of the fibrillar collagen family and mainly reported to play a role in calcification of cartilage and the transition of cartilage to bone. TNC is a glycoprotein and reported to be increased in NAFLD.⁴¹

Our data suggested that XZD treatment reduced the expression of genes of cytokines and ECMs simultaneously, which indicated that XZD probably inhibited inflammation and fibrosis in very early stage. The effects and mechanisms of XZD on inflammation and fibrosis in NAFLD deserve thorough investigations.

5. Conclusion

Our data have demonstrated that XZD ameliorated NAFLD induced by HFD, which probably correlated closely to the pathways of fatty acid degradation.

Taxonomy (classification by EVISE)

Identify the disease/health condition.

Funding

This work was supported by Science and Technology Commission Shanghai Municipality (17PJ1408900), Shuguang Hospital Affiliated to Shanghai University of Traditional Chinese Medicine (SGXZ-201911).

Declaration of competing interest

None.

Appendix A. Supplementary data

The following are the Supplementary data to this article: