Introduction

Non-alcoholic fatty liver disease (NAFLD) is a fatty lesion of the liver caused by liver injury factors other than alcohol.1 It has evolved from a relatively unknown disease to the most common chronic liver disease worldwide. It has been reported that nearly 25% of the global population suffers from NAFLD, which affects about 30% of the general population in Western countries.2 NAFLD has become the number one chronic liver disease in China.3,4 The pathogenesis of NAFLD has not yet been fully elucidated, and researchers have generally accepted the “multiple strike” theory, in which the enterohepatic axis and innate immune disorders are the main factors. The early stage of simple hepatic steatosis is a reversible lesion, and it is the best time for pharmacological intervention.5 NAFLD is mainly characterized by abnormal lipid metabolism, which affects the intestinal flora, and the intestinal flora also affects lipid metabolism and lipid levels, which are causative of each other.6 There is increasing evidence that metabolic disorders caused by intestinal flora dysbiosis play an important role in the pathogenesis of NAFLD.7 There exists an inextricable bi-directional connection between the liver and intestines in terms of anatomy and biological function, which is known as the gut-liver axis, which is consistent with the theory of “liver and intestines are connected”in traditional Chinese medicine(TCM).8

The liver and the large intestine are closely related anatomically, and they are interconnected through the portal venous system.9 In embryonic origin, both originate from the foregut, which provides a structural basis for “gut and liver connectedness”. Modern medicine has proposed “the gut-liver axis” theory,10 which suggests that there is an interaction between intestinal microorganisms and the liver. Products derived from gut microbes are transported to the liver through the portal vein and influence liver metabolism through multiple pathways. Bile and antibodies from the liver can, in turn, be secreted into the intestine and influence the intestinal flora.11,12 As an important component of the enterohepatic axis, the intestinal flora not only participates in the host’s nutrient metabolism, but also maintains the functional integrity of the intestinal mucosal barrier and immune regulation.13 The liver derives 75% of its blood supply from the portal vein, making it the first organ to be exposed to intestinal flora and metabolites via portal blood.14 A normal intestinal barrier prevents the transfer of microorganisms and metabolites or toxins from the intestinal lumen to the extra-intestinal lumen, and disruption of the intestinal barrier may cause translocation of intestinal microorganisms, over activate the immune system, and trigger or promote the development of hepatic inflammation. Thus, the development of NAFLD can affect the homeostasis of intestinal flora, and abnormalities of intestinal flora and metabolites can also affect the development of NAFLD.15

Although there is an increasing amount of basic research on NAFLD, there is still a lack of specific drugs for the treatment of NAFLD.16 In this context, it is noteworthy that the efficacy of traditional Chinese medicine (TCM) in treating NAFLD through intestinal flora is being affirmed by clinical practice and research.17 Some studies have shown that Chinese medicine enters the digestive tract through oral intake, part of it is directly absorbed into the blood through the gastric mucosa, and the other parts with lower bioavailability are not absorbed from the upper gastrointestinal tract, reach the large intestine through the small intestine, and come into contact with intestinal flora and have interactions with each other, and the components of the effective parts of Chinese medicine not only change the metabolism of the intestinal flora, but also are transformed or metabolized by the intestinal flora.18 Chinese medicine can also prevent the displacement of intestinal bacteria and endotoxins by regulating the intestinal flora, balancing the beneficial and pathogenic bacteria, and can significantly improve the regeneration ability of mucosal cells and reduce the excessive mucosal permeability, thus improving the barrier function of the mucosa.19 According to TCM, NAFLD caused by long-term overeating is characterized by intestinal symptoms such as spleen deficiency, phlegm stagnation, blood stasis and diarrhea and loose stools. Modern studies have shown that intestinal flora dysfunction plays an important role in the pathogenesis of various chronic liver diseases, and the digestive symptoms are consistent with the evidence of spleen deficiency in TCM.20

Previous studies have shown that early NAFLD can be improved and treated by strengthening the spleen and resolving blood stasis. Shenling Baizhu San(SLBZS) is a classic formula for strengthening the spleen and eliminating dampness from the Tai Ping Hui Min He Yao Bao Fang, which has the effect of benefiting qi, tonifying the spleen, resolving dampness and stopping diarrhea. Modern pharmacological studies have confirmed that Renshen and Baizhu not only have good antibacterial, antioxidant and immune regulation effects, but also can adjust the structure of intestinal flora by reducing the number of enterococci, lowering the value of Firmicutes / Bacteroidetes and increasing the composition of probiotic bacteria.21 The therapeutic effect of SLBZS on NAFLD may be related to the regulation of hepatic mitochondrial energy metabolism by intestinal flora and its metabolites, and may be associated with the UCP2/AMPK/IF1 signaling pathway.22 Quzhi Ruangan Fang(QZRGF) is an empirical formula of Professor Su Lian, an expert in liver diseases and an honorary famous Chinese medicine practitioner in Yunnan Province, China, which has the efficacy of softening the liver and strengthening the spleen, regulating qi and activating blood circulation. Previous studies have demonstrated that QZRGF can improve blood lipids, liver lipids, hepatic tissue steatosis and haemorheology-related indexes in high-fat diet induced fatty liver rats.23,24 As one of the famous formulas for removing blood stasis in Wang Qingren’s “Reform and Error Correction of Medical Forests” in the Qing Dynasty, the group found that the mechanism of action of Gexia Zhuyu Tang(GXZYT) removing formula in improving lipid deposition in hepatocytes of rats with NAFLD might be related to the modulation of lipid metabolism, the reduction of the expression of CD62p, and the reduction of platelet aggregation rate.25,26 The group’s previous research on the three compound formulas found that the high-dose group had the best effect, therefore, the high-dose group was chosen for all the comparative studies in this experiment.27–29

The complex composition and diverse targets of TCM compounding have led to more difficulties in studying the mechanisms of TCM compounding. It was found that there are many similarities between the core concepts of TCM holistic treatment and network pharmacology.30 So some scholars proposed the research strategy of Chinese medicine network to study Chinese medicine pharmacology.31 In recent years, the study of network pharmacology of Chinese medicine has achieved vigorous development. It has emerged as an important intersection between TCM and network pharmacology. By exploring the mechanism of action of TCM at the molecular level and in the context of biological networks, it can provide a new way to study the mechanism of action of TCM.32 In this study, we systematically analyzed the mechanism of action of SLBZS, QZRGF and GXZYT in the treatment of NAFLD as well as their interfering effects on the intestinal flora at the molecular level by means of network pharmacology, bioinformatics, molecular docking and other means.

Materials and Methods Screening of Active Ingredients and Targets of Action of Shenling Baizhu San, Quzhi Ruangan Fang and Gexia Zhuyu Tang

Firstly, the active ingredients of Shenling Baizhu San, Quzhi Ruangan Fang and Gexia Zhuyu Tang were screened from the Chinese herbal tablets (https://www.example.comtcmsp.php)33 using the criteria of Oral Bioavailability (OB) ≥ 30% and Drug Likeness (DL) ≥ 0.18. Meanwhile, with the help of Bioinformatics Analysis Tool for Molecular Mechanism of Action of Traditional Chinese Medicines (BATMAN-TCM),34 we searched for the Chinese medicine Wulingzhi,35 which was not available in the TCMSP platform, with the search criteria of Score ≥ 20 and P-value < 0.05 to complete the whole formula of GXZYT. Literature was searched through CNKI to find the compositional information of Lanhuasheng,36 and the targets corresponding to the active ingredients were predicted and organized through the SwissTarget website (http://swisstargetprediction.ch/). Then, the protein targets of these screened active ingredients were retrieved in TCMSP and their chemical composition and targets were supplemented from CNKI and PubMed. Subsequently, the protein targets were entered into the UniProt database (www.uniprot.org)37 for standardization using Homo sapiens, and the target information of all the drugs obtained earlier was annotated by means of an annotation file, which converts the full name of the gene into the symbol of the gene.

Screening for NAFLD-Related Targets

Using “NAFLD” as the keyword, potential therapeutic targets for NAFLD were explored from the databases of OMIM (https://omim.org/),38 DrugBank (https://go.drugbank.com/),39 DisGeNET (https://www.disgenet.org/),40 Gene Recognition Cards (http://www.genecards.org/) databases41 to mine potential therapeutic targets for NAFLD. Finally, the install. Packages (“venn”) package was installed through R. The four database disease targets were removed from duplicate targets and concatenated sets were taken to obtain the final disease-associated genes, and the Wayne diagram was plotted.

Construction of the Relationship Between Effective Active Ingredients and NAFLD

Therapeutic targets of Shenling Baizhu San, Quzhi Ruangan Fang and Gexia Zhuyu Tang Cytoscape 3.9.1 was used to visualize the data and construct the maps of the active ingredient-NAFLD therapeutic targets of Shenling Baizhu San, Quzhi Ruangan Fang and Gexia Zhuyu Tang respectively.

Construction of a PPI Network of Effective Active Ingredients-NAFLD Targets in Shenling Baizhu San, Quzhi Ruangan Fang and Gexia Zhuyu Tang

The acquired drug-disease common target genes were inputted into the String database42 to construct the interaction networks of Shenling Baizhu San, Quzhi Ruangan Fang and Gexia Zhuyu Tang NAFLD target proteins, respectively. Firstly, select multiple proteins, then set the biological category as “Homo Sapiens” and the interaction score as “Highest confidence > 0.9”. In the network display option, we chose to hide unconnected nodes, and the rest of the settings used the system default values. The protein association score is set to 0.4, and each node represents all the proteins produced by a protein-coding locus. The colors of the lines in the graph include blue, red and yellow, which represent the evidence of gene co-evolution, gene fusion and text mining, respectively. The “string.tsv” file was downloaded to visualize and analyze the PPI network data using Cytoscape 3.9.1, and the CytoNCA plugin was used to find the core of the PPI network. The topological parameters include node connectivity, meso centrality and proximity to centrality, the larger the parameter value the closer the node is to the center of the network, indicating that the node is more critical in the network.43,44

Gene Enrichment Analysis

We used the R language in combination with the gene data analysis tools provided by Bioconductor to perform gene enrichment analysis. Specifically for Gene Ontology (GO) enrichment, the GO enrichment analysis includes three aspects: cellular component (CC), molecular function (MF) and biological process (BP). The results of all GO enrichment analyses performed were determined by P < 0.05.

Molecular Docking Validation

Based on these analyses, molecular docking of the core target and the main active ingredient was performed. The 3D crystal structures of the target proteins were obtained in the Protein Data Bank (PDB) (https://www.rcsb.org/).45 Their PDB format files were downloaded, dehydrated and hydrogenated in AutoDockTools (version: 1.5.7) (https://ccsb.scripps.edu/mgltools/downloads/),46 and then selected as a receptor and saved as a pdbqt file. The mol 2 file of the active component was obtained via TCMSP database and the chemical structure of the active ingredient was obtained in PubChem database, hydrogenated in AutoDockTools, selected as a ligand and exported as a pdbqt file. Molecular docking was performed by AutoDock (version: 1.5.7)47 to obtain the affinity of each component for the target. Active ingredients and targets with good binding activity were screened based on affinity and visualised by PyMOL (version: 2.4.0) (http://www.pymol.org/pymol).

Preparation of Drugs

The formula “Shenling Baizhu San” is derived from the book “Taiping Huimin Hejiaobu Fang” and is composed as follows: 12g of Dangshen, 12g of Baizhu, 15g of Fulin, 15g of Baibiandou, 15g of Yiyiren, 15g of Shanyao, 10g of Chenpi, 10g of Sharen, 12g of Lianzi, 10g of Jiegeng, 6g of Gancao. SLBZS (Production batch number: 23101056) from Beijing Tong ren tang, a well-known brand of traditional Chinese medicine, was selected for this study. SLBZS routine daily dose for adults 20g. The dosage of the drug is based on the clinical equivalent dose and the body surface area of human and rats. The daily adult dosage was about 20g. Based on the clinical equivalent dose, the daily requirement of Shenling Baizhu San per 200g of rat is 20×0.018×5 = 1.8g/kg. Our previous study27 found that a high dose was the most effective in the treatment of NAFLD, and the high dose was four times the equivalent dose (7.2 g/kg), so we chose a high dose for the experiment this time.

Gexia Zhuyu Tang: Wulingzhi 9g, Danggui 9g, Chuanxiong 6g, Taoren 9g, Danpi 6g, Chishao 6g, Wuyao 6g, Yanhuosuo 3g, Gancao 9g, Xiangfu 3g, Honghua 9g, Zhike 5g. The daily dosage for adults is about 80 g. Based on the clinically equivalent dose, the daily requirement of GXZYT per 200 g of rat is 80×0.018 × 5 = 7.2 g/kg. Our previous study28 found that the high dose, which was four times the equivalent dose (28.8 g/kg), was the most effective in treating NAFLD, so we chose the high dose for this experiment.

Quzhi Ruangan Fang is mainly composed of 15g of Lanhuashen, 15g of Ezhu, 15g of Baizhu, 15g of Shanzha, 10g of Juhua, 6g of Qingpi and 3g of Sanqi. Sanqi is Saned into fine San and added into the formula when used. The minimum clinical dosage of the formula for human is 77g, 77×0.018 × 5 =6.93 g/kg. The high dose, which was four times the equivalent dose (27.72 g/kg), was the most effective in treating NAFLD.29

Fenofibrate capsules are 200mg daily for adults, Also converted from clinically equivalent doses, therefore the daily dose for rats is 200×0.018×5 =18mg/kg by gavage. Purchased from French pharmaceutical company RECIPHARMFONTAINE. Production batch number: 33145.

Raw herbs were purchased from the pharmacy of Yunnan Provincial Hospital of Traditional Chinese Medicine. The above prescriptions of QZRGF and GXZYT were boiled in the Pharmacy Department of Yunnan Provincial Hospital of Traditional Chinese Medicine and made into extracts in strict accordance with the production standards for in-hospital preparations. The extracts were cooled and stored at 4 °C. When used, the extracts were prepared with saline according to the concentration of the SLBZS, QZRGF, GXZYT and fenofibrate required for each group and the weight of the rats.

The identification of the Shenling Baizhu San, Gexia Zhuyu Tang, Quzhi Ruangan Fang components were outsourced to Wuhan metware Metabolic Biotechnology Co. (Figure 1)

Figure 1 HPLC Chromatogram.

Co-Detection of Active Ingredients in Shenling Baizhu San, Gexia Zhuyu Tang, Quzhi Ruangan Fang Based on UPLC-MS/MS Detection Platform and Self-Constructed Database Dry Sample Extraction

(1) Biological samples were placed in a lyophiliser (Scientz-100F) for 63h of vacuum freeze-drying;

(2) Grinding (30 Hz, 1.5 min) to powder form using a grinder (MM 400, Retsch);

(3) Use an electronic balance (MS105DΜ) to weigh 50 mg of sample powder, add 1200 μL of −20 °C pre-cooled 70% methanol water internal standard extract (less than 50 mg according to the proportion of every 50 mg of sample to add 1200 μL of extractant), the internal standard extract was prepared as 1 mg of standard dissolved in 1 mL of 70% methanol water to formulate 1000 μg / mL standard The internal standard extract was prepared by dissolving 1 mg of standard in 1 mL of 70% methanol water to make 1000 μg/mL standard mother liquor, and the 1000 μg/mL mother liquor was further diluted with 70% methanol to make 250 μg/mL internal standard solution;

(4) Vortex every 30 minutes, each lasting 30 seconds, for a total of 6 vortexes;

(5) After centrifugation (12000 rpm, 3 min), the supernatant was aspirated and the sample was filtered through a microporous filter membrane (0.22 μm pore size) and stored in the injection vial for UPLC-MS/MS analysis.

Chromatography and Mass Spectrometry Acquisition Conditions

The data acquisition instrumentation consists primarily of Ultra Performance Liquid Chromatography (UPLC) (ExionLC™ AD, https://sciex.com.cn/) and Tandem mass spectrometry (MS/MS).

Liquid phase conditions mainly include:

(1) Column: Agilent SB-C18 1.8 µm, 2.1 mm * 100 mm;

(2) Mobile phase: ultra-pure water (with 0.1% formic acid added) for phase A, and acetonitrile (with 0.1% formic acid added) for phase B. The mobile phase was 0.00 min and the elution gradient was 0.00 min;

(3) Elution gradient: the proportion of phase B was 5% at 0.00 min, the proportion of phase B increased linearly to 95% within 9.00 min and was maintained at 95% for 1 min, and the proportion of phase B decreased to 5% from 10.00 to 11.10 min and equilibrated at 5% for 14 min;

(4) Flow rate 0.35 mL/min; column temperature 40°C; injection volume 2 μL.

Animal Models

120 male SD rats of SPF grade, weighing 180–220g, were kept in the animal room of Yunnan University of Traditional Chinese Medicine. Production license number: SCXK (Beijing) 2019–0010, Number of Experimental Facilities Certification: SYXK(Yunnan)K2022-0004. The rats were purchased from SPF(Beijing)biotechnology co.,Ltd. All rats were given free access to feed and water throughout the experimental period at a controlled temperature of 21°C±2°C and a 12-hour light/dark cycle. (Normal feed: 29% primary secondary flour, 4.1% high gluten flour, 40.8% puffed corn, 2% fishmeal, 16% 46 soybean meal, 2% soybean oil, 6% premixes, which were purchased from SPF(Beijing)biotechnology co.,Ltd Production batch number: 06148H06152133C). High fat feed: 78.8% basal feed + 10% lard + 10% egg yolk powder + 1% cholesterol + 0.2% sodium cholate, which were purchased from SPF(Beijing)biotechnology co,Ltd. After one week of normal feeding, the experimental rats were randomly divided into 6 groups, Normal (n=10), Model (n=10), GXZYT (n=10), SLBZS (n=10), QZRGF (n=10), Fenofibrate (n=10). After 6 weeks of modeling, the livers of rats in the model group were taken for paraffin section HE and oil red O staining to observe the degree of hepatocellular steatosis. If the number of fat vacuoles in the liver cells observed on the pathological section accounts for one-third or more of the field of view, it indicates successful modeling. After the success of modeling was confirmed, the amount of drug administered to each rat was calculated according to the body weight of the rats, and the drug was diluted and dissolved in physiological saline. SLBZS, GXZYT, QZRGF, and Fenofibrate groups of rats were orally administered with a drug volume of 18mg/kg, respectively. The normal group and model group were orally administered with 2mL/d physiological saline. This study was approved by the Animal Ethics Committee of Yunnan University of Traditional Chinese Medicine (Ethics Committee approval number: R-062021LH013).

Sample Preparation

On the day before euthanasia, we collected fecal samples for 16S rDNA sequencing. After the last administration, fasted for 12 hours and the rats were anesthetized by intraperitoneal injection of 3% pentobarbital (0.1 mL/100 g body weight). The entire livers were excised and stored at - 80°C for histological content analysis. Blood was taken from the abdominal aorta, and the whole blood was centrifuged at 3000×g at 4°C for 15 min, and the supernatant was collected and dispensed into 1.5 mL centrifuge tubes and stored at - 20°C for biochemical analysis.

Biochemical Analysis

A Roche automatic biochemical analyzer (Roche, Switzerland; model: Cobas c311) and Biochemical test kits (Nanjing Jiancheng Bioengineering Company) were used. Blood was taken from the abdominal aorta of rats, and venous blood and whole blood were collected. It was left for about 2 h, centrifuged at 3000 rpm for 15 min, and the supernatant was taken to evaluate the liver function and lipid indexes, including ALT, AST, TC, TG, high-density lipoprotein cholesterol (HDL-C), and low-density lipoprotein cholesterol (LDL-C). Each step was performed according to the instructions of the kit.

Pathological Examination of the Liver

Livers were fixed with tissue fixative for >24 hours, then removed, dehydrated and embedded, cut into two 5×5 × 2 cm cubes and stained with H&E staining in paraffin sections and O staining in frozen sections according to the instructions of the staining kit, each of which was about 5 μm in thickness. Finally, the membrane was sealed with gelatine, and the histopathological changes of the liver specimens were observed under the microscope and calculated lipid volume fraction (lipid area/total area×100%).

Detection of Bacterial Diversity and Abundance in Rat Feces Faeces

We collected rat feces, transferred them to sterile eppendorf (EP) tubes, froze them in liquid nitrogen and stored them at - 80°C. Stored intestinal contents were sent on dry ice to GemVision in Suzhou, China for testing. Construction of high-throughput sequencing libraries and sequencing based on the Illumina MiSeq platform were performed by GENEWIZ (Suzhou, China). DNA was extracted using the Magen Hipure Soil DNA Kit following the instructions for use. We used a Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, CA) to detect the concentration of DNA samples, and the MetaVx™ Library Construction Kit (GENEWIZ, Inc., South Plainfield, NJ, USA) to construct sequencing libraries. The two highly variable regions of prokaryotic 16S rDNA, including V3 and V4, were amplified using a series of PCR primers designed by GemViz using 30–50ng of DNA as template according to the manufacturer’s instructions. The two highly variable regions including V3 and V4 on prokaryotic 16S rDNA were amplified using a series of PCR primers designed by Viwit Jin as a template of 30–50ng DNA. The upstream primer containing the sequence “CCTACGGRRBGCASCAGKVRVGAAT” and the downstream primer containing the sequence “GGACTACNVGGGGTWTCTAATCC” were used to amplify the V3 and V4 regions. In addition, the PCR product of 16S rDNA was spliced by PCR with an Index at the end for NGS sequencing. Library quality was checked using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) and library concentration was checked using a Qubit2.0 Fluorometer (Invitrogen, Carlsbad, CA). After DNA library mixing, PE250/300 double-end sequencing was performed according to the Illumina MiSeq (Illumina, San Diego, CA, USA) instrument instruction manual, and the sequence information was read by MiSeqControlSoftware (MCS), which comes with the MiSeq.48–51

Statistical Analysis

Data were statistically analyzed using SPSS 26.0 software, and the measurement information was expressed as the mean and standard deviation . Firstly, the variance of the data in each group was tested for chi-square, one-way ANOVA was performed for chi-square between groups, and the two comparisons were made using the LSD test; if the variance was not chi-square, the rank-sum test was used, and the two comparisons were made using the Dunnett’s T3 test. The test level was α=0.05. GraphPad Prism 9.0 was used to draw the images.

Results Network Pharmacology Analysis Active Ingredients and Targets of SLBZS, GXZYT and QZRGF

A total of 132 active components were isolated from SLBZS, including 1 Baibiandou, 3 Baizhu, 4 Fulin, 83 Gancao, 4 Jiegeng, 9 Lianzi, 16 Renshen, 9 Shanyao, 4 Sharen, and 5 Yiyiren; among them, 6 components were shared. A total of 34 active ingredients were isolated from QZRGF, including 3 from Baizhu, 1 from Ezhu, 16 from Juhua, 3 from Qingpi, 5 from Sanqi, 2 from Shanzha, and 4 from Lanhuasheng. 272 active ingredients were isolated from GXZYT, including 2 from Danggui, 23 from Taoren, 22 from Honghua, and 22 from Xiangfu. Honghua 22, Xiangfu 18, Wuyao 9, Danpi 11, Chishao 29, Chuanxiong 7, Yanhusuo 49, Zhike 5, Wulingzhi 6, Gancao 92. (See Table 1)

Table 1 Traditional Chinese Medicines and Main Active Ingredients of SLBZS, GXZYT and QZRGF

NAFLD Disease Target Genes Combined with Drug Target Genes of SLBZS, GXZYT and QZRGF

Nonalcoholic fatty liver disease (NAFLD) targets were screened with the keyword “Nonalcoholic fatty liver disease” from GeneCards, OMIM, Drug Bank, and DisGeNET databases, and median values were screened, and a total of 1792 NAFLD-related disease genes were obtained by running a concatenation operation in R version 4.2.0 software, a total of 1792 disease genes associated with NAFLD were obtained (Figure 2A). In further studies, a total of 4982 TCM target genes for GXZYT were obtained from TCMSP and BATMAN-TCM databases, and 251 non-duplicated TCM target genes were obtained after de-duplication. Comparing these TCM target genes with NAFLD-related disease genes, 144 intersecting genes were obtained. A total of 2416 TCM target genes of SLBZS were obtained from TCMSP database, and 233 non-duplicated TCM target genes were obtained after de-duplication. Comparing these TCM target genes with NAFLD-related disease genes, 136 intersecting genes were obtained. A total of 1144 TCM target genes of QZRGF were obtained from the TCMSP database and CNKI literature. After deduplication, 289 non repetitive traditional Chinese medicine target genes were obtained. Comparing these TCM target genes with NAFLD-related disease genes, 169 intersecting genes were obtained (Figure 2B).

Figure 2 (A) NAFLD-related disease genes (B) Intersection of GXZYT, QZRGF and SLBZS Chinese medicine target genes and NAFLD disease genes.

Graph of SLBZS, GXZYT and QZRGF Effective Active Ingredients-NAFLD Therapeutic Target Modulation Relationship

Cytoscape 3.9.1 software was used to construct the regulatory network of Chinese medicine compound and analyze the active ingredients and NAFLD disease-related genes of SLBZS, GXZYT and QZRGF. In the GXZYT active ingredient NAFLD network, Yanhusuo was identified by purple, Xiangfu was identified by light purple, Wuyao was identified by pink, Taoren was identified by dark yellow, Honghua was identified by light grey, Chuanxiong was identified by light yellow, Chishao was identified by green, Danpi was identified by red, Zhike was identified by light green, Gancao was identified by grey, Danggui was identified by green, and Wulingzhi was identified by yellow. In the overlapping colours, the main drugs associated with NAFLD treatment in GXZYT are Honghua, Xiangfu, Gancao and Zhike. In the QZRGF active active ingredient-NAFLD network, Lanhuashen is identified in red, Sanqi in yellow, Shanzha in green, Baizhu in blue, Ezhu in light yellow, Juhua in purple, and Qingpi in dark blue. In the overlapping colours, the main drugs associated with NAFLD treatment in QZRGF are Lanhuashen, Juhua and Sanqi. In the SLBZS active active ingredient-NAFLD network, Renshen is identified in green, Shanyao is identified in pink, Sharen is identified in grey, Yiyiren is identified in purple, white Baibiandou is identified in blue, Baizhu is identified in light yellow, Fulin is identified in light purple, Gancao is identified in red, Jiegeng is identified in blue, and Lianzi is identified in yellow. In the overlapping colors, the main drugs associated with NAFLD treatment in SLBZS are Renshen, Shanyao, Gancao and Lianzi. Node size indicates the degree of relevance (Figure 3).

Figure 3 Effective active ingredients of GXZYT, QZRGF and SLBZS-NAFLD treatment target regulatory network diagram.

Protein Interaction Networks (PPI)

The common targets of SLBZS, GXZYT and QZRGF were imported into the STRING database and analyzed to obtain the PPI network diagrams (Figure 4A). The PPI network graphs were imported into Cytoscape 3.9.1 software for visualization and analysis, “Degree”, “Eigenvector”, “LAC” and “Network”. “The core targets of SLBZS were CCND1, MAPK1, AKT1, CASP3, HSP90AA1, FOS, ESR1, TP53, STAT3, MAPK3, MYC, and GXZYT were MAPK1, AKT1, CASP3, FOS, MAPK3, MYC, CASP3, FOS, HSP90AA1, BCL2, ESR1, TP53, STAT3, MAPK3, CTNNB1, MYC. The core targets of QZRGF are: BCL2, TP53, ESR1, STAT3, MAPK3, FOS, MYC, CCND1, MAPK1, AKT1, CASP3. The three party intersection targets of the core targets were MAPK1, AKT1, CASP3, FOS, ESR1, TP53, STAT3, MAPK3, MYC. (Figure 4B)

Figure 4 (A) Network map of protein interactions between GXZYT, QZRGF and SLBZS targets and NAFLD disease genes (B) Treatment of NAFLD target PPI networks using STRING’s GXZYT, QZRGF and SLBZS. And use Cytoscape to treat the PPI network and get the PPI network core targets.

GO Enrichment Analysis

The tripartite and NAFLD co-targets were introduced into Cytoscape and subjected to GO enrichment and KEGG pathway analysis via R language. The analysis results demonstrated GO enrichment results at three levels: biological process (BP), molecular function (MF) and cellular composition (CC). The results of the study showed that by treating NAFLD with SLBZS, GXZYT and QZRGF, respectively, we found that we could modulate the response to xenobiotic stimulus, response to oxidative stress, response to molecule of bacterial, originate response to lipopolysaccharide, and many other biological processes. The results ranked according to the P value from smallest to largest showed that Diaphragm Downward Expelling Blood Stasis Tang was closely related to response to molecule of bacterial in the treatment of NAFLD. The relationship is close (Figure 5).

Figure 5 GO enrichment analysis: top 10 P values for Biological process(BP), Cellular component(CC) and Molecular function(MF).

Molecular Docking Validation

Based on the above analysis, 9 core components (baicalein, luteolin, nobiletin, quercetin, kaempferol, naringenin, licochalcone a, diosgenin, the acacetin) were molecularly docked with core targets of MAPK1, AKT1, CASP3, FOS, TP53, STAT3, and MYC (Table 2 and Figure 6). Affinity < −4.25 kcal-mol-1 indicates the presence of binding activity between the ligand and the target; affinity < −5.0 kcal-mol-1 indicates good binding activity; affinity < - 7.0 kcal-mol-1 indicates strong docking activity.52 The best docking activities were screened at AKT1-diosgenin (−9.51), AKT1-beta-carotene (−9.99), AKT1-baicalein (−7.6), TP53-baicalein (−8.41), TP53-luteolin (−8.48), TP53- quercetin (−8.33), TP53-nobiletin (−7.99), TP53-acacetin (−6.61), CASP3-luteolin (−7.11), CASP3-baicalein (−6.7), FOS-quercetin (−6.44), STAT3-licochalconea (−5.89), MAPK1-Chryseriol (−6.66), MAPK1-chryseriol (−6.66), and MAPK1-chryseriol (−6.66), MAPK1-isorhamnetin (−6.68), MYC-quercetin (−6.17).

Table 2 Molecular Docking Information of Core Components and Core Targets

Figure 6 The results of molecular docking between the core targets and core ingredients.

Animal Experiment Verification The SLBZS, QZRGF and GXZYT All Reduced Lipid Deposition and Liver Injury in Rats with NAFLD

We gave the rats in the Normal group a normal diet, the Model, SLBZS, QZRGF, and GXZYT groups were given a high-fat diet, the Normal and Model groups were given an equal amount of saline, and the TCM groups were given their respective TCM compounds, which were taken at 8W and 12W, respectively (Figure 7A). The body weight and liver weight of rats in the Model group were significantly higher than those in the normal group (P< 0.01). Compared with the model group, the body weight and liver weight of rats in all dosing groups were significantly lower (P< 0.01). The liver-to-body ratio of rats in the Model group was significantly higher than that of the normal group (P< 0.01), and the liver-to-body ratios of rats in all groups were significantly lower than that of the model group (P< 0.01; Figure 7B–D). Compared with the normal group, serum ALT and AST levels of rats in the Model group were significantly higher (P< 0.01). Compared with the Model group, ALT and AST levels of rats in the Fenofibrate group were significantly lower (P< 0.01). AST and ALT levels of rats in the SLBZS, QZRGF and GXZYT groups were significantly lower (P< 0.01) (Figure 7E and F). As for lipid levels, compared with the normal group, serum TC, TG and LDL-C levels were significantly higher and HDL-C levels were significantly lower in the Model group of rats (P< 0.01). Compared with the Model group, the serum TC, TG and LDL-C levels of rats in each administration group were significantly reduced, and the HDL-C level was significantly increased (P< 0.01) (Figure 7G–J). HE staining results (Figure 7K) showed that the liver of rats in the Normal group had an intact histological structure, and the hepatocytes were arranged in strips radially outward around the central vein. In contrast, rats in the Model group showed a large number of microvesicular and macrovesicular steatosis in the hepatocytes, severe damage to the lobular structure, ballooning hepatocytes. The area of fatty vacuoles and balloon-like lesions in the hepatocytes of rats in the herbal and fenofibrate groups was significantly reduced compared with the model group. The best effect was observed in the GXZYT group, which made the liver basically return to the normal level. The results of oil red O staining showed that the liver cells of rats in the Normal group had a clear and complete liver structure, the nuclei of the cells showed blue-purple color, and there was no deposition of red lipid droplets in the cytoplasm; the rats in the Model group had a large number of red lipid droplets were aggregated. The lipid droplets in the administered group were significantly reduced, the reddened area of liver tissue was reduced, and the liver was basically restored to the normal level. This indicated that all groups of herbal compound could reduce hepatic steatosis and liver injury (Figure 7L).

Figure 7 Experimental validation results in NAFLD rats. (A) Overall course of animal experiments (B) Comparison of body weight of rats in each group (n = 6). (C) Comparison of liver weight of rats in each group (n = 6). (D) Comparison of liver-body ratios of rats in each group (n = 6). (E) Comparison of ALT in rats of each group (n = 6). (F) Comparison of AST in rats of each group (n=6). (G) Comparison of TC of rats in each group (n=6). (H) Comparison of TG in each group (n=6). (I) Comparison of HDL-C in rats of each group (n=6). (J) Comparison of LDL-C in rats of each group (n=6). (K) Pathological staining of rat liver in each group (HE staining, 400×). (L) Pathological staining of rat liver in each group (Oil Red O staining, 400×). ##P < 0.01, Compared with Normal; *P < 0.05, **P< 0.01, compared with Model.

The Tripartite Formulae of SLBZS, QZRGF and GXZYT Can Improve the Flora Structure of Rats with NAFLD

In order to investigate the effect of tripartite on the diversity of rat intestinal flora, fecal 16S rDNA sequencing was used to compare the diversity of rat intestinal flora after herbal medicine intervention, including alpha diversity analysis and beta diversity analysis. The community richness dilution curve showed that the sequencing amount covered all the intestinal flora in the samples, which met the requirements of data analysis. (As shown in Figure 8A), each curve represents one sample, marked with different colors; the number of OTUs increased with increasing sequencing depth. The dilution curves of each group of rat intestinal flora firstly increased and then flattened out, ie, more data volume would only produce a very small number of new species, indicating that the sequencing depth was sufficient to meet the requirements of data analysis. Based on the results of OTU clustering analysis, we counted the OTUs common and specific to the identified microbial species in each group of rat fecal samples and plotted the petal diagrams in (Figure 8B). A total of 188 different bacterial species were identified, of which 64 strains were specifically identified in each group. 95 species were common to the 12 groups of rats, as shown in Figure 8B. The intestinal microorganisms in the fecal isolates of rats in each group were rich in species. Compared with the Normal group, there was a significant decrease in the bacterial species in the Model group and a significant increase in all bacterial species in the medicated group. Compared with the Normal group, the Chao1 index of intestinal contents decreased in the Model group, and increased in the GXZYT, QZRGF, SLBZS, and Fenofibrate groups, with the most significant improvement in the flora of the SLBZS group, followed by the GXZYT and QZRGF groups in the 8-week period, and in the 12-week period, the most significant improvement in the flora of the GXZYT group, followed by the SLBZS and QZRGF groups in the 12-week period. SLBZS and QZRGF groups (Figure 8C) Compared with the Normal group, the Shannon index of the intestinal contents of the rats in the Model group was reduced, and the GXZYT, QZRGF, SLBZS and Fenofibrate groups showed an increasing trend, and the most significant improvement in the flora was observed in the SLBZS group followed by the QZRGF and GXZYT groups among the 8-week groups. The most significant improvement of the flora was observed in the QZRGF and GXZYT groups among the 12-week groups. QZRGF group showed the most significant improvement in flora, followed by SLBZS and GXZYT groups (Figure 8D). The β-diversity analysis of intestinal flora was performed using the Unifrac distance method, which calculates the distance between samples and compares the magnitude of the differences in species diversity between samples and groups. PCA analysis revealed that the intestinal flora of the Normal group was gradually separated from that of the other groups of rat in 8W and 12W, and that there was no overlap between the drug-intervention group and the Normal group but there was a tendency for them to be close to each other, and that the SLBZS group was the closest to the Normal group, followed by the GXZYT group. The SLBZS group was the closest to the Normal group, followed by the GXZYT group and the QZRGF group (Figure 8E). MDS analysis revealed that the intestinal flora of the Normal group and the other groups of rat were also gradually separated in 8W and 12W. Although there was no overlap between the drug intervention group and the Normal group, all drug groups tended to be close to the Normal group. The SLBZS group was closest to the Normal group, followed by the GXZYT and QZRGF groups (Figure 8F).

Figure 8 Effect of SLBZS, QZRGF and GXZYT on colony structure in NAFLD rats: (A) Dilution curve of community richness of rat intestinal flora in each group (B) OTU petal plot of microbial species (C) Chao1 index of rat intestinal flora in each group (D) Shannon index of rat intestinal flora in each group (E) PCA analysis of rat intestinal flora in each group (F) MDS analysis of rat intestinal flora in each group.

Analysis of the Effects of SLBZS, QZRGF and GXZYT on the Phylum and Genus Levels of Intestinal Flora in NAFLD Rats

The top 15 species of rat intestinal flora in terms of maximum abundance at the phylum level were selected, and the highest abundance was found in Firmicutes and Bacteroidetes, which accounted for more than 80% of the total abundance, followed by Proteobacteria, Actinobacteria, and Fusobacteria (Figure 9A and B). The ratio of abundance of Firmicutes to Bacteroidetes (ie, F/B value) in each group, we can see that it was significantly higher in the Model group compared with the Normal group, and the F/B ratio was significantly lower after the intervention of Chinese medicine and Fenofibrate (see Figure 9C–E). Among the six groups of samples, Bacteroidetes, Fusobacteria, Firmicutes, and Fusobacteria were the most significantly different from each other. Firmicutes) showed the most significant differences. The highest percentage of Firmicutes was found in the Normal group, and the highest percentage of Bacteroidetes was found in the Model group. The smaller percentage of bacterial groups in the six groups were Proteobacteria, Patescibacteria, Epsilonbacteraeota, and Verrucomicrobia; of these, Proteobacteria was found in the Normal group, and the highest percentage of Bacteroidetes was found in the Model group. Proteobacteria are more highly represented in the Normal group, second only to Bacteroidetes in molecular evolutionary studies (Figure 9F and G). Phylogenetic inference can reveal the sequence of evolutionary processes of the organism in question and understand the history and mechanisms of biological evolution. In molecular evolutionary research, systematic inference can reveal the sequence of biological evolution processes, understand the history and mechanisms of biological evolution, and construct evolutionary trees based on the differences in bases between sequences at a certain taxonomic level (Figure 9H–J). We constructed an evolutionary tree based on the maximum likelihood method by selecting OTUs corresponding to the genus level, and found that Bacteroidetes and Firmicutes had the highest proportion and close relationship.

Figure 9 Effects of SLBZS, QZRGF and GXZYT on the levels of intestinal flora phylum and genus in NAFLD rats: (A) Percentage of horizontal abundance of rat intestinal bacterial phylum in each group (B) Percentage of horizontal abundance of rat intestinal bacterial genus in each group (C) Percentage of abundance of thick-walled bacterial phylum in each group (D) Percentage of abundance of anaplasmosis phylum in each group (E) Percentage of abundance of thick-walled bacterial phylum in each group of rat intestinal phylum/anaplasmosis phylum abundance ratio (F/B)(F) 8W Comparison of the top 6 bacterial genera ranked by the relative abundance in the rat intestinal tract in each group (G) 12W Comparison of the top 6 genera ranked by relative abundance in the rat intestine in each group (H) Network network analysis of rat intestinal flora in each group (I) Species evolution tree at genus level of rat intestinal flora in each group (J) Phylogenetic evolution tree of rat intestinal flora in each group.

LEfSe Analysis of the Effects of SLBZS, QZRGF and GXZYT on the Intestinal Flora of NAFLD Rats

LEfSe was then performed to identify specific microorganisms associated with pharmacological interventions with LDA>4. The results showed that (Figure 10A–F) the most characterized bacterial phylum was Firmicutes and the genus was Lachnospiraceae_NK4A136_group in the 8QZRGF compared to the 8GXZYT group, the 8SLBZS compared to the 8QZRGF group, the most characterized bacterial The most characteristic bacterial phylum was Firmicutes, genus Lachnospiraceae; 8QZRGF compared with 8GXZYT group, the most characteristic bacterial phylum was Firmicutes, genus Lachnospiraceae_NK4A136_group; 12QZRGF compared with 12GXZYT group, the most characteristic bacterial phylum was Helicobacter, genus Lachnospiraceae_NK4A136_group; 12QZRGF compared with 12GXZYT group, the most characteristic bacterial phylum was Helicobacter, genus Helicobacter. 12SLBZS compared to 12QZRGF group, the most characterized bacterial phylum is Firmicutes, genus weissella. 12QZRGF compared to 12GXZYT group, the most characterized bacterial phylum is Firmicutes, genus Lachnospiraceae_NK4A136_group.

Figure 10 LEfSe analysis showing biomarkers of SLBZS, QZRGF and GXZYT gut microbiota, LDA>4. (A) LEfSe analysis of the intestinal flora of 8SLBZS and 8GXZYT rats (B) LEfSe analysis of the intestinal flora of 12SLBZS and 12GXZYT rats (C) LEfSe analysis of the intestinal flora of 8SLBZS and 8QZRGF rats (D) LEfSe analysis of the intestinal flora of 12SLBZS and 12QZRGF rats (E) LEfSe analysis of the intestinal flora of 8QZRGF and 8GXZYT rats (F) LEfSe analysis of the intestinal flora of 12QZRGF and 12GXZYT rats (G) Columnar distribution of KEGG functional abundance of the intestinal flora of the rats in each group (H) Columnar distribution of the functional abundance of the intestinal flora of COG in each group of rats.

Note: Taxonomic units are denoted as p (phylum), c (class), o (order), f (family), g (genus) and s (species).

Finally, functional prediction analyses based on the homology clustering (COG) family information of the proteome and the Kyoto Encyclopedia of Genes and Genomes (KEGG) homology (KO) information revealed that these cellular functions play a key role in the statistical differences in the composition of the diversity of the phyla, which involve function unknown, posttranslational modification, and the most characteristic bacterial phylum is Firmicutes. posttranslational modification, protein turnover, chaperones, cell motility, transcription, lipid transport and metabolism, carbohydrate transport and metabolism. It indicates that exploring the mechanism of herbal compounding for the treatment of NAFLD in terms of the interrelationships between bacterial flora and signal transduction is a feasible strategy. PICRUST2 was used for KEGG pathway analysis of the intestinal microbiota based on 16S rDNA gene amplicon sequencing results.53 Cluster analysis of the KEGG pathway at the first level showed that the function of the gut flora in the Model group was mainly in cellular processes and environmental information processing. In contrast, the functions of the gut microbiota in the Medicated group were mainly reflected in human diseases, biological systems, metabolism and genetic information processing (Figure 10G and H).

Discussion

Early NAFLD has spleen deficiency, and as the disease progresses, it gradually manifests as blood stasis and phlegm stasis, SLBZS treats spleen deficiency, GXZYT treats phlegm stasis and phlegm stasis; QZRGF softens the liver and strengthens the spleen, regulates qi and promotes blood circulation. The efficacy of each of them has its own focus, so the three parties were selected for the study. SLBZS21 consists of Dangshen, Fulin, Baizhu, Shanyao, Lianzi, Baibiandou, Yiyiren, Sharen, Jiegeng, Gancao; GXZYT consists of Danggui, Chishao, Chuanxiong, Taoren, Honghua, Wulingzhi, Danpi, Xiangfu, Wuyao, Zhike, Yanhuosuo, Gancao; QZRGF consists of Ezhu, Qingpi, Juhua, Baizhu, Shanzha, Lanhuashen, Sanqi. It was found54 that Fulin and Baizhu have anti-inflammatory and intestinal flora regulating effects. Taoren, Honghua, Zhike and Yanhusuo are capable of anti-inflammatory and antioxidant in order to protect the liver.55–58 Wulingzhi,59 Salvinorin,60 Xiangfu,61 Wuyao,62 Fulin,63 Gancao,63 Shanyao,64 Baizhu polysaccharide,65 Baizhu lactone,66 Ginsenosides,67 Shanyao polysaccharides,68 Chishao,69 Danggui,70 Jiegeng,71 Lianzi,72 Qingpi,73 Sanqi,74 Shanzha75 have found improvements in both gut flora. Acacetin76 can regulate the composition of intestinal flora, including increasing the content of probiotics and decreasing the content of harmful bacteria. Kaempferol77 can, at the genus level, the relative abundance of the genera Anabacterium spp. and Lactobacillus spp. Luteolin78 can increase the abundance of the genera of HFD rats Parvibacter, Faecalitalea, Allobaculum sp. and Bacteroids dorei relative abundance. Naringin, quercetin, baicalin, diosgenin, and licochalcone aincrease the diversity of intestinal flora and improve the abundance of beneficial flora.79–83 Nobiletin84 can improve intestinal permeability affecting the composition of the intestinal flora in models of intestinal disorders. AKT1,85 CASP3,86 FOS87 can improve the intestinal flora, enhance the mucosal barrier function and significantly increase the abundance of beneficial bacteria. The MAPK signaling pathway has been associated with the maintenance of intestinal microbial homeostasis.88,89 The intestinal MYC90 is involved in the regulation of systemic metabolism and improves hepatic steatosis and steatohepatitis induced by high-fat diet. Expression of STAT391 protects the intestinal mucosal barrier and alleviates intestinal flora imbalance. TP5392 is a key target for improving intestinal flora dysbiosis, and is closely associated with the production of SCFAs and improving inflammatory responses. We focused on the intersecting core targets of SLBZS, QZRGF and GXZYT for the treatment of NAFLD as MAPK1, AKT1, CASP3, FOS, TP53, STAT3, and MYC through the study of network pharmacology; and the core constituents as baicalein, luteolin, nobiletin, quercetin, kaempferol, naringenin, licochalcone a, diosgenin, acetin. Molecular docking of core targets and core components validated the results of the network pharmacology screening. Based on the molecular docking results, affinity < - 7.0 kcal-mol-1 indicates strong docking activity.52 We targeted AKT1- diosgenin (−9.51), AKT1- kaempferol (−7.38), AKT1- luteolin (−7.48), AKT1- quercetin (−7.4), CASP3- luteolin (−7.11), TP53- baicalein (- 8.41), TP53- luteolin (−8.48), TP53- nobiletin (−7.99), TP53- quercetin (−8.33). Among them, AKT1- diosgenin had the best docking activity. Therefore, we speculate that the core target AKT1, the active ingredient diosgenin may play an important role in the treatment of NAFLD.

According to traditional Chinese medicine theory, the large intestine is the organ of conduction, and its function depends on the liver’s ability to relieve stress. The liver and large intestine can be connected through the lung meridian, further demonstrating that the liver and large intestine are interconnected.93 In his book the Chinese Ming Dynasty physician Li Fu proposed the theory of “Zangfu (Viscera) Connectedness”, which states: “The heart is connected to the gallbladder, the liver is connected to the large intestine, the spleen is connected to the small intestine, the lungs are connected to the bladder, the kidneys are connected to the Sanjiao, and the kidneys are connected to the Mingmen”. He also proposed94 the treatment idea of “the liver is connected to the large intestine, and liver diseases should be unblocked from the large intestine”. With the concept of “the gut-liver axis” proposed, the association between the occurrence and development of NAFLD and gut microbiota has gradually attracted attention. Numerous studies have shown that gut microbiota plays an important role in the occurrence, development, and treatment of NAFLD.95 Factors such as intestinal blood influx promote the occurrence and development of NAFLD;96 Previous studies have shown that beneficial bacterial strains such as lactobacilli can alleviate symptoms of NAFLD by regulating body metabolism, reducing inflammatory responses, and restoring gut microbi