Modified Chaishao Liujunzi Decoction inhibits bile acid-induced gastric intestinal metaplasia: from network prediction to experimental verification

Chemical profiling in the water extract of MCLD identified by UPLC-Q-TOF-MS/MS

MCLD was prepared from Radix bupleuri, Radix Paeoniae Alba, Codonopsis pilosula, Poria cocos, Atractylodes macrocephala, Pinellia ternata, tangerine peel, Coix seed, Fructus aurantii, Rhizoma cyperi, Radix notoginseng, Rhizoma zedoariae, and Herba hedyotis diffusa at a ratio of 10: 12: 15: 15: 15: 9: 10: 20: 10: 10: 3: 9: 20. Next, we obtained the total ion chromatograms (TIC) of MCLD using ultraperformance liquid chromatography coupled with quadrupole time of flight tandem mass spectrometry (UPLC-Q-TOF-MS/MS) analysis, as shown in Figure 2. By comparing retention time, precise relative molecular weight, mass charge ratio, and secondary mass spectrometry fragmentation information provided by the chromatogram with the literature, we identified the stable components of MCLD (Supplementary Table 1), which included quercetin, kaempferol, isorhamnetin, luteolin, and naringenin.

Figure 2. Mass spectrum chromatograms of MCLD. (A) Positive total ion chromatogram of MCLD; (B) Negative total ion chromatogram of MCLD.

Screening of active compounds and related target proteins

Oral bioavailability (OB) and drug similarity (DL) are two important parameters to evaluate the absorption, distribution, metabolism, and excretion (ADME) characteristics of drugs in the human body [21]. Based on the UPLC-Q-TOF-MS/MS identification results and database screening results, we obtained 141 MCLD-active compounds using the parameters of OB and DL (Supplementary Table 2). Using PubChem and SwissTargetPrediction databases, 815 target proteins corresponding to active compounds of MCLD were predicted (Supplementary Table 3). Furthermore, 1255 GIM-related target genes were obtained from the OMIM and GeneCards databases, as shown in Figure 3A. A total of 227 MCLD and GIM intersection target proteins were obtained using Venny v.2.1 (Figure 3B), representing the potential target proteins for drugs that act on GIM evaluated in this study. By constructing the drug-compound-common target network (Figure 3C), we obtained unique active ingredients with high degree values, including quercetin, kaempferol, isorhamnetin, luteolin, and naringenin. Therefore, we consider the above compounds as the main active ingredients of MCLD.

Figure 3. GIM-related targets, common targets, and drug-compound-common target network. (A) GIM target information is obtained through relevant databases. (B) Venny diagram of common targets for MCLD and GIM. (C) Drug-compound-common target network. The red nodes represented the targets of MCLD compounds, whereas the blue and yellow nodes represented MCLD compounds. MOL000098: Quercetin; MOL000422: Kaempferol; MOL000354: Isorhamnetin; MOL000006: Luteolin; MOL004328: Naringenin.

Protein-protein interaction network analysis and core target screening

To further understand the interaction between potential target proteins, we combined the STRING database and Cytoscape software to build a protein-protein interaction (PPI) network diagram (Figure 4A), where the degree value size reflected the correlation between nodes, and the size and intensity of the color of the nodes in the network were proportional to the degree value size. After two median screenings based on the relevant parameters of the topological analysis, 33 core action targets with strong interaction relationships were obtained (Figure 4C), suggesting that these may play an important biological role in the GIM intervention. According to the degree value, the top 10 core targets with the strongest correlations were AKT1, TNF, IL6, VEGFA, EGFR, STAT3, IL1B, SRC, CASP3, and MAPK3 (Figure 4B).

Figure 4. Analysis of the PPI network of potential target proteins and selection of core targets. (A) PPI network diagram of the GIM-related protein of the MCLD intervention constructed by Cytoscape software. (B) The top 10 core targets were selected according to degree value. (C) Selection of core targets based on the relevant topology analysis parameters.

Enrichment analysis using gene ontology and Kyoto encyclopedia of genes and genomes tools

To better understand the potentially complex mechanism of MCLD acting on GIM, we used Metascape and Bioinformatics online tools to conduct Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses using the potential target proteins identified above and visualize the top 20 enrichment results with the highest significance according to the log P-value (Figure 5 and Supplementary Table 4). The results of the GO enrichment analysis suggested that 227 potential targets were enriched in a variety of biological processes related to GIM regulation, including positive regulation of cell migration, cellular response to nitrogen compounds, positive regulation of protein phosphorylation, and positive regulation of the MAPK cascade (Figure 5A) [22–24].

Figure 5. Enrichment analysis of the GO and KEGG pathways of potential targets. (A) Bubble diagram of the top 20 GO enrichment analyses identified according to biological process. The size and color of the bubbles represented the number of genes and the threshold value of the P-value of potential target enrichment in the biological process. (B) The top 20 pathways enriched by the KEGG pathway. The x-axis represents the number of targets enriched in the pathway, and the color represents the P-value.

The enrichment analysis of the KEGG pathways showed that 137 potential targets were enriched in the top 20 pathways (Figure 5B). Interestingly, the PI3K/Akt signaling pathway, FoxO signaling pathway, and MAPK signaling pathway have been reported to be closely related to GIM, and intervention in these pathways can effectively improve GIM [4, 25, 26]. Combined with the core targets and the available literature, EGFR, and mTOR have been reported to be the upstream and downstream targets of the PI3K/Akt signaling pathway, respectively, and the EGFR/PI3K/Akt/mTOR pathway is an important pathway related to cell proliferation, growth, and differentiation. TCM prescriptions have been reported to effectively intervene in PLGC (including chronic atrophic gastritis [CAG], GIM, and dysplasia) via the PI3K/Akt/mTOR or EGFR/PI3K/Akt pathways [4, 27]. Therefore, whether MCLD acts on GIM through the EGFR/PI3K/Akt/mTOR pathway became the focus of experimental validation studies.

Molecular docking

Using molecular docking simulations, the interaction between the active components of drugs and target proteins can be identified. In this study, we used the main active components of MCLD, quercetin, kaempferol, isorhamnetin, luteolin, and naringenin, to define the binding properties with the top ten core targets (AKT1, TNF, IL6, VEGFA, EGFR, STAT3, IL1B, SRC, CASP3, and MAPK3). In general, the lower the binding energy between the ligand and the receptor, the more stable the binding conformation [28]. Binding energy less than -4.25 kcal/mol (1 kcal ≈418,585 kJ), -5.0 kcal/mol, or -7.0 kcal/mol indicates a certain, good, or strong binding activity between the ligand and the receptor, respectively [28]. Our research results indicate that the main active components of MCLD have a good binding ability to the ten core targets mentioned above (Table 1). Among them, the binding energies of EGFR with kaempferol, isorhamnetin, quercetin, luteolin, and naringenin are -8.9, -9.0, -8.7, -9.0, and -8.8 (kcal/mol), respectively. The visualization results are shown in Figure 6, indicating that the high binding activity of the main active components of MCLD with EGFR may support the potential ability of MCLD to regulate the EGFR/PI3K/Akt/mTOR pathway predicted by network pharmacology.

Figure 6. Molecular docking models that combined the main active ingredients with the core target EGFR. EGFR-PDB ID: 5UG9. Green dotted line: Conventional Hydrogen Bond; Light green dotted line: Carbon Hydrogen Bond; Purple dotted line: Pi Sigma; Light purple dotted line: Pi Alkyl. (A) Kaempferol. (B) Isorhamnetin. (C) Quercetin. (D) Luteolin. (E) Naringenin.

Effects of DCA and MCLD on the viability/proliferation of GES-1 cells

To evaluate the effects of DCA and MCLD on GES-1 human gastric mucosal epithelial cell viability, cells were treated with different doses of DCA (100–800 μM) and MCLD (25–600 μg/mL) at different times, followed by cell viability measurements using the MTT assay. Our results indicate that, compared to the blank control group, stimulation of GES-1 cells with 100 μM DCA over any period of time can enhance cell viability and promote cell proliferation. Treatment with 200 μM DCA also exhibited proliferative activity after stimulating cells for 12 or 24 h, but cell viability decreased significantly after 48 h, whereas 400 μM or above concentrations of DCA showed significant cytotoxicity at any time period (Figure 7A). Regardless of the treatment period, 200 μg/mL or lower concentrations of MCLD did not have an effect on cell viability (Figure 7B). To avoid the toxic effect of drugs on cells, low dose DCA (≤200 μM) and MCLD (≤200 μg/mL) were selected to treat GES-1 cells for 24 h in subsequent experiments.

Figure 7. Effects of different concentrations of DCA and MCLD on the viability of GES-1 cells at different times and DCA induced the expression of CDX2 and MUC2 mRNA and protein in GES-1 cells. (A) Effects of different concentrations of DCA on the viability of GES-1 cells at different times. (B) Effects of different concentrations of MCLD on the viability of GES-1 cells at different times. (C) Effects of DCA on the expression of CDX2 and MUC2 mRNA in GES-1 cells detected by RT-qPCR. (D, E) Effects of DCA on the expression of CDX2 and MUC2 proteins in GES-1 cells detected by western blotting. Data are expressed as mean±standard deviation (SD). Compared to the control group, ^*P <0.05, ^**P <0.01, ^***P <0.001.

DCA induced the expression of CDX2 and its downstream intestinal marker MUC2 in GES-1 cells

GES-1 cells were treated with different concentrations of DCA (0, 50, 100, and 150 μM) for 24 h, and then the expression of CDX2 and its downstream intestinal marker MUC2 were detected by RT-qPCR and western blotting. As shown in Figure 7C–7E, treatment of GES-1 cells with DCA increased the expression of CDX2 and MUC2 mRNA and proteins in a dose-dependent manner, among which the 150 μM DCA treatment showed a significantly higher expression than the blank control group (P <0.01). These results showed that DCA could induce the expression of CDX2 and the intestinal marker MUC2 in GES-1 cells, and thus 150 μM DCA was selected for further study.

MCLD inhibited DCA-induced hyperproliferation of GES-1 cells and levels of inflammatory cytokines

To determine whether MCLD could improve DCA-induced GES-1 cell proliferation and levels of inflammatory cytokines, GES-1 cells were treated with 150 μM DCA for 24 h in the presence or absence of MCLD (50, 100, and 150 μg/mL), respectively. The results of MTT and ELISA showed that the proliferation activity of cells and the level of inflammatory markers (including interleukin [IL]-6, IL-1β, and tumor necrosis factor α [TNF-α]) induced by 150 μM DCA were significantly higher than those of the control group (P<0.01). After administering MCLD, the above indicators were inhibited, with 150 μg/mL MCLD treatment showing particularly significant (P<0.01) inhibition, as shown in Figure 8A–8D. These results suggest that MCLD may have a certain inhibitory effect on DCA-induced GES-1 cell proliferation and chronic inflammation.

Figure 8. MCLD inhibited DCA-induced GES-1 cell hyperproliferation and levels of inflammatory cytokines. (A) The effect of MCLD on DCA-induced proliferation of GES-1 cells. (B–D) Effects of MCLD on the levels of inflammatory cytokines induced by DCA. Data are expressed as mean±SD (n=3). Compared to the control group, ^###P <0.001; Compared to the DCA group, ^*P <0.05, ^**P <0.01, ^***P <0.001.

MCLD inhibited the target genes of the EGFR/PI3K/AKT/mTOR pathway and the expression of CDX2 and MUC2 mRNA

Based on the results of the previous network pharmacology and molecular docking analyses, we preliminarily validated the effect of MCLD on target genes of the EGFR/PI3K pathway and the expression of CDX2 and MUC2 mRNA. The expression of these genes was detected by RT-qPCR. As shown in Figure 9, 150 μM DCA promoted the expression levels of EGFR, PI3K, AKT, mTOR, CDX2, and MUC2 mRNA, and 150 μg/mL MCLD significantly inhibited the above situation. In a subsequent experiment, 150 μg/mL MCLD was selected for further validation of the mechanism.

Figure 9. MCLD inhibited the target genes of the EGFR/PI3K/AKT/mTOR pathway and the expression of CDX2 and MUC2 mRNA. (A) Effects of MCLD on the expression of DCA-induced CDX2 and MUC2 mRNA. (B–E) Effects of MCLD on the target genes of DCA-induced EGFR/PI3K/AKT/mTOR pathway. Data are expressed as mean±SD (n=3). Compared to the control group, ^###P <0.001; Compared to the DCA group, ^*P <0.05, ^**P <0.01, ^***P <0.001.

MCLD inhibited DCA-induced GES-1 cell proliferation and GIM through the EGFR/PI3K/AKT/mTOR pathway

To further determine whether MCLD could improve GIM by regulating the EGFR/PI3K/AKT/mTOR pathway, as predicted by our network pharmacology findings, we treated GES-1 cells with different combinations of 150 μM DCA, 150 μg/mL MCLD, 100 μM EGFR agonist (NSC228155), and 2 μM EGFR inhibitor (Erlotinib) for 24 h (Figure 10A). Western blotting analysis showed that, compared to the control group, DCA treatment significantly increased the protein expression of CDX2, MUC2, p-EGFR, PI3K, p-Akt, p-mTOR, as well as Ki67 and Cyclin D1 proteins related to cell proliferation, while MCLD and EGFR inhibitor negatively regulated the high expression of DCA-induced proteins, contrary to the result of EGFR agonist treatment. Significantly, compared to the EGFR inhibitor, MCLD exhibited a stronger inhibitory ability on the expression of the aforementioned proteins. Furthermore, compared to DCA-induced cells treated with EGFR agonist, MCLD combined with EGFR agonist significantly decreased the expression of the above proteins, as shown in Figure 10B. These data suggest that MCLD may inhibit DCA-induced cell proliferation and GIM through the EGFR/PI3K/AKT/mTOR pathway.

Figure 10. MCLD inhibited cell proliferation and GIM through the EGFR/PI3K/AKT/mTOR pathway. (A, B) Western blotting was used to detect the protein expression of EGFR, p-EGFR, PI3K, p-AKT, AKT, mTOR, p-mTOR, Ki67, Cyclin D1, CDX2, and MUC2 in DCA-induced GES-1 cells. Data are expressed as mean±standard deviation (SD) (n=3). Compared with the control group, ^###P <0.001; Compared with the DCA group, ^*P <0.05, ^**P <0.01, ^***P <0.001; Compared with the DCA+NSC228155 group, ^{^}P <0.05, ^{^ ^}P <0.01, ^{^ ^ ^}P <0.001.

Drug preparation

MCLD was provided by the Affiliated Hospital of Changchun University of Traditional Chinese Medicine and was prepared from Radix bupleuri, Radix Paeoniae Alba, Codonopsis pilosula, Poria cocos, Atractylodes macrocephala, Pinellia ternata, tangerine peel, Coix seed, Fructus aurantii, Rhizoma cyperi, Radix notoginseng, Rhizoma zedoariae, and Herba hedyotis diffusa at a ratio of 10: 12: 15: 15: 15: 9: 10: 20: 10: 10: 3: 9: 20. After various Chinese herbal medicines were combined, an aqueous extract of MCLD was obtained using the Chinese herbal decoction method. The decoction was then filtered and centrifuged, and a rotary evaporator and a freeze-dryer were used successively to obtain the freeze-dried powder sample. Before the experiment, the freeze-dried powder was diluted in the culture medium according to the proportion required for the experiment. DCA (Merck Millipore, USA) is a cytotoxic bile acid that induces GIM. NSC228155 (MedChemExpress, USA) is an EGFR agonist. Erlotinib (MedChemExpress, USA) is an EGFR inhibitor. All treatments were dissolved in dimethylsulfoxide (DMSO) for the in vitro studies.

Screening of active compounds and related target proteins from MCLD

We first accurately weighed MCLD and added 500 μL extraction solution (methanol:water=4:1, internal standard concentration of 10 μg/mL). After ultrasonic treatment in an ice water bath, the solution was centrifuged for 15 minutes at a speed of 12,000 rpm and a temperature of 4° C. The supernatant was then filtered using a 0.22 μm filter membrane to prepare each test sample separately. Next, the components of MCLD were identified using UPLC-Q-TOF-MS/MS [4]. Based on two important parameters of ADME (OB and DL), the components were screened using the database of Traditional Chinese Medicine Systems Pharmacology (TCMSP). Compounds that met the criteria of OB ≥30% and DL ≥0.18 criteria were considered active candidate compounds in MCLD [49]. Subsequently, the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) [50] and the Swiss Target Prediction database (http://www.swisstargetprediction.ch/) [51] were combined to predict relevant targets corresponding to active compounds, and the UniProt database (http://www.uniprot.org/) [52] was used to standardize the target genes, after which duplicate targets were deleted.

Identification of GIM-related target proteins

The GeneCards database (https://www.genecards.org) [53] and the OMIM database (https://www.omim.org/ and https://www.omim.org/search/advanced/geneMap) [54] were used to identify GIM-related target proteins. The targets obtained were based on GeneCards database scores greater than the median as the filter condition. We then merged and sorted the GIM-related target information obtained from the above databases and deleted duplicate targets.

Construction of the protein-protein interaction network and screening of the core targets

We used Venny online software (https://bioinfogp.cnb.csic.es/tools/venny/index.html) to obtain potential targets for MCLD that act on GIM and entered these targets into the STRING database tool (https://string-db.org/) [55] to construct the PPI. Here, the species was limited to Homo sapiens, and the confidence score was set at greater than 0.400. Next, we saved the PPI results in the TSV file format and imported them into Cytoscape V3.7.2 [56] for the construction and analysis of the PPI network. Finally, the CytoNCA plug-in was used for the topological analysis of relevant parameters, and targets with values greater than the median of degree centrality, betweenness centrality, and closeness centrality values were taken as core targets.

Bioinformatics analysis based on target proteins

Metascape (http://metascape.org/gp/index.html) [57] is a full-featured gene function annotation and analysis platform that updates data monthly to improve their reliability. The platform was used for enrichment analyses of the GO and KEGG pathways. The study set species as “Homo sapiens” and P <0.01 as screening conditions for enrichment analysis and we used the online bioinformatics tool (http://bioinformatics.com.cn/) to visualize the enrichment results.

Molecular docking between main active ingredients and core targets

To clarify the interaction between the main active ingredients and the pathway target proteins predicted by network pharmacology, we used AutoDock Vina software (http://vina.scripps.edu/) [58] to perform molecular coupling analyses using the main active ingredients and the potential pathway target proteins. First, we obtained the three-dimensional crystal structure of the target proteins from the Protein Data Bank (PDB) database (https://www.rcsb.org/) and performed pretreatment evaluations, including water removal, hydrogenation, charge calculation, and storage conditions (PDB format). The structures of the compounds obtained from the PubChem database were optimized and saved in pdb format. Finally, AutoDock Vina was used to calculate the docking activity of target and compound structures. The lower the binding energy, the more stable the binding conformation of the receptor and the ligand [28].

Cell culture

GES-1 human gastric mucosal epithelial cells were purchased from Fenghui Biotechnology Co., Ltd. (Hunan, China) and identified using short tandem repeat DNA analysis. GES-1 cells were cultured in Dulbecco's Modified Eagle Medium (Thermo Fisher Scientific, USA) containing 10% fetal bovine serum (FBS) (Thermo Fisher Scientific, USA) and supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin. Subsequently, they were incubated in a humidified incubator at 37° C and 5% CO₂, and cells with logarithmic growth period and good condition were taken for subsequent experiments. To explore the effects of MCLD and its mechanism of action, GES-1 cells from the control group were cultured in DMEM medium for 24 h. Meanwhile, cells from the experimental group were cultured with different combinations of DCA (50, 100, and 150 μM), MCLD (550, 100, and 150 μg/mL), erlotinib (2 μM), and NSC228155 (100 μM) for 24 h.

Determination of cell viability

GES-1 cells were seeded in 96-well plates at a density of 1×10⁴ cells/well. After the cells adhered to the plates, we added different concentrations of DCA (100–800 μM) and MCLD (25–600 μg/mL) to incubate the cells at different times and set three wells for each concentration. Next, after removing and discarding the culture medium, we added a 10 μL MTT solution (Solarbio, Beijing) for 4 h to obtain a precipitate. The resulting solution in each well was discarded and 150 μL DMSO was added. Subsequently, the plates were placed on a table shaker at room temperature + 5° C–60° C for 10 min before measuring the absorbance of each well at 490 nm using a microplate reader (BioTek, USA).

Enzyme-linked immunosorbent assays

GES-1 cells were inoculated in a 6-well plate at a density of 1×10⁶ cells/well for 24 h. After the cells adhered to the plate, GES-1 cells were treated with 150 μM DCA for 24 h in the presence or absence of MCLD (50, 100, and 150 μg/mL). Next, we collected the cell supernatant, which was centrifuged before performing the ELISA with strict adherence to the manufacturer’s instructions to determine the contents of IL-6, IL1β, and TNF-α in the cell supernatant.

Total RNA extraction and RT-qPCR

TRIZOL^® reagent (TIANGEN, Beijing, China) was used to extract total RNA from each treatment group of GES-1 cells. A reverse transcription reaction solution was then prepared to reverse transcribe the RNA into cDNA. RevertAid reverse transcriptase was purchased from Thermo Fisher Scientificm USA. Next, SYBR Green Master Mix (Thermo Scientific, USA) and a Real-Time PCR Detection system were used for qPCR. Finally, using GAPDH as the internal reference gene, we used the 2^-∆∆Ct method to calculate the relative mRNA expression of each gene to be tested [28]. The PCR primers were as follows: CDX2, 5’-GGTTTCAGAACCGCAGAGCA-3’-(Forward), and 5’-GAAGACACCGGACTCAAGGG-3’-(Reverse); MUC2, 5’-TCAAAAGCAGCGTGTTCAGC-3’-(Forward), and 5’-AGCAGAAGCACTCACAGTCC-3’-(Reverse); EGFR, 5’-CAGATCGCAAAGGGCATGAA-3’-(Forward), and 5’-TTGCCTCCTTCTGCATGGTA-3’-(Reverse); PI3K, 5’-TTATAAACGAGAACGTGTG-3’-(Forward), and 5’-AATAGCTAGATAAGCC-3’-(Reverse); AKT, 5’-CAGCATCGCTTCTTTGCCGGTA-3’-(Forward), and 5’-CCTGGTGTCAGTCTCCGACGTGA-3’-(Reverse); mTOR, 5’-GTGGTGGCAGATGTGCTTAG-3’-(Forward), and 5’-TTCAGAGCCACAAACAAGGC-3’-(Reverse); GAPDH, 5’-CACCATCTTCCAGGAGCGAGA-3’-(Forward), and 5’-CATGACGAACATGGGGGCAT-3’-(Reverse).

Western blotting analysis

Proteins were extracted from treated GES-1 cells using a radioimmunoprecipitation assay (RIPA) buffer (Beyotime, Shanghai, China) containing protease and phosphatase inhibitors. A BCA protein concentration determination kit (Beyotime, Shanghai, China) was then used to determine the protein concentration in each group. Proteins were separated by SDS-PAGE and transferred to a methanol-pre-activated PVDF membrane (Merck Millipore, USA). Finally, the PVDF membrane was completely immersed in 5% milk-PBST at room temperature for 60 min for blocking, and then the membrane and the first antibody were incubated overnight at 4° C. A secondary antibody was added to the membrane after washing for 1 h the next day. After washing again, the enhanced chemiluminescence (ECL) method was used to image the chemiluminescence (Shanghai Qinxiang Scientific Instrument Co., Ltd., China) and densitometric analysis of the protein bands was performed using ImageLab software. The primary antibodies used were anti-EGFR (# 4267, CST), anti-p-EGFR (# 3777 CST), anti-PI3K (# 60225-1-Ig, Protentech), anti-AKT (# 4691, CST), anti-p-AKT (# 4060, CST), anti-mTOR (# 2983, CST), anti-p-mTOR (# 5536, CST), anti-CDX2 (# ab76541, Abcam), anti-MUC2 (# ab272692, Abcam), anti-Ki67 (# ab92742, Abcam), anti-Cyclin-D1 (# ab134175, Abcam), and anti-GAPDH (# 60004-1-Ig, Protentech).

Statistical analysis

Statistical analysis of all data was performed using the SPSS software (v.19.0, IBM Corp., Armonk, NY, USA). The experimental data were compared using one-way analysis of variance and expressed as mean±standard deviation (SD). Differences were considered statistically significant at P-values < 0.05.