Exploring the Regulatory Mechanism of Modified Huanglian Maidong Decoction on Type 2 Diabetes Mellitus Biological Network Based on Systematic Pharmacology

Objective To explore the mechanism of modified Huanglian Maidong decoction (Maidong-Sanqi-Huanglian Compounds, MSHCs) intervention in type 2 diabetes mellitus (T2DM). Method This study used PubChem and SciFinder to collect the molecular structure of MSHCs, used PharmMapper to predict the potential targets of MSHC, and combined them with the T2DM gene to construct MSHC-T2DM protein-protein interaction (PPI) network. The plugin MCODE in Cytoscape 3.7.1 was then used to perform cluster analysis on the MSHC-T2DM PPI network. The genes and targets were input into DAVID for Gene Ontology (GO) and pathway enrichment analysis. Finally, animal experiments were performed to verify the therapeutic effect of MSHC on T2DM. Results Several T2DM-related targets, clusters, signaling pathways, and biological processes are found. The experimental results showed that compared with the blank group, the content of fasting blood glucose (FBG) in the model group was higher (P < 0.01). Compared with the model group, the content of FBG decreased and the insulin level increased in the MSHC medium-dose (0.15 g/kg) and high-dose (0.45 g/kg) groups and metformin group after 4 weeks of drug administration (P < 0.05). MSHC can also improve blood liquid levels and inflammatory factor levels (P < 0.05). Conclusion MSHC may achieve therapeutic effects through regulating the T2DM-related targets, biological processes, and pathways, such as insulin resistance, energy metabolism, oxidative stress, and inflammation, found in this research.


Introduction
Diabetes is a series of metabolic disorder syndrome caused by islet dysfunction, insulin resistance, etc., and often accompanied by a variety of complications [1,2]. At present, the global adult diabetic patients (aged 20-79) have increased from 151 million in 2000 to 425 million in 2017, a nearly double increase [3]. Due to urbanization, population aging, obesity, and other reasons, the number of patients with diabetes continues to grow rapidly. It is estimated that by 2045, diabetes patients may reach 629 million [4]. So far, the allopathic medicine was usually based on injection of insulin or oral glucose-lowering drugs such as metformin and acarbose. However, due to the expensive prices and side effects of these drugs, patient compliance has gradually decreased [5,6]. Traditional Chinese medicine (TCM) has a unique understanding of diabetes as a systemic endocrine disease. Meanwhile, TCM can well intervene and regulate diabetes from the perspective of the overall system [7,8]. Herbal medicine has gradually attracted attention due to its mildness, multiple targets, and fewer adverse reactions [9,10].
Huanglian Maidong decoction comes from Dou Ke Lei Bian Volume III (Written by Zhai Liang in the Ming Dynasty), which is composed of Coptis chinensis Franch. (CopidisRhizoma) and Ophiopogon japonicus (L. f.) Ker GawL (Ophiopogon japonicus). e herbal formula has been used in ancient times to treat Xiaoke Bing, namely, diabetes, and its effect is exact. Our previous research showed that modified Huanglian Maidong decoction (also called Ophiopogon japonicus-Panax notoginseng (Burk.) F. H. Chen ex C. Chow-Copidis Rhizoma Compounds (Maidong-Sanqi-Huanglian Compounds, MSHCs)) can significantly improve fasting blood glucose and glycated hemoglobin levels in patients with type 2 diabetes mellitus (T2DM) [11]. We hope to clarify its mechanism for treating T2DM through new strategies.
Due to the complex compounds and targets of TCM, the classical pharmacology presents difficulties analyzing its complex intervention mechanisms [12]. Hopkins proposed the concept of "network pharmacology" in 2007 and gave new ideas for TCM research, especially in combination with multidirectional pharmacology [13][14][15]. Network pharmacology can study the therapeutic mechanism of drugs from the holistic and systemic aspects of the interaction between Chinese medicine compounds, targets, and diseases.
is is particularly suitable for studying the role of TCM multi-componentsmulti-targets and is conducive to revealing its complex mechanism [16][17][18][19]. erefore, this study will combine the methods of experimental pharmacology and network pharmacology to explore the therapeutic mechanism of MSHC on T2DM.

MSHC's Potential Target Prediction and T2DM Gene
Collection.
e mol2 files of all compounds were input into PharmMapper (http://lilab-ecust.cn/pharmmapper/) to predict the potential targets of MSHC [24]. e reverse docking prediction results for each compound were downloaded and arranged in descending order according to the docking score (Z) values. e first 300 targets of each compound were collected for subsequent research. e GeneCards database (http://www.genecards.org/) and the OMIM database (http://www.ncbi.nlm.nih.gov/omim) were used to collect T2DM-related genes [25,26]. e search function of UniProt KB (https://www. uniprot.org/uniprot/) was utilized to import the PDB ID number of the MSHC target, with the species restricted to human. After the above operation, the "official symbol" of the MSHC target is obtained. e MSHC targets are shown in Table S1 and the T2DM genes are shown in Table S2 (see supplementary materials).

Network Construction and Analysis.
e data of proteinprotein interaction (PPI) were obtained from String 11.0 (http://string-db.org/) [27] and InAct (http://www.ebi.ac. uk/intact/) [28]. e T2DM gene, MSHC and potential targets, and PPI data were imported into Cytoscape 3.7.1 software (https://cytoscape.org/) for network construction [29]. e networks were analyzed by the plugin molecular complex detection (MCODE) to obtain cluster [14,15,[17][18][19]. MCODE can be used to discover closely related regions in the PPI network, and these regions may represent molecular complexes. According to the given parameters, dense regions are separated, which has advantages over other cluster methods because other methods rarely consider the rest of the network.

Gene Ontology (GO) and Pathway Enrichment Analysis.
DAVID (https://david.ncifcrf.gov/summary.jsp, Version 6.8) provides systematic comprehensive biological function annotation information for large-scale genes or proteins, which can find the most significantly enriched biological annotations [30]. e targets and genes were imported into the DAVID database, the Identifier was set to official gene symbol, the List Type was set to GeneList, and the species was restricted to "Homo sapiens." en, the Gene Ontology (GO) analysis and KEGG pathway analysis of MSHC targets and T2DM genes can be performed. e animal license number is SCXK (Xiang) 2016-0017. All rats were given conventional animal feed and water, and the feed was full-scale solid pellet feed for breeding rats (contains crude protein 18%, crude fat 5%, crude fiber 5%, crude ash 8%, amino acid 9.23%, calcium 1.00%, phosphorus 0.85%, potassium 0.55%, sodium 0.25%, vitamin A 17 IU/g, vitamin D 4 U/g, vitamin E 45 ppm, vitamin K 2.0 ppm, vitamin B12 0.08 ppm, etc.) (Liaoning Changsheng Biotechnology Co., Ltd.). All animals' care and experimental procedures were approved by the Ethical Committee on According to the method of reference [31], the diabetic rat model was made by intraperitoneal injection of STZ. One week after the adaptive feeding of rats, 10 rats were randomly selected as the blank group according to the random number table method, and the rest were intraperitoneally injected with STZ 0.055 mg/gbody weight. After 1 week, the fasting blood glucose (FBG) of the rats was examined. Rats with FBG ≥11.1 mmol/L were considered as a successful model. Rats with successful modeling were randomly assigned to 5 groups according to the random number table method: (1) model group (10 rats); (2) metformin group (10 rats); (3) MSHC high-dose group (10 rats); (4) MSHC medium-dose group (10 rats); (5) MSHC low-dose group (10 rats). e male to female ratio was 1 : 1 in all 6 groups. MSHC high-, medium-, and low-dose groups were administered intragastrically with MSHC at 0.45, 0.15, and 0.05 g/kg, respectively. e metformin group was given oral metformin 0.05 mg/kg intragastrically. e model group and the blank group were intragastrically administered with an equal volume of saline solution (0.9%). e intragastric administration volume in each group was 10 mL/kg.

Detection of FBG, Insulin, and Blood Liquid Levels.
Tail vein blood was taken 1 d before and 4 weeks after the administration, and FBG was detected by Blood Glucose Test Strips (glucose oxidase method). e FBG was detected strictly according to the kit instructions. After 4 weeks of drug administration, the rats were euthanized and blood were collected under 2% sodium pentobarbital anesthesia. e blood was allowed to stand for 60 minutes and centrifuged at 1000 g/min for 15 minutes to separate the serum. e TC, TG, HDL-C, and LDL-C levels were detected by 7180 automatic biochemical analyzer. e insulin level was detected by radioimmunoassay. e operation was strictly performed according to the kit instructions. Insulin and resistance index are calculated based on fasting blood glucose and fasting insulin, and the calculation formula is insulin resistance index (IRI) = (fasting blood glucose × fasting insulin)/22.5.

Determination of Inflammatory Factor Levels in Serum.
Serum TNF-α, IL-6, and MCP-1 levels were detected strictly in accordance with the ELISA kit instructions.

Pathological Observation.
After blood collection, the pancreas was harvested, fixed in 4% paraformaldehyde solution for 12 h, embedded in conventional paraffin, sectioned, and stained with hematoxylin and eosin (H&E). e pathological changes of the islet tissue were observed under an optical microscope.

Statistical Analysis.
e SPSS22.0 statistical software was used for analysis. e experimental data are expressed as mean ± SD. e data consistent with normality and homogeneity of variance were analyzed by analysis of variance (ANOVA). For multiple comparisons, the comparison between the two groups was performed using the LSD method. Non-parametric analysis was used for non-conformances. P < 0.05 indicates that the difference is statistically significant.

T2DM Genes and MSHC Potential Targets.
A total of 5719 T2DM genes were obtained. e genes with relevance score ≥9 were selected to construct the MSHC-T2DM PPI network. After the potential target protection, totally 420 human potential targets are collected (Figure 1). In the outer circle, purple, red, blue, and green represent T2DM genes, Ophiopogon polysaccharide targets, Coptis alkaloids targets, and PNS targets, respectively. In the inner circle, the greater the number of purple links and the longer the dark orange arc, the greater the number of overlapping genes. e blue link indicates the amount of functional overlap between the input target lists.

Clusters of MSHC-T2DM PPI Network.
e MSHC-T2DM PPI network was analyzed by MCODE and returns 15 clusters (Figure 3). en, the genes and targets were input into DAVID for GO enrichment analysis.
Cluster 1 is associated with calcium metabolism, ERBB2, and Wnt signaling. Cluster 2 is related to glucose metabolism, insulin and its related signaling pathways, vascular smooth muscle proliferation, and lipid metabolism. Cluster 3 is involved in hypoxia, glucose homeostasis, insulin resistance, adiponectin-related pathways, inflammation, and energy metabolism. Cluster 4 is related to cholesterol metabolism, glucose metabolism, lipid metabolism, and MAPK cascade. Cluster 6 is involved in glucose homeostasis and insulin secretion. Cluster 7 is related to redox. Cluster 15 is associated with lipid metabolism, glucose metabolism, and energy metabolism. Cluster 10 failed to return any human biological processes. Clusters 5,8,9,11,12,13, and 14 did not return any T2DM-related biological processes. e details are shown in Table S3.
Since the biological process of cluster 2 is representative, the main biological process in cluster 2 is selected as an example for demonstration ( Figure 4).

Effect of MSHC on Body Weight in Rats.
e weight difference of rats in each group before intervention was not significant (P > 0.05). After 4 weeks of intervention, compared with the model group and each treatment group, the differences were significant (P < 0.01). It shows that MSHC and metformin can effectively inhibit the weight loss trend of type 2 diabetic rats (Figure 7).

Effect of MSHC on FBG in Rats.
Compared with the blank group, the FBG in the model group was significantly increased before and after 4 weeks (P < 0.01). Compared with the model group, FBG was significantly reduced after 4 weeks of administration in the MSHC medium-and highdose groups and metformin group (P < 0.05) (Figure 8).

Effects of MSHC on Serum Insulin Level and IRI in Rats.
Compared with the blank group, the insulin level of the model group was significantly reduced, and the IRI was significantly increased (P < 0.01). Compared with the model group, the insulin level in the MSHC medium-and highdose groups and metformin group was significantly increased, and the IRI was significantly decreased (P < 0.01) (Figure 9).

Effects of MSHC on Serum TC, TG, HDL-C, and LDL-C in Rats.
Compared with the blank group, the TC, TG, and LDL-C levels of the model group increased, while HDL-C level decreased (P < 0.05). Compared with the model group, the TC, TG, and LDL-C levels in the MSHC medium-and high-dose groups and metformin group were significantly decreased, while HDL-C increased (P < 0.01) ( Figure 10). It shows that MCSH can reduce serum TC, TG, and LDL-C, increase HDL-C, and improve lipid metabolism in T2DM rats.   Evidence-Based Complementary and Alternative Medicine

Effects of MSHC on Serum Inflammatory Factor in Rats.
Compared with the blank group, the TNF-α, IL-6, and MCP-1 of the model group increased (P < 0.05). Compared with the model group, the TNF-α, IL-6, and MCP-1 in the MSHC medium-and high-dose groups and metformin group were significantly decreased (P < 0.05 or 0.01) (Figures 11 and 12).

Pathological
Changes. e islet β cells of the blank control group are tightly arranged, the size is the same, the nuclear chromatin is clear, and the cytoplasmic boundary is clear; the islet β cells of the model group are loosely arranged, the cell sizes are different, and the cytoplasmic vacuolation is serious. Compared with the model group, the pancreatic islet β cells in the MSHC medium-and low-dose groups are less loosely arranged and cytoplasmic vacuolation is less. e morphological structure of pancreatic islet β cells in MSHC high-dose group and metformin group showed that the cell size was more uniform and the arrangement and the cytoplasm was more uniform. It shows that MSHC has a protective effect on the islet β cells of T2DM rats (Figure 13).

Discussion
Insulin resistance is the main pathological basis of T2DM, which refers to the decrease of insulin utilization rate and decrease of the body's sensitivity to insulin [2,32]. Current research suggests that the main cause of insulin resistance is an increase in inflammatory cytokines, which interferes with the normal phosphorylation of IRS in insulin signaling pathway and blocks a series of cascade amplification reactions activated by downstream signals, thereby affecting physiological functions such as insulin production and transport [4,33,34].
Inflammatory cytokines and other factors can interfere with the normal insulin signaling pathway and activate the insulin resistance signaling pathway, leading to insulin resistance and T2DM [35,36]. Currently, the insulin-related signaling pathways those have been deeply studied are PI3K signaling pathway and Ras-RafMAPK signaling pathway [37,38]. Meanwhile,insulin resistance-related    [39][40][41][42][43]. e essence of the insulin signaling pathway and insulin resistance signaling pathway is the result of joint effects of inflammatory cytokines, transcription factors, phosphorylation of various protein kinases, and protein-protein interactions [42,43]. In this study, through the T2DM gene collection, MSHC target prediction, and the construction and analysis of networks such as the MSHC-T2DM PPI network, the potential regulatory mechanism of MSHC on T2DM disease networks was discovered. For example, we found that MSHC can regulate various functional modules (namely, cluster) in the T2DM's complex bionetwork. Each module contains many biological processes, suggesting that MSHC may insulin resistance and insulin signaling pathways. e energy metabolism-related signaling pathways are the adipocytokine signaling pathway and the PPAR signaling pathway. e oxidative stress-related signaling pathways include FoxO signal pathway, AMPK signal pathway, and so on. e inflammatory-related signaling pathways include PI3K-Akt, Wnt, TGF-β, TNF signaling pathways, and so on. ese signaling pathways mediate all aspects of the development of T2DM. ey are both independent and connected to each other and form a close regulatory network through extensive interaction to jointly promote the development of T2DM.
Current research also shows that Coptis alkaloids can improve insulin resistance, promote insulin secretion, inhibit gluconeogenesis in the liver, stimulate glycolysis of surrounding tissue cells, reduce intestinal absorption of glucose, and regulate lipid metabolism by inhibiting oxidative stress and inflammation in a variety of tissues including the liver, adipose tissue, kidney, and pancreas [44][45][46][47][48]. Its specific mechanism mainly involves a variety of cell kinases and signaling pathways, such as the AMPK, MAPK, Nrf2, and NF-κB signaling. Ophiopogon polysaccharide can reduce pancreatic islet damage in diabetic mice induced by alloxan [49]. Ophiopogon polysaccharide can also reduce FBG levels and increase oral glucose tolerance in diabetic mice [50]. It can also reduce insulin resistance by regulating the InsR/IRS-1/PI3K/Akt/GSK-3/Glut-4 signaling pathways [51,52]. e results of these studies also illustrate the accuracy of target prediction based on the pharmacophore-based response molecule docking technology.

Conclusion
In summary, this study used TCMSP, GeneCards, Cytoscape, and other databases/tools to study the mechanism of MSHC for T2DM from multiple aspects (biological processes, pathways, and so on). e results of animal pharmacological experiments show that MSHC can reduce FBG and increase basal insulin levels in diabetic rats; MSHC can also improve blood liquid levels and inflammatory factor levels.

Data Availability
e data used to support the findings of this study are included within the article and the supplementary information files.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
Qi He, Tianqing Zhang, Yonghe Wu, Jiamin Wu, Bing Jin, and Shiwei Wu were responsible for the study concept and design. Qi He, Tianqing Zhang, Yonghe Wu, Jiamin Wu, and Shiwei Wu were responsible for data analysis and interpretation in the systematic pharmacological section. Pu Gao and Bing Jin were responsible for data analysis and interpretation in experiments. Qi He, Tianqing Zhang, Yonghe Wu, and Jiamin Wu drafted the paper. Shiwei Wu supervised the study. All authors participated in the analysis and interpretation of data and approved the final paper. Qi He, Tianqing Zhang, and Bing Jin should be considered co-first authors. Table S1: potential targets for each compound. Table S2: type 2 diabetes mellitus genes. Table S3: enrichment analysis of clusters based on Gene Ontology (GO) annotation.  Evidence-Based Complementary and Alternative Medicine 11