Discussing the Mechanism of Dahuang Huanglian Xiexin Decoction in the Treatment of Type 2 Diabetes Mellitus via Network Pharmacology and Molecular Docking

Background This study used network pharmacological analysis and molecular docking approach to explore the mechanism of DHXD on T2DM. Firstly, the compounds in DHXD were obtained from TCMSP and TCMID databases, the potential targets were determined based on TCMSP and UniProt databases. Next, Genecards, Digenet and UniProt databases were used to identify the targets of T2DM. Then, the protein-protein interaction (PPI) network was established with overlapping genes of T2DM and compounds, and the core targets in the network were identied and analyzed. Then, the David database was used for GO and KEGG enrichment analysis. Finally, the target genes were selected and the molecular docking was completed by Autodock software to observe the binding level of active components with target genes. b action with action quercetin with target d action of target action beta-sitosterol (MOL000358) with target GLUT4;


Background
Type 2 diabetes mellitus (T2DM), as a common chronic metabolic disease globally, has brought huge burden to social nance and health, which the number of patients with T2DM will increase to 693 million by 2045 [1]. Its main characteristics of persistent hyperglycemia and obesity. Due to the progress of the disease, it will produce a series of complications to damage health, such as large or small blood vessels damage, diabetic peripheral neuropathy, diabetic foot, diabetic nephropathy, and even cardiovascular and cerebrovascular diseases [2]. It has discovered that the core pathogenesis of T2DM is insulin resistance and islet cell damage [3]. There are many therapeutic schemes in clinic, such as exercises, diet control [4], oral medication, insulin injection and stem cell replacement therapy [5]. But, due to the high costs and uncurable results, scientists accelerate the pace of exploring the mechanism of T2DM and develop new drugs and alternative therapy.
With a high safety and small side effects, Traditional Chinese Medicine (TCM) originated in ancient China, have been veri ed providing effective and safety methods for the treatment of T2DM. Through systematic analysis and meta-analysis, it was found that TCM could intervene T2DM by regulating intestinal ora [6]. Some studies have shown that Liuwei Dihuang pill is effective in the treatment of T2DM. The mechanism may be that quercetin and β -sitosterol act on VEGFA, MAPK1 and other targets through AGE-RAGE, TNF, NF -κ B and other signaling pathways [7]. It has been reported that berberine, quercetin and other major components in Huanglian Jiedu Decoction can act on AKT1, VEGFA and other targets, regulate apoptosis, in ammatory response and other biological processes, treat T2DM through AGE-RAGE, insulin resistance and other signaling pathways [8]. Under the action of berberine, the core component of Gegen Qinlian decoction, it can signi cantly improve the overall structure of intestinal ora, increase the number of butyrate-producing bacteria, so as to reduce intestinal in ammation and reduce blood glucose [9]. Some articles also mentioned that quercetin, rhein, berberine and other effective components of Da-chai-hu decoction may intervene in T2DM through TNF signaling pathway and PI3K / Akt signaling pathway [10].
DHXD is derived from Treatise on Febrile Diseases, it is used to treat spleen stomach excess heat syndrome, which is just to cut the pathogenesis of Pi Dan disease in TCM. It has been proved that Huanglian (coptis chinensis, HL) and Dahuang (rhubarb, DH) in DHXD play an important role in regulating the disorder of glucose and lipid metabolism, and are commonly used in the treatment of T2DM in clinic [11]. However, The mechanism of DHXD on T2DM is not clear, so this study tries to explain it from the perspective of network pharmacology. This paper used network pharmacology to establish a herb-compound-target-pathway network, carried out Gene Ontology (GO) and Kyoto Encyclopedia of genes and genomes (KEGG) analysis. It further veri ed the potential chemical components and core targets through molecular docking, in order to try to clarify the mechanism of DHXD on T2DM. The detailed ow chart of the research design is shown in Fig. 1 [13], one of the largest comprehensive TCM platforms, to determine the chemical components of DHXD.

Screening of active ingredients OB assessment
Oral bioavailability (OB) is one of the most important pharmacokinetic parameters in drug screening, which refers to the percentage of oral drug components reaching systemic circulation. it obtained the OB value from the OBioavail1.1 system of TCMSP platform [14], which has the concordancy of P450, 3A4 and P-glycoprotein information. Simultaneously, it set the OB threshold at 30% for determining the active ingredients used for the subsequent research.

DL prediction
Drug-likeness (DL) refers to the "drug like" degree of the target compound, which is used to remove the chemical inappropriate compounds, and is a qualitative indicator [12]. Comparing the target compounds with all the molecules from TCMSP platform and the DL index was calculated by Tanimoto similarity method, compounds with DL ≥ 0.18 were included.

Target prediction of active ingredients
The compound components in the prescription in uence on the biological function of the target. Therefore, it use TCMSP platform to predict the target of active ingredients, and Uniprot Database (Uniprot, http://www.uniprot.org) [15] was used to screen the target of "Homo sapiens".

Common targets of DHXD and T2DM
The intersection of drug target and disease target was represented by Venn diagram. It set the condition as "Homo sapiens" to construct PPI network with common targets in STRING database (STRING, http://www.string-db.org) [5].

Key Genes screening of common targets of DHXD and T2DM
All the genes were screened by CytoHubba program of Cytoscape 3.7.1 software[18], following the herbcomponent-target network constructed.
Page 5/26 2.7 Biological process and pathway enrichment analysis DAVID Bioinformatics Resources 6.8 database (DAVID, https: //david.ncifcrf.gov/) [19]was used for the GO and KEGG pathway enrichment analysis of common targets [20]. The screening was carried out under the condition of "p < 0.01, gene number > 8", and Bubble Plot was drawn by R Studio software. In addition, following common genes labeled, it draw the relevant path diagram according to the KEGG Mapper (https://www.genome.jp/kegg/mapper.html) [21].

Component-target molecule docking
It used Autodock 4.2.6 software for semi exible molecular docking [22]. Firstly, the name, molecular weight and 2D structure of the compound were determined in PubChem database(https://pubchem.ncbi.nlm.nih.gov/), and its 3D structure was constructed by ChemO ce software. Under the condition of the highest degree centrality (DC), the target protein was selected from different clusters, and the 3D structure of the protein receptor was obtained in RCSB PDB database (http://www.rcsb.org/) [23], the original ligand structure of the protein was extracted by using PyMOL 2.3 software. Then, AutoDock Tools (ADT) was used to add polar hydrogen and Gasteiger charge to the processed receptor and ligand. AutoGrid tool was used to set the parameters of docking frame. Lamarckian genetic algorithm (LGA) was used to nd the best docking conditions to make it exible docking and record the docking position of receptor and ligand. Finally, PyMOL software was used to analyze and observe the docking results[8].

Disease target acquisition
With the key words of "Diabetes Mellitus, Non-Insulin-Dependent", it integrated the disease-related genes obtained from multi-source databases (including GeneCards, DisGenet, UniProt). At last, it identi ed 3272 related genes, and UniProt database was used to standardize the selected targets.

Network construction and analysis
Taking the intersection of drug target and diabetes target, 128 overlapping targets were obtained, which were represented by Venn diagram (Fig. 2a). Through the interaction of T2DM, HL and DH targets, it was found that HL alone had 90 related targets for T2DM, suggesting that the main effective components of DHXD in the prevention and treatment of T2DM mostly came from HL. As shown in Fig. 2b, red represented T2DM, yellow represented DH, and blue represented HL. The composition and target information of DH and HL and the related target information of T2DM were made into herb-componenttarget network diagram, as shown in Fig. 2c, yellow nodes represented HL and DH, green represented active ingredient, purple represents central target, and lines represent their interaction. According to the network analysis, quercetin in HL was considered to be the most effective compound interacting with target genes. Input target genes into the STRING database for PPI network analysis, as shown in Fig. 3a.
In this network, there were 126 nodes and 1888 edges in total. When the setting conditions were degree ≥ 17, closeness ≥ 0.459 and betweenness ≥ 0.002, the PPI network with 69 nodes and 1110 edges would be obtained in Fig. 3b. Under the condition of degree ≥ 28, closeness ≥ 0.532 and betweenness ≥ 0.004, the PPI network with 42 nodes and 632 edges remained would be build in Fig. 3c after clustering analysis of these related targets ,in this network, the color and size change of all nodes is based on the pertinency from high to low, the most in uential target was VEGFA. The detailed information of 42 genes was shown in Table 2.
Venn diagram and herb-component-target network. a 128 intersection genes; b red, yellow and blue areas represented T2DM, DH and HL respectively; c a complete herb-component-target network
The bubble chart about whole 128 genes. In these bubble charts, the larger the bubble, the more genes were enriched in the pathway; the redder the color, the smaller the P value, the more signi cant the result is; Rich factor referred to the ratio of the number of genes belonging to the pathway in the target gene set to the number of genes belonging to the pathway in the background gene set, and the higher the value, the higher the enrichment degree. a
After labeling the common genes with KEGG-MAPPER, insulin resistance signaling pathway (hsa04931) and HIF-1 signaling pathway (hsa04066) were the most correlated pathways with DHXD, according to KEGG-MAPPER, their most related targets were shown in Fig. 5 and Fig. 6 as red nodes.
HIF-1 signaling pathway in uenced by DHXD. The red nodes represented the hub genes.
Insulin resistance signaling pathway in uenced by DHXD. The red nodes represented the hub genes.

Component-target molecule docking of DHXD
INSR and GLUT4, two target genes, showed strong association with other targets, pathways and active components, in this study, we combined them with 5 putative components to test their binding ability as shown in Table 3. According to the binding energy (Δ gbind) of molecular docking results, their binding activities were good (Δ gbind < − 5kj · mol − 1). The docking results of INSR with 5 active components were shown in Fig. 7, Fig. 7a-e represented their action mode, and the highly relevant target-pathway network was presented in Fig. 7f, Among them, HIF-1 signaling pathway (hsa04066) were the most correlated pathway (Fig. 7g) and quercetin (MOL000098) had the strongest binding force with INSR ( Fig. 7h). Figure 8a-e meant the action mode of GLUT4 with 5 active ingredients, INSR,VEGFA,PPARG and other targets had been proved to be related again (Fig. 8f), HIF-1 signaling pathway (hsa04066) was also the most signi cant related pathway (Fig. 8g), berberine (MOL001454) had a strong interaction with GLUT4 (Fig. 8h).  analyze the complex mechanism of TCM by knowing their effective components or formulas. This study found that DHXD could treat T2DM by regulating multiple targets of multiple components.
The mechanism of DHXD on T2DM was analyzed by network pharmacology and related databases, Studies have shown that berberine, berberrubine, epiberberine, EUPATIN, beta-sitosterol, rhein and other components play a major role in the treatment of T2DM and its complications. At the same time, we analyzed the MF, BP and CC of DHXD in T2DM, found that the BP of gene expression and apoptosis process were most affected, these processes occur mainly in the extracellular space, plasma membrane and so on. the MF shows that these components can bind to enzyme and protein well to maintain the activity of cells and receptors, we further examined 42 target genes and top 10 key pathways about DHXD on T2DM. Studies have shown that berberine can regulate glucose and lipid metabolism by regulating IL-6, TNF and other targets for anti-oxidation, reducing in ammatory response and increasing serum SOD activity [24], some researches have also con rmed this result of promoting insulin secretion and improving insulin resistance (IR) by regulating AMPK, MAPK and NF -κ B pathways [25]. In addition, other alkaloids such as epiberberine, magno orine, palmatine, jatrorrhizine and coptisine also intervene T2DM and its complications through related mechanisms [26]. Some network pharmacological studies have found that β -sitosterol may resist diabetic retinopathy by regulating key targets such as VEGFA in HIF-1 pathway [27], in addition to insulin secretagogue, β -sitosterol also has insulin-like activity, regulating PI3K / Akt and GLUT4 [28]. Rhein can play an anti-in ammatory role by regulating NF -κ B and reducing the expression of pro-in ammatory factors TNF -α and IL-6, which is closely related to T2DM; clinically, oral rhubarb rhizome extract can signi cantly reduce glycosylated hemoglobin, fasting blood glucose and body weight of patients with T2DM [29].
HIF-1 signaling pathway is closely related to hypoxia, HIF-1 is a heterodimer composed of HIF-1α and HIF-1β [30]. Both HIF-1α and HIF-1β play a regulatory role in metabolic diseases. HIF-1α may inhibit or increase obesity and IR, play an important role in islet β cells, and HIF1β has a separate function in metabolic abnormalities [31]. In this paper, KEGG enrichment analysis based on 128 overlapping genes and 42 core targets, as well as molecular docking results showed that HIF-1 signaling pathway played an important role, which might be the potential mechanism of DHXD in the treatment of T2DM. Some experiments had shown that the activation of HIF-1α could control the weight and blood glucose of the model mice, the insulin sensitivity, lipid metabolism level and albuminuria of the model mice were improved, which might suggest that HIF-1 signaling pathway has a potential impact on diabetes and its complications [32]. Some studies found that HIF-1α gene expression in T2DM patients with poor metabolic control was decreased [33]. However, some studies showed that the level of HIF-1α in T2DM patients was signi cantly higher than that in the control group, and insulin secretion increased when hypoxia was inhibited [34]. Therefore, the effect of HIF-1 signaling pathway on T2DM is not completely clear, which needs to be further con rmed. This paper also laid a research foundation for its elucidation.
INSR is a protein coding gene, which encodes two isoforms of INSR-A and INSR-B, the former is bene cial to prenatal development and tissue growth, expressed in undifferentiated cells; the latter is responsible for the systemic metabolism of insulin, growing in differentiated cells and post-mitotic cells, INSR plays an important role in maintaining glucose homeostasis and insulin sensitivity [35]. It was found that INSR increased in IR rats, and the increase of soluble INSR might be an early sign of metabolic syndrome [36].
This may suggest that the increase of INSR could be found in the early stage of T2DM. There was a research reported that the activation of INSR in liver and skeletal muscle of LEPR db/db mice can improve insulin sensitivity and reduce blood glucose [37]. An article also showed that up regulation of INSR can reduce blood glucose and increase insulin level in mice[38].
GLUT4, a member of glucose transporter family, is a major insulin regulated glucose transporter. The decrease of GLUT4 content will affect glucose uptake in adipose tissue and fat formation [39], activating the expression of GLUT4, promoting the fusion of GLUT4 plasma membrane and subsequent glucose uptake could effectively improve hyperglycemia and protect islet function [40]. It has been proved that GLUT4 can stay on the plasma membrane for a long time in order to maximize glucose absorption [41].
There was a study showed that berberine could reduce blood glucose, total cholesterol, low density lipoprotein by increasing the expression of GLUT4 in rat skeletal muscle to reduce IR [42]. In this paper, the results of molecular docking showed that berberine had the strongest binding capacity with GLUT4.

Conclusion
DHXD is effective in the treatment of T2DM, but the speci c active ingredients and mechanisms of intervention have not been clari ed. Therefore, using network pharmacology and molecular docking veri cation to establish a herbs-components-targets-pathways network and observe the potential targets with the highest correlation will help to clarify the direction of further research and provide the basis for explaining the exact mechanism. It was explored in this paper that berberine and other components might Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Competing interests
The authors declare that they have no competing interests.