A Network Pharmacology-Based Strategy for Predicting Active Ingredients and Potential Targets of Shuilu Erxian Dan in Treating Diabetic Kidney Disease


 Background and objective:

Recent years, some Chinese scholars have applied Shuilu Erxian Dan (SED) to the treatment of treating diabetic kidney disease (DKD) and achieved well curative effect. However, these studies are mostly limited to clinical observation. This study aimed to explore the molecular mechanisms of SED in treating DKD.
Methods

The active components of SED were retrieved in TCMSP database and BATMAN-TCM database, and the herbal targets were obtained by drugbank database and SwissTargetPrediction platform. The gene expression data of DKD patients were downloaded from GEO database and analyzed to obtain DKD-related targets. The ingredient-target network and the PPI network were constructed by Cytoscape software. The clusterProfiler package of R software is used for bioinformatic analysis. Molecular docking was further applied to verify the interaction between compounds and targets by Autodock Vina software.
Results

610 differential expressed genes of DKD patients were obtained, and 29 potential targets of SED against DKD were screened out (including PPTGS2, FABP3, HSD17B2, FABP1, HSD11B2, CYP27B1, JUN, UGT2B7, VCAM1, CA2, MAOA, MMP2, CXCR1, SLC22A6, EPHX2, SLC47A1, FOS, EGF, CCL2, COL3A1, GSTA1, GSTA2, HSPA1A, DAO, ALDH2, ALB, GPR18, FPR2, and LPL). All the active ingredients in SED can act on the DKD-related targets, among which quercetin, Ellagic acid, and kaempferol may be the key active compounds. SED may play a therapeutic role in DKD by regulating pathways including “Fluid shear stress and atherosclerosis”, “AGE − RAGE signaling pathway in diabetic complications” and “IL-17 signaling pathway”.
Conclusion

This study suggests that the mechanism of SED treating DKD is a complex network with multi-target and multi-pathway, which provides a reference for future experimental studies.


Construction of ingredient-target network and PPI network
The compound targets were intersected with the DKD-related targets to obtain the SED-DKD intersection targets. Then, the data of the compounds and its corresponding intersection targets was imported into the Cytoscape 3.7.2 software [40] to construct the ingredient-target network, and the network topology parameters were calculated by the "NetworkAnalyzer" tool [41]. The PPI network of SED-DKD intersection targets was constructed by BisoGenet 3.0.0 plugin [42] of Cytoscape software, which was set as "input nodes and its neighbors". PPI data sources include "Biological General Repository for Interaction Datasets" [43] (BioGRID), "IntAct Molecular Interaction Database" [44] (IntAct), "Database of Interacting Proteins" [45] (DIP), "Human Protein Reference Database" [46] (HPRD), "Biomolecular Interaction Network Database" [47] (BIND) and "The Molecular INTeraction Database" [48] (MINT). The parallel edges and sel oops in the PPI network were then removed and the topology parameters were calculated using the CytoNCA plugin [49]. Degree centrality (DC) and betweenness centrality (BC) were used as screening indicators to obtain the nal core PPI network.

Bioinformatic analysis
First, gene symbols were converted to Entrez IDs using the "org.Hs.eg.db" package (Version 3.8.2) of R software, and then the "clusterPro ler" package [50] (Version 3.12.0) was used for GO and KEGG pathway enrichment analysis of the SED-DKD intersection targets with FDR < 0.05. The pathway-target network were visualized by Cytoscape software.

Molecular docking
The SED-DKD intersection targets in the core PPI network and the targets with the highest degree in the pathway-target network were used as receptors for molecular docking with the corresponding compounds. The 3D structures of the compounds were constructed by ChemBio3D Ultra 14.0 software and optimized with MMFF94 force eld. The 3D structures of the target proteins were downloaded from RCSB Protein Data Bank [51] (http://www.rcsb.org). And AutoDockTools 1.5.6 software was used to pretreat the proteins, including removal water molecules, removal ligand molecules, and protonate 3D hydrogenation. The semi-exible docking calculation was carried out using Autodock Vina 1.1.2 [52], the exhaustiveness set to 20, and the rest use default parameters. The conformations with the lowest binding energy were analyzed and plotted by Molecular Operating Environment (MOE) 2015 [53].

DKD-related targets
610 DKD-related targets were identi ed by analyzing the gene expression arrays data downloaded from GEO database. The distribution of the differentially expressed genes is shown in Fig. 1, the red dots in the gure represent the up-regulated genes in DKD patients, the green dots represent the down-regulated genes, and the black dots represent the genes with insigni cant differences.

Ingredient-target network
There are twenty-nine intersection genes of SED targets and DKD-related targets, including PPTGS2, FABP3, HSD17B2, FABP1, HSD11B2, CYP27B1, JUN, UGT2B7, VCAM1, CA2, MAOA, MMP2, CXCR1, SLC22A6, EPHX2, SLC47A1, FOS, EGF, CCL2, COL3A1, GSTA1, GSTA2, HSPA1A, DAO, ALDH2, ALB, GPR18, FPR2, and LPL. The ingredient-target network is shown in Fig. 2, the edges between the nodes represent the compound-target interaction. All of the 10 active ingredients we retrieved could act on the corresponding DKD-related targets. Degree is one of the most important topological parameters, the node degree of a node is the number of edges linked to it [54]. The three compounds with the highest degree value in the ingredient-target network are quercetin, Ellagic acid and kaempferol, which respectively act on 12, 9, and 8 targets.

Molecular docking
As shown in Fig. 3C, the SED-DKD intersection targets in the core PPI network include JUN, FOS, VCAM1 and HSPA1A. The three target proteins with the highest degree value in the pathway-target network ( Fig. 6) are JUN, FOS and CCL2, and the degrees are 15, 14 and 7, respectively. FOS, JUN, VCAM1, HSPA1A and CCL2 were used as receptor proteins for molecular docking with the corresponding compounds (Fig. 2). The results of molecular docking are shown in Table 2, and the speci c docking modes are shown in Fig. 6.

Discussion
In ancient China, there was no disease name of DKD, which was mostly classi ed as "edema", "consumptive disease", "Guan Ge" and other categories [56]. According to theory of TCM, the key pathogenesis of DKD lies in the de ciency of kidney qi and the loss of storing function [56][57][58]. Fructus Rosae Laevigatae can astringe and preserve the kidney essence [59], Semen Euryales can bene t the kidney to preserve the essence [60]. The compatibility of the two herbs provides synergistic effect. The increase of urinary albumin excretion is one of the important diagnostic and evaluation indicators of DKD, strategies that can reduce albuminuria are associated with renal protection [61][62][63][64][65][66] and additional cardiovascular protection [67] in DKD patients. TCM believes that the kidney as the storehouse of essential qi stores essence and is the root of storage [68]. Albumin is one of the essence of human body. Due to the de ciency of kidney qi and the loss of storing function, the albumin of DKD patients is lost in urination. By tonifying kidney and astringing, SED can gradually restore the renal physiological function of storing essence and reduce the albuminuria. Jinsong Jin et al. [69] proved that SED extract could effectively reduce the albuminuria and improve the nutritional status in adriamycin-inducded nephropathy rats. In addition, it was found that Fructus Rosae Laevigatae played a protective role in the kidney of streptozotocin-induced DKD rats by inhibiting oxidative stress [70]. Semen Euryales may reduce albuminuria and delay the progress of DKD by up-regulating the expression of renal SOCS-3 and inhibiting the overexpression of renal IGF-1 in rats [71]. However, due to the "multi-ingredient, multi-target" characteristics of Chinese herbal compound, these studies cannot reveal the mechanisms of SED acting on DKD comprehensively and systematically.
By constructing the ingredient-target network (Fig. 2), we found that all the active compounds in SED can act on the DKD-related targets, which indicated that SED has a strong pertinence in treating DKD. All of the 10 active compounds can affect multiple targets, among which quercetin, Ellagic acid and kaempferol can act on the most targets, so these three compounds may be the crucial active ingredients of SED in treating DKD. In addition, many of these compounds have common targets, suggesting that different compounds may provide synergistic effects. Quercetin belongs to the avonol group of polyphenolic compounds, which functions as antibacterial, antiviral, anti-in ammatory, antioxidant, anticancer, anti-diabetic, immunomodulatory, etc [72]. The existing studies [73][74][75]have provided convincing evidence on the renoprotective effects of quercetin in both animal and cell models of DKD. In addition to its antiviral, anti-in ammatory, antioxidant, anti-cancer and anti-diabetic properties, Ellagic acid also exerts hepatoprotective effect [76]. Ellagic acid has been shown to ameliorate renal function and renal pathology in streptozotocin-induced DKD rats by inhibiting the NF-кB pathway and the accumulation of AGEs (advanced glycation end products) in kidney [77][78][79]. Kaempferol has similar pharmacological effects to quercetin and is used in the treatment of diabetes, metabolic syndrome, liver injury, cancer, etc [80]. Sharma D et al. [81] found that Kaempferol can attenuate DKD by inhibiting RhoA/Rho-kinase mediated in ammatory signaling in vitro.
In this study, we identi ed 29 potential targets of SED acting on DKD, and constructed the PPI networks of these 29 target proteins and their related proteins. The result shows that the target proteins can interact with each other, and there are as many as 1,399 proteins related to them, which form a complex interaction network. By calculating topological parameters of the network and screening the proteins, we obtained a core PPI network containing 209 proteins (Fig. 4C). Four SED-DKD intersection targets are included in the core PPI network: JUN, FOS, VCAM1 and HSPA1A. Activator protein-1 (AP-1) belongs to the basic-leucine-zipper family of transcription factors, and the most common form of AP-1 is a dimer of JUN protein and FOS protein [82]. High glucose/Ang -mediated activation of AP-1 can lead to the proliferation of mesangial cells and the excessive accumulation of extracellular matrixs, which is a key pathologic feature of DKD [83,84]. The full name of VCAM1 is vascular cell adhesion molecule 1. During the activation or damage of vascular endothelial cells, VCAM1 can be shed from the cell surface into the circulation, and soluble VCAM1 in the blood of DKD patients was found to be signi cantly increased, which was positively correlated with UACR (urine albumin: creatinine ratio) [85,86]. HSPA1A is an in ammation related protein, which is incriminated in the renal in ammation of DKD as the endogenous TLR ligand [87,88].
The GO analysis of SED-DKD intersection targets enriched some interesting GO terms, such as "fatty acid metabolic process", "regulation of lipid metabolic process", "peroxisome", "SMAD binding" and "glutathione transferase activity". In nonadipose tissues, excess cytosolic free fatty acids (FFAs) can lead to cell dysfunction and death, a process known as "lipotoxicity". Disturbed FFA metabolism and renal lipid accumulation are thought to be associated with DKD glomerulosclerosis and tubulointerstitial damage in DKD [89][90][91]. Peroxisome is a kind of microbody whose main function is to catalyze the βoxidation of fatty acids and the hydrolysis of hydrogen peroxide [92]. The inactivation of peroxisomal catalase in DKD animal models can cause alterations of mitochondrial membrane potential, which stimulate the generation of mitochondrial reactive oxygen species (ROS) [93]. Oxidative stress caused by excessive ROS production is considered to be an important factor in the occurrence and development of diabetic complications including diabetic nephropathy [94][95][96][97]. Members of the SMAD protein family act as signal integrators and interact with several DKD-related signaling pathways, among which Smad3 is pathogenic, Smad2 and Smad7 are protective [98,99]. Glutathione-S-transferases represent a superfamily of enzymes involved in cell protection and detoxi cation, play an important role in protecting the body from oxidative stress products [100]. Glutathione-S-transferase activity is considered as one of the markers of severity in DKD patients [101].
Multiple signaling pathways were signi cantly enriched by KEGG analysis, among which "Fluid shear stress and atherosclerosis" is an important atherosclerosis-related pathway. DKD is closely related to cardiovascular disease. Microalbuminuria re ects generalized endothelial damage and is regarded as an early event in atherosclerosis [102]. Vascular endothelial dysfunction is also considered to be involved in the pathogenesis of DKD [103]. In addition, several traditional risk factor for atherosclerosis has been identi ed in DKD patients including high blood pressure, hyperlipaemia and procoagulatory state associated with endothelial dysfunction [104]. A lot of evidence supports the signi cance of "AGE − RAGE signaling pathway" in the pathogenesis of DKD, and its blockade seems to be an attractive therapeutic target [105]. "IL-17 signaling pathway" is believed to play a pro-in ammatory role in podocyte injury, mesangial expansion and renal brosis in DKD patients [106].
In our study, molecular docking was further applied to verify the interaction between compounds and targets. The combination with the lower binding energy scores is more stable, and the binding energy ≤ − 5.0 kcal·mol − 1 was de ned as the standard of well binding between ligands and receptors in some studies [107][108][109]. As shown in Table 2, the binding energies of all docking are less than − 5 kcal·mol − 1 . Take the complex with the lowest binding energy as an example, "Ellagic acis-HSPA1A" (Fig. 6I)

Conclusion
This study explored the mechanism of SED in the treatment of DKD by means of network pharmacology. We identi ed 29 potential targets of SED acting on DKD, among which FOS, JUN, VCAM1, HSPA1A and CCL2 were the key targets. SED may play a therapeutic role in DKD by regulating pathways including "Fluid shear stress and atherosclerosis", "AGE − RAGE signaling pathway in diabetic complications" and "IL-17 signaling pathway". Quercetin, Ellagic acid, and kaempferol in SED may be the key active ingredients. The mechanism of SED treating DKD is a complex network with multi-target and multipathway. This study provides a scienti c theoretical basis for the prevention and treatment of SED acting on DKD, and also provides a reference for researchers in related elds to further carry out experimental work.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data are available in the manuscript and they are showed in gures, tables and supplement le.

Competing interests
The authors declare that they have no con ict of interest.

Funding
This study was supported by the National Natural Science Foundation of China (no. 81774279).
Authors' contributions TW and RY: study design; acquisition of data; analysis and interpretation of data; drafting of the manuscript; critical revision of the manuscript for important intellectual content; statistical analysis. MH: revision of the manuscript and study supervision. The author(s) read and approved the nal manuscript. Ingredient-target network. Note: The blue-green rectangles represent intersection targets, the orange ovals represent compounds from Fructus Rosae Laevigatae, the yellow ovals represent compounds from Semen Euryales.  Pathway-target network.