Introduction

Diabetes mellitus (DM) is a highly prevalent chronic metabolic disease that is expected to affect 783.2 million people worldwide by 20451. Notably, nephropathy develops in 20% to 40% of patients with both type 1 and type 2 diabetes mellitus (T1 and T2DM), and approximately 5% of patients with T2DM already have diabetic kidney injury at the time of diagnosis2. DM-induced kidney damage may progress to end-stage renal disease (ESRD), requiring long-term dialysis or kidney transplantation and resulting in significant socioeconomic costs. Current therapies focus on lowering and managing blood glucose levels to minimize the development of renal, neurological, and vascular complications resulting from glucose accumulation3. However, serious side effects, including hypoglycemia4, lactic acidosis5, and gastrointestinal damage6 are inevitable with prolonged administration of hypoglycemic agents. Therefore, the discovery of effective drugs with fewer side effects for patients with diabetic kidney injury is of great importance.

Despite the lack of precise elucidation regarding the mechanisms of kidney injury in DM, existing studies suggest that hyperglycemia-induced fibrosis is a critical contributor to this injury7,8. In the diabetic state, glucose metabolism is severely impaired in the liver, muscles, and brain, while it is significantly increased in the kidneys, nerves, and eyes9. At this point, approximately 50% of the glucose is metabolized in the kidneys, far exceeding its normal glycemic load10. Excess glucose stimulates renal cells to overexpress transforming growth factor beta (TGF-β), a key mediator of renal fibrosis8. Renal fibrosis is characterized by excessive accumulation of extracellular matrix (ECM) in the renal parenchyma11. Numerous studies have shown that overexpressed TGF-β induces uncontrolled buildup of ECM and collagen fibers in the kidney, which in turn accelerates the process of renal fibrosis12,13. In addition, hyperglycemia promotes the accumulation of pathogenic mediators such as angiotensin II and advanced glycosylation end products (AGEs), which enhance TGF-β-mediated fibrotic effects14. Unfortunately, the addition or combination of hypoglycemic agents fails to diminish the risk of kidney damage in diabetic patients15. Therefore, it is imperative to explore new therapeutic strategies to prevent or mitigate DM-induced kidney injury.

Accumulating evidence suggests that Dioscoreaceae plants and their extracts have great potential in the treatment of DM and diabetic kidney injury, especially Dioscorea nipponica Makino16. D. nipponica Makino, a traditional medicinal plant for DM therapy in Korea, is the main active pharmaceutical ingredient (API) of Di’ao Xinxuekang capsules (DXXK)17. DXXK has been listed in the Chinese Pharmacopoeia (Ch.P.) 2020 as a Chinese herbal medicine with minimal adverse effects and a high safety profile18. It is also the first phytopharmaceutical outside the European Union (EU) to be approved by the Medicines Evaluation Board (MEB). It was demonstrated that DXXK effectively reduced fasting blood glucose (FBG) levels and lipid parameters, improved glucose and insulin tolerance, and ameliorated renal, pancreatic, and hepatic lesions in DM rats19,20. DXXK consists mainly of steroidal saponins, including dioscin, diosgenin (DG), and protodioscin, which have been extensively used in DM-related drug studies21. Specifically, dioscin significantly improved renal function in DM mice by regulating FBG and inhibiting glomerular mesangial matrix expansion; DG effectively attenuated peripheral neuropathy in DM animals via lowering FBG levels22,23. Likewise, protodioscin demonstrated stable glycemic control and ameliorated DM-induced renal pathology in DM rats24. Furthermore, clinical studies have shown that DXXK may aid in the therapy of DM and kidney disease by regulating FBG, sensitizing insulin and improving blood rheology (plasma viscosity, erythrocyte electrophoretic index, and hematocrit)25,26. In patients with coronary artery disease combined with DM, the co-administration of DXXK and conventional therapy was clinically more effective, with significant reductions in FBG and low-density lipoprotein cholesterol (LDL-C) levels, and a markedly lower incidence of adverse effects such as nausea, vomiting, and dizziness27. Thus, DXXK has the potential to serve as a renoprotective agent in DM, but its specific mechanism remains to be explored.

Network pharmacology is a new approach to recognize the interaction between drugs and organism by using the holistic view of network balance, which is characterized by "multi-gene, multi-target", and has important application value in the research of pathogenesis and therapeutic targets of complex diseases28. In this study, network pharmacology was used for the first time to screen the potential therapeutic ingredients and targets of DXXK, and to experimentally validate its key signaling pathways for the treatment of DM. The specific workflow is shown in Fig. 1. Our study elucidates the mechanisms underlying the nephroprotective properties of DXXK and provides empirical support for DXXK in the treatment of diabetic kidney injury.

Fig. 1
figure 1

A flowchart to explore the therapeutic mechanisms of DXXK.

Full size image

Materials and methods

Collection and verification of DXXK ingredients

Since DXXK capsules are herbal preparations, we searched for the corresponding ingredients using its most prominent API (D. nipponica Makino) as the keyword. The ingredients of DXXK were mainly obtained from the Encyclopedia of Traditional Chinese Medicine (ETCM) and Bioinformatics Analysis Tool for Molecular mechANism of Traditional Chinese Medicine (Batman-TCM) databases, both of which contain comprehensive and standardized information on the specific ingredients of Chinese herbs and formulas. To ensure the completeness and accuracy of the ingredients in DXXK, some ingredients were hand-collected from other literature21. Their chemical formulas were determined by Pubchem and stored in 2D.sdf format. The pharmacological properties of each ingredient were evaluated separately in accordance with the Swiss ADME database. We also reviewed the literature on the antidiabetic activity of these ingredients, followed by literature verification of the ingredients excluded according to the Swiss ADME screening criteria. If the pharmacological effects of these ingredients were clearly relevant to the subject of the study, they were included directly in the study.

Prediction of DXXK-related targets

Each ingredient of DXXK was individually loaded into the PharmMapper database (a comprehensive pharmacophore matching platform) to find potential targets. Based on the number of targets obtained, genes with a normalized fit (NF) score ≥ 0.3 were screened as DXXK-related targets.

Prediction of DM-related targets

To obtain a comprehensive and reliable list of DM-related targets, the keyword “Diabetes mellitus” was selected to search the GeneCards, OMIM, DisGeNET, and DrugBank databases. All DM-related targets in the OMIM and DrugBank databases were collected. In addition, the median relevance score was used for multiple screening in the GeneCards database, while Score_gda > 0.1 was used as a filtering criterion in the DisGeNET database. Diabetic targets were obtained by removing duplicate values and merging the targets extracted from the above databases.

Construction of protein–protein interaction (PPI) networks

Predicted targets for DM and DXXK were converted to gene symbols using the UniProt database. These gene symbols were then mapped onto a Venn diagram to identify core therapeutic targets. To further elucidate the interactions between overlapping genes, their PPI networks were generated using the STRING database. The species was set to Homo sapiens with a confidence score of 0.4, and the other settings were default. Moreover, the interaction network of DXXK-related components and corresponding targets was constructed using Cytoscape (version 3.7.1).

Enrichment analysis

Kyoto Encyclopedia of Genes and Genomes (KEGG)29,30,31 and Gene Ontology (GO) enrichment analyses (p ≤ 0.01) were performed on the overlapping targets using the Metascape database. The GO enrichment analyses included biological process (BP), molecular function (MF), and cellular component (CC), respectively. After sorting the enrichment analysis results by p-value, the top 20 pathways were selected for visualization.

Molecular docking

The PDB files of the proteins were downloaded from the RCSB database (PDB IDs: 1PY5-TGF-β1, 5E8Y-TGF-β2, 1KHX-p-Smad2, 1MK2-p-Smad3, 1DEV-Smad2, and 1MJS-Smad3). To eliminate the effect of water of crystallization on ligand-target protein interactions and optimize the docking energy, ligands (ingredients of the DXXK) and protein receptors were saved in PDBQT format in the Autodock Vina software (version 1.1.2) after removing water molecules and adding hydrogen bonds, respectively. The appropriate coordinate positions and grid boxes were set to cover the active sites of the proteins by adjusting the number of points in the x, y, and z dimensions and the spacing (angstroms), and then Vina was run. After docking, the conformations with the highest number of hydrogen bonds were analyzed using PyMOL (version 2.5). Table 1 lists the web links to all databases and software.

Table 1 Database and software (URL).
Full size table

Medications and reagents

DXXK was manufactured by Di’ao Pharmaceutical Group Co., Ltd (2,106,036, Chengdu, China). Metformin was produced by Xinyi Tian ping Pharmaceutical Co., Ltd (67,210,972, Shanghai, China). Control samples of dioscin, protodioscin, and pseudoprotodioscin were purchased from Yuanye Biotechnology Co., Ltd (B21176, B21621, B21303, Shanghai, China). High-fat diet (HFD) was purchased from Xietong Pharmaceutical Bio-engineering Co., Ltd (XTHF60, Nanjing, China). Streptozotocin (STZ) was purchased from Aladdin (G2115339, Shanghai, China). Citric acid and sodium citrate were purchased from Macklin (C12022793, C12931024, Shanghai, China). Insulin was purchased from Wan bang Pharmaceutical Technology Co., Ltd (H10890001, Hefei, China). Hematoxylin–eosin (H&E) staining kit, Masson’s trichrome (M-T) staining kit, reverse transcription kit, and SYBR Green qPCR kit were purchased from Biosharp (BL700A, BL1059A, BL699A, BL698A, Hefei, China).

Constituent analysis

DXXK and control samples were dissolved separately in methanol, vortexed and filtered. A volume of 10 μL was injected into a high-performance liquid chromatography (HPLC) system (Agilent1260, USA) for analysis (203 nm). DXXK was separated on a ZORBAX SB-C18 column (4.6 mm × 250 mm, 5 μm) using acetonitrile (A) and water (B) as mobile phases. The detection wavelength was 203 nm and the elution conditions were 0–30 min: 15–35% A, 85–65% B; 30–55 min: 35–68% A, 65–32% B.

Animal

Eight-week-old male C57BL/6 mice were purchased from Ziyuan Experimental Animal Technology Co., Ltd (license number: SCXK [Zhe] 2019–0004, Hangzhou, China). All mice were housed in an animal room (24 ± 2 °C, 12-h light/dark cycle) with access to water and food. Animal experiments were conducted under the approval of the animal ethics committee of Anhui University of Chinese Medicine (No. AHUCM-mouse-2022092). Throughout the experimental process, we strictly adhere to international and national animal welfare guidelines as well as ARRIVE guidelines.

DM modeling and drug therapy

After 1 week of acclimatization, mice were fasted for 12 h and then injected intraperitoneally with STZ (50 mg/kg/d) for 4 days. In the control group, the corresponding volume of saline was injected intraperitoneally. Metformin was used as a positive control in this study. Following 3 days, mice with FBG ≥ 16.8 mmol/L were randomly divided into DM, low-dose, high-dose of DXXK, and metformin groups (n = 6). The low and high doses of DXXK were 75 mg/kg/d and 150 mg/kg/d, respectively, administered once daily by oral gavage. The intragastric administration of metformin was at a dose of 180 mg/kg/d. The control and model groups received equal volumes of saline by gavage daily. All doses were calculated using the Meeh-Rubner formula, with a conversion factor of 9.1X mg/kg (X is the human dose). Treatment with DXXK and metformin for 10 weeks and the specific animal procedures were shown in Fig. 2.

Fig. 2
figure 2

Procedure for the treatment of DM mice with DXXK.

Full size image

DM indicators assessment

During the administration of DXXK, the mice were monitored biweekly for body weight, water and food consumption, and FBG levels. In addition, post-treatment oral glucose and insulin tolerance tests (OGTT and ITT) were performed. All mice were fasted for 12 h prior to testing. Blood samples were collected from the tail at 0, 15, 30, 45, 60, 90, and 120 min after oral glucose administration (2 g/kg) and then measured (Roche, Switzerland). Similarly, mice were injected intraperitoneally with insulin (1.5 UI/kg) after an 8-h fast and FBG data were recorded as previously described.

Serum biochemical assay

At the completion of medication, mice were anesthetized with 20% urethane and then the eyes were removed for blood collection. Serum was collected by centrifugation at 2500 g for 25 min and then analyzed using an automatic analyzer (Hitachi 7020, Japan).

Histopathologic assay

Following euthanasia, the right kidney and liver were immersed in 4% paraformaldehyde, then dehydrated in graded alcohol, followed by hyalinization with xylene, and finally embedded in paraffin. Paraffin Sects. (4 μM) were stained with H&E and M-T stain, respectively. Moreover, all left kidneys were cryopreserved for subsequent experiments.

Quantitative real-time PCR

Total RNA from kidneys was extracted with trizol and quantified by ultraviolet–visible (UV–Vis) spectroscopy (Molecular Devices, USA). After reverse transcription, the cDNA was subjected to qRT-PCR using SYBR Green chemistry in a real-time fluorescence quantitative PCR instrument (Applied Biosystems, USA). Primers were listed in Table 2.

Table 2 Primer sequences for qRT-PCR.
Full size table

Western blot analysis

Proteins were extracted from the kidney with pre-cooled RIPA buffer and quantified with bicinchoninic acid (P0010S, Beyotime, China). Chemically denatured proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membranes. Membranes were blocked with TBST which containing 5% skim milk for 2 h at room temperature and then incubated for 12 h at 4 °C with diluted primary antibodies: GAPDH (ab8245, Abcam, UK), TGF-β1 (AF1027, Affinity, China), phospho-Smad2/3 (AF3367, Affinity, China), and Smad2/3 (8685, Cell Signaling Technology, USA). Subsequently, incubation with the secondary antibody was performed for 1 h at room temperature. Between blocking and antibody incubation, the membranes were washed with TBST solution (3 times for 10 min each). After the enhanced chemiluminescence reaction on the membrane, the protein expression levels were detected in an imaging system (GelView 6000Plus, China) and the results were processed using Image J (version 1.8).

Statistical analysis

All data were expressed as mean ± standard deviation (SD). Differences between groups were compared by one-way analysis of variance (ANOVA) with Dunnett’s post hoc test. Statistical significance was considered when p < 0.05. Statistical analyses were performed using GraphPad Prism 6.0 software (California, USA).

Results

Screening of active ingredients and targets of DXXK

We retrieved 13 ingredients of DXXK from the ETCM, Batman-TCM and the literature (Supplementary Table 1). Considering the different origins of the ingredients of DXXK in this study, each ingredient was evaluated for its druggability in the Swiss ADME database. However, by analyzing their pharmacokinetics (GI absorption) and drug-likeness (Lipinski, Ghose, Veber, Egan, Muegge), the Swiss ADME screening criteria (GI absorption as “High” and both parameters of drug-likeness as “Yes”) excluded almost all ingredients with antidiabetic activity. Therefore, we finalized the list of 7 ingredients with antidiabetic potential by using whether the ingredient has antidiabetic activity (literature validation) as a special criterion (Table 3). DXXK-related targets were obtained from the PharmMapper database and sorted by NF score (Supplementary Table 2). The corresponding targets of each ingredient were screened separately using NF score ≥ 0.3 as the threshold, and the top 20 targets were shown in Table 3. Then, an interaction network between DXXK1-7 and targets was constructed as shown in Fig. 3A, in which dioscin (DXXK1), DG (DXXK2), and protodioscin were enriched for more targets.

Table 3 Active ingredients of DXXK (after Swiss ADME screening).
Full size table
Fig. 3
figure 3

Network pharmacology analysis revealed therapeutic targets of DXXK. (A) Component-target network of DXXK. Orange circles indicate components of DXXK and pink ovals indicate targets of DXXK. (B) Venn diagram of 152 overlapping targets between DXXK and DM. (C) PPI network of overlapping genes based on STRING. (D) Interaction of fibrosis-related genes.

Full size image

Construction of the PPI network

To obtain therapeutic targets for DXXK, 2,216 DM-related genes were collected from 4 disease datasets (Supplementary Table 3). These genes were intersected with 356 DXXK-related genes using the Venn tool, yielding 152 potential targets (Fig. 3B). The overlapping targets were entered into the STRING database to generate the PPI network, which contained 150 genes and 1683 interactions (Fig. 3C and Supplementary Table 4). The area and color of the nodes were positively correlated with the degree, and the thickness and brightness of the lines in the network were also positively correlated with the combined scores. Interestingly, several proteins associated with fibrosis were identified, including TGFBR-1, TGFBR-2, TGFB-2, BMP-2, and BMP-7 (Fig. 3D). Although their degree and combined scores were not outstanding, they had high neighborhood connectivity, all greater than 45, suggesting that these proteins act as connectivity links for interactions with other proteins (Supplementary Table 4).

KEGG/GO enrichment analyses

KEGG and GO (MF, BP, CC) analyses were performed on 150 core targets using the Metascape data platform. Notably, the insulin signaling pathway, insulin resistance, TGF-beta signaling pathway, and glycolysis/gluconeogenesis were significantly enriched (Fig. 4A). Furthermore, GO analysis revealed that the regulation of lipid metabolic processes, insulin receptor binding, and extracellular matrix were closely related to the therapeutic mechanism of DXXK (Fig. 4B–D). Notably, TGFBR-1, TGFBR-2 and TGFB-2 are key genes in the TGF-beta and extracellular matrix signaling pathway. Therefore, multiple active ingredients in DXXK may exert therapeutic effects on diabetic kidney injury through these signaling pathways.

Fig. 4
figure 4

Functional analysis (p < 0.05) of potential therapeutic targets of DXXK. (A) KEGG analysis of therapeutic targets. (B–D) GOMF, GOBP, and GOCC analysis of therapeutic targets.

Full size image

Molecular docking

To further verify the effect of DXXK on the TGF-beta signaling pathway, proteins in this pathway were simulated to dock with DXXK1-7, respectively. Binding activity is usually considered to be present when the affinity is < -5 kcal/mol. The simulation results revealed that DXXK1-7 tightly bound to TGF-β1, TGF-β2, p-Smad2, p-Smad3, Sma2, and Smad3, especially dioscin, with affinities of -8.4, -9.1, -9.4, -8.2, -8.1, and-8.5 kcal/mol (Fig. 5A). Specifically, Fig. 5B and C showed the active pockets of dioscin and DG, respectively, as well as the hydrogen bonds attached to the proteins.

Fig. 5
figure 5

Molecular docking of the active ingredient in DXXK with pro-fibrotic proteins. (A) Heat map of the binding energy of DXXK1-7 to proteins. (B) Docking of dioscin with proteins. b1) dioscin-TGFβ1. b2) dioscin-TGFβ2. b3) dioscin-Smad2. b4) dioscin-Smad3. b5) dioscin-p-Smad2. b6) dioscin-p-Smad3. (C) Docking of DG with proteins. c1) DG-TGFβ1. c2) DG-TGFβ2. c3) DG-Smad2. c4) DG-Smad3. c5) DG-p-Smad2. c6) DG-p-Smad3. The blue and purple colors represent the position of the bound amino acids.

Full size image

Verification of DXXK ingredients

To validate the accuracy of the predicted components in network pharmacology, the main ingredients of DXXK were analyzed by HPLC, and the chromatogram is shown in Fig. 6. Consistent with the predicted results, dioscin, prodioscin, and pseudoprodioscin were the major ingredients, with pseudoprodioscin being the most abundant at a concentration of 36.8 mg/capsule in accordance with Ch.P. 2020. This result will contribute to the quality control of DXXK and support further pharmacodynamic studies.

Fig. 6
figure 6

Chromatographic analysis of DXXK.

Full size image

DXXK ameliorated the diabetic symptoms induced by STZ

Over a period of 10 weeks, mice in the control group had shiny coats and were responsive. In contrast, all DM mice had matted fur and were unresponsive, accompanied by an increase in urination. The average body weight of the control group exhibited a steady increase, while the DM group experienced a decrease from 20 to 17 g. However, the weight loss of the DM mice was improved after administration of DXXK, which significantly suppressed this decline at week 10 (Fig. 7A). Excessive nutrient requirement is a typical symptom of DM, and our study revealed that the food and water intake of the DM group far exceeded that of the control group. On the contrary, after drug treatment, high doses of DXXK and metformin were effective in alleviating the excessive intake, with metformin having a more pronounced effect (Fig. 7B and C). Moreover, the average level of FBG continued to rise to 28 mmol/L in the DM group (Fig. 7D), whereas the FBG levels in DM mice were markedly reduced at the week 10 and week 6 under the intervention of DXXK and metformin, respectively.

Fig. 7
figure 7

Effects of DXXK on physiological parameters and glucose metabolism in DM mice. (A) Effect of DXXK on weight. (B, C) Changes in food and water intake of mice. (D) FBG levels of mice during therapy. (E–F) OGTT and ITT. Data (n = 6) are represented as the mean ± standard deviation (SD). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 vs. control (a) or DM (b) group.

Full size image

OGTT and ITT were performed to further evaluate the effect of DXXK on FBG. The results of OGTT showed that the glucose metabolic function was severely impaired in the DM group compared with the control group, which was not improved by DXXK (Fig. 7E). The ITT results are shown in Fig. 7F. Notably, DXXK significantly ameliorated this adverse condition and enhanced insulin sensitivity in DM mice. These results suggest that DXXK has the potential to ameliorate the symptoms of DM, lower FBG levels and alleviate insulin resistance.

DXXK improved lipid metabolism and liver function in DM mice

Hyperlipidemia tends to occur in DM due to inadequate insulin secretion. To determine whether DXXK regulates lipid metabolism in DM, we examined some lipid indices. As shown in Fig. 8 A-C, serum levels of total cholesterol (TC), triglycerides (TG), and LDL-C were significantly elevated in the DM group compared with the control group. Conversely, DXXK intervention reduced these abnormalities in the treatment group. Furthermore, both DXXK and metformin increased serum levels of high-density lipoprotein cholesterol (HDL-C), but there was no significant difference compared with the DM group (Fig. 8D). Since the liver is the primary regulator of lipids and blood glucose, we also measured serum levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT). Aminotransferase levels were markedly reduced in the high-dose DXXK group compared to the DM group (Fig. 8E and F). These findings suggest that DXXK improves lipid metabolism and liver function in DM mice.

Fig. 8
figure 8

Regulation of lipid metabolism and liver function by DXXK in DM mice. (A–D) Serum concentrations of TC, TG, LDL-C, and HDL-C (n = 6). (E–F) Serum levels of AST and ALT in mice (n = 6). All data are expressed as the mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 vs. control (a) or DM (b) group.

Full size image

DXXK attenuated kidney and liver lesions in DM mice

As mentioned above, diabetics with uncontrolled blood glucose are highly susceptible to kidney injury. To assess whether DXXK has renoprotective properties, we performed pathological examination on of the kidneys. H&E staining showed that the kidneys in the control group were structurally intact, whereas mesangial hyperplasia and glomerular basement membrane thickening were observed in the DM group. By contrast, the glomerular and tubular structures tended to be normal in both the DXXK and metformin-treated groups (Fig. 9A). Furthermore, excessive collagen fibers accumulated in the kidney structures of the DM group, as observed by M-T staining. However, DXXK and metformin significantly reduced this accumulation in the glomerulus and tubules. Notably, the therapeutic effect of DXXK was dose-dependent, and the anti-renal fibrosis activity of high-dose DXXK was comparable to that of metformin (Fig. 9B). To further investigate the improvement in liver function with DXXK, we also performed H&E staining of the liver. The livers of the DM group showed obvious ballooning degeneration and steatosis compared to the control group. Notably, both DXXK and metformin effectively reduced these pathological changes (Fig. 9C). These results suggest that DXXK attenuated kidney and liver injury and prevented the exacerbation of renal fibrosis in DM mice.

Fig. 9
figure 9

Effects of DXXK on kidney and liver pathology in DM mice. (A) H&E staining of the kidneys (n = 6). (B) M-T staining of the kidneys (n = 6). (C) H&E staining of the livers (n = 3). Representative images were selected for presentation (bars, 100 µm, × 400 (kidneys) and bars, 5 µm, × 100(livers)).

Full size image

DXXK attenuated fibrosis in DM mice through the TGF-β1/Smad signaling pathway

Fibrosis is an important pathogenic mechanism of diabetic kidney injury, and the fibrotic response may be mediated in part by the TGF-β signaling pathway. Based on the results of network pharmacology and molecular docking, we found that the therapeutic effect of DXXK on diabetic kidney injury may be related to the TGF-β1/Smad signaling pathway. Therefore, the TGF-β signaling pathway was investigated. The protein expression of TGF-β1, p-Smad2/3, and Smad2/3 were significantly higher in the DM group than in the control group. Conversely, both DXXK and metformin showed significant inhibition of these protein expressions (Fig. 10A and B). Furthermore, the gene expression levels followed a similar trend to that observed in the protein blots (Fig. 10C–E). Notably, the modulatory effect of DXXK was found to be comparable to that of metformin. Taken together, we conclude that DXXK exerts renoprotective effects by inhibiting fibrosis through partial inactivation of the TGF-β1/Smad2/3 signaling pathway.

Fig. 10
figure 10

Effect of DXXK on the TGF-β1/Smad pathway. (A) Protein expression of TGF-β1, total Smad2/3, and p-Smad2/3. (B) Quantification was normalized to GAPDH (n = 5). (C–E) Relative gene expression of TGF-β1, Smad2, and Smad3 (n = 3). All data are expressed as the mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 vs. control (a) or DM (b) group. The blot results have been cropped and the original blot is shown in Supplementary Fig. 1.

Full size image

Discussion

Diabetic kidney injury is a glomerular and tubulointerstitial lesion caused by prolonged uncontrolled hyperglycemia. It is mostly asymptomatic in the early stages and its incidence gradually increases as the disease progresses, leading to proteinuria, edema, uremia, and ESRD, which severely affects the lives of patients32. Increasing evidence suggests that traditional Chinese medicine (TCM), as a classical approach to DM treatment, is effective in minimizing treatment-related side effects33,34. DXXK is a modern herbal capsule that can be used safely and effectively as an adjunctive treatment for DM and nephropathy. Clinical studies have shown that compared with oral hypoglycemic agents, DXXK can effectively reduce urinary β2-microglobulin levels, attenuate kidney damage, correct dyslipidemia, and improve insulin resistance in patients with T2DM35. In addition, DXXK improves chronic renal failure in nephropathic patients by lowering serum urinary nitrogen, urinary creatinine and platelet aggregation rate, and can be administered for long periods of time with little adverse effects26. However, its pharmacological mechanism is still unclear. Therefore, this study fills this knowledge gap through a combination of network pharmacology and experimental validation.

Network pharmacology identified 7 components and 356 targets of DXXK to construct an ingredient-target network. In this network, most of the chemical components associated with DM are steroidal saponins, with dioscin, DG, and protodioscin strongly interacting with other targets. In the treatment of DM, studies have demonstrated the antidiabetic activity of the above components. Specifically, dioscin attenuated hyperglycemia-induced apoptosis by stimulating the proliferation of pancreatic beta-cells36. DG and protodioscin exhibited ameliorative effects against diabetic complications, including diabetic nephropathy, diabetic hepatopathy, and diabetic neuropathy24,37. To verify the accuracy of the predicted ingredients, we performed HPLC analysis of DXXK. Consistent with the network pharmacology results, dioscin, protodioscin, and pseudoprotodioscin serve as the basic building blocks of DXXK. Notably, DG was not detectable via chromatography because of its low presence in DXXK. However, DG can be formed in the gastrointestinal tract from dioscin, which is the glycosidic form of DG38.

Through network pharmacology, we obtained 2,216 DM-related genes and finalized 150 overlapping targets. The PPI network of core targets revealed a number of fibrosis-associated genes, such as TGFBR1, TGFBR2, and BMP7, which are key hubs in the TGF-β signaling pathway. Studies have shown that the expression levels of pro-fibrotic genes in the TGF-β pathway are positively correlated with the risk of diabetic kidney injury11. In the diabetic state, renal cells stimulated by hyperglycemia secrete excess TGF-β, of which TGF-β1 is one of the predominant isoforms39. Overexpressed TGF-β1 mediates signaling through the heterotetramer composed of TGFBR1 and TGFBR2, which promotes the accumulation of ECM in the kidney, thereby exacerbating renal fibrosis40. Moreover, TGF-β1 accelerates the fibrotic process by inducing the activation and phosphorylation of Smad2 and Smad3, whereas BMP-7 inhibits Smad signaling and protects the kidney through negative feedback regulation41. Interestingly, we found that the targets of DXXK for DM treatment were related to the TGF-beta signaling pathway and ECM in KEGG and GO enrichment analysis. Combined with the above results, it is clear that the TGF-β-related signaling pathway is particularly important in explaining the therapeutic mechanism of DXXK. Remarkably, we found that components of DXXK tightly binds to the TGF-β1/Smad signaling pathway during molecular docking. These analyses suggest that DXXK may attenuate diabetic kidney injury by modulating the TGF-β1/Smad pathway. Therefore, we established a mouse model of DM to verify the effect of DXXK on the fibrotic response and the role of the TGF-β1/Smad signaling pathway in DM.

Typical symptoms of DM include polyphagia, polydipsia, polyuria, and weight loss. Dioscin, DG, and protodioscin have all been shown to attenuate the adverse consequences of hyperglycemia by lowering blood glucose and modulating lipid metabolism42,43,44. In contrast, DXXK significantly ameliorated food and water intake, FBG levels, and insulin resistance in DM mice. Furthermore, DXXK modulated lipid metabolism in DM mice, which may be attributed to the potent lipid-lowering activities of dioscin and DG.

Severe kidney damage, such as glomerular basement membrane (GBM) thickening, mesangial expansion, glomerular sclerosis, and tubulointerstitial fibrosis is common in diabetic patients45,46. The critical ingredient of DXXK has been shown to have significant nephroprotective effects. Dioscin alleviates kidney injury by promoting ECM degradation and reducing urinary protein excretion23. DG reduced kidney lesions by inhibiting glomerular mesangial matrix expansion and decreasing GBM thickness47. Protodioscin improved kidney function in DM rats by lowering serum creatinine and blood urea nitrogen levels24. In our study, DXXK could protect the kidney by attenuating GBM thickening and inhibiting fibrosis. As a key regulator of glucose metabolism, the liver is vulnerable to hyperglycemia48. Dioscin and DG have been shown to be hepatoprotective by significantly reducing markers of liver injury (AST and ALT)49,50. Furthermore, DXXK was shown to attenuate lipid accumulation in the liver of hyperlipidemic mice51. Likewise, DXXK decreased AST and ALT levels and reduced lipid droplets in the livers of mice with DM.

TGF-β1 has been consistently shown to be an important pathogenic mediator of renal fibrosis, and its downstream proteins are primarily Smad2 and Smad311. Smad3 has been demonstrated to primarily mediate renal fibrosis, whereas Smad2 may play a nephroprotective role12. Studies have shown that downregulation of Smad2/3 phosphorylation delays podocyte injury and inhibits renal fibrosis by reducing TGF-β1 expression52. In contrast, Smad2 deficiency promotes fibrosis by enhancing TGF-β/Smad3 signaling while increasing TGF-β1 autoinduction53. To verify whether the mechanism of DXXK to protect the diabetic kidney is related to the TGF-β1/Smad signaling pathway, we examined the protein and mRNA expression of TGF-β1, p-Smad2/3, and Smad2/3. The results indicated that the TGF-β1/Smad2/3 signaling pathway was abnormally activated in the STZ-induced DM model. Nevertheless, the expression of TGF-β1, p-Smad2/3, and Smad2/3 in the renal tissues of the DXXK and metformin groups was significantly lower than that of the DM group, which was consistent with the results of the previous network pharmacology analysis. Taken together with the network pharmacology analyses, molecular docking, and in vivo experiments, the results suggest that DXXK has antifibrotic properties and may be useful in preventing and ameliorating diabetic kidney injury. However, our study was limited to capsules with complex compositions, and more specific experiments are needed to elucidate the antidiabetic effects and mechanisms of the individual components within DXXK.

Conclusion

This study used network pharmacology to identify potential therapeutic targets of DXXK in DM, which were further validated by in vivo experiments. Our results suggest that DXXK inhibits renal fibrosis by regulating TGF-β1/Smad2/3 signaling. These findings indicate that DXXK may be a useful adjunct in the prevention and treatment of DM.