Network pharmacology and molecular docking of endogenous active metabolites in diabetic kidney disease

Abstract

Objectives

Network pharmacology and molecular docking were used to predict endogenous active metabolites with protective effects in diabetic kidney disease (DKD).

Methods

We utilized metabolomics to screen differentially expressed metabolites in kidney tissues of mice with type 2 DKD and predicted potential targets using relevant databases. The interaction network between endogenous active metabolites and target proteins was established by integrating differentially expressed metabolites and proteins associated with DKD identified through proteomics. Gene ontology (GO) and signaling pathway enrichment analysis were performed. The biological functions of the active candidate metabolites and their effects on downstream pathways were also verified.

Results

Metabolomics revealed 130 differentially expressed metabolites. Through co-expression network analysis coupled with the investigation of differentially expressed proteins in proteomics, 2-hydroxyphenylpropionylglycine (2-HPG) emerged as a key regulator of DKD. 2-HPG was found to modulate the progression of DKD by regulating the conformation and activity of synaptophysin 1 (SYNJ1), with a correlation coefficient of 0.974. In vivo experiments revealed that SYNJ1 expression was significantly downregulated in the Macroalbuminuria Group compared to the Control Group and negatively correlated with proteinuria (r = −0.7137), indicating its important role in DKD progression. Immunofluorescence demonstrated that treatment with 2-HPG restores the expression of the foot process marker protein Wilms tumor-1 (WT-1) in podocytes injured by high glucose levels. Western blot and polymerase chain reaction support the involvement of SYNJ1 in this process.

Conclusions

This study demonstrated the significance of the 2-HPG/SYNJ1 signaling axis in safeguarding the foot process of podocytes in DKD.

Introduction

Diabetes is the leading cause of end-stage kidney disease worldwide. Since 2011, diabetes-related chronic kidney disease (CKD) cases have consistently exceeded glomerulonephritis-related CKD cases, accounting for 30–50% of all CKD cases and affecting 285 million (6.4%) adults worldwide [1]. This figure is expected to increase by 69% in high-income countries and 20% in low- and middle-income countries by 2030 [2]. Additionally, diabetic kidney disease (DKD) is also the strongest predictor of mortality in patients with diabetes. Currently, the main treatment of DKD involves effective management of blood sugar and blood pressure, as well as inhibition of the renin-angiotensin-aldosterone system.

In recent years, new therapeutic methods, including sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 analogs, have proven crucial in delaying the progression of DKD. However, there is currently no effective clinical treatment to reverse DKD, highlighting the urgent need to elucidate the molecular mechanisms underlying its progression.

DKD originates from abnormal glucose metabolism, leading to the reprogramming of various metabolic pathways and the synthesis disorder or abnormal accumulation of functional metabolites. These metabolites can serve as biomarkers of disease progression and potential drug targets. Additionally, enzymes, receptors, ion channels, and nucleic acids involved in these metabolic pathways might also be effective targets, becoming key avenues for drug discovery.

Metabolomics, through urine and plasma analysis, is an important tool for diagnosing DKD. Currently, identified metabolomic biomarkers for early DKD diagnosis mainly comprise fatty acids, amino acids, lipids, and nucleotides. For example, plasma tyrosine levels are negatively associated with an increased risk of DKD, being potential DKD biomarkers [3]. Increased sphingomyelin levels are associated with podocyte injury and disruption of the glomerular slit diaphragm, further associated with proteinuria [4]. The levels of 1, 5-anhydroglucitol are higher in urine samples from patients with type 2 diabetes and DKD compared to those without DKD [5]. Serum levels of L-dihydroorotic acid, 6-methylmercaptopurine, and piperidine were lower in patients with type 2 diabetes and DKD than in those without [6]. Using metabolomics and considering metabolic reprogramming, we anticipate identifying active metabolites that play a protective role against DKD, providing new targets for clinically diagnosing and treating this pathology.

In this study, we utilized metabolomics and proteomics to identify differential metabolites in the kidney tissue of mice with type 2 diabetes. Using bioinformatics, we predicted endogenous active metabolites and their regulatory targets. The protective effect of these metabolites against high glucose-induced podocyte injury in mouse kidney tissue was verified via immunohistochemistry and cellular assays. This study identified 2-HPG as a potential protective molecule against DKD, offering a new strategy for preventing and treating this disease.

Materials and methods

Type 2 diabetic kidney disease (T2DKD) model development

Db/db (C57BL/6J, BKS. Cd Dockm+/+Lepr db) mice and their litter control, db/m mice, were purchased from the Model Animal Research Center of Nanjing University (Nanjing, China) and kept in the Laboratory Animal Center of Shanghai Jiao Tong University Affiliated Tongren Hospital, Shanghai, PR China.

All animals were housed in air-conditioned rooms with a 12-h light/12-h dark cycle and free access to water and food. At 24 weeks of age, we collected 24-h urine to measure urinary albumin excretion (UAE) using indirect competition enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions (Albuwell M, Exocell, Philadelphia, PA). After urine collection, fasting blood samples were taken to measure blood glucose, glycated serum protein (GSP), blood urea nitrogen (BUN), and serum creatinine (Scr). Subsequently, mice were euthanized, and kidney samples were stored in liquid nitrogen for real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) and Western blot.

MPC-5 cell culture and treatment

The mouse kidney podocytes MPC-5 were purchased from the Chinese Academy of Sciences (Shanghai, China). Cells were incubated with Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Grand Island, NY; BRL), supplemented with 10% fetal bovine serum (FBS, Gibco, BRL), 100 U/mL penicillin (Amresco, Cleveland, OH), and 100 U/mL streptomycin (Amesco, USB) at 37 °C with 5% CO₂.

Following a 48-h incubation in a medium containing 5.5 mM D-glucose (Control Group) or 30 mM D-glucose (high-glucose [HG] group), cells were collected to assess the impact of high levels of glucose on MPC-5 cells. To examine the effect of 2-hydroxyphenylpropionylglycine (2-HPG) on HG-induced MPC-5 cells, cells in the HG group were treated with 2-HPG at different concentrations (1 μM, 10 μM, and 20 μM) for 48 h before being collected and analyzed.

Histology and immunohistochemical staining

The kidney tissue was fixed overnight in 10% formalin at room temperature, embedded in paraffin, and sectioned into 2–5 μm slices. Staining with hematoxylin and eosin (HE), periodic acid-Schiff (PAS), 3,3′-diaminobenzidine, and 4′,6-diamidino-2-phenylindole was performed to evaluate the morphological changes in renal sections. Additionally, a sample of the renal cortex (1 mm³) was stored in 2.5% glutaraldehyde for transmission electron microscope analysis.

Metabolomic analysis

Comprehensive metabolomic analyses in this study were performed using gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS) as previously described [7].

Chemicals

All utilized chemicals and solvents were of analytical or HPLC grade. Water, acetonitrile, methanol and formic acid were purchased from Thermo Fisher Scientific (Waltham, MA). Pyridine, methoxylamine hydrochloride (97%), n-hexane, and N, O-Bis(trimethylsilyl)trifluoroacetamide (BSTFA) with 1% trimethylchlorosilane (TMCS) were purchased from CNW Technologies GmbH (Düsseldorf, Germany). Chloroform was acquired from Titan Chemical Reagent Co., Ltd. (Shanghai, China), and L-2-chlorophenyl alanine from Shanghai Heng Chuang Bio Technology Co., Ltd. (Shanghai, China).

Sample preparation

GC-MS

The 30 mg sample was accurately weighed and transferred to a 1.5 mL Eppendorf tube along with two small steel balls. Subsequently, 20 µL of L-2-chlorophenylalanine (0.3 mg/mL) dissolved in methanol, serving as an internal standard, and 600 µL of extraction solvent (methanol/water, 4/1, vol/vol) was added to each sample. The samples were stored at −20 °C for 2 min and then ground at 60 HZ for 2 min. Subsequently, 120 µL of chloroform was added to the samples, and the mixtures were vortexed, underwent extraction by sonication in an ice-water bath for 30 min, and were stored at −20 °C for 20 min.

The samples were centrifuged at 4 °C (13,000 rpm) for 10 min, and 150 μL of the supernatant was dried in a glass vial using a freeze concentration centrifugal dryer. Subsequently, 80 μL of 15 mg/mL methoxylamine hydrochloride in pyridine was added. The resultant mixture was vortexed vigorously for 2 min and incubated at 37 °C for 90 min. Volumes of 50 μL BSTFA (containing 1% TMCS) and 20 μL n-hexane were added to the mixture, which was vortexed for 2 min and derivatized at 70 °C for 60 min. Samples were allowed to reach ambient temperature for 30 min before GC-MS analysis.

LC-MS

The 30 mg sample was accurately weighed and placed in a 1.5 mL Eppendorf tube with two small steel balls. Subsequently, 20 μL of L-2-chlorophenylalanine (0.3 mg/mL) dissolved in methanol, serving as the internal standard, and 0.6 mL of a methanol-water mixture (4/1, vol/vol) were added to each sample. Samples were stored at −20 °C for 2 min and ground at 60 Hz for 2 min. The whole samples underwent ultrasonic extraction for 10 min in an ice-water bath and were stored at −20 °C for 30 min. The extracts were centrifuged at 4 °C and 13,000 rpm for 10 min. Supernatants (150 μL) from each tube were collected using crystal syringes, filtered through 0.22 μm microfilters, and transferred to LC vials. These vials were stored at −80 °C until LC-MS analysis. Quality control samples were prepared by mixing aliquots of all samples to create a pooled sample.

Data preprocessing and statistical analysis

The GC/MS raw data in .D format were converted to .abf format using the Analysis Base File Converter software for efficient data retrieval. Subsequently, the data were imported into MS-DIAL software, which performs tasks, such as peak detection, peak identification, MS2Dec deconvolution, characterization, peak alignment, wave filtering, and missing value interpolation. Metabolite characterization is based on the LUG database. The resulting data matrix is 3D and includes sample information, peak names for each substance, retention time, retention index, mass-to-charge ratio (M/z), and signal intensity. Within each sample, all peak signal intensities were segmented and normalized based on internal standards with an RSD greater than 0.3 after screening. Following normalization, redundancy removal, and peak merging were conducted to obtain the final data matrix.

The original LC-MS data were processed using Progenesis QI version 2.3 software (Nonlinear, Dynamics, Newcastle, UK) for baseline filtering, peak identification, integration, retention time correction, peak alignment, and normalization. The key parameters included a 5 ppm precursor tolerance, 10 ppm product tolerance, and a 5% product ion threshold. Compound identification relied on precise M/z ratio, secondary fragments, and isotopic distribution using databases, such as The Human Metabolome Database (HMDB), Lipidmaps (V2.3), Metlin, The Electron Microscopy Data Bank (EMDB), The Protein Model DataBase (PMDB), and self-built databases for qualitative analysis. The extracted data underwent additional processing, involving the removal of peaks with a missing value (ion intensity = 0) in more than 50% of groups, replacement of zero values with half of the minimum value, and screening based on qualitative compound results. Compounds with scores below 36 (out of 60) points were considered inaccurate and excluded. A data matrix was then created by combining the positive and negative ion data. The matrix was imported into R for Principle Component Analysis to observe the overall sample distribution and assess the stability of the whole analysis process. Orthogonal Partial Least-Squares-Discriminant Analysis (OPLS-DA) and Partial Least-Squares-Discriminant Analysis were utilized to identify differing metabolites between groups. To prevent overfitting, the model’s quality was assessed using 7-fold cross-validation and 200 Response Permutation Testing.

Variable Importance of Projection (VIP) values from the OPLS-DA model ranked each variable’s overall contribution to group discrimination. A two-tailed Student’s T-test was applied to confirm the significance of inter-group metabolite differences. Differential metabolites were selected based on VIP values exceeding 1.0 and p values less than 0.05.

Functional analysis

Gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis

GO is a database established by the GO Consortium, aiming to establish a standard semantic vocabulary applicable across species. It defines and describes gene and protein functions, categorized into three categories: molecular function (MF), biological process (BP), and cellular components (CC). GO enrichment analysis offers insight into the biological functions, pathways, or cellular locations of differentially enriched genes.

KEGG is a comprehensive database integrating genomic, chemical, and system functional information. Among the seventeen sub-databases of KEGG, the KEGG pathway stores information about gene pathways in various species. Pathway analysis of differentially expressed genes provides valuable insights into significantly altered metabolic pathways, especially in mechanistic studies.

To elucidate the role of the predicted target in gene function and signaling pathways in response to 2-HPG, GO enrichment analysis and KEGG pathway analysis were performed on the identified key target genes using R-language. This analysis aimed to determine the MF of the protein, its intracellular localization, biological response, and the metabolic pathways involved.

Correlation analysis of differentially expressed metabolites enriched in KEGG pathway and revealed by proteomics

The Pearson correlation analysis was performed between differentially expressed metabolites and proteomic results, and the coefficient with r absolute value > 0.95 was selected for polar cluster visualization using OriginLab 2022.

Drug-active ingredient-key target-disease network construction and core target screening

Using Cytoscape version 3.6.3 software (CA, USA), we created the network illustrating the relationship between key targets and diseases affected by the drug. The weight value was determined by the Pearson coefficient. Subsequently, Hubba clustering analysis was applied to the network. The eccentricity Top10 was selected as the clustering method.

Molecular docking

To further explore the binding mode of 2-HPG with related targets, we used molecular docking. We downloaded the 3D structure of 2-HPG from Pubchem and the 3D structure of synaptophysin 1 (SYNJ1) from the PDB database. PyMOL version 2.3 software (NY, USA) was used to delete irrelevant small molecules in core target protein molecules, import protein molecules into AutoDockTools version 1.5.6 software (CA, USA), erase water molecules, and add hydrogen atoms. The files were saved under pdbqt format. The molecular structures of nuclear components were imported into AutoDockTools version 1.5.6 software (CA, USA), water molecules were erased, atomic charges were added, atomic types were assigned, and all flexible keys were rotated by default. Files were saved under pdbqt format and imported into PyMOL version 2.3 system analysis software (NY, USA).

Semi-quantitative analysis

Initially, Image Pro Plus (IPP) version 6.0 software (American Media Cybernetics Inc., Rockville, MD) was employed to preprocess the acquired image, creating a standardized image for subsequent quantitative evaluation. The most important techniques in image preprocessing, including contrast enhancement, color normalization, and denoising, were applied.

Western blot and RT-PCR analysis

The expression levels of SYNJ1 and Wilms tumor-1 (WT-1) after exposure of MPC-5 cell to 20 μM 2-HPG intervention were detected using Western blot and qRT-PCR. SYNJ1 immunohistochemistry first antibody information (Proteintech, Rosemont, IL; 24677-1-AP, 1:500), SYNJ1 WB first antibody information (Servicebio, Hubei, China; GB11382, 1:500).

Results

Db/db mouse model

Based on UAE value, db/db mice were divided into the Macroalbuminuria Group (UAE ≥ 300 μg/24 h) and Microalbuminuria Group (UAE = 30–299 μg/24 h). At 24 weeks, both the microalbuminuria (31.7 ± 1.028 mmol/L, n = 6) and the macroalbuminuria (32.93 ± 0.2361 mmol/L, n = 6) groups had significant hyperglycemia (p < 0.0001) compared to the Control Group (8.6 ± 0.2113 mmol/L, n = 6; Figure 1(A)). Similar results were obtained for GSP, BUN, and Scr (Figure 1(A)).

Figure 1.

A: Determination of blood glucose level, GSP, BUN, Scr, and UAE of DKD animal model; quantification of glomerular surface area, relative mesangial area, and GBM thickness. B: HE staining (400x), PAS staining (400x), electron microscopy (12000x) of DKD animal model. (*p < 0.05, **p < 0.01, and ***p < 0.001, n = 6). The red arrows in Figure 1(B) indicate the fusion of the podocyte foot processes.

Regarding UAE, statistical differences were observed between the Microalbuminuria Group (30.33 ± 4.613 μg/24 h, n = 6) and the Macroalbuminuria Group (531 ± 81.02 μg/24 h, n = 6) compared to the Control Group (24.66 ± 1.12 μg/24 h, n = 6) (p < 0.05) (Figure 1(A)). Representative histopathological and electron microscopic data from the three groups are illustrated in Figure 1(B). HE and PAS staining revealed more pronounced kidney histopathological changes in the Macroalbuminuria Group compared to Control Group. Thus, HE staining revealed that the former exhibited more pronounced kidney histopathological changes, including glomerular hypertrophy, increased glomerular surface area, interstitial infiltration of multiple nuclear cells, and renal tubule atrophy (Figure 1(A,B)). PAS staining further revealed that the Macroalbuminuria Group exhibited increased mesangial matrix and mesangial cell proliferation (Figure 1(A,B)). Electron microscopy displayed apparent fusion and thickening of the base membrane in the Macroalbuminuria Group compared to the Control Group, while the Microalbuminuria Group exhibited kidney pathology between the two (Figure 1(A,B)).

Metabolomic analysis of the renal tissue

We conducted a comprehensive metabonomic study using GC-MS and LC-MS techniques in the three groups of mice to enhance the accuracy, integrity, and sensitivity of phenotype-related metabolite identification [8]. GC-MS identified 40 differentially expressed metabolites, including 21 upregulated and 19 downregulated metabolites. LC-MS identified 90 differentially expressed metabolites, including 54 up- and 36 down-regulated metabolites (Figure 2(A–D)). KEGG pathway enrichment analysis revealed 29 significantly different pathways (p < 0.05). Key metabolites were mainly involved in glycine, serine, and threonine metabolism, lipid metabolism, insulin resistance, valine, leucine, and isoleucine biosynthesis, nuclear factor kappa B (NF-κB) signaling pathway, and cyclic adenosine 3′,5′-monophosphate (cAMP) signaling pathway (Figure 2(E)).

Figure 2.

Differential metabolite screening. A: Differential metabolite heat map of the Macroalbuminuria Group and Microalbuminuria Group (GC-MS) metabolites in mouse animal models of DKD. B: Volcano plots of the Macroalbuminuria Group and Microalbuminuria Group (GC-MS) metabolites in mouse animal models of DKD. C: Differential metabolite heat map of the Macroalbuminuria Group and Microalbuminuria Group (LC-MS) metabolites in mouse animal models of DKD. D: Volcano plots of the Macroalbuminuria Group and Microalbuminuria Group (LC-MS) metabolites in mouse animal models of DKD. E: Enrichment analysis of KEGG pathway of p < 0.05 differential metabolite.

We analyzed the correlation between the 26 metabolites significantly enriched in the KEGG pathway, up-regulated in the Microalbuminuria Group, and proteomic results, identifying 698 pairs of significant correlations (absolute value of the Pearson correlation coefficient ≥ 0.95) (Figure 3(A)). A co-expression network was created using Cytoscape, with the Pearson coefficient as the weight value, resulting in a network diagram (Figure 3(B)). Subsequently, the network underwent Hubba clustering analysis, selecting Eccentricity Top10 as the clustering method to screen 2-HPG and Q8CHC4 (Figure 3(C)) with a positive correlation of 0.974246977. The coding gene of Q8CHC4 is SYNJ1. Correlation analysis between metabolites corresponding to endogenous protection and proteomic results was performed, establishing positive and negative correlations (> 0.95 or < −0.95). Trend screening was based on fold change (FC), where an increase was determined if FC > 1.2, stability was indicated when 0.833 ≤ FC ≤ 1.2, and the decrease was FC < 0.833.

Figure 3.

Screening of active metabolites and targets. A: Correlation analysis between differential metabolites of KEGG signal enrichment and proteomics. B: Co-expression network analysis of joint analysis. C: Co-expression network Hubba cluster analysis (eccentricity Top10). D: 2-HPG expression in metabonomics (*p < 0.05). E: 2-HPG 3D structure. F: SYNJ1 3D structure. G: 2-HPG and SYNJ1 molecular docking.

The endogenous metabolite 2-HPG was expressed in both the control and microalbuminuria groups and was down-regulated in the Macroalbuminuria Group (Figure 3(D)). Comprising 11 C, 13 H, 1 N, and 4 O, this metabolite belongs to the n-acyl group. It is an α-organic amino acid resulting from the metabolic reprogramming of phenylalanine and glycine. In vitro, the compound can be obtained by heating and refluxing 3-phenylpropionamide, glyoxylic acid, and tetrahydrofuran, followed by spinning, drying, and the addition of ethers and filters.

The 3D structure of 2-HPG sourced from Pubchem is illustrated in Figure 3(E) and that of SYNJ1 sourced from PDB is presented in Figure 3(F). Molecular docking revealed that 2-HPG might interact with GLY871, GLY873, and GLN827 of SYNJ1 through hydrogen bonding (Figure 3(G)), with a comprehensive score of 6.1.

Expression of SYNJ1 in animal models

Immunohistochemistry of the kidney tissue collected from the three groups of mice indicated that podocyte SYNJ1 expression was significantly down-regulated in the Macroalbuminuria Group compared to the Control Group. Furthermore, SYNJ1 expression level was negatively correlated with proteinuria (Figure 4).

Figure 4.

Expression of SYNJ1 in animal models. A: Immunohistochemistry of SYNJ1 mouse DKD animal model (400x). Red arrows indicate SYNJ1 expression in podocytes. B: Immunohistochemistry semiquantitative analysis of SYNJ1 mouse DKD animal model (^****p < 0.0001). C: Correlation analysis of SYNJ1 semiquantitative results and urine protein (r = −0.7137).

2-HPG improves HG-induced MPC5 damage through SYNJ1 regulation

In HG-treated MPC-5 cells, 2-HPG increased the expression of the foot cell marker WT-1 compared to the control. Furthermore, the increase was dose-dependent. Western blot and PCR also indicated that 2-HPG could regulate the expression of SYNJ1 and alleviate HG-induced podocyte injury (Figure 5).

Figure 5.

2-HPG improves HG induced MPC5 damage through SYNJ1. A: High glucose-induced 48 h cellular immunofluorescence under intervention of 2-HPG in different concentrations (green fluorescence represents WT-1, blue fluorescence represents DAPI stained nucleus (200x)). B: Semiquantitative analysis of cellular immunofluorescence. C: Western blot results of SYNJ1 and WT-1 after 20 μM 2-HPG intervention. D: WB semiquantitative analysis of SYNJ1 and WT-1. E: RT-PCR results after 20 μM 2-HPG intervention (*p < 0.05 compared to Control Group; ^#p < 0.05 compared to HG group, n = 3).

Discussion

Thickening of the glomerular base membrane (GBM) is the initial lesion observed using electron microscopy in patients with DKD. Mesangial dilation represents the histopathological lesion most frequently observed under light microscopy. Nodular glomerulosclerosis, particularly typical Kimmelstein-Wilson nodes, stands out as a highly specific DKD lesion [9]. In this study, these pathological alterations were observed in the kidney tissue of mice with DKD in both groups compared to the Control Group. Notably, the Macroalbuminuria Group exhibited more pronounced pathological changes than the Microalbuminuria Group.

Metabolomics holds promise for advancing nephrology research. In this study, 130 differentially expressed metabolites, comprising 75 up- and 55 down-regulated metabolites, were examined using metabolomics. When enrichment analysis of differential metabolites was performed, there were significant differences observed in the metabolic pathways, particularly in the glycine, serine, and threonine metabolism, lipid metabolism, insulin resistance, and other signaling pathways. Coupled with differentially expressed proteins identified through proteomics, 2-HPG appears to play a crucial regulatory role in the DKD process, as suggested by co-expression network analysis.

We analyzed the correlation between the 26 metabolites significantly enriched in the KEGG pathway, up-regulated in the Microalbuminuria Group, and proteomic results, and constructed a Cytoscape co-expression network for the obtained correlation. Subsequent Hubba cluster analyses in the network identified 2-HPG and Q8CHC4 (coding gene SYNJ1) as being significantly and positively correlated. Therefore, 2-HPG can protect against DKD by regulating the conformation and activity of SYNJ1.

SYNJ1, a member of the synaptojanin protein family, comprises three domains: actin-1 inhibitor (SAC1), 5-phosphatase, and proline-rich domain (PRD) [10].

Unlike most proteins, SYNJ1 features two enzymatic domains crucial for lipid homeostasis and Synj2-mediated signal transduction and membrane transport. Previous studies have linked SYNJ1 abnormalities to various neurological and neuropsychiatric disorders, including Parkinson’s disease, Alzheimer’s disease, and Down syndrome [11–13]. However, there is limited research on the role of SYNJ1 in kidney disease [14].

KEGG enrichment analysis of 2HPG co-expressed genes revealed significant enrichment in glycine, serine, and threonine metabolism, lipid metabolism, insulin resistance, valine, leucine, and isoleucine biosynthesis, NF-κB signaling pathway, and cAMP signaling pathway, etc. Recent studies emphasized the crucial role of metabolites in directly binding to proteins and regulating biological functions. For example, 3-oxolithocholic acid inhibited Th17 cell differentiation by binding to the retinoic acid-related orphan receptor gamma-t (RORγt) ligand binding domain [15].

Furthermore, 2-HPG had similar biological activities, providing 3 hydrogen bond acceptors and 4 hydrogen bond donors. This suggests that 2-HPG may participate in the progression of DKD by binding to a target protein and regulating its steric conformation and activity. The Swiss Target Prediction database predicted SYNJ1 as the binding protein of 2-HPG, and network analysis demonstrated the highest correlation between 2-HPG and SYNJ1.

The spatial structure prediction of 2-HPG in SYNJ1 revealed hydrogen bond interactions between 2-HPG and GLY871, GLY873, and GLN827 residues of SYNJ1, with a comprehensive score of −6.1, indicating a potential interaction. These results suggest that SYNJ1 may be involved in the protective mechanism of 2-HPG on renal function. Immunohistochemistry further revealed a significant down-regulation of SYNJ1 expression in the kidney tissue of the Macroalbuminuria Group compared to the Control Group. The level of expression of SYNJ1 was negatively correlated with proteinuria (r = −0.7137), suggesting that SYNJ1 may play an important role in the progression of DKD.

Podocytes are highly specialized cells in the capillary ring. Together with the thin diaphragm, they are crucial in forming the renal filtration barrier through the foot processes [16]. The protein network governing podocyte foot process development mirrors that of synaptic development in neurons [17–19]. Genetic studies highlight the significance of actin-regulatory proteins and upstream signaling factors for the integrity of podocyte foot processes, emphasizing actin’s central role [20]. Dynein’s function in podocytes may be mediated by its role in the actin cytoskeleton [21,22]. Additionally, the presence of clathrin-coated vesicles in podocyte foot processes raises the possibility that endocytosis defects contribute to abnormal foot processes [23]. SYNJ1-KO mice exhibit normal embryonic development but fail to establish a proper filtration barrier, resulting in severe proteinuria due to abnormal podocyte foot process formation [14]. This reveals that SYNJ1 plays an important role in podocyte foot process development.

Cellular immunofluorescence revealed that 2-HPG restored the expression of the foot process marker protein WT-1 down-regulated by high glucose levels. Western blotting and RT-PCR indicated that SYNJ1 may be involved in this process.

However, this study has some limitations. Metabolomics detection has a limited dynamic range, making it challenging to detect substances with large content differences simultaneously in the same sample. Metabolites are also susceptible to disturbances from factors, such as diet, environment, and age. Moreover, the small number of kidney samples in each group restricts this study, highlighting the need for larger sample sizes in future research. Additionally, the herein presented in vivo experiment should be further developed to evaluate the safety of this active metabolite, allowing its clinical use.

Conclusion

Differentially expressed metabolites in the kidney tissue of mice with DKD were studied using metabolomics and proteomics. The endogenous active metabolite 2-HPG was predicted by bioinformatics to regulate the conformation and activity of SYNJ1 and the progression of DKD. Furthermore, its protective effect against HG-induced podocyte injury was verified by immunohistochemistry and other relevant tests. Our results provide a novel target for investigating the underlying mechanisms, diagnosis, and treatment of DKD.

Funding Statement

This study was supported by the National Natural Science Foundation of China (Grant No. 82000687, 82170745, and 82100766) and the Research Fund of Shanghai Tongren Hospital, Shanghai Jiaotong University School of Medicine (2020TRYJ(JC)03).

Statement of ethics

The animal study was reviewed and approved by the Animal Care and Use Committee of Shanghai Jiao Tong University School of Medicine affiliated with Tongren Hospital (Ethical Approval No. A2023-020-01).

Disclosure statement

The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this article.

Author contributions

WXX conceived and supervised the study and acquired funding. XXM, WYZ, and LYY performed the AFADESA-MSI and metabolomics and data analyses. LFQ, ZCC, and CSJ performed the animal study and/or contributed to the materials. CSJ and WYZ performed the validation experiments. XXM prepared the manuscript draft. ZL and WXX revised the manuscript. All authors participated in the discussion and editing of the manuscript. XXM and WYZ were co-first authors of the article.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material, and further inquiries can be directed to the corresponding author.