Safflower injection against obesity-induced mice podocyte injury by improving insulin resistance through increasing renal INSR and eNOS expression

Abstract

Background

Podocyte injury is a common pathologic mechanism in obesity-related glomerulopathy (ORG). Safflower injection (SFI), scientifically extracted and refined from safflower, is used to treat diabetic kidney disease according to clinical guideline. Our previous study confirmed that the main active compounds of SFI ameliorated high glucose-induced podocyte injury. It is uncertain whether SFI has an effect on ORG-related podocyte injury.

Objectives

This study aimed to explore the pharmacological effects and related mechanisms of SFI on podocyte injury of ORG mice.

Methods

First, by combining ultra-high performance liquid chromatography tandem mass spectrometry analysis with online databases, the pathway enrichment, target-pathway analysis, and human protein–protein interaction network were conducted to discover the possible crucial mechanism of SFI against ORG. Then, ORG mice model was established by high-fat diet and biochemical assays, histopathology and western blot were used to explore the effects of SFI on obesity and podocyte injury. Finally, system pharmacology-based findings were evaluated in ORG mice.

Results

The results of system pharmacology suggested that SFI could alleviate ORG through insulin resistance (IR)-related pathway by regulating insulin receptor (INSR) and endothelial nitric oxide synthase (eNOS) expressions. The in vivo experiment confirmed that SFI ameliorated obesity, lipid metabolism-related indicators, podocyte injury of ORG mice. The mechanism relationships among IR, INSR, and eNOS were further verified in ORG mice.

Conclusions

Our findings imply that by up-regulating the expression of renal INSR and eNOS, thereby inhibiting IR, SFI may be a promising candidate for the treatment of ORG.

1. Introduction

Recently, the global prevalence of obesity has shown a rapidly increasing trend [1]. Epidemiological studies have shown that obesity has emerged as a significant independent risk factor for chronic kidney disease (CKD) [2]. Obesity-related glomerulopathy (ORG) clinically characterized by proteinuria, is a glomerular disease secondary to obesity [3]. The main etiology of ORG is podocyte damage, but contributing theories include dysfunctional renin–angiotensin–aldosterone system activation, hyperinsulinemia, and lipid deposition [4]. Traditional Chinese medicine (TCM) has already demonstrated its efficacy in treating podocyte injury [5].

Safflower is the dried tubular flower of Carthamus tinctorius, with a long history of use for promoting blood circulation and alleviating blood stasis. Safflower injection (SFI), a scientifically extracted and refined injection from safflower, has been used for the treatment of cardio-cerebrovascular diseases [6]. Safflower yellow (SY), the main ingredient in SFI, can reduce body fat and blood lipid levels, improve glucose tolerance, and enhance insulin sensitivity in obese mice [7]. Meanwhile, SY is recommended in early CKD patients clinically based on its improvement on blood glucose, insulin resistance (IR), blood viscosity, and renal function [8]. Our previous study demonstrated that some active compound in SFI has a protective effect against lipopolysaccharide and high glucose-induced podocyte injury in vitro [9]. However, the efficacy of SFI in ameliorating podocyte injury in ORG and the related mechanism is limited.

The complex chemical composition and wide range of targets of SFI make it relatively difficult to study. Therefore, it is urgent to find more effective methods to study the mechanism of SFI against ORG. Studies have been conducted to investigate the pharmacokinetic markers of SFI for ameliorating cardiomyocyte injury by integrating systems biology and ultra-high performance liquid chromatography tandem mass spectrometry (UHPLC–MS/MS) techniques, but the mechanisms have not been thoroughly investigated [10]. Therefore, the combination of UHPLC–MS/MS, the TCM databases, ingredient potential target prediction platforms, and protein–protein interaction (PPI) network analyses will be beneficial to reveal the key regulatory pathways and potential targets of SFI for ameliorating renal injury in ORG [11].

In this study, the components of SFI were first characterized by UHPLC–MS/MS. Based on the system pharmacology approach, the mechanisms of SFI to improve ORG were focused on IR-related pathways and targets, including insulin receptor (INSR) and endothelial nitric oxide synthase (eNOS). Next, the model of ORG mice was established using a high-fat diet (HFD) to study the effects of SFI on obesity, abnormal lipid metabolism, and renal impairment in ORG mice. Finally, the role and mechanism of SFI in improving podocyte injury in ORG mice were verified by immunohistochemistry. This is a first trying to explore the effects and possible mechanisms of SFI on alleviating HFD-induced podocyte injury.

2. Materials and methods

2.1. Chemicals and reagents

SFI was purchased from Shineway Pharmaceutical Group Co., Ltd. (Shijiazhuang, China, LOT No. 220903c1). The both standard diet and HFD were obtained from Boaigang Biotechnology Company (Beijing, China). The energy ratio of standard diet was as follows: 12.0% fat, 20.6% protein, and 67.4% carbohydrates, and the HFD was 60% fat, 20% protein, and 20% carbohydrates. The analytical grade organic solvents, such as methanol and acetonitrile, were bought from Thermo Fisher Scientific (Waltham, MA).

Total cholesterol (TC) test kit, total triglyceride (TG) test kit, and urine protein quantitative test kit were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Glycogen periodic acid-Schiff (PAS) staining kit and RIPA lysis buffer were from Solarbio Biomedical Technology Co., Ltd. (Beijing, China). SDS-PAGE Protein Sampling Buffer (denaturing, 5×) and rapid Gel Preparation Kit were obtained from Yase Biomedical Technology Co., Ltd. (Shanghai, China).

2.2. SFI components collection by UHPLC–MS/MS

The sample of SFI 200 μL and 600 μL methanol were taken in an EP tube which was then dried by vacuum at 35 °C. The 100 μL 50% methanol–water was added to the dried sample. After filtration with 0.22 μm organic filter membrane, the sample was analyzed by an UPLC system (Shimadzu, Kyoto, Japan).

The chromatographic separation was carried out on Waters XSelect HSS T3 XP column (2.1 × 100 mm, 1.8 μm, Milford, MA) with the mobile phase composed of 0.1% formic acid in water and 0.1% formic acid in acetonitrile under gradient elution. The AB 5600 Triple TOF mass spectrometer (AB SCIEX, Framingham, MA) was capable of primary and secondary mass spectrometry data acquisition based on IDA functionality under Analyst TF 1.7, AB Sciex software control. In each acquisition cycle, the molecular ions with the maximum intensity and over 100 were selected for the acquisition of the corresponding secondary mass spectrometry data. MS detection range was m/z 50–1200; electrospray bombardment energy was 30 eV; secondary mass spectra were acquired every 50 ms. The ESI settings were as follows: gas pressure: 60 Psi; auxiliary air pressure: 60 Psi; capillary temperature, 350 °C; in positive ion mode: spray voltage, 5 kV; in negative ion mode: spray voltage, 4 kV.

The composition of the precursor and fragment ions were processed and sequenced by MassHunter software. After the data were converted to mzML format, MS-DIAL software was utilized for peak identification, peak filtration, and peak alignment, to obtain the data matrix, including m/z, intensity and retention time. Components identification was first confirmed on the basis of accurate molecular weight (Δ ≤ 30 ppm) and then further validation was based on the secondary mass spectrometry database (MassBank, GNPS, mzCloud, etc.).

2.3. Safflower components collection from TCMSP

Based on Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (https://old.tcmsp-e.com/tcmsp.php, TCMSP), using ‘Honghua’ as the keyword search, the components with the Latin name ‘Carthami Flos’ were selected. The screening criteria were oral bioavailability (OB) ≥30%, drug likeness (DL) ≥0.18.

2.4. SFI potential active targets collection

The 2D structural information of the components in SFI was imported into PharmMapper (http://lilab-ecust.cn/pharmmapper/index.html), and the human protein data were selected for prediction to obtain the potential active targets. The Norm fit score ≥0.5 was used as the screening criterion. Finally, in the UniProt database (https://www.uniprot.org/), the UniProt ID of human targets were obtained.

2.5. ORG therapeutic targets collection

ORG therapeutic targets and corresponding UniProt IDs were obtained from the GeneCards database (https://www.genecards.org/) using the search term ‘obesity-related glomerulopathy’. Next, in the GEO database, the transcriptome data GSE53996 of the kidney tissues of C57/BJ mice fed with HFD for 20 weeks were obtained. And based on GEO2R online analysis tool, the differentially expressed genes (DEGs) were screened by adjusting p < .05 and absolute value of fold change >1. The obtained DEGs were imported into the Ensembl database (https://asia.ensembl.org/) to obtain human gene IDs. Finally, the UniProt IDs of DEGs were obtained in the UniProt database. The targets for SFI treatment ORG were obtained by taking the intersection of the active targets corresponding to the SFI components with the ORG therapeutic targets.

2.6. Pathway enrichment analysis and PPI network construction

The intersecting targets for SFI treatment ORG were subjected to gene ontology biological process (GO-BP) functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis in Database for Annotation, Visualization and Integrated Discovery (DAVID). The key pathways and key targets related to IR obtained from GO-BP functional enrichment analysis and KEGG pathway enrichment analysis were subjected to pathway-target analysis.

Protein–Protein Interaction Networks Functional Enrichment Analysis (https://cn.string-db.org/, STRING) online database was used to obtain the PPI network of intersecting targets for SFI treatment ORG, and the data were imported into Cytoscape 3.10.1 software to construct the PPI network, and the node connectivity was analyzed to screen out the core targets.

2.7. Animal experiments

Male C57BL/6J mice (4-week-old) were purchased from Laboratory Animal Center of Shanxi Provincial People’s Hospital (Taiyuan, China). Animals were kept under specific pathogen-free conditions (12-h light/dark cycle, 25 ± 2 °C) with free access to food and water. After one week of normal dietary acclimatization feeding, the mice were randomly divided into two groups: ordinary food normal control group (NOR, n = 6) and HFD group (n = 18), and weighed once a week. After 8 weeks of feeding, the mean body weight (BW) of HFD group mice was greater than 20% for NOR, which was confirmed as obesity [12].

Then, the HFD group was randomly divided into three groups: the ORG model group (ORG, n = 6), the SFI low-dose group (SFI-L, n = 6), and the SFI high-dose group (SFI-H, n = 6). The mice in ORG group, SFI-L group, and SFI-H group were continued to be fed with HFD until the end of the experiment. Starting from week 9, distilled water was gavaged once a day in the NOR and ORG groups, and SFI was gavaged once a day in the SFI-L and SFI-H groups at a dose of 0.33 g/kg and 0.99 g/kg, respectively. The duration of the intervention was 12 weeks (the flowchart is shown in Figure 1) [13].

Figure 1.

Schematic diagram of animal experiment for safflower injection (SFI) improvement high-fat diet (HFD)-induced obesity-related glomerulopathy (ORG) mice. FBG: fast blood glucose; FINS: fast insulin; TC: total cholesterol; SCr: serum creatinine; BUN: blood urea nitrogen; TG: triglycerides; NOR: normal control; SFI-L: ORG mice treated with 0.33 g·kg⁻¹·d⁻¹ SFI; SFI-H: ORG mice treated with 0.99 g·kg⁻¹·d⁻¹ SFI.

The BW of the mice at weeks 0, 4, 8, 12, 16, and 20 was recorded, and the body weight ratio (BWR) between NOR group and other groups was calculated. The calculation formula is as follows [14]:

$$\text{BWR\hspace{0.33em}(\%)=}\frac{\begin{array}{l}\text{the\hspace{0.33em}average\hspace{0.33em}BW\hspace{0.33em}of\hspace{0.33em}other\hspace{0.33em}groups}-\\ \text{the\hspace{0.33em}average\hspace{0.33em}BW\hspace{0.33em}of\hspace{0.33em}NOR\hspace{0.33em}group}\end{array}}{\text{the\hspace{0.33em}average\hspace{0.33em}BW\hspace{0.33em}of\hspace{0.33em}NOR\hspace{0.33em}group}}\times 100$$ (1)

At the end of experiment, Lee’s index was calculated to assess obesity in each group. The calculation formula is as follows [15]:

$$\text{Lee}\prime \text{s\hspace{0.33em}index\hspace{0.33em}=\hspace{0.33em}}\frac{\sqrt[3]{\text{BW\hspace{0.33em}(g)}}\times 1000}{\text{Body\hspace{0.33em}length\hspace{0.33em}(cm)}}$$ (2)

2.8. Samples collection and biochemical assay

After the experiment beginning, 24-h urine was collected and recorded from each group of mice by metabolic cages at weeks 10, 12, 16, and 20, and urinary protein was determined by the Coomassie Brilliant Blue method.

At the end of the experiment, blood, bilateral kidney, and liver tissues were taken. Blood was centrifuged to obtain serum, which was used to measure serum creatinine (SCr), blood urea nitrogen (BUN), fasting insulin (FINS) content (Jiangsu Meimian Industrial Co. Ltd., Yancheng, China, MM-0880R1, MM-0692M1, MM-0579M1), fasting blood glucose (FBG) content (Nanjing Jiancheng Bioengineering Institute, Nanjing, China, A154-1-1), and TC content (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Liver tissues were used for histopathological examination and homogenized for determination of TGs content by GPO-PAP kit assay (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Perirenal fat was weighed and kidney tissues were used for histopathological examination and western blot experiments. Based on FBG and FINS, the homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as follows [16]:

$$\text{HOMA‐IR =}\frac{\text{FBG}\left(\text{mmol/L}\right)\mathrm{\times }\text{FINS}\mathrm{(}\text{mIU/L}\mathrm{)}}{\text{22.5}}$$ (3)

2.9. Histopathological examination

2.9.1. Light microscopy

Freshly dissected kidneys were fixed in 10% neutral formalin buffer, embedded in paraffin. The embedded kidney was cut into 3 μm thick sections with Leica Microtome (Wetzlar, Germany). The sections were stained with PAS reagent [17], and the morphological changes of glomerulus were observed with KF-PRO-005-EX digital scanner (Motic, Xiamen, China).

The images were selected randomly in which the glomerulus number was no more than 20 and were analyzed with K-Viewer (1.5.3.1) image analysis software (×400, KFMI, Yuyao, China). The length of the two longest perpendicular diameters in each glomerular capillary tuft without Bowman’s space was measured in μm, and then the mean value was calculated from 10 images. The areas of the glomerular mesangial region and capillary tuft were measured. The relative area of the mesangial region (%) was calculated according to the formula [18]:

$$\text{The\hspace{0.33em}relative\hspace{0.33em}area\hspace{0.33em}}\left(\text{\%}\right)\text{\hspace{0.33em}=\hspace{0.33em}}\frac{\text{area\hspace{0.33em}of\hspace{0.33em}the\hspace{0.33em}\hspace{0.17em}mesangial\hspace{0.33em}region}}{\text{area\hspace{0.33em}of\hspace{0.33em}the\hspace{0.33em}capillary\hspace{0.33em}tuft}}\times 100$$ (4)

2.9.2. Electron microscopy

Freshly dissected kidneys were fixed using 2.5% glutaraldehyde. The ultrathin sections were stained with uranium acetate–lead citrate for electron microscopy. For each specimen, 10 separate photographs (20,000× magnification) covering different regions in the glomerular cross section were taken. The thickness of glomerular basement membrane (GBMT) and the width of slits were all measured [19]. Images were analyzed using the RADIUS Control & Imaging software (EMSIS ASIA, Münster, Germany).

2.10. Western blot

Renal tissues (about 20 mg) were homogenized with RIPA lysis buffer (1:10, w/v) containing protease inhibitors (Yase, Shanghai, China). After centrifugation at 14,000 × g for 10 min, the supernatants were collected and protein levels were determined using the BCA protein assay kit (Yase, Shanghai, China). Proteins were separated by 10% SDS-PAGE and then transferred to a PVDF membrane (Merck, Dublin, Ireland). Membranes were blocked with 5% nonfat milk in PBST buffer (25 mM Tris pH 7.6, 150 mM NaCl, 2.5 mM KCl, and 0.1% Tween 20) for 2 h and probed with primary antibodies overnight at 4 °C, including nephrin (Abcam, Waltham, MA) and synaptopodin (Proteintech, Wuhan, China). GAPDH (Proteintech, Wuhan, China) was used as a loading control. After the membrane was washed with PBST three times, it was incubated with HRP-conjugated anti-rabbit or anti-mouse antibody for 90 min at room temperature. The bands were visualized by gel imaging system (Bio-Rad, Hercules, CA) after using a ultra-sensitive ECL luminescent solution (Yakoin, Wuhan, China) [18].

2.11. Immunohistochemistry

Renal tissues were fixed in 10% buffered formalin for three days, embedded in paraffin and then sectioned at a thickness of 4 mm. After deparaffinization, antigen retrieval and blocking with goat serum, the section was incubating with anti-INSR or anti-eNOS rabbit polyclonal antibodies (Servicebio, Wuhan, China) overnight at 4 °C. After washed with PBS, the sections were incubated with HRP-conjugated goat anti-rabbit IgG secondary antibody (Servicebio, Wuhan, China) and the peroxidase reaction was developed with diaminobenzidine (Servicebio, Wuhan, China) [20]. Subsequently, the sections were counterstained with hematoxylin, and the images were captured using an inverted microscope.

2.12. Statistical analysis

SPSS 22.0 software (SPSS Inc., Chicago, IL) was used for statistical analysis. Measurements were expressed as mean ± standard deviation (mean ± SD), and comparisons of means between groups were analyzed using a one-way ANOVA, with differences between the two groups were considered statistically significant at p < .05.

3. Results

3.1. Collection of SFI components and corresponding targets

Pipeline of network pharmacology and experimental verification is shown in Figure 2. Through UHPLC–MS/MS, 104 components of SFI were identified and listed in Table S1. The chromatograms of their positive and negative ion modes are shown separately in Figure 3(A,B). The TCMSP database was searched for 189 components, and 22 active components were obtained and screened, as shown in Table S2. After deduplication and merging of the active components obtained through both methods, a total of 123 active components were identified.

Figure 2.

Pipeline of network pharmacology and experimental verification and qualitative analysis of SFI components by UHPLC–MS/MS. UHPLC–MS/MS: ultra-high performance liquid chromatography tandem mass spectrometry; TCMSP: Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform; HFD: high-fat diet; LFD: low-fat diet; SFI: safflower injection; ORG: obesity-related glomerulopathy; KEGG: Kyoto Encyclopedia of Genes and Genomes; GO: Gene Ontology; PPI: protein–protein interaction.

Figure 3.

Experimental verification and qualitative analysis of SFI components by UHPLC–MS/MS. (A) Ion chromatograms in positive mode (POS) of SFI. (B) Ion chromatograms in negative mode (NEG) of SFI.

Considering the limitations of UHPLC–MS/MS detection methods, such as sensitivity and separation efficiency of some compounds in safflower, the active components predicted by the website were supplemented. The potential targets of 123 active ingredients of SFI were obtained through the PharmMapper database. A total of 118 active components corresponded to 17,729 potential targets, and 371 potential targets of SFI were finally obtained by deleting the duplicate targets.

3.2. Collection of ORG therapeutic targets

A total of 190 therapeutic targets related to ORG were obtained from the GeneCards database, as shown in Table S3. From GSE53996, 486 DEGs corresponding to human gene IDs were identified. Among them, 319 genes were up-regulated, and 167 genes were down-regulated, as shown in Table S4 and Table S5, respectively. These DEGs were visualized in a volcano plot as shown in Figure 4(A). A total of 662 targets were obtained by intersecting the ORG therapeutic targets obtained in both methods.

Figure 4.

The key pathways and related targets for safflower injection (SFI) treatment of obesity-related glomerulopathy (ORG). (A) Volcano plot of differential expressed genes. (B) Venn diagram of SFI potential targets and ORG therapeutic targets. (C) KEGG pathway enrichment bubble plot of intersection targets. (D) GO-BP functional enrichment bubble plot of intersection targets. (E) Chord diagram of target and insulin resistance-related pathways. (F) Protein–protein interaction network. As the color gets darker and the nodes get bigger, the degree gets larger. SFI: safflower injection; KEGG: Kyoto Encyclopedia of Genes and Genomes; GO-BP: gene ontology-biological process; INSR: insulin receptor; eNOS: endothelial nitric oxide synthase.

Based on above results, the 371 SFI potential targets were intersected with 662 ORG therapeutic targets by intersecting and plotting the Venn diagram as shown in Figure 4(B). Consequently, a total of 45 intersected targets were obtained, which might be potential key targets of SFI against ORG, as shown in Table S6.

3.3. The key pathways and related targets for SFI treatment of ORGs

Through KEGG pathway enrichment analysis, the top three pathways were ‘lipid and atherosclerosis’, ‘diabetic cardiomyopathy’, and ‘chagas disease’, which were divided in the ‘Human Diseases’ network of KEGG pathway maps. And then, pathways related to IR were significantly enriched, such as ‘Insulin resistance’ and ‘Insulin signaling pathway’, as shown in Figure 4(C). The results of GO-BP gene function enrichment analysis also showed that insulin pathways related to IR were significantly enriched, such as ‘Positive Regulation of Glycogen Metabolic Process’, ‘Response to Insulin’, and ‘Cellular Response to Insulin Stimulus’, as shown in Figure 4(D). Through combining the results of KEGG and GO pathway enrichment, it was found that the IR-related pathway was significantly enriched, and previous clinical studies have also found that the pathogenesis of ORG is related to IR [21,22]. Therefore, it can be hypothesized that the effect of SFI in the treatment of ORG may be related to the improvement of IR.

From the KEGG and GO-BP pathway enrichment analysis, five pathways associated with IR were selected to identify the regulated targets. As shown in Figure 4(E), INSR was associated with the enriched five pathways related to IR. Additionally, the binding of INSR and insulin triggers downstream signaling cascades, which is a crucial mechanism of the blood glucose regulation and IR development [23]. The results showed that INSR might be the key regulatory targets for SFI to improve ORG-induced IR.

Our previous research indicated that in the model of rats with diabetic kidney disease, the reduced expression of renal eNOS was associated with the elevated blood glucose and the presence of IR [18]. In this study, we continued to pay attention on the role of eNOS in ORG mice. The comprehensive analysis of Figure 4(F) and Table S6 showed that in the PPI network, eNOS had a degree of 14 and ranked 11, and INSR had a degree of 5 and ranked 28. Therefore, we hypothesized that eNOS might also be one of the key regulatory targets for SFI to improve ORG-induced IR.

3.4. SFI improves HFD-induce obesity

As shown in Figure 5(A), the BW of mice in each group was measured every four weeks. At the beginning of the experiments, there was no significant difference in BW between the ORG and NOR groups (p = ns). After 8 weeks of feeding, the average BW of mice in the HFD group was 20% higher than that of mice in the standard diet group, and thus the mice in the HFD group were identified as obese mice [12]. At 12, 16, and 20 weeks, the BWs of the ORG group were significantly higher than those of the NOR group (12th week: p = .011, 16th week: p = .004, 20th week: p = .002). The BWs of the SFI-L and SFI-H groups seemed to have a decreasing trend compared to the ORG group, but there was no significant difference among the three groups (all: p = ns).

Figure 5.

Effects of safflower injection (SFI) on obesity, lipid metabolism disorders. (A) Body weight (BW) at 0th, 4th, 8th, 12th, 16th, and 20th weeks. (B) Lee’s index. (C) Perirenal fat weight. (D) BW ratio (%). (E) Serum TC: serum total cholesterol. (F) Hepatic TG: hepatic triglyceride; NOR: normal control; ORG: obesity-related glomerulopathy; SFI-L: ORG mice treated with 0.33 g·kg⁻¹·d⁻¹ SFI; SFI-H: ORG mice treated with 0.99 g·kg⁻¹·d⁻¹ SFI; data were presented as mean ± SD. *p < .05 ORG vs. NOR; ^#p < .05 vs. ORG.

As shown in Figure 5(B), Lee’s index was significantly higher in the ORG group compared to the NOR group (p < .001). Lee’s index decreased in both the SFI-L and SFI-H groups compared to the ORG group and was statistically different in the SFI-H group compared to the ORG group (p = .034). Moreover, there was a significant difference between SFI-L and SFI-H group (p = .048).

The results of the perirenal fat weight of mice in each group are shown in Figure 5(C). The perirenal fat weight of the ORG group was significantly higher than that of the NOR group (p < .001). Compared with the ORG group, the perirenal fat weight was reduced in SFI-L group and SFI-H group, with statistically significant difference observed between the SFI-H group and the ORG group (p = .014).

The results of the BWR of mice in each group are shown in Figure 5(D). From 8th week, the BWR values of mice in the ORG group were statistically higher than the NOR group (8th week: p = .016, 12th week: p = .011, 16th week: p = .004, 20th week: p = .002). There was no statistically significant difference in the BWR values of mice in both the SFI-L and SFI-H groups compared with the ORG group (all: p = ns).

As shown in Figure 5(E), the serum TC level of the ORG group was significantly higher than that of the NOR group (p < .001). The serum TC levels decreased following the SFI intervention in dose-dependent approach, with a statistically significant difference observed between the SFI-H group and the ORG group (p < .001). Moreover, the improvement in the SFI-H group was significantly better than that in the SFI-L group (p = .002). Of interest, serum TC level did not differ from the NOR group after high-dose SFI intervention (p = ns).

As shown in Figure 5(F), the hepatic TG level of the ORG group was significantly higher than that of the NOR group (p < .001). Compared with the ORG group, the hepatic TG levels of SFI-L group and SFI-H group were significantly decreased (p < .001 and p = .003, respectively). The results showed that the improved TG levels in SFI-H and SFI-L groups were not statistically different compared to the NOR group (p = ns).

3.5. SFI improves renal function injury in ORG mice

As shown in Figure S1, the levels of SCr and BUN did not show a significant change between the NOR and ORG groups (p = ns). In addition, the 24-h urine protein levels at the 12th, 14th, and 20th weeks were higher in the ORG group than those in the NOR group, but there were no statistical differences (p = ns).

In Figure 6(A), kidney tissue sections were stained with PAS. As shown in Figure 6(B), the glomerular diameters of mice in the ORG group were significantly increased compared with that in the NOR group (p < .001). The glomerular diameters of the both SFI-L and SFI-H groups were significantly lower than the ORG group (two: p < .001).

Figure 6.

Effects of safflower injection (SFI) on pathological examination. (A) Kidney tissue sections were stained with PAS (×400). (B) Glomerular diameter. (C) Mesangial relative area (%). (D) Observation of GBMT and podocyte slits width by electron microscopy (×20,000). (E) GBMT: glomerular basement membrane thickness. (F) Podocyte slits width. NOR: normal control; ORG: obesity-related glomerulopathy; SFI-L: ORG mice treated with 0.33 g·kg⁻¹·d⁻¹ safflower injection; SFI-H: ORG mice treated with 0.99 g·kg⁻¹·d⁻¹ safflower injection; data were presented as mean ± SD. *p < .05 ORG vs. NOR; ^#p < .05 vs. ORG. PAS staining and electron microscopy results were obtained from 10 images randomly selected from six individuals in each group.

As shown in Figure 6(C), the relative area of mesangial matrix was significantly larger in the ORG group than the NOR group (p < .001). After SFI intervention, the relative areas of mesangial matrix of the two SFI groups were significantly reduced in a dose-dependent manner (two: p < .001). Additionally, there was no significant difference of the glomerular diameters and the relative areas of mesangial matrix between the two SFI-administrated groups and NOR group (p = ns).

As shown in Figure 6(D), GBMT and podocyte slits width were observed using electron microscopy. The results of GBMT in each group are shown in Figure 6(E). The GBMT in the ORG group was significantly greater than that in the NOR group (p < .001). Compared with the ORG group, the GBMT was significantly lower in the SFI-L and SFI-H groups (two: p < .001).

The results of podocyte slits width in each group are shown in Figure 6(F). In the ORG group, the podocyte slits width was significantly higher than that in the NOR group (p < .001). Compared with the ORG group, the slits width in the SFI-L and SFI-H groups was significantly reduced (p = .003 and p < .001, respectively), and the high dose of SFI had significantly better effect than the low dose (p < .001). The results showed that the improved GBMT and slits width of the SFI-H and SFI-L groups were not statistically different from those of the NOR group (p = ns).

3.6. SFI improves podocyte specific proteins in ORG mice

The expressions of mouse kidney podocyte specific proteins, nephrin and synaptopodin, are shown in Figure 7. After SFI intervention, the expression of nephrin significantly increased compared to the ORG group, in a dose-dependent manner. There was a significant change between SFI-L and SFI-H group (p = .018). Compared with the ORG group, the SFI-L and SFI-H groups increased in the expression of synaptopodin. Additionally, the SFI-H group had a statistically significant difference compared with the ORG group (p < .001), and the beneficial effect of SFI appeared to be dose-dependent (p = .007).

Figure 7.

Effects of safflower injection (SFI) on renal nephrin and synaptopodin. NOR: normal control; ORG: obesity-related glomerulopathy; SFI-L: ORG mice treated with 0.33 g·kg⁻¹·d⁻¹ SFI; SFI-H: ORG mice treated with 0.99 g·kg⁻¹·d⁻¹ SFI; data were presented as mean ± SD. *p < .05 ORG vs. NOR; ^#p < .05 vs. ORG.

According to the results, after SFI intervention, the expression of nephrin in the SFI-H group did not differ from the NOR group (p = ns), and the expression of synaptopodin in the SFI-L and SFI-H groups was not significantly different from that in the NOR group (p = ns).

3.7. SFI improves FBG and HOMA-IR in ORG mice

The results of FBG in each group of mice are shown in Figure 8(A). The FBG in the ORG group was significantly higher (p = .026) compared with that in the NOR group. After the SFI intervention, the FBG levels in the SFI-L and SFI-H groups were significantly decreased compared with the ORG group (p = .003 and p < .001, separately).

Figure 8.

Effects of safflower injection (SFI) on insulin resistance. (A) FBG: fasting blood glucose. (B) FINS: fasting insulin. (C) HOMA-IR: homeostasis model assessment of insulin resistance. NOR: normal control; ORG: obesity-related glomerulopathy; SFI-L: ORG mice treated with 0.33 g·kg⁻¹·d⁻¹ SFI; SFI-H: ORG mice treated with 0.99 g·kg⁻¹·d⁻¹ SFI; data were presented as mean ± SD. *p < .05 ORG vs. NOR; ^#p < .05 vs. ORG.

The results of FINS in mice for each group are shown in Figure 8(B). There was no significant difference in FINS of mice in the ORG group compared with the NOR group. There was a significant difference between the mice in the SFI-L and SFI-H groups (p = .018).

The results of HOMA-IR in mice for each group are shown in Figure 8(C). The HOMA-IR in the ORG group was significantly higher than that in the NOR group (p = .025). After SFI administration, the HOMA-IR of mice in the SFI-L and SFI-H groups was significantly lower compared with that of the ORG group, with a statistically significant difference (p = .029 and p < .001).

Also of interest was that after SFI intervention, FBG and HOMA-IR did not differ from the NOR group (p = ns).

3.8. SFI increases renal INSR and eNOS expression in ORG mice

As shown in Figure 9(A), the expression of INSR in the kidneys of ORG mice was significantly lower compared with that in the NOR group (p < .001). The expression of INSR was significantly increased in the SFI-L and SFI-H groups compared with the ORG group (p = .048 and p = .004, respectively).

Figure 9.

Effects of safflower injection (SFI) on the expression of renal INSR and eNOS. (A) The expression of INSR in the kidney was detected by immunohistochemical staining (×400). (B) The expression of eNOS in the kidney was detected by immunohistochemical staining (×400). INSR: insulin receptor; eNOS: endothelial nitric oxide synthase; NOR: normal control; ORG: obesity-related glomerulopathy; SFI-L: ORG mice treated with 0.33 g·kg⁻¹·d⁻¹ SFI; SFI-H: ORG mice treated with 0.99 g·kg⁻¹·d⁻¹ SFI; data were presented as mean ± SD. *p < .05 ORG vs. NOR; ^#p < .05 vs. ORG. Immunohistochemical results were obtained from 10 fields of vision randomly selected from each group of six individuals.

As shown in Figure 9(B), the expression of eNOS in the kidneys of ORG mice was significantly lower compared with that of the NOR group (p < .001). The expression of eNOS was significantly increased in the SFI-L and SFI-H groups compared with the ORG group (p = .007 and p = .013). Additionally, there was no significant difference among SFI-L, SFI-H, and NOR groups.

UPLC–MS/MS analysis showed that 23 compounds of SFI were absorbed into the bloodstream, of which 17 compounds were absorbed as prototypes and six compounds were absorbed as their own metabolites (Figure S2A-B and Table S7). Among the 23 compounds, there were four flavonoids components, consistent with the present research [24]. On this basis, molecular docking of hydroxysafflor yellow A (HSYA) and theophylline (ThP) with INSR (PDB ID: 4D1P) and eNOS (PDB ID: 1GAG) respectively is shown in Figure S2C-D. The binding energies of HSYA with INSR and eNOS were −8.7 and −7.1, and the binding energies of ThP with INSR and eNOS were −5.1 and −7.1, respectively. These results confirmed that the flavonoid components of SFI could directly bind to INSR and eNOS, to regulate the IR signaling pathway, and improve the ORG.

4. Discussion

In this study, based on system pharmacology, a hypothesis was proposed that SFI may alleviate ORG through IR related pathways by regulating the expressions of INSR and eNOS. The in vivo experiment confirmed that SFI ameliorated glomerular hypertrophy, abnormal GBMT and podocyte slits width and increased nephrin and synaptopodin expressions, the podocyte-specific proteins. The obesity and lipid metabolism-related indicators of ORG, such as Lee’s index, perirenal fat weight, serum TC, and liver TG had also been improved by SFI. Finally, the immunohistochemical experiments validated that SFI could up-regulate the expressions of INSR and eNOS in the kidneys of ORG mice, consistent with the system pharmacology results and the mechanism relationships among IR, INSR and eNOS were verified.

The ORG mouse model established by a long-term HFD had pathological features of obesity, abnormal lipid metabolism, renal pathological changes, and renal podocyte injury. Related studies have found that SY and HSYA can improve leptin resistance by reducing leptin in white adipose tissue, achieving the effect of reducing fat content [25]. HSYA is included in the compositional information of SFI, so the decrease in perirenal fat weight may be due to the action of HSYA and other compounds. The specific mechanism needs further exploration.

Proteinuria is a clinical manifestation of ORG [26]. In our study, the typical pathological changes and the decreased expressions of podocyte specific proteins indicated podocyte injury but there were no significant changes in 24-h urinary protein. Pathological studies have confirmed that less than 30% podocyte damage does not cause glomerulosclerosis, which associated with the severity of CKD [27]. This may explain why our model has not yet developed proteinuria despite podocyte injury. In addition, the low sensitivity of the biochemical assay kits and individual differences among mice may result in an insignificant result in the final 24-h urinary protein level. Other trace proteins in urine, such as microalbumin or more sensitive α-acid glycoprotein, might also be considered for detection [28]. Proteinuria may be observed with prolonged HFD or improved assays. Based on clinical practice, the mice podocyte injury model established in this study could be regarded as an early stage of ORG [29].

After confirming the efficacy of SFI, IR assessment was performed. Presently, methods for assessing IR can be broadly categorized into two types: the methods with precision assessment and the methods with simple substitution index. Hyperinsulinemic–euglycemic clamp (HEC) is the gold standard in the endocrinology for assessing IR [30]. However, HOMA-IR is a simple and effective method to evaluate IR by combining FBG and FINS levels in clinic [31]. We plan to carry out HEC in future studies to allow for more precise work around the topic of IR.

Currently, the proven mechanisms of ORG include IR, glomerular hypertrophy, hemodynamic abnormalities, and lipid metabolism disorders [6]. IR is a systemic disorder that is characterized by a reduced action of insulin despite increased insulin concentrations [32]. On the kidney and vasculature, IR could impair the function of vascular endothelial cells, increase capillary permeability, and interfere with nitric oxide (NO) signaling, lead to obstruction of vasodilatation, reduction of renal blood flow and aggravation of kidney injury renal injury [33]. In this study, pathway enrichment analysis of intersecting targets obtained from SFI targets and ORG targets revealed that ORG is closely related to IR-induced renal injury.

By analyzing the targets of the IR pathway, combined with the PPI results, we speculated that SFI may improve IR by regulating INSR. INSR is a tetramer consisting of two extracellular α-subunits and two intracellular β-subunits [34]. Upon insulin binding, the INSR triggers downstream signaling cascades, which are the most essential mechanism in monitoring blood glucose level [23]. Mutations of INSR are known to cause inherited severe IR syndromes. In addition to classical insulin target tissues (liver, skeletal muscle, and white adipose tissue), insulin also acts on the arterial vasculature and the kidney [33,35]. Mechanism studies have shown that TCM, such as Huangqi decoction, functioned via up-regulating the INSR-related signaling pathway, alleviated renal dysfunction in db/db mice with diabetic nephropathy [36]. It follows that INSR may be a key regulatory target for IR in the development of ORG.

In addition to INSR, eNOS has been found in IR-related pathways and PPI networks. eNOS is mainly located in vascular endothelial cells and activated by a variety of mediators to increase NO content [37]. Our previous studies indicated that in the model of rats with diabetic kidney disease, the reduced expression of renal eNOS was associated with the elevated blood glucose levels and the presence of IR [18]. Through activation of the phosphatidylinositol 3-kinase (PI3K) pathway, insulin activates eNOS and promotes endogenous NO production. In endothelial cells, NO dilates blood vessels, improves blood flow and blood pressure, and facilitates glucolipid metabolism [38]. Accordingly, the expression of renal eNOS in the ORG model and in the SFI-treated group deserves to be further explored.

The immunohistochemistry experiments of renal tissue confirmed the hypothesis proposed by systems pharmacology. The results showed that the expressions of INSR and eNOS in the ORG mice kidneys were significantly decreased, which led to IR in the kidneys of ORG mice [39]. Down-regulation of INSR expression inhibits insulin entry into the cells and fails to decrease blood glucose, resulting in IR. On the other hand, through the PI3K-AKT2 signaling pathway, INSR inhibited eNOS expression which further aggravated IR [40]. According to some related research, both eNOS and INSR were expressed in glomeruli [41,42]. The expression of INSR in glomeruli also could be observed in our research. The low expression of eNOS in this study, we will continue to try immunofluorescence in the subsequent study. On the other hand, we observed high expression of eNOS and INSR in renal tubules, do they play an equally important role in ORG-related kidney injury? We will continue our research in this point.

Under normal physiological conditions, insulin binds and activates INSR, which recruits insulin receptor substrates 1 (IRS-1) to bind to PI3K [43]. After IRS binding with PI3K, phosphatidylinositol (PI) is phosphorylated, resulting in phosphorylation of phosphatidylinositol 3,4-bisphosphate (PIP2) to phosphatidylinositol 3,4,5-trisphosphate (PIP3). PIP3 interacts with phosphoinositide-dependent protein kinase (PDK1), which is AKT regulatory factor, to activate AKT2 [44]. AKT2 performs a series of phosphorylation to promote eNOS synthesis, result in promoting NO production and maintaining vasodilation [45]. Under IR, the expression of INSR and eNOS decreased in podocytes, which are insulin-sensitive cells. Our previous studies have also shown that down-regulation of eNOS expression would impair podocytes [46]. After SFI intervention, the increased expressions of INSR and eNOS suggest that SFI improves the IR of ORG mice. Figure 10 shows the potential mechanisms by which SFI counteracts ORG and protects podocytes by improving the expression of INSR and eNOS.

Figure 10.

Effects of SFI on podocyte INSR and eNOS proteins in the IR signaling pathway. Akt: protein kinase B; eNOS: endothelial nitric oxide synthase; INSR: insulin receptor; IRS-1: insulin receptor substrate-1; NO: nitric oxide; PI3K: phosphoinositide 3-kinase; SFI: safflower injection.

In summary, this study still has some limitations, especially in terms of experimental design. First, the staging of ORG disease was not taken into account. Second, no control experiments were conducted on the active ingredients of SFI. Therefore, the different stages of podocyte injury in ORG and the pharmacological effects and mechanisms of active ingredients in SFI still need to be elucidated.

5. Conclusions

In this study, we presented a system pharmacology method that incorporated UHPLC–MS/MS data from SFI, online TCM databases, drug target databases, and human PPI, along with animal experimental validations. We demonstrated that the crucial regulation pathways and network analysis from shared targets between SFI and ORG for understanding the renal protective mechanism of SFI. We validated the role of SFI in ameliorating ORG podocyte injury by improving IR and modulating INSR and eNOS. In summary, the system pharmacology approach developed herein may facilitate mechanistic discovery of SFI therapeutics and help drive innovation in its clinical application.

Funding Statement

This study was supported by Demonstration Project on Reformation and Quality Development of Public Hospitals (No. SCP-2023-8), the Local Science and Technology Development Funds Projects Guided by Central Government (No. YDZJSX2021C027), Basic Research Program of Shanxi Province (No. 202103021224370), and the Basic Research Program of Shanxi Province (No. 202303021222359).

Author contributions

HL and ZDH designed the study and provided funding. JL and XYW conducted experiments. ZDH, JL, and XYW were responsible for data organization and analysis. ZDH contributed to manuscript writing. XYW was responsible for manuscript revision. HL was responsible for manuscript revision and proofreading. ZDH, RW, and WYZ participated in experiment. All the authors read and approved the submitted manuscript.

Ethical approval

The animal study was reviewed and approved by the Ethics Review Committee for Animal Experimentation of Shanxi Provincial People’s Hospital (Approval No. 2021-306).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data and materials supporting this study are available with the corresponding author upon request.