Circulating exosome-circRNAs mediated

Full text

Turn on search term navigation

Introduction

Diabetic nephropathy (DN) is one of the microvascular complications of diabetes and the main cause of end-stage renal disease [1]. Progressive tubulointerstitial fibrosis is the key pathological change observed in DN. The pathophysiological mechanism of DN is complex, including immune regulation, inflammation, oxidative stress and other mechanisms. However, effective treatments which are able to reduce or even reverse tubulointerstitial fibrosis are limited. Therefore, it is of utmost urgency to identify the therapeutic targets of DN to delay or reverse its progression.

Exosomes originate in the extracellular vesicles of endosome and are the most typical type of extracellular EVs [2,3]. Exosomes can be isolated from a variety of mammalian cells and are present in almost all types of biological fluids, including blood, urine, prostate fluid and saliva [4,5]. Exosomes contain a variety of signaling molecules, including proteins, DNA, mRNAs, miRNAs, long-chain non-coding RNAs, circRNAs, etc. They are the key mediators of paracrine function, and can fuse with the cell membrane, transfer nucleic acid substances and proteins to target cells, and mediate functions in cells [6,7]. circRNAs are a class of non-coding RNAs (ncRNAs) produced by reverse splicing, which are highly stable in structure and evolutionarily conserved, and are stably enriched in exosomes. Exosomal circular RNAs (circRNAs) have the dual characteristics of exosome delivery function and circRNA transcriptional regulation, and have the potential to become DN biomarkers due to their high stability, expression specificity and easy detection. circRNAs can function as sponges for miRNAs, indirectly regulating mRNA expression, known as competitive endogenous RNA (ceRNA) function [8]. The dysregulation of ceRNAs disrupts the balance of cellular processes and functions. circRNAs have become key regulators of a number of diseases, including cancer, cardiovascular disease, kidney disease etc. [9,10]. Studies have shown that circRNAs have been detected in the blood of patients with DN [11]. However, studies on the role of circulating exosomes-circRNAs in DN are limited.

Fibroblast growth factor (FGF) is one of the most diverse groups of growth factors in vertebrates. FGFs plays a key role in metabolism by regulating the kidneys, liver, brain, gut and fatty tissue. FGF9, also known as glial activator, belongs to the FGF9 subfamily. FGF9 has a high affinity for FGFR3 and FGFR2 [12], and is involved in the growth and development of the lungs, kidneys and bone tissue [13,14]. Among the growth factors known to be expressed in embryonic kidneys, FGF9 has been found to promote the proliferation of nephron progenitor cells in vitro [15]. It has been shown that the knockout of FGF9 and FGF20 alone or in combination leads to the apoptosis of nephron progenitors and subsequent renal stunting [16]. However, the effects of FGF9 on DN remain unclear.

Exosome circRNA is associated with inflammatory mechanism [17]. Besides, type 2 DM [18] and diabetic complications such as diabetic kidney disease [19–25] and diabetic neuropathy [26] are also characterized with chronic inflammatory burden. Hence, studying exosome-circRNAs mediated downregulation of FGF9 through ceRNA mechanism in DN is reasonable. In the present study, a circRNA-miRNA-FGF9 regulatory network was constructed by the high-throughput sequencing of serum circRNAs, miRNA binding site prediction, and the WGCNA co-expression of DN, and the role of FGF9 in the epithelial-mesenchymal transition (EMT) of renal tubular epithelial cells was further verified by in vitro cell experiments. A graphical abstract of the present study was illustrated in (Fig 1).

[Figure omitted. See PDF.]

Materials and methods

Human sample inclusion principle

All human specimens were obtained from the Affiliated Hospital of Binzhou Medical University from April 22, 2021 to March 11, 2023. The patients were provided written informed consent to participate in this study. The six serum samples were from three healthy people and three DN patients. The criteria for inclusion of healthy people were disease-free. All experiments were all pathologically proven with an estimated glomerular filtration rate (eGFR) no less than 30 ml/min/ 1.73 m² (based on the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation) [27]. All experiments involving human subjects were approved by the Ethics Committee of Binzhou Medical University (protocol no. 2020-405).

Exosome isolation

The purified exosomes were obtained from serum by ultrafast centrifugation. Specific process: The sample was transferred to a new centrifuge tube at 37°C, centrifuged at 2000 × g at 4°C for 30 min. Then the supernatant was carefully transferred to a new centrifuge tube at 10000 × g, 4°C, 45 min, and centrifuged again to remove the large vesicles. Filter the supernatant with 0.45 μm filter membrane and collect the filtrate. The filtrate was transferred to a new centrifuge tube, the overspeed rotor was selected, and centrifuged at 100,000 × g at 4°C for 70 min. After removing the supernatant and re-suspension with 10 mL pre-cooling 1 × PBS, the overspeed rotor was selected and centrifuged again for 70 min at 4°C, 100,000 × g. Finally, the supernatant was pre-cooled with 100 μL 1 × PBS, 20 μL electron microscopy, 10 μL particle size, re-suspension, and the remaining exosomes were stored at −80°C for subsequent sequencing [28].

Transmission electron microscopy (TEM)

10μL exosomes were absorbed, deposited on the copper net for 1 min, and the floating liquid was absorbed by filter paper. Then 10 μL uranyl acetate was added to the copper mesh and precipitated for 1 min. The filter paper absorbed the float. After drying at room temperature for a few minutes, electron microscope imaging (HT-7700 Hitachi microscope, Hitachi, Tokyo, Japan).

Nanoparticle tracking analysis (NTA)

Test according to manufacturer’s instructions. After the standard sample had been tested for instrument performance and found to be qualified. The 30 μL exosomes were extracted for particle size analysis.

Exosome- circRNAs sequencing

After RNA extraction, RNA detection, rRNA and linear RNA removal, sequencing library construction, and passing the library inspection, Illumina PE150 sequencing was performed after pooling different libraries according to the effective concentration and target data volume requirements. The transcriptome sequencing Reads were compared with reference genome using Hierarchical Indexing for Spliced Alignment of Transcripts (HISAT2) software, and circRNA was screened [29]. The login number of the data set: PRJNA1184123.

Differential expression of circRNAs

Two circRNA prediction software find_circ [30] and CIRI2 [31] were used to predict and screen circRNAs. Reads Count values of different circRNAs were counted according to the comparison results of each sample. The calculated transcripts per million (TPM) values were used to normalize expression levels, and dimensionality of the data was reduced by principal component analysis (PCA) [32]. According to the set threshold: | log₂(Fold Change, FC) | > 1 and adjusted P- (Padj-) value < 0.05, circRNA differences between samples were screened. The corresponding Fold Change and Padj- value of up-regulated and down-regulated genes were obtained, and the volcano map was used for visualization. According to the TPM values of the different circRNAs, the clustering heat map between the samples was drawn.

CircRNA-miRNA regulatory network analysis

The miRNA-specific binding sites of differentially expressed circRNAs were predicted using miRanda [33] and psRobot [34]. Based on the binding site relationship between miRNAs and circRNAs, Cytoscape software was used to screen the differentially expressed circRNAs and binding information [35]. The visualization of circRNA-miRNA regulatory network was carried out.

Patient datasets

Six GEO datasets related to DN were screened: GSE30528 (GPL571) included 9 DN cases and 13 controls; GSE30529 (GPL571) included 10 DN cases and 22 control cases. The matrix data of GSE30528 and GSE30529 were combined and normalized using the “sva” package in R software for differential gene analysis. GSE30122 (GPL571) included 19 DN cases and 50 control cases. GSE96804 (GPL17586) included 41 DN cases and 20 control cases. As an external validation data set. GSE142025 (GPL20301) included 27 DN cases and 9 controls for weighted correlation network analysis (WGCNA).

Identification of Differential Expression Analysis

PCA was performed to observe the distribution among the groups. The “limma” software package in R was used to identify differential expression genes (DEGs) between normal and DN samples, Padj < 0.05 and |log₂ FC| > 1 gene was DEGs. The “Pheatmap” R software package and “ggplot2” R software package were used to draw the heatmap and volcano map of DEGs respectively.

WGCNA co-expression Analysis

The gene cluster co-expression network was constructed by the “WGCNA” R software package. Select an appropriate soft threshold using the pickSoftThreshold function in the WGCNA software package. Each gene was divided into modules according to the weighted correlation coefficient, and then the correlation analysis between modules and traits and between modules was carried out, and finally epigenes were screened.

Functional enrichment analysis

GO and KEGG pathway analysis was performed using the “clusterProfiler” R software package to evaluate gene-related biological processes (BP), molecular functions (MF), cell components (CC), and gene-related signaling pathways.

Protein ‐protein interaction (PPI) network analysis construction

The STRING database (http://string-db.org/) [36] and Cytoscape [37] were used for protein-protein interaction (PPI) analysis and modular analysis of differential genes. Import DEGs into STRING database and analyze PPI.

ROC evaluation

Use the “pROC” R software package to construct the ROC curve and calculate the area under the curve (AUC). The higher the AUC, the higher the diagnostic value of this gene.

Cell culture and treatment

Normal rat kidney-52 epithelial (NRK-52E) were cultured in 1640 medium containing 10% fetal bovine serum and 1% penicillin-streptomycin mixture in a 5% CO₂ incubator at 37°C. NRK-52E cells were stimulated by the high-concentration glucose (HG, 20–40 mM) and the same concentration of mannitol as osmolarity control. After HG stimulation, FGF9 recombinant protein (50 ng/ml) treatment followed.

CCK-8 tests

In the CCK-8 experiment, NRK-52E cells were planted in 96-well plates (5 × 10³ cells/well), incubated for 48 h. Replaced the medium with CCK8 working solution, and continued to incubate in 37°C for 4 h. Finally, the absorbance at 450 nm was measured with an enzyme-labeled instrument (Instrument models: SpectraMax Mini).

Western blot assays

Cell homogenates were lysed in an ice-cold radioimmunoprecipitation (RIPA, Beyotime, Beijing, China) lysis buffer with 1 mM of phenylmethanesulphonyl fluoride (PMSF, Beyotime, Beijing, China), and then were seperated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene fluoride membranes (PVDF, Millipore, Bedford, MA). And then the membranes were incubated with primary antibodies at 4°C overnight, washed, and incubated with secondary antibodies. Finally, the blots were visualized by an enhanced chemiluminescence detection system (Tanon, Shanghai, China). Primary antibodies included those against FGF9 (1:1000, product number: A6374, ABclonal), α-SMA (1:1000, product number: #19245T, Cell Signaling Technology, Inc.), vimentin (1:4000, product number: 10366-1-AP, ProteinTech Group, Inc.) and β-actin (product number: 60008-1-Ig, ProteinTech Group, Inc.). Second antibodies included HRP-conjugated Affinipure Goat Anti-Mouse IgG (H + L), (1:5000, product number: SA00001-1, ProteinTech Group, Inc.), HRP-conjugated Affinipure Goat Anti-Rabbit IgG (H + L), (1:10000, product number: SA00001-2, ProteinTech Group, Inc.).

Immunofluorescence of α-SMA

NRK-52E cells were seeded with a density of 10⁶cells/ well and incubated at 37°C, 5% CO₂ for 24h. The cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100, and blocked with nonspecific antigens with 3% BSA. The cells were incubated with α-SMA antibody (1:200, product number: #19245T, Cell Signaling Technology, Inc.) overnight at 4°C, followed by Anti-rabbit lgG Fab 2 Alexa Fluor 488 (Cat. #4412, 1:200, Cell Signaling Technology, Inc.) at room tem-perature for a further 2h, and the fluorescent nuclear DAPI stain (Beyotime, Beijing, China) for 5 min. Then fluorescent images were observed under a confocal laser scanning microscope (ZEISS, Germany).

Statistical analysis

Data were expressed as means ± s.e.m. The significance of the differences in mean values between multiple groups were examined by one-way analysis of variance via GraphPad Prism 9.4 software. P < 0.05 was considered statistically significant.

Results

Expression of serum circRNAs in patients with DN

Serum was obtained from 3 healthy individuals and 3 patients with DN, and exosomes were isolated and purified using the ultrafast centrifugation method. Transmission electron microscopy (TEM) was used to observe spherical particles with a diameter of 30–150 nm (Fig 2A). The size of the particles analyzed by nanoparticle tracking analysis (NTA) technology was 81.35nm and the concentration was 1.93x10⁹ particles/ml (Fig 2B). The exosomes were extracted successfully. The purified exosomes were sequenced by circRNAs. The length distribution of circRNAs in each sample was statistically analyzed and plotted, the results showed that the length of most circRNAs was distributed between 149–15149nt (Fig 2C). The type distribution map of the source regions of circRNAs showed circRNAs were derived from the splicing of exons, introns and intergenic regions (Fig 2D). The chromosomal distribution map showed that the circRNA host genes were located on chromosome 1, 10, and 18. (Fig 2E). The results of sample correlation heat map showed that the correlation between DN1 and DN2 was 0.997, indicating that the two samples were highly similar (Fig 2F). The results of PCA diagram showed that the control group (red) and DN (blue) clustered on both sides of PC1 (46.97%), indicating that circRNA expression was significantly different between the two groups (Fig 2G). The differential expression of circRNAs was analyzed, and the volcano map revealed that the expression of differentially expressed circRNAs was significantly decreased in DN (Fig 2H). The clustering heatmap analysis revealed that circRNAs were generally green (low expression) in DN and red (high expression) in control group, indicating that low expression of circRNA might be related to DN (Fig 2I). The top eight circRNAs were listed in descending order according to the | log₂FC | value (Fig 2J).

[Figure omitted. See PDF.]

(A) Representative images of exosome morphology by transmission electron microscopy. (B) The vesicle size distribution of a representative exosome was measured by nanoparticle tracking analysis (NTA). (C) CircRNA length distribution statistics. (D) CircRNA source distribution diagram. (E) CircRNA chromosome distribution. (F) Sample correlation heat map. (G) Principal component analysis (PCA) diagram. (H) Volcanic plot of different circRNAs. (I) Heat map of differential circRNAs expression profiles. (J) The first 8 meaningful differences presented by circRNAs.

Functional enrichment analysis of source genes with differentially expressed circRNAs

In order to further examine the biological function of differentially expressed circRNAs, GO, KEGG and Reactome were used to enrich the source genes of differentially expressed circleRNAs. Cellular component (CC) enrichment analysis revealed that these source genes were related to the N-terminal protein acetyltransferase complex, the site of DNA damage, the site of double-strand break, etc. Molecular function (MF) enrichment analysis revealed that these source genes were related to ATPase activator activity, RNA stem-loop binding, alpha-mannosidase activity, etc. (Fig 3A). Similarly, KEGG pathway analysis revealed enrichment in protein processing in the endoplasmic reticulum (Padj < 0.01), porphyrin and chlorophyll metabolism (Padj < 0.05), N-Glycan biosynthesis (Padj < 0.05), Aminoacyl-tRNA biosynthesis (Padj < 0.05). (Fig 3B). The results of Reactome enrichment analysis revealed that the circRNAs were related to tRNA modification in the nucleus and cytoplasm, and tRNA aminoacylation (Padj < 0.05), etc. (Fig 3C).

[Figure omitted. See PDF.]

(A) GO Bubble Chart enrichment analysis of differentially circRNAs derived genes. (B) KEGG enrichment analysis of differentially derived circRNAs genes. (C) Reactome enrichment analysis of differentially derived circRNAs genes.

Identification of DEGs

The GSE30528 and GSE30529 dataset samples were combined after batch effect removal, including 19 DN tubule samples and 25 control samples. The data box plot after processing revealed that the median of each sample was almost on the same line, which meant the data were standardized successfully (Fig 4A). PCA results showed PC1 (21%) and PC2 (16.9%), which indicated that samples were clustered according to DN and control group (Fig 4B). The volcano map showed 202 upregulated genes and 64 downregulated genes in DN according to the analysis of DEGs (Fig 4C). The heatmaps showed the top 10 upregulated (RARRES1, COL1A2, IGJ, IGLV1–44, LTF, C3, LYZ, ALOX5, CXCL6, MMP7) and downregulated genes (TYRP1, EGF, APOLD1, FOSB, NR4A3, ALB, ETNPPL, APOH, G6PC, CYP27B1) in DN (Fig 4D).

[Figure omitted. See PDF.]

(A) Boxplot of the combined data of the GSE30528 database and the GSE30529 database sample after batch-effect removal. (B) PCA diagram of GSE30528 database and GSE30529 database after data merger. (C) Volcanic plot of differential genes. (D) Heat map of the expression profiles of the first 10 up-regulated and down-regulated genes were taken.

WGCNA identifies eigengenes associated with DN

In order to ensure the scale-free topology and high fit of the co-representation network, the optimal soft threshold was determined to be 12 (Fig 5A). The weighted co-expression network was constructed, and the hierarchical clustering tree of genes was drawn. Genes clustered to the same branch were divided into the same module, and different colors represented different modules (Fig 5B). The correlation and significance between modules and DN were calculated and the heatmap of the correlation between modules and traits was drawn. The results revealed that 15 gene modules with similar co-expression traits were found, among which the pink module, turquoise module and black module had the highest correlation with DN traits and exhibited a negative correlation (Fig 5C). Further analysis of the correlation between gene membership (MM) and gene significance (GS) in the three modules revaled the positive correlation between MM and GS, in the pink module (MM and GS correlation = 0.34; P = 3E-07, Fig 5D), in the turquoise module (MM and GS correlation = 0.51; P < 1e-200, Fig 5E) and the black module (MM and GS correlation = 0.67; P = 1.1e-49, Fig 5F), indicating that the more important genes in the module played a more significant role in DN. A total of 845 eigengenes were screened according to the criteria of |MM| > 0.8 and |GS| > 0.6.

[Figure omitted. See PDF.]

(A) The optimal soft threshold was determined to be 12. (B) Hierarchical clustering tree shows different modules related to DN. (C) Heat map of correlation between characteristic genes of different modules and DN. (D) Scatter plot of correlation between Gene and Module Membership (MM) and Gene Significance (GS) in pink module. (E) Scatter plot of correlation between turquoise module MM and GS. (F) Scatter plot of correlation between black module membership and DN gene traits.

FGF9 is the key gene of DN

A total of 34 common genes were obtained by the intersection of eigengenes and DEGs (Fig 6A). A PPI network of 34 common genes was constructed, and 34 nodes and 67 edges were identified (PPI enrichment P-value < 1.01e-10) (Fig 6B). The thicker the cable, the higher the reliability of the cable. The larger the node, the greater the significance of the protein in the key biological processes of DN. Among the negatively correlated genes, the FGF9 node was the largest; thus, FGF9 was considered to be the key gene of DN. The functional enrichment analysis of 34 common genes was performed. The results of KEGG analysis revealed that they were significantly associated with the PI3K-AKT signaling pathway, Rap1 signaling pathway and AGE-RAGE signaling pathway in diabetic complications (Fig 6C). The results of GO analysis revealed that they were closely related to extracellular matrix structural constituent, collagen-containing extracellular matrix and smad binding (Fig 6D). The Network Analyst 3.0 online network tool was used to predict the miRNAs for FGF9. The mRNA-miRNA co-expression network based on the association between mRNAs and miRNAs was constructed using Cytoscape (Fig 6E).

[Figure omitted. See PDF.]

(A) Venn diagram of intersection of DEGs and characteristic genes. (B) Protein Interaction (PPI) networks of 34 differentially expressed proteins. (C) GO Bubble Chart of differentially expressed genes. (D) KEGG enrichment of differentially expressed genes. (E) Network regulatory diagram of FGF9 target miRNAs predicted by Network Analyst.

Construction of the circRNA/miRNA-FGF9 regulatory network

The ceRNA mechanism of circRNA/miRNA is a key strategy for determining the function of circRNAs. Bioinformatic analysis using the miRanda and psRobot databases identified 676 miRNAs that may interact with differentially expressed circRNAs. The top 10 miRNAs are presented in order from the highest to lowest Max-Score (Fig 7A). The visualization of the circRNA/miRNA interaction network was performed using Cytoscape (Fig 7B). The intersection of 676 miRNAs with 24 FGF9 target miRNAs yielded four common miRNAs, namely, hsa-miR-34a-5p, hsa-miR-766-3p, hsa-miR-147a and hsa-miR-27a-3p (Fig 7C). Their circRNAs were hsa-circ-0006382 and hsa-circ-0019539, respectively, and the circRNA/miRNA interaction network was visualized using Cytoscape (Fig 7D).

[Figure omitted. See PDF.]

(A) The top ten miRNAs were displayed in order of Max-Score from higher to lower for differentially expressed circRNAs and miRNAs binding sites. (B) Regulatory network diagram of differentially expressed circRNAs and their predicted miRNAs. (C) Venn diagram of common miRNAs predicted by Network Analyst and miRanda and psRobot. (D) CeRNA regulatory network diagram of the four common miRNAs and their corresponding serum exosomes circRNAs and FGF9.

FGF9 external database validation and functional analysis

The expression of FGF9 was verified using the DN kidney tissue GSE30122 and GSE96804 datasets. The results revealed that the expression of FGF9 in the kidney tissues of patients with DN in both the GSE30122 and GSE96804 datasets was significantly decreased (Fig 8A and B). ROC curves were further drawn to verify the diagnostic value of FGF9. The AUC values of FGF9 in the GSE30122 and GSE96804 datasets were 0.902 and 0.699, respectively (Fig 8C and D), indicating that FGF9 had a high diagnostic value for DN. The function of FGF9 in renal tubular epithelial cells was further verified using cell experiments in vitro. The results revealed that the expression of FGF9 in NRK-52E cells was significantly decreased under high glucose stimulation in a concentration-dependent manner (Fig 8E and F). In NRK-52E cells, the expression of α-SMA and vimentin was significantly decreased after FGF9 treatment. (Fig 8G–J). The CCK8 test showed that, compared with control, NRK cell viability decreased significantly after 48h of high glucose stimulation, which was significantly improved by FGF9 intervention for 48h (Fig 8K). Consistently, immunofluorescence staining revealed that a-SMA expression level was obviously increased in HG-treated cells compared with normal glucose, a-SMA expression decreased in NRK-52 cells of HG + FGF9 group (Fig 8L), indicating that FGF9 could alleviate EMT in NRK-52E cells.

[Figure omitted. See PDF.]

(A) The expression of FGF9 in the GSE30122 database was significantly decreased. (B) The expression of FGF9 in the GSE96804 database was significantly decreased. (C) AUC value of FGF9 in the GSE30122 database. (D) The AUC value of FGF9 in the GSE96804 database. (E) Representative western blots showed that FGF9 expression of NRK-52E under HG stimulation. (F) Summarized data of western blots showed the FGF9 expression was downregulated under HG stimulation. (G) Representative western blots showed that α-SMA protein expression after FGF9 treatment. (H) Summarized data of western blots showed that α-SMA protein expression was decreased after FGF9 treatment. (I) Representative western blots showed that vimentin protein expression after FGF9 treatment. (J) Summarized data of western blots showed that vimentin protein expression was decreased after FGF9 treatment. (K) Cell viability was measured by CCK8. (L) Representative images of immunofluorescence staining for a-SMA (green) and DAPI (blue) (Scale bars: 75 μm). * p < 0.05, ** p < 0.01, *** p < 0.001.

Discussion

In the present study, circRNA profiles in serum-derived exosomes of DN were revealed by circRNA high-throughput sequencing, which clarified the differential expression of circulating exosome-circRNAs in patients with DN and healthy individuals, and predicted the binding sites of miRNAs to search for downstream mRNA. At the same time, bioinformatics analysis and WGCNA co-expression analysis were used to screen the renal tissue-specific gene, FGF9, in DN, predict the miRNA binding site of FGF9, and the intersection with differential circRNAs to construct the circRNA-miRNA-FGF9 regulatory network. Circulating exosome-circ-0006382/-circ-0019539-FGF9 was found to promote hyperglycemia-induced renal tubular cell fibrosis.

circRNAs are a novel type of RNA which is different from traditional linear RNA. circRNAs have a closed ring structure and exist in the eukaryotic transcriptome in large numbers [38,39]. The dysregulation of circRNA expression under pathological conditions promotes the pathogenesis and progression of various diseases, including kidney disease. In recent years, there has been increasing interest in the role of circRNAs in kidney diseases, including renal cell carcinoma, acute kidney injury, DN and other glomerular diseases [11,40,41]. Studies have shown that exosome hsa_circ_0125310 promotes cell proliferation and fibrosis in DN by sponging miR-422a and targeting the IGF1R/p38 axis [42]. Exosome circ_DLGAP4 promotes the progression of DN by absorbing miR-143 and targeting the ERBB3/NF-KB/MMP-2 axis [43]. In the present study, by comparing the circRNA expression profiles in healthy individuals and patients with DN, it was found that the expression of differentially expressed circRNAs in DN was significantly reduced. The majority of these dysregulated circRNAs originate from exons or introns. It has been reported that both exon circRNAs and intron circRNAs may have potential functions in gene regulation [8]. The biological functions of differentially expressed circRNAs were further studied, and it was found that they were related to endoplasmic reticulum stress and DNA damage. Endoplasmic reticulum stress is closely related to the development of DN [44].

Exosome-circRNAs function as sponges for miRNAs, competitively binding to miRNAs, thereby reducing the inhibitory effect of miRNAs on their target mRNAs, which is termed the ceRNA mechanism [45]. Research has revealed the role of circRNAs in DN. The deletion of circRNA_010383 promotes the expression of TRPC1 protein through the sponge of miRNA-135a, resulting in the formation of DN proteinuria, mesangial cell proliferation and renal fibrosis [46]. Another study found that circRNA_15698 acted as a sponge for miR-185, inhibiting its expression and thereby increasing the expression of transforming growth factor-β1 protein, leading to the accumulation of glomerular extracellular matrix [47]. miRNAs in serum and urine exosomes have been shown to be associated with renal tubule interstitial injury. hsa-miR-34a-5p has been found to regulate oxidative stress damage and promote diabetic kidney damage [48]. hsa-miR-147 has been shown to be upregulated in rats with doxorubicin-induced renal fibrosis [49]. The circulation hsa-27a-3p has been found to be involved in fibrosis in DN [50]. In the present study, four common miRNAs were obtained by intersecting the miRNAs specifically bound to differentially expressed circRNAs with the upstream miRNAs of FGF9. These were hsa-miR-34a-5p, hsa-miR-766-3p, hsa-miR-147a and hsa-miR-27a-3p. Respectively, the sponge circRNAs in serum exosomes were hsa-circ-0006382 and hsa-circ-0019539, based on which, the circRNA-miRNA-FGF9 ceRNA network was constructed.

FGF is one of the most diverse groups of growth factors in vertebrates [51]. The FGF9 subfamily is a member of the FGF families, including FGF9, 16 and 20, which have similar biochemical functions due to their high sequence homology [52]. The FGF9 signaling pathway is involved in the occurrence and development of multiorgan diseases [53,54]. To date, FGF9 has been extensively studied in heart disease. FGF9 signaling is a key mediator of myocardial growth and development, regulating coronary artery development. Exogenous FGF9 protein, or its specific expression in the heart under specific conditions, may improve function following myocardial infarction [55]. A previous study demonstrated that FGF9 attenuated myocardial infarction in diabetic mice by increasing the levels of anti-inflammatory cytokines and M2 macrophage differentiation, thereby reducing adverse remodeling and improving cardiac function [56]. FGF9 is selectively expressed in adults, including the central nervous system and kidneys [57]. However, the role of FGF9 in kidney disease has been rarely reported. In the present study, it was verified that HG stimulation decreased the expression of FGF9 in renal tubular epithelial cells in in vitro experiments. The results also revealed that the expression of FGF9 was downregulated though ceRNA mechanisms by hsa-circ-0006382 and hsa-circ-0019539, which were transported to the kidneys by circulating exosomes, exacerbating the EMT of the renal tubular epithelium.

The ERK signaling pathway played an important role in DN, especially in the progression of renal tubulointerstitial fibrosis. Studies had shown that ERK signaling pathway was directly involved in renal tubule fibrosis by regulating multiple processes such as cell proliferation, apoptosis and matrix deposition [58]. In the presence of high glucose, the ERK pathway of tubular epithelial cells was activated, and the ERK pathway was involved in the epithelial-mesenchymal transformation (EMT) process of tubular cells, which played a key role in the progression of DN [59]. Other studies had shown that FGF9 promotes neurite growth in Huntington’s disease models through the ERK signaling pathway [60]. By binding to FGFR1, FGF9 activated FAK, AKT and ERK/MAPK signaling pathways to promote tumor cell proliferation and epithelial-mesenchymal transformation (EMT), thereby enhancing tumor cell migration and invasion [61]. However, whether FGF9 inhibited renal tubulointerstitial fibrosis through ERK pathways had not been investigated, and future studies will explore the role of these pathways.

Conclusions

In conclusion, in the present study, the circRNA-miRNA-FGF9 mRNA network was constructed by high-throughput sequencing of circulating exosome-circRNAs, miRNA binding site prediction and WGCNA co-expression analysis, and the role of the key gene, FGF9, in the EMT of renal tubular epithelial cells was validated in vitro. The findings presented herein may provide new evidence of the role of the ceRNA mechanisms of the circRNA-miRNA interaction in renal fibrosis in DN. However, the clinical sample in this study was limited. The limited sample size might indeed have had an impact on the generality of the results. In order to eliminate the uncertainty brought by it, we further verified it through in vitro cell experiment and animal experiment. More samples would be collected in future studies to verify the generalizability of our results.

Supporting information

S1 Raw Images.

https://doi.org/10.1371/journal.pone.0326217.s001

S1 File.

https://doi.org/10.1371/journal.pone.0326217.s002

(XLSX)

S2 File. Raw data.

https://doi.org/10.1371/journal.pone.0326217.s003

(TIF)

References

1. 1. Tonneijck L, Muskiet MHA, Smits MM, van Bommel EJ, Heerspink HJL, van Raalte DH, et al. Glomerular hyperfiltration in diabetes: mechanisms, clinical significance, and treatment. J Am Soc Nephrol. 2017;28(4):1023–39. pmid:28143897

* View Article

* PubMed/NCBI

* Google Scholar

2. 2. Salomon C, Guanzon D, Scholz-Romero K, Longo S, Correa P, Illanes SE, et al. Placental exosomes as early biomarker of preeclampsia: potential role of exosomal MicroRNAs across gestation. J Clin Endocrinol Metab. 2017;102(9):3182–94. pmid:28531338

* View Article

* PubMed/NCBI

* Google Scholar

3. 3. Kalluri R, LeBleu VS. The biology, function, and biomedical applications of exosomes. Science. 2020;367(6478):eaau6977. pmid:32029601

* View Article

* PubMed/NCBI

* Google Scholar

4. 4. He X, Kuang G, Wu Y, Ou C. Emerging roles of exosomal miRNAs in diabetes mellitus. Clin Transl Med. 2021;11(6):e468. pmid:34185424

* View Article

* PubMed/NCBI

* Google Scholar

5. 5. Colombo M, Raposo G, Théry C. Biogenesis, secretion, and intercellular interactions of exosomes and other extracellular vesicles. Annu Rev Cell Dev Biol. 2014;30:255–89. pmid:25288114

* View Article

* PubMed/NCBI

* Google Scholar

6. 6. Altevogt P, Bretz NP, Ridinger J, Utikal J, Umansky V. Novel insights into exosome-induced, tumor-associated inflammation and immunomodulation. Semin Cancer Biol. 2014;28:51–7. pmid:24769223

* View Article

* PubMed/NCBI

* Google Scholar

7. 7. Pillay P, Vatish M, Duarte R, Moodley J, Mackraj I. Exosomal microRNA profiling in early and late onset preeclamptic pregnant women reflects pathophysiology. Int J Nanomedicine. 2019;14:5637–57. pmid:31413567

* View Article

* PubMed/NCBI

* Google Scholar

8. 8. Greene J, Baird A-M, Brady L, Lim M, Gray SG, McDermott R, et al. Circular RNAs: biogenesis, function and role in human diseases. Front Mol Biosci. 2017;4:38.

* View Article

* Google Scholar

9. 9. Ju J, Song Y-N, Chen X-Z, Wang T, Liu C-Y, Wang K. circRNA is a potential target for cardiovascular diseases treatment. Mol Cell Biochem. 2022;477(2):417–30. pmid:34780000

* View Article

* PubMed/NCBI

* Google Scholar

10. 10. van Zonneveld AJ, Kölling M, Bijkerk R, Lorenzen JM. Circular RNAs in kidney disease and cancer. Nat Rev Nephrol. 2021;17(12):814–26. pmid:34381199

* View Article

* PubMed/NCBI

* Google Scholar

11. 11. Jin J, Sun H, Shi C, Yang H, Wu Y, Li W, et al. Circular RNA in renal diseases. J Cell Mol Med. 2020;24(12):6523–33. pmid:32333642

* View Article

* PubMed/NCBI

* Google Scholar

12. 12. Santos-Ocampo S, Colvin JS, Chellaiah A, Ornitz DM. Expression and biological activity of mouse fibroblast growth factor-9. J Biol Chem. 1996;271(3):1726–31.

* View Article

* Google Scholar

13. 13. Colvin JS, Feldman B, Nadeau JH, Goldfarb M, Ornitz DM. Genomic organization and embryonic expression of the mouse fibroblast growth factor 9 gene. Dev Dyn. 1999;216(1):72–88.

* View Article

* Google Scholar

14. 14. Garofalo S, Kliger-Spatz M, Cooke JL, Wolstin O, Lunstrum GP, Moshkovitz SM, et al. Skeletal dysplasia and defective chondrocyte differentiation by targeted overexpression of fibroblast growth factor 9 in transgenic mice. J Bone Miner Res. 1999;14(11):1909–15. pmid:10571691

* View Article

* PubMed/NCBI

* Google Scholar

15. 15. Brown AC, Adams D, de Caestecker M, Yang X, Friesel R, Oxburgh L. FGF/EGF signaling regulates the renewal of early nephron progenitors during embryonic development. Development. 2011;138(23):5099–112. pmid:22031548

* View Article

* PubMed/NCBI

* Google Scholar

16. 16. Barak H, Huh S-H, Chen S, Jeanpierre C, Martinovic J, Parisot M, et al. FGF9 and FGF20 maintain the stemness of nephron progenitors in mice and man. Dev Cell. 2012;22(6):1191–207.

* View Article

* Google Scholar

17. 17. Lei T, Zhang Y, Wang X, Liu W, Feng W, Song W. Integrated analysis of the functions and clinical implications of exosome circRNAs in colorectal cancer. Front Immunol. 2022;13:919014. pmid:35924235

* View Article

* PubMed/NCBI

* Google Scholar

18. 18. Aktas G, Duman TT, Atak Tel B. Controlling Nutritional Status (CONUT) score is a novel marker of type 2 diabetes mellitus and diabetic microvascular complications. Postgrad Med. 2024;136(5):496–503. pmid:38921178

* View Article

* PubMed/NCBI

* Google Scholar

19. 19. Aktas G. Association between the prognostic nutritional index and chronic microvascular complications in patients with type 2 diabetes mellitus. J Clin Med. 2023;12(18):5952. pmid:37762893

* View Article

* PubMed/NCBI

* Google Scholar

20. 20. Bilgin S, Kurtkulagi O, Atak Tel BM, Duman TT, Kahveci G, Khalid A, et al. Does C-reactive protein to serum albumin ratio correlate with diabetic nephropathy in patients with type 2 diabetes mellitus? The CARE TIME study. Prim Care Diabetes. 2021;15(6):1071–4. pmid:34497035

* View Article

* PubMed/NCBI

* Google Scholar

21. 21. Kocak MZ, Aktas G, Erkus E, Duman TT, Atak BM, Savli H. Mean platelet volume to lymphocyte ratio as a novel marker for diabetic nephropathy. J Coll Physicians Surg Pak. 2018;28(11):844–7. pmid:30369376

* View Article

* PubMed/NCBI

* Google Scholar

22. 22. Kocak MZ, Aktas G, Duman TT, Atak BM, Kurtkulagi O, Tekce H, et al. Monocyte lymphocyte ratio as a predictor of diabetic kidney injury in type 2 diabetes mellitus; the MADKID Study. J Diabetes Metab Disord. 2020;19(2):997–1002. pmid:33553019

* View Article

* PubMed/NCBI

* Google Scholar

23. 23. Kocak MZ, Aktas G, Atak BM, Duman TT, Yis OM, Erkus E, et al. Is Neuregulin-4 a predictive marker of microvascular complications in type 2 diabetes mellitus? Eur J Clin Invest. 2020;50(3):e13206. pmid:31999832

* View Article

* PubMed/NCBI

* Google Scholar

24. 24. Taslamacioglu Duman T, Ozkul FN, Balci B. Could systemic inflammatory index predict diabetic kidney injury in type 2 diabetes mellitus? Diagnostics (Basel). 2023;13(12):2063. pmid:37370958

* View Article

* PubMed/NCBI

* Google Scholar

25. 25. Aktas G, Yilmaz S, Kantarci DB, Duman TT, Bilgin S, Balci SB, et al. Is serum uric acid-to-HDL cholesterol ratio elevation associated with diabetic kidney injury? Postgrad Med. 2023;135(5):519–23. pmid:37170820

* View Article

* PubMed/NCBI

* Google Scholar

26. 26. Aktas G. Serum C-reactive protein to albumin ratio as a reliable marker of diabetic neuropathy in type 2 diabetes mellitus. Biomol Biomed. 2024;24(5):1380–6. pmid:38635449

* View Article

* PubMed/NCBI

* Google Scholar

27. 27. Levey AS, Inker LA, Coresh J. GFR estimation: from physiology to public health. Am J Kidney Dis. 2014;63(5):820–34. pmid:24485147

* View Article

* PubMed/NCBI

* Google Scholar

28. 28. Théry C, Amigorena S, Raposo G, Clayton A. Isolation and characterization of exosomes from cell culture supernatants and biological fluids. Curr Protoc Cell Biol. 2006;Chapter 3:Unit 3.22. pmid:18228490

* View Article

* PubMed/NCBI

* Google Scholar

29. 29. Kim D, Langmead B, Salzberg SL. HISAT: a fast spliced aligner with low memory requirements. Nat Methods. 2015;12(4):357–60. pmid:25751142

* View Article

* PubMed/NCBI

* Google Scholar

30. 30. Memczak S, Jens M, Elefsinioti A, Torti F, Krueger J, Rybak A, et al. Circular RNAs are a large class of animal RNAs with regulatory potency. Nature. 2013;495(7441):333–8. pmid:23446348

* View Article

* PubMed/NCBI

* Google Scholar

31. 31. Gao Y, Wang J, Zheng Y, Zhang J, Chen S, Zhao F. Comprehensive identification of internal structure and alternative splicing events in circular RNAs. Nat Commun. 2016;7:12060. pmid:27350239

* View Article

* PubMed/NCBI

* Google Scholar

32. 32. Zhou L, Chen J, Li Z, Li X, Hu X, Huang Y, et al. Integrated profiling of microRNAs and mRNAs: microRNAs located on Xq27.3 associate with clear cell renal cell carcinoma. PLoS One. 2010;5(12):e15224. pmid:21253009

* View Article

* PubMed/NCBI

* Google Scholar

33. 33. John B, Enright AJ, Aravin A, Tuschl T, Sander C, Marks DS. Human MicroRNA targets. PLoS Biol. 2004;2(11):e363. pmid:15502875

* View Article

* PubMed/NCBI

* Google Scholar

34. 34. Wu H-J, Ma Y-K, Chen T, Wang M, Wang X-J. PsRobot: a web-based plant small RNA meta-analysis toolbox. Nucleic Acids Res. 2012;40:W22–8. pmid:22693224

* View Article

* PubMed/NCBI

* Google Scholar

35. 35. Kohl M, Wiese S, Warscheid B. Cytoscape: software for visualization and analysis of biological networks. Methods Mol Biol. 2011;696:291–303. pmid:21063955

* View Article

* PubMed/NCBI

* Google Scholar

36. 36. Szklarczyk D, Franceschini A, Wyder S, Forslund K, Heller D, Huerta-Cepas J, et al. STRING v10: protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 2015;43:D447–52. pmid:25352553

* View Article

* PubMed/NCBI

* Google Scholar

37. 37. Doncheva NT, Morris JH, Gorodkin J, Jensen LJ. Cytoscape StringApp: network analysis and visualization of proteomics data. J Proteome Res. 2019;18(2):623–32. pmid:30450911

* View Article

* PubMed/NCBI

* Google Scholar

38. 38. Chen L, Wang C, Sun H, Wang J, Liang Y, Wang Y, et al. The bioinformatics toolbox for circRNA discovery and analysis. Brief Bioinform. 2021;22(2):1706–28. pmid:32103237

* View Article

* PubMed/NCBI

* Google Scholar

39. 39. Kristensen LS, Andersen MS, Stagsted LVW, Ebbesen KK, Hansen TB, Kjems J. The biogenesis, biology and characterization of circular RNAs. Nat Rev Genet. 2019;20(11):675–91.

* View Article

* Google Scholar

40. 40. Wang Y, Zhang Y, Wang P, Fu X, Lin W. Circular RNAs in renal cell carcinoma: implications for tumorigenesis, diagnosis, and therapy. Mol Cancer. 2020;19(1):149. pmid:33054773

* View Article

* PubMed/NCBI

* Google Scholar

41. 41. Chen X-T, Li Z-W, Zhao X, Li M-L, Hou P-F, Chu S-F, et al. Role of circular RNA in kidney-related diseases. Front Pharmacol. 2021;12:615882. pmid:33776764

* View Article

* PubMed/NCBI

* Google Scholar

42. 42. Zhu Y, Zha F, Tang B, Ji T-T, Li X-Y, Feng L, et al. Exosomal hsa_circ_0125310 promotes cell proliferation and fibrosis in diabetic nephropathy via sponging miR-422a and targeting the IGF1R/p38 axis. J Cell Mol Med. 2022;26(1):151–62. pmid:34854210

* View Article

* PubMed/NCBI

* Google Scholar

43. 43. Bai S, Xiong X, Tang B, Ji T, Li X, Qu X, et al. Exosomal circ_DLGAP4 promotes diabetic kidney disease progression by sponging miR-143 and targeting ERBB3/NF-κB/MMP-2 axis. Cell Death Dis. 2020;11(11):1008. pmid:33230102

* View Article

* PubMed/NCBI

* Google Scholar

44. 44. Ni L, Yuan C, Wu X. Endoplasmic reticulum stress in diabetic nephrology: regulation, pathological role, and therapeutic potential. Oxid Med Cell Longev. 2021;2021(1).

* View Article

* Google Scholar

45. 45. Wang Y, Liu J, Ma J, Sun T, Zhou Q, Wang W, et al. Exosomal circRNAs: biogenesis, effect and application in human diseases. Mol Cancer. 2019;18(1):116. pmid:31277663

* View Article

* PubMed/NCBI

* Google Scholar

46. 46. Peng F, Gong W, Li S, Yin B, Zhao C, Liu W, et al. circRNA_010383 acts as a sponge for miR-135a, and its downregulated expression contributes to renal fibrosis in diabetic nephropathy. Diabetes. 2021;70(2):603–15. pmid:33472945

* View Article

* PubMed/NCBI

* Google Scholar

47. 47. Hu W, Han Q, Zhao L, Wang L. Circular RNA circRNA_15698 aggravates the extracellular matrix of diabetic nephropathy mesangial cells via miR-185/TGF-β1. J Cell Physiol. 2019;234(2):1469–76. pmid:30054916

* View Article

* PubMed/NCBI

* Google Scholar

48. 48. Xue M, Li Y, Hu F, Jia Y-J, Zheng Z-J, Wang L, et al. High glucose up-regulates microRNA-34a-5p to aggravate fibrosis by targeting SIRT1 in HK-2 cells. Biochem Biophys Res Commun. 2018;498(1):38–44. pmid:29371016

* View Article

* PubMed/NCBI

* Google Scholar

49. 49. Xu Y, Zhan X. lncRNA KCNQ1OT1 regulated high glucose-induced proliferation, oxidative stress, extracellular matrix accumulation, and inflammation by miR-147a/SOX6 in diabetic nephropathy (DN). Endocr J. 2022;69(5):511–22. pmid:34911869

* View Article

* PubMed/NCBI

* Google Scholar

50. 50. Ji J-L, Shi H-M, Li Z-L, Jin R, Qu G-T, Zheng H, et al. Satellite cell-derived exosome-mediated delivery of microRNA-23a/27a/26a cluster ameliorates the renal tubulointerstitial fibrosis in mouse diabetic nephropathy. Acta Pharmacol Sin. 2023;44(12):2455–68. pmid:37596398

* View Article

* PubMed/NCBI

* Google Scholar

51. 51. Ornitz DM, Itoh N. The fibroblast growth factor signaling pathway. Wiley Interdiscip Rev Dev Biol. 2015;4(3):215–66. pmid:25772309

* View Article

* PubMed/NCBI

* Google Scholar

52. 52. Xie Y, Zinkle A, Chen L, Mohammadi M. Fibroblast growth factor signalling in osteoarthritis and cartilage repair. Nat Rev Rheumatol. 2020;16(10):547–64. pmid:32807927

* View Article

* PubMed/NCBI

* Google Scholar

53. 53. Maddaluno L, Urwyler C, Werner S. Fibroblast growth factors: key players in regeneration and tissue repair. Development. 2017;144(22):4047–60. pmid:29138288

* View Article

* PubMed/NCBI

* Google Scholar

54. 54. Liu Y, Liu Y, Deng J, Li W, Nie X. Fibroblast growth factor in diabetic foot ulcer: progress and therapeutic prospects. Front Endocrinol (Lausanne). 2021;12:744868. pmid:34721299

* View Article

* PubMed/NCBI

* Google Scholar

55. 55. Korf-Klingebiel M, Kempf T, Schlüter K-D, Willenbockel C, Brod T, Heineke J, et al. Conditional transgenic expression of fibroblast growth factor 9 in the adult mouse heart reduces heart failure mortality after myocardial infarction. Circulation. 2011;123(5):504–14. pmid:21262993

* View Article

* PubMed/NCBI

* Google Scholar

56. 56. Singla DK, Singla RD, Abdelli LS, Glass C. Fibroblast growth factor-9 enhances M2 macrophage differentiation and attenuates adverse cardiac remodeling in the infarcted diabetic heart. PLoS One. 2015;10(3):e0120739. pmid:25768089

* View Article

* PubMed/NCBI

* Google Scholar

57. 57. Yin H, Staples SCR, Pickering JG. The fundamentals of fibroblast growth factor 9. Differentiation. 2024;139:100731. pmid:37783652

* View Article

* PubMed/NCBI

* Google Scholar

58. 58. Hu Q, Yang Q, Gao H, Tian J, Che G. Immunoglobulin heavy constant gamma 1 silencing decreases tonicity-responsive enhancer-binding protein expression to alleviate diabetic nephropathy. J Diabetes Investig. 2024;15(5):572–83. pmid:38268239

* View Article

* PubMed/NCBI

* Google Scholar

59. 59. Dong Z, Sun Y, Wei G, Li S, Zhao Z. A nucleoside/nucleobase-rich extract from cordyceps sinensis inhibits the epithelial-mesenchymal transition and protects against renal fibrosis in diabetic nephropathy. Molecules. 2019;24(22):4119. pmid:31739543

* View Article

* PubMed/NCBI

* Google Scholar

60. 60. Yusuf IO, Chen H-M, Cheng P-H, Chang C-Y, Tsai S-J, Chuang J-I, et al. FGF9 induces neurite outgrowth upon ERK signaling in knock-in striatal Huntington’s disease cells. Life Sci. 2021;267:118952. pmid:33383048

* View Article

* PubMed/NCBI

* Google Scholar

61. 61. Chang M-M, Wu S-Z, Yang S-H, Wu C-C, Wang C-Y, Huang B-M. FGF9/FGFR1 promotes cell proliferation, epithelial-mesenchymal transition, M2 macrophage infiltration and liver metastasis of lung cancer. Transl Oncol. 2021;14(11):101208. pmid:34438248

* View Article

* PubMed/NCBI

* Google Scholar

Citation: Yang D, Yin R, Zhang X, Wang X, Pei X, Guo Z, et al. (2025) Circulating exosome-circRNAs mediated downregulation of FGF9 through ceRNA mechanism aggravates renal fibrosis in diabetic nephropathy. PLoS One 20(6): e0326217. https://doi.org/10.1371/journal.pone.0326217

About the Authors:

Donglin Yang

Contributed equally to this work with: Donglin Yang, Rongjiang Yin, Xiaomin Zhang

Roles: Data curation, Software, Writing – original draft

¶These authors contributed equally to this work.

Affiliation: College of Basic Medical, Binzhou Medical University, Yantai, Shandong, P.R. China

Rongjiang Yin

Contributed equally to this work with: Donglin Yang, Rongjiang Yin, Xiaomin Zhang

Roles: Data curation, Software

¶These authors contributed equally to this work.

Affiliation: Department of Thoracic Surgery, Yantai Affiliated Hospital of Binzhou Medical University, Yantai, Shandong, P.R. China

Xiaomin Zhang

Contributed equally to this work with: Donglin Yang, Rongjiang Yin, Xiaomin Zhang

Roles: Data curation, Software

¶These authors contributed equally to this work.

Affiliation: Department of Nephrology, Binzhou Medical University Hospital, Binzhou, Shandong, P.R. China

Xiaohui Wang