Hsa_circRNA_101015 and hsa_circRNA_104310 as novel biomarkers of childhood steroid-resistant nephrotic syndrome
Article information
Abstract
Background
Nephrotic syndrome (NS) is the leading cause of glomerular diseases in pediatric patients, among whom corticosteroid responsiveness has been shown to be closely associated with prognosis. Circular RNAs (circRNAs) play crucial roles in various pathophysiological processes and hold significant promise as biomarkers. This study investigated circRNAs for their ability to discriminate patients with steroid-resistant nephrotic syndrome (SRNS) from steroid-sensitive nephrotic syndrome (SSNS), and to explore the pathogenesis underlying NS.
Methods
Microarray analysis was performed to detect circRNA profiles. Total RNA was purified from peripheral blood mononuclear cells obtained from three children, each with SRNS and SSNS. The seven identified candidate circRNAs were validated among 31 SRNS and 30 SSNS patients utilizing real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR). The clinical diagnostic value of the circRNAs was assessed by the receiver operating characteristic curve. Bioinformatic analysis was further performed to understand the intricate biological functions of circRNAs.
Results
Overall, 209 downregulated and 65 upregulated differentially expressed circRNAs were identified in patients with SRNS vs. SSNS. Validation by qRT-PCR revealed that the expression levels of hsa_circRNA_101015 and hsa_circRNA_104310 were considerably lower in the SRNS than the SSNS group. These two circRNAs had area under the curve values of 0.90 and 0.84, respectively, which validated their diagnostic power to discriminate SRNS from SSNS. Further, bioinformatic analysis revealed enrichment of the Wnt signaling pathway.
Conclusion
Hsa_circRNA_101015 and hsa_circRNA_104310 are novel predictive biomarkers for distinguishing SRNS from SSNS, and may participate in the pathogenesis of NS.
Introduction
Nephrotic syndrome (NS), a clinical entity that presents with massive proteinuria, edema, hypoalbuminemia, and hyperlipidemia [1], is the most prevalent glomerulopathy among children, affecting between 14 and 61 of every 10,000 pediatric patients [2]. Although approximately 80% to 85% of children with NS can achieve complete remission following a standard course (4 weeks) of corticosteroids (termed steroid-sensitive nephrotic syndrome [SSNS]), 15% to 20% of children with NS who do not reach remission are accordingly diagnosed with steroid-resistant nephrotic syndrome (SRNS) [3]. Although prognosis is strongly associated with responsiveness to corticosteroids and the kidney histology, the latter association is not always parallel, and routine kidney biopsies are not necessary in children. Thus, early identification to predict the response to corticosteroids could diminish the exposure to the adverse effects of corticosteroids and attenuate disease progression. Therefore, the discovery of noninvasive and specific biomarkers to predict SRNS would greatly benefit this patient group. Additionally, biomarkers participating in molecular pathways could provide novel approaches for NS.
Circular RNAs (circRNAs) represent a group of non-coding RNAs generated through the process of backsplicing [4]. Their covalently closed loop structure makes them more stable in body fluids and tissues [5]. CircRNAs exhibit tissue- and cell-specific expression patterns [6]. Multiple studies have implicated circRNAs in many diseases [7–9]. These circRNAs perform vital biological functions by acting as sponges to bind RNA and microRNAs (miRNAs) [10], as well as protein translation templates [11]. Disease-specific circRNA profiles have also been identified in many kidney diseases, including diabetic kidney disease [12], lupus nephritis [13], acute kidney injury [14], and membranous nephropathy [15]. Further, plasma levels of hsa_circ_0001230 and hsa_circ_0023879 have been identified as predictive indicators in adults diagnosed with focal segmental glomerulosclerosis (FSGS) [16]. However, the function of circRNAs in the peripheral blood mononuclear cells (PBMCs) of children with NS remains unknown.
Our study aimed to assess the utility of circRNAs as predictors of steroid resistance in PBMCs from children with SRNS and SSNS. First, a circRNA microarray analysis was carried out on PBMCs extracted from children diagnosed with NS. Second, some candidate circRNAs were validated among the entire cohort utilizing quantitative reverse transcription polymerase chain reaction (qRT-PCR). To assess the clinical diagnostic efficiency of circRNAs, we plotted receiver operating characteristic (ROC) curves. Finally, bioinformatics analyses were employed to investigate latent molecular mechanisms.
Methods
Subjects and study design
Sixty-seven individuals who underwent treatment at the Department of Pediatric Nephrology, Guangzhou Women and Children’s Medical Center from May 2021 to December 2023 were included in the present study. We obtained blood specimens from three patients, each with SSNS and SRNS patients to perform circRNA microarray analysis. Further, 31 and 30 patients with SRNS and SSNS, respectively, were enrolled for candidate circRNA validation. The following inclusion criteria were applied: diagnosis of NS according to the KDIGO (Kidney Disease: Improving Global Outcomes) 2021 guidelines definition; patients with new-onset NS without steroid or other immunosuppressant treatment; and individuals in the acute phase, presenting with nephrotic-range proteinuria (defined as a 24-hour protein-creatinine ratio ≥2 g/g, urinary protein excretion ≥50 mg/kg/day, or ≥3+ on dipstick for 3 consecutive days) and serum albumin ≤30 g/L. SSNS was defined as reaching complete remission with 4-week standard dosage steroid therapy, while SRNS referred to the failure to reach complete remission after 4-week standard dosage steroid therapy. The exclusion criteria were as follows: secondary NS, such as immunoglobulin A (IgA) nephropathy, IgA vasculitis, lupus nephritis, or kidney damage caused by infections or drugs; monogenic SRNS; coexistence of serious infection; having received steroid or other immunosuppressant treatment; and patients with reduced kidney function characterized by an estimated glomerular filtration rate (eGFR) below 60 mL/min/1.73 m2. All blood samples were obtained from patients in the acute phase of the disease before the steroid therapy. Kidney biopsy was routinely conducted in pediatric SRNS cases.
This study received ethical approval from the Ethics Committee of Guangzhou Women and Children’s Medical Center (No. 201940801). Legally effective informed consent was obtained from the legal guardians of all subjects.
Peripheral blood mononuclear cell isolation and total RNA extraction
For PBMC isolation, 2 mL of venous blood specimens were collected from each individual in the study. PBMCs were isolated from blood specimens using human peripheral blood lymphocyte separation medium. PBMCs were treated with TRIzol Reagent (Invitrogen) to extract total RNA. Subsequently, we used the NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific) to measure RNA concentrations. The qualified RNA was preserved at –80 °C for subsequent experimentation.
Circular RNA microarray analysis
RNA preparation and microarray hybridization were performed according to the standard Arraystar protocol. The extracted RNA was digested using Ribonuclease R (Epicenter) to deplete linear RNA and obtain enriched circRNAs. Subsequently, we transcribed the circRNAs into complementary DNA (cDNA), which was labeled using SYBR Green Ⅰ stain. Subsequently, we hybridized the labeled cDNAs to circular RNA arrays (8 × 15 K; Arraystar). R software limma package was used for the quantile normalization. Empirical Bayes analysis was used to assess differential expression in microarrays. CircRNAs were classified as differentially expressed circRNAs (DEcircRNAs) if their fold changes were above 2.0 with a p-value below 0.05.
Validation of circular RNAs by quantitative reverse transcription polymerase chain reaction
We selected seven downregulated/upregulated circRNAs for validation in 61 subjects (30, SSNS and 31, SRNS), following the order for circRNAs with a lower false discovery rate (FDR) and higher expression. Applied Biosystems real-time PCR system was used for circRNA amplification. β-actin was applied to normalize circRNA expression levels. The relative circRNA expression was calculated using the 2–ΔΔCT method. The sequence-specific primers designed for each circRNA were synthesized by Kangcheng Biotech and are outlined in Supplementary Table 1 (available online).
Bioinformatic analysis
Principal component analysis (PCA) and hierarchical clustering analysis for circRNA expression profiling were conducted using the R software. Gene Ontology (GO) analysis was applied to determine the functions of genes targeted by the DEcircRNAs. Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed to identify the biological pathways related to DEcircRNAs. The TargetScan 8.0 and miRBD 6.0 were applied to predict miRNA targets. DIANA TOOLS (http://diana.imis.athena-innovation.gr/) were also used to predict miRNA-messenger RNA (mRNA) interactions. Cytoscape version 3.10.2 (Cytoscape Consortium) was employed to generate the network of circRNA-miRNA-mRNA.
Statistical analyses
All statistical analyses and visualization were conducted using GraphPad Prism version 9.0. Categorical variables are shown as frequencies (percentages), while continuous variables are presented as means ± standard deviation or medians (minimum, maximum). The chi-square test was used to determine correlations among categorical data. The Student t test was applied for comparisons of normally distributed data between experimental groups. If the variable was not normally distributed, the Mann-Whitney U test was applied. ROC curves were applied to assess the diagnostic efficacy of circRNAs. A two-sided p < 0.05 was defined as statistical significance.
Results
Patient characteristics
Overall, 67 pediatric patients were consecutively enrolled in this investigation over 2 years. The clinical characteristics of the three patients with SRNS and three with SSNS enrolled for circRNA microarray analysis are shown in Table 1. The validation cohort comprised 31 children with SRNS and 30 children with SSNS. Table 2 provided an overview of the clinical features of the validation cohort. FSGS (25 of 31) was the most prevalent histological pattern in the SRNS group. Two patients with SRNS declined to undergo kidney biopsy. No statistically significant differences in sex, age, serum albumin, serum cholesterol, serum creatinine, proteinuria, or eGFR were found between the SRNS and SSNS groups.
Clinical characteristics of patients with SRNS and SSNS who underwent microarray analysis
Clinical characteristics of patients with SRNS and SSNS used for validation
Circular RNA profiling in peripheral blood mononuclear cells
The circRNA microarray profiling identified 274 DEcircRNAs in the PBMCs of patients with SRNS compared with those with SSNS based on the established thresholds. PCA revealed that the samples were clearly segregated between two groups but clustered within their respective groups (Fig. 1A). Hierarchical clustering (Fig. 1B) revealed distinct expression levels between the two groups. A total of 209 downregulated and 65 upregulated circRNAs were detected in the SRNS group compared with the SSNS group (Fig. 1C, D). These circRNAs were located on the autosomal chromosomes, sex chromosome X, and mitochondrial chromosomes (Fig. 1E). CircRNAs originated from diverse genomic regions. In our study, 1.46%, 77.37%, 13.87%, 5.84%, and 1.46% of DEcircRNAs had antisense, exonic, intronic, sense overlapping, and intergenic origins, respectively (Fig. 1F). Tables 3 and 4 illustrate the top 10 up- and downregulated circRNAs, respectively.
Sequencing data of DEcircRNAs in peripheral blood mononuclear cells from patients with SRNS and SSNS.
(A) Principal component analysis of DEcircRNAs. (B) Hierarchical clustering of DEcircRNAs. Red and green represent relatively higher and lower expression, respectively. (C) Volcano plot of DEcircRNAs. (D) Scatter plots of DEcircRNAs. (E) Chromosomal distribution of DEcircRNAs. (F) Genomic origins of DEcircRNAs.
DEcircRNAs, differentially expressed circular RNAs; SRNS, steroid-resistant nephrotic syndrome; SSNS, steroid-sensitive nephrotic syndrome.
Top 10 PBMC circular RNAs (circRNAs) upregulated in the SRNS group
Top 10 PBMC circular RNAs (circRNAs) downregulated in the SRNS group
Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway analysis
We predicted the potential functions of DEcircRNAs through GO enrichment analysis. Ubiquitin-like modifiers activating the enzyme activity, enzyme binding, and biological process regulation of protein import into the nucleus were found to be enriched in GO analysis of the upregulated circRNAs (Fig. 2A). The downregulated circRNAs proved to be significantly abundant in the progress of cellular component organization or biogenesis and protein binding (Fig. 2B). As only a few circRNAs were upregulated, KEGG pathway analysis to assess enrichment could not be conducted. We consequently only used downregulated circRNAs to perform KEGG pathway analysis. Our analysis revealed enrichment of the Wnt signaling pathway (Fig. 2C), which has been shown to be linked to the pathogenesis of NS.
GO and KEGG pathway analysis of DEcircRNAs for their host genes between the SRNS and SSNS groups.
(A) GO analysis of upregulated DEcircRNAs. (B) GO analysis of downregulated DEcircRNAs. (C) KEGG pathway enrichment analysis of DEcircRNAs.
DEcircRNAs, differentially expressed circular RNAs; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; SRNS, steroid-resistant nephrotic syndrome; SSNS, steroid-sensitive nephrotic syndrome.
Validation of circular RNAs by quantitative reverse transcription polymerase chain reaction
CircRNAs were chosen as candidates based on a higher fold change and lower FDR. We validated seven candidate circRNAs by qRT-PCR in 31 patients with SRNS and 30 patients with SSNS. The candidate circRNAs were: hsa_circRNA_103670, hsa_circRNA_406237, hsa_circRNA_104310, hsa_circRNA_102838, hsa_circRNA_001490, hsa_circRNA_0067773, and hsa_circRNA_101015. The results indicated that the expression levels of hsa_circRNA_101015 and hsa_circRNA_104310 were considerably reduced in patients with SRNS compared to those with SSNS (p < 0.001) (Fig. 3G, C). The other five circRNAs did not show any statistical differences between the two groups (Fig. 3A, B, D, E, F).
Real-time quantitative reverse transcription polymerase chain reaction validation of candidate circRNAs in peripheral blood mononuclear cells of patients with SRNS and SSNS.
(A) Hsa_circRNA_103670, (B) hsa_circRNA_406237, (C) hsa_circRNA_104310, (D) hsa_circRNA_102838, (E) hsa_circRNA_001490, (F) hsa_circ_0067773, and (G) hsa_circRNA_101015. ***p < 0.001. Receiver operating characteristic curves assessing the diagnostic value of (H) hsa_circRNA_101015 and (I) hsa_circRNA_104310 to distinguish patients with SRNS from SSNS.
AUC, area under the curve; circRNAs, circular RNAs; SRNS, steroid-resistant nephrotic syndrome; SSNS, steroid-sensitive nephrotic syndrome.
Diagnostic efficacy of circular RNAs
The diagnostic efficiency of hsa_circRNA_101015 and hsa_circRNA_104310 as biomarkers for discriminating between SRNS and SSNS was assessed using ROC curve analysis. Hsa_circRNA_101015 exhibited a great area under the curve (AUC) of 0.90 (95% confidence interval [CI], 0.81–0.99; p < 0.001) (Fig. 3H), while hsa_circRNA_104310 showed an AUC of 0.84 (95% CI, 0.74–0.94; p < 0.001) (Fig. 3I). However, we were unable to detect any linear relationship between the levels of either circRNA and clinical parameters, such as age, serum albumin, serum cholesterol, serum creatinine, proteinuria, or eGFR.
Circular RNA-miRNA-mRNA network
The circRNA-miRNA-mRNA interaction network was established to extensively explore the biological functions of hsa_circRNA_101015 and hsa_circRNA_104310. MiRNA targets were predicted with TargetScan 8.0 (Whitehead Institute for Biomedical Research) and miRBD 6.0 (Wang Lab, Washington University). We first predicted targets through validated interactions and machine learning algorithms, and then integrated the data and visualized the circRNA-miRNA-mRNA interactions with Cytoscape 3.10.2 (Fig. 4). The results revealed that hsa-mir-543, hsa-mir-135a/135b-5p, hsa-mir-640, and hsa-mir-499a-3p were predicted targets of hsa_circRNA_101015, while hsa_circRNA_104310 was predicted to inhibit the expression of hsa-mir-499a-3p, hsa-mir-489-3p, hsa-mir-1271-3p, and hsa-mir-179-3p. Hsa-mir-499a-3p is a common target of hsa_circRNA_101015 and hsa_circRNA_104310. In prior studies, hsa-mir-135a was found to be involved in some kidney disorders.
The circRNA-microRNA-messenger RNA network for hsa_circRNA_101015 and hsa_circRNA_104310.
circRNAs, circular RNAs.
Discussion
The majority of children with NS are sensitive to steroids, and only approximately 20% of affected patients present with SRNS. SRNS cases generally progress to end-stage renal disease within 5 to 10 years, depending on the origin [17]. Up to 33% of SRNS cases are of monogenic origin and contribute to steroid resistance [17]. Other SRNS cases are considered to have an immune-mediated origin associated with immune cell dysfunction [18]. Prior studies have indicated that circRNAs may affect the function of immune cells in various illnesses and serve as potential biomarkers for various diseases [19–21]. In the present study, we used PBMCs as the sample material as they directly reflect alterations in immune cells. Furthermore, patients with SRNS of monogenic causes were excluded from the study so that the subjects investigated would represent typical cases of immune-mediated NS. Moreover, we collected blood samples from patients without infection before steroid treatment to avoid the influence of infection or steroids on PBMCs.
In the present study, we detected 209 downregulated and 65 upregulated DEcircRNAs in the SRNS group compared to the SSNS group. We subsequently implemented GO and KEGG pathway analyses based on these DEcircRNAs, finding associations with protein post-translational modifications, including ubiquitin-like modifier-activating enzyme activity and Small Ubiquitin-like Modifier (SUMO) binding. It has further been reported that SUMOylation and deSUMOylation can affect the development of kidney fibrosis [22]. SUMOs also participate in inflammatory responses, fibrosis, and apoptosis in kidney diseases [23]. KEGG pathway analysis further indicated enrichment of the Wnt signaling pathway. As has been well-established, the Wnt signaling pathway is crucial in the development of many kidney disorders, including NS [24], lupus nephritis [25], and diabetic nephropathy [26].
We further narrowed down the list of candidate circRNAs to seven, and validated their expression levels in 61 patients. Among these, hsa_circRNA_101015 and hsa_circRNA_104310 levels were dramatically decreased in the SRNS group. The diagnostic efficiency of these two circRNAs in discriminating SRNS from SSNS was therefore assessed using ROC curve analysis. Hsa_circRNA_101015 was aligned with LRP6 and named hsa_circ_0000378 or circLRP6. Liu et al. [27] found that hsa_circRNA_101015 was elevated in the plasma of acute pancreatitis and was correlated with the severity of acute pancreatitis. Several prior studies have shown that hsa_circRNA_101015 is upregulated in neoplastic diseases and exerts a pro-tumoral pathogenetic effect [28,29]. Hsa_circRNA_101015 was elevated under high glucose conditions in vitro and could induce mesangial cell injury through the toll-like receptor 4/nuclear factor kappa B (NF-κB) pathway [30]. However, only one study focused on hsa_circRNA_104310; in this study, it exhibited decreased expression in tissues from patients with infantile hemangioma [31]. Several studies have previously been conducted on circRNAs in NS. CircZNF609 was increased in the kidney tissues of patients with lupus nephritis, and participated in the pathogenesis of FSGS in FSGS mice [32]. For example, Ran et al. [16] indicated that elevated plasma concentrations of hsa_circ_0001230 and hsa_circ_0023879 were observed in patients with FSGS compared with healthy controls [16]. To our knowledge, circRNA expression profiles have not yet been demonstrated in pediatric NS.
Exo-derived circRNAs may modulate gene expression by binding to and suppressing miRNA function [10]. Hsa_circRNA_101015 and hsa_circRNA_104310 are both located on the exons of host genes; therefore, the network of circRNA-miRNA-mRNA was constructed to illustrate the regulatory relationships between circRNAs and their targets. Hsa_circRNA_101015 was predicted to target five miRNAs, among which mir-135a/135b-5p has been implicated in the development of many diseases [33]. Further, a number of studies have shown that mir-135a/b participates in multiple signaling pathways, including the Wnt/β-catenin, transforming growth factor-β/SMAD, and NF-κB pathways [33,34]. Yang et al. [35] further demonstrated that mir-135a was elevated in the podocytes of FSGS and could aberrantly activate the Wnt/β-catenin signaling pathway to disrupt podocyte function. Moreover, recent studies have indicated that miR-135a was associated with unresponsiveness to treatment in various tumors [36]. Additionally, a relatively higher level of miR-135a-5p expression was observed in the PBMCs of NS patients compared to healthy controls [37]. miR-135a targeted the serine/threonine kinase (glycogen synthase kinase-3, GSK-3), a crucial enzyme for multiple physiological processes. The inactivation of GSK-3β could thus lead to reduced sensitivity to steroids in patients with chronic obstructive pulmonary disease [38]. However, determining whether glucocorticoid responsiveness in NS is regulated by hsa_circRNA_101015 targeting GSK-3βvia mir-135a-5p requires further experimentation. Hsa-mir-499a-3p, a common target of hsa_circRNA_101015 and hsa_circRNA_104310, is predicted to target NFKBIA, which encodes inhibitor of nuclear factor kappa B alpha (IκBα), a member of the NF-κB inhibitor family. IκBα can inhibit NF-κB signaling to suppress inflammation [39]. NF-κB activation is widely known as a critical regulator of cytokine expression in the inflammatory response in the development of NS [40]. Consequently, patients with SRNS exhibit a more intense inflammatory response and require higher levels of immunosuppression.
The current investigation had certain limitations that should be considered. First, the sample size was insufficient. Therefore, it would be beneficial to expand the validation cohort. Second, we did not validate the candidate biomarkers in urine samples; however, such studies are currently underway. Third, in addition to bioinformatic analysis, functional experiments are warranted to clarify the molecular mechanisms in NS.
In conclusion, the present study represents the first investigation into the differential circRNA expression in PBMCs from children with SRNS and SSNS. Through this analysis, we identified hsa_circRNA_101015 and hsa_circRNA_104310 as promising novel diagnostic biomarkers to distinguish SRNS from SSNS, which could help to predict the corticosteroid response at an early stage. Hsa_circRNA_101015 and hsa_circRNA_104310 have been implicated in the pathogenesis of NS. Further investigation to clarify the precise regulatory relationships between circRNAs and their target genes is warranted.
Supplementary Materials
Supplementary data are available at Kidney Research and Clinical Practice online (https://doi.org/10.23876/j.krcp.24.308).
Notes
Conflicts of interest
All authors have no conflicts of interest to declare.
Funding
This study was supported by grants from the Internal Fund of Guangzhou Women and Children’s Medical Center (NKE-PRE-2019-021).
Data sharing statement
The microarray data in this publication have been deposited in Gene Expression Omnibus (GSE 295499).
Authors’ contributions
Conceptualization: YC, HYD, XDW
Formal analysis, Data curation: JYZ, MT, XQL
Funding acquisition: XYC
Investigation: XYC, WFM
Writing–original draft: XYC
Writing–review & editing: All authors
All authors have read and approved the final manuscript.