Transcriptomic analysis of differential expression and correlation of coding and non-coding RNAs in urine from patients with idiopathic membranous nephropathy

Abstract

Background

The incidence rate of idiopathic membranous nephropathy (IMN) has been increasing, and its pathogenesis is still unclear. Exploring new diagnostic molecular markers and the molecular mechanisms of IMN.

Methods

Fifteen healthy controls and fifteen IMN patients were recruited. Clinical baseline data of all participants, including age, gender, and body mass index (BMI), along with histopathological staging (Ehrenreich-Churg classification) and immunohistochemical staining results for phospholipase A2 receptor (PLA2R), were collected and analyzed. Meanwhile, peripheral blood and urine samples were obtained to detect renal function-related biochemical indicators, including serum creatinine (Scr), estimated glomerular filtration rate (eGFR), serum albumin, 24-hour urinary protein (24 h UPT), serum total cholesterol (TC), anti-PLA2R antibody titer, and serum immunoglobulin G (IgG). Urine samples (n = 10) were also used for high-throughput sequencing analysis.

Results

Serum albumin was significantly downregulated in the IMN group, while 24 h UPT, TC, anti-PLA2R Ab titer, and IgG were upregulated. MicroRNA (miRNA) transcriptomic analysis identified 34 differentially expressed (DE) miRNAs. Hsa-miR-576-3p, hsa-miR-766-5p, and NovelmiRNA-837 exhibited strong diagnostic potential via ROC curve analysis (AUC > 0.8). Sixty-four DE mRNAs were identified in IMN tissues, with enrichment in pathways such as Ribosome and MAPK signaling-fly. ENST00000481739 (RXRA) was upregulated in IMN, correlated with anti-PLA2R Ab titer, and exhibited diagnostic potential (AUC = 0.929). Thirty-five DE lncRNAs were identified in IMN, enriched in autophagy, mitophagy, and apoptosis pathways. TCONS_00180924, TCONS_00176280, TCONS_00208611 correlated significantly with TC, 24-h urinary protein, and anti-PLA2R Ab titer.

Conclusions

This study provides a foundation for developing non-invasive diagnostic tools and targeted therapies for IMN.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13062-026-00864-7.

Introduction

Membranous nephropathy (MN) is an kidney autoimmune disease characterized by subepithelial immune complex deposition in glomeruli with glomerular basement membrane thickening, clinically presenting as insidious-onset proteinuria or nephrotic syndrome [1]. In China, MN is one of the main types of glomerular diseases, with an incidence rate of approximately 28.1%, and its incidence has continued to increase from 2004 to 2014 [2]. Similarly, in northwestern China, the proportion of MN among primary glomerulonephritis reached 17.4%, with a significant increase in its incidence observed from 1989 to 2018 [3]. This rising trend may be attributed to factors such as population aging, relaxed indications for renal biopsy, the prevalence of obesity, and increased environmental pollution [2, 4]. Although 30% of patients with membranous nephropathy (MN) experience spontaneous remission, approximately one-third of them progress to end-stage renal disease (ESRD). As a result, MN patients are at high risk of adverse health outcomes and face a substantial economic burden associated with treatment [5]. These findings highlight the growing burden of MN in China, with multiple potential contributing factors underlying its increasing incidence, underscoring the need for further research to address this evolving clinical challenge.

MN is classified into idiopathic membranous nephropathy (IMN) and secondary membranous nephropathy (SMN) based on etiology and pathogenesis [6]. IMN accounted for about 70% of MN [7]. Since the identification of phospholipase A2 receptor (PLA2R) involvement in MN in 2009 [8], accumulating evidence has indicated that anti-PLA2R antibodies are associated with IMN activity. These antibodies can provide prognostic information regarding disease severity and serve as useful biomarkers for evaluating treatment efficacy [1]. Jiang et al. [6] reported that after treatment, 25.3% of patients experienced a 30% decline in renal function, and 7% progressed to ESRD. However, the pathogenesis of IMN remains unclear. Therefore, more clinical analyses are needed to advance research on the pathogenesis of IMN and provide data support for optimizing prevention and treatment strategies.

MiRNAs are small non-coding RNAs that post-transcriptionally regulate gene expression by targeting mRNA 3’-untranslated regions, playing pivotal roles in cell differentiation, inflammation, and apoptosis [9]. Their stability in bodily fluids and tissue-specific expression make them promising diagnostic and prognostic biomarkers. In IMN, Chen et al. [10] have identified distinct miRNA expression patterns. Analysis of peripheral blood mononuclear cells (PBMCs) from IMN patients revealed 326 differentially expressed miRNAs. Hu et al. [11] used single-cell RNA sequencing and found novel therapeutic targets of hepatitis-B virus-associated MN.

Additionally, Zhou et al. [12] identified miR-195-5p, miR-192-3p, miR-328-5p, and their target genes in PBMCs and urine as potential biomarkers involved in inflammation and apoptosis in MN. Barbagallo et al. [13] further documented differential expression of 10 miRNAs (let-7a-5p, let-7b-5p, let-7c-5p, let-7d-5p, miR-107, miR-129-3p, miR-423-5p, miR-516-3p, miR-532-3p, miR-1275) in renal biopsy tissues from MN patients versus controls. Long non-coding RNA (lncRNAs), another class of non-coding RNAs (> 200 nucleotides), regulate gene expression through epigenetic, transcriptional, or post-transcriptional mechanisms, participating in processes such as cell proliferation, autophagy, and immune response. Szeto et al. [14] reported that urinary lncRNAs have the potential to serve as biomarkers for lupus nephritis. Evidence also supports the involvement of lncRNAs in the development of MN [9]. Additionally, lncRNA XIST regulates the miR-217 signaling pathway to antagonize podocyte apoptosis [15]. However, the molecular mechanisms by which lncRNAs and miRNAs participate in the pathogenesis and progression of idiopathic membranous nephropathy (IMN) remain unclear, and there is no clear theory regarding the etiology of IMN. Further analysis of clinical samples is still required.

Genome-wide transcriptional analysis using high-throughput sequencing has become a commonly used tool in biological process research, widely applied in etiological exploration, subtype analysis, disease diagnosis, and the identification of new therapeutic targets [16]. The rapid advancement of transcriptomics has revolutionized the exploration of disease mechanisms by enabling global profiling of gene expression. Transcriptomic analyses, including those focusing on miRNAs, lncRNAs, and messenger RNAs (mRNAs), have provided critical insights into the molecular pathways underlying MN. In this study, 15 healthy participants who underwent physical examinations and 15 IMN patients who attended outpatient clinics at our hospital were recruited. Clinical baseline data and characteristic information were collected, and urine samples were obtained. A transcriptomic analysis of 10 urine samples from each group was performed using high-throughput sequencing technology to identify differentially expressed miRNAs, mRNAs, and lncRNAs, respectively. Then, common genes associated with clinical biochemical indicators were screened. Receiver operating characteristic (ROC) curve analysis was used to determine their ability to identify IMN. This study aims to provide data for the diagnosis and pathogenesis research of IMN.

Methods

Participant recruitment

From January 2024 to January 2025, patients with idiopathic membranous nephropathy (IMN) were recruited from the the institution. Simultaneously, healthy controls were enrolled from individuals who underwent physical examinations during the same period. Participants were divided into two groups (n = 15, each group): the IMN group and the health control group. The diagnosis of IMN was confirmed by renal biopsy in accordance with clinical guidelines. Health controls were defined as individuals without renal diseases or other systemic disorders and were matched for age and gender where appropriate. All subjects had no prior history of glucocorticoid or other immunosuppressive therapy. Exclusion criteria included participants under 18 years of age, those with severe illnesses, or those who had undergone major surgeries within three months. Clinical baseline data of all participants, including age, gender, and body mass index (BMI), were collected. The histopathological staging of IMN patients was performed based on the Ehrenreich-Churg classification. Immunohistochemical staining for PLA2R was simultaneously performed on the biopsy tissues using Anti-PLA2R antibody (1:100; ab211490, abcam, the United Kingdom) and mouse two-step assay kit (PV-6002, ZSGB-BIO, China). All participants provided written informed consent, and the study was approved by the Institutional Review Board of First Affiliated Hospital of Huzhou Normal University (The First People’ Hospital of Huzhou), Huzhou, Zhejiang, China.

Biochemical index detection

Before receiving any treatment, all subjects provided urine samples and underwent fasting peripheral blood sampling (5 mL) before 10:00 AM. These samples were analyzed using an automated biochemical analyzer to measure various biochemical parameters, including, including serum creatinine (Scr), estimated glomerular filtration rate (eGFR), serum albumin, 24-hour urinary protein, serum total cholesterol (TC), anti-phospholipase A2 receptor antibody (anti-PLA2R Ab) titer, and serum immunoglobulin G (IgG), using an automatic biochemical analyzer (Hitachi 7020, Marunouchi, Japan). The measurement threshold of anti-PLA2R Ab was defined as <14 RU/mL for negative, and >20 RU/mL for positive [17].

High throughput sequencing

Within 2 hours of collection, urine samples (n = 10) were centrifuged at 1500×g for 5 minutes at room temperature. The supernatants were transferred to new sterile centrifuge tubes and were immediately used for RNA isolation with RNeasy Mini Kit (74106, Qiagen, China). For samples that could not be processed immediately, they were stored at -80°C for no longer than 2 weeks, but in this study all samples were processed within 1 h after collection. Isolated RNA was aliquoted and stored at -80°C, and all RNA samples were only thawed before detection, with no repeated freeze-thaw cycles, and detected within one year, then, the concentration and integrity of the extracted total RNA were assessed. Once the RNA passes the quality check, library construction can begin. This typically involves steps such as adapter ligation, cDNA synthesis, and PCR amplification (Supplementary Method 1). Reverse transcription was conducted using M-MLV Reverse Transcriptase (28025021, Thermo Fisher Scientific, USA) with the RT primer from the TruSeq Small RNA Library Prep Kit (RS-200-0012, Illumina, USA), which specifically binds to the 3’ adaptor to synthesize cDNA from miRNA [15]. The reaction was incubated at 42 °C for 60 min, then inactivated at 70℃ for 15 min. Finally, the prepared libraries were subjected to high-throughput sequencing on platforms like Illumina HiSeq 2000 (Illumina, CA, USA). Trimmomatic (v0.30) or cutadapt (v1.9.1) was used to remove technical sequences and get high quality clean data.

Hisat2 (v2.0.1) was used to index reference genome sequence; clean data were aligned to reference genome via software Hisat2 (v2.0.1). For miRNA, miRDeep2 was used to identify microRNA and their expression data. Differential gene expression analysis was performed using the DESeq2 (v1.26.0) package from Bioconductor. Genes with a fold change ≥ 2 and a p-value ≤ 0.05 were defined as differentially expressed (DE) genes.

Data analysis

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed on these target genes to explore their biological functions. Gene Ontology (GO) encompasses three ontologies that describe the molecular function (MF), cellular component (CC), and biological process (BP) of genes, respectively. GO enrichment analysis was performed using GOseq (v1.34.1), while KEGG enrichment analysis was conducted with R software. The threshold for significant enrichment was set at a p/adjusted p (Q)-value <0.05. Correlation analysis was performed using OmicShare tool (Spearma’s correlation method) (https://www.omicshare.com/tools, accessed on 13 August 2025). Difference analysis between two groups. Furthermore, the Mann-Whitney U test, a non-parametric test, was used to compare gene expression levels (mRNA, miRNA, lncRNA) between the IMN and control groups, as data did not follow a normal distribution. The Mann-Whitney U test and ROC curve analysis were performed using SPSS 20.0 (IBM Corp., Armonk, NY, USA), with statistical significance set at p < 0.05.

Results

Statistical analysis results of clinical baseline data and characteristic data

To explore the diagnostic markers of IMN and its potential pathogenesis, this study recruited 15 patients with IMN (IMN group) and 15 healthy participants (control group) in the same period and collected urine for high-throughput sequencing. In control group, gender (male/female: 5/10), age (50.33±10.44), and BMI (22.98±3.03) were no different compared to IMN group, including gender (male/female: 10/5], age (56.13±4.47], and BMI (22.83±2.29) (all p>0.05) (Table 1). It suggests that two groups possessed comparability.

Table 1.

Clinical baseline and characteristics of healthy participants and IMN patients

	Control	IMN	χ2/t/Z	p
Total	15	15
Male/Female (n)	5/10	10/5	χ2 = 0.068	0.143
Age (year) (Mean±SD)	50.33 ± 10.44	56.13 ± 4.47	t=-1.977	0.063
BMI (kg/m2) (Mean±SD)	22.98 ± 3.03	22.83 ± 2.29	t = 0.151	0.881
Histologic stagea (I/II/III) (n)	-	3/11/1	-	-
PLA2R (-/+/++) (n)	-	5/9/1	-	-
Scr (µmol/L) (Mean±SD)	68.76 ± 7.83	75.80 ± 11.52	t=-1.958	0.06
eGFR (mL/min/1.73m2)	94.81 ± 11.28	89.47 ± 14.70	t = 1.116	0.274
Serum albumin (g/L)	41.61 ± 10.76	33.81 ± 6.96	Z=-3.381	<0.001
24-h Urinary protein (g/24 h)	0.00 ± 0.00	460.91 ± 109.04	t=-16.371	<0.001
Serum TC (mmol/L)	4.89 ± 0.93	5.95 ± 1.40	t=-2.449	0.021
Anti-PLA2R Ab titer (RU/mL)	2.66 ± 1.95	56.72 ± 17.54	t=-11.863	<0.001
Serum IgG (g/L)	3.07 ± 1.43	6.27 ± 1.95	t=-5.126	<0.001

^a Staging system according to Ehrenreich and Churg

Abbreviation: Ab: antibody; BMI: body mass index; eGFR: estimated glomerular filtration rate; IgG: immunoglobulin G; PLA2R: phospholipase A2 receptor; Scr: serum creatinine; TC: total cholesterol

Kidney tissue samples were obtained from patients by renal needle biopsy, followed by pathological diagnosis and PLA2R IHC staining. Based on Ehrenreich and Churg, there included 3 phase I, 11 phase II, and 1 phase III. Based on the intensity and distribution of staining, IHC results showed 5 patients were negative, 9 patients were weakly positive, and one patient was strongly positive (Table 1).

Additionally, we collected peripheral blood healthy participants for biochemical tests. There were no significant differences in Scr (68.76±7.83 vs. 75.80±11.52, p = 0.06) and eGFR (94.81±11.28 vs. 89.47±14.70, p = 0.274) levels between healthy group and IMN group (Table 1). In healthy group, mean levels of albumin (41.61±10.76), 24-h urinary protein (0.00±0.00), TC (4.89±0.93), anti-PLA2R Ab titer (2.66±1.95) and IgG (3.07±1.43) were showed in Table 1. In IMN group, the levels of 24-h urinary protein (460.91±109.04), TC (5.95±1.40), anti-PLA2R Ab titer (56.72±17.54), and IgG (6.27±1.95) were significantly higher than those in the control group, while the serum albumin level (33.81±6.96) in IMN was significantly lower than that in the control group (all p<0.05) (Table 1).

IMN-related miRNAs and their targeted mRNA

Subsequently, we randomly selected urine samples from 10 patients for transcriptomics. We collected 34 differentially expressed (DE) miRNAs: 15 downregulated and 19 upregulated (Fig. 1A, B). The heat map showed the transcripts per kilobase million (TPM) levels of all DE microRNAs (Fig. 1C). Miranda (v3.3a) software was used to predict the potential miRNA-targeted genomic cDNA sequences. Subsequently, GO and KEGG enrichment were performed on DE miRNA-targeted genes. Based on an adjusted p (FDR) < 0.05, 630 GO terms in the Molecular Function (MF) category, 263 in the Cellular Component (CC) category, and 1763 in the Biological Process (BP) category can be screened out. Figure 2A shows the top 10 GO terms ranked by the number of genes. GO and KEGG items are sorted from small to large according to the adjusted P value or Q-value. The top 3 items for BP are as follows: adrenal gland development (94 counts), amyloid precursor protein metabolic process (94), response to insulin (98). The top 3 items for CC are as follows: intracellular membrane-bounded organelle (939 counts), cell periphery (92 counts), and specific granule lumen (97 counts). The top 3 items for MF are as follows: sequence-specific DNA binding RNA polymerase II transcription factor activity (883 counts), actin filament binding (378 counts), metal ion binding (2266 counts). Subsequently, KEGG enrichment analysis was performed. Figure 2B shows the top 30 pathways with the smallest Q-values, among which the top 5 are as follows: Fatty acid elongation (130 counts), ABC transporters (140 counts), Hypertrophic cardiomyopathy (HCM) (212 counts), Notch signaling pathway (144 counts), Lysosome (312 counts), and Amyotrophic lateral sclerosis (ALS) (124 counts).

Fig. 1.

Differentially expressed microRNA (miRNA) between healthy and idiopathic membranous nephropathy (IMN) group. Differentially expressed miRNAs were screed using miRNA transcriptomics. (A) Bar chart and (B) volcanic map show that there are 15 downregulated and 19 upregulated in IMN group compared to the control. (C) Heat map shows the expression of 34 differential miRNAs. Red indicates upregulation and blue indicates downregulation

Fig. 2.

Gene ontology (GO) and kyoto encyclopedia of genes and genomes (KEGG) enrichment of differentially expressed miRNA-targeted genes. (A) Histogram shows top 10 items in the biological process (BP), cellular component (CC) and molecular function (MF) subgroups of GO (adjusted P<0.05). (B) Bubble chart shows top 30 pathway in KEGG enrichment (Q-value<0.05). GO and KEGG items are sorted from small to large according to the adjusted P value or Q-value. The histogram and bubble chart were made by SRplot (https://www.bioinformatics.com.cn, last accessed on 14 August 2025), an online platform for data analysis and visualization

Spearman correlation and ROC curve analysis of DE miRNAs and clinical biochemical data

Spearman correlation analysis was performed between DE miRNA and clinical biochemical data DE miRNA. As shown in Supplementary Fig. 1, the correlation heatmap visually demonstrated the correlation direction and magnitude, in which red represented positive correlation and blue represented negative correlation. A total of 15 DE miRNAs showed significant correlation with at least one clinical parameter (p<0.05). Notably, the most robust correlations were observed for hsa-miR-576-3p and hsa-miR-766-5p, which showed strong positive correlations with TC, IgG, 24-hour urinary protein, and anti-PLA2R antibody titer (all p < 0.05). Conversely, hsa-miR-30e-3p displayed significant negative correlations with TC, IgG, 24-hour urinary protein, and anti-PLA2R antibody titer (p < 0.05 to p < 0.001). Several novel miRNAs, including NovelmiRNA-1911, NovelmiRNA-2346, NovelmiRNA-796, and NovelmiRNA-2296 also showed significant positive correlations with Scr (all p < 0.05).

To evaluate the diagnostic potential of these 15 candidate miRNAs for idiopathic membranous nephropathy (IMN), receiver operating characteristic (ROC) curve analysis was performed (Supplementary Fig. 2). The area under the curve (AUC) of each miRNA reflects its overall discriminative ability, with values ranging from 0.5 (no predictive power, represented by the diagonal reference line) to 1.0 (perfect discrimination). Consistent with their high correlation with clinical parameters, several miRNAs, including hsa-miR-576-3p, hsa-miR-766-5p, and NovelmiRNA-837, showed ROC curves positioned clearly above the reference line, indicating promising diagnostic potential. In contrast, other miRNAs, such as NovelmiRNA-2296, hsa-miR-187-3p, and NovelmiRNA-472, exhibited curves close to or even below the reference line, suggesting limited or no discriminative power for IMN.

Based on the AUC values and statistical significance, the four miRNAs with the highest diagnostic potential were selected and their optimal threshold, sensitivity, and 1-specification data are summarized in Table 2. Hsa-miR-576-3p (AUC = 0.903, p = 0.020) and hsa-miR-766-5p (AUC = 1.000, p = 0.004) had a moderate-to-high discriminative ability for IMN diagnosis. Additionally, NovelmiRNA-2587 (AUC = 0.750, p = 0.150) and NovelmiRNA-837 (AUC = 0.833, p = 0.055), their p-values were slightly greater than 0.05 exhibit relatively high AUC values. Hsa-miR-576-3p (Sensitivity, 0.833; 1-Specificity, 0.000), hsa-miR-766-5p (Sensitivity, 0.833; 1-Specificity, 0.000) and NovelmiRNA-837 (Sensitivity, 0.833; 1-Specificity, 0.167) have the potential to serve as a diagnostic biomarker for IMN (Table 2).

Table 2.

ROC curve analysis of candidate miRNAs for IMN

MicroRNA	AUC	95% CI	p	Optimal threshold	Sensitivity	1-Specificity
hsa-miR-766-5p	1.00	1.000–1.000	0.004	63.475	0.833	0.000
hsa-miR-576-3p	0.903	0.700-0.987	0.020	0.845	0.833	0.167
NovelmiRNA-837	0.833	0.570-1.000	0.055	37.335	0.833	0.167
NovelmiRNA-2587	0.750	0.456–0.987	0.150	64.345	0.500	0.000

Abbreviation: ROC, Receiver Operating Characteristic Curve; AUC, Area Under the Curve; miRNA, microRNA; IMN, idiopathic membranous nephropathy

DE mRNAs were screed based on mRNA transcriptomics

Total 64 DE mRNAs were identified through mRNA transcriptome analysis: 28 downregulated and 36 upregulated (Fig. 3A-C). The GO enrichment results are shown in Supplementary Fig. 3A. GO and KEGG items are sorted from small to large according to the adjusted P value or Q-value. The top 3 terms in the BP category were: regulation of transcription from RNA polymerase II promoter (8 counts), cytokine-mediated signaling pathway (3 counts), and cytoplasmic translation (3 counts); top 3 of CC: nucleus (24 counts), cytosol (24 counts), and macromolecular complex (6 counts); top 3 of MF: chaperone binding (4 counts), ubiquitin-protein transferase activity (4 counts), and structural constituent of ribosome (4 counts). The results of KEGG enrichment analysis are shown in Supplementary Fig. 3B. There are only 9 KEGG pathways with an adjusted p-value < 0.05, and the top 3 pathways in terms of the number of included genes are: Ribosome (6 counts), MAPK signaling pathway-fly (4 counts), and non-small cell lung cancer (3 counts).

Fig. 3.

Differentially expressed mRNA between healthy and idiopathic membranous nephropathy (IMN) group. Differentially expressed mRNAs were screed using mRNA transcriptomics. (A) Bar chart and (B) volcanic map show that there are 28 downregulated and 36 upregulated in IMN group compared to the control. (C) Heat map shows the expression of 64 differential mRNAs. Red indicates upregulation and blue indicates downregulation

Correlation analysis and ROC curve of DE mRNA with miRNA and clinical indicators

The intersection analysis of genes between DE mRNAs and miRNA-targeted mRNAs identified 6 common genes (Fig. 4A). The Mann-Whitney U Test was used to analyze the expression levels of these 6 shared genes. Compared to the control group, the level of ENST00000481739 (RXRA) was increased in the IMN group (Z=-2.796, p = 0.005), whereas ENST00000361729 (E2F2) level significantly decreased (Z=-2.163, p = 0.031) (Fig. 4B). They might be involved in the pathogenesis of the disease.

Fig. 4.

Results of Mann-Whitney U test, Spearman correlation and ROC curve analysis of common genes between differentially expressed (DE) mRNA and miRNA-targeted mRNA. (A) Veen analyzed common genes between DE mRNAs and DE miRNA-targeted genes. There were 6 genes in intersection dataset. (B) Mann-Whitney U test was used to analyze difference genes between the control and idiopathic membranous nephropathy (IMN) groups (n = 10). (C) Heat map shows the Spearman correlation analysis results between DE mRNA and the biochemical indicators. (D) ROC curve of 6 common DE genes. Veen and heat map were both created using the OmicStudio tools at https://www.omicstudio.cn/tool (last accessed on 14 August 2025). SPSS 20.0 was used for the Mann-Whitney U test and ROC curve

In addition, we analyzed the correlation between DE mRNA and patients’ biochemical indicators. A total of 24 DE mRNA showed significant correlations with biochemical indicators. The Fig. 4C showed the correlation heatmap. Among them, three of them were also miRNA-targeted genes: ENST00000270460 (vs. anti-PLA2R Ab titer, cor=-0.6095, p = 0.021), ENST00000361729 (vs. IgG, cor=-0.7992, p = 0.0006; vs. anti-PLA2R Ab titer, cor=-0.5396, p = 0.046), and ENST00000481739 (vs. anti-PLA2R Ab titer, cor = 0.5836, p = 0.028).

To further explore their sensitivity as diagnostic indicators for IMM, we performed ROC curve analysis on the 6 genes (Fig. 4D). Three genes (ENST00000304218, ENST00000624585, and ENST00000361729) have identical AUC values (0.286) and overlapping ROC trajectories, thus appearing as a single curve in the plot. Only ENST00000481739 (RXRA) performed well in differentiating healthy individuals from those with IMN (AUC = 0.929 [0.767-1.000], p = 0.007; at the optimal threshold of 2.82, sensitivity = 0.857 and 1-specificity = 0.000). These results suggest that ENST0000481739 (RXRA) has the potential as a diagnostic indicator to aid clinical diagnosis.

DE lncRNAs were screed based on lncRNA transcriptomics

Furthermore, we also conducted LncRNA transcriptomics, identifying 33 downregulated differentially expressed (DE) lncRNAs and 2 upregulated DE lncRNAs, totaling 35 (Fig. 5A-C). Subsequently, we performed GO and KEGG enrichment analyses on lncRNA-targeted mRNAs. With a false discovery rate (FDR) < 0.05, 55 GO terms were obtained, which are shown in Fig. 6A, B. The top 3 terms in the BP category were: cellular response to hypoxia (27), negative regulation of NF-kappaB transcription factor activity (21), and cellular response to epidermal growth factor stimulus (20). For the CC category, the top 3 terms were: nucleus (713), ripoptosome (12), CD95 death-inducing signaling complex (12). In the molecular function (MF) category, the top 3 terms were: RNA binding (247), death receptor binding (13), and cysteine-type endopeptidase activity involved in execution phase of apoptosis (12). With a Q-value (the adjusted p-value) < 0.05, 20 KEGG pathways were identified. The top 3 KEGG pathways were the Autophagy-animal (79), Mitophagy-animal (57), and Apoptosis (53).

Fig. 5.

Differentially expressed long non-coding RNA (lncRNA) between healthy and idiopathic membranous nephropathy (IMN) group. Differentially expressed lncRNAs were screed using lncRNA transcriptomics. (A) Bar chart and (B) volcanic map show that there are 33 downregulated and 2 upregulated in IMN group compared to the control. (C) Heat map shows the expression of 35 differentia lncRNAs. Red indicates upregulation and blue indicates downregulation

Fig. 6.

Gene ontology (GO) and kyoto encyclopedia of genes and genomes (KEGG) enrichment of differentially expressed lncRNA-targeted genes. (A) Histogram shows top 10 items in the biological process (BP), cellular component (CC) and molecular function (MF) subgroups of GO (adjusted P<0.05). (B) Bubble chart shows 20 pathways in KEGG enrichment: all of their Q-value<0.05. GO and KEGG items are sorted from small to large according to the adjusted P value or Q-value. The histogram and bubble chart were made by SRplot (https://www.bioinformatics.com.cn, last accessed on 14 August 2025), an online platform for data analysis and visualization

Correlation analysis of miRNAs and lncRNA

Spearman correlation analysis was then conducted. Figure 7A showed a heatmap of the correlation analysis between DE miRNAs and DE lncRNAs. miRNAs and lncRNA with significant correlations were presented in Table 3. Hsa-miR-138-5p was significantly positively correlated with TCONS_00059442 (cor = 0.56, p = 0.01), TCONS_00101716 (cor = 0.88, p<0.001), TCONS_00176280 (cor = 0.51, p = 0.023), TCONS_00226100 (cor = 0.61, p = 0.004), TCONS_0025985 (cor = 0.68, p = 0.0009), TCONS_00367623 (cor = 0.56, p = 0.01), and TCONS_00486222 (cor = 0.56, p = 0.01). Hsa-miR-187-3p exhibited significant positive correlations with TCONS_00101716 (cor = 0.49, p = 0.02) and TCONS_00226100 (cor = 0.69, p = 0.0008). Hsa-miR-766-5p was significantly negatively correlated with TCONS_00176280 (cor=-0.55, p = 0.03). NovelmiRNA-1062 was significantly positively correlated with TCONS_00180924 (cor = 0.48, p = 0.03). NovelmiRNA-1694 showed significant positive correlations with TCONS_00176280 (cor = 0.57, p = 0.008), TCONS_00180924 (cor = 0.83, p<0.001), TCONS_00258182 (cor = 0.994, p<0.001), and TCONS_00323665 (cor = 0.994, p<0.001). NovelmiRNA-2594 was significantly positively correlated with TCONS_00059442 (cor = 0.45, p = 0.05), TCONS_00176280 (cor = 0.47, p = 0.04), TCONS_00367623 (cor = 0.45, p = 0.046), and TCONS_00486222 (cor = 0.45, p = 0.046).

Fig. 7.

Correlation analysis results of microRNA (miRNA), long non-coding RNA (lncRNA) and the biochemical indicators. (A) A heatmap of the spearman correlation analysis between differentially expressed DE miRNA and DE LncRNA. (B) A heatmap of the spearman correlation analysis between DE miRNA and biochemical indicators. (C) A heatmap of the spearman correlation analysis between lncRNA and biochemical indicators. *p<0.05, **p<0.01, ***p<0.05: Correlation is significant. This correlation analysis was performed using OmicShare tool (Spearman’s correlation method) (https://www.omicshare.com/tools, accessed on 13 August 2025). Column order is determined by hierarchical clustering. (D) The violin plot was used to show expressions of TCONS00176280, hsa-miR-766-5p, and their comment targeted mRNA (ENST00000440650) in Control and idiopathic membranous nephropathy groups. Mann-Whitney U test was used for differences analysis between two groups

Table 3.

LncRNA and miRNA lists that are significantly correlated and have the same target mRNA

Number	LncRNA_ID	microRNA_ID	Gene symbol	cor	p
1	TCONS_00059442	hsa-miR-138-5p	DNAH14	0.5603	0.0102
	TCONS_00059442	hsa-miR-138-5p	PAX7
	TCONS_00059442	hsa-miR-138-5p	LRSAM1
	TCONS_00059442	hsa-miR-138-5p	RAF1
2	TCONS_00101716	hsa-miR-138-5p	H3P37	0.8754	<0.001
3	TCONS_00367623	hsa-miR-138-5p	LRSAM1	0.5603	0.0102
4	TCONS_00176280	hsa-miR-766-5p	PIK3CG	-0.5467	0.0126
5	TCONS_00180924	NovelmiRNA-1062	PRX	0.4810	0.0318
	TCONS_00180924	NovelmiRNA-1062	TNKS
	TCONS_00180924	NovelmiRNA-1062	RRP12
	TCONS_00180924	NovelmiRNA-1062	UBE2O
	TCONS_00180924	NovelmiRNA-1062	SSH1
	TCONS_00180924	NovelmiRNA-1062	SPPL3
	TCONS_00180924	NovelmiRNA-1062	GLRX3
	TCONS_00180924	NovelmiRNA-1062	VGLL4
	TCONS_00180924	NovelmiRNA-1062	KDM5C
	TCONS_00180924	NovelmiRNA-1062	MAP2K3
	TCONS_00180924	NovelmiRNA-1062	TNKS
	TCONS_00180924	NovelmiRNA-1062	CASP16P
	TCONS_00180924	NovelmiRNA-1062	RAB4B
	TCONS_00180924	NovelmiRNA-1062	SMAD6
	TCONS_00180924	NovelmiRNA-1062	ING5
	TCONS_00180924	NovelmiRNA-1062	FBRSL1
	TCONS_00180924	NovelmiRNA-1062	KLHL38
	TCONS_00180924	NovelmiRNA-1062	REEP1
	TCONS_00180924	NovelmiRNA-1694	TAF4	0.8334	<0.001
	TCONS_00180924	NovelmiRNA-1694	PRX
	TCONS_00180924	NovelmiRNA-1694	MARVELD3
	TCONS_00180924	NovelmiRNA-1694	HSP90AB1
	TCONS_00180924	NovelmiRNA-1694	MED12
6	TCONS_00323665	NovelmiRNA-1694	TAF4	0.9945	<0.001
	TCONS_00323665	NovelmiRNA-1694	PRX
	TCONS_00323665	NovelmiRNA-1694	MARVELD3
	TCONS_00323665	NovelmiRNA-1694	HSP90AB1
	TCONS_00323665	NovelmiRNA-1694	MED12
7	TCONS_00059442	NovelmiRNA-2594	SEPTIN3	0.4515	0.0457
	TCONS_00059442	NovelmiRNA-2594	KRIT1
	TCONS_00059442	NovelmiRNA-2594	CYP46A1
	TCONS_00059442	NovelmiRNA-2594	FAM86HP

Correlation analysis to screen IMN-related miRNAs and lncRNAs

We performed Spearman correlation analysis between the miRNAs and lncRNAs listed in Table 3 and clinical biochemical indicators, respectively (Fig. 7B, C; Table 4). A total of 8 miRNAs were correlated with clinical biochemical data (Fig. 7B; Table 4). Hsa-miR-187-3p exhibited significant negative correlations with IgG (cor=-0.5848, p = 0.0280). NovelmiRNA-1062 was significantly positively correlated with albumin (cor = 0.6304, p = 0.0157), while exhibiting significant negative correlations with TC (cor=-0.6821, p = 0.0072) and IgG (cor=-0.5512, p = 0.0411). NovelmiRNA-472 had significant negative correlations with IgG (cor=-0.6029, p = 0.0225). Hsa-miR-766-5p exhibited significant negative correlations with albumin (cor=-0.6126, p = 0.0199), while was significantly positively correlated with TC (cor = 0.5850, p = 0.0280), IgG (cor = 0.7331, p = 0.0029), 24 h Urinary protein (cor = 0.9279, p<0.0001) and anti-PLA2R Ab titer (cor = 0.9177, p<0.0001). NovelmiRNA-837 was significantly positively correlated with IgG (cor = 0.7043, p = 0.0049), 24 h Urinary protein (cor = 0.5601, p = 0.0373) and anti-PLA2R Ab titer (cor = 0.7395, p = 0.0025). NovelmiRNA-441 was significantly positively correlated with IgG (cor = 0.6101, p = 0.0205) and 24 h Urinary protein (cor = 0.5631, p = 0.0360). Additionally, NovelmiRNA-1911 (cor = 0.5481, p = 0.0424) and NovelmiRNA-2346 (cor = 0.5626, p = 0.03624) were only significantly positively correlated with Scr. Meanwhile, TCONS_00180924 (ENSG00000260507) exhibited significant negative correlations with TC (cor=-0.6031, p = 0.0224), TCONS_00176280 was significant negative correlations with 24 h Urinary protein (cor=-0.5957, p = 0.0246) and anti-PLA2R Ab titer (cor=-0.6327, p = 0.0152), and TCONS_00208611 (ENSG00000260440) was significant negative correlations with anti-PLA2R Ab titer (cor=-0.6095, p = 0.0207) (Fig. 7C; Table 4).

Table 4.

LncRNA and miRNA lists that are significantly correlated with clinical biochemical data

Biochemical data	Gene		cor	p_value
TC	miRNA	hsa-miR-766-5p	0.585039	0.027971
TC	miRNA	NovelmiRNA-1062	-0.6821	0.007203
serum albumin	miRNA	hsa-miR-766-5p	-0.61256	0.019867
serum albumin	miRNA	NovelmiRNA-1062	0.630421	0.015653
IgG	miRNA	hsa-miR-187-3p	-0.58475	0.028067
IgG	miRNA	hsa-miR-766-5p	0.733061	0.002856
IgG	miRNA	NovelmiRNA-1062	-0.55119	0.041051
IgG	miRNA	NovelmiRNA-441	0.61012	0.020504
IgG	miRNA	NovelmiRNA-472	-0.6029	0.022477
IgG	miRNA	NovelmiRNA-837	0.704323	0.004922
Anti-PLA2R Ab titer	miRNA	hsa-miR-766-5p	0.917666	3.76E-06
Anti-PLA2R Ab titer	miRNA	NovelmiRNA-837	0.739513	0.002504
24 h Urinary protein	miRNA	hsa-miR-766-5p	0.927886	1.74E-06
24 h Urinary protein	miRNA	NovelmiRNA-441	0.563104	0.036021
24 h Urinary protein	miRNA	NovelmiRNA-837	0.560081	0.037252
Scr	miRNA	NovelmiRNA-1911	0.548144	0.042417
Scr	miRNA	NovelmiRNA-2346	0.562569	0.036237
TC	LncRNA	TCONS_00180924	-0.60312	0.022414
Anti-PLA2R Ab titer	LncRNA	TCONS_00176280	-0.63266	0.015178
Anti-PLA2R Ab titer	LncRNA	TCONS_00208611	-0.60945	0.020681
24 h Urinary protein	LncRNA	TCONS_00176280	-0.59568	0.024591

Furthermore, the correlation heatmaps Fig. 7B and C were generated not only to visualize the relationships between differentially expressed miRNAs/lncRNAs and clinical biochemical parameters, but also to identify underlying correlation patterns through row- and column-wise hierarchical clustering. In Fig. 7B, the clinical parameters clustered into two main groups: renal function markers include serum albumin, eGFR, Scr; disease activity indicators include TC, IgG, 24-hour urinary protein, anti-PLA2R antibody titer, reflecting distinct correlation patterns with miRNAs. In contrast, Fig. 7C showed a different clustering structure, in which serum albumin, TC, and IgG formed one cluster, while eGFR, 24-hour urinary protein, and anti-PLA2R antibody titer formed another, indicating that lncRNAs exhibit different correlation patterns with clinical indices compared to miRNAs.

Notably, hsa-miR-766-5p and NovelmiRNA-837 were identified as miRNAs with potential as biological indicators through the previous ROC curve analysis of miRNAs. Their gene expression levels in urine of participants were shown in Fig. 7D. Results of the Mann-Whitney U test revealed that NovelmiRNA-837 (Z=-2.616, p = 0.009) and hsa-miR-766-5p (Z=-3.069, p = 0.003) were significantly increased in IMN. TCONS_00176280 and hsa-miR-766-5p share the common target gene PIK3CG. Although the TCONS_00176280 and ENST00000440650 (PIK3CG) showed no statistical significance, their expression trends were opposite to that of hsa-miR-766-5p, which is consistent with the regulatory mechanism of the lncRNA-miRNA-mRNA axis.

Discussion

The present study observed distinct alterations in key clinical indicators between IMN patients and healthy controls, which align with the well-recognized pathophysiological features of IMN. Serum albumin was significantly downregulated in the IMN group, while 24-h urinary protein, serum TC, anti-PLA2R Ab titer, and serum IgG were markedly upregulated. These findings are consistent with the core pathological processes of IMN, a disease characterized by autoimmune-mediated glomerular basement membrane damage, leading to increased glomerular permeability and subsequent proteinuria. Hypoalbuminemia in IMN primarily results from excessive loss of albumin through the damaged glomerular filtration barrier, as reflected by elevated 24-h urinary protein levels, an established hallmark of IMN and a critical indicator of disease severity. Additionally, serum TC in IMN is also higher than health control and Patients with type A immunoglobulin nephropathy, which was the clinical feature of IMN [18]. Ye et al. [19] reported that serum albumin decreased in IMN group compared to healthy control group, while TC and 24-h urinary protein increased. Also, IgG subtypes increased significantly in MN patients and related to THSD7A and percentage of plasmablasts [20, 21]. Notably, anti-PLA2R Ab titer, a specific serological marker for primary IMN, was significantly elevated in our IMN cohort, reinforcing its role in disease pathogenesis and diagnosis [22]. Collectively, these clinical indicators not only confirm the diagnostic relevance of traditional markers but also underscore the systemic nature of IMN, involving immune dysregulation, metabolic disturbance, and renal function impairment.

MiRNA in bodily fluids and tissue-specific expression make them promising non-invasive biomarkers for various diseases. Multiple miRNAs have been identified as biomarkers of kidney injury, such like miR-21 in acute kidney injury and miR-30 in diabetic nephropathy [23]. In this study, miRNA transcriptomic analysis identified 34 differentially expressed (DE) miRNAs between IMN and control groups, among which hsa-miR-576-3p, hsa-miR-766-5p, and NovelmiRNA-837 exhibited strong diagnostic potential via ROC curve analysis (AUC > 0.8). Although the roles of these miRNAs in IMN have not been fully elucidated, their dysregulation may reflect pathogenic processes such as podocyte injury or immune activation. MiR-576-3p was involved in the regulation of mouse IFN-induced protein with tetratricopeptide repeats-1 (IFIT1) protein [24]; IFIT1 is associated with lupus nephritis [25]. MiR-766-5p related inflammation and cell apoptosis of mouse neural cells [26]. We found through KEGG enrichment analysis that miR-576-3p is associated with the RNA degradation signaling pathway; miR-766-5p is significantly associated with the cholinergic synapse, mTOR signaling pathway, VEGF signaling pathway, apoptosis, acute myeloid leukemia, type II diabetes mellitus, and neurotrophin signaling pathway. Additionally, novelmiRNA-837 has not been studied yet. Further studies are needed to clarify its target genes and functional relevance. These findings expand the research status of IMN biomarkers and highlight miRNAs as a complementary tool for traditional early diagnosis and disease monitoring.

In our study, 64 DE mRNAs were identified in IMN tissues, with enrichment in pathways such as Ribosome and MAPK signaling-fly. Wan et al. [27] also found that Ribosome signaling pathway was related with IMN. Ribosome protein S6 phosphorylation mediates the hypertrophic growth of kidney proximal tubule cells and related to focal segmental glomerulosclerosis [28]. The MAPK signaling pathway, a conserved regulator of cell survival and inflammation, is known to mediate podocyte apoptosis in glomerular diseases [29], suggesting its potential role in IMN progression.

Notably, 6 DE mRNAs were found to interact with DE miRNAs, among which ENST00000481739 (RXRA) was upregulated in IMN, correlated with anti-PLA2R Ab titer, and exhibited diagnostic potential (AUC = 0.929). RXRA, a member of the nuclear receptor superfamily, regulates lipid metabolism and anti-inflammatory responses [30]. Conversely, ENST00000361729 (E2F2), a transcription factor involved in cell cycle regulation. E2F2 is mostly involved in promoting the malignant development of cancer cells [31]; research evidence also supports its involvement in regulating macrophage energy metabolism [32]. Their research in IMN still has gaps, and this study innovatively proposes their correlation with IMN, which is worth further investigation.

LncRNAs are emerging as crucial regulators of gene expression, often acting as competing endogenous RNAs (ceRNAs) to sequester miRNAs and modulate target mRNA levels [33].Our study identified 35 DE lncRNAs in IMN, enriched in autophagy, mitophagy, and apoptosis pathways. Dysregulation of podocyte autophagy is one of the key mechanisms in chronic kidney disease, involved in causing podocyte damage and urinary protein [34, 35]. It suggests that these lncRNAs may influence IMN progression by regulating cellular stress responses. Three lncRNAs (TCONS_00180924, TCONS_00176280, TCONS_00208611) showed significant correlations with clinical indicators, including TC, 24-h urinary protein, anti-PLA2R Ab titer, indicating their potential as indirect markers of disease severity. TCONS_00176280, in particular, interacted with hsa-miR-766-5p and shared the common target gene PIK3CG (phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit gamma). a key component of the PI3K/Akt pathway. PIK3CG can the metabolism and stemness of acute myeloid leukemia stem cells [36]. We predicted that TCONS_00176280/hsa-miR-766-5p/PIK3CG ceRNA network may be involved in IMN renal dysfunction requiring validation through functional experiments.

Although Zhou et al. [11] previously identified miR‑195‑5p, miR‑192‑3p, and miR‑328‑5p as potential MN biomarkers using urine sediments and PBMCs, our study differs in several key aspects. We focused on urine supernatants, a readily accessible biofluid enriched in extracellular RNAs that may better reflect renal changes. Our miRNA panel (hsa‑miR‑576‑3p, hsa‑miR‑766‑5p, NovelmiRNA‑837) is novel and distinct, with strong diagnostic potential (AUC > 0.8). Unlike Zhou’s miRNA‑mRNA pair analysis, we further integrated lncRNA data to predict a ceRNA network, highlighting potential regulatory axes such as TCONS_00176280/hsa‑miR‑766‑5p/PIK3CG. We also correlated these dysregulated RNAs with key clinical parameters, including anti‑PLA2R titer and 24‑h urinary protein. Collectively, our findings complement Zhou et al.’s work and contribute to a better understanding of IMN’s molecular landscape.

This study has limitations, including a small sample size and reliance on bioinformatic predictions without experimental validation of molecular mechanisms. Future work should validate the diagnostic efficacy of identified miRNAs/mRNAs in larger cohorts and explore the functional roles of the TCONS_00176280/hsa-miR-766-5p/PIK3CG axis in podocyte biology. Additionally, longitudinal studies are needed to assess the utility of these markers in monitoring treatment response.

Conclusions

In conclusion, our multi-omics analysis identifies novel biomarkers and potential regulatory networks in IMN, bridging clinical indicators with molecular mechanisms. These findings provide a foundation for developing non-invasive diagnostic tools and targeted therapies for IMN.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

IMN

Idiopathic membranous nephropathy

BMI

Body mass index

PLA2R

Phospholipase A2 receptor

Scr

Serum creatinine

eGFR

Estimated glomerular filtration rate

24h

UPT 24-hour urinary protein

TC

Total cholesterol

IgG

Immunoglobulin G

miRNA

MicroRNA

DE

Differentially expressed

MN

Membranous nephropathy

ESRD

End-stage renal disease

SMN

Secondary membranous nephropathy

PBMCs

Peripheral blood mononuclear cells

lncRNAs

Long non-coding RNA

mRNAs

Messenger RNAs

Author contributions

Jingwen Li: Conceptualization, Investigation, Project administration, Writing-original draft. Zhelun Zhou: Formal analysis, Validation, Writing-original draft. Rong Wu: Data curation, Visualization. Qi Chen: Conceptualization, Funding acquisition, Project administration, Supervision, Writing-review & editing.

Funding

This work was supported by Huzhou Municipal Science and Technology Bureau [grant number 2021GY20].

Data availability

The data are available in the Materials and Methods, Results, and/or Supplemental Material of this article.

Declarations

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Institutional Review Board of First Affiliated Hospital of Huzhou Normal University (The First People’ Hospital of Huzhou), Huzhou, Zhejiang, China. Informed consent was obtained from all individual participants included in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.