The microarray data were downloaded from the gene expression omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo) using the keywords “IgA nephropathy,” “Homo sapiens” and “Expression profiling by array” were included in the next round of screening. The gene expression microarray datasets GSE104948 (9), GSE93798 (10) GSE37463 (11) were selected and downloaded for DEG analysis. Probes were matched to the gene symbols using the annotation files provided by the manufacturer. DEGs were identified by sva/limma R package (12) (version 4.1.0) after consolidation and batch normalization, with threshold |log₂FC| >1 and adjusted p < 0.05. Heatmap (https://CRAN.R-project.org/package=pheatmap) and volcano plot were subsequently performed using R software to visualize the DEGs.

GO enrichment and KEGG pathway analyses were performed to explore the potential biological processes and utilities, cellular components, and molecular functions of DEGs using colorspace (13) and stringi (https://stringi.gagolewski.com/) R packages. Significantly relevant signal pathways were identified via clusterProfiler package (14) in R software with p <0.05 and q < 0.05. The STRING database (https://string-db.org) (15) was used to explore potential protein-protein interactions (PPI) interactions between DEGs, and a network interaction matrix was built. An interaction with a combined score > 0.7 was set as the cut-off value. Densely connected clusters in the PPI network were visualized by R.

To validate the expression of hub genes in other glomerulonephritis types and IgAN, we use GSE116626 (16) as an independent validate cohort, comprising 52 IgAN patients, 22 non-IgAN glomerulonephrites (non-IgAN GN), and 7 kidney living donors (LD). The relative mRNA expression levels of hub genes in each patient were extracted from the raw data, which were then analyzed by Graphpad Prism 6. Statistical differences were analyzed by Student's t-test, and p < 0.05 was considered statistically significant. Analyses were conducted with GraphPad Prism 6.0 statistical software.

Formaldehyde-fixed and paraffin-embedded tissue sections (4 mm thick) were used for immunohistochemistry as previously described (17). After dehydration, slides in citrate solution were subjected to a microwave antigen retrieval process. Primary antibodies were employed against the integrin beta 2 (ITGB2) (Santa Cruz Biotechnology, sc-8420), fc fragment of IgE receptor Ig (FCER1G) (Santa Cruz Biotechnology, sc-390222), and colony-stimulating factor 1 receptor (CSF1R) (Santa Cruz Biotechnology, sc-46662), periodic acid-Schiff (PAS) was used instead of primary antibodies as negative controls for all staining. Finally, horseradish peroxidase (HRP)-conjugated polymer (Abcam, USA) was used for the visualized detection under light microscopy (Nikon Tokyo, Japan). OD values were analyzed by ImageJ software (National Institute of Health, USA). The use of renal specimens of IgAN patients was approved by the Ethics Committee of the Second Xiangya Hospital of Central South University (2019SNK1222000) according to the Declaration of Helsinki.

The CMap (https://portals.broadinstitute.org/cmap) is an open resource that links disease, genes, and drugs by similar or opposite gene expression profiles (18), and was used to explore potential drugs targeting IgAN. Upregulation and the downregulation DEGs were converted to probe sets in the chip type to query the CMap. A list of CMap instances predicted to reverse IgAN DEGs was obtained, with connectivity score and p-value.

The human mesangial cells (HMCs) collected from (CBR130735, Cellbio, China) were cultivated in Dulbecco's Modified Eagle's Medium (DMEM)/F-12 medium (Gibco, USA) supplemented with 10% fetal bovine serum (FBS) under 5% CO₂ at 37 °C. Monomeric human IgA1 (Abcam, ab91020) was heated and aggregated at 65°C for 150 min on a dry plate heater to obtain aggregated IgA1 (aIgA1) as previously described (19). After cooling at room temperature, the proteins were centrifuged at 11,000g for 5 min to remove insoluble precipitant. The supernatants were then used as heat-aggregated IgA. Mesangial cells were treated with or without IgA aggregates at a final concentration of 25 μg/ml for 24 h. The dose of IgA aggregates was selected based on our previous study in mesangial cells (17). Tetrandrine (10 mM in 1 ml dimethyl sulfoxide) was purchased from ApexBio Technology (Houston, USA). Cells were incubated with aIgA1 alone or with a combination of tetrandrine (0–10 μM) concentrations for 24 h.

Cell proliferation was accessed by cell counting kit-8 (CCK8) assay (Dojindo, Japan) following the manufacturer's recommended procedure. 10 μl of CCK8 was added to treated cells in 96-wells and incubated at 37 °C. The absorbance was detected at 450 nm using spectrophotometer.

Cell cycle was measured by cell cycle and apoptosis kit (Wellbiology, China), and the treated cells were fixed overnight with 70% cold ethanol at 4°C, then stained with a mixture of propidium iodide and RNase A at 37°C for 30 min. The distribution of the cell cycle phase was measured by FlowJo software.

This study was conducted according to the flow chart shown in Figure 1. The microarray data archives for glomeruli of IgAN were searched and downloaded from the Gene Expression Omnibus (GEO) (https://www.ncbi.nlm.nih.gov/geo/) using the “IgA nephropathy,” with “Homo sapiens” and “Expression profiling by array” filters in the next round of screening. Three relevant microarray expression profiles (GSE104948, GSE93798, GSE37463) were selected between IgAN and normal glomeruli compartment. A total of 144 human glomerular compartment tissue samples were screened, of which 74 were IgAN patients and 70 were healthy donors. Details of the GEO datasets in this study were shown in Table 1. A total of 134 DEGs were identified after consolidation and batch normalization (Figure 2A; Supplementary Table S1), including 54 upregulated and 80 downregulated genes. The volcano plot (Figure 2B) showed the distribution of DEGs between IgAN and normal controls.

To identify comprehensive information on biological processes, cellular components, and molecular functions gene function, GO and KEGG pathway analyses of DEGs were performed. The top twelve GO terms with the highest degree of enrichment were shown in Figure 3A. GO terms mainly concentrated on anion transmembrane transporter activity, extracellular matrix structural constituent, and organic anion transmembrane transporter activity. Figure 3B presented seven enriched KEGG pathways of the DEGs, protein digestion and absorption, pertussis, complement, and coagulation cascades were identified as the top three significantly relevant signal pathways.

The STRING database (https://string-db.org) was used to explore potential interactions between DEGs. High confidence score>0.7 serves as the cut-off point for the construction of protein-protein interaction (PPI) network, visualization of the PPI network was shown in Figure 4A. A total of 91 nodes and 194 interactions were calculated by R software, the top 30 candidate hub genes which may play a central role in this network were identified in Figure 4B. Five genes with the highest score were obtained as ITGB2, FCER1G, complement C3a receptor 1 (C3AR1), CSF1R, transmembrane immune signaling adaptor TYROBP (TYROBP).

To verify the reliability of the data and the link between key genes and IgAN, an external dataset GSE116626 (16), which was uploaded by Cox et al. in 2018 and updated in 2020, was used to validate the top 5 hub genes. ITGB2, FCER1G, C3AR1, CSF1R, TYROBP expression profiles were screened in archival FFPE biopsy samples of 52 IgAN patients, 22 non-IgAN glomerulonephrites (non-IgAN GN), and 7 kidney living donors (LD). The transcriptional levels of ITGB2, FCER1G, and CSF1R were significant higher in IgAN compared to either healthy donors or non-IgAN (Figure 5A), while C3AR1 and TYROBP transcriptional levels were not significant (data not shown). We further investigated the ITGB2, FCER1G, and CSF1R transcriptional and protein levels in IgAN with different composite renal lesions, patients were stratified into 4 groups based on the E, C,and T (but not M or S) scores: minimal (E0, C0, T0); active (E1 and/or C1, C2, T0); chronic (T1 or T2, E0, C0);mixed group composed of active and chronic lesions (E1 and/or C1, C2, T1 or T2) as previously described (16). Compared with minimal lesions or active lesions, these three genes mRNA expression were significantly increased in the mixed lesions group. Also, FCER1G mRNA level was significantly upregulated in the mixed lesion group compared to chronic lesions (Figure 5B). Moreover, biopsy specimens of minimal change disease (MCD) and IgAN were collected in our hospital. Through immunohistochemistry, ITGB2, FCER1G, and CSF1R were significantly elevated in IgAN patients with different renal pathology compared with MCD (Figure 6).

To screen potential therapeutic small molecular compounds for IgAN, we performed CMap (http://portals.broadinstitute.org/cmap/) analysis and obtained a list of drug candidates predicted to reverse altered expression of DEGs (Figure 7A). Tetrandrine, an antirheumatic and analgesic agent, was the top-ranked candidate with the highest negative connectivity score (Figure 7B) and was selected for subsequent validation experiments.

Considering that mesangial cell proliferation is a prominent pathological feature of IgAN, to further investigate the renal protective effect of tetrandrine, we focused its role on the proliferation of human mesangial cells (HMCs). 25 μg/ml aIgA1-stimulated HMCs was used to mimic the IgAN HMC model in vitro, as previously described (17). The CCK8 assay showed that tetrandrine at concentrations of 2.5–10 μM for 24 h had strong inhibitory effects on aIgA1-induced mesangial proliferation, and did not alter the viability of HMCs (Figure 8A). The representative light microscopy was shown in Figure 8B. To further confirm the effect of tetrandrine on the proliferation of HMCs, cell cycle analysis was carried out by flow cytometry. After 24 h incubation with aIgA1, the proportion of cells in S and G2/M phases increased, while 5 μM tetrandrine retained more cells in G0/G1 phase (Figure 8C).