The culture protocol of the primary human GMC (HRMC) and the human GMC line (C2M12) was described in detail in our previous study (17).

HRMC and C2M12 were seeded uniformly on six-well plates. When the subcultured cells reached 70% confluency, GMCs were cultured in a medium containing 0.5% FBS overnight for synchronization, followed by treatment with 1 μg/mL of SEB (Nanjing Aiyi Biotechnology Co., Ltd., Nanjing, China), Staphylococcus aureus Cowan 1 (SAC) (1:1,000), or lipopolysaccharide (LPS) (5 μg/mL) in a serum-free culture medium for 48 h.

Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc. Massachusetts, USA) was used for the transfection with small interfering RNA (siRNA). siRNAs directed against different regions of the constant region of the IGHA1 (siRNA-1 and siRNA-2) and the non-specific, scrambled, control siRNA (NC) were designed by GenePharma Company (Shanghai, China) (Supplementary Table S1). The C2M12 was transfected with siRNA in a serum-free culture medium for 48 h following the Lipofectamine 3000 protocol.

Total RNA of cultured cells was extracted with the RaPure Total RNA Kit (Magen, Inc., Guangzhou, China), and its concentration was assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Inc.). Then, 2 μg of total RNA was reverse-transcribed to cDNA using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Inc.). The FastStart Universal SYBR Green Master Mix kit (Roche, Mannheim, Germany) was used for quantitative polymerase chain reaction (qPCR), with the primers targeting complement C3 (C3), interleukin 6 (IL-6), collagen type IV alpha 1 (COLIVA1), fibronectin 1 (FN1), and the constant region of immunoglobulin heavy constant alpha 1 (IGHA1). The reaction conditions regulated by the 7500 FAST real‐time PCR system (Applied Biosystems, USA) and the reaction conditions for standard PCR are listed in Supplementary Table S2. HRMC cDNA was also used to amplify the hinge region of IGHA1 covering the serine/threonine-rich region by PCR with primers targeting the CH1-H-CH2 region. The sequences of the primers are listed in Supplementary Table S3. PCR products of the IGHA1 CH1-H-CH2 region amplification were sent to BGI.Tech (Beijing, China) for sequencing, and the sequence was analyzed by the Chromas software.

The cell lysis solution, extracellular matrix component (ECM), and culture supernatant were harvested for Western blot analysis. Cultured cells were detached from the bottom of the six-well plate by incubating with 1 mL of Versene solution (NaCl 8 g/L, KCl 0.2 g/L, Na₂HPO₄*12H₂O 2.3 g/L, KH₂PO₄ 0.2 g/L, EDTANa₂ 0.2 g/L, pH 7.2–7.4) in each well (37°C, 10 min) and collected into 15 mL centrifuge tubes. Thus, the ECM remained intact and deposited at the bottom of the plate. The cell pellet was obtained after centrifugation for cell lysis. An isochoric TSD lysis buffer (TSD lysis buffer, 1% SDS; 50 mmol/L, pH 7.5 Tris–HCl) was added into each well of the six-well plates to collect the ECM. The culture supernatant was concentrated by the ultrafiltration centrifuge tube (Merck, Darmstadt, Germany) into an isochoric solution. The samples after denaturation were immediately used for Western blot analysis in non-reduced and reduced conditions. Serum from IgAN patients and the culture medium were used as positive control and negative control, respectively.

Renal biopsy specimens from IgAN patients and para-cancerous renal cortices from patients with renal carcinoma were obtained with informed consent from patients and approval of the Research Ethics Committee, Peking University Third Hospital, Beijing.

The inclusion criteria for IgAN patients were listed as follows: 1) pathologically confirmed primary IgAN, 2) 18–45 years old, 3) eGFR >30 mL/min/1.73 m², and 4) urinary protein <3.5 g/day. The exclusion criteria for IgAN patients were as follows: 1) secondary IgAN such as disease associated with viral infections, autoimmune diseases, hepatic cirrhosis, and tumors; 2) IgAN accompanying other renal diseases such as crescentic nephritis, minimal change disease, membranous nephropathy or other nephropathy, and transplanted kidney IgAN; 3) use of immunosuppressive agents; and 4) renal replacement therapy has been performed.

The inclusion criteria for para-cancerous renal cortices from patients with renal carcinoma included the following: 1) the renal cortices were obtained far enough from the tumor boundary (at least 5 cm away) to prevent transcription alteration due to malignancy differentiation and 2) patients without hypertension, diabetes, hepatitis B, tuberculosis, amyloidosis, systemic lupus erythematosus vasculitis, or Sjogren’s syndrome.

Renal tissues were mixed and minced into small pieces (<1 mm³ in size) in ice-cold HBSS^−/− using a razor blade. These small pieces were passed through a 100-μm cell strainer, the flow-through was then passed through a 70-μm cell strainer, and the retained glomeruli were transferred onto the slides for isolation. Single glomeruli were isolated under a microscope by capillary pipette and digested in PBS containing 1 mg/mL of collagenase I (Sigma Chemicals, St. Louis, MO, USA) for 5 min at 37°C. Glomerular intrinsic cells were released, and single glomerular cells were manually picked out under the microscope by a capillary pipette.

RNA extraction, cDNA synthesis, and amplification of these single glomerular cells were carried out following the protocol we previously described (19). ITGA8 and ACTA2 were considered as mesangial cell-specific markers (25). The sequences of the primers are listed in Supplementary Table S3. The specific V_HDJ_H rearrangement sequence of IGHA mRNA in these single GMCs was amplified by nested PCR and compared with those in the basic local alignment search tool (IgBLAST, https://www.ncbi.nlm.nih.gov/igblast/) to identify the best-matching germline gene segments and junctions. Detailed protocols including reaction conditions and primers used for nested PCR were described in the previous study by our team (19).

Kidney single-cell RNA sequencing (scRNA-seq) data were downloaded from the public scRNA-seq database (DISCO, DISCO (immunesinglecell.org). The data processing, quality control, and filtering were conducted in R using the Seurat package (Version 4.3.0). Each IGHA1-expressing cell cluster was assigned to a corresponding cell type based on known cell-type markers. Mesangial cell-type markers include PDGFRB, ITGA8, LUM, and COL1A2 (25), and CD19, CD27, and CD38 were considered specific markers for B and plasma cells.

Gene set enrichment analysis was performed to explore the IGHA1-associated biological functions and pathways in IGHA1-expressing GMCs, in which the c2.cp.kegg_legacy.v2024.1.Hs.symbols.gmt, c2.cp.reactome.v2024.1.Hs.symbols.gmt, and c5.go.bp.v2024.1.Hs.symbols.gmt were used as reference genomes. An adjusted P-value <0.05 acquired through the Benjamini–Hochberg approach was considered statistically significant.

FOXD1^GC mice and IGHA^flox (IGHA^fl) mice were purchased from the Jackson Lab and the Biocytogen Pharmaceuticals (Beijing), respectively. IGHA homozygous and Foxd1 heterozygous mice were bred to generate FOXD1^GCIGHA^fl/fl mice. As Foxd1 is mainly expressed in GMCs, we considered FOXD1^GCIGHA^fl/fl mice as the GMC conditional IGHA knockout mice (cKO mice), and the IGHA homozygous mice (IGHA^fl/fl) without the Foxd1 cre allele, named as wild-type mice (WT mice), were used as littermate controls. The PCR conditions and primers used for mice genotyping are listed in Supplementary Tables S2, S3.

Ighmtm1Cgn (μMT) mice, which lack serum IgA and B cells (26), were a kind gift from Professor Qiu (Peking University Health Science Center, Beijing, China). Balb/c mice were purchased from the Charles River Laboratories (Beijing, China). Animal experimentation was performed following the NIH Guide for the Care and Use of Laboratory Animals.

Male mice aged 8–10 weeks were utilized for the experiment. In the IgAN model group, each mouse was given 0.4 mL of 6 mmol/L hydrochloric acid solution containing 0.1% BSA (1 g/L) (Sigma-Aldrich, Inc., USA) every 2 days via oral gavage in the first 5 weeks, followed by tail intravenous injection of 0.2 mL of 0.01 mol/L PBS containing 2% BSA for three consecutive days in the sixth week. From the 9th to the 11th week, 0.1 mL of SEB solution (diluted in PBS) was administered to each mouse (0.5 mg/kg) via tail intravenous injection once a week. In the control group, 0.4 mL of 6 mmol/L hydrochloric acid solution was given in the first 5 weeks, and PBS was intravenously injected on the 6th and 9th to 11th weeks.

Mice were sacrificed on the 12th week, and 24-h urine was collected in the metabolic cage 1 day before the harvest. Hematuria, calculated by red blood cell count per microliter in urine, was measured by flow cytometry, and urine albumin–creatinine ratio (UACR) and serum urea and serum creatinine levels were measured using a biochemical analyzer, respectively, in the clinical laboratory in our hospital. Cardiac perfusion with PBS was performed to remove the intrarenal blood before kidney harvest.

RNA-fluorescence in situ hybridization (FISH) was performed on human para-cancerous kidney tissue using the RNASweAMI fluorescence in situ hybridization detection kit (GF007; Servicebio, Hubei, China), and the FISH probe targeting human IGHA1 was designed and synthesized by Servicebio. After hybridization following the recommended protocol, we co-stained the Pdgfrb protein to display glomeruli at 37°C for 1 h, followed by a 1-h secondary antibody incubation at room temperature.

For indirect immunofluorescence (IF) staining, cells cultured on confocal dishes were fixed and permeabilized with 4% paraformaldehyde and 0.1% Triton X-100. Mice kidney samples embedded in OCT were fixed and permeabilized by cold acetone. Fluorescent signals were detected with a confocal laser scanning microscope (Carl Zeiss, Germany), and the intraglomerular mean fluorescence intensity was quantitatively measured by ZEN 2012 SP2.

Immunohistochemistry (IHC) staining was performed on the formalin-fixed paraffin-embedded kidney tissues. We incubated the sections with 10 mM of citrate buffer (pH = 6.0) in a pressure cooker for 3 min for antigen retrieval. The intraglomerular mean optical density (mOD) of the specific antibody staining was quantitatively analyzed by the IHC Toolbox plugin in the ImageJ software.

Masson’s trichome staining and hematoxylin and eosin (HE) staining were performed following the protocols in the Masson Trichrome Staining Kit and the HE Stain Kit (Solarbio, Beijing, China), respectively. At least 10 glomeruli with maximal cross section for each mouse were randomly chosen for histological statistical analysis.

The following antibodies were used: Gd-IgA1 (1:1,000 in WB, 1:20 in IF; KM55; Immuno-Biological Laboratories, Gunfecture, Japan), IgA (1:1,000 in WB, 1:200 in IF; ab124716; Abcam, Cambridge, UK), COLIV (1:1,000 in WB, 1:2,000 in IHC; ab236640; Abcam), FN (1:1,000 in WB, 1:2,000 in IHC; ab268022; Abcam), PDGFRB (1:100; ab32570; Abcam), C3 (1:1,000 in WB, 1:2,000 in IHC; ab200999; Abcam), IL-6 (1:1,000; ab290735; Abcam), and C1GALT1 (1:3,000 in WB, 1:300 in IF; ab237734; Abcam).

Aside from using the anti-IgA (1:20; Gene Tech Company Limited, Shanghai, China) antibody for direct IF was officially used for the detection of glomerular IgA deposition in IgAN diagnosis in the Pathology Department at Peking University Third Hospital.

At least three independent biological repeats were performed in every experiment. The significant differences between groups were calculated by Student’s unpaired t-test and one-way ANOVA with Tukey’s multiple comparison test. A P-value <0.05 was considered statistically significant. All statistical analyses were performed using GraphPad Prism version 8.0 for Windows (GraphPad Software, San Diego, CA).

Nested PCR on individual glomerular cells was used to demonstrate the expression of IGHA in isolated GMCs. Single glomerular cells were isolated from para-cancerous renal cortices to detect IGHA V_HDJ_H gene rearrangement and transcription (Figure 1a). Fifteen single cells with transcription features of ITGA8 and ACTA2 were identified as GMCs (Figure 1b), of which four GMCs were detected to express the variable region of the IGHA transcripts. Transcription of the variable region of IgHE, IgHG, and IgHM was also detected in other GMCs (Figure 1c).

The transcription of IGHA1 in GMCs was reconfirmed in the public single-cell RNA-sequencing database (DISCO, DISCO (immunesinglecell.org), in which fibroblasts with transcription features of PDGFRB+ITGA8+LUM+COLIA2+ were considered as GMCs (Figures 2a, b). It was reported that GMCs present with fibroblast features (27). This group of cells was not B cells or plasma cells, as they did not express CD19, CD27, or CD38 (Supplementary Figure S1a). Although the IGHA transcriptional level in GMCs was relatively lower compared with that in plasma cells, IGHA1 was the dominant one in both cells (Figures 2c, d). Reclustering of the IGHA1-expressing cells in the database based on the known cell-type markers additionally revealed that not only immune cells but also renal parenchymal cells including mesangial cells, podocytes, tubular epithelial cells, pericytes, and endothelial cells could express IGHA1 (Supplementary Figure S1b). Further investigation of the differential expression pattern among these IGHA1-expressing cells could prevent these IGHA1-expressing renal parenchymal cells from B-cell or plasma-cell contamination. Supplementary Figure S1c lists the top 8 genes differentially expressed in each IGHA1-expressing cell type.

Following our finding that expression of IGHA was identified in single GMCs, we detected IgA expression in glomerular mesangium in situ in human para-cancerous renal tissue and in μMT mice.

IgA protein deposition in glomerular mesangium in para-cancerous renal tissues was detected clearly although IgA staining was light and sparse compared to the thick and dense staining in IgAN renal biopsies (Figure 3a). More importantly, fluorescent hybridization of IGHA1 RNA (orange) and Pdgfrb protein co-staining (green) reconfirmed the IGHA1 transcription by GMCs in situ (Figure 3b). To exclude the presence of IgA from local B lymphocytes or the infiltrated serum IgA, IgA expression in glomerular mesangium was detected in μMT mice lacking B lymphocytes and serum IgA (28), and it was usually used to study the effect of B-lymphocyte-originated immunoglobulin (29). The IgA expression in μMT mice was clear and precise, even with a higher intensity than that in wild-type mice (possibly due to the weakened inhibitory effect of systemic immunity on local immunity (30)) (Figure 3c).

Gd-IgA1 deposition in glomerular mesangium is the most important factor that initiates the pathogenesis of IgAN. A serine/threonine-rich region was detected in the IGHA1 transcripts in HRMC (Supplementary Figure S2). While HRMC transcribed the CH1-H-CH2 sequence in the IGHA1 constant region with a similar length to PBMC (Supplementary Figures S2a, b), the amino acid composition of its serine/threonine-rich region (containing 6 serine/threonine in 18 amino acids) is different from that in B lymphocytes (Supplementary Figure S2c) (31). The presence of a serine/threonine-rich region in IGHA1 transcripts provided the possibility of galactose deficiency in IgA1 produced by cultured GMCs.

Western blot demonstrated that Gd-IgA1 in unreduced condition in both HRMC and C2M12 cytoplasm was larger than 250 kDa, similar to that in the serum from IgAN patients, suggesting that Gd-IgA1 produced by cultured GMCs could agglomerate to form poly-Gd-IgA1. Gd-IgA1 produced by cultured GMCs presented at approximately 67 kDa in reduced conditions, while Gd-IgA1 in the serum from IgAN patients was located at 55 kDa (Figure 4a). The absence of a band in the MCM eliminates the possibility of Gd-IgA1 in the culture medium. Different molecular weights of cultured GMCs compared with the Gd-IgA1 from the serum of IgAN patients might be attributed to different glycosylation of Gd-IgA1. Immunofluorescence demonstrated the expression of Gd-IgA1 and C1GALT1 in HRMC and C2M12 (Figure 4b), suggesting that IgA in GMCs could be transformed into Gd-IgA1 in a similar mechanism as in plasma cells. Mean immunofluorescent intensity of the Gd-IgA1 staining in C2M12 was significantly lower after IGHA1 knockdown by siRNA-2 transfection (Figure 4c), which further confirmed the expression of Gd-IgA1 in GMCs.

The top 20 biological functions or pathways enriched by GSEA with the highest nominal enrichment score (NES) significantly associated with IGHA1 expression in GMCs are listed in Figures 5a–c. IGHA1 in GMCs was positively associated with response to bacterium, innate immune response, and complement activation and was negatively associated with galactose metabolism (Figure 5d), suggesting that bacterium infection-induced innate immune response might be responsible for the IgA transformation into Gd-IgA1, and Gd-IgA1 might trigger the subsequent complement activation in GMCs.

In vitro, SEB could significantly upregulate the transcription of IGHA1 after 24 h of stimulation (Figure 6a) and induced overexpression of Gd-IgA1 in the intracellular compartment in HRMC after 48 h compared with the control group, while the IgA expression level remained relatively constant (Figures 6b, c). In addition to SEB, LPS and SAC could also induce Gd-IgA1 overproduction in HRMC (Figure 6d).

Accompanied by Gd-IgA1 upregulation in HRMC after SEB stimulation, C1GALT1 was significantly downregulated, and transcription and expression of COLIV, FN, C3, and IL-6 were significantly elevated (Figures 6a, c). TLRs, important members of the innate immune response, were constitutively expressed in HRMCs (Supplementary Figure S3), and TLR4 appeared to have the highest expression level. TAK242 (an antagonist of TLR4) could block the elevated transcription of IGHA1, and the expression of Gd-IgA1 relieved the abnormal expression of COLIV, FN, C3, IL-6, and C1GALT1 induced by SEB (Figures 6e–g).

IGHA1 knockdown in C2M12 was used to investigate the pro-inflammatory effect of Gd-IgA1. siRNA-2 significantly knocked down the expression of IgA and Gd-IgA1 in C2M12 after 48 h of transfection in the condition of serum-free culture with or without SEB stimulation (Figures 7a, c). Gd-IgA1 deficiency reduced the production of COLIV, C3, and IL-6 in the cell lysate (Figure 7a) and reduced the deposition of COLIV and FN in the extracellular compartment and their release into the supernatant (Figures 7b, c).

We first established the IgAN mouse model in μMT mice to exclude the effect of B-lymphocyte-originated IgA on IgAN formation, after which we managed to observe the causality between GMC-expressed IgA and the formation of IgAN in mice by conditionally knocking out IGHA expression in GMCs.

Similar to Balb/c mice, mesangial IgA staining was significantly enhanced after IgAN establishment (Figure 8a) in μMT mice, accompanied by increased inflammatory cell infiltration into glomeruli (increased intraglomerular cell count in HE staining) and ECM expansion (Figure 8b), while mesangial proliferation in μMT mice was not significant after model construction (Figure 8a).

PCR confirmed the IGHA homozygous genotype, and mice were classified into the FOXD1 cre+ group (cKO) and the FOXD1 cre− group (WT) (Supplementary Figure S4a). Mean fluorescence intensity of intraglomerular IgA staining in cKO mice was significantly lower than that in WT mice, confirming the IGHA knockout efficiency in GMC, in which platelet-derived growth factor receptor beta (PDGFRB) served as a marker of GMCs (Supplementary Figure S4b). As we used the FOXD1 cre system for conditional knockout, IgA expression in CD19⁺ cells in the spleen and the serum IgA level remained constant between the WT and cKO groups (Supplementary Figure S4c).

In WT mice, intensified IgA in glomerular mesangium was observed in the IgAN group compared with the control group, with significant mesangial hypercellularity proven by PDGFRB staining and HE staining (Figures 9a, b). Masson’s trichome staining revealed more collagen deposition, and IHC demonstrated more deposition of COLIV, FN, and C3 in the mesangium in the IgAN group compared with the control group (Figure 9b). Significantly deteriorated hematuria and enhanced UACR were also found in the IgAN group (Figure 9c). Stronger IgA expression in the mesangium and significant hematuria and proteinuria with enhanced intraglomerular inflammation response indicated the successful construction of the IgAN mouse model.

Compared with the WT IgAN model, less IgA expression in the mesangium was detected in cKO IgAN mice (Figure 9a). GMC proliferation indicated by PDGFRB staining (Figure 9a) and HE staining (Figure 9b) and extracellular matrix deposition by Masson’s trichome staining and IHC (Figure 9b) were all alleviated in the cKO group compared with the WT group. Moreover, deficient IgA expression in GMCs also led to reduced C3 deposition in the mesangium (Figure 9b) and alleviation of hematuria and proteinuria (Figure 9c) in the cKO IgAN model. However, the kidney function of mice in the IgAN group did not change significantly which might be attributed to the early stage of the disease (Figure 9c). To sum up, these results indicated that IgA expression in GMCs was responsible for inducing intraglomerular inflammation and alleviating the hematuria and proteinuria in IgAN in mice.