Spatial transcriptomic profiling reveals molecular signatures of endocapillary hypercellularity in IgA nephropathy

Abstract

Background

Immunoglobulin A nephropathy (IgAN) is a leading cause of chronic kidney disease, marked by mesangial IgA1 immune-complex deposition and a multi-hit pathogenesis. The Oxford classification grades histopathological lesions such as endocapillary hypercellularity (E0 and E1), which are associated with disease severity and prognosis. Despite advances in single-cell and spatial transcriptomics, the molecular basis of endocapillary hypercellularity in human IgAN remains unclear.

Methods

Spatial transcriptomic data obtained using the GeoMx DSP platform from kidney biopsy specimens of East Asian patients were analyzed to compare gene expression profiles across normal, E0, and E1 groups. Differentially expressed genes (DEGs) were identified, and subsequent gene ontology and protein–protein interaction network analyses were performed to investigate functional pathways and molecular interactions.

Results

Analysis identified 50 DEGs between E0 and controls, with more extensive changes in E1 compared with both E0 and controls (342 and 223 DEGs, respectively). E1 lesions displayed distinct molecular signatures, including the novel downregulation of amyloid beta precursor protein, cluster of differentiation 47, and integrin alpha V, along with well-known interactions among involving dual-specificity phosphatase 1, early growth response 1, and Fos proto-oncogene. Functional enrichment highlighted pathways related to ion response and mitogen-activated protein kinase/extracellular signal-regulated kinase signaling.

Conclusions

This study provides new insights into the molecular changes associated with endocapillary hypercellularity in IgAN progression. The identified gene signatures and pathways may serve as potential biomarkers of disease activity and targets for therapeutic intervention.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1186/s12882-026-05171-x.

Background

Immunoglobulin A nephropathy (IgAN) is an immune-complex glomerulonephritis characterized by mesangial deposition of IgA1-containing immune complexes. Its pathogenesis is traditionally described by a multistep “four-hit” model (1): excessive mucosal production of galactose-deficient IgA1 (Gd-IgA1) (2), generation of anti-glycan autoantibodies (IgA or IgG) targeting the Gd-IgA1 hinge region (3), formation of circulating Gd-IgA1–autoantibody immune complexes, and (4) mesangial deposition of these complexes, leading to complement activation and glomerular injury [1]. Genetic susceptibility—particularly human leukocyte antigen region alleles, complement factor H variants, and loci involved in mucosal immunity—along with environmental factors such as gut microbiome composition, modulate these steps. Cytokines including B-cell activating factor and a proliferation-inducing ligand promote survival and proliferation of mucosal B cells responsible for Gd-IgA1 production [1]. Histopathologically, IgAN exhibits variable lesions. The Oxford (MEST-C) classification is widely used to grade these findings: M for mesangial hypercellularity (≥4 mesangial cells per area), E for endocapillary hypercellularity (inflammatory cells within capillary lumina), S for segmental glomerulosclerosis, T for tubular atrophy/interstitial fibrosis, and C for cellular or fibrocellular crescents [2]. The presence of endocapillary hypercellularity (E1) indicates active glomerular inflammation and is associated with more aggressive disease. Prior studies have shown that E1 lesions correlate with higher proteinuria, faster renal function decline, and an increased risk of progression to end-stage kidney disease. In contrast, the absence of endocapillary hypercellularity (E0) generally reflects a less active or more slowly progressive course [3–5].

Recent single-cell and spatial transcriptomic studies have begun to map the cellular and molecular landscape of IgAN at high resolution [6]. Single-cell RNA sequencing of IgAN kidneys has delineated diverse immune and renal cell populations, including expanded peripheral B-cell and T-helper subsets, relative depletion of cytotoxic T cells, and multiple kidney-infiltrating immune cell types. These studies have also revealed altered mesangial and endothelial cell gene expression profiles in early disease [7]. Complementing these findings, spatial transcriptomics in human IgAN biopsy tissue has localized gene expression changes to specific glomerular regions. One study reported that glomeruli with mesangial proliferation (Oxford M1 lesions) showed upregulation of the podocyte injury marker transcription factor 21 and enrichment of genes involved in cell adhesion, vascular development, and extracellular matrix processes, distinguishing M1 IgAN from both non-proliferative IgAN and controls [8]. Similarly, in a rat Thy-1.1 IgAN model evaluated with spatial multi-omics, mesangial proliferation was linked to metabolic reprogramming. In particular, the polyamine pathway genes Oat and Odc1 were downregulated in glomeruli, accompanied by reduced spermidine and L-proline levels, correlating with suppressed mesangial cell growth under treatment [9]. Although these spatial approaches illustrate how mesangial expansion and glomerular injury map to specific transcriptomic signatures, no spatial transcriptomic changes have yet been defined according to endocapillary hypercellularity status in human IgAN. Recent integrative analyses combining single-cell RNA sequencing and spatial transcriptomics have also identified early activation of inflammatory pathways and novel inflammation-associated molecules in glomerular endothelial cells from mild IgAN specimens, suggesting that endothelial cell responses may precede mesangial alterations in disease pathogenesis [10].

In this study, we conducted spatial profiling of IgAN specimens obtained from East Asian patients to examine transcriptional changes associated with the degree of endocapillary hypercellularity. We hypothesized that glomerular spatial transcriptomic data would identify molecular markers relevant to disease development, as well as key pathways that differ between E0 and E1 lesions, while providing a basis for future validation in more diverse populations.

Methods

Study participants

Patients diagnosed with IgAN by kidney biopsy at Kyung Hee University Medical Center were screened for inclusion. Eligible participants were those with adequate formalin-fixed paraffin-embedded biopsy tissue available and who provided informed consent for research use. Inclusion criteria were preserved renal function (estimated glomerular filtration rate ≥ 60 mL/min/1.73 m²), absence of diabetes mellitus, no coexisting glomerulonephritis, no prior immunosuppressive therapy, more than 10 glomeruli per section, biopsy performed between 2018 and 2023, and established Oxford classification. To minimize confounding, only patients aged 20–59 years were included, and cases with cellular or fibrocellular crescents, significant tubular atrophy, or interstitial fibrosis were excluded. Biopsy specimens were reviewed by a pathologist to assess endocapillary hypercellularity, classifying glomeruli as E0 (absent) or E1 (present). Three IgAN cases with E0 lesions and four with E1 lesions were selected for analysis.

For the control group, kidney tissues were obtained from four patients who underwent nephrectomy for renal malignancies at the same institution and consented to the use of their specimens for research. Non-tumorous kidney tissue, verified by a pathologist to be free of malignancy and other pathological changes, was selected for analysis.

All study procedures complied with the ethical principles of the Declaration of Helsinki and were approved by the Institutional Review Board at each participating hospital. The approval number for Kyung Hee University Medical Center was 2023–05-084. Written informed consent was obtained from all participants.

Statistical analysis

Statistical analyses were conducted to evaluate differences in clinical variables between groups. Continuous variables were analyzed using the Kruskal–Wallis test. Categorical variables were analyzed using Fisher’s exact test.

Spatial transcriptomic data generation

Areas of interest (AOIs) were manually delineated by the pathologist to include glomeruli representative of E0 or E1 lesions, while excluding those with segmental sclerosis or other pathological abnormalities. AOI selection was based on histomorphologic identification of glomeruli under brightfield and nuclear staining, ensuring inclusion of entire glomerular tufts and exclusion of surrounding tubulointerstitial regions. The typical AOI size was 36,032 ± 8,094 µm², and each AOI contained a median of 195.5 nuclei (interquartile range, 177.0–218.5). Endocapillary regions were selected based on established histopathologic criteria for E1 lesions, whereas non-representative or structurally distorted glomeruli were excluded from AOI selection.

Libraries were generated according to the GeoMx DSP NGS Readout User Manual (MAN-10153–01), and sequencing data were processed using the GeoMx NGS DRAGEN Pipeline. After sample indexing, libraries were amplified by polymerase chain reaction and sequenced on a NovaSeq 6000 platform (Illumina, San Diego, California, USA). Paired-end sequencing was performed with a read length of 27 bp for both Read 1 and Read 2. Dual indexing was used with Index 1 (i7) and Index 2 (i5), each 8 bp in length, allowing for a single mismatch.

Data quality control and processing

FASTQ files were processed using GeoMx NGS v2.3.3.10. Quality control was performed using the following thresholds: raw reads ≥ 1,000; aligned reads ≥ 80%; stitched reads ≥ 80%; trimmed reads ≥ 80%; sequencing saturation ≥ 50%; negative probe count geomean ≤ 10; and no-template control ≤ 1,000. AOIs with a surface area ≥ 16,000 µm² and more than 200 nuclei were included in downstream analyses. Additional quality control included assessment of relative log expression to evaluate potential batch effects across samples. To eliminate possible probe-capture artifacts, six genes (FOXN2, GPX3, IGHG2, IGHG3, IGHG4, IGKC) were excluded from the analysis when the difference between the highest and second-highest counts per million mapped reads (CPM)-normalized expression values within each sample group exceeded 1,000.

Downstream bioinformatics analysis

Principal component analysis was performed to visualize similarities and differences among sample groups, and differentially expressed genes (DEGs) were identified using DESeq2 v.1.40.2. Comparisons were made across the normal (control), E0, and E1 groups. Genes were designated as DEGs based on Bonferroni correction with an adjusted p-value (false discovery rate, FDR) <0.1 and an absolute log₂ fold change (LFC) > |1|. DEGs were further classified into upregulated and downregulated categories. Expression levels were compared across samples using log₂ CPM-normalized and CPM-normalized values.

Gene ontology (GO) term enrichment was performed using clusterProfiler v.4.8.3 with Benjamini–Hochberg correction; terms with FDR < 0.05 were considered significant. Protein–protein interaction (PPI) network analysis of DEGs was conducted using the STRING database, including only interactions with a confidence score ≥ 0.5 (medium confidence). DEGs without recorded interactions were excluded. The interaction network was visualized using the confidence-based network setting of the STRING API for Homo sapiens.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this manuscript, the authors used GPT-4 (OpenAI) to improve the readability and language. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the final manuscript.

Results

Clinical and pathological characteristics of study participants

The clinical and pathological characteristics of patients with IgAN classified by endocapillary hypercellularity (E0 and E1) and of control subjects are summarized in Table 1. All participants were between their 20s and 50s and included both men and women. All patients showed preserved renal function with an estimated glomerular filtration rate of at least 60 mL/min/1.73 m². Proteinuria was higher in the IgAN groups compared to the controls. Because urine PCR values were not available for the control group, statistical comparisons were limited to the E0 and E1 groups. Using the Kruskal–Wallis test, no significant difference was observed between E0 and E1 for random urine PCR with a p-value of 0.2888. Similarly, 24-hour urine PCR did not differ significantly between the two groups with a p-value of 0.2752. Hypertension was more frequently observed in the IgAN groups compared with the control group. However, Fisher’s exact test showed no statistically significant difference in hypertension prevalence across groups with a p-value of 0.3091. Sex distribution did not differ significantly between groups based on Fisher’s exact test. Age was evaluated using the Kruskal–Wallis test and showed no statistically significant difference across groups with a p-value of 0.09881. According to the Oxford classification, all IgAN cases showed mesangial hypercellularity, segmental sclerosis, and tubular atrophy or interstitial fibrosis. Endocapillary hypercellularity was observed only in the E1 group.

Table 1.

Participant information. Urine RBC (/HPF): 0 = 0–1, 1 = 2–4, 2 = 5–9, 3 = 10–29, 4 = many. G, glomerulus; E, endocapillary hypercellularity; C, crescent formation; M, mesangial proliferation; S, glomerulosclerosis; T, tubular atrophy; PCR, protein-to-creatinine ratio; RBC, red blood cell; HPF, high-power field

Group	Tissue	Sex	Age	Diabetes	Hypertension	Dyslipidemia	Oxford Classification	Creatinine (mg/dL)	eGFR (mL/min/1.73 m2)	Cystatin–C (mg/dL)	IgA (mg/dL)	C3 (mg/dL)	Urine RBC (/HPF)	random urine PCR (g/gCr)	24 hr urine PCR (g/gCr)
Control	G	F	50s	(-)	(-)	(-)		0.65	102.0743
Control	G	M	50s	(+)	(-)	(-)		1.01	81.0363
Control	G	M	50s	(-)	(-)	(-)		1.17	70.2621
Control	G	F	50s	(-)	(-)	(-)		0.55	106.337
E0	G	M	30s	(-)	(-)	(-)	M0, E0, S1, T0, C0	1.09	89.96	0.87	229	100	1	0.372	0.25774
E0	G	F	50s	(-)	(+)	(+)	M0, E0, S1, T0, C0	0.91	70.5289	1.01	339	96.1	0	0.686364	0.428021
E0	G	F	40s	(-)	(+)	(-)	M0, E0, S1, T0, C0	0.85	83.9152	1.09	1390	98.6	1	0.684058	0.566772
E1	G	M	30s	(-)	(-)	(-)	M0, E1, S1, T0, C0	1.22	74.1307	1.02	404	154	2	0.264706	0.182609
E1	G	F	50s	(-)	(+)	(-)	M0, E1, S1, T0, C0	0.78	77.86	0.95	158	99.8	1	0.350781	NA
E1	G	F	20s	(-)	(-)	(-)	M0, E1, S1, T0, C0	0.46	141.2018	0.48	287	93.3	1	0.805063	0.331915
E1	G	F	50s	(-)	(+)	(+)	M0, E1, S0, T0, C0	0.92	71.586	0.99	1480	115	0	0.29898	0.283146

Quality control and sample characteristics

Manually delineated AOIs with irregular outlines were drawn by an experienced pathologist. Each AOI was defined to include sufficient cellular content and glomerular area to meet the quality control thresholds required for spatial transcriptomic analysis, ensuring robust signal detection without reliance on standardized AOI shapes.

One representative periodic acid–Schiff-stained image per replicate was provided for each sample (Supplementary Figure 1). Glomerular AOIs were consistently defined using morphology-guided segmentation, encompassing entire glomerular tufts and excluding adjacent tubulointerstitial compartments (Supplementary Figure 2). Potential batch effects were evaluated by examining the median nuclei count for each group and reviewing the relative log expression plot. The median nuclei counts were 209 for controls, 200 for the E0 group, and 180 for the E1 group, suggesting minimal risk of batch effects (Fig. 1A). Using the expression levels of renal microstructure marker genes, samples showed the expected pattern of higher expression in glomerulus (PODXL, NPHS1, NPHS2), followed by proximal convoluted tubule (SLC34A1), distal convoluted tubule (SLC12A3), collecting duct (AQP2), and vascular smooth muscle cell markers (CNN1; Fig. 1B). Overall, the separation of clusters were consistent with the disease grade (control, E0, and E1) except for the one E0 sample with 4 replicates (3–1 to 3–4) closer to E1 sample cluster (Fig. 1C). For the clear identification of transcriptional differences among the groups, we removed the sample for the downstream analyses.

Fig. 1.

Overview of initial quality control across all samples. (A) RLE plot for identification of batch effects. (B) Expression box plot for renal microstructure marker genes: glomerulus (PODXL, NPHS1, NPHS2), proximal convoluted tubule (SLC34A1), distal convoluted tubule (SLC12A3), collecting duct (AQP2), and vascular smooth muscle cell (CNN1). (C) PCA of all samples across control, E0, and E1 groups. RLE, relative log expression; PCA, principal component analysis

To reduce inter-sample variation, six genes (FOXN2, GPX3, IGHG2, IGHG3, IGHG4, IGKC) with a difference ≥ 1,000 between the highest and second-highest CPM-normalized expression values within the same group were excluded because they represented apparent outliers (Supplementary Figure 3). Downstream analyses were conducted using 18,669 genes.

Differential expression analysis and DEG patterns

DEGs comparisons were identified using FDR < 0.1 and LFC > |1| criteria (Supplementary Table 1). In total, 50 genes (24 downregulated and 26 upregulated) were differentially expressed in E0 compared with controls (E0/Control). The E1 group showed a more substantial transcriptional shift, with 311 downregulated and 31 upregulated genes relative to E0 (E1/E0), and 183 downregulated and 40 upregulated genes relative to controls (E1/Control; Fig. 2A). Analysis of DEG overlap across comparisons revealed no gene consistently increased or decreased across all three group comparisons (Fig. 2B). Among upregulated DEGs, nucleolar protein 9 (NOP9) was shared between E0/Control and E1/Control, while interferon alpha-inducible protein 6 (IFI6) and orofacial cleft 1 candidate 1 (OFCC1) were shared between E1/E0 and E1/Control. Among downregulated DEGs, 14 genes overlapped between E0/Control and E1/Control, and 72 genes overlapped between E1/E0 and E1/Control (Fig. 2B and Supplementary Material 1). DEG expression patterns were further compared using log₂ CPM-normalized expression values (Fig. 2C). To assess how these genes change with IgAN progression, CPM-normalized expression values were examined within each section of the Venn diagram. Four major expression patterns emerged (Supplementary Figure 4A–B) [1]: genes increased in E0 relative to controls but returned to control levels in E1 (BMERB1, CBLC, CPLX2, DEGS2, ELN, ENHO, EVX1, KRTAP12-3, KRTAP5-8, NDUFAF4, NXF2, PTK6, SPDEF, TLR5, WNT8A, ZIC3, ZNF587, ZNF616) [2]; genes decreased in E0 and then returned to control levels in E1 (ADAMTS9, CRABP2, CREG1, EFNA2, LBR, LECT2, MT1E, MT2A, NFKB2, NFKBIA, PABIR1, PGK1, PPP2CA, PPP2R2B, PXN, RNPEP, TAL1, WDR89, XKR4) [3]; genes showing a continuous increase across groups (CDH8, CYLC2, FAM171B, H2AC14, NXPE2, PLGLB1, SORCS3, STATH, ZFHX4, ZNF888); and [4] genes showing a continuous decrease (APP, FOSL2, ITM2B, MYH9, PLAT, RHOB, TIMP3, TSC22D1, VEGFA, ZFP36). Among overlapping DEGs, some genes were consistently upregulated (NOP9) or downregulated (ARRDC4, CEBPD, CHI3L1, DUSP1, EGR1, FOS, NR4A1, PDK4, PTGDS, TPPP3) in both E0 and E1 relative to controls, whereas others changed only at the E1 stage (IFI6, OFCC1, BMERB1, C4orf46, CRYAB, DUX4, PHAX, PRMT8, TSC22D3, ZNF587, ZNF776, ZNF845; Supplementary Figure 4C–D). Collectively, these findings demonstrate distinct gene expression trajectories associated with IgAN progression based on endocapillary hypercellularity status. The relatively small number of DEGs in E0 suggests that most transcriptional alterations emerge at the E1 stage, a pattern broadly consistent with the mesangial lesion-associated transcriptomic shifts reported by Park et al.

Fig. 2.

Degs among normal (control), E0-stage (E0), and E1-stage (E1) IgA nephropathy groups. (A) Volcano plots of DEGs between sample groups (left to right: E0/Control, E1/Control, E1/E0). Red and blue dots indicate upregulated and downregulated DEGs, respectively (LFC > |1|, FDR < 0.1). (B) Venn diagrams of upregulated (left) and downregulated (right) DEGs across comparisons. (C) Box plots of overlapping DEGs: upregulated (left); downregulated genes shared between E0/Control and E1/Control (middle); and downregulated genes shared between E1/E0 and E1/Control (right). The top 10 genes in the E1/Control comparison based on LFC are shown. DEG, differentially expressed gene; LFC, log₂ Fold change; FDR, false discovery rate

GO and PPI network analysis of DEGs

To assess pathway and functional enrichment, GO term analysis was performed on DEGs, with terms considered significant at FDR < 0.05. In the E0/Control comparison, downregulated genes were significantly enriched in biological processes related to ion responses, including “response to metal ion,” “response to zinc ion,” “cellular response to copper ion,” and “re sponse to cadmium ion” (Fig. 3A). In the E1/Control comparison, downregulated genes were enriched in categories associated with the mitogen-activated protein kinase/extracellular signal-regulated kinase signaling pathway (Fig. 3B). To explore potential interactions and underlying molecular mechanisms among these DEGs, PPI analysis was conducted using medium-confidence thresholds. The PPI networks revealed interactions among dual-specificity phosphatase 1 (DUSP1), early growth response 1 (EGR1), and Fos proto-oncogene (FOS), which appeared in both E0/Control and E1/Control comparisons, highlighted by navy boxes (Fig. 4A–B). In the E1/Control comparison, an additional interaction module involving amyloid beta precursor protein (APP), cluster of differentiation 47 (CD47), and integrin alpha V (ITGAV) was identified, highlighted by red boxes (Fig. 4B). These genes showed progressive downregulation with increasing disease severity, suggesting their potential as biomarkers (Fig. 4C). Overall, functional categories enriched among downregulated genes in E0/Control were also observed in E1/Control, with additional downregulated pathways emerging specifically in the E1/Control comparison.

Fig. 3.

Gene ontology term enrichment for downregulated genes (top 20 categories, false discovery rate threshold < 0.05). (A) E0/Control (genes = 24). (B) E1/Control (genes = 183)

Fig. 4.

Protein–protein interaction analysis. (A) Downregulated genes in E0/Control. (B) Downregulated genes in E1/Control. The navy box indicates the DUSP1–EGR1–FOS interaction. The red box indicates the APP–CD47–ITGAV interaction. (C) Expression patterns of genes involved in these interactions

Discussion

IgAN is a leading cause of chronic kidney disease worldwide, characterized by mesangial IgA1 deposition and a multi-hit pathogenesis [11]. Although numerous studies have examined IgAN, research on diagnostic or therapeutic strategies and their associated molecular markers remains limited. This study is the first to apply spatial transcriptomic profiling of glomeruli from patients with IgAN to investigate molecular changes associated with E0 and E1 classifications. Our aim was to identify potential therapeutic or prognostic biomarkers linked to disease progression. By elucidating molecular alterations across the E-class spectrum, this study offers insights that may contribute to a deeper understanding of IgAN pathogenesis.

Based on the overall differential expression analysis, a predominant pattern of downregulated gene expression was observed. The transition from E0 to E1 was accompanied by more extensive transcriptional alterations than the transition from control to E0, suggesting accelerated transcriptional reprogramming during disease progression. These transcriptomic changes provide important insights into the molecular dynamics underlying IgAN progression associated with endocapillary hypercellularity. Representative DEGs followed two distinct expression patterns. The first group—including EGR1, FOS, and NR4A1—showed altered expression as early as the E0 stage compared with controls, with dysregulation persisting into E1. This indicates that transcriptional abnormalities arise early in disease development and continue throughout progression. In contrast, the second group—comprising TSC22D3, ZNF587, and PRMT8—showed no significant differences between control and E0 but demonstrated marked alterations only in E1, suggesting activation of late-stage transcriptional programs characteristic of endocapillary hypercellularity. Spatial transcriptomic analysis identified NOP9 as the only consistently upregulated DEG across comparisons. NOP9 encodes a nucleolar protein essential for 18S ribosomal RNA maturation [12]. Although its CPM expression increased, absolute expression levels remained low, and no prior studies have linked NOP9 to IgAN, indicating the need for further investigation. IFI6 and OFCC1 were identified as upregulated exclusively in the E1 group. IFI6 is associated with antiviral activity [13], whereas OFCC1 has been implicated in orofacial clefts and schizophrenia [14]. Given the absence of prior evidence connecting either gene to IgAN, additional studies are needed to clarify their potential roles in disease progression. In contrast, many genes were downregulated in IgAN, including DUSP1, EGR1, and FOS—previously reported as potential biomarkers in earlier IgAN studies [8, 15–17]. FOS, a component of the AP-1 transcription factor complex involved in cell growth, apoptosis, inflammation, and differentiation [18], has been consistently shown to be downregulated in IgAN, consistent with our findings [8, 15]. EGR1, a zinc-finger transcription factor regulating cell growth and proliferation [19], is known to increase in diabetic kidney disease, yet this study confirmed its decreased expression in IgAN [20], in line with previous reports [8, 15]. DUSP1, a dual-specificity phosphatase that dephosphorylates threonine and tyrosine residues and is involved in apoptosis, tumor growth, inflammatory responses, and metabolic disorders, is reduced in diabetic nephropathy under hyperglycemic conditions [21]; prior IgAN studies have similarly reported its downregulation, which our results corroborate [8, 15].

This study also identified APP, CD47, and ITGAV as potential biomarkers specific to the E1 classification, none of which have been highlighted in previous IgAN research. APP was among the top 10 downregulated genes in the E1/Control comparison (Supplementary Figure 3B). It is known to participate in extracellular signal-regulated kinase-related signaling pathways, and the PPI network analysis demonstrated multiple interactions involving APP, suggesting its potential relevance as a biomarker. APP encodes a type I membrane glycoprotein involved in neuronal development, signaling, and intracellular transport [22]. Although it has primarily been investigated in the context of Alzheimer’s disease, APP has been recognized as a regulator of cell adhesion and MAPK/ERK signaling [23–25]. Its progressive downregulation in E1 lesions, together with enrichment of MAPK/ERK-related pathways, may indicate altered stress-responsive signaling in glomerular endothelial or mesangial cells, potentially contributing to impaired cellular turnover and increased intracapillary cell accumulation. To date, however, the role of APP in IgA nephropathy has not been directly investigated, and further studies are required to clarify its disease-specific relevance. Exploratory immunohistochemical staining for APP was additionally performed in this study; however, no significant differences were observed among the control, E0, and E1 groups. This finding suggests possible discordance between transcriptomic and protein-level alterations. Such discrepancies may reflect post-transcriptional regulation, disease heterogeneity, or the limited sample size available for tissue validation. CD47 encodes an integrin-associated glycoprotein of the immunoglobulin superfamily. Through interaction with signal regulatory protein α, CD47 contributes to preventing autoimmune reactions, maintaining immune tolerance, and regulating immune responses [26]. CD47 has been reported to be associated with leukocyte recruitment and renal inflammation and has also been implicated in the maintenance of endothelial homeostasis [27–29]. Consistent with these roles, altered CD47 expression may induce dysregulated leukocyte trafficking and endothelial dysfunction, thereby contributing to inflammatory cell accumulation within glomerular capillaries. Given the absence of prior studies examining CD47 in IgAN, the observed downregulation highlights a potential role that merits further investigation. ITGAV is a member of the integrin alpha-chain family and has been implicated in the regulation of angiogenesis, cell proliferation, migration, and invasion in various cancers [30]. Decreased ITGAV expression observed in E1 glomeruli may reflect disruption of endothelial–mesangial adhesion and altered extracellular matrix signaling, which could lead to maladaptive repair responses and expansion of intracapillary cellular components. Although αVβ6 knockout protects against renal inflammation and macrophage infiltration via Hippo/YAP–IL-34 signaling in ischemia-reperfusion injury [31], a previous study reported increased expression of ITGAV in IgAN [8]. These contrasting observations underscore the context- and stage-specific regulation of ITGAV in IgAN, warranting further investigation into its diagnostic and therapeutic implications. Taken together, APP, CD47, and ITGAV emerge as E1-specific biomarkers that have not been fully characterized in IgAN, potentially linking impaired endothelial regulation, dysregulated immune cell trafficking, and altered cell–matrix interactions to endocapillary hypercellularity. Further studies are required to clarify their stage- and context-specific roles in IgAN.

From a clinical and translational perspective, stratifying patients with IgAN according to the presence or absence of endocapillary hypercellularity (E1 vs. E0) may yield valuable insights for therapeutic decision-making. Although the Oxford classification has established the prognostic significance of the E score, its use in guiding treatment strategies remains inconsistent in current practice. In this study, transcriptomic profiling revealed distinct molecular signatures associated with E1 lesions, including downregulation of APP, CD47, and ITGAV. These molecular changes may occur alongside—or even precede—histological alterations, suggesting their potential value as early biomarkers of disease activity. Further validation in larger and more ethnically diverse cohorts is needed to confirm their clinical relevance. Ultimately, such biomarkers could refine risk stratification and help guide future therapeutic approaches in IgAN.

This study has several limitations. First, to reduce variation between samples, six genes (FOXN2, GPX3, IGHG2, IGHG3, IGHG4, and IGKC) were excluded during the quality control process. Some of these genes have been reported as differentially expressed genes in previous IgA nephropathy studies. In particular, IGHG2 and IGKC were reported as downregulated DEGs in comparisons between M1 lesions and controls [8]. However, these findings were based on the M1 classification, which differs from the E0/E1 classification used in our study, limiting direct comparability. GPX3 was reported as a downregulated DEG in a rat model of IgAN, specifically in renal tubule rather than glomerulus [9]. Given the differences in species and tissue context, the relevance of GPX3 to human glomerular E1 lesions remains uncertain. These differences in classification, species, and tissue context indicate that further validation is required to clarify the biological relevance of these genes. Second, AOIs were selected using the segmented profiling method on the GeoMx DSP platform. In this study, immunofluorescence markers were not sufficiently informative for precise AOI delineation. However, glomerular regions were reliably identified based on histomorphologic features, allowing consistent selection of regions of interest across samples. Despite this, reliance on morphology-based segmentation may limit the ability to capture finer spatial heterogeneity or molecular gradients within the glomerulus. Future studies incorporating optimized immunofluorescence staining protocols and higher-resolution imaging approaches are expected to improve AOI definition. In addition, alternative AOI selection strategies such as gridded profiling enable unbiased spatial assessment across entire tissue sections. Integrating spatial transcriptomics with single-cell transcriptomics using matched samples could also provide a more comprehensive and granular understanding of disease mechanisms. Third, independent validation of key spatial transcriptomic findings remains limited. Although exploratory immunohistochemical staining for APP was performed using human biopsy specimens, no significant differences were observed among groups. Further validation using larger cohorts and complementary experimental approaches will be necessary to confirm the biological relevance of these findings in IgAN. Fourth, the study included only East Asian participants, limiting the generalizability and cross-ethnic applicability of the findings. Addressing these limitations through larger, multi-ethnic, and methodologically integrated studies is expected to advance the identification of therapeutic or prognostic biomarkers for the early detection and management of IgAN. Last, although the sex distribution was comparable between groups, strict age matching was not fully achieved. While statistical testing did not identify a significant age difference across groups, the potential influence of aging on the glomerular transcriptome cannot be completely excluded. Therefore, residual confounding may remain in comparisons between IgAN and control groups. Future studies incorporating age- and sex-matched control cohorts will be important to further validate and generalize our findings.

Conclusions

This study applied spatial transcriptomic profiling to characterize molecular alterations associated with the progression of IgAN from E0 to E1. GO and PPI analyses confirmed the involvement of previously reported genes such as FOS, DUSP1, and EGR1, while also identifying novel genes specifically associated with the E1 stage. These findings provide insights into the molecular basis of disease progression and highlight potential biomarkers that may aid in the diagnosis and management of IgAN.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

IgAN

Immunoglobulin A nephropathy

E0

Absence of endocapillary hypercellularity

E1

Presence of endocapillary hypercellularity

AOI

Area of interest

DEG

Differentially expressed gene

GO

Gene ontology

PPI

Protein–protein interaction

CPM

Counts per million mapped reads

FDR

False discovery rate

LFC

Log₂ fold change

NOP9

Nucleolar protein 9

IFI6

Interferon alpha-inducible protein 6

OFCC1

Orofacial cleft 1 candidate 1

DUSP1

Dual-specificity phosphatase 1

EGR1

Early growth response 1

FOS

Fos proto-oncogene

APP

Amyloid beta precursor protein

CD47

Cluster of differentiation 47

ITGAV

Integrin alpha V

Author contributions

Hyeon Do Jeon: Writing – review & editing, Writing – first draft, Visualization, Software, Methodology, Investigation, Formal analysis, Data curation. Kyung Hwan Jeong: Supervision, Formal analysis, Conceptualization. Hyeon Seok Hwang: Supervision, Formal analysis, Conceptualization. Ji-Youn Sung: Supervision, Visualization, Data curation. Kyungsik Ha: Methodology, Investigation, Formal analysis, Supervision. Woochang Choi: Methodology, Investigation, Formal analysis, Data curation. Jin Sug Kim: Writing – review & editing, Writing – first draft, Supervision, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Hwajin Lee: Writing – review & editing, Writing – first draft, Supervision, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Funding

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT; 2022R1F1A1074102).

Data availability

The data are available in the Gene Expression Omnibus (GEO) database under the accession number GSE311745.

Declarations

Ethics approval and consent to participate

The study was approved by the Institutional Review Board of Kyung Hee University Medical Center (approval No. 2023-05-084) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Consent for publication

Not applicable.

Author information

Hyeon Do Jeon is the sole first author, with no co-first authors.

Footnotes

Not applicable.

Competing interests

Hwajin Lee is a co-founder of UPPThera, inc., but this work was not funded by the company.