IgM Hyposialylation Modulates Podocyte Vulnerability in Patients With Idiopathic Nephrotic Syndrome

Abstract

Introduction

Altered Ig glycosylation has been implicated in antibody-mediated podocytopathies; however, the functional relevance of IgM sialylation remains poorly defined. Previous evidence suggests that circulating cationic or hyposialylated IgM may contribute to podocyte vulnerability in idiopathic nephrotic syndrome (iNS).

Methods

Serum IgM from 86 pediatric and adult patients with podocytopathies, 20 patients with membranous nephropathy (MN), 20 patients with lupus nephritis (LN), and 30 healthy controls were analyzed by lectin-based enzyme-linked immunosorbent assay (ELISA) using biotinylated lectins to assess terminal N-glycan residues, including Sambucus nigra agglutinin (SNA) and Ricinus communis agglutinin I (RCA-I). Serum levels of the sialyltransferase ST6GAL1 and the sialidases neuraminidase-1 (NEU1) and neuraminidase-3 (NEU3) were quantified. Cultured human podocytes were exposed to native, desialylated, or resialylated IgM and analyzed by confocal microscopy, quantitative proteomics, phosphoproteomics, and metabolic assays.

Results

IgM from patients with iNS showed reduced SNA binding, which inversely correlated with proteinuria and circulating NEU1/NEU3 levels. In paired samples, SNA reactivity decreased during relapse and increased during remission. ST6GAL1 was undetectable across all groups, whereas phospholipase A2 receptor 1(PLA2R1)-positive MN displayed reduced RCA-I binding. Podocytes exposed to hyposialylated or desialylated IgM exhibited disorganization of the actin cytoskeleton, reduced nephrin signal, increased lipid peroxidation, and decreased ATP levels. Resialylated IgM displayed podocyte morphological and metabolic features not statistically distinguishable from those observed under control conditions. Proteomic and phosphoproteomic analyses highlighted modulation of mitogen-activated protein kinase (MAPK)-, mechanistic Target of Rapamycin (mTOR-), adenosine monophosphate-activated protein kinase (AMPK), and cytoskeleton-related pathways.

Conclusion

IgM sialylation status tracks disease activity and modulates podocyte structural, metabolic, and signaling responses, supporting immune glycan remodeling as a disease-associated modifier of podocyte vulnerability in iNS.

Graphical abstract

iNS represents a heterogeneous group of glomerular disorders characterized by podocyte injury, foot-process effacement, and proteinuria. Despite decades of investigation, the immunopathogenesis of iNS remains only partially understood. Both clinical observations and experimental studies indicate that circulating immune factors can directly alter podocyte structure and function, challenging the traditional distinction between immune-mediated and so-called “nonimmune” forms of iNS. Although the role of IgG autoantibodies has been established in antibody-mediated glomerular diseases such as MN, the contribution of IgM to podocyte injury has received comparatively little attention.

Early studies from our group demonstrated that cationic IgM from patients with iNS binds to the glomerular basement membrane and induces proteinuria when administered in vivo, supporting a potential pathogenic role for these antibodies.¹ More recently, IgM hyposialylation has been identified as a distinctive biochemical signature in pediatric patients with steroid-dependent or frequently relapsing iNS, where hyposialylated IgM persists on T-cell surfaces and is associated with altered steroid responsiveness.² Together, these observations suggest that IgM glycan remodeling may modulate both immune behavior and tissue interactions.

Sialic acid is a terminal monosaccharide that confers a negative charge to glycoproteins, modulating molecular interactions, complement activation, and immune recognition. In the kidney, multiple lines of evidence indicate that impaired sialylation is sufficient to disrupt glomerular perm-selectivity and podocyte architecture. In mice with mutations in the sialic-acid biosynthetic enzyme GNE (UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase), supplementation with N-acetylmannosamine ameliorates glomerular injury.³ Similarly, reduced sialylation of podocyte podocalyxin or secreted angiopoietin-like 4 (ANGPTL4) has been shown to alter glomerular charge selectivity and promote proteinuria.⁴ These data support sialylation as a regulatory element of podocyte integrity at the interface between immune regulation and structural barrier function.

Beyond sialylation, other glycan alterations, such as IgG afucosylation, have recently been linked to antibody-mediated kidney injury. In patients with iNS harboring antinephrin autoantibodies, circulating and urinary IgG displayed markedly reduced fucosylation, which correlates with disease activity and proteinuria.⁵ Thus, antibody glycan remodeling may represent a broader mechanism shaping humoral effector properties in glomerular disease.

In this study, we investigated whether IgM hyposialylation correlates with podocyte injury and metabolic alterations. We combined lectin-based glycoprofiling of IgM from patients with podocytopathies with functional assays in cultured human podocytes exposed to native, desialylated, or resialylated IgM. We also quantified circulating sialidase (NEU1/NEU3) levels in patients’ sera to examine their relationship with IgM desialylation. Using this integrated approach, we tested whether IgM sialylation status may act as a modifiable determinant of podocyte signaling, oxidative stress, and cellular energy metabolism.

Methods

Serum IgM from patients with minimal change disease (MCD), focal segmental glomerulosclerosis (FSGS), MN, LN, and healthy controls (CTR) were analyzed by lectin-based ELISA to assess IgM N-glycan composition. Serum levels of enzymes involved in sialylation and desialylation (ST6GAL1, NEU1, and NEU3) were measured by ELISA. Purified IgM from selected patients and controls, including enzymatically desialylated and resialylated IgM, were used to stimulate cultured human podocytes for confocal imaging, quantitative proteomic and phosphoproteomic analyses, and metabolic assays of oxidative stress and ATP synthesis. Detailed experimental procedures, including patient selection criteria,^5,6 IgM purification and enzymatic modification, lectin- and enzyme-based ELISA conditions,^7,8 cell culture and confocal microscopy,9, 10, 11 proteomic and kinase array workflows,12, 13, 14 and statistical analyses,14, 15, 16 are provided in the Supplementary Material.

Results

Clinical and Biochemical Characteristics of the Study Population

Serum samples were obtained from 86 pediatric and adult patients with biopsy-proven podocytopathies, including 46 with MCD and 40 with FSGS. As summarized in Table 1, patients displayed a wide range of proteinuria, from normal values to overt nephrotic-range levels. According to clinical history and treatment response, patients were classified into steroid-dependent, multidrug-dependent, and multidrug-resistant phenotypes. Consistent with the study design, all patients with MCD and FSGS tested were negative for circulating antinephrin antibodies, and no pathogenic or likely pathogenic variants were identified in the genes included in the diagnostic genetic panel.

Table 1.

Clinical and demographic characteristics of the study cohort

Basic characteristics	FSGS	MCD	MN	LN
Number (%)	40 (47)	46 (53)	20	20
Female (%)	18 (45)	21 (46)	10 (50)	10 (50)
Male (%)	22 (55)	25 (55)	10 (50)	10 (50)
Age median (range) yrs	10 (1–51)	10 (1–51)	45 (31–51)	39 (28–49)
UPCR mg/mg	4 (0.05–13)	3.5 (0.05–9)	4 (1.1–13)	3 (0.5–7)
Serum/urine anti-nephrin positive	0	0	0	0
Serum anti-PLA2R positive	Not tested	Not tested	10	Not tested

FSGS, focal segmental glomerulosclerosis; LN, lupus nephritis; MCD, minimal change disease; MN, membranous nephropathy; PLA2R1, phospholipase A2 receptor 1; UPCR, urine protein-to-creatinine ratio.

Summary of demographic, clinical, and histopathological features of patients with MCD, FSGS, MN (including PLA2R1-positive and -negative cases), and LN. All FSGS and MCD patients included in the study were negative for circulating antinephrin antibodies by study design and showed no pathogenic or likely pathogenic variants in the genes included in the diagnostic nephropathy panel. Values are expressed as median (interquartile range or range) or number (percentage) as appropriate.

Lectin-Based Profiling of N-glycan Residues on Serum IgM

Lectin specificity was verified by pretreating purified serum IgM with N-glycosidase F (PNGase F), which resulted in complete loss of lectin binding (Supplementary Figure S1A–C).

The glycosylation profile of serum IgM was assessed by lectin ELISA using 6 biotinylated lectins with defined glycan specificities as follows: SNA(terminal α2,6-linked sialic acid), Aleuria aurantia lectin (fucose), Ulex europaeus agglutinin I (α1,2-linked fucose), Lotus tetragonolobus lectin (fucose/mannose), RCA-I (β1,4-linked galactose), and Concanavalin A (mannose/hybrid-type residues).

As shown in Figure 1a, SNA binding was markedly reduced in patients with iNS (499 [480–527] RU/ml) compared with CTR (614 [554–654] RU/ml), LN (583 [555–612] RU/ml), and MN (581 [568–591] RU/ml) (P < 0.0001). Within the iNS cohort, patients with FSGS showed lower SNA reactivity than those with MCD (479 [466–496] vs. 517 [499–540] RU/ml; P < 0.0001). When stratified by therapeutic response, SNA reactivity differed among steroid-dependent (478 [459–504]), multidrug-dependent (504 [480–530]), and multidrug-resistant (530 [517–551]) groups (P < 0.05).

Figure 1.

Lectin-based IgM N-glycan profiling and serum neuraminidase levels. (a) Boxplots showing lectin-ELISA results for IgM N-glycan residues across control (30 CTR; white), minimal change disease (46 MCD; magenta), focal segmental glomerulosclerosis (40 FSGS; red), membranous nephropathy PLA2R1 negative (10 MN^-; blue) or positive (10 MN⁺; cyan), and lupus nephritis (20 LN; gray). Binding to SNA was significantly reduced in FSGS and MCD (P < 0.0001), whereas RCA-I binding was lowest in PLA2R1-positive MN (P < 0.0001). (b) Paired analysis of SNA reactivity in ten patients with podocytopathies at 2 clinical time points, during active disease (proteinuria > 3.5 g/d) and complete remission (proteinuria < 0.2 g/d). SNA binding was significantly reduced during the active phase compared with remission (P < 0.01). (c) Heatmap correlogram displaying Pearson correlation coefficients among SNA reactivity, NEU1 and NEU3 serum levels, and urinary protein-to-creatinine ratio (UPCR). The color scale indicates the strength and direction of the correlation: red represents a strong positive, blue a strong negative, and white no correlation; color intensity is proportional to the absolute value of the correlation coefficient (∗∗∗P < 0001; ∗∗P < 0.001; ∗P < 0.05). (d) Boxplots of NEU1 and NEU3 serum concentrations showing significant increases in podocytopathies versus other groups (P < 0.05). ST6GAL1 was undetectable in all samples (data not shown). AAL, Aleuria aurantia lectin; CTR, healthy controls; ELISA, enzyme-linked immune sorbent assay; FSGS, focal segmental glomerulosclerosis; LN, lupus nephritis; LTL, Lotus tetragonolobus lectin; MCD, minimal change disease; MN, membranous nephropathy; NEU1, neuraminidase-1; PLA2R1, phospholipase A2 receptor 1; RCA-I, Ricinus communis agglutinin I; SNA, Sambucus nigra agglutinin; UEA-I, Ulex europaeus agglutinin I; UPCR, urinary protein-to-creatinine ratio.

All lectin signals were normalized to total IgM levels, excluding differences because of IgM concentration. In 10 iNS patients with paired samples obtained during active disease and complete remission, SNA binding was significantly reduced during active proteinuria (470 [455–492] vs. 534 [507–551] RU/ml; P < 0.01; Figure 1b).

Pearson correlation analysis showed that IgM SNA reactivity correlated inversely with circulating NEU1 (R = -0.67, P < 0.001), NEU3 (R = -0.59, P < 0.001), and urinary protein-to-creatinine ratio (R = -0.69, P < 0.0001) (Figure 1c). No significant differences were observed for Aleuria aurantia lectin, Lotus tetragonolobus lectin, Ulex europaeus agglutinin I, or Concanavalin A among groups.

In contrast, RCA-I binding revealed a disease-specific pattern, with PLA2R1-positive patients with MN displaying markedly reduced reactivity (1581 [1555–1594] RU/ml) compared with PLA2R1-negative MN, LN, podocytopathies, and CTR (P < 0.0001; Figure 1a).

Serum Expression Levels of Sialylation- and Desialylation-Related Enzymes in Serum

To examine the relationship between IgM sialylation and circulating enzymes involved in sialylation or desialylation, a direct ELISA was performed to quantify ST6GAL1, NEU1, and NEU3 in serum samples from all study groups. ST6GAL1 levels were below the detection limit in all samples (data not shown).

In contrast, both NEU1 and NEU3 serum levels were significantly higher in patients with iNS (NEU1 = 0.65 [0.53–0.81]; NEU3 = 0.69 [0.59–0.85]) compared with CTR (NEU1 = 0.39 [0.34–0.43]; NEU3 = 0.548 [0.45–0.628]), LN (NEU1 = 0.36 [0.33–0.43]; NEU3 = 0.549 [0.44–0.66]), and MN (NEU1 = 0.366 [0.32–0.41]; NEU3 = 0.569 [0.46–0.69]) (P < 0.05 for both enzymes).

Within the iNS cohort, patients with FSGS exhibited higher serum levels of both NEU1 and NEU3 than patients with MCD (NEU1 = 0.71 [0.59–0.81] vs. 0.62 [0.50–0.80]; NEU3 = 0.72 [0.63–0.88] vs. 0.67 [0.58–0.81]; P < 0.05), mirroring the more pronounced reduction in IgM sialylation observed in FSGS (Figure 1d).

No significant differences in NEU1 or NEU3 levels were detected between patients with LN or MN and CTR subjects. Across the entire cohort, NEU1 and NEU3 serum concentrations were positively correlated (R = 0.75, P < 0.05).

Effects of IgM Sialylation Status on Podocyte Ccytoskeletal Organization and Nephrin Expression Profile

As shown in Supplementary Figure S1D, enzymatic desialylation and resialylation efficiency was validated by SNA lectin ELISA. To exclude residual sialyltransferase or sialidase activity in IgM preparations, CTR-desialylated and CTR-resialylated IgM were tested by ELISA for ST6GAL1, NEU1, and NEU3 with no detectable signal, indicating the absence of enzyme carryover (Supplementary Figure S1D).

To assess whether IgM derived from patients with iNS induces structural alterations in podocytes, conditionally immortalized human podocytes were incubated with IgM purified from 10 independent iNS serum pools or matched CTR pools, processed in parallel under identical conditions. Each iNS pool comprised sera from 8 distinct patients with comparably low IgM sialylation, minimizing interpool variability and ensuring broad cohort representation. In selected experiments, iNS-derived IgM was enzymatically resialylated before podocyte exposure.

Preliminary dose-response and time-course experiments showed concentration- and time-dependent cytoskeletal modulation (Supplementary Figure S2). Based on these analyses, 0.12 μg IgM per well (1:50 serum dilution) and overnight incubation were selected as standard conditions.

As shown in Figure 2a and c, untreated podocytes and podocytes exposed to CTR IgM displayed comparable morphology, actin organization, and adhesion. In contrast, iNS-derived IgM induced marked cytoskeletal rearrangement, characterized by reduced phalloidin fluorescence, cell shrinkage, and decreased cell number (P < 0.0001). CTR-desialylated IgM recapitulated this phenotype, whereas CTR-resialylated IgM did not induce detectable alterations, resulting in actin organization and cell counts comparable to controls. A similar preservation of cytoskeletal architecture was observed following exposure to resialylated iNS-derived IgM.

Figure 2.

Confocal immunofluorescence analysis of podocytes exposed to IgM with different sialylation status. (a, b) Representative confocal images showing the actin cytoskeleton stained with phalloidin–AF488 (green) and nephrin distribution detected with antinephrin–AF568 (red); nuclei are shown in blue (DAPI). Untreated podocytes and podocytes treated with control (CTR) IgM display preserved morphology, organized actin stress fibers, and intense nephrin staining. In contrast, exposure to iNS-derived IgM or enzymatically CTR-desialylated IgM is associated with actin cytoskeleton disorganization, loss of stress fibers, cell shrinkage, reduced adhesion, and a marked decrease in nephrin signal intensity. Podocytes treated with resialylated IgM show actin organization and nephrin localization comparable to those observed under control conditions. Scale bars: 10 μm; (c–d) Quantitative analysis corresponding to panels (a, b), showing total phalloidin–AF488 fluorescence (c) and nephrin–AF568 fluorescence (d) per field, computed using EBImage as the sum of pixel intensities under identical acquisition settings. Quantitative data confirm lower actin and nephrin fluorescence signals in podocytes treated with iNS-derived and desialylated IgM, whereas fluorescence levels in resialylated IgM–treated podocytes are comparable to those measured in control cells. iNS, idiopathic nephrotic syndrome.

Nephrin immunostaining mirrored these findings. Podocytes treated with CTR IgM or resialylated IgM showed preserved nephrin intensity and membrane localization, whereas iNS-derived or CTR-desialylated IgM induced a significant reduction in nephrin signal and altered membrane distribution (P < 0.0001; Figure 2b and d). Nephrin staining in podocytes exposed to resialylated iNS-derived IgM was comparable to control conditions.

IgM Binding to Podocytes is Modulated by Sialylation Status

To assess whether IgM sialylation influences podocyte interaction, IgM binding was analyzed by confocal microscopy. Fixed podocytes were incubated at 4 °C with IgM from patients with iNS or CTR subjects, either untreated, desialylated, or resialylated.

As shown in Figure 3, CTR IgM displayed limited, focal binding, whereas iNS-derived IgM showed markedly increased and diffused binding. CTR IgM desialylation recapitulated the binding pattern observed with iNS IgM, whereas enzymatic resialylation markedly reduced IgM binding, resulting in fluorescence intensity and distribution not statistically distinguishable from CTR conditions. Importantly, enzymatic resialylation of iNS-derived IgM similarly reduced podocyte binding, resulting in fluorescence levels not statistically distinguishable from those observed with native control IgM.

Figure 3.

IgM binding to podocytes is modulated by sialylation status. (a) Representative confocal images showing binding of human IgM to podocytes. After fixing, podocytes are incubated with IgM from CTR subjects or patients with iNS, either untreated or enzymatically desialylated, or resialylated. IgM derived from CTR shows limited, discrete binding, whereas IgM derived from patients with iNS is associated with a stronger, more diffuse podocyte signal. Enzymatically desialylated IgM from CTR results in a binding pattern comparable to that observed with IgM from untreated iNS. In contrast, resialylated IgM shows reduced binding, resulting in fluorescence intensity and distribution not statistically distinguishable from CTR conditions. IgM binding was detected using anti-human IgM–FITC immunofluorescence (green); nuclei were counterstained with DAPI (blue). (b) Quantitative image analysis demonstrated increased fluorescence of podocyte-associated IgM–FITC signal in podocytes exposed to iNS-derived or desialylated IgM compared with CTR or resialylated IgM (P < 0.0001). Each experimental condition was analyzed in three independent biological replicates. CTR, healthy controls; iNS, idiopathic nephrotic syndrome.

Quantitative image analysis confirmed significantly increased IgM signal in podocytes exposed to iNS-derived or CTR-desialylated IgM compared with CTR, CTR-resialylated, or iNS-resialylated IgM (P < 0.0001).

Podocyte Proteome Profile

Label-free quantitative proteomic analysis identified 6159 proteins across all samples. Principal component analysis of the whole proteome dataset revealed clear separation among podocytes treated with CTR IgM, iNS IgM, or CTR-desialylated IgM, indicating distinct global proteomic profiles (Figure 4a).

Figure 4.

Proteomic analysis of podocytes exposed to IgM of different sialylation states. (a) Principal component analysis (PCA) showing distinct clustering of podocytes treated with CTR, iNS, or CTR-Desialylated IgM. (b) Venn diagram summarizing the overlap of statistically significant proteins identified by ANOVA and unpaired t tests. Numbers and circles represent the distinct statistically significant proteins in each comparison, respectively. (c-e) Volcano plots comparing (c) iNS versus CTR, (d) CTR-Desialylated versus CTR, and (e) CTR-Desialylated versus iNS. In the volcano plot, the x-axis reports log₂ fold change, and the y-axis: –log₁₀P-value. Black dots indicate nonsignificant proteins; red and blue denote significantly up- or down-regulated proteins, respectively. Each experimental condition was analyzed in four independent biological replicates. CTR, healthy controls; iNS, idiopathic nephrotic syndrome.

An analysis of variance test for unpaired samples identified 5250 proteins significantly modulated across conditions (Supplementary Table S1, Figure 4b). Subsequent unpaired t tests showed that iNS IgM treatment modulated 363 proteins compared with CTR IgM, whereas CTR-desialylated IgM induced a broader response, with 3037 proteins differing from CTR IgM–treated cells. Direct comparison between CTR-desialylated and iNS IgM identified 3411 differentially expressed proteins, highlighting marked differences between partial and complete loss of IgM sialylation (Supplementary Table S1, Figure 4b–e).

Gene Ontology Enrichment Analysis

To characterize the biological processes associated with the observed proteomic changes, gene ontology enrichment analysis was performed on statistically significant proteins from the 3 pairwise comparisons. This analysis identified 38 significantly enriched gene ontology terms (Supplementary Table S2), grouped into 5 major biological clusters as follows: MAPK signaling and stress response, cell–cell and cell–matrix interactions, mitochondrial and metabolic processes, oxidative stress, and inflammatory signaling (Figure 5).

Figure 5.

Gene ontology (GO) enrichment analysis of differentially expressed proteins. Each circle represents a significantly enriched GO biological process term. The x-axis indicates the fold enrichment score; circle size is proportional to the number of associated proteins (range: 3–39), and color intensity reflects statistical significance (–log₁₀ Adj P-value; white = 1.3, red = 38). Enriched functional clusters include MAPK and stress-response signaling, cytoskeletal organization, mitochondrial and metabolic pathways, oxidative stress, and inflammatory signaling. MAPK, mitogen-activated protein kinase.

Enriched pathways included coordinated modulation of intracellular signaling (MAPK, Ras, mTOR, and AMPK), cytoskeletal organization and cell adhesion, mitochondrial and redox metabolism, and multiple stress-response mechanisms, including autophagy, mitophagy, cellular senescence, and ferroptosis. Immune- and inflammation-related pathways (e.g., TNF, IL-17, Toll-like receptor, NOD-like receptor, HIF-1, TGF-β, and AGE–RAGE signaling) were also represented.

Importantly, all 38 gene ontology terms were detected across the 3 comparisons, with differences in enrichment magnitude, statistical significance, and number of associated proteins, indicating a shared biological framework whose modulation varied with IgM sialylation status.

Phosphoproteomic and Kinase Activity Analysis

To validate and extend the proteomic findings (Figure 6a), phosphorylation pathway profiling was performed in podocytes exposed to CTR, iNS, CTR-desialylated, or CTR-resialylated IgM. Distinct phosphorylation signatures were observed across conditions, consistent with the differences detected at the total proteome level (Figure 6b).

Figure 6.

Phospho-array and kinase activity profiling of podocytes exposed to IgM with different sialylation status. (a) The entire kinase family identified by mass spectrometry in podocytes treated with IgM derived from iNS, control (CTR), CTR-desialylated, or CTR-resialylated preparations is shown. Each colored node represents an individual kinase whose change in expression among the four conditions is either statistically significant (red) or not (gray). Node size is proportional to the −log10 P-value. (b) Heatmap of phospho-array signals showing relative phosphorylation levels in podocytes treated with control (CTR), iNS, CTR-desialylated, and CTR-resialylated IgM. Color intensity represents normalized signal intensity for each phospho-site. In the heatmap, the rows and columns correspond to the kinases and the experimental conditions, respectively. Each experimental condition was analyzed in three independent biological replicates. iNS, idiopathic nephrotic syndrome.

Kinase-Substrate Enrichment Analysis revealed differential modulation of kinase activities. In podocytes treated with iNS IgM, MAPK-related kinases (ERK1/2, p38, and JNK) showed increased activity, accompanied by reduced activity of the AKT–mTOR axis. CTR-desialylated IgM induced a broader and more pronounced response, with increased inferred activity of MAPK14, MAPK8, PRKCA, and SRC, and decreased activity of AKT1, GSK3β, and PAK1.

In contrast, CTR-resialylated IgM produced a kinase activity profile comparable to control conditions, with no significant differences relative to CTR IgM–treated cells. Overall, graded changes in phosphorylation-dependent signaling pathways were observed across conditions differing in IgM sialylation status (Supplementary Figure S4).

Hyposialylated IgM Increases Lipid Peroxidation and Reduces ATP Synthesis in Podocytes

To determine whether IgM sialylation–dependent signaling alterations were associated with metabolic changes, lipid peroxidation, and ATP synthesis were measured in podocytes exposed to IgM with different sialylation status. Cells were treated with IgM purified from patients with CTR, iNS, MN (anti-PLA2R negative), and LN, as well as with CTR or iNS IgM subjected to enzymatic desialylation or resialylation. All IgM preparations were processed under identical conditions and extensively washed to minimize enzyme or contaminant carryover.

As shown in Figure 7a, intracellular malondialdehyde (MDA) levels were significantly increased in podocytes treated with CTR-desialylated IgM compared with all other conditions (P < 0.0001). In contrast, MDA levels remained comparable to controls in podocytes exposed to CTR- or iNS-resialylated IgM, as well as to IgM from CTR, MN, or LN sera, whereas iNS IgM induced intermediate MDA levels.

Figure 7.

Effects of IgM sialylation status on oxidative stress and ATP production. (a) Boxplots showing malondialdehyde (MDA) concentrations, indicating higher levels of lipid peroxidation in podocytes exposed to iNS-derived IgM and enzymatically desialylated control IgM compared with CTR IgM. MDA levels measured in podocytes treated with resialylated iNS IgM, as well as IgM derived from patients with MN and LN, are comparable to those observed under control conditions (P < 0.0001). (b) Boxplots showing cellular ATP levels, which are lower in podocytes treated with iNS-derived IgM and enzymatically desialylated control IgM. In contrast, ATP levels measured in podocytes exposed to resialylated iNS IgM, CTR-derived IgM, and IgM from patients with MN and LN are comparable to those observed under control conditions (P < 0.0001). Values are expressed as nmol/mg total protein; each condition was analyzed using ten independent biological replicates, corresponding to ten distinct IgM pools, each generated by combining sera from different patients with comparable IgM sialylation levels. Individual patients were not repeated across pools.

A complementary pattern was observed for ATP synthesis (Figure 7b). Desialylated IgM markedly reduced ATP levels, whereas CTR- or iNS-resialylated IgM restored ATP concentrations to values comparable to CTR IgM (P < 0.0001). MN- and LN-derived IgM did not significantly affect ATP levels, whereas iNS IgM again showed intermediate effects.

These findings were consistent across biological replicates from 10 independent iNS IgM pools, indicating graded metabolic alterations associated with IgM sialylation status, with desialylated IgM exerting the strongest effects and resialylated IgM resembling control conditions.

Immunoprecipitation

To investigate IgM–podocyte protein interactions, membrane- and cytoplasm-enriched fractions from untreated podocytes were incubated with IgM purified from patients with CTR or iNS, followed by immunoprecipitation and mass spectrometry. A total of 4371 proteins were identified, and, after normalization and removal of nonspecific binders, enrichment analysis was performed separately for plasma membrane and cytoplasmic fractions.

In the plasma membrane–enriched fraction, 58 proteins were significantly enriched. Based on annotation against a curated mass spectrometry–derived cell surface protein atlas,¹⁷ 42 proteins (72.4%) were classified as membrane or cell surface proteins, with no nuclear proteins detected (Supplementary Figure S5A, Supplementary Table S3). In parallel, 75 proteins were significantly enriched in the cytoplasmic fraction (Supplementary Figure S5B, Supplementary Table S3). No overlap was observed between proteins enriched in the membrane and cytoplasmic fractions.

Functionally, enriched proteins clustered into broad categories. Membrane-associated proteins were mainly involved in membrane organization and vesicular trafficking, cell adhesion and cytoskeletal organization, extracellular matrix/secreted components, metabolic and redox-associated processes, and signaling or immune-related functions. Cytoplasmic interactors were enriched for cytoskeletal organization, metabolic enzymes, protein turnover and post-translational regulation, signaling and immune-related regulators, and proteins linked to RNA/DNA processing, cell cycle control, and stress responses.

Discussion

In this study, we identify IgM hyposialylation as a disease-associated and potentially modifiable feature associated with increased podocyte vulnerability. Patient-derived IgM with reduced SNA reactivity, and more markedly enzymatically desialylated IgM, induced actin disorganization, loss of stress fibers, reduced adhesion, decreased nephrin expression, increased lipid peroxidation, and diminished ATP synthesis. In contrast, enzymatic resialylation of IgM resulted in structural and metabolic features not statistically distinguishable from those observed under control conditions.

Our findings extend and mechanistically link 2 previously independent lines of evidence. First, IgM has been implicated as a pathogenic effector in podocytopathies— Musante et al.¹ demonstrated that anti-actin/β-ATP-synthase IgM from patients with iNS is cationic, deposits in glomeruli, and induces proteinuria in vivo, supporting a direct antibody-mediated mechanism of podocyte injury. Second, IgM hyposialylation on T cells identifies a severe steroid-dependent iNS phenotype— Colucci et al.² showed that less-sialylated IgM accumulates on T-cell surfaces, resists internalization, abrogates steroid-induced inhibition, and promotes the release of podocyte-damaging factors, with total IgM sialylation inversely correlating with T-cell–bound IgM.

Our data help unify these observations by demonstrating that the sialylation state of IgM per se modulates podocyte responses in vitro, supporting the concept that antibody glycosylation represents an additional layer of regulation of antibody-tissue interactions. Recent studies have similarly highlighted how distinct Ig glycosylation patterns can shape antibody effector properties and disease phenotypes in immune-mediated kidney diseases.¹⁸

To further explore the molecular basis of this interaction, immunoprecipitation-based proteomic analyses were performed using membrane- and cytoplasm-enriched podocyte fractions. These experiments were not designed to identify a single cognate receptor, but rather to characterize the spectrum of podocyte proteins preferentially associated with patient-derived IgM under native conditions. Notably, more than 70% of proteins enriched in membrane-associated fractions were experimentally annotated as membrane-related, whereas cytoplasmic interactors represented a distinct and nonoverlapping set. Functional annotation highlighted pathways related to membrane organization, vesicular trafficking, cell adhesion, cytoskeletal dynamics, extracellular matrix interactions, and redox- and signaling-related processes. Together, these findings support a model in which IgM hyposialylation enhances interactions with membrane-associated podocyte protein complexes through glycan- and charge-dependent mechanisms.

The importance of sialic acid metabolism in maintaining glomerular perm-selectivity is well established. Genetic deficiency of GNE (UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase), the key enzyme in sialic-acid biosynthesis, causes severe proteinuria in mice and is rescued by N-acetylmannosamine supplementation, demonstrating that deficient sialylation is sufficient to disrupt the filtration barrier.³ Selective desialylation of podocyte glycoproteins, including podocalyxin, similarly induces foot-process effacement and FSGS-like lesions.^19,20 Moreover, altered sialylation of angiopoietin-like-4 modulates its isoelectric point and glomerular charge selectivity, correlating with proteinuria in steroid-sensitive models.⁴ Importantly, the glomerular filtration barrier is an integrated structure composed of endothelial glycocalyx, glomerular basement membrane, and podocytes, all of which are rich in sialylated glycoconjugates; the present in vitro model specifically interrogates podocyte responses and does not define where IgM desialylation occurs in vivo.

At the signaling level, phosphoproteomic and Kinase-Substrate Enrichment Analysis analyses revealed activation of MAPK pathways (ERK, p38, and JNK) with concomitant down-regulation of AKT/mTOR signaling. Importantly, modulation of the AKT/mTOR axis should not be interpreted as a linear surrogate of podocyte stress severity. Accumulating evidence indicates that the relationship between mTOR signaling and podocyte injury is highly context-dependent and nonlinear. Sustained hyperactivation of mTORC1 in podocytes represents a maladaptive stress response, leading to cytoskeletal disorganization, impaired autophagy, podocyte loss, and progressive proteinuria.²¹ Conversely, excessive genetic or pharmacological suppression of mTOR signaling is also detrimental, as a basal level of mTOR activity is required for podocyte metabolic homeostasis, adaptive stress responses, and survival.²² In this framework, the signaling changes observed in response to hyposialylated IgM are best interpreted as context-dependent adaptations of stress-response pathways rather than as a direct measure of injury intensity.

The observed association between IgM hyposialylation and increased serum neuraminidase abundance provides a potential mechanistic link. NEU3, a plasma-membrane sialidase, removes sialic acids from gangliosides and modulates β1-integrin trafficking and EGFR activation, thereby influencing adhesion and MAPK/PI3K signaling.²³ Similarly, NEU1, the major lysosomal neuraminidase, regulates growth-factor receptor desialylation and renal homeostasis.²⁴ Although the elevation of NEU1 and NEU3 in patient sera does not demonstrate circulating enzymatic activity per se, it aligns with growing evidence that neuraminidases are dynamically upregulated during inflammatory states. Further studies will be required to determine the cellular sources and in vivo contribution of neuraminidase activity in the glomerular environment.

Indeed, NEU1 and NEU3 are increasingly recognized as active mediators of inflammation-driven desialylation. In idiopathic pulmonary fibrosis, NEU3 expression and activity are markedly increased in fibrotic lesions and bronchoalveolar lavage fluid, paralleling reduced α2,6-linked sialylation and enhanced SNA reactivity; pharmacologic or genetic inhibition of NEU3 attenuates inflammation and fibrosis.^25,26 NEU1 is similarly upregulated in activated microglia during neuroinflammation, promoting desialylation of neuronal surface glycoproteins and amplification of oxidative and metabolic stress.²⁷ Additional evidence from airway epithelial cells and experimental colitis further supports a role for neuraminidase-driven desialylation in tissue injury and immune activation.^28,29 Collectively, these studies define an inflammation-driven neuraminidase axis that parallels the pattern observed in our patients and provides a mechanistic bridge between systemic inflammatory activation, altered serum neuraminidase levels, and IgM hyposialylation–dependent podocyte injury.

Consistent with this model, serum NEU1 and NEU3 abundance correlated inversely with IgM sialylation and directly with proteinuria, despite the absence of direct enzymatic activity measurements. Longitudinal analysis of paired samples further demonstrated reduced IgM sialylation during active disease and normalization during remission, supporting the concept that IgM sialylation is dynamically modulated with disease activity. ST6GAL1 was not detected in serum, consistent with its predominant intracellular Golgi localization, and does not contradict a systemic imbalance favoring desialylation. Rather, IgM hyposialylation may reflect exposure to desialylating environments within immune or tissue compartments, with modified IgM subsequently detectable in the circulation.

Functionally, enzymatically desialylated IgM recapitulated and intensified the podocyte alterations observed with patient-derived IgM in vitro, whereas resialylated IgM restored cytoskeletal and metabolic features comparable with controls. These findings align with classical observations that cationic or desialylated proteins display enhanced interaction with the glomerular filtration barrier and increased proteinuric potential.³⁰

Distinct glycan signatures also emerged across glomerular diseases. Antinephrin–negative podocytopathies were characterized by selective IgM hyposialylation, whereas PLA2R1-positive MN displayed reduced IgG galactosylation, consistent with previous reports.³¹ Together with IgG afucosylation in antinephrin–positive nephrotic syndrome,⁵ these observations delineate 3 converging modes of immune glycan remodeling that shape antibody effector properties and podocyte injury mechanisms.

This study has limitations. The in vitro system cannot fully recapitulate the complexity of the in vivo environment, including complement activation and systemic mediators. Pooling IgM samples reduced variability but may have masked patient-specific glycan differences, and lectin-based assays do not resolve linkage- or branch-specific glycan structures. Future studies employing liquid chromatography-mass spectrometry/mass spectrometry glycoproteomics and kidney tissue analyses will be required to define the precise glycan architecture and anatomical site of IgM desialylation.

In conclusion, IgM hyposialylation, in the context of increased serum neuraminidase abundance, is linked to MAPK-associated cytoskeletal remodeling and mitochondrial dysfunction in podocytes. Rather than acting as a primary pathogenic driver, IgM hyposialylation emerges as a disease-associated modifier that amplifies podocyte vulnerability within an immune-activated milieu. These findings provide mechanistic insight into how antibody glycosylation may influence podocyte responses and suggest that targeting the NEU–sialic acid axis could represent a potential therapeutic strategy.

Disclsoure

All the authors declared no competing interests.

Supplementary Material

Supplementary Methods.

Figure S1. Validation of IgM purity, enzymatic desialylation/resialylation efficiency, absence of residual enzyme activity, and lectin specificity.

Figure S2. Optimization of purified IgM treatment conditions on human podocytes.

Figure S3. Two-dimensional plot of ANOVA-significantly modulated proteins in podocytes.

Figure S4. Kinome enrichment analysis.

Figure S5. Volcano plot analysis of proteins associated with hyposialylated IgM in podocyte subcellular fractions.

Table S1. List of significantly modulated proteins identified by quantitative proteomics.

Table S2. Gene Ontology (GO) enrichment analysis of significantly modulated proteins.

Table S3. List of significantly enriched proteins by immunoprecipitation.