Ultrasensitive Immunoassay Using a Novel Galactose-Deficient IgA1 Antibody and Its Clinical Application in the Diagnosis of IgAN

Abstract

Introduction

IgA nephropathy (IgAN) diagnosis remains dependent on invasive renal biopsy because currently available serologic biomarkers, including galactose-deficient IgA1 (Gd-IgA1), have limited diagnostic performance. This study aimed to develop and evaluate a high-sensitivity time-resolved fluorescence immunoassay (TRFIA) based on a novel anti-Gd-IgA1 monoclonal antibody (i.e., MH33) for Gd-IgA1 detection in IgAN.

Methods

MH33 glycoepitope specificity was characterized using surface plasmon resonance (SPR). An MH33-based TRFIA was developed and evaluated for analytical and diagnostic performance in healthy controls (HCs), non-IgAN renal disease controls, and patients with IgAN, with comparison to a KM55-based assay.

Results

SPR analysis showed that MH33 recognized a GalNAc-modified IgA1 hinge-region glycopeptide distinct from the KM55 epitope. The MH33-based TRFIA achieved a detection limit of 0.20 response units (RU)/ml, a quantification range of 0.20 to 600 RU/ml, and acceptable analytical performance (intra-/inter-assay coefficients of variation: 2.32%–5.49% and 3.97%–5.23%; recoveries: 89.80%–103.76%). Correlation with KM55 was moderate (IgAN, r = 0.23, P = 0.09). For distinguishing IgAN from combined controls (HCs and non-IgAN renal diseases), the MH33-based TRFIA yielded an area under the curve (AUC) of 0.90 (95% confidence interval [CI]: 0.86–0.93), compared with 0.78 (95% CI: 0.71–0.86) for KM55. At the optimal cutoff value (22.0 RU/ml), the assay achieved a specificity of 93.82% and sensitivity of 64.44%.

Conclusion

MH33-based TRFIA provides a sensitive approach for Gd-IgA1 detection and demonstrates favorable diagnostic performance for IgAN.

Graphical abstract

IgAN is the most common primary glomerular disease worldwide,^1,2 with approximately 20% to 40% of affected individuals progressing to end-stage kidney disease within 20 years of diagnosis.^3,4 Despite decades of investigation, substantial challenges remain in the diagnosis, treatment, and prognostic monitoring of IgAN; and the disease continues to impose a considerable global health burden. Renal biopsy remains the current “gold standard” for diagnosis, because it provides essential histopathological information to guide clinical decision-making.^5,6 However, biopsy is invasive, carries multiple contraindications, and is associated with the risk of complications, limiting its broader clinical applicability. Although biomarkers such as proteinuria and estimated glomerular filtration rate assist in risk stratification, they lack disease specificity and often change only after irreversible renal damage has occurred.⁷ Therefore, disease-specific biomarkers reflecting the underlying pathogenic process are urgently needed to improve early detection, reduce reliance on invasive testing, and facilitate dynamic disease monitoring.

The “multi-hit” hypothesis has substantially advanced our understanding of IgAN pathogenesis by emphasizing its complex and multistep nature.^8,9 Aberrant O-glycosylation of IgA1, generation of antiglycan autoantibodies, formation of circulating immune complexes, and subsequent mesangial deposition with complement activation constitute a sequential cascade of pathogenic events.¹⁰ Among these processes, Gd-IgA1, the initiating molecular abnormality, is considered both a pathogenic hallmark and one of the most promising biomarker candidates for IgAN. Gd-IgA1 results from deficient galactosylation of O-glycans within the IgA1 hinge region, leading to increased exposure of GalNAc-containing structures that promote autoantibody recognition and immune complex formation. Importantly, increasing evidence suggests that Gd-IgA1 does not represent a single molecular species but rather a heterogeneous population of glycoforms with variable site-specific O-glycosylation patterns, which may differ in biological activity and disease relevance.

Several analytical approaches have been developed for Gd-IgA1 detection, although important limitations remain. Early studies primarily relied on Helix aspersa agglutinin–based lectin assays, which provided initial proof of concept but suffered from substantial nonspecific binding and batch variability.¹¹ The introduction of the KM55 monoclonal antibody–based enzyme-linked immunosorbent assay (ELISA) in 2015 represented an important advancement and has since become the most widely used approach for Gd-IgA1 quantification.¹² Compared with lectin-based assays, KM55-based detection exhibits improved stability and reproducibility; however, substantial overlap in Gd-IgA1 levels among healthy individuals, patients with IgAN, and those with other renal diseases, remains a challenge for broader clinical application.¹³ One possible explanation is the highly heterogeneous nature of Gd-IgA1 O-glycosylation, raising the possibility that antibodies recognizing distinct glycoepitopes may capture different disease-associated subsets of circulating Gd-IgA1 and therefore provide complementary clinical information.

Given the potential heterogeneity of Gd-IgA1 glycosylation, the identification of antibodies recognizing distinct disease-relevant glycoepitopes may improve diagnostic performance. Recent advances in single-cell reverse transcription polymerase chain reaction technology have substantially enhanced the screening throughput and efficiency of high-affinity antibody discovery compared with traditional hybridoma methods. In parallel, TRFIA offers high analytical sensitivity, low background signal, wide dynamic range, and compatibility with automated detection systems, making it particularly suitable for low-abundance biomarker quantification.14, 15, 16, 17, 18, 19, 20 Previous studies have also demonstrated the advantages of TRFIA in improving diagnostic performance for clinically relevant biomarkers.

In this study, we developed a human–mouse chimeric monoclonal antibody, MH33, recognizing a distinct GalNAc-associated glycoepitope within the IgA1 hinge region. We established an MH33-based TRFIA platform and systematically evaluated its glycoepitope specificity, analytical performance, and clinical utility in patients with IgAN. In particular, we investigated the relationship between MH33- and KM55-based detection and explored whether recognition of distinct Gd-IgA1 glycoepitopes may improve discrimination between IgAN and disease controls, thereby providing new insights into the heterogeneity and clinical relevance of Gd-IgA1 glycoforms in IgAN.

Methods

Materials

The anti-Gd-IgA1 monoclonal antibody (clone MH33, isotype: Human-mouse chimeric antibody, IgG1 type) was generously provided by Peking University First Hospital (Beijing, China). Specifically, Balb/c mice were immunized with an i–synthesized human IgA1 hinge–region glycopeptide bearing GalNAc modifications and conjugated to KLH (KLH-C-VPSTPPTPSPSTPPT∗PSPS-NH2), from which the Gd-IgA1–specific monoclonal antibody MH33 was generated. The goat antihuman IgA1 polyclonal antibody, iron oxide polystyrene nanospheres, Eu³⁺-labeled N1-(p-isothiocyanatobenzyl)-diethylenetriamine-N1,N2,N3,N4-tetraacetate, and enhancer solution were obtained from Zhejiang Boshi Biotechnology Co., Ltd. (Hangzhou, China). Tris(hydroxymethyl)aminomethane, bovine serum albumin, and Triton X-100 were purchased from Sigma-Aldrich. Sephadex G-50 was sourced from Amersham Pharmacia Biotech (USA). EDC and NHS were acquired from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). The MES buffer was procured from Yuanye Bio, and the Ultracel IM-50 ultrafiltration tube was from Millipore. A peristaltic pump and an automatic fraction collector were supplied by Shanghai Qingpu Huixi Instrument Factory (Shanghai, China). Ninety-six–well microplates were purchased from Xiamen Yunpeng Technology Development Co., Ltd. A TRFIA analyzer was provided by Shanghai Hu Gu Science Instrument Co., Ltd. SPR Biacore 8K⁺ (12914229, Cytiva). All other chemical reagents used were of analytical grade, and all instruments were used according to the manufacturers' instructions.

Patients

Serum samples for this study were obtained from 268 individuals randomly recruited from the Department of Nephrology at Peking University First Hospital. The cohort included 90 patients with biopsy-confirmed IgAN, 88 patients with other biopsy-confirmed non-IgAN renal diseases (non-IgAN), and 90 HCs. The non-IgAN renal disease control group comprised patients with membranous nephropathy (n = 37), minimal change disease (n = 17), and diabetic nephropathy (n = 34). Serum samples from patients were collected at the time of renal biopsy or before the initiation of immunosuppressive therapy. All blood samples were centrifuged immediately after collection at 1000g for 15 minutes, and the resulting supernatants were collected for subsequent analysis. All participants provided written informed consent. The study protocol was approved by the Ethics Committee of Peking University First Hospital (Approval No.: 2025-124).

Conjugation of Antibodies to Magnetic Beads

A total of 10 mg of magnetic beads was transferred into a centrifuge tube and placed on a magnetic separation rack for approximately 1 minute to remove the supernatant. The beads were resuspended in 200 μl of MES buffer using a vortex mixer. NHS and EDC were prepared at concentrations of 10 mg/ml, and for every 10 mg of beads, 28 μl of NHS and 40 μl of EDC were added for activation. The mixture was incubated with gentle rotation at room temperature for 30 minutes. After activation, the tube was placed on the magnetic separation rack for 1 minute to remove the supernatant, and this washing step was repeated twice with MES buffer.

The activated beads were then resuspended in MES buffer, and the designated antibody was added. The mixture was vortexed thoroughly to ensure complete suspension and incubated overnight under gentle rotation to allow conjugation. Following incubation, the beads were placed on the magnetic rack for 1 minute, the supernatant was removed, and the beads were washed twice with blocking buffer. After the final wash, the beads were resuspended in blocking buffer and incubated at room temperature for 2 hours to complete the blocking process.

Preparation of Eu³⁺-Labeled Antibodies

A total of 0.3 mg of goat antihuman IgA antibody was buffer-exchanged to sodium carbonate buffer (pH: 9.0) using a 50-kDa ultrafiltration membrane. The filtered antibody was then mixed with isothiocyanatobenzyl-DTTA–Eu³⁺ and incubated at 25 °C for 16 hours to allow conjugation. The reaction mixture was subsequently separated on a Sephadex G-50 column, and fractions of 2 ml were collected. Each fraction was analyzed using a TRFIA analyzer, and the first fluorescence peak was collected as the purified Eu³⁺-labeled goat antihuman IgA antibody.

Analytical Principle and Procedure

As illustrated in Figure 1, the automated assay procedure was as follows: MH33-conjugated magnetic beads were first dispensed into the reaction wells. A magnetic field was applied automatically to immobilize the beads, and the supernatant was removed. Diluted samples were then added and incubated for 18 minutes. After incubation, the magnetic field was reapplied to secure the beads, followed by washing to remove residual sample components. Next, diluted Eu³⁺-labeled antihuman IgA antibody was added and incubated for another 18 minutes. Following a final washing step to eliminate unbound conjugates, an enhancement solution was added to dissociate Eu³⁺ ions, producing a measurable fluorescence signal.

Figure 1.

Schematic overview of MH33 epitope characterization, assay principle, analytical performance, and diagnostic evaluation of the MH33-based magnetic separation TRFIA for Gd-IgA1 detection in IgAN. CV, coefficient of variation; Gd-IgA1, galactose-deficient IgA1; SPR, surface plasmon resonance; TRFIA, time-resolved fluorescence immunoassay.

Evaluation of Methodological Performance

The methodological evaluation included assessments of analytical sensitivity, precision, and accuracy. Sensitivity was determined by measuring the 0-concentration sample 10 consecutive times, and the mean + 2 SDs was defined as the limit of detection. Precision was evaluated by repeatedly testing serum samples with low, medium, and high concentrations 10 times/d over 3 consecutive days, and both intraassay and interassay coefficients of variation were calculated. Accuracy was assessed through recovery experiments. Briefly, standard solutions of different concentrations were spiked into low-concentration serum samples at a ratio of 1:9. After repeated measurements, the recovery rate was calculated as the ratio of the measured concentration to the theoretical concentration.

SPR Analysis

SPR experiments were performed on a Biacore 8K+ system (Cytiva, Cat. No. 12914229). Biotinylated glycopeptides or native peptides were immobilized on a streptavidin-coated sensor chip. After a 60-second immobilization step, serial dilutions of antibodies (0.78125, 3.125, 6.25, 12.5, 25, and 50 nM) were injected over the chip surface. The association phase was set to 120 seconds, followed by a dissociation phase of 600 seconds. Response signals (in RU) generated during molecular interactions were recorded in real time at room temperature using Biacore 8K Control Software (v4.0.8.19879, Cytiva). The raw sensorgrams were further analyzed using Biacore 8K Evaluation Software 3.0 (v4.0.8.20368, Cytiva). Binding kinetics were fitted using a 1:1 Langmuir binding model to determine the equilibrium dissociation constant (KD).

Statistics

All statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, San Diego, CA). Data are presented as mean ± SD unless otherwise specified. The Mann–Whitney test was used to determine whether significant differences existed between 2 groups. Receiver operating characteristic curve analysis was conducted to evaluate the diagnostic performance of Gd-IgA1 detection, and the AUC with corresponding 95% CIs were reported as a measure of test accuracy. A p-value < 0.05 was considered statistically significant. All statistical tests were 2-tailed.

To internally validate the diagnostic performance, a 10-fold cross-validation procedure was performed. In each iteration, 9 folds were used to determine the optimal cutoff based on the Youden index, and the remaining fold was used for validation. All statistical analyses were performed using R (version 4.3.2).

Results

Characterization of MH33 Glycoepitope Specificity

As shown in Figure 2a, to further characterize the binding specificity of MH33, SPR analysis was performed using synthetic IgA1 hinge–region glycopeptides carrying site-specific GalNAc modifications. MH33 demonstrated strong binding toward the glycopeptide containing the VPSTPPTPSPSTPPT(GalNAc)PSPS motif (Figure 2b). In contrast, no measurable binding was observed toward the corresponding nonglycosylated peptide (Figure 2c), indicating that glycosylation is essential for antibody recognition.

Figure 2.

Characterization of MH33 and KM55 recognition of glycosylated peptides derived from the IgA1 hinge region by surface plasmon resonance. (a) Schematic illustration of the Gd-IgA1 hinge–region glycopeptide recognized by MH33. (b) Representative SPR sensorgrams of MH33 binding to the corresponding glycosylated hinge-region peptide. (c) Representative SPR sensorgrams of MH33 binding to the corresponding nonglycosylated hinge–region peptide. (d) Schematic illustration of the Gd-IgA1 hinge–region glycopeptide preferentially recognized by KM55. (e) Representative SPR sensorgrams of KM55 binding to the corresponding glycosylated hinge-region peptide. (f) Representative SPR sensorgrams of KM55 binding to the corresponding nonglycosylated hinge–region peptide. Gd-IgA1, galactose-deficient IgA1; SPR, surface plasmon resonance.

By comparison, KM55 preferentially recognized a distinct glycosylated motif, VPSTPPTPSPST (GalNAc) PPTPSPS (Figure 2d), suggesting that MH33 and KM55 target different Gd-IgA1–associated glycoepitopes within the IgA1 hinge region (Figure 2e and f). These findings support that MH33 recognizes a distinct GalNAc-associated glycoepitope rather than simply replicating the recognition pattern of KM55.

SPR kinetic analysis further demonstrated that MH33 exhibited slow association and slow dissociation characteristics. Fitting using a 1:1 binding model yielded an equilibrium dissociation constant (KD) of 8.69 × 10⁻¹⁴ M, indicating high-affinity binding toward the target glycopeptide.

Optimization of Experimental Conditions

To achieve optimal assay performance, key experimental parameters were systematically optimized, including conjugation temperature and duration, antibody-to-bead ratio, working concentration of antibody-conjugated magnetic beads, and dilution of Eu³⁺-labeled goat antihuman IgA antibody.

As shown in Figure 3a, conjugation performed at 20 °C and 37 °C produced comparable fluorescence signals; however, extending the conjugation time from 4 hours to 18 hours significantly improved signal intensity (P < 0.001). Therefore, 20 °C and 18 hours were selected as the optimal conjugation conditions.

Figure 3.

Optimization of the galactose-deficient IgA1–time-resolved fluorescence immunoassay conditions (a) optimization of reaction temperature and time when conjugation magnetic beads with antibodies (b) optimize the ratio when conjugation magnetic beads with antibodies (c) optimize the usage amount of coupled antibody magnetic beads (d) optimize the dilution ratio of the Eu³⁺ labeled antibody. CPS, counts per second; galactose-deficient IgA1; ns, nonsignificant.

Optimization of the antibody-to-bead ratio demonstrated that fluorescence intensity increased progressively and reached saturation at a ratio of 1:100 (Figure 3b), which was therefore selected for subsequent experiments. Likewise, fluorescence signals increased with the concentration of antibody-conjugated magnetic beads and plateaued at 750 ng/ml while maintaining low background signals (Figure 3c). Thus, 750 ng/ml was selected as the optimal working concentration.

As shown in Figure 3d, a 1:800 dilution of Eu³⁺-labeled goat antihuman IgA antibody generated the highest signal-to-background ratio and was selected for subsequent experiments.

Analytical Performance Evaluation of MH33-Based TRFIA

The analytical performance of the MH33-based TRFIA was systematically evaluated.

As shown in Figure 4, the standard curve exhibited excellent linearity across the tested concentration range and followed the equation: Log(Y) = 0.41 Log(X) + 4.19 with an R² value of 0.99, indicating robust quantitative performance. The analytical sensitivity, calculated as mean + 2SD of 0-concentration measurements, was 0.2 RU/ml. The intraassay coefficient of variation ranged from 2.32% to 5.49%, whereas the interassay coefficient of variation ranged from 3.97% to 5.23% (Supplementary Table S1). Recovery rates ranged from 89.80% to 103.76% (Supplementary Table S2), demonstrating good precision, reproducibility, and analytical accuracy.

Figure 4.

Standard curve of galactose-deficient IgA1–time-resolved fluorescence immunoassay. CPS, counts per second.

Comparison Between MH33- and KM55-Based Gd-IgA1 Detection

To compare the detection characteristics of MH33 and KM55, paired measurements were performed using samples for which both assays were available. Because KM55 measurements and total serum IgA data were not available for all participants, subset analyses were performed using all available samples within the same study cohort for assay comparison and correlation analyses.

As shown in Figure 5a, KM55-based measurements revealed Gd-IgA1 concentrations of 3411.43 (2052.91–4353.24) ng/ml in HCs, 4352.62 (2860.07–5830.91) ng/ml in the non-IgAN group, and 6268.50 (4411.25–8223.39) ng/ml in the IgAN group. Significant differences were observed between HCs and IgAN, as well as between non-IgAN and IgAN (P < 0.001).

Figure 5.

Comparison of Gd-IgA1 detection characteristics between MH33-based TRFIA and KM55-based assays. (a) Distribution of Gd-IgA1 levels measured by the KM55-based assay in healthy controls, patients without IgA nephropathy (IgAN), and patients with IgAN. (b) Distribution of Gd-IgA1 levels measured by MH33-based TRFIA in healthy controls, non-IgAN, and IgAN groups. (c) Correlation analysis between MH33-based TRFIA and KM55-based measurements in the IgAN group. (d) Correlation analysis between MH33-based TRFIA and KM55-based measurements in the non-IgAN group. (e) Correlation analysis between MH33-derived Gd-IgA1 levels and total serum IgA concentrations in the IgAN group. (f) Correlation analysis between MH33-derived Gd-IgA1 levels and total serum IgA concentrations in the non-IgAN group. ELISA, enzyme-linked immunosorbent assay; Gd-IgA1, galactose-deficient IgA1; TRFIA, time-resolved fluorescence immunoassay.

Similarly, MH33-based TRFIA demonstrated Gd-IgA1 levels of 8.75 (4.47–14.6) RU/ml in HCs, 6.45 (3.75–11.60) RU/ml in the non-IgAN group, and 27.0 (16.43–52.33) RU/ml in the IgAN group (Figure 5b). Significant differences were observed between HCs versus IgAN and non-IgAN versus IgAN (P < 0.001).

To further evaluate the relationship between MH33- and KM55-based measurements, correlation analyses were performed separately in IgAN and non-IgAN groups. As shown in Figure 5c and d, relatively weak correlations were observed between the 2 assays in both the IgAN group (r = 0.23, P = 0.09) and the non-IgAN group (r = 0.11, P = 0.80), indicating that MH33- and KM55-based assays may not detect identical Gd-IgA1–associated signals.

To further investigate whether MH33-derived Gd-IgA1 measurements were influenced by overall IgA abundance, correlation analyses were performed between MH33-derived Gd-IgA1 levels and total serum IgA concentrations using all available samples with complete total IgA measurements from the same study cohort. As shown in Figure 5e and f, weak correlations were observed between MH33-derived Gd-IgA1 levels and total serum IgA concentrations in both the IgAN group (r = 0.14, P = 0.25) and the non-IgAN group (r = 0.23, P = 0.052). These findings suggest that MH33-based detection is not merely driven by total serum IgA abundance.

Diagnostic Performance of MH33-Based Gd-IgA1 TRFIA and Comparison With KM55

To evaluate the diagnostic performance of MH33-based Gd-IgA1 TRFIA, serum Gd-IgA1 levels were first compared among HCs, non-IgAN renal disease controls, and patients with IgAN. As shown in Figure 6a, Gd-IgA1 concentrations measured by MH33-based TRFIA were 9.90 (4.87–14.65) RU/ml in HCs, 7.25 (4.37–11.80) RU/ml in the non-IgAN group, and 27.40 (17.47–48.65) RU/ml in patients with IgAN. Gd-IgA1 levels in the IgAN group were significantly higher than those in both HCs and non-IgAN groups (both P < 0.001), whereas no significant difference was observed between HCs and non-IgAN groups (P = 0.056).

Figure 6.

Diagnostic performance of MH33-based Gd-IgA1 TRFIA and comparison with KM55-based ELISA. (a) Distribution of serum Gd-IgA1 levels measured by MH33-based TRFIA in HC, patients with non-IgAN, and patients with IgAN. Data are presented as scatter plots with median and interquartile range. (b) Receiver operating characteristic curve analysis of MH33-based Gd-IgA1 TRFIA for distinguishing IgAN from HC, non-IgAN renal diseases, and combined controls (HC + non-IgAN). (c) Diagnostic classification performance of MH33-based Gd-IgA1 TRFIA at the optimal cutoff value (22.0 RU/ml). The proportions of positive and negative test results are shown for HC, non-IgAN renal disease controls, and patients with IgAN. (d–f) Comparison of receiver operating characteristic curves between MH33-based Gd-IgA1 TRFIA and KM55-based ELISA for differentiating (d) IgAN from HC, (e) IgAN from non-IgAN renal diseases, and (f) IgAN from all controls (HC + non-IgAN). AUC, area-under-the-curve; ELISA, enzyme-linked immunosorbent assay; HC, healthy controls; Gd-IgA1, galactose-deficient IgA1; IgAN, IgA nephropathy; TRFIA, time-resolved fluorescence immunoassay.

Receiver operating characteristic analysis demonstrated the diagnostic performance of MH33-based Gd-IgA1 TRFIA (Figure 6b). The assay yielded an AUC of 0.88 (95% CI: 0.84–0.93) for differentiating IgAN from HCs, 0.91 (95% CI: 0.87–0.95) for differentiating IgAN from non-IgAN renal diseases, and 0.90 (95% CI: 0.86–0.93) for differentiating IgAN from the combined HCs + non-IgAN group.

Because this study was conducted at a single center and an independent external validation cohort was unavailable, internal validation was performed using a 10-fold cross-validation strategy in R (version 4.3.2). The mean cross-validated AUC was 0.89, closely matching the AUC obtained from the full dataset (0.90), thereby demonstrating the robustness of the model and suggesting minimal overfitting.

To further evaluate diagnostic classification performance, the optimal cutoff value determined by receiver operating characteristic analysis (22.0 RU/mLl was applied (Figure 6c). At this threshold, positive test results were observed in 58 of 90 patients with IgAN (64.4%), 6 of 90 HCs (6.7%), and 5 of 88 non-IgAN renal disease controls (5.7%). The corresponding specificities were 93.33% for IgAN versus HCs, 94.32% for IgAN versus non-IgAN renal diseases, and 93.82% for IgAN versus the combined control group. The agreement between assay results and biopsy-based diagnosis across different cutoff values is summarized in Supplementary Table S3, showing progressively increased specificity at higher thresholds.

The diagnostic performance of MH33-based Gd-IgA1 TRFIA was further compared with that of the KM55-based ELISA. As shown in Figure 6d, the AUC values for differentiating IgAN from HCs were 0.88 for MH33-based TRFIA and 0.83 for KM55-based ELISA. For differentiating IgAN from non-IgAN renal diseases (Figure 6e), the AUC values were 0.91 and 0.75, respectively. When IgAN was compared with the combined HCs + non-IgAN group (Figure 6f), the corresponding AUC values were 0.90 and 0.78, respectively.

Discussion

IgAN is a complex immune-mediated kidney disease characterized by aberrant O-glycosylation of IgA1, formation of circulating immune complexes, and subsequent mesangial deposition. Despite extensive research, the lack of reliable noninvasive biomarkers remains a major clinical challenge, and renal biopsy continues to be required for definitive diagnosis.²¹ Therefore, identifying biomarkers that more accurately reflect the underlying pathogenic process remains a priority for improving diagnostic precision and reducing dependence on invasive procedures. In the present study, we established a magnetic separation TRFIA based on a novel human–mouse chimeric monoclonal antibody, MH33, for the detection of Gd-IgA1 and systematically evaluated its analytical performance and clinical utility in IgAN.

Gd-IgA1 is widely recognized as a key pathogenic molecule in IgAN and has become one of the most promising biomarker candidates for disease diagnosis and monitoring. However, accumulating evidence suggests that Gd-IgA1 does not represent a single molecular species but rather a heterogeneous population of glycoforms with variable site-specific O-glycosylation patterns. Because O-glycosylation is highly heterogeneous and posttranslationally regulated, differences in GalNAc exposure and glycosylation site occupancy may generate structurally distinct Gd-IgA1 molecules with potentially different biological and clinical relevance. This heterogeneity may partly explain why currently available assays, including lectin-based methods and KM55-based ELISAs, still exhibit considerable overlap among HCs, patients with IgAN, and disease controls.^22,23

Compared with traditional Helix aspersa agglutinin lectin–based assays, KM55-based ELISA substantially improves assay stability and analytical reproducibility.²⁴ However, our findings suggest that different antibodies recognizing distinct glycoepitopes may capture different subsets of circulating Gd-IgA1 molecules. Using site-specific synthetic glycopeptides, we demonstrate that MH33 preferentially recognizes the PSTPPT(GalNAc)PS motif, whereas KM55 preferentially recognizes a distinct glycosylated motif, PST(GalNAc)PPTPS. Importantly, MH33 showed no measurable binding toward the corresponding non-glycosylated peptide, indicating that recognition depends on GalNAc-associated glycan exposure rather than the peptide backbone itself. These findings support that MH33 is not functioning as a general IgA1-binding antibody but instead preferentially recognizes a distinct subset of Gd-IgA1-associated glycoepitopes.

Importantly, our intention is not to suggest that MH33 is universally more specific than KM55 for all forms of Gd-IgA1. Rather, both antibodies appear to possess strong specificity toward Gd-IgA1-associated GalNAc-exposed glycoepitopes while recognizing different glycosylation sites within the heterogeneous Gd-IgA1 population. This concept is supported by our clinical findings. Although MH33 and KM55 showed comparable discrimination between IgAN and HCs, MH33 demonstrated substantially improved discrimination between IgAN and disease controls. Furthermore, MH33- and KM55-based measurements showed only a relatively weak correlation, suggesting that the 2 antibodies may not detect identical populations of circulating Gd-IgA1 molecules. Considering that MH33 and KM55 recognize distinct GalNAc-associated glycoepitopes within the IgA1 hinge region, this finding further supports the possibility that the 2 antibodies capture different subsets of circulating Gd-IgA1 glycoforms.

Interestingly, MH33-derived Gd-IgA1 measurements showed poor correlation with total serum IgA levels, suggesting that MH33 recognition is not simply driven by overall IgA abundance but may instead preferentially reflect a disease-relevant glycoepitope subset of circulating IgA1 molecules. Because normally glycosylated IgA1 contains GalNAc residues that are typically masked by terminal galactose, whereas Gd-IgA1 exhibits increased GalNAc exposure due to deficient galactosylation, antibodies such as MH33 and KM55 likely recognize exposed GalNAc-associated glycoepitopes rather than the mere presence of IgA1 itself. This may explain why MH33 demonstrated improved disease-control discrimination despite only moderate correlation with KM55-based measurements.

The mechanisms underlying the improved diagnostic performance of MH33 remain incompletely understood. One plausible explanation is that the glycoform subset recognized by MH33 may be more closely associated with disease-relevant pathogenic processes in IgAN. Recent evidence increasingly supports the possibility that IgAN may involve distinct site-specific Gd-IgA1 glycosylation patterns; however, whether IgAN harbors disease-associated glycoforms remains incompletely defined. In our ongoing studies, a multienzyme coupled mass spectrometry strategy for site-specific O-glycosylation analysis has identified differences in GalNAc exposure patterns among patients with IgAN, disease controls, and HCs, providing preliminary mechanistic support for this possibility. Nevertheless, these findings remain under investigation and require further validation.

Beyond its diagnostic value, Gd-IgA1 may have potential utility for therapeutic monitoring in IgAN. Previous studies have suggested that circulating Gd-IgA1 levels tend to decrease following treatment and clinical improvement; however, the magnitude of reduction is often modest, generally in the range of approximately 10% to 20%. This presents an analytical challenge because the reported coefficient of variation of conventional ELISA-based assays may approach a similar range (approximately 15%–20%), potentially limiting their ability to reliably detect subtle longitudinal changes.^25,26 In contrast, the relatively low intra- and inter-assay variability observed for the present MH33-based TRFIA may provide improved analytical precision for monitoring dynamic changes in Gd-IgA1 levels over time. Although therapeutic monitoring was not evaluated in the current study, this possibility warrants further investigation in longitudinal cohorts of treated IgAN patients.

This study highlights several remaining challenges. First, given the ethnic variation and clinical heterogeneity of IgAN, large-scale multicenter validation studies are necessary to determine the generalizability of MH33 across populations and disease stages. Second, although MH33 demonstrated improved disease discrimination, further enhancement may be achieved through antibody combinations, multimarker strategies, or integration with assays measuring circulating immune complexes. Third, though our findings support that MH33 and KM55 recognize different Gd-IgA1 glycoepitope subsets, further biochemical and structural studies will be necessary to fully define these site-specific antibody–glycan interactions and determine their relationship to disease biology.

In conclusion, the MH33-based magnetic separation TRFIA provides a sensitive, specific, and noninvasive approach for Gd-IgA1 detection and represents a promising tool for the diagnosis of IgAN. Notably, the relatively high specificity and improved discrimination against disease controls suggest that MH33 may be particularly valuable as a rule-in diagnostic biomarker rather than a population screening tool. By combining improved analytical performance with compatibility for automated platforms, this assay may facilitate earlier diagnosis, refine risk stratification, and support individualized disease monitoring. More importantly, our findings suggest that recognition of distinct Gd-IgA1 glycoepitopes may provide new insights into the biological heterogeneity of Gd-IgA1 and the identification of clinically relevant glycoform subsets in IgAN.

Disclosure

All the authors declared no competing interests.

Supplementary Material

Supplementary File (PDF)

Table S1. Accuracy of Gd-IgA1-TRFIA. Table S2. Precision of Gd-IgA1-TRFIA. Table S3. Gd-IgA1-TRFIA with different cutoff values and the positive distribution of pathological diagnosis.