Abstract
Objectives
IgG4-related disease (IgG4-RD) is an immune-mediated condition with diagnostic challenges in clinical practice. Sensitive biomarkers are needed for diagnosis, disease activity and disease progression assessment. This study aimed to identify and validate novel serum protein biomarkers of IgG4-RD to improve clinical management.
Methods
Using Olink proteomics, we analyzed the expression of 92 serum immuno-oncology-related proteins in 11 treatment-naïve IgG4-RD patients and 11 healthy controls (HCs). Candidate biomarkers capable of distinguishing IgG4-RD patients from HC were subsequently validated by enzyme-linked immunosorbent assay in a large independent cohort (n = 220). Diagnostic performance was assessed via 5-fold cross-validation. Correlations between biomarkers and clinical characteristics were also investigated, as well as predictive value for disease relapse.
Results
Olink proteomic analysis identified 27 differentially expressed proteins between IgG4-RD and HCs. Through longitudinal follow-up, serum PD1, OX40, CCL19, and MMP12 were significantly upregulated in IgG4-RD patients and closely associated with disease progression, serum IgG4 levels and inflammatory indicators. A comprehensive diagnostic model incorporating the four biomarkers was developed and showed high discriminatory capacity. Elevated baseline PD1 level served as an independent risk factor to predict clinical relapse of IgG4-RD patients.
Conclusion
This study identifies four novel serum biomarkers that effectively assist diagnosis, assess disease activity, and one capable for clinical relapse prediction. These findings provide a practical approach for large-scale clinical screening and monitoring of IgG4-RD patients. Furthermore, the identified proteins may offer new insights into disease pathogenesis and represent potential therapeutic targets for this challenging condition.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13075-026-03787-w.
Keywords: IgG4-related disease, Biomarker, Proteomics, Relapse
Introduction
IgG4-related disease (IgG4‐RD) is a chronic fibro-inflammatory disease characterized by excessive immune system activation. Multiple organs are involved in IgG4-RD, which typically manifested as localized masses that lead to organ damage and functional impairment [1]. Early diagnosis and treatment are crucial for improving clinical outcomes in IgG4-RD. Currently, definitive diagnosis relies on a combination of clinical manifestations, elevated serum IgG4 levels, and pathological findings [2]. However, the clinical manifestations often overlap with other inflammatory conditions and malignancies, often delaying accurate diagnosis [3]. Moreover, IgG4-RD exhibits a remission-relapse disease course, with approximately 10.66% of patients suffer clinical relapse within one year despite treatment, which can result in progressive organ dysfunction [4]. Disease relapse prediction before initiating therapy could assist clinicians in selecting optimal therapeutic strategies, thereby improving long-term disease management.
Several biomarker have been explored for the clinical diagnosis, treatment monitoring, and prognosis of autoimmune disorders such as rheumatoid arthritis, systemic lupus erythematosus, systemic sclerosis [5–11]. Though several candidate biomarkers have been preliminarily proposed to distinguish the disease from other immune disorders and healthy controls, they remain inadequately validated and lack comprehensive clinical exploration [12–14]. Currently, there remains a critical need for non-invasive, clinically applicable protein markers capable of accurately diagnosing IgG4-RD, evaluating disease activity, and predicting long-term outcomes. To address this gap, we utilized high-sensitivity proteomics to comprehensively analyze a profile of 92 serum proteins in IgG4-RD patients and HCs. Our systematic approach identified multiple differentially expressed proteins, and subsequently evaluated their diagnostic performance, correlation with disease activity, and prognostic value for clinical outcomes.
Materials and methods
Study design
An exploratory proteomic analysis was conducted using the highly sensitive Olink Explore proteomics on a preliminary cohort of 11 treatment-naïve IgG4-RD patients and 11 age/sex-matched HCs. Clinical and proteomic data were collected from IgG4-RD patients throughout treatment to identify candidate biomarkers for diagnosis and progression. Candidate biomarkers were subsequently validated in an independent cohort using enzyme-linked immunosorbent assay (ELISA). This study was approved by the Ethics Committee of Beijing Friendship Hospital (No. 2024-P2-245-01). Written informed consent was obtained from all participants, and the study complied with the principles of the Declaration of Helsinki.
Patients and HCs
Patients diagnosed with IgG4-RD according to the 2019 American College of Rheumatology/European League Against Rheumatism classification criteria were enrolled [15]. The preliminary cohort consisted of 11 treatment-naïve IgG4-RD patients who subsequencly received standardized treatment (glucocorticoid [GC] alone or combined with immunosuppressant [IM]) following a uniform GC tapering regimen. Among these patients, eight experienced clinical relapse, whereas the other three maintained stable conditions under one-year treatment. Clinical examinations, routine blood parameter evaluations, and serum biomarker detections were also performed during follow-up. Clinical disease activity measures for IgG4-RD included serum IgG4 levels, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) levels, and the IgG4-RD responder index (IgG4-RD RI) [16]. To evaluate the specificity of biomarkers, expression levels of candidate biomarkers were evaluated in an extented cohort comprising 50 IgG4-RD patients, 20 lymphoma patients, 30 patients with antineutrophil cytoplasmic antibody-associated vasculitis (AAV) patients, and 28 patients with pancreatic cancer (PC) prior to surgery. Adiitionally, formalin-fixed paraffin-embedded salivary gland tissues were examined, including IgG4-RD involved tissues (n = 3) and benign tumor-adjacent normal tissues as disease control(n = 3). An independent validation cohort of 140 IgG4-RD patients and 80 HCs was assembled to assess the performance of candidate biomarkers. The treatment regimens in this cohort were GC monotherapy, GC combined with IM (cyclophosphamide, mycophenolate mofetil, leflunomide, iguratimod, tofacitinib, baricitinib, azathioprine, and tacrolimus), or biologic agents (rituximab and dupilumab).
Sample collection and ELISA detection
Peripheral blood samples were collected and processed within two hours. Serum was separated by centrifugation at 2500× g for 10 min at 4 °C and stored at -80 °C until detection. Serum levels of candidate biomarkers were detected using the human programmed cell death protein 1 (PD1) (Invitrogen), tumor necrosis factor receptor superfamily member 4 (OX40) (Proteintech), C-C motif chemokine 19 (CCL19) (Neobioscience), and macrophage metalloproteinase-12 (MMP12) (RayBiotech) ELISA kits according to the manufacturers’ instructions.
Olink proteomic assays
Serum samples were profiled in-house using proximity extension assays, specifically the 96-plex Immuno-Oncology assay developed by Olink Proteomics and provided by Shanghai Biotechnology Corporation. The protein assay list is available online (https://olink.com/content/uploads/2021/09/1047-v1.3-immuno-onc-panel-content-final.pdf). Quality control and data normalization according to reference controls were performed using the Normalized Protein Expression (NPX) Software, Olink NPX Manager. The NPX values were used as a relative quantification method to compare the expression levels of individual proteins under different conditions. Proteins with>75% missing data were excluded from analysis. Differentially expressed proteins (DEPs) were defined as those with a significant change in NPX (P < 0.05). The full names, abbreviations, and analytical details for the 92 assayed biomarkers are provided in Table S1.
Immunohistochemistry (IHC)
IHC staining was performed on all tissue specimens to localize the expression of PD1, OX40, CCL19 and MMP12 using the following autibodies: anti-PD1 (1:200 dilution, catalogue #HA750103; HUABIO), anti-OX40 (1:100 dilution, catalogue #61637S; Cell Signaling Technology), anti-CCL19 (1:200 dilution catalogue #13397-1-AP; Proteintech), and anti-MMP12 (1:100 dilution, catalogue #66930-1-Ig; Proteintech). Primary antibodies against PD1, OX40, CCL19 and MMP12 were incubated at 4°C overnight, rinsing with phosphate-buffered saline (PBS) three times and allowed to react with a secondary antibody at room temperature for 50 minutes. Colorimetric detection was completed with 3,3’-diaminobenzidine, and the slides were counterstained with haematoxylin. Images were captured from immunoreactive areas. Positive staining was indicated by a brown precipitate, while its absence indicated negativity. The staining results were semi-quantitatively expressed as the percentage of positive staining area which analysed by ImageJ.
Statistical analysis
Statistical analysis was performed using GraphPad Prism version 9, IBM SPSS Statistics version 25, and R version 4.3.3. Data normality was assessed with the Shapiro-Wilk test. Normally distributed continuous data are presented as mean ± standard deviation (SD) and analyzed using Student’s t-test; non-normally distributed continuous data are presented as median with interquartile range (IQR) and analyzed using the Mann-Whitney U test for two-group comparisons. Categorical parameters were compared using the chi-squared test or Fisher’s exact test. Correlations were evaluated with Pearson’s (normal data) or Spearman’s (non-normal data) rank correlation tests. P values from correlation analyses and multiple comparisons were adjusted using the Benjamini-Hochberg False Discovery Rate (FDR) method. Pathway enrichment analysis was conducted using the Kyoto Encyclopedia of Genes and Genomes (KEGG) library, and Gene Ontology (GO) annotations were performed using Blast2Go. A logistic regression model of four candidate proteins was constructed, and the diagnostic efficacy was validated using the glm function and forward step method from the R package stats. Model diagnostic performance was assessed by receiver operating characteristic (ROC) curve analysis, reporting the area under the curve (AUC), sensitivity, specificity, and statistical significance (P < 0.05). Model robustness was assessed via K-fold cross-validation. Hazard ratios (HRs) for clinical relapse were calculated using univariate and multivariate Cox regression models. Multivariate models were adjusted for age, sex and clinical covariates with P < 0.1 by backward methods. Statistical significance was defined as P value < 0.05.
Results
Identification of differentially expressed proteins in IgG4-RD and novel candidate biomarkers that correlated with disease progression
A flowchart of the biomarker screening performed in this study was shown in Fig. 1. Initially, Olink proteomic profiling was conducted between 11 untreated IgG4-RD patients and 11 age/sex-matched HCs (Table 1). The relative expression of 92 assayed proteins for both groups were listed in Table S1. A total of 27 DEPs were identified, and the top 10 most significantly altered proteins between IgG4-RD patients and HCs were visualized in the Volcano plot (Fig. 2a). Among the 27 DEPs, 11 were significantly upregulated and 16 were downregulated in IgG4-RD patients. Comparative analysis of treatment-naïve and post-treatment samples from IgG4-RD patients revealed significantly decreased expression levels for 10 proteins and elevated levels for 4 proteins following therapy (Fig. 2b). Pathway enrichment analysis revealed that the DEPs were primarily involved in biological regulation pathways, suggesting that treatment may help restore the expression of key immune-regulatory proteins towards levels observed in HCs.A correlation heatmap was constructed to analyze the relationships between the expression levels of the 27 DEPs and clinical characteristics (Figure S1). From the above analysis, we identified four proteins—PD1, OX40, CCL19, and MMP12—whose expressions were not only significantly elevated in IgG4-RD patients but also showed strong correlations with established inflammatory indicators. We found that the expressions of PD1, OX40, CCL19, and MMP12 were significantly upregulated in IgG4-RD patients and closely correlated with inflammatory indicators. By comparing the levels throughout the treatment period, we surprisingly found that the changes of four proteins, PD1, OX40, CCL19, and MMP12 were parallelled with disease progression (Fig. 2c). The expressions of PD1, OX40, CCL19, and MMP12 decreased after treatment and re-elevated when patients suffered clinical relapse, closely paralleled with serum IgG4 levels and IgG4-RD RI (Fig. 2d). Correlation analyses confirmed that the levels of these four proteins were positively associated with serum IgG4, eosinophil count, disease activity scores, and other inflammatory markers, while demonstrating a negative correlation with complement components C3 and C4 (Fig. 2e). In contrast, the remaining DEPs exhibited weaker or less consistent correlations with disease progression and key clinical parameters.
Fig. 1.
Study design. Differential serum protein landscape between IgG4-RD patients and HCs were firstly detected using Olink proteomics. Candidate biomarkers were further validated by ELISA in an independent cohort. The diagnostic value, disease activity, correlation with clinical characteristics, and predictive value of the novel biomarkers were further explored. ELISA, enzyme-linked immunosorbent assay; HC, healthy control; IgG4-RD, Immunoglobulin G4-related disease
Table 1.
Baseline characteristics of IgG4-RD patients and HCs for Olink proteomics
| IgG4-RD (N = 11) |
HC (N = 11) |
|
|---|---|---|
| Male/Female | 1.75/1 | 1.75/1 |
| Age (year) | 61.91 ± 8.92 | 60.00 ± 5.76 |
| History of allergy, n (%) | 6 (54.5%) | NA |
| IgG, g/L | 1850 (1610–2400) | NA |
| IgM, g/L | 78.13 ± 46.95 | NA |
| IgA, g/L | 186.24 ± 76.36 | NA |
| IgE, IU/ml | 323.79 ± 190.29 | NA |
| IgG1, g/L | 9.78 (8.60–12.20) | NA |
| IgG2, g/L | 5.34 ± 2.66 | NA |
| IgG3, g/L | 0.41 (0.23–0.69) | NA |
| IgG4, g/L | 8.49 (5.81–15.70) | NA |
| ESR, mm/H | 16.00 (12.00–46.00) | NA |
| CRP, mg/L | 2.00 (0.59–3.75) | NA |
| IgE, IU/ml | 323.79 ± 190.29 | NA |
| C3, g/L | 0.787 ± 0.258 | NA |
| C4, g/L | 0.163 ± 0.100 | NA |
| EO, 109/L | 0.18 (0.14–0.56) | NA |
| IgG4-RD RI | 8.6 ± 5.1 | NA |
CRP C-reactive protein, EO eosinophils count, ESR erythrocyte sedimentation rate, IgG4-RD IgG4-related disease, NA not available
Fig. 2.
Identification of novel candidate biomarkers in IgG4-RD through proteomic sequencing analysis. (a) Volcano plots for 27 DEPs between IgG4-RD (n = 11) and HCs (n = 11). The 27 DEPs were marked as red dots, with the top 10 DEPs annotated with protein names. GO classification in biological processes, molecular function, and cellular components among the DEPs. (b) Volcano plots for 14 DEPs in IgG4-RD before and after treatment. (c) Changes of serum levels of PD1, OX40, CCL19, and MMP12 during one-year treatment in IgG4-RD patients. (d) Changes of serum IgG4 concentration during one-year treatment in IgG4-RD patients. (e) Correlation map between candidate biomarkers and clinical parameters by Pearson or Spearman analysis. The correlation coefficients were indicated in the matrix marked with color intensity. X = non-significant (adjusted for FDR). HC, healthy control; IgG4-RD, Immunoglobulin G4-related disease; DEP, differential protein; GO, Gene Ontology; FDR, false discovery rate. ns, P>0.05, *P<0.05, **P<0.01, ***P<0.001
Validation of the candidate biomarkers in IgG4-RD and control groups
IgG4-RD is frequently misdiagnosed as AAV due to overlapping manifestations, including multi-system involvement and similar histopathological features. Notably, AAV patients can exhibit elevated serum IgG4 levels and tissue infiltration of IgG4-positive plasma cells, further complicating the differential diagnosis [17]. Similarly, IgG4-RD patients with pancreatic involvement often present with abdominal masses, leading to potential confusion with pancreatic cancer. Given that lymphoma represents a malignant lymphoproliferative disorder that can clinically mimic IgG4-RD, it was also included as a critical disease control to rigorously evaluate the diagnostic specificity of the candidate biomarkers. No significant differences in age or sex distribution were observed among the IgG4-RD patients, the disease control groups (lymphoma, AAV, and PC), and the HCs. The expressions of the four candidate biomarkers were systematically compared across these five groups (Fig. 3a-d). The results demonstrated that all four proteins were elevated in IgG4-RD compared to those with lymphoma, AAV, PC and HCs. These findings align with the initial proteomic findings, suggesting their possible involvement in the underlying mechanisms of IgG4-RD. Notably, MMP12 showed marked upregulation specifically in IgG4-RD patients, while maintains comparable levels among AAV, PC and HCs.
Fig. 3.
Expressions of four biomarkers in IgG4-RD and control groups. (a) The differences in the expressions of PD1 among IgG4-RD and control groups. (b) The differences in the expressions of OX40 among IgG4-RD and control groups. (c) The differences in the expressions of CCL19 among IgG4-RD and control groups. (d) The differences in the expressions of MMP12 among IgG4-RD and control groups. Kruskal-Wallis test was conducted for comparison between groups. AAV, ANCA-associated vasculitis; HC, healthy controls; PC, pancreatic cancer. ns: P > 0.05, *P < 0.05, **P < 0.01, ***P < 0.001
Subsequently, we created a five-fold cross-validation procedure in an independent cohort containing 140 IgG4-RD patients and 80 HCs to assess the diagnostic performance of each biomarker (Fig. 4a). Among the four candidates, PD1 demonstrated the highest diagnostic value, yielding a mean AUC of 0.902 (P < 0.001), with a high specificity of 98.7% and a sensitivity of 77.5%. OX40 achieved a mean AUC of 0.869 with 84.0% sensitivity and 81.4% specificity; CCL19 showed a mean AUC of 0.879, with a sensitivity of 82.2% and specificity of 84.8%; and MMP12 provided a lower but still significant discriminatory power, with an AUC of 0.699, with a sensitivity and specificity of 69.4% and 73.9%, respectively. Importantly, PD1, OX40, and CCL19 retained their ability to effectively distinguish IgG4-RD patients who had serum IgG4 levels below the conventional diagnostic threshold (< 1.35 g/L), highlighting their added diagnostic value in seronegative cases that are challenging to identify with current criteria (Figure S2).
Fig. 4.
Diagnostic performance of biomarkers and the protein model in IgG4-RD. (a) ROC curves of 5-fold cross-validation of PD1, OX40, CCL19 and MMP12 in IgG4-RD. (b) ROC curves of protein model in IgG4-RD by 5-fold cross-validation. IgG4-RD, Immunoglobulin G4-related disease; ROC, receiver operating characteristic
A logistic regression model containing the four proteins by a stepwise selection model was established (Fig. 4b). This combined model showed superior discrimination performance in the independent cohort (AUC = 0.939 ± 0.018, P<0.001) compared with each protein alone, with a high sensitivity of 87.1% and specificity of 93.3%, respectively.
To further validate these findings at the tissue level, IHC was performed to detect the expression of PD1, OX40, CCL19, and MMP12 in salivary gland tissues from IgG4-RD patients and benign tumor-adjacent normal controls (Fig. 5a). The expression of PD-1 [7.730 ± 2.441 vs. 0.299 ± 0.242, P = 0.006], OX40 [10.920 ± 5.200 vs. 0.912 ± 0.122, P = 0.029], CCL19 [10.450 ± 1.139 vs. 5.194 ± 0.834, P = 0.003], and MMP12 [26.52 ± 7.349 vs. 3.447 ± 0.642, P = 0.006] were all significantly elevated in IgG4-RD–involved salivary glands compared to normal tissues (Fig. 5b). Furthermore, the IHC analysis revealed that OX40 was predominantly expressed within lymphoid follicle areas, suggesting its potential role in activating germinal center responses—a key immunological feature of IgG4-RD.
Fig. 5.
The expression of biomarkers in salivary gland tissue from IgG4-RD patients and controls. (a) Representative immunohistochemical staining of PD1, OX40, CCL19 and MMP12 in salivary gland tissue. (b) The differential expressions were estimated by semiquantitative analysis and shown as percentages of positive area between two groups. *P < 0.05, **P < 0.01
Association of serum biomarker levels with disease activity and clinical parameters in IgG4-RD
Correlation analyses were performed between serum levels of the four biomarkers and the clinical variables across the entire IgG4-RD cohort. Serum concentrations of PD1, OX40, CCL19 and MMP12 were positively correlated with serum IgG4 levels and IgG4-RD RI (P < 0.001). The strong correlations were also observed between the levels of four biomarkers and the inflammatory indicators (ESR and CRP) (Table 2). These findings indicate that elevated levels of PD1, OX40, CCL19, and MMP12 are associated with more severe systemic inflammation and higher overall disease activity. In addition, serum levels of four proteins showed significant positive correlations with the number of affected organs, a crucial prognostic indicator that guides clinical management strategies in IgG4-RD. These biomarker levels also exhibited positive correlations with other relevant clinical parameters, such as serum IgG, IgG1, globulin (GLB), IgE, and eosinophil count. Conversely, a negative correlation was observed with complement components C3 and C4, suggesting a potential link to complement consumption during active disease. Additional analysis revealed that serumlevels of PD1 and MMP12 were correlated with patients’ age at disease onset. A positive correlation was observed between the four biomarkers and serum IgG3 levels, whilePD1 and CCL19 were additionally correlated with IgG2. Given the correlation between biomarker levels and the number of affected organs, we further investigated the relationship between biomarkers and specific organ involvement as well as organ involvement subtypes (Table S2). The results indicated that none of the four biomarkers demonstrated a statistically significant, organ-specific association. This suggests that PD1, OX40, CCL19, and MMP12 are more likely to be general biomarkers of systemic disease activity in IgG4-RD, rather than being specific to pathology in any single organ.
Table 2.
Association between serum biomarkers and clinical parameters in IgG4-RD
| PD1 | OX40 | CCL19 | MMP12 | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| r | FDR-adjusted P value |
r | FDR-adjusted P value |
r | FDR-adjusted P value |
r | FDR-adjusted P value |
||||
| Onset age (year) | 0.181 | 0.040* | 0.128 | 0.150 | 0.134 | 0.131 | 0.225 | 0.010* | |||
| Number of involved organs | 0.441 | <0.001*** | 0.367 | <0.001*** | 0.407 | <0.001*** | 0.378 | <0.001*** | |||
| EO, 109/L | 0.422 | <0.001*** | 0.317 | <0.001*** | 0.296 | <0.001*** | 0.215 | 0.014* | |||
| GLB, g/L | 0.607 | <0.001*** | 0.470 | <0.001*** | 0.601 | <0.001*** | 0.409 | <0.001*** | |||
| ESR, mm/H | 0.559 | <0.001*** | 0.388 | <0.001*** | 0.489 | <0.001*** | 0.389 | <0.001*** | |||
| CRP, mg/L | 0.252 | 0.004** | 0.260 | 0.003** | 0.316 | <0.001*** | 0.252 | 0.004** | |||
| IgG, g/L | 0.671 | <0.001*** | 0.520 | <0.001*** | 0.626 | <0.001*** | 0.492 | <0.001*** | |||
| IgA, g/L | -0.100 | 0.262 | -0.127 | 0.150 | 0.013 | 0.892 | -0.056 | 0.544 | |||
| IgM, g/L | 0.037 | 0.695 | 0.011 | 0.899 | -0.021 | 0.830 | -0.056 | 0.544 | |||
| IgG1, g/L | 0.508 | <0.001*** | 0.436 | <0.001*** | 0.463 | <0.001*** | 0.427 | <0.001*** | |||
| IgG2, g/L | 0.250 | 0.004** | 0.163 | 0.065 | 0.261 | 0.003** | 0.155 | 0.079 | |||
| IgG3, g/L | 0.328 | <0.001*** | 0.238 | 0.006** | 0.344 | <0.001*** | 0.295 | <0.001*** | |||
| IgG4, g/L | 0.715 | <0.001*** | 0.491 | <0.001*** | 0.611 | <0.001*** | 0.392 | <0.001*** | |||
| IgE, IU/ml | 0.446 | <0.001*** | 0.331 | <0.001*** | 0.402 | <0.001*** | 0.297 | <0.001*** | |||
| C3, g/L | -0.356 | <0.001*** | -0.296 | <0.001*** | -0.250 | 0.004** | -0.197 | 0.024* | |||
| C4, g/L | -0.312 | <0.001*** | -0.303 | <0.001*** | -0.277 | 0.001** | -0.156 | 0.077 | |||
| IgG4-RD RI | 0.700 | <0.001*** | 0.389 | <0.001*** | 0.568 | <0.001*** | 0.354 | <0.001*** | |||
EO Eosinophils, GLB Globulin, ESR erythrocyte sedimentation rate, CRP C-reactive protein, FDR false discovery rate, IgG4-RD RI IgG4-related disease responder index
*P<0.05, **P<0.01,*** P<0.001
PD1 as a potential protein to predict IgG4‑RD relapse
To explore the predictive biomarkers for clinical relapse, we classified IgG4-RD patients who received standardized treatment and regular follow-up for at least one year. Clinical relapse was defined as exacerbation of symptoms or imaging findings, with or without re-elevation of serum IgG4 concentrations. In total, 66 eligible patients were enrolled and divided into relapse (n = 20) and non-relapse (n = 46) groups (Table S3). There were no significant differences in baseline covariates between the two groups. Univariate and multivariate Cox regression analyses were performed to identify the risk factors for relapse. Univariate Cox regression analysis revealed that the baseline levels of PD1 (HR = 1.007, 95% CI = 1.002–1.012, P = 0.009) and OX40 (HR = 1.001, 95% CI = 1.000-1.002, P = 0.030) were risk factors for relapse (Table 3). Treatment of GC combined with IM (HR = 0.352, 95% CI = 0.134–0.928, P = 0.035) during the maintenance period was a protective factor for relapse. Multivariate Cox regression analysis revealed that high PD1 level at baseline (HR = 1.007, 95% CI = 1.001–1.013, P = 0.015) were an independent risk factor. Other variables were not associated with relapse. High baseline PD1 levels are linked to a greater propensity for disease recurrence, potentially reflecting an underlying state of heightened T-cell activation and inflammatory state.
Table 3.
Prediction of clinical relapse in patients with IgG4-RD patients
| Univariable | Multivariable | ||||||
|---|---|---|---|---|---|---|---|
| HR, 95% CI | P value | HR, 95% CI | P value | ||||
|
Sex Age (year) |
0.693 (0.430–1.118) 1.038 (0.987–1.091) |
0.133 0.147 |
|||||
| ESR, mm/H | 1.011 (0.999–1.024) | 0.082 | |||||
| CRP, mg/L | 1.018 (0.995–1.042) | 0.133 | |||||
| EO, 109/L | 1.710 (0.942–7.797) | 0.064 | |||||
| IgG1, g/L | 1.043 (0.961–1.132) | 0.311 | |||||
| IgG4, g/L | 1.003 (0.981–1.025) | 0.775 | |||||
| IgE, IU/ml | 1.001 (0.999–1.002) | 0.344 | |||||
| C3, g/L | 0.984 (0.969-1.000) | 0.052 | |||||
| C4, g/L | 0.957 (0.913–1.004) | 0.070 | |||||
| IgG4-RD RI | 0.957 (0.856–1.069) | 0.435 | |||||
| Visceral organ involvement | 0.599 (0.230–1.558) | 0.293 | |||||
| PD1, pg/ml | 1.007 (1.002–1.012) | 0.009 | 1.007 (1.001–1.013) | 0.015 | |||
| OX40, pg/ml | 1.001 (1.000-1.002) | 0.030 | |||||
| CCL19, pg/ml | 1.001 (0.999–1.003) | 0.199 | |||||
| MMP12, ng/ml | 1.123 (0.729–1.729) | 0.599 | |||||
| Treatment | |||||||
| GC monotherapy | Reference | Reference | |||||
| GC + IM | 0.352 (0.134–0.928) | 0.035 | 0.406 (0.152–1.087) | 0.073 | |||
| Biologic agents | 0.150 (0.020–1.146) | 0.067 | 0.137 (0.018–1.057) | 0.057 | |||
EO Eosinophils, ESR erythrocyte sedimentation rate, CRP C-reactive protein, GC glucocorticoid, HR Hazard ratio, IgG4-RD RI IgG4-related disease responder index, IM immunosuppressant. Hazard ratios (HRs) for clinical relapse were calculated using univariate and multivariate Cox regression models. Multivariate models were adjusted for age, sex and clinical covariates with P < 0.1 by backward methods. Other covariates (age, sex, ESR, C3, OX40) were excluded in multivariate models
Discussion
An elevated serum IgG4 level is a key diagnostic criterion for IgG4-RD; however, its specificity is limited. Elevated serum IgG4 levels can also occur in other conditions, such as malignancies, infections, and allergic disorders. Conversely, a subset of IgG4-RD patients presents with normal IgG4 levels, often necessitating invasive biopsy for definitive diagnosis [18]. In addition, identifying reliable predictors of relapse prior to treatment initiation is crucial for risk stratification and personalized therapeutic planning for clinicians. To address these unmet clinical needs, we leveraged high-sensitivity Olink proteomic sequencing coupled with experimental validation to discover and characterize novel serum protein biomarkers for both diagnosis and long-term outcome prediction in IgG4-RD.
Our exploratory proteomic analysis revealed a distinct dysregulation of immune regulatory pathways in IgG4-RD, with 27 proteins showing differential expression compared to HCs. From this panel, four serum proteins (PD1, OX40, CCL19, and MMP12) were upregulated in IgG4-RD patients. Among these, PD1 demonstrated the highest diagnostic accuracy, characterized by excellent specificity. OX40 and CCL19 also exhibited good diagnostic performance. A comprehensive protein model consisting of the four proteins achieved superior diagnostic value, which was validated through cross-validation. Notably, the longitudinal dynamics of these four biomarkers closely mirrored key clinical parameters, including serum IgG4 levels, IgG4-RD RI, and systemic inflammatory markers. Their elevated expressions were further corroborated at the tissue level in affected salivary glands. Collectively, these findings strongly implicate PD1, OX40, CCL19, and MMP12 in the pathogenesis of IgG4-RD and support their potential as promising clinical biomarkers.
PD1, an inhibitory immune checkpoint expressed on activated T cells, plays a pivotal role in maintaining self-tolerance in immunity. Zhang et al. demonstrated the elevated expression of PD1 and its receptor in plasma, submandibular gland and the surface of regulatory T cells [19]. A correlation between PD1+ immune cells and the IgG4/IgG ratio has been reported in IgG4-RD, suggesting the regulatory role of PD1 in B cell differentiation and antibody production [20]. Our finding of a strong correlation between serum PD1 and inflammatory burden aligns with this result. Persistent high expression of PD1 may lead to excessive suppression of the function of immune cells, allowing persistent inflammation. Whether the soluble PD1 affects the functions of T cell subsets in IgG4-RD remains to be investigated. OX40 functions as a co-stimulatory molecule that regulates T-cell function and proliferation by activating multiple signaling pathways upon ligand binding. OX40 agonists have emerged as novel therapies in cancer immunotherapy [21]. Limited researches on OX40 in IgG4-RD have primarily focused on the T-cell expression, with no report of soluble OX40 [22, 23]. Our study firstly report that elevated serum OX40, correlating with disease activity and the number of affected organs. OX40 is known to enhance the Th2-type immune response and allergic process [24]. The blockage of OX40-OX40L signaling can inhibit inflammatory and allergic response, providing a potential therapeutic target for IgG4-RD [25]. According to researchers, nearly half of IgG4-RD patients have allergic symptoms and enhanced type 2 immune response, in which OX40 may play an important role [26]. CCL19 has been reported to promote dendritic cell migration and facilitate osteoclast migration in bone resorption of rheumatoid arthritis as a chemokine [27]. Under the chronic inflammation condition, immune cells accumulate in tertiary lymphoid structures and complete immunoglobulin subtype switching, and CCL19 can drive the formation of peripheral lymphoid organs [28]. Tertiary lymphoid structures are present in the tissues of IgG4-RD [29]. In this study, we found that elevated serum CCL19 was correlated with the disease activity, which suggests that CCL19 may serve as a potential biomarker for evaluating the disease activity and may be a key driver of the ectopic lymphoid neogenesis characteristic of IgG4-RD. We hypothesize that the chronic inflammatory environment in IgG4-RD may stimulates stromal cells such as fibroblasts to secrete CCL19, which in turn recruits lymphocytes, perpetuating tertiary lymphoid structures formation and sustaining local immune dysregulation. However, the cellular source and precise regulatory mechanism of CCL19 in IgG4-RD remain to be clarified. MMP12, a macrophage-derived metalloproteinase involved in extracellular matrix remodeling, is associated with fibrosis in conditions like systemic sclerosis [30]. The MMP12 mRNA was upregulated in IgG4-related ophthalmic disease tissue [31]. The high expression of MMP12 may be driven by the activation of macrophages by the IL33/ST2 pathway [32]. In our study, serum MMP12 was firstly reported to be upregulated in IgG4-RD, suggesting its involvement in the fibrotic progression of the disease. Although PD1, OX40, CCL19, and MMP12 are involved in immunological and fibrotic processes in multiple disease, our study firstly demonstrated their elevated expression especially its significant association with disease activity, highlighting their specific role in IgG4-RD. Besides, although the four biomarkers were significantly positively correlated with the number of involved organs, no specificity was showed to the specific organ involvement, indicating that these biomarkers mainly reflect the systemic inflammatory response of IgG4-RD patients.
Although serum IgG4 hold a significant value in disease diagnosis, the specificity is limited [33]. The accuracy of IgG4/IgG RNA ratio was also poor for IgG4-RD diagnosis [34]. Our biomarkers, particularly PD1, OX40, and CCL19, maintained diagnostic utility even in seronegative patients. Four autoantigens including prohibitin, annexin A11, laminin 511-E8, and galectin-3 have previously been identified in IgG4-RD [35–38]. However, only a small portion of IgG4-RD patients presented the high autoantibody response in a diverse validation cohort [39]. Besides, these proteins are present only in a minority of patients and lack correlation with disease activity. Similarly, other reported factors such as B cell activating factor and IL-2, though elevated, show weak clinical correlations [40]. Our panel thus offers a superior combination of diagnostic accuracy and activity assessment.In addition, we found that patients who suffered clinical relapse within one year had high levels of PD1, OX40, CCL19, and MMP12. Higher serum IgG4 concentrations, number of organ involvement, higher disease activity, history of allergy, and increased eosinophil levels at baseline have been reported to be predictive of relapse [4]. Patients who accepted the combination of GC with IM and biologics tend to have a lower relapse rate in our study. Yan et al. established a model that combined five metabolites for relapse prediction with good performance [14]. No protein has been reported as a predictor for relapse in IgG4-RD. Notably, our finding that high baseline PD1 is an independent risk factor for relapse provides novel insight into the proteomic drivers of disease recurrence, implicating heightened T-cell co-stimulation and active T-B cell collaboration as key mechanisms. This also suggests that patients with high baseline PD1 may benefit from more intensive initial therapy. This is the first study to integrate Olink proteomic discovery with extensive validation to identify a novel four-protein signature (PD1, OX40, CCL19 and MMP12) for IgG4-RD. These biomarkers demonstrate strong potential for improving diagnosis, monitoring disease activity, and predicting clinical relapse, offering a multidimensional view of the proteomic and immune dysregulation in IgG4-RD. Nevertheless, there are some limitations that need to be addressed. First, the sample size was relatively small owing to the low incidence rate of IgG4-RD. External validation is needed to test the clinical applicability of these findings in large, multicenter prospective cohorts. Second, owing to the infrequent occurrence of fibrotic subtypes, comparisons of biomarker levels across different clinical subtypes are limited. Third, deeper mechanistic studies to elucidate the precise roles of these proteins are needed. Future research addressing these points will be essential to translate these biomarkers into routine clinical practice and to explore their potential as therapeutic targets.
Conclusions
In conclusion, our study delineates a distinct proteomic signature associated with IgG4-RD. We identified and validated four key proteins—PD1, OX40, CCL19, and MMP12—that not only demonstrate diagnostic utility but also exhibit strong correlations with clinical disease activity and laboratory markers of inflammation. Importantly, we established that elevated baseline serum PD1 serves as an independent and valuable predictor of clinical relapse. These insights advance our understanding of the molecular underpinnings of IgG4-RD and highlight promising biomarkers for improving diagnosis, monitoring, and prognostic stratification in clinical practice.
Supplementary Information
Authors’ contributions
Conceptualization, design, and writing of the article: Qiyuan Hao and Yanying Liu. Performance of ELISA and data curation: Qiyuan Hao and Difei Lian. Sample and data collection: Qiyuan Hao, Difei Lian, Mingzhu Zhou, Hang Zhou, Xia Zhang, Tianqi Wang, Huilan Liu, Lan Gao, Mingxin Bai, Yuetong Xu, Fan Yang, Yini Wang, Zhitao Ying. Statistical analysis: Qiyuan Hao and Yanying Liu. Literature review and evaluation: Qiyuan Hao, Difei Lian, Mingzhu Zhou, Hang Zhou, Xia Zhang, Tianqi Wang, Huilan Liu, Lan Gao, Mingxin Bai, Yuetong Xu, Fan Yang, Guang Yang, Xiaolin Sun, Yanying Liu. Critical manuscript editing and funding acquisition: Yanying Liu.
Funding
This work was supported by the Beijing Municipal Natural Science Foundation (Grant No. 7232029) and National Natural Science Foundation of China (Grant No. 82572048).
Data availability
The datasets used during the current study are available from the corresponding author Yanying Liu ( [Liuyanying6850@126.com](mailto: Liuyanying6850@126.com) ) on reasonable request.
Declarations
Ethics approval and consent to participate
Written informed consent was obtained from all participants, and the study complied with the principles of the Declaration of Helsinki. This study was approved by the Ethics Committee of Beijing Friendship Hospital (No. 2024-P2-245-01).
Consent for publication
All the authors have read and approved the final version of this manuscript.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Perugino CA, Stone JH. IgG4-related disease: an update on pathophysiology and implications for clinical care. Nat Rev Rheumatol. 2020;16(12):702–14. [DOI] [PubMed] [Google Scholar]
- 2.Umehara H, Okazaki K, Masaki Y, et al. Comprehensive diagnostic criteria for IgG4-related disease (IgG4-RD), 2011. Mod Rheumatol. 2012;22(1):21–30. [DOI] [PubMed] [Google Scholar]
- 3.Peyronel F, Della-Torre E, Maritati F et al. IgG4-related disease and other fibro-inflammatory conditions. Nat Rev Rheumatol Published online April 7, 2025. [DOI] [PubMed]
- 4.Peng Y, Li JQ, Zhang PP, et al. Clinical outcomes and predictive relapse factors of IgG4-related disease following treatment: a long-term cohort study. J Intern Med. 2019;286(5):542–52. [DOI] [PubMed] [Google Scholar]
- 5.Sanges S, Rice L, Tu L, et al. Biomarkers of haemodynamic severity of systemic sclerosis-associated pulmonary arterial hypertension by serum proteome analysis. Ann Rheum Dis. 2023;82(3):365–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zhang Y, Liang Y, Feng L, Cui L. Diagnostic performance of 14-3-3η and anti-carbamylated protein antibodies in Rheumatoid Arthritis in Han population of Northern China. Clin Chim Acta. 2020;502:102–10. [DOI] [PubMed] [Google Scholar]
- 7.Mathian A, Mouries-Martin S, Dorgham K, et al. Ultrasensitive serum interferon-α quantification during SLE remission identifies patients at risk for relapse. Ann Rheum Dis. 2019;78(12):1669–76. [DOI] [PubMed] [Google Scholar]
- 8.Cesaroni M, Seridi L, Loza MJ, et al. Suppression of Serum Interferon-γ Levels as a Potential Measure of Response to Ustekinumab Treatment in Patients With Systemic Lupus Erythematosus. Arthritis Rheumatol. 2021;73(3):472–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Yuan CT, Huang WT, Hsu CL, et al. CXCL9 as a Reliable Biomarker for Discriminating Anti-IFN-γ-Autoantibody-Associated Lymphadenopathy that Mimics Lymphoma. J Clin Immunol. 2023;44(1):35. [DOI] [PubMed] [Google Scholar]
- 10.Li X, Luo T, Yan H, et al. Proteomic Analysis of Pediatric Hemophagocytic Lymphohistiocytosis: a Comparative Study with Healthy Controls, Sepsis, Critical Ill, and Active Epstein-Barr virus Infection to Identify Altered Pathways and Candidate Biomarkers. J Clin Immunol. 2023;43(8):1997–2010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wang Q, Gao QC, Wang QC, et al. A compendium of mitochondrial molecular characteristics provides novel perspectives on the treatment of rheumatoid arthritis patients. J Transl Med. 2023;21(1):561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Cai S, Chen Y, Lin S, et al. Multiple Processes May Involve in the IgG4-RD Pathogenesis: An Integrative Study via Proteomic and Transcriptomic Analysis. Front Immunol. 2020;11:1795. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Gong Y, Zhang P, Liu Z, et al. UPLC-MS based plasma metabolomics and lipidomics reveal alterations associated with IgG4-related disease. Rheumatology (Oxford). 2021;60(7):3252–61. [DOI] [PubMed] [Google Scholar]
- 14.Yan S, Peng Y, Wu Z, et al. Distinct metabolic biomarkers to distinguish IgG4-related disease from Sjogren’s syndrome and pancreatic cancer and predict disease prognosis. BMC Med. 2022;20(1):497. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Wallace ZS, Naden RP, Chari S, et al. The 2019 American College of Rheumatology/European League Against Rheumatism classification criteria for IgG4-related disease. Ann Rheum Dis. 2020;79(1):77–87. [DOI] [PubMed] [Google Scholar]
- 16.Wallace ZS, Khosroshahi A, Carruthers MD, et al. An International Multispecialty Validation Study of the IgG4-Related Disease Responder Index. Arthritis Care Res (Hoboken). 2018;70(11):1671–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Houghton DC, Troxell ML. An abundance of IgG4 + plasma cells is not specific for IgG4-related tubulointerstitial nephritis. Mod Pathol. 2011;24(11):1480–7. [DOI] [PubMed] [Google Scholar]
- 18.Carruthers MN, Khosroshahi A, Augustin T et al. The diagnostic utility of serum IgG4 concentrations in IgG4-related disease. Ann Rheum Dis. 2015;74(1). [DOI] [PubMed]
- 19.Zhang X, Lu H, Peng L, et al. The role of PD-1/PD-Ls in the pathogenesis of IgG4-related disease. Rheumatology (Oxford). 2022;61(2):815–25. [DOI] [PubMed] [Google Scholar]
- 20.Arora K, Kulkarni A, Pankaj A, et al. IgG4-related disease is characterised by the overexpression of immunomodulatory proteins. Histopathology. 2022;81(4):486–95. [DOI] [PubMed] [Google Scholar]
- 21.Kraehenbuehl L, Weng CH, Eghbali S et al. Enhancing immunotherapy in cancer by targeting emerging immunomodulatory pathways. Nat reviews Clin Oncol. 2022;19(1). [DOI] [PubMed]
- 22.Akiyama M, Suzuki K, Yoshimoto K, et al. Peripheral TIGIT + T Follicular Helper Cells That Produce High Levels of Interleukin-21 via OX40 Represent Disease Activity in IgG4-Related Disease. Front Immunol. 2021;12:651357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lu H, Wu X, Peng Y, et al. TSLP promoting B cell proliferation and polarizing follicular helper T cell as a therapeutic target in IgG4-related disease. J Transl Med. 2022;20(1):414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ito T, Wang YH, Duramad O, et al. TSLP-activated dendritic cells induce an inflammatory T helper type 2 cell response through OX40 ligand. J Exp Med. 2005;202(9):1213–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Gracias DT, Sethi GS, Mehta AK, et al. Combination blockade of OX40L and CD30L inhibits allergen-driven memory TH2 cell reactivity and lung inflammation. J Allergy Clin Immunol. 2021;147(6):2316–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhao Z, Liu Y, Bai M et al. Clinical profiles differ in IgG4-related disease with and without allergy: a large case-control study in China. Clin Exp Rheumatol. Published online February 11, 2023. [DOI] [PubMed]
- 27.Lee J, Park C, Kim HJ, et al. Stimulation of osteoclast migration and bone resorption by C-C chemokine ligands 19 and 21. Exp Mol Med. 2017;49(7):e358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Steines L, Poth H, Herrmann M, et al. B Cell Activating Factor (BAFF) Is Required for the Development of Intra-Renal Tertiary Lymphoid Organs in Experimental Kidney Transplantation in Rats. Int J Mol Sci. 2020;21(21):8045. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Aoyagi R, Maehara T, Koga R et al. Single-cell transcriptomics reveals granzyme K-expressing cytotoxic Tfh cells in tertiary lymphoid structures in IgG4-RD. J Allergy Clin Immunol Published online August 29, 2023; S0091-6749(23):01072-2. [DOI] [PubMed]
- 30.Manetti M, Guiducci S, Romano E, et al. Increased serum levels and tissue expression of matrix metalloproteinase-12 in patients with systemic sclerosis: correlation with severity of skin and pulmonary fibrosis and vascular damage. Ann Rheum Dis. 2012;71(6):1064–72. [DOI] [PubMed] [Google Scholar]
- 31.Asakage M, Usui Y, Nezu N, et al. Comprehensive Gene Analysis of IgG4-Related Ophthalmic Disease Using RNA Sequencing. J Clin Med. 2020;9(11):3458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Ding X, Yu Y, Su D, et al. IL-33/ST2 enhances MMP-12 expression by macrophages to mediate inflammatory and immune response in IgG4-Related Ophthalmic Disease. Cytokine. 2024;184:156754. [DOI] [PubMed] [Google Scholar]
- 33.Carruthers MN, Khosroshahi A, Augustin T, et al. The diagnostic utility of serum IgG4 concentrations in IgG4-related disease. Ann Rheum Dis. 2015;74(1):14–8. [DOI] [PubMed] [Google Scholar]
- 34.de Vries E, Tielbeke F, Hubers L, et al. IgG4/IgG RNA ratio does not accurately discriminate IgG4-related disease from pancreatobiliary cancer. JHEP Rep. 2020;2(4):100116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Herta T, Kersten R, Chang JC, et al. Role of the IgG4-related cholangitis autoantigen annexin A11 in cholangiocyte protection. J Hepatol. 2022;76(2):319–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Shiokawa M, Kodama Y, Sekiguchi K, et al. Laminin 511 is a target antigen in autoimmune pancreatitis. Sci Transl Med. 2018;10(453):eaaq0997. [DOI] [PubMed] [Google Scholar]
- 37.Perugino CA, AlSalem SB, Mattoo H, et al. Identification of galectin-3 as an autoantigen in patients with IgG4-related disease. J Allergy Clin Immunol. 2019;143(2):736–e7456. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Du H, Shi L, Chen P, et al. Prohibitin Is Involved in Patients with IgG4 Related Disease. PLoS ONE. 2015;10(5):e0125331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Liu H, Perugino CA, Ghebremichael M, et al. Disease Severity Linked to Increase in Autoantibody Diversity in IgG4-Related Disease. Arthritis Rheumatol. 2020;72(4):687–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Kiyama K, Kawabata D, Hosono Y, et al. Serum BAFF and APRIL levels in patients with IgG4-related disease and their clinical significance. Arthritis Res Ther. 2012;14(2):R86. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used during the current study are available from the corresponding author Yanying Liu ( [Liuyanying6850@126.com](mailto: Liuyanying6850@126.com) ) on reasonable request.