In our study, 40 pSS patients and 26 healthy controls (HCs) were recruited as research subjects. All pSS patients were diagnosed according to the classification criteria of Sjögren’s syndrome revised by American-European consensus group (AECG) in 2002 (16), and none of them received medication. We exclusively enrolled treatment-naïve patients at initial diagnosis, with documented absence of previous immunomodulatory therapy (including systemic corticosteroids, cytokine inhibitors, or other immunosuppressants). Clinical data and laboratory parameters of pSS patients were collected. Based on the ESSDAI score (17), pSS patients with ESSDAI≥5 were classified as active group, while those with ESSDAI<5 were considered inactive group. Fresh venous peripheral blood samples were collected from pSS patients and HCs with fasting in the morning after an overnight fast. Partial patients underwent pathological examination of salivary gland(SG) biopsy, and the remaining SG tissue was preserved after inspection for our study. Our study was approved by the Ethics Committee of the First Affiliated Hospital of Soochow University in 2020 (Ethical No. 2020105).Informed consent was obtained from all study participants.

Non-obese diabetic (NOD) mice were utilized as mice models for pSS, and institute of cancer research(ICR) mice were taken as mice models for control mice (18, 19). The NOD(n=5) and ICR(n=5) mice(female, aged 6 weeks) were purchased from Changzhou Cavens Experimental Animal Co., Ltd. The animal research was approved by the Animal Ethics Committee of Soochow University(Ethics No. SUDA20240724A03).

Fluorescence-conjugated anti-human monoclonal antibodies were adopted in this study as following: anti-CD3-FITC, anti-CD8-PE-Cy5, anti-CD19-PE-Cy5, anti-CD56-FITC, anti-CD14-FITC, anti-CD226-PE, anti-CD27-PE-Cy7, anti-IgD-FITC, anti-CD21-APC-Cy7, anti-CD38-AF700, anti-CD80-PE-Cy7, anti-CD86-FITC, anti-CD137-APC-Cy7, anti-CD69-AF700, anti-IgG-PE-Cy7, anti-IgA-FITC, anti-Ki67-AF700, anti-IFN-γ-PE-Cy7, anti-TNF-α-PE-Cy7. And fluorescence-conjugated anti-mouse monoclonal antibodies were utilized as following: anti-CD3-PE-Cy5, anti-CD4-FITC, anti-CD8-PE-Cy7, anti-NK1.1-APC-Cy7, anti-B220-FITC, anti-CD11b-PerCP-Cy5.5, anti-CD226-PE. These above anti-human and anti-mouse monoclonal antibodies were all purchased from Biolegend (USA).

The human anti-CD3, anti-CD8, anti-CD19, anti-CD56, anti-CD14, anti-CD27, anti-IgD, anti-CD21, anti-CD38, anti-CD80, anti-CD86, anti-CD137, anti-CD69, anti-IgG, anti-IgA, and anti-CD226 antibodies were added into 50µL peripheral blood samples and incubated for 30 minutes. Then erythrocyte lysis was added into the samples for red blood cell lysis. For intracellular staining, peripheral blood mononuclear cells(PBMCs) were isolated from human samples, and added phorbol 12‐myristate 13‐acetate (PMA) and ionomycin for stimulating cells, then stained by human anti-CD19 and anti-CD226 antibodies. Subsequently, Fixation and permeabilization buffers were added, followed by staining with human anti-Ki67, anti-IFN-γ, and anti-TNF-α antibodies. All samples were conducted by flow cytometry. For NOD and ICR mice, spleen single-cell suspensions were prepared and stained with mouse anti-CD3, anti-CD4, anti-CD8, anti-NK 1.1, anti-B220, anti-CD11b and anti-CD226 antibodies for processing by fow cytometry(Cyto FLEX, Beckman Coulter, USA).All the data were analyzed by FlowJo 10.8.1.

Paraffin sections of SG tissues from pSS patients and HCs were deparaffinized, underwent antigen retrieval, and were blocked. Subsequently, the sections were incubated with a human anti-CD226 antibody as the primary antibody, followed by incubation with a goat anti-mouse IgG antibody as the secondary antibody. Then 3,3’-Diaminobenzidine(DAB) staining was utilized to detect the expression of CD226 in SGs. Using tyramide signal amplification(TSA) fluorescence labeling technique, the co-expression of CD226 and CD20 in SGs tissue was detected according to the aforementioned method, utilizing human anti-CD226 and human anti-CD20 antibody respectively. The nucleus was stained by 4’,6-diamidino-2-phenylindole(DAPI), CD226 was labeled by red fluorescence, and CD20 was labeled by green fluorescence.

Using a fluorescence-activated cell sorter (FACS) to sort CD226⁺ CD19⁺ B cells and CD226⁻ CD19⁺ B cells from splenocytes of NOD mice. Total RNA was extracted with Trizol (Invitrogen) and assessed with Agilent 2100 BioAnalyzer(Agilent Technologies, Santa Clara, CA, USA) and Qubit Fluorometer (Invitrogen). RNA-seq libraries were generated and sequenced by CapitalBio Technology (Beijing, China). The cDNA was synthesized from the extracted RNA using a reverse transcription kit, and labeled with fluorescent dyes. The purified DNA was denatured at high temperature (typically at 95°C for 3 minutes), followed by hybridization of the labeled cDNA to the mouse gene expression microarray chip(Agilent) according to the manufacturer’s instructions. The hybridization was then carried out in a hybridization oven at 42°C for 16–20 hours. The microarray chip was washed to remove unbound cDNA using the recommended buffers and then scanned to detect fluorescence signals.

The gene expression analyses were performed with StringTie(v1.3.3b). DESeq(v1.28.0) was used to analyze the differentially expressed genes(DEGs) between samples. Thousands of independent statistical hypothesis testing was conducted on DEGs, separately. Then a p-value was obtained, which was corrected by FDR method. The corrected P-value (q-value) was calculated by correcting using BH method and used to conduct significance analysis. Parameters for classifying significantly DEGs were≥2-fold differences (|log2FC|≥1, FC: the fold change of expressions) in the transcript abundance and p ≤ 0.05. The annotation of the DEGs were performed based on the information obtained from the database of ENSEMBL, NCBI, Uniprot, GO, and Kyoto encyclopedia of genes and genomes(KEGG).

Flow cytometry data were processed and analyzed using FlowJo software (version 10.8.1). Images were processed using ImageJ software. Statistical analysis and data visualization were performed using GraphPad Prism software (version 8.0.2). The Shapiro-Wilk test was applied to assess data normality (n<50). Normally distributed measurement data were expressed as mean ± standard deviation ( x ¯ ± s), while non-normally distributed data were presented as median (minimum, maximum). Comparisons between two groups were analyzed using the Student’s t-test for normally distributed data and the Mann-Whitney U test for non-normally distributed data. For correlation analysis, the Pearson correlation coefficient was used for normally distributed data, and the Spearman correlation coefficient was applied for non-normally distributed data. P-value<0.05 was considered statistically significant, with significance levels denoted as follows: *P<0.05, **P<0.01, and ***P<0.001, ****P<0.0001.

The pSS patients who participated in this study were matched with HCs by gender, and there was no statistical difference in age (46.75 ± 13.10 vs. 44.35 ± 9.28 (years), P>0.05).The basic information of pSS patients and HCs enrolled in this study is shown in Table 1 .

Flow cytometry was utilized to identify the expression levels of CD226 on T cells, B cells, CD56⁺ NK cells and CD14⁺ monocytes in the peripheral blood of pSS patients and HCs. The gating strategies are depicted in Supplementary Figure 1 . The results revealed that, compared with HCs, the percentage of CD226 on CD3⁺ T cells, CD3⁺CD8^- T cells,CD8⁺ T cells as well as CD19⁺ B cells and CD14⁺ monocytes was significantly increased in patients with pSS( Figure 1B ). In contrast, CD226 percentage on CD56⁺ NK cells was significantly reduced in pSS patients compared to HCs. Representative FACS plots are shown in Figure 1A . Next, we conducted a correlation analysis between the expression levels of CD226 on CD3⁺CD8^- T cells, CD8⁺ T cells, CD19⁺ B cells, CD56⁺ NK cells, and CD14⁺ monocytes and clinical activity parameters in patients with pSS. The findings indicated that in pSS patients, the percentage of CD226⁺CD3⁺CD8^- T cells did not exhibit any significant correlation with the clinical parameters of pSS. The percentage of CD226⁺ CD8⁺ T cells demonstrated a positive correlation with erythrocyte sedimentation rate (ESR), while it showed a negative correlation with complement C3 (C3). However, no significant correlation was observed between the percentage of CD226⁺ CD8⁺ T cells and rheumatoid factor (RF), immunoglobulin G (IgG), European League Against Rheumatism Sjögren’s Syndrome Patient-Reported Index (ESSPRI), European Sjögren’s Syndrome Disease Activity Index (ESSDAI), or complement C4 (C4). The percentage of CD226⁺CD56⁺ NK cells was positively correlated with C3 and C4, while negatively correlated with ESSPRI and ESSDAI, and not correlated with ESR, RF and IgG. There was a positive correlation between the percentage of CD226⁺CD14⁺ monocytes and RF, IgG and ESSDAI, and a negative correlation with C3, but no correlation with ESR, ESSPRI and C4. A significant positive correlation was exhibited between the percentage of CD226⁺CD19⁺ B cells and ESR, RF, IgG, ESSPRI and ESSDAI. Conversely, a significant negative correlation was identified between CD226⁺CD19⁺ B cells and C3 and C4 ( Figure 1C ). In brief, the findings demonstrated that CD226⁺CD19⁺ B cells were more closely related to the disease activity and clinical parameters of pSS patients. We further discovered that CD226 was expressed in the SG tissue of pSS patients and co-expressed with CD20⁺ B cells ( Figures 1D, E ).

Then, we analyzed the differences of CD226⁺ CD19⁺ B cells percentage in the peripheral blood of pSS patients with different clinical features and autoantibodies positive or negative. We found that the percentage of CD226⁺ CD19⁺ B cells in pSS patients with arthralgia, fatigue, decayed tooth, xerostomia, interstitial lung disease(ILD), leukopenia and high IgG was elevated compared to those without these clinical manifestations ( Figure 2A ). Among patients who were positive for anti-SSA/Ro52 antibody, anti-SSB/La antibody, and anti-centromere antibody, the percentage of CD226⁺ CD19⁺ B cells was significantly higher than that in patients who were negative for these antibodies ( Figure 2B ). The representative FACS plots of CD226⁺ CD19⁺ B cells percentage in the peripheral blood of pSS patients with different clinical features and auto-antibody statuses were displayed in Supplementary Figure 2 . Moreover, compared to patients with inactive disease, those with active disease exhibited a marked increase in the percentage of CD226⁺ CD19⁺ B cells ( Figure 2C ). The ROC curve analysis revealed that CD226⁺ CD19⁺ B cells have significant discriminative capability in distinguishing pSS patients from HCs and in evaluating disease activity ( Figures 2E, F ). In addition, we followed up with seven patients who had received treatment for over two months and observed a downward trend in the percentage of CD226⁺ CD19⁺ B cells after treatment, however, there was no statistical significance ( Figure 2D ).

We detected the expression levels of CD226 in different B cell subsets in the peripheral blood of pSS patients and HCs by multi-color flow cytometry. B cell subsets were classified according to CD27/IgD and CD21/CD38 as follows: CD27⁻ IgD⁺ as naive B cells, CD27⁺ IgD⁺ as unswitched memory B cells, CD27⁺ IgD⁻ as switched memory B cells, CD27⁻ IgD⁻ as double-negative B cells and CD21⁻ CD38⁺ as plasmablasts (20, 21). The gating strategies of CD27, IgD, CD21 and CD38 on CD226⁺ CD19⁺ B cells are shown in Supplementary Figure 3A .

The results showed that the percentage of CD226⁺ cells on switched memory B cells and unswitched memory B cells was higher than that of CD226⁻ cells in pSS patients(switched memory B cells%: 20.16% ± 12.38% vs. 10.53% ± 5.80%, P=0.03; unswitched memory B cells%: 52.63% ± 18.87% vs. 35.02% ± 10.11%, P=0.03)( Figure 3A ). There was no significant difference in the expression of CD226 between pSS patients and HCs on naive B cells, unswitched memory B cells, switched memory B cells and double negative B cells ( Supplementary Table 1 ). No significant difference was observed in the expression of CD226⁺ and CD226⁻ cells on naive B cells and double-negative B cells in the peripheral blood of pSS patients ( Supplementary Table 2 ). Compared to CD226⁻cells, the percentage of CD226⁺ cells on plasmablasts was significantly elevated (1.30% ± 0.81% vs. 0.40% ± 0.22%,P=0.002, Figure 3B ). In comparison with HCs, the percentage of CD226 on plasmablasts and CD21⁻ CD38⁻ B cells in pSS patients significantly increased ( Supplementary Table 1 ). In addition, the distribution of CD226⁺ and CD226⁻ cells on B cell subsets did not exhibit a significant difference among HCs ( Supplementary Table 3 ).

In order to evaluate the association between CD226⁺ B cells expression and activated/inflammatory phenotypes in pSS patients, we measured the expression levels of costimulatory molecules(CD80,CD86 and CD137), activation markers(CD69), proliferation markers(Ki67), immunoglobulins (IgG and IgA), and proinflammatory cytokines (TNF-α and IFN-γ) in the peripheral blood of patients with pSS by fow cytometry.

The gating strategies of CD80, CD86, CD137, CD69, Ki67, IgG, IgA, TNF-α and IFN-γ on CD226⁺ CD19⁺ B cells are presented in Supplementary Figures 3B-E . It was observed that in patients with pSS, CD226⁺ CD19⁺ B cells exhibited significantly elevated expression of costimulatory molecules(CD80, CD86, CD137), activation markers(CD69), proliferation markers(Ki67), immunoglobulin production (IgG, IgA), and pro-inflammatory cytokines (TNF-α, IFN-γ) compared to their CD226⁻ CD19⁺ B cell counterparts in the peripheral blood ( Figure 4 ). No significant difference was detected in the expression levels of above indicators between CD226⁺ CD19⁺ B cells and CD226⁻ CD19⁺ B cells subsets in HCs ( Figure 4 ). In comparison with HCs, the percentage of costimulatory molecules(CD86, CD137), proliferation markers(Ki67), and pro-inflammatory cytokines (TNF-α, IFN-γ) on CD226⁺ CD19⁺ B cells was significantly upregulated in the peripheral blood of pSS patients ( Figure 4 ). The representative FACS plots of costimulatory molecules(CD80, CD86, CD137), activation markers(CD69), proliferation markers(Ki67), immunoglobulins (IgG, IgA), and pro-inflammatory cytokines (TNF-α, IFN-γ) of CD226⁺ CD19⁺ B cells and CD226⁻ CD19⁺ B cells in the peripheral blood of pSS patients and HCs were exhibited in Supplementary Figure 4 . The representative FACS plots of costimulatory molecules (CD86, CD137), proliferation markers(Ki67), and pro-inflammatory cytokines (TNF-α, IFN-γ) of CD226⁺ CD19⁺ B cells in the peripheral blood of pSS patients and HCs are shown in Supplementary Figure 5 .

To investigate the functional role of CD226⁺ CD19⁺ B cells in pSS, CD226⁺ CD19⁺ and CD226⁻ CD19⁺ B cell subsets were isolated from splenic mononuclear cells of NOD mice via FACS, followed by transcriptome-wide microarray profiling to delineate their differential molecular signatures. Using multi-color flow cytometry, we first detected CD226 expression levels on major immunocyte subsets, including CD3⁺ T cells (CD4⁺ and CD8⁺ subsets), B220⁺ B cells, NK1.1⁺ NK cells, and CD11b⁺ monocytes in splenic mononuclear cells of NOD mice and ICR controls. In NOD mice, the percentage of CD226 on CD3⁺ T cells, CD4⁺ T cells and B220⁺ B cells was significantly higher than that of ICR mice, whereas the percentage of CD226 on NK1.1⁺ NK cells was markedly reduced in NOD mice relative to ICR mice(CD3⁺ T cells%:46.30% ± 4.48% vs. 35.16% ± 7.46%, P=0.0210; CD4⁺ T cells%: 22.06% ± 3.03% vs. 11.02% ± 4.08%, P=0.0013; B220⁺ B cells%: 12.81% ± 3.76% vs. 2.74% ± 1.15%,P=0.0004; NK1.1⁺ NK cells%: 17.38% ± 4.45% vs. 41.54% ± 3.97%,P=0.0032, Figure 5A ). The representative FACS plots of CD226 on CD3⁺ T cells CD4⁺ T cells, B220⁺ B cells, and NK1.1⁺ NK cells in splenic mononuclear cells of NOD and ICR mice are displayed in Supplementary Figure 6 . Through microarray analysis, we found that there were 129 upregulated DEGs and 8 downregulated DEGs between CD226⁺ CD19⁺ B cells and CD226⁻ CD19⁺ B cells ( Figure 5B ). The differential expression of Cd79a, Il17d, Il1b, and Cd8a between CD226⁺ CD19⁺ B cells and CD226⁻ CD19⁺ B cells held marked biological significance, potentially reflecting their distinct functional roles in immune regulation or disease pathogenesis ( Figure 5C ). KEGG pathway analysis demonstrated that CD226⁺ CD19⁺ B cells were significantly enriched in multiple critical biological pathways, including but not limited to complement and coagulation cascades, PI3K-Akt signaling pathway, extracellular matrix (ECM)-receptor interactions, and platelet activation mechanisms. Furthermore, these cells exhibited obvious associations with neoplastic pathologies (notably breast cancer and acute myeloid leukemia),as well as autoimmune disorders such as systemic lupus erythematosus (SLE)( Figures 5D-F ). Gene set enrichment analysis (GSEA) revealed that CD226⁺ CD19⁺ B cells mediated their biological functions through significantly enriched signaling pathways, including ECM-receptor interaction, cytokine-cytokine receptor interaction, and cell adhesion molecules (CAMs), with false discovery rate (FDR)-adjusted p-values <0.05( Figure 5G ).