We collected the GSE83632 dataset from the GEO database, which includes expression data from 76 DLBCL samples and 83 normal samples. After standardization and log₂ transformation, the data were analyzed for differences between normal and tumor tissues using the “limma” R package. Additionally, we calculated the expression levels of stomatin in various tissues.

The data for the stomatin gene were obtained from a comprehensive meta-analysis of blood eQTLs (25). Disease-related data were obtained from the OPEN GWAS database, specifically using the dataset labeled “Diffuse large B-cell lymphoma” (finn-b-C3_DLBCL). We established a significance threshold of 5 × 10–⁸ for the instrumental variables of the stomatin gene, including an LD coefficient (r²) of < 0.45 and independent SNPs with a physical distance greater than 10,000 kilobases (26). For our MR analysis, we used methods such as random effects inverse variance weighting (IVW), MR-Egger, and weighted median (WM) analyses. Primary results were derived using the IVW method, with statistical significance defined as P < 0.05. To assess pleiotropy, we employed MR-Egger regression, where a significant intercept (P < 0.05) indicates substantial pleiotropy (27). We also used Cochran’s Q statistic to test for heterogeneity; a significant Cochran’s Q (P < 0.05) indicates considerable heterogeneity in the effects of various genetic variations (28).

Between December 2008 and December 2014, thirty-six patients diagnosed with orbital DLBCL and treated with surgery at the Department of Orbital Diseases of the Third Medical Center of the Chinese PLA General Hospital (Beijing, China) and the Ophthalmology Center of Shandong Second Medical University Affiliated Hospital were enrolled in the study and followed up until January 2022. All studies involving human subjects were conducted following the principles of the Declaration of Helsinki and received approval from the Medical Ethics Committee of Weifang Medical University. All clinical samples were collected with written informed consent from the patients or their immediate family members.

Diagnosis of orbital DLBCL was confirmed by experienced pathologists for all patients. Among the patients, 29 (80.6%) had the ABC subtype and 7 (19.4%) had the GCB subtype. The cohort included 20 male and 16 female patients, aged 24 and 85 years (median age 60 years). Patients were followed up in person for at least 5 years to monitor recurrences.

Surgically excised orbital tissues (classified as surgical discard) were fixed in 10% formalin for 12 hours and embedded in paraffin. Each sample was sectioned into two 4-μm slices and stained with an anti-stomatin antibody (Proteintech, Rosemont, USA). Erythrocyte staining served as an internal positive control. To verify the specificity of the staining, an immunohistochemical negative control was performed by replacing the primary antibody with PBS under the same experimental conditions. Two blinded senior pathologists independently assessed the stained sections. Stomatin-positive cells and staining intensity were evaluated in 10 randomly selected fields (100×magnification). The percentage of positive cells was categorized as 0 (≤5%), 1 (6%–25%), 2 (26%–50%), 3 (51%–75%), and 4 (75%). Staining intensity was scored as 1 (weak), 2 (moderate), and 3 (strong). The final score was calculated as the product of percentage and intensity scores, with low (final score <6) and high (final score ≥6) categories for stomatin expression (29).

OCI-Ly3 was cultured in OCI-Ly3 cell-specific medium (Pricella Life Science & Technology Co., Ltd, Wuhan, China), which contained 20% fetal bovine serum and 1% penicillin-streptomycin, mixed with RPMI 1640 medium. Cells were maintained at 37°C in a 5% CO₂ atmosphere with saturated humidity. Small interfering RNA (siRNA) targeting stomatin (GenePharma, Shanghai, China) was transfected into OCI-Ly3 cells using Lipofectamine 3000 (Thermo Fisher Scientific, MA, USA). Negative control (NC) siRNA (GenePharma, Shanghai, China) was transfected into cells as the control group. Si-stom (sense 5′-GCAGUCUACUCUGGAUGAUTT-3′, antisense 5′-AUCAUCCAGAGUAGACUGCTT-3′), and NC siRNA (sense 5′-UUCUCCGAACGUGUCACGUTT-3′, antisense 5′-ACGUGACACGUUCGGAGAATT-3′) were designed and synthesized by GenePharma.

After 24 hours of transfection with siRNA negative control (NC) and siRNA targeting stomatin (si-stom), total RNA was extracted using TRnaZol Reagent (New Cell & Molecular Biotechnology, Suzhou, China). Reverse transcription was performed to synthesize cDNA using the Primescript Master Mix (Perfect Real Time) kit (Takara Bio Inc, Shiga, Japan). Quantitative Reverse-Transcription PCR (qRT-PCR) amplification was conducted using the TB Green Premix Ex Taq II kit (Takara Bio Inc, Shiga, Japan). Equal amounts of reverse transcription products were mixed with other components to form a 10μl reaction system. Stomatin mRNA expression in OCI-Ly3 cells was normalized to GAPDH as the internal reference gene. Specific primers were synthesized by GenePharma (Shanghai, China): stomatin, forward 5′-AAAGGTGGAGCGTGTGGAAA-3′, reverse 5′-CTTCGGCTGCAATAACCTTGG-3′; GAPDH, forward 5′-AAGCTCATTTCCTGGTATGACAA-3′, reverse 5′-CTTACTCCTTGGAGGCCATGT-3′.

After 48 hours of transfection with siRNA, NC cells and si-stom cells were seeded in 96-well plates at a density of 5 × 10³ cells per well, with a volume of 100 μl per well. Ten microliters of CCK-8 (Beyotime Biotechnology, Shanghai, China) was added to each well, and the plates were incubated at 37°C for 4 hours. The absorbance at 450 nm was measured using a microplate reader.

After transfection, NC and si-stom cells were collected by centrifugation, resuspended in RPMI-1640 serum-free medium (Gibco, Thermo Fisher Scientific, MA, USA), and adjusted to a density of 1.5 × 10⁶ cells/ml. A 200 μl aliquot of the cell suspension was then added to the Transwell chambers (Corning, New York, USA). The cells were incubated at 37°C for 48 hours, after which they were collected for migration analysis. For the invasion assay, cells were resuspended in RPMI-1640 serum-free medium and seeded onto a polycarbonate membrane pre-coated with a uniform layer of Matrigel (Corning, New York, USA) in the Transwell chambers. The cell density was adjusted to 1.5 × 10⁶ cells/ml, and 200 μl of the cell suspension was added to each chamber. The cells were incubated again at 37°C for 48 hours. After incubation, the cells were fixed in 4% paraformaldehyde (Solarbio, Beijing, China) for 10 minutes and stained with 0.1% crystal violet (Solarbio, Beijing, China) for 15 minutes. The stained cells were observed under an inverted microscope at 200× magnification and photographed for analysis.

After OCI-Ly3 cells were transfected with siRNA, NC and si-stom cells were collected, refreshed with new medium, and re-cultured for 48 hours. Apoptosis in the NC and si-stom cells was then assessed using the Annexin V-FITC/7-AAD Fluorescence Double Staining Apoptosis Detection Kit (Pricella, Wuhan, China).

After transfection of OCI-Ly3 cells, NC and si-stom cells were collected, and proteins were extracted using RIPA lysis buffer (Beyotime Biotechnology, Shanghai, China). The concentration of the target protein was determined using the BCA method, and equal amounts of protein were separated by SDS-PAGE. The proteins were transferred to a PVDF membrane at a constant current of 400 mA and blocked with blocking buffer (Boster, Wuhan, China). Membranes were incubated overnight with antibodies against stomatin (1:1000, Proteintech, Rosemont, USA) and β-actin (1:1000, Proteintech, Rosemont, USA). After washing with Tris-HCl-Tween buffered saline solution (TBST), the membranes were incubated at room temperature for 1 hour with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Proteintech, Rosemont, USA). Finally, the membranes were developed using Super ECL Plus chemiluminescent substrate (Applygen, Beijing, China), and images were analyzed using Image J.

Data analysis was conducted using GraphPad Prism v10.0, SPSS 25, and ImageJ software. Statistical analyses included unpaired t-tests and two-way ANOVA, with significance levels indicated in the corresponding figures. Spearman correlation coefficients were calculated to evaluate the relationships between two variables. The strength of correlation was calculated (r). P value < 0.05 was considered statistically significant.

To better understand the role of stomatin in different tumor types, we first conducted a pan-cancer analysis of the stomatin gene. The results showed that the expression levels of stomatin vary across different tumor types ( Supplementary Figure S1 ). The goal of this study is to explore the specific expression pattern of stomatin in orbital DLBCL, particularly in comparison to other tumor types, where stomatin exhibits relatively higher expression levels. Differential expression analysis of the GSE83632 dataset identified 1531 differentially expressed genes (adjusted p-value < 0.05 and |log₂FC| ≥ 0.5), and their expression is shown in a heatmap ( Figures 1A, B ). Among them, stomatin was highly expressed in tumor tissues, demonstrating a degree of specificity in DLBCL ( Figure 1C ).

To explore the causal relationship between the stomatin gene and DLBCL, we performed MR analysis. The aim of this analysis was to assess the potential causal relationship between stomatin and orbital DLBCL through genetic variation, thereby avoiding confounding factors that may exist in traditional observational studies. At the genome-wide level, we identified and screened 20 independent stomatin SNPs that were not in linkage disequilibrium. The IVW analysis showed a positive causal association between stomatin gene and DLBCL [OR = 1.40, 95% CI (1.01, 1.95), P = 0.04, which is statistically significant at P < 0.05]. As shown in Supplementary Table S1 , the Q-pvalue from the IVW analysis was 0.91, indicating no evidence of heterogeneity. Additionally, the MR-Egger intercept test suggested that directional pleiotropy was unlikely to introduce bias (P = 0.14), confirming that pleiotropy was not present in our data ( Figures 1D, E ).

After a minimum follow-up of 5 years, 50% of patients (18/36) experienced tumor recurrence, and 44.4% (16/36) died from their tumors. The median disease-free survival was 27.5 months (range, 8–70 months), and the median overall survival was 49 months (range, 9–89 months). Only three patients received Rituximab treatment, of whom two patients received the R-CHOP regimen and one patient received the R-CVP regimen. None of the patients has a poor response to Rituximab. Immunohistochemical analysis showed that stomatin protein expression was localized to the cell membrane in all orbital DLBCL tissues, with varying rates of positive expression ( Figures 2A, B ). The percentage of the stomatin stained cells varied from 15% to 56% in these samples. The median proportion of stomatin stained cells was 37.5%. Out of 36 patients with ocular DLBCL, 18 showed low expression (final score <6) and 18 showed high expression (final score ≥6) of stomatin. Intraorbital tumor sections of patients with orbital MALT lymphoma, follicular lymphoma, orbital lymphoreactive hyperplasia, and orbital inflammatory pseudotumor were also stained. In orbital MALT lymphoma and follicular lymphoma, staining intensity was weaker and the proportion of stomatin stained cells was less than 10% ( Figures 2C, D ). In orbital lymphoreactive hyperplasia and orbital inflammatory pseudotumor, the proportion of stomatin-positive cells was less than 5% ( Figures 2E, F ).

As shown in Table 1 , there were 18 patients each with high and low stomatin expression. Stomatin expression was significantly higher in Ann Arbor stage II-IV tumors compared to stage I tumors (P=0.002). Furthermore, patients with low stomatin expression had a significantly lower recurrence rate than those with high stomatin expression (P=0.002). Additionally, high stomatin expression was associated with higher lactate dehydrogenase (LDH) levels (P=0.000).

We also examined the expression of stomatin mRNA using qRT-PCR in the biopsied specimen of orbital lymphoma. The expression of stomatin mRNA and protein in orbital tumor tissue of patients with orbital DLBCL was significantly higher than that of orbital MALT lymphoma and orbital follicular lymphoma (P<0.05) ( Figures 3A, B ). It was found that the expression of stomatin mRNA is closely related to the expression of stomatin protein (r=0.9381, P<0.0001) ( Figure 3C ). The expression of stomatin mRNA was closely related to overall survival (r = -0.6272, P<0.0001) and disease-free survival (r = -0.6479, P<0.0001) ( Figure 3D ).

Univariate analysis showed significant associations between stomatin protein expression, Ann Arbor stage, and LDH levels with disease-free and overall survival in patients ( Table 2 ). Multivariate analysis was first performed using a Cox regression model ( Table 3 ) and further adjusted using a Bayesian regression model ( Supplementary Tables S2, S3 ). The adjusted results indicated that Ann Arbor stage and stomatin expression were independent prognostic factors for both disease-free and overall survival ( Figure 4 ). Survival analysis further revealed that patients with high stomatin expression had significantly shorter survival times than those with low expression ( Figure 5 ).

To verify the promoting effect of stomatin protein on orbital DLBCL, we performed in vitro experiments to knock down the expression of stomatin in two DLBCL cell line. The transfection efficiency was validated at both the molecular and protein levels. The results of qRT-PCR and Western blot analyses showed that the expression of stomatin at both the mRNA and protein levels was significantly reduced in si-stom cells compared to NC cells after siRNA transfection. The difference was statistically significant ( Figure 6 ).

We used the CCK-8 assay to assess the cell proliferation ability of si-stom and NC cells after transfection. The results are shown in Figure 7A . The proliferation ability of si-stom cells was significantly reduced compared to NC cells within 5 days following stomatin gene knock down, indicating that stomatin has a promoting effect on the growth of DLBCL cells.

To further verify the promoting effect of stomatin protein on DLBCL, we conducted transwell experiments to assess the invasion and migration abilities of cells. The results, as shown in Figure 7B , indicate that after transfection, the invasion and migration abilities of si-stom cells were significantly weakened. These findings suggest that stomatin promotes the invasion and migration capabilities of DLBCL.

To further elucidate the promoting effect of stomatin protein on DLBCL, we stained cells transfected with the stomatin gene using Annexin V-FITC and 7-AAD, and detected the effect of stomatin protein on cell apoptosis using flow cytometry. As shown in Figure 7C , Q1 represents necrotic cells, Q2 represents late apoptotic cells, Q3 represents early apoptotic cells, and Q4 represents live cells. The results showed that the apoptosis rate of si-stom cells was 30.21%, while the control group had an apoptosis rate of 20.74%. There was a significant difference between the experimental group and the control group, and the difference was statistically significant (P < 0.05).