Identification of GALNT18, HOMER3, and NRP2 as shared molecular signatures associated with stromal remodeling and immune suppression in oral mucosal malignancy-associated disorders

Identification of GALNT18, HOMER3, and NRP2 as shared molecular signatures associated with stromal remodeling and immune suppression in oral mucosal malignancy-associated disorders | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Identification of GALNT18, HOMER3, and NRP2 as shared molecular signatures associated with stromal remodeling and immune suppression in oral mucosal malignancy-associated disorders Jiabin Zhao, Yuanyi Chen, Ting Tang, Weike Lu, Yue Pan, Zihua Qi, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9127273/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 12 You are reading this latest preprint version Abstract Background. Oral submucous fibrosis (OSF), oral leukoplakia (OLK), and oral squamous cell carcinoma (OSCC) are prevalent disorders associated with oral mucosal malignancy. However, the common mechanisms underlying the progression and shared characteristics of these malignancy-associated disorders remain unclear. This study aims to investigate the core common differentially expressed genes (DEGs) shared by OSF, OLK, and OSCC, providing novel targets for the diagnosis and evaluation of oral mucosal malignancy-associated disorders. Methods. We mined the Gene Expression Omnibus (GEO) database to pinpoint overlapping transcriptomic signatures across OSF, OLK, and OSCC. Functional enrichment analyses, utilizing Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), were performed to map biological pathways. Protein–protein interaction (PPI) networks were constructed to identify candidate genes. These candidates were further screened via Least Absolute Shrinkage and Selection Operator (LASSO) regression and Receiver Operating Characteristic (ROC) validation to prioritize core markers. Additionally, immune infiltration assessments and single-gene Gene Set Enrichment Analysis (GSEA) were conducted to explore mechanistic links, while survival analysis was employed to evaluate prognostic value. Results. We identified 94 co-expressed genes, which were primarily clustered in biological processes related to extracellular matrix (ECM) remodeling, epithelial–mesenchymal transition (EMT), and immune regulation. Through rigorous screening, GALNT18 , HOMER3 , and NRP2 were prioritized as the final core genes. These markers demonstrated consistent correlations with specific infiltrating immune cells and ECM-related signaling pathways. Notably, while all three genes served as robust diagnostic markers, high NRP2 expression was specifically associated with poor overall survival. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Introduction Oral squamous cell carcinoma (OSCC) is a common head and neck malignancy, with nearly 400,000 new cases diagnosed annually [1] . Despite therapeutic advances, the overall five-year survival rate remains approximately 50% [2] . OSCC can arise from clinically normal-appearing mucosa or coexist with/succeed oral potentially malignant disorders (OPMDs) [3] . Oral submucous fibrosis (OSF) and oral leukoplakia (OLK) are the two predominant types of OPMDs. OSF, strongly associated with long-term betel nut chewing, is characterized by subepithelial fibrosis and carries a malignant transformation rate of approximately 4%–10% [4] . OLK presents as white patches on the oral mucosa, with a prevalence of about 4% and a malignant transformation rate ranging from 1% to 10% [3] . While the majority of OSF and OLK cases do not eventually progress to OSCC, they frequently exhibit varying degrees of cellular atypia and alterations in the immune microenvironment, reflecting molecular features similar to those of OSCC [5-7] . The molecules mediating these early changes represent potential targets for risk stratification and prognostic assessment; however, they are often difficult to identify effectively through routine histopathological examination [8,9] . The development of high-throughput sequencing technologies has facilitated in-depth investigations into disease mechanisms. Nevertheless, most previous studies have focused on single disease entities, lacking a systematic analysis across multiple conditions [10,11] . By integrating transcriptomic data from multiple diseases, this study screened for core genes commonly differentially expressed in OSF, OLK, and OSCC. Our objective is to reveal shared molecular signatures and enhance the understanding of the pathogenesis of oral malignancy-associated disorders [12] , thereby providing a theoretical foundation for the development of early assessment strategies and therapeutic interventions [13,14] . Materials & Methods 2.1. Data Acquisition and Processing Transcriptomic profiles were retrieved from the Gene Expression Omnibus (GEO) database [15] . The study included: (1) an OSF dataset (GSE64216) comprising 4 OSF cases and 2 normal mucosal controls; (2) an OLK dataset (GSE246050) containing 6 OLK cases and 3 controls; and (3) an OSCC dataset (GSE30784) with 167 OSCC cases and 45 controls. These datasets served as the foundation for identifying DEGs. Additionally, three independent datasets (GSE25099 for OSCC, GSE12586 for OSF, and GSE227919 for OLK) were utilized for external validation. 2.2. Differential Expression Analysis DEGs were identified using the GEO2R tool. To ensure statistical rigor, we applied a significance threshold of P 0.585 [16] . Genes were subsequently stratified into upregulated and downregulated categories. To pinpoint shared molecular drivers, we utilized the bioinformatics platform (https://www.bioinformatics.com.cn/) to generate Venn diagrams, isolating genes with consistent expression trends across all three pathological states. 2.3. Functional Enrichment Analysis Biological insights into the co-expressed genes were gained through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses using the DAVID platform [17] (https://davidbioinformatics.nih.gov/). A P-value of < 0.05 was considered statistically significant. 2.4. Protein–Protein Interaction (PPI) Network Construction A PPI network was constructed using the STRING database [18] (https://string-db.org/) with a minimum interaction confidence score of 0.4. Isolated nodes were excluded to focus on functional connectivity. The network was visualized in Cytoscape, and the CytoNCA plugin [19] was employed to calculate topological properties. The top 15 genes ranked by Degree centrality were identified as candidate hub genes. 2.5. Feature Selection and Diagnostic Validation To identify the most robust biomarkers, Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression was performed using the “glmnet” R package [20] . A 10-fold cross-validation was employed to determine the optimal penalty parameter. The diagnostic performance of the selected core genes was evaluated using Receiver Operating Characteristic (ROC) curves, with the Area Under the Curve (AUC) serving as the primary metric of discrimination. 2.6. Immune Landscape Profiling The immune microenvironment was characterized using the CIBERSORT algorithm via the bioinformatics platform [21] . The LM22 signature matrix was used to deconvolute gene expression data and estimate the relative abundance of 22 distinct infiltrating immune cell types. Spearman’s correlation analysis was then conducted in R to assess the relationship between core gene expression and specific immune cell subsets. 2.7. Single-Gene Gene Set Enrichment Analysis (GSEA) To elucidate the specific pathways regulated by each core gene, we performed single-gene GSEA. This involved calculating the Spearman correlation between the target gene and all other genes in the genome, ranking them, and testing for pathway enrichment using R software. 2.8. Survival Analysis The clinical prognostic value of the core genes was assessed using the GEPIA tool (http://gepia.cancer-pku.cn/), utilizing data from the TCGA-HNSC cohort. Patients were stratified into high- and low-expression groups based on the upper (75%) and lower (25%) quartile thresholds. Overall Survival (OS) was compared between groups using the Log-rank test. Results 3.1. Identification of DEGs in OSF, OLK, and OSCC Our analysis yielded 3506 differentially expressed genes (DEGs) in OSF (1622 up-regulated, 1884 downregulated), 3020 in OLK (1575 upregulated, 1445 downregulated), and 4764 in OSCC (2558 upregulated, 2206 downregulated). Volcano plots and hierarchical clustering (Fig. 1 A–F) confirmed that these distinct expression profiles could effectively differentiate pathological tissues from normal mucosa. 3.2. Identification of Co-Expressed Genes Intersection analysis revealed 94 co-expressed genes (54 upregulated and 40 downregulated) shared across OSF, OLK, and OSCC (Fig. 2 A-B). These overlapping genes represent potential common denominators in the pathogenesis of these conditions. 3.3. Functional Enrichment of Co-Expressed Genes GO analysis indicated that these genes are primarily involved in extracellular matrix organization, specifically skin barrier establishment, inflammatory responses, and bone remodeling (BP). In terms of cellular components (CC), they are enriched in cell junctions and filopodia. Molecular functions (MF) highlighted activities such as protease inhibition and growth factor binding (Fig. 3 A-B). KEGG pathway analysis further linked these genes to metabolic reprogramming (central carbon metabolism in cancer, arachidonic acid metabolism) and structural integrity (cornified envelope formation) (Fig. 3 C). 3.4. PPI Network and Hub Gene Screening The PPI network analysis, refined by CytoNCA, identified the top 15 genes based on topological importance (Fig. 4 A–B). These hub genes are positioned at critical junctions of the interaction network, suggesting pivotal regulatory roles. 3.5. Selection of Core Genes via LASSO Regression To filter for genes most relevant to malignant transformation, we applied LASSO regression to the 15 hub genes within the OSCC cohort. This process narrowed the list to five core genes: C1GALT1 , GALNT18 , HOMER3 , KRT16 , and NRP2 (Fig. 5 A–B). In the OSCC training cohort, all five genes exhibited robust diagnostic performance, with AUC values of 0.797 for C1GALT1 , 0.951 for GALNT18 , 0.983 for HOMER3 , 0.917 for KRT16 , and 0.937 for NRP2 (Fig. 6 A). When evaluated in the OSF and OLK cohorts used for initial differential expression analysis, the AUCs for most genes ranged between 0.88 and 1.00. Specifically, in the OSF cohort, all five genes achieved an AUC of 1.000 (Fig. 6 B); in the OLK cohort, C1GALT1 and KRT16 yielded AUCs of 0.944 and 1.000, respectively (Fig. 6 C). We additionally retrieved new OSF, OLK, and OSCC cohorts for validation, where the diagnostic performance of the five genes showed divergence. In the new OSCC cohort, GALNT18 , HOMER3 , and NRP2 maintained high AUCs of 0.862, 0.775, and 0.963, respectively, whereas C1GALT1 and KRT16 dropped to 0.594 and 0.535 (Fig. 6 D). In the new OSF cohort, C1GALT1 and KRT16 showed AUCs of 1.000 and 0.375, respectively, while the other three genes ranged from 0.875 to 0.938 (Fig. 6 E). In the new OLK cohort, AUCs for all five genes ranged from 0.719 to 0.772 (Fig. 6 F). Consequently, GALNT18 , HOMER3 , and NRP2 , which maintained AUC values > 0.7 across all datasets, were identified as core genes for subsequent analysis. 3.6. Immune Infiltration in OSCC Given the critical role of the immune microenvironment in OSCC, immune cell infiltration analysis was further conducted on the OSCC cohort. The results revealed that the relative abundance of memory B cells, resting dendritic cells, M2 macrophages, resting mast cells, naive CD4 + T cells, and CD8 + T cells was significantly reduced in OSCC tissues. Conversely, the infiltration levels of M0 and M1 macrophages, activated mast cells, and follicular helper T cells were significantly elevated (Fig. 7 ). Overall, the OSCC immune microenvironment was characterized by a significant enrichment of myeloid cells and highly activated mast cells, accompanied by a reduction in lymphoid components such as CD8 + T cells, memory B cells, and resting dendritic cells. 3.7. Correlation with Immune Cells Correlation analysis between core genes and immune cell infiltration levels revealed that GALNT18 , HOMER3 , and NRP2 were all significantly positively correlated with M0/M1 macrophages, neutrophils, and activated mast cells, while exhibiting significant negative correlations with memory B cells and γ δ T cells (Fig. 8 ). The consistent correlation patterns among these three genes suggest they may collectively participate in the remodeling of the OSCC tumor immune microenvironment. 3.8. Pathway Resolution via Single-Gene GSEA Single-gene GSEA was employed to analyze the specific functional pathways associated with GALNT18 , HOMER3 , and NRP2 . For GALNT18 , GO enrichment results indicated that GALNT18 -related genes were primarily enriched in biological processes such as collagen fibril organization, collagen metabolic processes, epithelial–mesenchymal transition (EMT), and response to TGF-β stimulus (Fig. 9 A). KEGG enrichment suggested associations with 26S proteasome-mediated protein degradation, ITGA–β–FAK–RAC signaling, Type I interferon–JAK/STAT signaling, and fatty acid beta-oxidation (Fig. 9 B). These findings indicate that GALNT18 may be involved in extracellular matrix remodeling, EMT, and inflammatory/immune-related signaling pathways, accompanied by metabolic reprogramming. Analysis of HOMER3 revealed that its related genes were significantly enriched in GO terms including collagen fibril organization, collagen metabolism, EMT, extracellular matrix disassembly, and fibroblast proliferation (Fig. 9 C). KEGG pathways mainly involved KEAP1–NRF2 signaling, mitochondrial complex I electron transport, and protein translation initiation (Fig. 9 D), suggesting links between HOMER3 and oxidative stress response, energy metabolism, and matrix-associated invasion processes. For NRP2 , GO analysis highlighted enrichment in collagen fibril organization, EMT, external encapsulating structure organization, extracellular matrix organization, positive regulation of fibroblast proliferation, and response to TGF-β stimulus (Fig. 9 E). KEGG analysis indicated involvement in 26S proteasome-mediated protein degradation, PRC2 complex activation (H2AK119 ubiquitination), ITGA–β–FAK–RAC signaling, and voltage-dependent Ca2 + channel-mediated apoptosis (Fig. 9 F). This suggests that NRP2 may play important roles in protein degradation/epigenetic regulation, cell migration, and apoptosis regulation. 3.9. Prognostic Value in HNSC Based on the TCGA-HNSC cohort, the relationship between the expression of GALNT18 , HOMER3 , and NRP2 and Overall Survival (OS) was analyzed. Patients were stratified into groups based on the quartiles of gene expression levels. The results showed no statistically significant difference in OS between the high and low expression groups for GALNT18 and HOMER3 (P > 0.05). In contrast, patients in the high NRP2 expression group (upper quartile) exhibited significantly lower OS compared to the low expression group (lower quartile) (Log-rank P = 0.024, HR = 1.60), indicating its value as a predictor of poor prognosis (Fig. 10 A–C).. Discussion By integrating GEO transcriptomic data from OSF, OLK, and OSCC, we identified 94 common differentially expressed genes (DEGs) with consistent expression trends. GO and KEGG enrichment analyses indicated that the functions of these genes were primarily concentrated in biological processes related to extracellular matrix (ECM) remodeling, epithelial–mesenchymal transition (EMT), and inflammatory immune responses. Furthermore, through PPI network analysis and LASSO regression within the OSCC cohort, three core genes— GALNT18 , HOMER3 , and NRP2 —were ultimately determined. Multi-cohort ROC validation demonstrated that their discriminatory performance was relatively stable. Subsequent immune cell infiltration analysis, single gene GSEA, and survival analysis further clarified their modes of action. The GALNT18 gene encodes the polypeptide N-acetylgalactosaminyltransferase 18, which is one of the initiating enzymes of the mucin-type O-glycosylation pathway. Abnormal expression of this family of enzymes can affect diverse cellular processes, including immune responses, signal transduction, and cell adhesion [ 22 ] . Dysregulated glycosylation is very common in malignant tumors and is closely associated with tumor immune suppression and tumor progression. A study pointed out that in non-alcoholic steatohepatitis (NASH)-associated hepatocellular carcinoma, GALNT18 activates the EGR1/transforming growth factor-β1 (TGF-β1)/Smad pathway by catalyzing the O-glycosylation of amyloid precursor protein (APP), thereby promoting tumor occurrence and development [ 23 ] . In addition, other members of the GALNT family play similar roles in head and neck tumors. GALNT2 mediates the glycosylation of epidermal growth factor receptor (EGFR), enhancing the invasive capability of oral squamous cell carcinoma cells [ 24 ] . Thus, it can be seen that GALNT18 -mediated aberrant O-glycosylation may play a role in tumor carcinogenesis and immune regulation through pathways such as influencing TGF-β signaling and ECM remodeling. HOMER3 belongs to the Homer protein family and acts as a scaffold protein in the assembly of cell signal transduction complexes. HOMER3 possesses pro-tumorigenic characteristics in various tumors, and its overexpression is closely related to enhanced tumor cell proliferation and invasion capabilities, as well as poor prognosis [ 25 , 26 ] . For example, in triple-negative breast cancer, highly expressed HOMER3 can promote tumor invasion and metastasis. Its mechanism lies in the fact that HOMER3 , as a scaffold protein, can simultaneously bind c-Src kinase and β-catenin, enhancing growth factor signal-induced tyrosine phosphorylation and activation of β-catenin, thereby driving epidermal growth factor-mediated malignant progression [ 27 ] . In hepatocellular carcinoma, it has been reported that the upregulation of HOMER3 mediates tumor progression via the EZH2/miR-361/GPNMB axis, whereas knocking down HOMER3 can significantly inhibit the proliferation and migration of hepatocellular carcinoma cells and improve prognosis [ 25 ] . Furthermore, the expression level of HOMER3 is also linked to the tumor immune microenvironment. In colorectal cancer, high HOMER3 expression is not only associated with advanced tumor stages and elevated CEA levels but is also accompanied by increased immune cell infiltration in tumor tissues, showing a positive correlation specifically with checkpoint molecules related to T-cell exhaustion, such as PD-1, CTLA-4, and LAG3 [ 26 ] . Previous studies have also indicated a potential link between HOMER3 and macrophage regulation [ 28 ] . In our OSCC cohort, we observed a distinct positive correlation between HOMER3 and M0/M1 macrophages. This suggests that HOMER3 may be involved in the recruitment of macrophage precursors or the maintenance of a pro-inflammatory microenvironment, contributing to the complex immune landscape of OSCC. NRP2 (Neuropilin-2) is a transmembrane co-receptor protein that was initially discovered for its roles in axon guidance and angiogenesis. Due to the very short intracellular domain of NRP2 , signal transduction requires the participation of co-receptors such as VEGF receptors or Plexins. Under physiological conditions, NRP2 is expressed at high levels in lymphatic endothelial cells and various immune cells, participating in the regulation of immune functions such as cell migration, antigen presentation, apoptotic clearance (efferocytosis), and cell contact [ 29 ] . Aberrantly high expression of NRP2 in tumors has become one of the important factors promoting malignant phenotypes. NRP2 can synergize with multiple growth factors and signaling molecules to enhance the proliferation, survival, and metastatic capabilities of tumor cells. Among these, the promoting effect of NRP2 on lymphangiogenesis is particularly significant: in various solid tumors (including head and neck squamous cell carcinoma), high NRP2 expression is closely related to an increased risk of lymph node metastasis [ 30 ] . Studies focusing on head and neck tumors have shown that stromal-derived TGF-β1 can induce tumor cells to upregulate NRP2 , thereby enhancing their invasive capability and increasing the tendency for lung metastasis [ 31 ] . In oral squamous cell carcinoma, NRP2 is similarly highly expressed, corresponding to a higher incidence of lymph node metastasis in patients and significantly shortened survival times [ 32 ] . Beyond directly promoting tumor progression, NRP2 is also a key molecule affecting tumor immune escape. On one hand, NRP2 highly expressed in tumor-associated macrophages (TAMs) can improve the efficiency of macrophages in phagocytosing and clearing tumor cell apoptotic debris, removing cell fragments in an immunologically silent manner, thereby maintaining the immune tolerance state of the tumor microenvironment. In mouse models, knocking out NRP2 in macrophages leads to increased accumulation of apoptotic cells and induces secondary necrosis, accompanied by massive infiltration of effector CD8 + T cells and NK cells within the tumor [ 33 ] . This suggests that NRP2 is a key molecule linking efferocytosis in tumor tissues with the immunosuppressive state. On the other hand, NRP2 can also function as a co-inhibitory receptor on T lymphocytes. Research has found that NRP2 expression levels are very high on exhausted human and mouse effector CD4 + T cells, and it is often co-expressed with classical checkpoint molecules such as PD-1, CTLA-4, LAG3, and TIM-3. Functionally, the loss of NRP2 leads to overactive CD4 + T cell responses and enhanced inflammatory reactions. In transplant rejection models, this situation manifests as accelerated immune attack and graft loss [ 34 ] . While these findings are promising, several limitations must be acknowledged. First, although the sample sizes for the OSF and OLK datasets were relatively small, they exhibited pronounced inter-group heterogeneity. This robust biological distinction reinforces the reliability of the identified core genes despite the limited sample numbers. Second, while GALNT18 and HOMER3 showed strong diagnostic potential, they were not significant prognostic factors for overall survival; this discrepancy suggests their roles might be more critical in the early establishment of the tumor microenvironment (initiation) rather than in driving late-stage mortality. Finally, our conclusions rely on bioinformatic predictions. Although supported by literature, definitive causal relationships require further validation through in vitro or in vivo experimental models. Conclusions In this study, we systematically integrated transcriptomic data from OSF, OLK, and OSCC to map the shared molecular landscape bridging oral mucosal potentially malignant disorders and invasive carcinoma. By utilizing OSCC as a malignant reference, we filtered out non-essential reactive changes and isolated a core gene signature capable of driving the entire disease continuum. Our analysis identified GALNT18 , HOMER3 , and NRP2 as key molecular markers in this process. These genes do not merely mark the presence of disease; they likely function as critical drivers that orchestrate extracellular matrix remodeling and shape a pro-tumorigenic immune microenvironment—specifically through macrophage regulation and the facilitation of immune evasion. We propose that these shared molecular signatures offer a dual clinical value: they serve as sensitive biomarkers for risk stratification in patients with OSF or OLK, and, more importantly, represent potential therapeutic targets. Targeting these early drivers provides a compelling strategy for chemoprevention, aiming to intercept the malignant transformation process before it progresses to irreversible invasive cancer. Declarations Funding This research received no external funding. Clinical trial number Not applicable. Declaration of Artificial Intelligence Use The authors declare that artificial intelligence (AI) was used solely for English language refinement and grammar improvement during manuscript preparation. No generative AI was used for data analysis, image creation, or content generation. The authors take full responsibility for the integrity and accuracy of the final manuscript. authorship contribution statement Conceptualization, J.Z. and C.X.; methodology, J.Z.; software, J.Z.; validation, T.T., W.L., Y.P., Z.Q. and Y.C.; formal analysis, J.Z.; investigation, J.Z. and Y.C.; resources, C.X.; data curation, J.Z.; writing—original draft preparation, J.Z.; writing—review and editing, C.X.; visualization, J.Z.; supervision, C.X.; project administration, C.X. All authors have read and agreed to the published version of the manuscript. Ethics, Consent to Participate, and Consent to Publish Declarations Not applicable. Data Availability The datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/). The specific accession numbers utilized for the analyses are: GSE64216, GSE246050, GSE30784, GSE25099, GSE12586, and GSE227919. Declaration of competing interest All authors declare that No conflict of interest exists. References Zhi Y, Wang Q, Zi M, et al. Spatial transcriptomic and metabolomic landscapes of oral submucous fibrosis-derived oral squamous cell carcinoma and its tumor microenvironment. Adv Sci (Weinh). 2024;11(12):e2306515. 10.1002/advs.202306515 . He Y, Shao F, Pi W, et al. Largescale transcriptomics analysis suggests over-expression of BGH3, MMP9 and PDIA3 in oral squamous cell carcinoma. PLoS ONE. 2016;11(1):e0146530. 10.1371/journal.pone.0146530 . Cui Y, Ding C, Peng C, et al. Progress in clinicopathological diagnosis of oral potentially malignant disorders. Hua Xi Kou Qiang Yi Xue Za Zhi. 2025;43(3):314–24. 10.7518/hxkq.2025.2024418 . Naik V, Kumar M, Solomon MC, et al. Clinical and histological indicators for malignant transformation of oral submucous fibrosis: an analysis of cases from a tertiary care center. BMC Oral Health. 2025;25(1):1053. 10.1186/s12903-025-06443-y . Zhu X, Zhang J, Li J, et al. Diagnosing malignant transformation potential in oral submucous fibrosis with leukoplakia: an approach utilizing deoxyribonucleic acid image cytometry. Sci Rep. 2025;15(1):22085. 10.1038/s41598-025-06585-x . Tilakaratne WM, Jayasooriya PR, Jayasuriya NS et al. 2019. Oral epithelial dysplasia: Causes, quantification, prognosis, and management challenges. Periodontol 2000 80(1):126–147 10.1111/prd.12259 Liu W, Wang YF, Zhou HW, et al. Malignant transformation of oral leukoplakia: a retrospective cohort study of 218 Chinese patients. BMC Cancer. 2010;10:685. 10.1186/1471-2407-10-685 . Alsarraf A, Kujan O, Farah CS. A systematic review of oral biopsies, sample types, and detection techniques applied in relation to oral cancer detection. BioTech (Basel). 2021;11(1):5. 10.3390/biotech11010005 . Poh CF, Ng S, Berean KW, et al. Biopsy and histopathologic diagnosis of oral premalignant and malignant lesions. J Can Dent Assoc. 2008;74(3):283–8. Natarajan PM, Umapathy VR. Role of transcriptomics in the study of oral cancer. Front Oral Health. 2025;6:1524364. 10.3389/froh.2025.1524364 . Singh P, Rai A, Dohare R, et al. Network-based identification of signature genes KLF6 and SPOCK1 associated with oral submucous fibrosis. Mol Clin Oncol. 2020;12(4):299–310. 10.3892/mco.2020.1997 . Mahmoudi Anzabi R, Davodpour AH, Ghodratizadeh S, et al. Recent knowledge on squamous cell carcinoma of the oral cavity: Contributing factors, underlying molecular pathways, and current attitudes in the therapeutic approaches. Int J Mol Cell Med. 2025;14(3):928–48. 10.22088/IJMCM.BUMS.14.3.928 . Li Y, Tang Z, Xiao C, et al. SOX11 and GLIS2: Novel biomarkers for understanding the progression of oral leukoplakia. Int Dent J DOI. 2025. 10.1016/j.identj.2025.100872 . Zhang Y, Zhang J, Zhao S, et al. Single-cell RNA sequencing highlights the immunosuppression of IDO1 + macrophages in the malignant transformation of oral leukoplakia. Theranostics. 2024;14(12):4787–805. 10.7150/thno.96273 . Barrett T, Wilhite SE, Ledoux P, et al. NCBI GEO: archive for gene expression and epigenomics data sets: 23-year update. Nucleic Acids Res. 2024;52(D1):D138–44. 10.1093/nar/gkad1013 . Häyrinen MJ, Kiiskilä J, Ranki A, et al. The transcription factor Twist1 has a significant role in Mycosis Fungoides (MF) cell biology: an RNA sequencing study of 40 MF cases. Cancers (Basel). 2023;15(5):1527. 10.3390/cancers15051527 . Sherman BT, Hao M, Qiu J, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50(W1):W216–21. 10.1093/nar/gkac194 . Szklarczyk D, Gable AL, Nastou KC, et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 2021;49(D1):D605–12. 10.1093/nar/gkaa1074 . Tang Q, Chen Y, Li X, et al. Prognostic model revealing pyroptosis-related signatures in oral squamous cell carcinoma based on bioinformatics analysis. Bioengineered. 2021;12(1):8565–76. 10.1080/21655979.2021.1988362 . Ma Y, Ming Y, Hou Z, et al. Deciphering the overlapping immune mechanism between depression and breast cancer. Int J Mol Sci. 2025;26(11):5229. 10.3390/ijms26115229 . Newman AM, Liu CL, Green MR, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453–7. 10.1038/nmeth.3337 . Indellicato R, Pisano A. Concerted regulation of glycosylation factors sustains tissue identity and function. Biomedicines. 2022;10(8):1805. 10.3390/biomedicines10081805 . Chen D, Nie Y, Shi L, et al. GALNT18-mediated O-glycosylation promotes development of nonalcoholic steatohepatitis-associated hepatocellular carcinoma through activation of EGR1/TGF-β1/Smad signaling. Mol Biotechnol DOI. 2025. 10.1007/s12033-025-01494-6 . Pucci M, Malagolini N, Dall'Olio F. Glycosyltransferases in cancer: prognostic biomarkers of survival in patient cohorts and impact on malignancy in experimental models. Cancers (Basel). 2022;14(9):2129. 10.3390/cancers14092129 . Chen L, Shan X, Wan X, et al. HOMER3 promotes liver hepatocellular carcinoma cancer progression by upregulating EZH2 and mediating miR-361/GPNMB axis. Pathol Res Pract. 2024;257:155280. 10.1016/j.prp.2024.155280 . Luo M, Zhao C, Zhao Y, et al. Identification of homer protein homolog 3 as a prognostic marker of colon adenocarcinoma. Heliyon. 2024;10(13):e33344. 10.1016/j.heliyon.2024.e33344 . Liu Q, He L, Lin C, et al. HOMER3 facilitates growth factor-mediated β-Catenin tyrosine phosphorylation and activation to promote metastasis in triple negative breast cancer. J Hematol Oncol. 2021;14(1):6. 10.1186/s13045-020-01020-8 . Luan R, Yang C, Niu X. HOMER3 and its correlation with immune-infiltrating M2 macrophages in colorectal cancer. Discov Oncol DOI. 2025. 10.1007/s12672-025-04247-8 . Schellenburg S, Schulz A, Poitz DM, et al. Role of neuropilin-2 in the immune system. Mol Immunol. 2017;90:239–44. 10.1016/j.molimm.2017.08.010 . Zhang M, Zhang W, Wu A, et al. Prognostic significance of VEGF-C, semaphorin 3F, and neuropilin-2 expression in oral squamous cell carcinomas and their relationship with lymphangiogenesis. J Surg Oncol. 2015;111(4):382–8. 10.1002/jso.23843 . Recalde-Percaz L, de la Guia-Lopez I, Linzoain-Agos P, et al. Neuropilin-2 upregulation by stromal TGFβ1 induces lung disseminated tumor cells dormancy escape and promotes metastasis outgrowth. Neoplasia. 2025. 10.1016/j.neo.2025.101220 . Kang Y, Zhang Y, Zhang Y, et al. NRP2, a potential biomarker for oral squamous cell carcinoma. Am J Transl Res. 2021;13(8):8938–51. Roy S, Bag AK, Singh RK, et al. Macrophage-derived neuropilin-2 exhibits novel tumor-promoting functions. Cancer Res. 2018;78(19):5600–17. 10.1158/0008-5472.CAN-18-0568 . Wedel J, Kochupurakkal N, Kong SW, et al. Neuropilin-2 functions as a coinhibitory receptor to regulate antigen-induced inflammation and allograft rejection. J Clin Invest. 2025;135(13):e172218. 10.1172/JCI172218 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 18 May, 2026 Reviews received at journal 12 Apr, 2026 Reviews received at journal 08 Apr, 2026 Reviewers agreed at journal 02 Apr, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers agreed at journal 31 Mar, 2026 Reviewers invited by journal 31 Mar, 2026 Editor invited by journal 30 Mar, 2026 Editor assigned by journal 27 Mar, 2026 Submission checks completed at journal 26 Mar, 2026 First submitted to journal 26 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9127273","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":615423181,"identity":"c6298a9a-438f-46ef-af00-58eec2cc9ab2","order_by":0,"name":"Jiabin Zhao","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Jiabin","middleName":"","lastName":"Zhao","suffix":""},{"id":615423182,"identity":"60db0d96-52f6-4fa3-be86-4d7941acb27a","order_by":1,"name":"Yuanyi Chen","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Yuanyi","middleName":"","lastName":"Chen","suffix":""},{"id":615423183,"identity":"e49fb5e6-aaa8-4fbd-9cfe-d43d7f6a17bd","order_by":2,"name":"Ting Tang","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Ting","middleName":"","lastName":"Tang","suffix":""},{"id":615423184,"identity":"cfbd3130-d4eb-49ab-b827-0d57f82e22f1","order_by":3,"name":"Weike Lu","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Weike","middleName":"","lastName":"Lu","suffix":""},{"id":615423185,"identity":"b7971b49-7f18-4ee5-b6ad-c8b6d7e16ec2","order_by":4,"name":"Yue Pan","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Yue","middleName":"","lastName":"Pan","suffix":""},{"id":615423186,"identity":"e4ea29ce-fc77-4cbe-ad7e-04db942e1888","order_by":5,"name":"Zihua Qi","email":"","orcid":"","institution":"Central South University","correspondingAuthor":false,"prefix":"","firstName":"Zihua","middleName":"","lastName":"Qi","suffix":""},{"id":615423187,"identity":"b90012eb-f675-4eba-9277-939751a5065f","order_by":6,"name":"Chunjiao Xu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyElEQVRIiWNgGAWjYDACCRA2ALGYDxz4UEGaFrbEgzPOEKsFAniMD/O2EKGDf3aP2QOLAjs5gxs5Hw7wNjDI84sdIGDJnTPmBhIGycaSM3I3HJDcwWA4c3YCfi0GEjlmEhIGzIn9EkAthmcYEgxuE6elPrFNIufBgcQ24rUcBtqSw3DgIDFaJG6klQG1HDeW7HlmcLDhjARhv/DPSN4mLfGnWs7gePLjz38qbOT5pQloAQFmRNwgRRNewPiBOHWjYBSMglEwUgEAvMpA5IAGWtMAAAAASUVORK5CYII=","orcid":"","institution":"Central South University","correspondingAuthor":true,"prefix":"","firstName":"Chunjiao","middleName":"","lastName":"Xu","suffix":""}],"badges":[],"createdAt":"2026-03-15 08:38:43","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9127273/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9127273/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106094826,"identity":"2f5f9fd5-df78-45c5-a637-8616291a74e5","added_by":"auto","created_at":"2026-04-03 11:43:22","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":482795,"visible":true,"origin":"","legend":"\u003cp\u003eAnalysis of differentially expressed genes (DEGs) in OSF, OLK, and OSCC. (A) Volcano plot of the OSF dataset (GSE64216). (B) Hierarchical clustering heatmap of the OSF dataset (GSE64216). (C) Volcano plot of the OLK dataset (GSE246050). (D) Hierarchical clustering heatmap of the OLK dataset (GSE246050). (E) Volcano plot of the OSCC dataset (GSE30784). (F) Hierarchical clustering heatmap of the OSCC dataset (GSE30784).\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/3468aa0ae9a67ead972c6538.png"},{"id":106009984,"identity":"000cbdc9-7f89-4716-9bca-45d597101065","added_by":"auto","created_at":"2026-04-02 11:30:40","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":97245,"visible":true,"origin":"","legend":"\u003cp\u003eIntersection analysis of differentially expressed genes (DEGs) across OSF, OLK, and OSCC. (A) Venn diagram illustrating the intersection of commonly upregulated genes(n=54). (B) Venn diagram illustrating the intersection of commonly downregulated genes(n=40).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/eb29d755f003b3e8d1d23833.png"},{"id":106095070,"identity":"bb33d4ee-fc45-4782-bc67-d56b4d7baddc","added_by":"auto","created_at":"2026-04-03 11:44:09","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":273375,"visible":true,"origin":"","legend":"\u003cp\u003eFunctional enrichment analysis of the common differentially expressed genes (DEGs). (A) Significantly enriched Gene Ontology (GO) Biological Process (BP) terms. (B) Significantly enriched GO Cellular Component (CC) and Molecular Function (MF) terms. (C) Significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/377058247cf347bd4a708e13.png"},{"id":106093819,"identity":"71b6e46b-ad15-4bf7-ae4b-0f056729d341","added_by":"auto","created_at":"2026-04-03 11:39:21","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":124930,"visible":true,"origin":"","legend":"\u003cp\u003eScreening of core genes. (A) The protein–protein interaction (PPI) network constructed using Cytoscape. Nodes highlighted in yellow represent the candidate core genes. (B) The PPI network of the 15 identified core genes.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/2d35d237eab9bb7e56d36d51.png"},{"id":106009988,"identity":"0659383e-0b78-4028-8e38-c95bd027e83e","added_by":"auto","created_at":"2026-04-02 11:30:40","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":92786,"visible":true,"origin":"","legend":"\u003cp\u003eLASSO regression analysis in the OSCC cohort. (A) LASSO coefficient profiles of the candidate genes. (B) Selection of the optimal tuning parameter (λ) using 10-fold cross-validation.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/6e066b69099f03f923f4563d.png"},{"id":106093971,"identity":"25ab447c-ef59-4efb-a5ee-32681f2b3ce2","added_by":"auto","created_at":"2026-04-03 11:40:25","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":177014,"visible":true,"origin":"","legend":"\u003cp\u003eResults of Receiver Operating Characteristic (ROC) analysis. (A) ROC curves in the OSCC training dataset (GSE30784). (B) ROC curves in the initial OSF dataset (GSE64216). (C) ROC curves in the initial OLK dataset (GSE246050). (D) ROC curves in the independent OSCC validation da-taset (GSE25099). (E) ROC curves in the independent OSF validation dataset (GSE12586). (F) ROC curves in the independent OLK validation dataset (GSE227919).\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/9c732d97388014e04c3609bf.png"},{"id":106094435,"identity":"cacfe178-ebc4-4562-9aa0-66aaae98f2d4","added_by":"auto","created_at":"2026-04-03 11:42:33","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":129495,"visible":true,"origin":"","legend":"\u003cp\u003eImmune infiltration analysis of the OSCC dataset (GSE30784). Asterisks indicate statistical significance: (* P \u0026lt; 0.05, ** P \u0026lt; 0.01, *** P \u0026lt; 0.001, and **** P \u0026lt; 0.0001).\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/b8dd2b533ce307c3c7a11d9c.png"},{"id":106009990,"identity":"6a44dde0-93f5-4631-b2a9-17e641480677","added_by":"auto","created_at":"2026-04-02 11:30:40","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":107137,"visible":true,"origin":"","legend":"\u003cp\u003eCorrelation analysis between the expression of core genes and immune cell infiltration levels. Red represents positive correlation, and blue represents negative correlation. (* P \u0026lt; 0.05, ** P \u0026lt; 0.01, *** P \u0026lt; 0.001).\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/20e329fc8d92315d0384aaf1.png"},{"id":106095077,"identity":"51ada2a2-70d6-420b-acd8-140ce370a6cd","added_by":"auto","created_at":"2026-04-03 11:44:11","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":426290,"visible":true,"origin":"","legend":"\u003cp\u003eSingle-gene Gene Set Enrichment Analysis (GSEA) of the three core genes. (A, B) Enriched GO terms and KEGG pathways associated with GALNT18. (C, D) Enriched GO terms and KEGG pathways associated with HOMER3. (E, F) Enriched GO terms and KEGG pathways associated with NRP2.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/be37938aae36f88e8629b8a5.png"},{"id":106093831,"identity":"d80287e4-34ee-4fa2-98c8-6333077ec108","added_by":"auto","created_at":"2026-04-03 11:39:26","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":189255,"visible":true,"origin":"","legend":"\u003cp\u003eSurvival analysis of the core genes based on the TCGA-HNSC cohort. (A) Kaplan-Meier survival curve for GALNT18. (B) Kaplan-Meier survival curve for HOMER3. (C) Kaplan-Meier survival curve for NRP2.\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/90a10c6ac8f55426fcc445f6.png"},{"id":106095801,"identity":"871b926f-63ff-4aa2-8bb6-867858327f24","added_by":"auto","created_at":"2026-04-03 11:51:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2818124,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9127273/v1/5aabbe08-7d18-49ee-8d94-f4ca58a0389b.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Identification of GALNT18, HOMER3, and NRP2 as shared molecular signatures associated with stromal remodeling and immune suppression in oral mucosal malignancy-associated disorders","fulltext":[{"header":"Introduction","content":"\u003cp\u003eOral squamous cell carcinoma (OSCC) is a common head and neck malignancy, with nearly 400,000 new cases diagnosed annually \u003csup\u003e[1]\u003c/sup\u003e. Despite therapeutic advances, the overall five-year survival rate remains approximately 50% \u003csup\u003e[2]\u003c/sup\u003e. OSCC can arise from clinically normal-appearing mucosa or coexist with/succeed oral potentially malignant disorders (OPMDs) \u003csup\u003e[3]\u003c/sup\u003e. Oral submucous fibrosis (OSF) and oral leukoplakia (OLK) are the two predominant types of OPMDs. OSF, strongly associated with long-term betel nut chewing, is characterized by subepithelial fibrosis and carries a malignant transformation rate of approximately 4%\u0026ndash;10% \u003csup\u003e[4]\u003c/sup\u003e. OLK presents as white patches on the oral mucosa, with a prevalence of about 4% and a malignant transformation rate ranging from 1% to 10% \u003csup\u003e[3]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eWhile the majority of OSF and OLK cases do not eventually progress to OSCC, they frequently exhibit varying degrees of cellular atypia and alterations in the immune microenvironment, reflecting molecular features similar to those of OSCC \u003csup\u003e[5-7]\u003c/sup\u003e. The molecules mediating these early changes represent potential targets for risk stratification and prognostic assessment; however, they are often difficult to identify effectively through routine histopathological examination\u003csup\u003e\u0026nbsp;[8,9]\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe development of high-throughput sequencing technologies has facilitated in-depth investigations into disease mechanisms. Nevertheless, most previous studies have focused on single disease entities, lacking a systematic analysis across multiple conditions \u003csup\u003e[10,11]\u003c/sup\u003e. By integrating transcriptomic data from multiple diseases, this study screened for core genes commonly differentially expressed in OSF, OLK, and OSCC. Our objective is to reveal shared molecular signatures and enhance the understanding of the pathogenesis of oral malignancy-associated disorders\u003csup\u003e\u0026nbsp;[12]\u003c/sup\u003e, thereby providing a theoretical foundation for the development of early assessment strategies and therapeutic interventions \u003csup\u003e[13,14]\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Materials \u0026 Methods","content":"\u003cp\u003e\u003cstrong\u003e2.1. Data Acquisition and Processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTranscriptomic profiles were retrieved from the Gene Expression Omnibus (GEO) database \u003csup\u003e[15]\u003c/sup\u003e. The study included: (1) an OSF dataset (GSE64216) comprising 4 OSF cases and 2 normal mucosal controls; (2) an OLK dataset (GSE246050) containing 6 OLK cases and 3 controls; and (3) an OSCC dataset (GSE30784) with 167 OSCC cases and 45 controls. These datasets served as the foundation for identifying DEGs. Additionally, three independent datasets (GSE25099 for OSCC, GSE12586 for OSF, and GSE227919 for OLK) were utilized for external validation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2. Differential Expression Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDEGs were identified using the GEO2R tool. To ensure statistical rigor, we applied a significance threshold of P \u0026lt; 0.05 and a fold-change cutoff of |log2FC| \u0026gt; 0.585 \u003csup\u003e[16]\u003c/sup\u003e. Genes were subsequently stratified into upregulated and downregulated categories. To pinpoint shared molecular drivers, we utilized the bioinformatics platform (https://www.bioinformatics.com.cn/) to generate Venn diagrams, isolating genes with consistent expression trends across all three pathological states.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3. Functional Enrichment Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBiological insights into the co-expressed genes were gained through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses using the DAVID platform\u003csup\u003e\u0026nbsp;[17]\u0026nbsp;\u003c/sup\u003e(https://davidbioinformatics.nih.gov/). A P-value of \u0026lt; 0.05 was considered statistically significant.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4. Protein\u0026ndash;Protein Interaction (PPI) Network Construction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA PPI network was constructed using the STRING database \u003csup\u003e[18]\u003c/sup\u003e (https://string-db.org/) with a minimum interaction confidence score of 0.4. Isolated nodes were excluded to focus on functional connectivity. The network was visualized in Cytoscape, and the CytoNCA plugin \u003csup\u003e[19]\u003c/sup\u003e was employed to calculate topological properties. The top 15 genes ranked by Degree centrality were identified as candidate hub genes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5. Feature Selection and Diagnostic Validation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo identify the most robust biomarkers, Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression was performed using the \u0026ldquo;glmnet\u0026rdquo; R package\u003csup\u003e\u0026nbsp;[20]\u003c/sup\u003e. A 10-fold cross-validation was employed to determine the optimal penalty parameter. The diagnostic performance of the selected core genes was evaluated using Receiver Operating Characteristic (ROC) curves, with the Area Under the Curve (AUC) serving as the primary metric of discrimination.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6. Immune Landscape Profiling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe immune microenvironment was characterized using the CIBERSORT algorithm via the bioinformatics platform \u003csup\u003e[21]\u003c/sup\u003e. The LM22 signature matrix was used to deconvolute gene expression data and estimate the relative abundance of 22 distinct infiltrating immune cell types. Spearman\u0026rsquo;s correlation analysis was then conducted in R to assess the relationship between core gene expression and specific immune cell subsets.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.7. Single-Gene Gene Set Enrichment Analysis (GSEA)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo elucidate the specific pathways regulated by each core gene, we performed single-gene GSEA. This involved calculating the Spearman correlation between the target gene and all other genes in the genome, ranking them, and testing for pathway enrichment using R software.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.8. Survival Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe clinical prognostic value of the core genes was assessed using the GEPIA tool (http://gepia.cancer-pku.cn/), utilizing data from the TCGA-HNSC cohort. Patients were stratified into high- and low-expression groups based on the upper (75%) and lower (25%) quartile thresholds. Overall Survival (OS) was compared between groups using the Log-rank test.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Identification of DEGs in OSF, OLK, and OSCC\u003c/h2\u003e \u003cp\u003eOur analysis yielded 3506 differentially expressed genes (DEGs) in OSF (1622 up-regulated, 1884 downregulated), 3020 in OLK (1575 upregulated, 1445 downregulated), and 4764 in OSCC (2558 upregulated, 2206 downregulated). Volcano plots and hierarchical clustering (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA\u0026ndash;F) confirmed that these distinct expression profiles could effectively differentiate pathological tissues from normal mucosa.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Identification of Co-Expressed Genes\u003c/h2\u003e \u003cp\u003eIntersection analysis revealed 94 co-expressed genes (54 upregulated and 40 downregulated) shared across OSF, OLK, and OSCC (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA-B). These overlapping genes represent potential common denominators in the pathogenesis of these conditions.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.3. Functional Enrichment of Co-Expressed Genes\u003c/h2\u003e \u003cp\u003eGO analysis indicated that these genes are primarily involved in extracellular matrix organization, specifically skin barrier establishment, inflammatory responses, and bone remodeling (BP). In terms of cellular components (CC), they are enriched in cell junctions and filopodia. Molecular functions (MF) highlighted activities such as protease inhibition and growth factor binding (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA-B). KEGG pathway analysis further linked these genes to metabolic reprogramming (central carbon metabolism in cancer, arachidonic acid metabolism) and structural integrity (cornified envelope formation) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.4. PPI Network and Hub Gene Screening\u003c/h2\u003e \u003cp\u003eThe PPI network analysis, refined by CytoNCA, identified the top 15 genes based on topological importance (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA\u0026ndash;B). These hub genes are positioned at critical junctions of the interaction network, suggesting pivotal regulatory roles.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.5. Selection of Core Genes via LASSO Regression\u003c/h2\u003e \u003cp\u003eTo filter for genes most relevant to malignant transformation, we applied LASSO regression to the 15 hub genes within the OSCC cohort. This process narrowed the list to five core genes: \u003cem\u003eC1GALT1\u003c/em\u003e, \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, \u003cem\u003eKRT16\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA\u0026ndash;B).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn the OSCC training cohort, all five genes exhibited robust diagnostic performance, with AUC values of 0.797 for \u003cem\u003eC1GALT1\u003c/em\u003e, 0.951 for \u003cem\u003eGALNT18\u003c/em\u003e, 0.983 for \u003cem\u003eHOMER3\u003c/em\u003e, 0.917 for \u003cem\u003eKRT16\u003c/em\u003e, and 0.937 for \u003cem\u003eNRP2\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). When evaluated in the OSF and OLK cohorts used for initial differential expression analysis, the AUCs for most genes ranged between 0.88 and 1.00. Specifically, in the OSF cohort, all five genes achieved an AUC of 1.000 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB); in the OLK cohort, \u003cem\u003eC1GALT1\u003c/em\u003e and \u003cem\u003eKRT16\u003c/em\u003e yielded AUCs of 0.944 and 1.000, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe additionally retrieved new OSF, OLK, and OSCC cohorts for validation, where the diagnostic performance of the five genes showed divergence. In the new OSCC cohort, \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e maintained high AUCs of 0.862, 0.775, and 0.963, respectively, whereas \u003cem\u003eC1GALT1\u003c/em\u003e and \u003cem\u003eKRT16\u003c/em\u003e dropped to 0.594 and 0.535 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). In the new OSF cohort, \u003cem\u003eC1GALT1\u003c/em\u003e and \u003cem\u003eKRT16\u003c/em\u003e showed AUCs of 1.000 and 0.375, respectively, while the other three genes ranged from 0.875 to 0.938 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE). In the new OLK cohort, AUCs for all five genes ranged from 0.719 to 0.772 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eF). Consequently, \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e, which maintained AUC values\u0026thinsp;\u0026gt;\u0026thinsp;0.7 across all datasets, were identified as core genes for subsequent analysis.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.6. Immune Infiltration in OSCC\u003c/h2\u003e \u003cp\u003eGiven the critical role of the immune microenvironment in OSCC, immune cell infiltration analysis was further conducted on the OSCC cohort. The results revealed that the relative abundance of memory B cells, resting dendritic cells, M2 macrophages, resting mast cells, naive CD4\u003csup\u003e+\u003c/sup\u003e T cells, and CD8\u003csup\u003e+\u003c/sup\u003e T cells was significantly reduced in OSCC tissues. Conversely, the infiltration levels of M0 and M1 macrophages, activated mast cells, and follicular helper T cells were significantly elevated (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). Overall, the OSCC immune microenvironment was characterized by a significant enrichment of myeloid cells and highly activated mast cells, accompanied by a reduction in lymphoid components such as CD8\u003csup\u003e+\u003c/sup\u003e T cells, memory B cells, and resting dendritic cells.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.7. Correlation with Immune Cells\u003c/h2\u003e \u003cp\u003eCorrelation analysis between core genes and immune cell infiltration levels revealed that \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e were all significantly positively correlated with M0/M1 macrophages, neutrophils, and activated mast cells, while exhibiting significant negative correlations with memory B cells and γ δ T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). The consistent correlation patterns among these three genes suggest they may collectively participate in the remodeling of the OSCC tumor immune microenvironment.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.8. Pathway Resolution via Single-Gene GSEA\u003c/h2\u003e \u003cp\u003eSingle-gene GSEA was employed to analyze the specific functional pathways associated with \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e. For \u003cem\u003eGALNT18\u003c/em\u003e, GO enrichment results indicated that \u003cem\u003eGALNT18\u003c/em\u003e-related genes were primarily enriched in biological processes such as collagen fibril organization, collagen metabolic processes, epithelial\u0026ndash;mesenchymal transition (EMT), and response to TGF-β stimulus (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eA). KEGG enrichment suggested associations with 26S proteasome-mediated protein degradation, ITGA\u0026ndash;β\u0026ndash;FAK\u0026ndash;RAC signaling, Type I interferon\u0026ndash;JAK/STAT signaling, and fatty acid beta-oxidation (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eB). These findings indicate that \u003cem\u003eGALNT18\u003c/em\u003e may be involved in extracellular matrix remodeling, EMT, and inflammatory/immune-related signaling pathways, accompanied by metabolic reprogramming. Analysis of \u003cem\u003eHOMER3\u003c/em\u003e revealed that its related genes were significantly enriched in GO terms including collagen fibril organization, collagen metabolism, EMT, extracellular matrix disassembly, and fibroblast proliferation (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eC). KEGG pathways mainly involved KEAP1\u0026ndash;NRF2 signaling, mitochondrial complex I electron transport, and protein translation initiation (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eD), suggesting links between \u003cem\u003eHOMER3\u003c/em\u003e and oxidative stress response, energy metabolism, and matrix-associated invasion processes. For \u003cem\u003eNRP2\u003c/em\u003e, GO analysis highlighted enrichment in collagen fibril organization, EMT, external encapsulating structure organization, extracellular matrix organization, positive regulation of fibroblast proliferation, and response to TGF-β stimulus (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eE). KEGG analysis indicated involvement in 26S proteasome-mediated protein degradation, PRC2 complex activation (H2AK119 ubiquitination), ITGA\u0026ndash;β\u0026ndash;FAK\u0026ndash;RAC signaling, and voltage-dependent Ca2\u0026thinsp;+\u0026thinsp;channel-mediated apoptosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eF). This suggests that \u003cem\u003eNRP2\u003c/em\u003e may play important roles in protein degradation/epigenetic regulation, cell migration, and apoptosis regulation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.9. Prognostic Value in HNSC\u003c/h2\u003e \u003cp\u003eBased on the TCGA-HNSC cohort, the relationship between the expression of \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e and Overall Survival (OS) was analyzed. Patients were stratified into groups based on the quartiles of gene expression levels. The results showed no statistically significant difference in OS between the high and low expression groups for \u003cem\u003eGALNT18\u003c/em\u003e and \u003cem\u003eHOMER3\u003c/em\u003e (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). In contrast, patients in the high \u003cem\u003eNRP2\u003c/em\u003e expression group (upper quartile) exhibited significantly lower OS compared to the low expression group (lower quartile) (Log-rank P\u0026thinsp;=\u0026thinsp;0.024, HR\u0026thinsp;=\u0026thinsp;1.60), indicating its value as a predictor of poor prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eA\u0026ndash;C)..\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eBy integrating GEO transcriptomic data from OSF, OLK, and OSCC, we identified 94 common differentially expressed genes (DEGs) with consistent expression trends. GO and KEGG enrichment analyses indicated that the functions of these genes were primarily concentrated in biological processes related to extracellular matrix (ECM) remodeling, epithelial\u0026ndash;mesenchymal transition (EMT), and inflammatory immune responses. Furthermore, through PPI network analysis and LASSO regression within the OSCC cohort, three core genes\u0026mdash;\u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e\u0026mdash;were ultimately determined. Multi-cohort ROC validation demonstrated that their discriminatory performance was relatively stable. Subsequent immune cell infiltration analysis, single gene GSEA, and survival analysis further clarified their modes of action.\u003c/p\u003e \u003cp\u003eThe \u003cem\u003eGALNT18\u003c/em\u003e gene encodes the polypeptide N-acetylgalactosaminyltransferase 18, which is one of the initiating enzymes of the mucin-type O-glycosylation pathway. Abnormal expression of this family of enzymes can affect diverse cellular processes, including immune responses, signal transduction, and cell adhesion \u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e. Dysregulated glycosylation is very common in malignant tumors and is closely associated with tumor immune suppression and tumor progression. A study pointed out that in non-alcoholic steatohepatitis (NASH)-associated hepatocellular carcinoma, \u003cem\u003eGALNT18\u003c/em\u003e activates the EGR1/transforming growth factor-β1 (TGF-β1)/Smad pathway by catalyzing the O-glycosylation of amyloid precursor protein (APP), thereby promoting tumor occurrence and development \u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/sup\u003e. In addition, other members of the \u003cem\u003eGALNT\u003c/em\u003e family play similar roles in head and neck tumors. \u003cem\u003eGALNT2\u003c/em\u003e mediates the glycosylation of epidermal growth factor receptor (EGFR), enhancing the invasive capability of oral squamous cell carcinoma cells \u003csup\u003e[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/sup\u003e. Thus, it can be seen that \u003cem\u003eGALNT18\u003c/em\u003e-mediated aberrant O-glycosylation may play a role in tumor carcinogenesis and immune regulation through pathways such as influencing TGF-β signaling and ECM remodeling.\u003c/p\u003e \u003cp\u003e \u003cem\u003eHOMER3\u003c/em\u003e belongs to the Homer protein family and acts as a scaffold protein in the assembly of cell signal transduction complexes. \u003cem\u003eHOMER3\u003c/em\u003e possesses pro-tumorigenic characteristics in various tumors, and its overexpression is closely related to enhanced tumor cell proliferation and invasion capabilities, as well as poor prognosis \u003csup\u003e[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e. For example, in triple-negative breast cancer, highly expressed \u003cem\u003eHOMER3\u003c/em\u003e can promote tumor invasion and metastasis. Its mechanism lies in the fact that \u003cem\u003eHOMER3\u003c/em\u003e, as a scaffold protein, can simultaneously bind c-Src kinase and β-catenin, enhancing growth factor signal-induced tyrosine phosphorylation and activation of β-catenin, thereby driving epidermal growth factor-mediated malignant progression \u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e. In hepatocellular carcinoma, it has been reported that the upregulation of \u003cem\u003eHOMER3\u003c/em\u003e mediates tumor progression via the EZH2/miR-361/GPNMB axis, whereas knocking down \u003cem\u003eHOMER3\u003c/em\u003e can significantly inhibit the proliferation and migration of hepatocellular carcinoma cells and improve prognosis \u003csup\u003e[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. Furthermore, the expression level of \u003cem\u003eHOMER3\u003c/em\u003e is also linked to the tumor immune microenvironment. In colorectal cancer, high \u003cem\u003eHOMER3\u003c/em\u003e expression is not only associated with advanced tumor stages and elevated CEA levels but is also accompanied by increased immune cell infiltration in tumor tissues, showing a positive correlation specifically with checkpoint molecules related to T-cell exhaustion, such as PD-1, CTLA-4, and LAG3 \u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e. Previous studies have also indicated a potential link between \u003cem\u003eHOMER3\u003c/em\u003e and macrophage regulation \u003csup\u003e[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e. In our OSCC cohort, we observed a distinct positive correlation between \u003cem\u003eHOMER3\u003c/em\u003e and M0/M1 macrophages. This suggests that \u003cem\u003eHOMER3\u003c/em\u003e may be involved in the recruitment of macrophage precursors or the maintenance of a pro-inflammatory microenvironment, contributing to the complex immune landscape of OSCC.\u003c/p\u003e \u003cp\u003e \u003cem\u003eNRP2\u003c/em\u003e (Neuropilin-2) is a transmembrane co-receptor protein that was initially discovered for its roles in axon guidance and angiogenesis. Due to the very short intracellular domain of \u003cem\u003eNRP2\u003c/em\u003e, signal transduction requires the participation of co-receptors such as VEGF receptors or Plexins. Under physiological conditions, \u003cem\u003eNRP2\u003c/em\u003e is expressed at high levels in lymphatic endothelial cells and various immune cells, participating in the regulation of immune functions such as cell migration, antigen presentation, apoptotic clearance (efferocytosis), and cell contact \u003csup\u003e[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e. Aberrantly high expression of \u003cem\u003eNRP2\u003c/em\u003e in tumors has become one of the important factors promoting malignant phenotypes. \u003cem\u003eNRP2\u003c/em\u003e can synergize with multiple growth factors and signaling molecules to enhance the proliferation, survival, and metastatic capabilities of tumor cells. Among these, the promoting effect of \u003cem\u003eNRP2\u003c/em\u003e on lymphangiogenesis is particularly significant: in various solid tumors (including head and neck squamous cell carcinoma), high \u003cem\u003eNRP2\u003c/em\u003e expression is closely related to an increased risk of lymph node metastasis \u003csup\u003e[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]\u003c/sup\u003e. Studies focusing on head and neck tumors have shown that stromal-derived TGF-β1 can induce tumor cells to upregulate \u003cem\u003eNRP2\u003c/em\u003e, thereby enhancing their invasive capability and increasing the tendency for lung metastasis \u003csup\u003e[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e. In oral squamous cell carcinoma, \u003cem\u003eNRP2\u003c/em\u003e is similarly highly expressed, corresponding to a higher incidence of lymph node metastasis in patients and significantly shortened survival times \u003csup\u003e[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/sup\u003e. Beyond directly promoting tumor progression, \u003cem\u003eNRP2\u003c/em\u003e is also a key molecule affecting tumor immune escape. On one hand, \u003cem\u003eNRP2\u003c/em\u003e highly expressed in tumor-associated macrophages (TAMs) can improve the efficiency of macrophages in phagocytosing and clearing tumor cell apoptotic debris, removing cell fragments in an immunologically silent manner, thereby maintaining the immune tolerance state of the tumor microenvironment. In mouse models, knocking out \u003cem\u003eNRP2\u003c/em\u003e in macrophages leads to increased accumulation of apoptotic cells and induces secondary necrosis, accompanied by massive infiltration of effector CD8\u003csup\u003e+\u003c/sup\u003e T cells and NK cells within the tumor \u003csup\u003e[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/sup\u003e. This suggests that \u003cem\u003eNRP2\u003c/em\u003e is a key molecule linking efferocytosis in tumor tissues with the immunosuppressive state. On the other hand, \u003cem\u003eNRP2\u003c/em\u003e can also function as a co-inhibitory receptor on T lymphocytes. Research has found that \u003cem\u003eNRP2\u003c/em\u003e expression levels are very high on exhausted human and mouse effector CD4\u0026thinsp;+\u0026thinsp;T cells, and it is often co-expressed with classical checkpoint molecules such as PD-1, CTLA-4, LAG3, and TIM-3. Functionally, the loss of \u003cem\u003eNRP2\u003c/em\u003e leads to overactive CD4\u003csup\u003e+\u003c/sup\u003e T cell responses and enhanced inflammatory reactions. In transplant rejection models, this situation manifests as accelerated immune attack and graft loss \u003csup\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWhile these findings are promising, several limitations must be acknowledged. First, although the sample sizes for the OSF and OLK datasets were relatively small, they exhibited pronounced inter-group heterogeneity. This robust biological distinction reinforces the reliability of the identified core genes despite the limited sample numbers. Second, while \u003cem\u003eGALNT18\u003c/em\u003e and \u003cem\u003eHOMER3\u003c/em\u003e showed strong diagnostic potential, they were not significant prognostic factors for overall survival; this discrepancy suggests their roles might be more critical in the early establishment of the tumor microenvironment (initiation) rather than in driving late-stage mortality. Finally, our conclusions rely on bioinformatic predictions. Although supported by literature, definitive causal relationships require further validation through in vitro or in vivo experimental models.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eIn this study, we systematically integrated transcriptomic data from OSF, OLK, and OSCC to map the shared molecular landscape bridging oral mucosal potentially malignant disorders and invasive carcinoma. By utilizing OSCC as a malignant reference, we filtered out non-essential reactive changes and isolated a core gene signature capable of driving the entire disease continuum.\u003c/p\u003e \u003cp\u003eOur analysis identified \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e as key molecular markers in this process. These genes do not merely mark the presence of disease; they likely function as critical drivers that orchestrate extracellular matrix remodeling and shape a pro-tumorigenic immune microenvironment\u0026mdash;specifically through macrophage regulation and the facilitation of immune evasion.\u003c/p\u003e \u003cp\u003eWe propose that these shared molecular signatures offer a dual clinical value: they serve as sensitive biomarkers for risk stratification in patients with OSF or OLK, and, more importantly, represent potential therapeutic targets. Targeting these early drivers provides a compelling strategy for chemoprevention, aiming to intercept the malignant transformation process before it progresses to irreversible invasive cancer.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research received no external funding.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eClinical trial number\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eDeclaration of Artificial Intelligence Use\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that artificial intelligence (AI) was used solely for English language refinement and grammar improvement during manuscript preparation. No generative AI was used for data analysis, image creation, or content generation. The authors take full responsibility for the integrity and accuracy of the final manuscript.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eauthorship contribution statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization, J.Z. and C.X.; methodology, J.Z.; software, J.Z.; validation, T.T., W.L., Y.P., Z.Q. and Y.C.; formal analysis, J.Z.; investigation, J.Z. and Y.C.; resources, C.X.; data curation, J.Z.; writing\u0026mdash;original draft preparation, J.Z.; writing\u0026mdash;review and editing, C.X.; visualization, J.Z.; supervision, C.X.; project administration, C.X. All authors have read and agreed to the published version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics, Consent to Participate, and Consent to Publish Declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets analyzed during the current study are publicly available in the Gene Expression Omnibus (GEO) repository (https://www.ncbi.nlm.nih.gov/geo/). The specific accession numbers utilized for the analyses are: GSE64216, GSE246050, GSE30784, GSE25099, GSE12586, and GSE227919.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of competing interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors declare that No conflict of interest exists.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eZhi Y, Wang Q, Zi M, et al. Spatial transcriptomic and metabolomic landscapes of oral submucous fibrosis-derived oral squamous cell carcinoma and its tumor microenvironment. Adv Sci (Weinh). 2024;11(12):e2306515. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/advs.202306515\u003c/span\u003e\u003cspan address=\"10.1002/advs.202306515\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe Y, Shao F, Pi W, et al. Largescale transcriptomics analysis suggests over-expression of BGH3, MMP9 and PDIA3 in oral squamous cell carcinoma. PLoS ONE. 2016;11(1):e0146530. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pone.0146530\u003c/span\u003e\u003cspan address=\"10.1371/journal.pone.0146530\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCui Y, Ding C, Peng C, et al. Progress in clinicopathological diagnosis of oral potentially malignant disorders. Hua Xi Kou Qiang Yi Xue Za Zhi. 2025;43(3):314\u0026ndash;24. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7518/hxkq.2025.2024418\u003c/span\u003e\u003cspan address=\"10.7518/hxkq.2025.2024418\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaik V, Kumar M, Solomon MC, et al. Clinical and histological indicators for malignant transformation of oral submucous fibrosis: an analysis of cases from a tertiary care center. BMC Oral Health. 2025;25(1):1053. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12903-025-06443-y\u003c/span\u003e\u003cspan address=\"10.1186/s12903-025-06443-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu X, Zhang J, Li J, et al. Diagnosing malignant transformation potential in oral submucous fibrosis with leukoplakia: an approach utilizing deoxyribonucleic acid image cytometry. Sci Rep. 2025;15(1):22085. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41598-025-06585-x\u003c/span\u003e\u003cspan address=\"10.1038/s41598-025-06585-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTilakaratne WM, Jayasooriya PR, Jayasuriya NS et al. 2019. Oral epithelial dysplasia: Causes, quantification, prognosis, and management challenges. Periodontol 2000 80(1):126\u0026ndash;147 \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/prd.12259\u003c/span\u003e\u003cspan address=\"10.1111/prd.12259\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu W, Wang YF, Zhou HW, et al. Malignant transformation of oral leukoplakia: a retrospective cohort study of 218 Chinese patients. BMC Cancer. 2010;10:685. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/1471-2407-10-685\u003c/span\u003e\u003cspan address=\"10.1186/1471-2407-10-685\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlsarraf A, Kujan O, Farah CS. A systematic review of oral biopsies, sample types, and detection techniques applied in relation to oral cancer detection. BioTech (Basel). 2021;11(1):5. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/biotech11010005\u003c/span\u003e\u003cspan address=\"10.3390/biotech11010005\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePoh CF, Ng S, Berean KW, et al. Biopsy and histopathologic diagnosis of oral premalignant and malignant lesions. J Can Dent Assoc. 2008;74(3):283\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNatarajan PM, Umapathy VR. Role of transcriptomics in the study of oral cancer. Front Oral Health. 2025;6:1524364. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/froh.2025.1524364\u003c/span\u003e\u003cspan address=\"10.3389/froh.2025.1524364\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSingh P, Rai A, Dohare R, et al. Network-based identification of signature genes KLF6 and SPOCK1 associated with oral submucous fibrosis. Mol Clin Oncol. 2020;12(4):299\u0026ndash;310. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3892/mco.2020.1997\u003c/span\u003e\u003cspan address=\"10.3892/mco.2020.1997\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMahmoudi Anzabi R, Davodpour AH, Ghodratizadeh S, et al. Recent knowledge on squamous cell carcinoma of the oral cavity: Contributing factors, underlying molecular pathways, and current attitudes in the therapeutic approaches. Int J Mol Cell Med. 2025;14(3):928\u0026ndash;48. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.22088/IJMCM.BUMS.14.3.928\u003c/span\u003e\u003cspan address=\"10.22088/IJMCM.BUMS.14.3.928\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi Y, Tang Z, Xiao C, et al. SOX11 and GLIS2: Novel biomarkers for understanding the progression of oral leukoplakia. Int Dent J DOI. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.identj.2025.100872\u003c/span\u003e\u003cspan address=\"10.1016/j.identj.2025.100872\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Zhang J, Zhao S, et al. Single-cell RNA sequencing highlights the immunosuppression of IDO1\u0026thinsp;+\u0026thinsp;macrophages in the malignant transformation of oral leukoplakia. Theranostics. 2024;14(12):4787\u0026ndash;805. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.7150/thno.96273\u003c/span\u003e\u003cspan address=\"10.7150/thno.96273\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarrett T, Wilhite SE, Ledoux P, et al. NCBI GEO: archive for gene expression and epigenomics data sets: 23-year update. Nucleic Acids Res. 2024;52(D1):D138\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkad1013\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkad1013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eH\u0026auml;yrinen MJ, Kiiskil\u0026auml; J, Ranki A, et al. The transcription factor Twist1 has a significant role in Mycosis Fungoides (MF) cell biology: an RNA sequencing study of 40 MF cases. Cancers (Basel). 2023;15(5):1527. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/cancers15051527\u003c/span\u003e\u003cspan address=\"10.3390/cancers15051527\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSherman BT, Hao M, Qiu J, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50(W1):W216\u0026ndash;21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkac194\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkac194\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSzklarczyk D, Gable AL, Nastou KC, et al. The STRING database in 2021: customizable protein-protein networks, and functional characterization of user-uploaded gene/measurement sets. Nucleic Acids Res. 2021;49(D1):D605\u0026ndash;12. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkaa1074\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkaa1074\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTang Q, Chen Y, Li X, et al. Prognostic model revealing pyroptosis-related signatures in oral squamous cell carcinoma based on bioinformatics analysis. Bioengineered. 2021;12(1):8565\u0026ndash;76. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/21655979.2021.1988362\u003c/span\u003e\u003cspan address=\"10.1080/21655979.2021.1988362\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa Y, Ming Y, Hou Z, et al. Deciphering the overlapping immune mechanism between depression and breast cancer. Int J Mol Sci. 2025;26(11):5229. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/ijms26115229\u003c/span\u003e\u003cspan address=\"10.3390/ijms26115229\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNewman AM, Liu CL, Green MR, et al. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nmeth.3337\u003c/span\u003e\u003cspan address=\"10.1038/nmeth.3337\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIndellicato R, Pisano A. Concerted regulation of glycosylation factors sustains tissue identity and function. Biomedicines. 2022;10(8):1805. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/biomedicines10081805\u003c/span\u003e\u003cspan address=\"10.3390/biomedicines10081805\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen D, Nie Y, Shi L, et al. GALNT18-mediated O-glycosylation promotes development of nonalcoholic steatohepatitis-associated hepatocellular carcinoma through activation of EGR1/TGF-β1/Smad signaling. Mol Biotechnol DOI. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s12033-025-01494-6\u003c/span\u003e\u003cspan address=\"10.1007/s12033-025-01494-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePucci M, Malagolini N, Dall'Olio F. Glycosyltransferases in cancer: prognostic biomarkers of survival in patient cohorts and impact on malignancy in experimental models. Cancers (Basel). 2022;14(9):2129. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/cancers14092129\u003c/span\u003e\u003cspan address=\"10.3390/cancers14092129\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen L, Shan X, Wan X, et al. HOMER3 promotes liver hepatocellular carcinoma cancer progression by upregulating EZH2 and mediating miR-361/GPNMB axis. Pathol Res Pract. 2024;257:155280. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.prp.2024.155280\u003c/span\u003e\u003cspan address=\"10.1016/j.prp.2024.155280\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo M, Zhao C, Zhao Y, et al. Identification of homer protein homolog 3 as a prognostic marker of colon adenocarcinoma. Heliyon. 2024;10(13):e33344. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.heliyon.2024.e33344\u003c/span\u003e\u003cspan address=\"10.1016/j.heliyon.2024.e33344\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Q, He L, Lin C, et al. HOMER3 facilitates growth factor-mediated β-Catenin tyrosine phosphorylation and activation to promote metastasis in triple negative breast cancer. J Hematol Oncol. 2021;14(1):6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s13045-020-01020-8\u003c/span\u003e\u003cspan address=\"10.1186/s13045-020-01020-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuan R, Yang C, Niu X. HOMER3 and its correlation with immune-infiltrating M2 macrophages in colorectal cancer. Discov Oncol DOI. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s12672-025-04247-8\u003c/span\u003e\u003cspan address=\"10.1007/s12672-025-04247-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchellenburg S, Schulz A, Poitz DM, et al. Role of neuropilin-2 in the immune system. Mol Immunol. 2017;90:239\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.molimm.2017.08.010\u003c/span\u003e\u003cspan address=\"10.1016/j.molimm.2017.08.010\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang M, Zhang W, Wu A, et al. Prognostic significance of VEGF-C, semaphorin 3F, and neuropilin-2 expression in oral squamous cell carcinomas and their relationship with lymphangiogenesis. J Surg Oncol. 2015;111(4):382\u0026ndash;8. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jso.23843\u003c/span\u003e\u003cspan address=\"10.1002/jso.23843\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRecalde-Percaz L, de la Guia-Lopez I, Linzoain-Agos P, et al. Neuropilin-2 upregulation by stromal TGFβ1 induces lung disseminated tumor cells dormancy escape and promotes metastasis outgrowth. Neoplasia. 2025. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.neo.2025.101220\u003c/span\u003e\u003cspan address=\"10.1016/j.neo.2025.101220\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKang Y, Zhang Y, Zhang Y, et al. NRP2, a potential biomarker for oral squamous cell carcinoma. Am J Transl Res. 2021;13(8):8938\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoy S, Bag AK, Singh RK, et al. Macrophage-derived neuropilin-2 exhibits novel tumor-promoting functions. Cancer Res. 2018;78(19):5600\u0026ndash;17. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.CAN-18-0568\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-18-0568\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWedel J, Kochupurakkal N, Kong SW, et al. Neuropilin-2 functions as a coinhibitory receptor to regulate antigen-induced inflammation and allograft rejection. J Clin Invest. 2025;135(13):e172218. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1172/JCI172218\u003c/span\u003e\u003cspan address=\"10.1172/JCI172218\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9127273/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9127273/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground.\u003c/h2\u003e \u003cp\u003eOral submucous fibrosis (OSF), oral leukoplakia (OLK), and oral squamous cell carcinoma (OSCC) are prevalent disorders associated with oral mucosal malignancy. However, the common mechanisms underlying the progression and shared characteristics of these malignancy-associated disorders remain unclear. This study aims to investigate the core common differentially expressed genes (DEGs) shared by OSF, OLK, and OSCC, providing novel targets for the diagnosis and evaluation of oral mucosal malignancy-associated disorders.\u003c/p\u003e\u003ch2\u003eMethods.\u003c/h2\u003e \u003cp\u003eWe mined the Gene Expression Omnibus (GEO) database to pinpoint overlapping transcriptomic signatures across OSF, OLK, and OSCC. Functional enrichment analyses, utilizing Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), were performed to map biological pathways. Protein\u0026ndash;protein interaction (PPI) networks were constructed to identify candidate genes. These candidates were further screened via Least Absolute Shrinkage and Selection Operator (LASSO) regression and Receiver Operating Characteristic (ROC) validation to prioritize core markers. Additionally, immune infiltration assessments and single-gene Gene Set Enrichment Analysis (GSEA) were conducted to explore mechanistic links, while survival analysis was employed to evaluate prognostic value.\u003c/p\u003e\u003ch2\u003eResults.\u003c/h2\u003e \u003cp\u003eWe identified 94 co-expressed genes, which were primarily clustered in biological processes related to extracellular matrix (ECM) remodeling, epithelial\u0026ndash;mesenchymal transition (EMT), and immune regulation. Through rigorous screening, \u003cem\u003eGALNT18\u003c/em\u003e, \u003cem\u003eHOMER3\u003c/em\u003e, and \u003cem\u003eNRP2\u003c/em\u003e were prioritized as the final core genes. These markers demonstrated consistent correlations with specific infiltrating immune cells and ECM-related signaling pathways. Notably, while all three genes served as robust diagnostic markers, high \u003cem\u003eNRP2\u003c/em\u003e expression was specifically associated with poor overall survival.\u003c/p\u003e","manuscriptTitle":"Identification of GALNT18, HOMER3, and NRP2 as shared molecular signatures associated with stromal remodeling and immune suppression in oral mucosal malignancy-associated disorders","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-02 11:30:30","doi":"10.21203/rs.3.rs-9127273/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"115410770944440916599247597812673810072","date":"2026-05-18T11:21:00+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-12T15:31:14+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-08T16:16:14+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"192164684961309783066632432666497209996","date":"2026-04-02T14:15:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"184574301173644650355604015491279776522","date":"2026-03-31T16:14:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"333053348943250922792378638054118271257","date":"2026-03-31T11:15:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"338162149506747930015827635809067527724","date":"2026-03-31T08:37:19+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-31T08:19:43+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-30T05:25:48+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-28T02:03:22+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-26T12:58:57+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Oncology","date":"2026-03-26T12:53:48+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2ba4303c-54af-4f6e-8837-f5be6c4ccb33","owner":[],"postedDate":"April 2nd, 2026","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"115410770944440916599247597812673810072","date":"2026-05-18T11:21:00+00:00","index":101,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-02T11:30:30+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-02 11:30:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9127273","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9127273","identity":"rs-9127273","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}