Significance of Liquid-Liquid Phase Separation-related Genes in the Prognostic Assessment of Oral Squamous Cell Carcinoma: A Multi-omics Analysis

Significance of Liquid-Liquid Phase Separation-related Genes in the Prognostic Assessment of Oral Squamous Cell Carcinoma: A Multi-omics Analysis | 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 Significance of Liquid-Liquid Phase Separation-related Genes in the Prognostic Assessment of Oral Squamous Cell Carcinoma: A Multi-omics Analysis Ding Luo, Huan Li, Jie Jing This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4129536/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Liquid-liquid phase separation (LLPS) is implicated in tumorigenesis and progression, yet its role in oral squamous cell carcinoma (OSCC) remains unexplored. This study aims to identify LLPS-associated genes in OSCC and develop a prognostic assessment model. Methods We analyzed 334 OSCC and 32 normal samples from the TCGA-HNSC cohort. Inclusion criteria encompassed histologically verified primary OSCC, mRNA profiles, and pertinent clinical data, while samples with no survival status or survival time less than 30 days were excluded. The final cohort consisted of 297 OSCC patients with complete data on age, gender, TNM staging, and grading. We utilized single-cell sequencing data from GEO (GSE103322), with GSE42743 as the validation cohort. LLPS-related genes from DrLLPS were employed, and key genes were identified through weighted co-expression network and clustering analysis. Prognostic models were developed using Coxboost, Lasso regression, and Stepcox regression. Additionally, immune infiltration analysis was conducted to study the immune microenvironment of OSCC. Results The study established a predictive model based on eight LLPS-related genes in OSCC (VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, GAS2L3, STIL). OSCC patients were stratified into two groups: high-risk and low-risk, with the high-risk group exhibiting significantly poorer prognosis ( p < 0.05). Furthermore, notable differences in the immune environment were also observed between the groups. Conclusions This study identified eight LLPS-associated genes critical for OSCC prognosis and immune status, leading to the development of a predictive model. This research holds significance for advancing OSCC diagnosis and treatment strategies. oral squamous cell carcinoma liquid-liquid phase separation single cell sequencing analysis machine learning immune microenvironment Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Introduction Among head and neck squamous cell carcinomas (HNSC), oral squamous cell carcinoma (OSCC) is considered one of the most prevalent diseases, accounting for up to 90% 1–3 . While surgical and adjuvant therapeutic techniques have seen improvements, leading to remarkable treatment outcomes for early-stage OSCC, the prognosis for advanced-stage cases remains unsatisfactory 4,5 . Even with complete surgical resection, some patients with advanced OSCC survive for less than 30 months, and the overall five-year survival rate ranges from only 50–60% 6,7 . Identifying OSCC biomarkers and establishing a relevant prognostic classification system are therefore crucial for enhancing prognostic prediction and developing targeted therapies. Liquid-liquid phase separation (LLPS), initially defined as an engineering technique in physics and chemistry, has been found to also play a role in biomolecular condensate formation, a universal mechanism for the spatiotemporal coordination of biological activities in cells 8–10 . Biomolecular condensates are widely observed to directly regulate key cellular processes involved in cancer cell pathology, and dysregulation of LLPS is an implicit driver of oncogenic activity, capable of inducing cancer through various mechanisms, including chromosome replication, cell signaling, and DNA repair 11–14 . For instance, SENP1 reduces RNF168 SUMOylation and LLPS, promotes DNA damage repair, protects genomic integrity, and drives chemotherapy resistance 15 . Additionally, biomolecular condensates and LLPS of heat-shock factor 1, a transcriptional regulator of chaperones, affect cancer development 16 . Furthermore, the guanine nucleotide exchange factor Son of Seven less participates in the occurrence and development of cancer by regulating RAS signaling 17 . Despite LLPS being closely linked to various cancers, its specific role in OSCC remains undefined. The significance of LLPS in OSCC has not been fully understood. This study aims to explore the impact of LLPS-related genes on OSCC. It primarily investigates the role of these genes in OSCC using both single-cell sequencing analysis and transcriptome sequencing analyses. Additionally, a predictive model was established to assess the prognosis and immune status of OSCC patients. This study is pivotal in uncovering new biomarkers and advancing treatment strategies for OSCC. Methods Download of LLPS-related genes A large set of LLPS phenotype-related genes was retrieved from the DrLLPS website ( http://llps.biocuckoo.cn/ ). Only 3611 protein cording genes were selected and used for following analysis. Transcriptome dataset download and processing RNA-seq data from the Cancer Genome Atlas (TCGA) and GEO databases were used for this study. Transcriptomic data of HNSC with clinical information from TCGA were obtained through database searching. A total of 327 samples with gene expression matrix and matched clinical data were used for future study. The GEO dataset GSE42743 for oral squamous cell carcinoma was used as an independent external validation cohort. Transcriptome data was log2 transformed for future analysis. Weighted Gene Co-expression Network Analysis (WGCNA) R package WGCNA was employed to classify co-expressed genes into modules 18 . Briefly, the "WGCNA" R package's pickSoftThreshold function was used to determine the optimal soft-threshold value. Step sizes of 1:10 and 12:20 were adjusted to 1 and 2, respectively. DeepSplit was set to 2, and the minimum module gene number was set to 30. KEGG/GO enrichment analysis KEGG and GO enrichment analyses were performed to analyze the biological activities and signaling pathways associated with the identified genes in different modules based on the WGCNA results. In this study, R package ClusterProfiler were used to perform enrichment analysis 19 . Single-sample gene set enrichment analysis (ssGSEA) The study utilized ssGSEA, an extension of GSEA, to calculate and obtain enrichment scores for each sample-gene set pair, depicting the degree of coordinated upregulation or downregulation of given gene set members in each sample. This analysis was employed to compute and acquire enrichment scores for the LLPS phenotype in each sample. Construction and validation of prognostic model Initially, single-factor COX regression identified genes linked to prognosis. Subsequently, LASSO regression was applied for a refined analysis of prognosis-related genes, Family was set as "Cox," and the value of Maxit was set to 500. The model's efficacy in accurately categorizing patients into risk groups was evaluated using both training and validation cohorts, focusing on survival differences between high-risk and low-risk groups. Single-cell sequencing data download and processing The Gene Expression Omnibus (GEO) database ( https://www.ncbi.nlm.nih.gov/geo/ ) provides transcriptional and single-cell sequencing data for various diseases. The single-cell sequencing dataset GSE103322 for OSCC was obtained from GEO. The dataset underwent quality control by the authors. DIMS value was set at 1:10, and UMAP method was employed for dimensionality reduction. K. Peram value was set to 10 with a random seed of 912036023, and KNN method was used for cell grouping. SingleR method was used for cell annotation. Immune infiltration analysis Immune cell infiltration was quantified using the cibersort R package, accessed via the cibersortX website ( https://cibersortx.stanford.edu/ ) 20 . Calculation results for each OSCC sample were retrieved from cibersortX. Variations in immune cell values between high-risk and low-risk groups were calculated in the model, and results were visualized using box plots. Statistical analysis Coxboost, Lasso, and Stepcox analyses were employed to screen for prognosis-related genes, and Kaplan-Meier survival analysis was used to evaluate patient outcomes. The wlicox test compared gene expression between high-risk and low-risk groups, with statistical significance set at p < 0.05. Results WGCNA screening for LLPS-related genes In the TCGA cohort, ssGSEA analysis quantified LLPS enrichment scores for each OSCC patient. WGCNA was then employed to identify genes associated with the LLPS phenotype in OSCC. The results revealed that when the soft threshold was set to 10, not only was R^2 > 0.9, indicating that the data suitable for WGCNA analysis, with mean connectivity stabilizing (Fig. 1 A). Next, the minimum gene number within the modules was set to 30, deepSplit was set to 3, and by setting cutHeight to 0.25 to merge similar modules, a total of 13 non-grey gene modules were obtained, as shown in Fig. 1 B. The turquoise and yellow modules exhibited the strongest correlation with LLPS (Cor = 0.65 & p < 0.001; Cor = 0.47 & p < 0.001, Fig. 1 C), suggesting a close association with LLPS in OSCC. Further analysis revealed a strong positive correlation between module membership and gene significance for both the turquoise and yellow modules (cor = 0.35 & p < 0.001; Cor = -0.57 & p < 0.001, Fig. 1 D, 1 E). We then selected genes from these modules for further analysis. Genes in turquoise and yellow modules associated with cell cycle To further investigate the function of genes in turquoise and yellow modules, we performed KEGG and GO enrichment analysis. Enrichment results for KEGG pathways are shown in Fig. 2 A. Cell cycle, cellular senescence, breast cancer, and thyroid hormone signaling pathway were the most enriched pathways. GO analysis results are depicted in Fig. 2 B. At the Biological Process (BP) level, chromosome segregation, mitotic cell cycle phase transition, and DNA replication were enriched. At the Cellular Component (CC) level, chromosomal region, spindle, and nuclear chromosome were enriched. At the Molecular Function (MF) level, protein kinase phosphatase activity, ATP hydrolysis activity, and protein serine kinase activity were enriched. All above suggested that genes in turquoise and yellow modules may involve in DNA replication and cell cycle progress. Single-cell sequencing analysis screening for LLPS-related genes Besides the bulk RNA-seq data used above, a single cell RNA-seq data from GEO database (GSE103322) were also used to screen LLPS-related genes. Using SingleR, cells were annotated into 8 categories, as illustrated in Fig. 3 A: fibroblasts, epithelial cells, B cells, smooth muscle cells, macrophages, endothelial cells, mast cells, and T cells. Figure 3 B depicts the expression of LLPS, which mainly expressed in epithelial cells. Based on LLPS-related gene enrichment scores, cells were categorized into high LLPS and low LLPS groups (Fig. 3 C). It was observed that low LLPS cells were mainly distributed in B cells, T cells, and mast cells, while high LLPS cells were predominantly found in epithelial cells and macrophages (Fig. 3 D). Considering the close association between OSCC and epithelial Cells, differential gene expression analysis was performed on epithelial cells from the high LLPS and low LLPS groups. A total of 692 LLPS-related genes were obtained by setting p < 0.05 (Fig. 3 E). Construction and Validation of LLPS-Related Prognostic Model Then, LLPS-related genes obtained from single-cell sequencing analysis were cross-referenced with module genes obtained from WGCNA analysis, resulting in 168 candidate genes (Fig. 4 A, 4 B). COXBOOST was then employed to screen out 67 candidate genes. Using LASSO regression, with a random seed set to 912034053 and maxit = 500, the optimal lambda value was determined as 0.03969583, resulting in a set of 15 genes forming the features, including "VRK1," "PLK1," "DEPDC1," "HMMR," "KIF15," "POLE2," "ESCO2," "CCNE2," "KLHL13," "NEIL3," "CENPI," "GAS2L3," "STIL," "YEATS2," and "AGO4" (Fig. 4 C, 4 D). Subsequently, Stepcox analysis identified 8 genes: VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, GAS2L3, and STIL. The risk values of the model are calculated as follows: RISK = -0.095362 * ESCO2–0.047149 * POLE2–0.072359 * STIL + 0.015412 * VRK1 + 0.011161 * PLK1 + 0.034623 * NEIL3 + 0.073969 * CENPI + 0.085009 * GAS2K3. Using the median value of the model (LLPS), all oral squamous cell carcinoma specimens were divided into LLPS_high-risk and LLPS_low-risk groups. The prognosis among different subgroups was then compared (Fig. 5 A). Survival curve analysis indicated a poorer prognosis in the LLPS_high-risk group (p < 0.001) (Fig. 5 B). The diagnostic values, represented by AUC values at 1, 3, and 5 years in the TCGA cohort, were 0.67, 0.70, and 0.60, respectively, demonstrating diagnostic significance (Fig. 5 B). This observation was further supported by Fig. 5 C. Similarly, in the external validation cohort GSE42743, survival analysis revealed a poorer prognosis in the LLPS_high-risk group (p < 0.001) (Fig. 5 D). The diagnostic values at 1, 3, and 5 years in the validation set were 0.74, 0.75, and 0.77, respectively, highlighting diagnostic significance (Fig. 5 E). This observation was also supported by Fig. 5 F. In Fig. 6 A, single-factor analysis, and Fig. 6 B, multi-factor COX analysis, only RISK and Age were identified as independent prognostic risk factors. Construction of the nomogram As shown in Fig. 7 A, we observed that the 1-year, 3-year, and 5-year mortality rates for patient TCGA-BA-4074 were 0.413, 0.815, and 0.905, respectively. To further evaluate the nomogram’s predictive accuracy, calibration curve analysis was performed, as depicted in Fig. 7 B. The predicted calibration curves for 1, 3, and 5 years closely aligned with the actual outcomes. Furthermore, continuous ROC analysis revealed that the AUC fluctuated around 0.7, notably higher than other clinical features such as age, gender, T, N, and stage, indicating the nomogram's high accuracy in predicting patient prognosis (Fig. 7 C). Decision curve analysis further suggested that clinical interventions based on the LLPS model values would be highly beneficial for patients (Fig. 7 D). Cellular localization of model genes Next, the expression of the eight genes in the model was investigated at the single-cell level. As depicted, VRK1 is expressed in epithelial cells, T cells, endothelial cells, and macrophages. PLK1 is expressed in epithelial cells, T cells, and macrophages. POLE2, ESCO2, NEIL3, and CENPI are expressed in epithelial cells. STIL is expressed in epithelial cells and T cells. GAS2L3 is expressed in fibroblasts, epithelial cells, and macrophages (Fig. 8 ). Immune Infiltration Analysis of LLPS-related genes Using the CIBERSORT method, we constructed immune cell profiles for 21 types of immune cells in OSCC cases to analyze the infiltration of immune subtypes in tumors (Fig. 9 ). A total of 10 tumor-infiltrating immune cells (TICs) showed significant associations with RISK gene expression (P < 0.001), T cells CD4 + memory resting, NK cells resting, and mast cells activated showed higher expression in high-risk group, while T cells CD8 + , T cells follicular helper, Tregs, NK cells activated, Dendritic cells resting, mast cells resting, and neutrophils were higher expressed in low-risk group(Fig. 10 ). The graphical results indicated a positive correlation between resting memory CD4 + T cells, resting NK cells, activated macrophages, and three TICs with RISK expression. Conversely, CD8 + T cells, follicular T cells, T regulatory cells, resting macrophages, mesenchymal stem cells, resting dendritic cells, neutrophils, and seven other TICs exhibited a negative correlation with RISK expression. Additionally, the majority of the 8 HUB genes were highly correlated with NK cells and CD4 + memory T cells, suggesting that liquid-liquid phase separation (LLPS) may influence OSCC pathogenesis through these cell types. In summary, these findings suggest that RISK genes may impact the immune response in OSCC by affecting the involvement of immune cells (Fig. 11 ). Discussion The study employs multi-omics analysis to investigate the expression, heterogeneity, prognostic value, and immunological evaluation of LLPS-related genes in OSCC. Initially, cells were categorized into groups with high and low LLPS scores based on single-cell sequencing cluster analysis. The differentially expressed genes suggest a close association between LLPS heterogeneity and these genes in OSCC. WGCNA identified the blue and yellow modules as highly correlated with LLPS in OSCC, indicating a close relationship between genes in these modules and LLPS regulation in OSCC. The intersection of these genes with those identified in previous single-cell sequencing studies resulted in the identification of key LLPS hub genes in OSCC, including VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, GAS2L3, and STIL. A prediction model was constructed using COXBOOST regression, Lasso regression, and STEPCOX based on these eight genes. Through this predictive model, OSCC patients were categorized into high-risk and low-risk groups, with significantly poorer prognosis observed in the high-risk group. The high-risk and low-risk groups exhibited marked differences in immune cell infiltration, potentially influencing their divergent prognoses and guiding immunotherapy strategies. Research has demonstrated that these eight genes play a pivotal role in cancer development. VRK1 has been shown to phosphorylate various transcription factors, including p53 and proteins related to the DNA damage response pathway, exerting a tumor-promoting effect (18) 21 . PLK1 is known to have a critical function in the cell cycle process, particularly in the G2-M checkpoint, mitosis, and cytokinesis, indicating a close association with uncontrolled proliferation of cancer cells (19) 22 . POLE2 promotes the malignant phenotype of glioblastoma by facilitating FOXM1 stability mediated by AURKA (20) 23 . ESCO2 is crucial for promoting malignant progression of hypopharyngeal carcinoma through the STAT1 pathway (21) 24 . NEIL3 activates the BRAF/MEK/ERK/TWIST pathway-mediated EMT and treatment resistance, leading to the progression of hepatocellular carcinoma (22) 25 . CENPI is overexpressed in colorectal cancer (CRC), acting as an oncogene in regulating CRC cell migration, invasion, and epithelial-mesenchymal transition (EMT) (23) 26 . GAS2L3 is found to be highly expressed in various cancers, but its specific mechanism of action remains unclear (24–26) 27–29 . Additionally, a multitude of LLPS-related biomarkers have been identified in various tumors. Wang et al. constructed a uterine endometrial cancer biomarker comprising four LLPS-related genes (27) 30 . Xie et al. identified PGAM1 as a crucial LLPS-related gene in breast cancer (28) 31 . Fang et al. utilized weighted co-expression network analysis and regression analysis to establish features in hepatocellular carcinoma consisting of six LLPS-related genes, with the aim of assessing patient prognosis and immune status (29) 32 . Sun et al. revealed the utility of LLPS-related genes for genotypic classification in bladder cancer (30) 33 . Presently, research indicates that OSCC is linked to the epithelial-to-mesenchymal transition (31) 34 . In comparison with these published studies, our research comprehensively integrates the biological context, with a specific focus on epithelial cells, offering valuable insights for OSCC diagnosis and treatment. Limitations This study has several limitations. Firstly, it lacks experimental validation of the results. Secondly, it did not utilize its own clinical samples to confirm the findings. Thirdly, there was a lack of demographic data, such as age, gender, and ethnicity, which may introduce bias. Fourthly, the role of LLPS in specific gene mutation-associated OSCC was not investigated. These areas will be a focus of our future research. Conclusion In conclusion, this study identified eight LLPS-associated genes and developed a predictive model for OSCC prognosis and immune status. However, further evaluation in clinical cohorts is necessary to assess the practicality and accuracy of this model, and adjustments will be made accordingly. These findings and the model developed herein hold significant potential for advancing OSCC diagnostics and therapeutic strategies. Future research will aim to refine and validate these outcomes, contributing to the enhancement of OSCC patient care. Declarations Data availability The datasets generated and/or analyzed during the present study can be found in the Gene Expression Omnibus datasets (GEO, https://www.ncbi.nlm.nih.gov/geo/, GSE10332, and GSE42743) and The Cancer Genome Atlas(TCGA, https://portal.gdc.cancer.gov/). Funding This work was supported by the National Natural Science Foundation of China (32160175), Guizhou Provincial Basic Research Program (Natural Science) (NO. ZK [2021]440), the PhD Fund of Scientific Research Foundation of School and Hospital of Stomatology, Zunyi Medical University (NO. KY2020-14). Author information Ding Luo, Huan Li and Jie Jing have contributed equally to this work. Authors and Affiliations School and Hospital of Stomatology, Zunyi Medical University, Zunyi, Guizhou, China No. 6, Xuefu West Road, Zunyi, 563000, Guizhou, China. Ding Luo，Huan Li，Jie Jing Author contributions Ding Luo, Huan Li and Jie Jing have contributed equally to this work DL .and JJ were responsible for designing the entire study. DL and HL performed data analysis and software operation. DL and HL prepared the figures and tables. DL and HL collected materials related to this research. DL and JJ wrote the manuscript. JJ was responsible for review and editing. All authors contributed to the article and approved the final version for submission. Ethics declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Declaration of competing interest The authors declare no conflict of interest. References Capparuccia, L. & Tamagnone, L. Semaphorin signaling in cancer cells and in cells of the tumor microenvironment – two sides of a coin. J. Cell Sci. 122 , 1723–1736 (2009). Epidemiology of oral cancer in Arab countries | Saudi Medical Journal. https://smj.org.sa/content/37/3/249. Tandon, P., Dadhich, A., Saluja, H., Bawane, S. & Sachdeva, S. The prevalence of squamous cell carcinoma in different sites of oral cavity at our Rural Health Care Centre in Loni, Maharashtra – a retrospective 10-year study. Contemp. Oncol. Onkol. 21 , 178–183 (2017). A, R. et al. The Challenges of OSCC Diagnosis: Salivary Cytokines as Potential Biomarkers. J. Clin. Med. 9 , (2020). I, P. et al. Oral and Oropharyngeal squamous cell carcinoma: prognostic and predictive parameters in the etiopathogenetic route. Expert Rev. Anticancer Ther. 19 , (2019). De Felice, F. et al. Radiotherapy Controversies and Prospective in Head and Neck Cancer: A Literature-Based Critical Review. Neoplasia N. Y. N 20 , 227–232 (2018). Gharat, S. A., Momin, M. & Bhavsar, C. Oral Squamous Cell Carcinoma: Current Treatment Strategies and Nanotechnology-Based Approaches for Prevention and Therapy. Crit. Rev. Ther. Drug Carrier Syst. 33 , 363–400 (2016). Biomolecular condensates: Organizers of cellular biochemistry - PMC. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7434221/. Shin, Y. & Brangwynne, C. P. Liquid phase condensation in cell physiology and disease. Science 357 , eaaf4382 (2017). Zhang, H. et al. Liquid-liquid phase separation in biology: mechanisms, physiological functions and human diseases. Sci. China Life Sci. 63 , 953–985 (2020). Mehta, S. & Zhang, J. Liquid-liquid phase separation drives cellular function and dysfunction in cancer. Nat. Rev. Cancer 22 , 239–252 (2022). Meng, F. et al. Induced phase separation of mutant NF2 imprisons the cGAS-STING machinery to abrogate antitumor immunity. Mol. Cell 81 , 4147-4164.e7 (2021). Rubio, K., Dobersch, S. & Barreto, G. Functional interactions between scaffold proteins, noncoding RNAs, and genome loci induce liquid-liquid phase separation as organizing principle for 3-dimensional nuclear architecture: implications in cancer. FASEB J. Off. Publ. Fed. Am. Soc. Exp. Biol. 33 , 5814–5822 (2019). Zhang, X. et al. The proline-rich domain promotes Tau liquid-liquid phase separation in cells. J. Cell Biol. 219 , e202006054 (2020). Wei, M., Huang, X., Liao, L., Tian, Y. & Zheng, X. SENP1 Decreases RNF168 Phase Separation to Promote DNA Damage Repair and Drug Resistance in Colon Cancer. Cancer Res. 83 , 2908–2923 (2023). Gaglia, G. et al. HSF1 phase transition mediates stress adaptation and cell fate decisions. Nat. Cell Biol. 22 , 151–158 (2020). Huang, W. Y. C. et al. A molecular assembly phase transition and kinetic proofreading modulate Ras activation by SOS. Science 363 , 1098–1103 (2019). Langfelder, P. & Horvath, S. WGCNA: an R package for weighted correlation network analysis. BMC Bioinformatics 9 , 559 (2008). Yu, G., Wang, L.-G., Han, Y. & He, Q.-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics J. Integr. Biol. 16 , 284–287 (2012). Newman, A. M. et al. Robust enumeration of cell subsets from tissue expression profiles. Nat. Methods 12 , 453–457 (2015). Campillo-Marcos, I., García-González, R., Navarro-Carrasco, E. & Lazo, P. A. The human VRK1 chromatin kinase in cancer biology. Cancer Lett. 503 , 117–128 (2021). Zitouni, S., Nabais, C., Jana, S. C., Guerrero, A. & Bettencourt-Dias, M. Polo-like kinases: structural variations lead to multiple functions. Nat. Rev. Mol. Cell Biol. 15 , 433–452 (2014). Zhang, P. et al. POLE2 facilitates the malignant phenotypes of glioblastoma through promoting AURKA-mediated stabilization of FOXM1. Cell Death Dis. 13 , 61 (2022). Hu, J. et al. ESCO2 promotes hypopharyngeal carcinoma progression in a STAT1-dependent manner. BMC Cancer 23 , 1114 (2023). Lai, H.-H. et al. NEIL3 promotes hepatoma epithelial-mesenchymal transition by activating the BRAF/MEK/ERK/TWIST signaling pathway. J. Pathol. 258 , 339–352 (2022). Ding, N., Li, R., Shi, W. & He, C. CENPI is overexpressed in colorectal cancer and regulates cell migration and invasion. Gene 674 , 80–86 (2018). Zhou, Y. et al. Elevated GAS2L3 Expression Correlates With Poor Prognosis in Patients With Glioma: A Study Based on Bioinformatics and Immunohistochemical Analysis. Front. Genet. 12 , 649270 (2021). Gao, X. et al. Genetic expression and mutational profile analysis in different pathologic stages of hepatocellular carcinoma patients. BMC Cancer 21 , 786 (2021). Liu, B. et al. RAB42 Promotes Glioma Pathogenesis via the VEGF Signaling Pathway. Front. Oncol. 11 , 657029 (2021). Wang, J., Meng, F. & Mao, F. Single cell sequencing analysis and transcriptome analysis constructed the liquid-liquid phase separation(LLPS)-related prognostic model for endometrial cancer. Front. Oncol. 12 , 1005472 (2022). Xie, J. et al. Significance of liquid-liquid phase separation (LLPS)-related genes in breast cancer: a multi-omics analysis. Aging 15 , 5592–5610 (2023). Fang, Z.-S. et al. Liquid-Liquid Phase Separation-Related Genes Associated with Tumor Grade and Prognosis in Hepatocellular Carcinoma: A Bioinformatic Study. Int. J. Gen. Med. 14 , 9671–9679 (2021). Sun, L. et al. Identification of molecular subtypes based on liquid–liquid phase separation and cross-talk with immunological phenotype in bladder cancer. Front. Immunol. 13 , 1059568 (2022). Ling, Z., Cheng, B. & Tao, X. Epithelial-to-mesenchymal transition in oral squamous cell carcinoma: Challenges and opportunities. Int. J. Cancer 148 , 1548–1561 (2021). Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted 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-4129536","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":284011713,"identity":"4fb9ad6d-bbf6-4d28-82bf-bd8b442c9a22","order_by":0,"name":"Ding Luo","email":"","orcid":"","institution":"Zunyi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ding","middleName":"","lastName":"Luo","suffix":""},{"id":284011714,"identity":"3d773399-c282-42e5-9ad7-557435486c01","order_by":1,"name":"Huan Li","email":"","orcid":"","institution":"Zunyi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Huan","middleName":"","lastName":"Li","suffix":""},{"id":284011715,"identity":"4290732a-a303-46aa-931b-4dadb066dac4","order_by":2,"name":"Jie Jing","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYBACNmb2ww8S//yrb2NvPkCcFj72njSDhw0HGPt4jiUQp0WO54CBJEjLPIkcAyIdJpGQYJC44w4zG8+ZjzfeMNjJ6TYQ1JJ44EHimWdsbOy9my3nMCQbmx0gxpYENmYeNp6z26R5GA4kbiNCi4EEUIsEm0TOMyK1AL0vkdh22ACohY1ILaBATjiTlsDGc8zYco4BEX6Rb2Y//PBHhU2CfHvzwxtvKuzkCGpBARI8REYNshZSdYyCUTAKRsGIAADWlUJZXzOP5QAAAABJRU5ErkJggg==","orcid":"","institution":"Zunyi Medical University","correspondingAuthor":true,"prefix":"","firstName":"Jie","middleName":"","lastName":"Jing","suffix":""}],"badges":[],"createdAt":"2024-03-19 10:35:58","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4129536/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4129536/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":53751780,"identity":"4a4d2dec-28cf-45b4-ae59-c39c9ec23384","added_by":"auto","created_at":"2024-03-29 18:47:57","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1263446,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eWeighted Gene Co-expression Network Analysis. \u003c/strong\u003e(A) The soft threshold of WGCNA analysis. (B) Cluster dendrogram of gene modules. The minimum gene number per module was set to 30, deepSplit was set to 3, and cutHeight was set to 0.25. Ultimately, 14 non-grey modules were obtained. (C) The heatmap of correlation between gene modules and LLPS. (D, E) In the brown and midnight blue modules, there was a strong positive correlation between module membership and gene significance (cor=0.35 \u0026amp; p\u0026lt;0.001, cor=-0.57 \u0026amp; p\u0026lt;0.001).\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/a9f3ae12c0f935ae4016664d.png"},{"id":53751779,"identity":"42fe14c9-1529-4d4d-a15f-3d0fe85359fa","added_by":"auto","created_at":"2024-03-29 18:47:57","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":421620,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eKEGG/GO enrichment analysis of genes in turquoise and yellow modules\u003c/strong\u003e. Burble plot of KEGG enrichment analysis (A) and GO enrichment analysis (B).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/a54874ecc8aba8956c0b7f6d.png"},{"id":53754085,"identity":"b2109c3e-e90b-4818-b33e-8a90168d1fda","added_by":"auto","created_at":"2024-03-29 18:55:58","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":590129,"visible":true,"origin":"","legend":"\u003cp\u003eSingle-Cell Sequencing Analysis of GSE103322 dataset. (A) Cells were divided into 8 cell types by SingleR and visualized by UMAP plot. (B) UMAP plot showing LLPS gene expression. (C) Cells categorized into high LLPS and low LLPS groups based on LLPS-related genes. (D) Expression of LLPS genes in various cell types were shown by violin plot. (E) Differential expressed genes between high and low LLPS groups of epithelial cells were shown by volcano plot. Top 5 up-regulated and down-regulated genes were marked in volcano plot.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/21cac05d3c6efa87fbb92221.png"},{"id":53751783,"identity":"86f8716d-ddc2-4157-8fec-39bcf05e962e","added_by":"auto","created_at":"2024-03-29 18:47:57","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":399731,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction of Prognostic Model.\u003c/strong\u003e(A) (B) COXBOOST Analysis. (C) (D) LASSO Regression.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/08d49113b5b9d4f7386cfd19.png"},{"id":53751789,"identity":"1a4ffaa7-c0b0-4a8d-a71a-2b43ce172751","added_by":"auto","created_at":"2024-03-29 18:47:58","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":712026,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eValidation of the prognostic model. \u003c/strong\u003e(A) Survival analysis in the TCGA cohort. The prognosis of the LLPS_high group is significantly worse (p\u0026lt;0.0001). (B) Diagnostic values (AUC) at 1, 3, and 5 years in the TCGA cohort are 0.67, 0.70, and 0.60, respectively, indicating diagnostic significance. (C) Risk factor interaction plot in the TCGA cohort. (D) Survival analysis in the validation set. The prognosis of the LLPS_high group is significantly worse (p\u0026lt;0.0001). (E) Diagnostic values (AUC) at 1, 3, and 5 years in the validation set are 0.74, 0.75, and 0.77, respectively, indicating diagnostic significance. (F) Risk factor interaction plot in the validation set.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/2c58c320c233d76857027d64.png"},{"id":53751787,"identity":"a7b9a21e-1457-420e-8c07-7d14c9207c33","added_by":"auto","created_at":"2024-03-29 18:47:58","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":257206,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eValidation of the prognostic model. \u003c/strong\u003e(A) Univariate COX Analysis. (B) Multivariate COX Analysis.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/7c526062d58dfa20bc3d41c0.png"},{"id":53751785,"identity":"99a46b95-3c68-4be7-9a2e-3ca4a01f7b37","added_by":"auto","created_at":"2024-03-29 18:47:57","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":626721,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and evaluation of the nomogram. \u003c/strong\u003e(A) Nomogram combining LLPS score and other clinical features. The 1-year, 3-year, and 5-year mortality rates for patient TCGA-BA-4074 were 0.413, 0.815, and 0.905, respectively. (B) Calibration curve. (C) Continuous ROC analysis revealed an AUC fluctuation of 0.7 for the nomogram. (D) Decision curve analysis showed that clinical interventions based on LLPS model values provided the maximum benefit to patients.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/0f9f326b5ca724af7dca8b55.png"},{"id":53751788,"identity":"8ca725e2-6185-4055-b8ab-2c7691a830b2","added_by":"auto","created_at":"2024-03-29 18:47:58","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":445750,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGene expression of VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, STIL, and GAS2L3 at single cell resolution.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/b2c99776bd9681c32b73b022.png"},{"id":53754086,"identity":"d2b20993-06fa-4106-8f01-e755b5ce841d","added_by":"auto","created_at":"2024-03-29 18:55:58","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":1396891,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmune cell infiltration status for each sample.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/64873075d5b6e8991b5e5e7e.png"},{"id":53751790,"identity":"bc1337f4-8cdc-4c98-881e-b658e4f774e7","added_by":"auto","created_at":"2024-03-29 18:47:58","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":357986,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmune cell infiltration status for high and low-risk groups.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/a064bf4209851bc863e39569.png"},{"id":53751786,"identity":"e3c7df82-30f0-4838-bd6f-c3bfdf45e1b9","added_by":"auto","created_at":"2024-03-29 18:47:58","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":592340,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelation between HUB genes and immune cells.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/ea361fede2a1b117c0284bef.png"},{"id":57281403,"identity":"b849478f-f0e1-4eb7-a040-3b817f28e0f8","added_by":"auto","created_at":"2024-05-28 15:07:39","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":9508717,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4129536/v1/110a9e99-2515-4b5e-abe5-f0a51068a01a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Significance of Liquid-Liquid Phase Separation-related Genes in the Prognostic Assessment of Oral Squamous Cell Carcinoma: A Multi-omics Analysis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAmong head and neck squamous cell carcinomas (HNSC), oral squamous cell carcinoma (OSCC) is considered one of the most prevalent diseases, accounting for up to 90% \u003csup\u003e1\u0026ndash;3\u003c/sup\u003e. While surgical and adjuvant therapeutic techniques have seen improvements, leading to remarkable treatment outcomes for early-stage OSCC, the prognosis for advanced-stage cases remains unsatisfactory \u003csup\u003e4,5\u003c/sup\u003e. Even with complete surgical resection, some patients with advanced OSCC survive for less than 30 months, and the overall five-year survival rate ranges from only 50\u0026ndash;60% \u003csup\u003e6,7\u003c/sup\u003e. Identifying OSCC biomarkers and establishing a relevant prognostic classification system are therefore crucial for enhancing prognostic prediction and developing targeted therapies.\u003c/p\u003e \u003cp\u003eLiquid-liquid phase separation (LLPS), initially defined as an engineering technique in physics and chemistry, has been found to also play a role in biomolecular condensate formation, a universal mechanism for the spatiotemporal coordination of biological activities in cells \u003csup\u003e8\u0026ndash;10\u003c/sup\u003e. Biomolecular condensates are widely observed to directly regulate key cellular processes involved in cancer cell pathology, and dysregulation of LLPS is an implicit driver of oncogenic activity, capable of inducing cancer through various mechanisms, including chromosome replication, cell signaling, and DNA repair \u003csup\u003e11\u0026ndash;14\u003c/sup\u003e. For instance, SENP1 reduces RNF168 SUMOylation and LLPS, promotes DNA damage repair, protects genomic integrity, and drives chemotherapy resistance \u003csup\u003e15\u003c/sup\u003e. Additionally, biomolecular condensates and LLPS of heat-shock factor 1, a transcriptional regulator of chaperones, affect cancer development \u003csup\u003e16\u003c/sup\u003e. Furthermore, the guanine nucleotide exchange factor Son of Seven less participates in the occurrence and development of cancer by regulating RAS signaling \u003csup\u003e17\u003c/sup\u003e. Despite LLPS being closely linked to various cancers, its specific role in OSCC remains undefined.\u003c/p\u003e \u003cp\u003eThe significance of LLPS in OSCC has not been fully understood. This study aims to explore the impact of LLPS-related genes on OSCC. It primarily investigates the role of these genes in OSCC using both single-cell sequencing analysis and transcriptome sequencing analyses. Additionally, a predictive model was established to assess the prognosis and immune status of OSCC patients. This study is pivotal in uncovering new biomarkers and advancing treatment strategies for OSCC.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eDownload of LLPS-related genes\u003c/p\u003e \u003cp\u003eA large set of LLPS phenotype-related genes was retrieved from the DrLLPS website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://llps.biocuckoo.cn/\u003c/span\u003e\u003cspan address=\"http://llps.biocuckoo.cn/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Only 3611 protein cording genes were selected and used for following analysis.\u003c/p\u003e \u003cp\u003eTranscriptome dataset download and processing\u003c/p\u003e \u003cp\u003eRNA-seq data from the Cancer Genome Atlas (TCGA) and GEO databases were used for this study. Transcriptomic data of HNSC with clinical information from TCGA were obtained through database searching. A total of 327 samples with gene expression matrix and matched clinical data were used for future study. The GEO dataset GSE42743 for oral squamous cell carcinoma was used as an independent external validation cohort. Transcriptome data was log2 transformed for future analysis.\u003c/p\u003e \u003cp\u003eWeighted Gene Co-expression Network Analysis (WGCNA)\u003c/p\u003e \u003cp\u003eR package WGCNA was employed to classify co-expressed genes into modules\u003csup\u003e18\u003c/sup\u003e. Briefly, the \"WGCNA\" R package's pickSoftThreshold function was used to determine the optimal soft-threshold value. Step sizes of 1:10 and 12:20 were adjusted to 1 and 2, respectively. DeepSplit was set to 2, and the minimum module gene number was set to 30.\u003c/p\u003e \u003cp\u003eKEGG/GO enrichment analysis\u003c/p\u003e \u003cp\u003eKEGG and GO enrichment analyses were performed to analyze the biological activities and signaling pathways associated with the identified genes in different modules based on the WGCNA results. In this study, R package ClusterProfiler were used to perform enrichment analysis\u003csup\u003e19\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSingle-sample gene set enrichment analysis (ssGSEA)\u003c/p\u003e \u003cp\u003eThe study utilized ssGSEA, an extension of GSEA, to calculate and obtain enrichment scores for each sample-gene set pair, depicting the degree of coordinated upregulation or downregulation of given gene set members in each sample. This analysis was employed to compute and acquire enrichment scores for the LLPS phenotype in each sample.\u003c/p\u003e \u003cp\u003eConstruction and validation of prognostic model\u003c/p\u003e \u003cp\u003eInitially, single-factor COX regression identified genes linked to prognosis. Subsequently, LASSO regression was applied for a refined analysis of prognosis-related genes, Family was set as \"Cox,\" and the value of Maxit was set to 500. The model's efficacy in accurately categorizing patients into risk groups was evaluated using both training and validation cohorts, focusing on survival differences between high-risk and low-risk groups.\u003c/p\u003e \u003cp\u003eSingle-cell sequencing data download and processing\u003c/p\u003e \u003cp\u003eThe Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) provides transcriptional and single-cell sequencing data for various diseases. The single-cell sequencing dataset GSE103322 for OSCC was obtained from GEO. The dataset underwent quality control by the authors. DIMS value was set at 1:10, and UMAP method was employed for dimensionality reduction. K. Peram value was set to 10 with a random seed of 912036023, and KNN method was used for cell grouping. SingleR method was used for cell annotation.\u003c/p\u003e \u003cp\u003eImmune infiltration analysis\u003c/p\u003e \u003cp\u003eImmune cell infiltration was quantified using the cibersort R package, accessed via the cibersortX website (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cibersortx.stanford.edu/\u003c/span\u003e\u003cspan address=\"https://cibersortx.stanford.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)\u003csup\u003e20\u003c/sup\u003e. Calculation results for each OSCC sample were retrieved from cibersortX. Variations in immune cell values between high-risk and low-risk groups were calculated in the model, and results were visualized using box plots.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eCoxboost, Lasso, and Stepcox analyses were employed to screen for prognosis-related genes, and Kaplan-Meier survival analysis was used to evaluate patient outcomes. The wlicox test compared gene expression between high-risk and low-risk groups, with statistical significance set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eWGCNA screening for LLPS-related genes\u003c/h2\u003e \u003cp\u003eIn the TCGA cohort, ssGSEA analysis quantified LLPS enrichment scores for each OSCC patient. WGCNA was then employed to identify genes associated with the LLPS phenotype in OSCC. The results revealed that when the soft threshold was set to 10, not only was R^2\u0026thinsp;\u0026gt;\u0026thinsp;0.9, indicating that the data suitable for WGCNA analysis, with mean connectivity stabilizing (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Next, the minimum gene number within the modules was set to 30, deepSplit was set to 3, and by setting cutHeight to 0.25 to merge similar modules, a total of 13 non-grey gene modules were obtained, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB. The turquoise and yellow modules exhibited the strongest correlation with LLPS (Cor\u0026thinsp;=\u0026thinsp;0.65 \u0026amp; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Cor\u0026thinsp;=\u0026thinsp;0.47 \u0026amp; p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC), suggesting a close association with LLPS in OSCC. Further analysis revealed a strong positive correlation between module membership and gene significance for both the turquoise and yellow modules (cor\u0026thinsp;=\u0026thinsp;0.35 \u0026amp; \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Cor = -0.57 \u0026amp; \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD, \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). We then selected genes from these modules for further analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eGenes in turquoise and yellow modules associated with cell cycle\u003c/h2\u003e \u003cp\u003eTo further investigate the function of genes in turquoise and yellow modules, we performed KEGG and GO enrichment analysis. Enrichment results for KEGG pathways are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA. Cell cycle, cellular senescence, breast cancer, and thyroid hormone signaling pathway were the most enriched pathways. GO analysis results are depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB. At the Biological Process (BP) level, chromosome segregation, mitotic cell cycle phase transition, and DNA replication were enriched. At the Cellular Component (CC) level, chromosomal region, spindle, and nuclear chromosome were enriched. At the Molecular Function (MF) level, protein kinase phosphatase activity, ATP hydrolysis activity, and protein serine kinase activity were enriched. All above suggested that genes in turquoise and yellow modules may involve in DNA replication and cell cycle progress.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eSingle-cell sequencing analysis screening for LLPS-related genes\u003c/h2\u003e \u003cp\u003eBesides the bulk RNA-seq data used above, a single cell RNA-seq data from GEO database (GSE103322) were also used to screen LLPS-related genes. Using SingleR, cells were annotated into 8 categories, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA: fibroblasts, epithelial cells, B cells, smooth muscle cells, macrophages, endothelial cells, mast cells, and T cells. Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB depicts the expression of LLPS, which mainly expressed in epithelial cells. Based on LLPS-related gene enrichment scores, cells were categorized into high LLPS and low LLPS groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). It was observed that low LLPS cells were mainly distributed in B cells, T cells, and mast cells, while high LLPS cells were predominantly found in epithelial cells and macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). Considering the close association between OSCC and epithelial Cells, differential gene expression analysis was performed on epithelial cells from the high LLPS and low LLPS groups. A total of 692 LLPS-related genes were obtained by setting p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eConstruction and Validation of LLPS-Related Prognostic Model\u003c/h2\u003e \u003cp\u003eThen, LLPS-related genes obtained from single-cell sequencing analysis were cross-referenced with module genes obtained from WGCNA analysis, resulting in 168 candidate genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). COXBOOST was then employed to screen out 67 candidate genes. Using LASSO regression, with a random seed set to 912034053 and maxit\u0026thinsp;=\u0026thinsp;500, the optimal lambda value was determined as 0.03969583, resulting in a set of 15 genes forming the features, including \"VRK1,\" \"PLK1,\" \"DEPDC1,\" \"HMMR,\" \"KIF15,\" \"POLE2,\" \"ESCO2,\" \"CCNE2,\" \"KLHL13,\" \"NEIL3,\" \"CENPI,\" \"GAS2L3,\" \"STIL,\" \"YEATS2,\" and \"AGO4\" (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC, \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Subsequently, Stepcox analysis identified 8 genes: VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, GAS2L3, and STIL.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe risk values of the model are calculated as follows: RISK = -0.095362 * ESCO2\u0026ndash;0.047149 * POLE2\u0026ndash;0.072359 * STIL\u0026thinsp;+\u0026thinsp;0.015412 * VRK1\u0026thinsp;+\u0026thinsp;0.011161 * PLK1\u0026thinsp;+\u0026thinsp;0.034623 * NEIL3\u0026thinsp;+\u0026thinsp;0.073969 * CENPI\u0026thinsp;+\u0026thinsp;0.085009 * GAS2K3. Using the median value of the model (LLPS), all oral squamous cell carcinoma specimens were divided into LLPS_high-risk and LLPS_low-risk groups. The prognosis among different subgroups was then compared (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). Survival curve analysis indicated a poorer prognosis in the LLPS_high-risk group (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). The diagnostic values, represented by AUC values at 1, 3, and 5 years in the TCGA cohort, were 0.67, 0.70, and 0.60, respectively, demonstrating diagnostic significance (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). This observation was further supported by Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC. Similarly, in the external validation cohort GSE42743, survival analysis revealed a poorer prognosis in the LLPS_high-risk group (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). The diagnostic values at 1, 3, and 5 years in the validation set were 0.74, 0.75, and 0.77, respectively, highlighting diagnostic significance (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE). This observation was also supported by Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA, single-factor analysis, and Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, multi-factor COX analysis, only RISK and Age were identified as independent prognostic risk factors.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of the nomogram\u003c/h2\u003e \u003cp\u003eAs shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA, we observed that the 1-year, 3-year, and 5-year mortality rates for patient TCGA-BA-4074 were 0.413, 0.815, and 0.905, respectively. To further evaluate the nomogram\u0026rsquo;s predictive accuracy, calibration curve analysis was performed, as depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB. The predicted calibration curves for 1, 3, and 5 years closely aligned with the actual outcomes. Furthermore, continuous ROC analysis revealed that the AUC fluctuated around 0.7, notably higher than other clinical features such as age, gender, T, N, and stage, indicating the nomogram's high accuracy in predicting patient prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC). Decision curve analysis further suggested that clinical interventions based on the LLPS model values would be highly beneficial for patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eCellular localization of model genes\u003c/h2\u003e \u003cp\u003eNext, the expression of the eight genes in the model was investigated at the single-cell level. As depicted, VRK1 is expressed in epithelial cells, T cells, endothelial cells, and macrophages. PLK1 is expressed in epithelial cells, T cells, and macrophages. POLE2, ESCO2, NEIL3, and CENPI are expressed in epithelial cells. STIL is expressed in epithelial cells and T cells. GAS2L3 is expressed in fibroblasts, epithelial cells, and macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eImmune Infiltration Analysis of LLPS-related genes\u003c/h2\u003e \u003cp\u003eUsing the CIBERSORT method, we constructed immune cell profiles for 21 types of immune cells in OSCC cases to analyze the infiltration of immune subtypes in tumors (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e). A total of 10 tumor-infiltrating immune cells (TICs) showed significant associations with RISK gene expression (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), T cells CD4\u003csup\u003e+\u003c/sup\u003e memory resting, NK cells resting, and mast cells activated showed higher expression in high-risk group, while T cells CD8\u003csup\u003e+\u003c/sup\u003e, T cells follicular helper, Tregs, NK cells activated, Dendritic cells resting, mast cells resting, and neutrophils were higher expressed in low-risk group(Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe graphical results indicated a positive correlation between resting memory CD4\u003csup\u003e+\u003c/sup\u003e T cells, resting NK cells, activated macrophages, and three TICs with RISK expression. Conversely, CD8\u003csup\u003e+\u003c/sup\u003e T cells, follicular T cells, T regulatory cells, resting macrophages, mesenchymal stem cells, resting dendritic cells, neutrophils, and seven other TICs exhibited a negative correlation with RISK expression. Additionally, the majority of the 8 HUB genes were highly correlated with NK cells and CD4\u003csup\u003e+\u003c/sup\u003e memory T cells, suggesting that liquid-liquid phase separation (LLPS) may influence OSCC pathogenesis through these cell types. In summary, these findings suggest that RISK genes may impact the immune response in OSCC by affecting the involvement of immune cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe study employs multi-omics analysis to investigate the expression, heterogeneity, prognostic value, and immunological evaluation of LLPS-related genes in OSCC. Initially, cells were categorized into groups with high and low LLPS scores based on single-cell sequencing cluster analysis. The differentially expressed genes suggest a close association between LLPS heterogeneity and these genes in OSCC. WGCNA identified the blue and yellow modules as highly correlated with LLPS in OSCC, indicating a close relationship between genes in these modules and LLPS regulation in OSCC. The intersection of these genes with those identified in previous single-cell sequencing studies resulted in the identification of key LLPS hub genes in OSCC, including VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, GAS2L3, and STIL. A prediction model was constructed using COXBOOST regression, Lasso regression, and STEPCOX based on these eight genes. Through this predictive model, OSCC patients were categorized into high-risk and low-risk groups, with significantly poorer prognosis observed in the high-risk group. The high-risk and low-risk groups exhibited marked differences in immune cell infiltration, potentially influencing their divergent prognoses and guiding immunotherapy strategies.\u003c/p\u003e \u003cp\u003eResearch has demonstrated that these eight genes play a pivotal role in cancer development. VRK1 has been shown to phosphorylate various transcription factors, including p53 and proteins related to the DNA damage response pathway, exerting a tumor-promoting effect (18)\u003csup\u003e21\u003c/sup\u003e. PLK1 is known to have a critical function in the cell cycle process, particularly in the G2-M checkpoint, mitosis, and cytokinesis, indicating a close association with uncontrolled proliferation of cancer cells (19)\u003csup\u003e22\u003c/sup\u003e. POLE2 promotes the malignant phenotype of glioblastoma by facilitating FOXM1 stability mediated by AURKA (20)\u003csup\u003e23\u003c/sup\u003e. ESCO2 is crucial for promoting malignant progression of hypopharyngeal carcinoma through the STAT1 pathway (21)\u003csup\u003e24\u003c/sup\u003e. NEIL3 activates the BRAF/MEK/ERK/TWIST pathway-mediated EMT and treatment resistance, leading to the progression of hepatocellular carcinoma (22)\u003csup\u003e25\u003c/sup\u003e. CENPI is overexpressed in colorectal cancer (CRC), acting as an oncogene in regulating CRC cell migration, invasion, and epithelial-mesenchymal transition (EMT) (23)\u003csup\u003e26\u003c/sup\u003e. GAS2L3 is found to be highly expressed in various cancers, but its specific mechanism of action remains unclear (24\u0026ndash;26)\u003csup\u003e27\u0026ndash;29\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAdditionally, a multitude of LLPS-related biomarkers have been identified in various tumors. Wang et al. constructed a uterine endometrial cancer biomarker comprising four LLPS-related genes (27)\u003csup\u003e30\u003c/sup\u003e. Xie et al. identified PGAM1 as a crucial LLPS-related gene in breast cancer (28)\u003csup\u003e31\u003c/sup\u003e. Fang et al. utilized weighted co-expression network analysis and regression analysis to establish features in hepatocellular carcinoma consisting of six LLPS-related genes, with the aim of assessing patient prognosis and immune status (29)\u003csup\u003e32\u003c/sup\u003e. Sun et al. revealed the utility of LLPS-related genes for genotypic classification in bladder cancer (30)\u003csup\u003e33\u003c/sup\u003e. Presently, research indicates that OSCC is linked to the epithelial-to-mesenchymal transition (31)\u003csup\u003e34\u003c/sup\u003e. In comparison with these published studies, our research comprehensively integrates the biological context, with a specific focus on epithelial cells, offering valuable insights for OSCC diagnosis and treatment.\u003c/p\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThis study has several limitations. Firstly, it lacks experimental validation of the results. Secondly, it did not utilize its own clinical samples to confirm the findings. Thirdly, there was a lack of demographic data, such as age, gender, and ethnicity, which may introduce bias. Fourthly, the role of LLPS in specific gene mutation-associated OSCC was not investigated. These areas will be a focus of our future research.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, this study identified eight LLPS-associated genes and developed a predictive model for OSCC prognosis and immune status. However, further evaluation in clinical cohorts is necessary to assess the practicality and accuracy of this model, and adjustments will be made accordingly. These findings and the model developed herein hold significant potential for advancing OSCC diagnostics and therapeutic strategies. Future research will aim to refine and validate these outcomes, contributing to the enhancement of OSCC patient care.\u003c/p\u003e "},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analyzed during the present study can be found in the Gene Expression Omnibus datasets (GEO, https://www.ncbi.nlm.nih.gov/geo/, GSE10332, and GSE42743) and The Cancer Genome Atlas(TCGA, https://portal.gdc.cancer.gov/).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the National Natural Science Foundation of China (32160175), Guizhou Provincial Basic Research Program (Natural Science) (NO. ZK [2021]440), the PhD Fund of Scientific Research Foundation of School and Hospital of Stomatology, Zunyi Medical University (NO. KY2020-14).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDing Luo, Huan Li and Jie Jing have contributed equally to this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors and Affiliations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSchool and Hospital of Stomatology, Zunyi Medical University, Zunyi, Guizhou, China\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNo. 6, Xuefu West Road, Zunyi, 563000, Guizhou, China.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDing Luo，Huan Li，Jie Jing\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDing Luo, Huan Li and Jie Jing have contributed equally to this work DL .and JJ were responsible for designing the entire study. DL and HL performed data analysis and software operation. DL and HL prepared the figures and tables. DL and HL collected materials related to this research. DL and JJ wrote the manuscript. JJ was responsible for review and editing. All authors contributed to the article and approved the final version for submission.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of competing interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCapparuccia, L. \u0026amp; Tamagnone, L. Semaphorin signaling in cancer cells and in cells of the tumor microenvironment \u0026ndash; two sides of a coin. \u003cem\u003eJ. Cell Sci.\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e, 1723\u0026ndash;1736 (2009).\u003c/li\u003e\n\u003cli\u003eEpidemiology of oral cancer in Arab countries | Saudi Medical Journal. https://smj.org.sa/content/37/3/249.\u003c/li\u003e\n\u003cli\u003eTandon, P., Dadhich, A., Saluja, H., Bawane, S. \u0026amp; Sachdeva, S. The prevalence of squamous cell carcinoma in different sites of oral cavity at our Rural Health Care Centre in Loni, Maharashtra \u0026ndash; a retrospective 10-year study. \u003cem\u003eContemp. Oncol. Onkol.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 178\u0026ndash;183 (2017).\u003c/li\u003e\n\u003cli\u003eA, R. \u003cem\u003eet al.\u003c/em\u003e The Challenges of OSCC Diagnosis: Salivary Cytokines as Potential Biomarkers. \u003cem\u003eJ. Clin. Med.\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, (2020).\u003c/li\u003e\n\u003cli\u003eI, P. \u003cem\u003eet al.\u003c/em\u003e Oral and Oropharyngeal squamous cell carcinoma: prognostic and predictive parameters in the etiopathogenetic route. \u003cem\u003eExpert Rev. Anticancer Ther.\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, (2019).\u003c/li\u003e\n\u003cli\u003eDe Felice, F. \u003cem\u003eet al.\u003c/em\u003e Radiotherapy Controversies and Prospective in Head and Neck Cancer: A Literature-Based Critical Review. \u003cem\u003eNeoplasia N. Y. N\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 227\u0026ndash;232 (2018).\u003c/li\u003e\n\u003cli\u003eGharat, S. A., Momin, M. \u0026amp; Bhavsar, C. Oral Squamous Cell Carcinoma: Current Treatment Strategies and Nanotechnology-Based Approaches for Prevention and Therapy. \u003cem\u003eCrit. Rev. Ther. Drug Carrier Syst.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 363\u0026ndash;400 (2016).\u003c/li\u003e\n\u003cli\u003eBiomolecular condensates: Organizers of cellular biochemistry - PMC. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7434221/.\u003c/li\u003e\n\u003cli\u003eShin, Y. \u0026amp; Brangwynne, C. P. Liquid phase condensation in cell physiology and disease. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e357\u003c/strong\u003e, eaaf4382 (2017).\u003c/li\u003e\n\u003cli\u003eZhang, H. \u003cem\u003eet al.\u003c/em\u003e Liquid-liquid phase separation in biology: mechanisms, physiological functions and human diseases. \u003cem\u003eSci. China Life Sci.\u003c/em\u003e \u003cstrong\u003e63\u003c/strong\u003e, 953\u0026ndash;985 (2020).\u003c/li\u003e\n\u003cli\u003eMehta, S. \u0026amp; Zhang, J. Liquid-liquid phase separation drives cellular function and dysfunction in cancer. \u003cem\u003eNat. Rev. Cancer\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 239\u0026ndash;252 (2022).\u003c/li\u003e\n\u003cli\u003eMeng, F. \u003cem\u003eet al.\u003c/em\u003e Induced phase separation of mutant NF2 imprisons the cGAS-STING machinery to abrogate antitumor immunity. \u003cem\u003eMol. Cell\u003c/em\u003e \u003cstrong\u003e81\u003c/strong\u003e, 4147-4164.e7 (2021).\u003c/li\u003e\n\u003cli\u003eRubio, K., Dobersch, S. \u0026amp; Barreto, G. Functional interactions between scaffold proteins, noncoding RNAs, and genome loci induce liquid-liquid phase separation as organizing principle for 3-dimensional nuclear architecture: implications in cancer. \u003cem\u003eFASEB J. Off. Publ. Fed. Am. Soc. Exp. Biol.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 5814\u0026ndash;5822 (2019).\u003c/li\u003e\n\u003cli\u003eZhang, X. \u003cem\u003eet al.\u003c/em\u003e The proline-rich domain promotes Tau liquid-liquid phase separation in cells. \u003cem\u003eJ. Cell Biol.\u003c/em\u003e \u003cstrong\u003e219\u003c/strong\u003e, e202006054 (2020).\u003c/li\u003e\n\u003cli\u003eWei, M., Huang, X., Liao, L., Tian, Y. \u0026amp; Zheng, X. SENP1 Decreases RNF168 Phase Separation to Promote DNA Damage Repair and Drug Resistance in Colon Cancer. \u003cem\u003eCancer Res.\u003c/em\u003e \u003cstrong\u003e83\u003c/strong\u003e, 2908\u0026ndash;2923 (2023).\u003c/li\u003e\n\u003cli\u003eGaglia, G. \u003cem\u003eet al.\u003c/em\u003e HSF1 phase transition mediates stress adaptation and cell fate decisions. \u003cem\u003eNat. Cell Biol.\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 151\u0026ndash;158 (2020).\u003c/li\u003e\n\u003cli\u003eHuang, W. Y. C. \u003cem\u003eet al.\u003c/em\u003e A molecular assembly phase transition and kinetic proofreading modulate Ras activation by SOS. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e363\u003c/strong\u003e, 1098\u0026ndash;1103 (2019).\u003c/li\u003e\n\u003cli\u003eLangfelder, P. \u0026amp; Horvath, S. WGCNA: an R package for weighted correlation network analysis. \u003cem\u003eBMC Bioinformatics\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 559 (2008).\u003c/li\u003e\n\u003cli\u003eYu, G., Wang, L.-G., Han, Y. \u0026amp; He, Q.-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. \u003cem\u003eOmics J. Integr. Biol.\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 284\u0026ndash;287 (2012).\u003c/li\u003e\n\u003cli\u003eNewman, A. M. \u003cem\u003eet al.\u003c/em\u003e Robust enumeration of cell subsets from tissue expression profiles. \u003cem\u003eNat. Methods\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 453\u0026ndash;457 (2015).\u003c/li\u003e\n\u003cli\u003eCampillo-Marcos, I., Garc\u0026iacute;a-Gonz\u0026aacute;lez, R., Navarro-Carrasco, E. \u0026amp; Lazo, P. A. The human VRK1 chromatin kinase in cancer biology. \u003cem\u003eCancer Lett.\u003c/em\u003e \u003cstrong\u003e503\u003c/strong\u003e, 117\u0026ndash;128 (2021).\u003c/li\u003e\n\u003cli\u003eZitouni, S., Nabais, C., Jana, S. C., Guerrero, A. \u0026amp; Bettencourt-Dias, M. Polo-like kinases: structural variations lead to multiple functions. \u003cem\u003eNat. Rev. Mol. Cell Biol.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 433\u0026ndash;452 (2014).\u003c/li\u003e\n\u003cli\u003eZhang, P. \u003cem\u003eet al.\u003c/em\u003e POLE2 facilitates the malignant phenotypes of glioblastoma through promoting AURKA-mediated stabilization of FOXM1. \u003cem\u003eCell Death Dis.\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 61 (2022).\u003c/li\u003e\n\u003cli\u003eHu, J. \u003cem\u003eet al.\u003c/em\u003e ESCO2 promotes hypopharyngeal carcinoma progression in a STAT1-dependent manner. \u003cem\u003eBMC Cancer\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1114 (2023).\u003c/li\u003e\n\u003cli\u003eLai, H.-H. \u003cem\u003eet al.\u003c/em\u003e NEIL3 promotes hepatoma epithelial-mesenchymal transition by activating the BRAF/MEK/ERK/TWIST signaling pathway. \u003cem\u003eJ. Pathol.\u003c/em\u003e \u003cstrong\u003e258\u003c/strong\u003e, 339\u0026ndash;352 (2022).\u003c/li\u003e\n\u003cli\u003eDing, N., Li, R., Shi, W. \u0026amp; He, C. CENPI is overexpressed in colorectal cancer and regulates cell migration and invasion. \u003cem\u003eGene\u003c/em\u003e \u003cstrong\u003e674\u003c/strong\u003e, 80\u0026ndash;86 (2018).\u003c/li\u003e\n\u003cli\u003eZhou, Y. \u003cem\u003eet al.\u003c/em\u003e Elevated GAS2L3 Expression Correlates With Poor Prognosis in Patients With Glioma: A Study Based on Bioinformatics and Immunohistochemical Analysis. \u003cem\u003eFront. Genet.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 649270 (2021).\u003c/li\u003e\n\u003cli\u003eGao, X. \u003cem\u003eet al.\u003c/em\u003e Genetic expression and mutational profile analysis in different pathologic stages of hepatocellular carcinoma patients. \u003cem\u003eBMC Cancer\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 786 (2021).\u003c/li\u003e\n\u003cli\u003eLiu, B. \u003cem\u003eet al.\u003c/em\u003e RAB42 Promotes Glioma Pathogenesis via the VEGF Signaling Pathway. \u003cem\u003eFront. Oncol.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 657029 (2021).\u003c/li\u003e\n\u003cli\u003eWang, J., Meng, F. \u0026amp; Mao, F. Single cell sequencing analysis and transcriptome analysis constructed the liquid-liquid phase separation(LLPS)-related prognostic model for endometrial cancer. \u003cem\u003eFront. Oncol.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 1005472 (2022).\u003c/li\u003e\n\u003cli\u003eXie, J. \u003cem\u003eet al.\u003c/em\u003e Significance of liquid-liquid phase separation (LLPS)-related genes in breast cancer: a multi-omics analysis. \u003cem\u003eAging\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 5592\u0026ndash;5610 (2023).\u003c/li\u003e\n\u003cli\u003eFang, Z.-S. \u003cem\u003eet al.\u003c/em\u003e Liquid-Liquid Phase Separation-Related Genes Associated with Tumor Grade and Prognosis in Hepatocellular Carcinoma: A Bioinformatic Study. \u003cem\u003eInt. J. Gen. Med.\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 9671\u0026ndash;9679 (2021).\u003c/li\u003e\n\u003cli\u003eSun, L. \u003cem\u003eet al.\u003c/em\u003e Identification of molecular subtypes based on liquid\u0026ndash;liquid phase separation and cross-talk with immunological phenotype in bladder cancer. \u003cem\u003eFront. Immunol.\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1059568 (2022).\u003c/li\u003e\n\u003cli\u003eLing, Z., Cheng, B. \u0026amp; Tao, X. Epithelial-to-mesenchymal transition in oral squamous cell carcinoma: Challenges and opportunities. \u003cem\u003eInt. J. Cancer\u003c/em\u003e \u003cstrong\u003e148\u003c/strong\u003e, 1548\u0026ndash;1561 (2021).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"oral squamous cell carcinoma, liquid-liquid phase separation, single cell sequencing analysis, machine learning, immune microenvironment","lastPublishedDoi":"10.21203/rs.3.rs-4129536/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4129536/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eLiquid-liquid phase separation (LLPS) is implicated in tumorigenesis and progression, yet its role in oral squamous cell carcinoma (OSCC) remains unexplored. This study aims to identify LLPS-associated genes in OSCC and develop a prognostic assessment model.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eWe analyzed 334 OSCC and 32 normal samples from the TCGA-HNSC cohort. Inclusion criteria encompassed histologically verified primary OSCC, mRNA profiles, and pertinent clinical data, while samples with no survival status or survival time less than 30 days were excluded. The final cohort consisted of 297 OSCC patients with complete data on age, gender, TNM staging, and grading. We utilized single-cell sequencing data from GEO (GSE103322), with GSE42743 as the validation cohort. LLPS-related genes from DrLLPS were employed, and key genes were identified through weighted co-expression network and clustering analysis. Prognostic models were developed using Coxboost, Lasso regression, and Stepcox regression. Additionally, immune infiltration analysis was conducted to study the immune microenvironment of OSCC.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eThe study established a predictive model based on eight LLPS-related genes in OSCC (VRK1, PLK1, POLE2, ESCO2, NEIL3, CENPI, GAS2L3, STIL). OSCC patients were stratified into two groups: high-risk and low-risk, with the high-risk group exhibiting significantly poorer prognosis (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Furthermore, notable differences in the immune environment were also observed between the groups.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eThis study identified eight LLPS-associated genes critical for OSCC prognosis and immune status, leading to the development of a predictive model. This research holds significance for advancing OSCC diagnosis and treatment strategies.\u003c/p\u003e","manuscriptTitle":"Significance of Liquid-Liquid Phase Separation-related Genes in the Prognostic Assessment of Oral Squamous Cell Carcinoma: A Multi-omics Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-03-29 18:47:52","doi":"10.21203/rs.3.rs-4129536/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ca013f28-26d7-48c0-b3a6-d1406f1e06e0","owner":[],"postedDate":"March 29th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-05-28T14:59:28+00:00","versionOfRecord":[],"versionCreatedAt":"2024-03-29 18:47:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4129536","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4129536","identity":"rs-4129536","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}