Integrated pan-cancer profiling highlights OSR2 as a prognostic indicator and immune-associated biomarker Authors: Lei Tang, Qiqi Chen, Chunmiao Han, Yuanyuan Wang, Jianbo Li, Yunsong Liu Published: 2026 ABSTRACT BackgroundOdd-skipped-related 2 (OSR2), encoded by the OSR2 gene, has been reported to function as a checkpoint associated with CD8⁺ T-cell exhaustion in the tumor microenvironment of solid malignancies, suggesting its potential as a therapeutic target to improve immunotherapeutic responses. Nevertheless, the molecular and clinical significance of OSR2 across diverse cancer types has not yet been systematically investigated, and its pan-cancer expression profile, prognostic implications, and associations with tumor immunity remain to be fully elucidated.MethodsIn this study, we integrated datasets from The Cancer Genome Atlas (TCGA), the Genotype-Tissue Expression (GTEx) portal, and the Human Protein Atlas to construct a systematic pan-cancer profile of OSR2. The prognostic value of OSR2 was comprehensively assessed using univariate Cox regression, survival analysis, and receiver operating characteristic (ROC) curve analysis. In addition, we performed an in-depth analysis of the relationships between OSR2 and multiple molecular and immunological features, including copy number variation (CNV), DNA methylation, tumor mutational burden (TMB), microsatellite instability (MSI), immune-related gene expression, immune cell infiltration, and drug sensitivity, with the aim of exploring its potential immunological associations with the tumor microenvironment.ResultsOSR2 expression was significantly upregulated or downregulated in the majority of tumor tissues relative to normal counterparts and exhibited distinct cancer-type–specific patterns across clinical stages. CNV alterations and aberrant DNA methylation were closely associated with abnormal OSR2 mRNA expression in multiple cancers. Prognostic analyses indicated that OSR2 expression was significantly associated with overall survival, disease-specific survival, disease-free interval, and progression-free interval across multiple cancer types, showing either risk-associated or protective associations in a tumor-context-dependent manner. Furthermore, OSR2 expression showed strong associations with immune cell infiltration, particularly T-cell subsets, and was significantly correlated with the expression of multiple immune checkpoint–related genes across diverse malignancies. OSR2 expression was also closely associated with TMB, MSI, and sensitivity to multiple anticancer agents.ConclusionTaken together, these findings suggest that OSR2 is associated with prognosis and immune-related features across multiple cancer types. OSR2 may be linked to features of the tumor immune microenvironment through its relationships with immune cell infiltration, immune checkpoint gene expression, and genomic instability, and thus may serve as a candidate biomarker for further investigation in cancer immunotherapy.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s12672-026-05360-y. Introduction Cancer remains one of the most significant global public health challenges [1, 2]. According to the most recent statistics released by the International Agency for Research on Cancer (IARC), nearly 20 million new cancer cases were diagnosed worldwide in 2022, with approximately 9.7 million cancer-related deaths, posing a substantial threat to human health and markedly reducing life expectancy [3]. To date, complete eradication of cancer remains elusive, and no universally effective therapeutic strategy has been achieved. In recent years, the advent of cancer immunotherapy has profoundly transformed therapeutic paradigms and has led to substantial clinical benefits across multiple malignancies [4, 5]. In particular, the successful clinical implementation of immune checkpoint inhibitors has driven the development and translation of antagonists targeting key regulatory molecules, including CTLA-4, PD-1, PD-L1, and LAG-3 [6–9]. Despite major advances in cancer immunotherapy, heterogeneous responses and immune resistance remain common [5, 8, 10]. One of the major contributors to therapeutic heterogeneity and resistance is the tumor microenvironment (TME), which plays a key role in immune evasion and strongly influences immunotherapeutic outcomes [11, 12]. Recent studies have shown that dysregulated cellular and molecular interactions within the TME promote immune evasion and treatment resistance, and that targeting these processes may enhance antitumor immunity [13, 14]. Accordingly, strategies to remodel the TME and restore immune cell function have emerged as promising approaches for cancer therapy [15]. Therefore, identifying genes that regulate the TME and antitumor immunity has become an important focus of cancer immunotherapy research. Recent studies suggest that the tumor immune microenvironment depends not only on immune cell abundance, but also on their functional states, spatial organization, and interactions with tumor-intrinsic signaling pathways. In particular, CD8⁺ T-cell exhaustion and immune checkpoint activation are central mechanisms of immune evasion and therapeutic resistance [16]. These findings support the search for novel regulators linking tumor-intrinsic programs to immune dysfunction in the TME [14]. The transcription factor odd-skipped related 2 (OSR2) has long been recognized for its roles in palatal development, synovial joint formation, and skeletal patterning [17–19]. Recent findings, however, have uncovered a previously unappreciated function of OSR2 in linking biomechanical signaling to the terminal exhaustion of tumor-reactive CD8⁺ T cells [20]. Loss-of-function studies demonstrate that OSR2 deficiency markedly attenuates exhaustion programs in tumor-specific CD8⁺ T cells, whereas forced OSR2 expression in solid tumor settings accelerates the acquisition of exhausted phenotypes [20]. At the mechanistic level, OSR2 facilitates CD8⁺ T-cell exhaustion through recruitment of histone deacetylase 3 (HDAC3), leading to extensive epigenetic remodeling that represses the transcription of cytotoxic effector molecules [20]. These findings suggest that OSR2 may be involved in mechanosensitive regulation linking tumor-associated mechanical stress to chromatin remodeling and the development of exhausted CD8⁺ T-cell states. This may help explain its pan-cancer associations with immune infiltration and immune checkpoint–related genes. Together, these data establish OSR2 as a biomechanically responsive checkpoint that amplifies CD8⁺ T-cell exhaustion, thereby highlighting it as a candidate for further investigation as a potential therapeutic target for reinvigorating antitumor immunity and improving cancer immunotherapy outcomes. However, current studies on OSR2 remain relatively limited, particularly with respect to systematic investigations based on multi-omics data that explore its potential roles across diverse human cancer types. Therefore, a comprehensive pan-cancer analytical strategy integrating multidimensional datasets is urgently needed to characterize the expression patterns and functional relevance of OSR2 in different malignancies, thereby advancing our understanding of its involvement in tumorigenesis and tumor immune regulation. In this study, we conducted a comprehensive pan-cancer analysis of OSR2 using multiple bioinformatics approaches across diverse analytical dimensions. By integrating data from several authoritative databases, including The Cancer Genome Atlas (TCGA), the Human Protein Atlas (HPA), and the Genotype-Tissue Expression (GTEx) project, we systematically characterized the mRNA expression patterns of OSR2 in tumor and normal tissues and examined its association with HDAC3 expression [21–23]. Furthermore, leveraging the Gene Set Cancer Analysis (GSCA) database, we investigated the potential relationships between OSR2 expression and genomic alterations as well as DNA methylation across 33 cancer types [24]. In addition, univariate Cox regression, survival analysis, receiver operating characteristic (ROC) curve analysis, immune cell infiltration analysis, immune-related gene correlation analysis, and drug sensitivity analysis were performed, along with an evaluation of the associations between OSR2 expression and tumor mutational burden (TMB) and microsatellite instability (MSI). Collectively, this study aimed to determine whether OSR2 is associated with tumor immune contexture, immune checkpoint activity, and CD8⁺ T-cell–related immune dysfunction across cancers, and, through integrated multi-omics and immunogenomic analyses, to establish a pan-cancer framework for understanding its context-dependent biological and immunological roles as well as its potential relevance to immunotherapeutic stratification. Materials and methods Pan-cancer analysis of OSR2 expression and its correlation with HDAC3 Bulk RNA-seq expression profiles for 33 cancer types from The Cancer Genome Atlas (TCGA), together with corresponding normal tissue data, were retrieved from the UCSC Xena platform and complemented with normal tissue expression data from the Genotype-Tissue Expression (GTEx) project and protein expression information from the Human Protein Atlas (HPA) [25]. Based on these datasets, we systematically evaluated the mRNA expression patterns of OSR2 across multiple tumor types and normal tissues, as well as its distribution among different human tissues. In addition, gene correlation analyses were conducted to quantify the association between OSR2 and HDAC3 expression across all 33 TCGA cancer cohorts. Furthermore, differential expression analyses were conducted to compare OSR2 expression levels among distinct clinical stages, aiming to explore its potential involvement in tumor progression. Integrated analysis of DNA methylation and copy number variation Gene Set Cancer Analysis (GSCA) is an integrated web-based platform for genomic, pharmacogenomic, and immunogenomic analyses across cancer types [24]. In this study, the Mutation module of GSCA was utilized to characterize OSR2 alterations and epigenetic regulation across 33 TCGA cancer cohorts, focusing on three main analyses: (1) the relationship between copy number variation (CNV) and OSR2 mRNA expression, (2) differential methylation of OSR2 between tumor and normal tissues, and (3) the association between OSR2 methylation levels and transcript abundance. For the “CNV and Expression” analysis, RSEM-normalized mRNA expression and CNV data from TCGA were matched by TCGA barcodes, and Spearman correlation was performed to assess the relationship between OSR2 copy-number alterations and transcript expression with FDR-adjusted P values. For the “Differential Methylation” and “Methylation and Expression” analyses, Illumina HumanMethylation450K Level 3 data and RSEM-normalized mRNA expression data were obtained from the TCGA database. Differential methylation between tumor and adjacent non-tumor tissues was evaluated using a t-test in a total of 14 cancer types—THCA, KIRP, BLCA, LIHC, HNSC, BRCA, LUAD, PRAD, ESCA, KICH, LUSC, KIRC, STAD, and COAD—each comprising more than 10 paired tumor and adjacent non-tumor samples. The resulting P values were adjusted using the false discovery rate (FDR) method. In the Methylation and Expression module, mRNA expression and methylation datasets were matched by TCGA barcode, followed by Spearman correlation analysis to quantify the association between gene mRNA expression and methylation levels, with FDR-adjusted P values. Prognostic evaluation using univariate Cox regression and survival analysis To evaluate the prognostic significance of OSR2 across 33 cancer types, univariate Cox proportional hazards regression models were applied to analyze associations with overall survival (OS), disease-specific survival (DSS), disease-free interval (DFI), and progression-free interval (PFI) [26]. A value of P < 0.05 was considered statistically significant. Furthermore, Kaplan–Meier survival curves were generated to assess the prognostic value of OSR2 within the TCGA cohort [27]. Patients were stratified into high- and low-expression groups based on the median OSR2 expression level. All statistical analyses were performed using the R computing environment, utilizing the “survival” (Version 3.8.3) and “survminer” (Version 0.5.1) packages. Receiver operating characteristic curve analysis Receiver operating characteristic (ROC) analysis was performed utilizing the “pROC” package (Version 1.19.0.1) to assess the discriminatory ability of OSR2 [28]. An area under the curve (AUC) threshold of > 0.7 was considered indicative of acceptable discriminatory performance. Immune infiltration estimation and correlation with immune-related genes We investigated the relationship between OSR2 expression and the immune landscape using CIBERSORT, MCP-counter, and ssGSEA methods to quantify immune cell infiltration [29–31]. The ESTIMATE algorithm (R package "estimate", Version 1.0.13) was utilized to calculate Stromal, Immune, and ESTIMATE scores for assessing tumor purity [32]. Additionally, we performed correlation analyses between OSR2 expression and several specific gene sets, comprising Mismatch Repair, immune checkpoints, Major Histocompatibility Complex molecules, chemokines, and chemokine receptors. Association of OSR2 expression with tumor mutational burden and microsatellite instability MSI scores were obtained from the “MSIsensor10k” dataset in the “BiocOncoTK” package (Version 1.24.0), and their association with OSR2 expression was evaluated using Spearman correlation. Somatic mutation data for all TCGA samples were downloaded using the “TCGAmutations” package (Version 0.4.0). Tumor mutational burden (TMB) was calculated with the tmb function in “maftools” (Version 2.24.0). The correlation between OSR2 expression and TMB was assessed using Spearman correlation. Evaluation of drug sensitivity associated with OSR2 expression To investigate the correlation between OSR2 expression and pan-cancer drug sensitivity, we utilized the “Drug” module. We integrated mRNA expression profiles with drug response data (IC50) from two pharmacogenomic databases: the Genomics of Drug Sensitivity in Cancer (GDSC) and the Cancer Therapeutics Response Portal (CTRP) [33, 34]. The analysis encompassed 265 small molecules across 860 cell lines from GDSC and 481 small molecules across 1001 cell lines from CTRP. Pearson correlation coefficients were calculated to assess the association between OSR2 expression and drug IC50 values, with P values adjusted using the False Discovery Rate. Statistical methods Differences in gene expression between tumor and normal tissues were evaluated using the Wilcoxon rank-sum test. Expression variations across clinical stages were compared by one-way ANOVA. The Spearman rank correlation test was used to evaluate the associations between OSR2 expression and immune infiltration scores, immune-related genes, MSI, and TMB. P < 0.05 was considered statistically significant. All statistical analyses were conducted using R software (Version 4.5.1). Results Expression profiles of OSR2 in human normal and tumor tissues Using expression data downloaded from the HPA and GTEx databases, we first analyzed the expression pattern of OSR2 across human normal tissues. The results showed that OSR2 exhibited relatively high expression levels in cervical tissues (Fig. 1A, B). Subsequently, based on transcriptomic data from TCGA covering 33 cancer types, a pan-cancer analysis was performed to compare OSR2 expression between tumor and corresponding normal tissues. As a result, statistically significant differences in OSR2 expression were observed in 18 cancer types, including BLCA, BRCA, CESC, CHOL, COAD, ESCA, GBM, HNSC, KICH, LIHC, LUAD, LUSC, PCPG, PRAD, READ, STAD, THCA, and UCEC (Fig. 1C). Furthermore, to improve the robustness of normal tissue comparisons, GTEx data were incorporated to expand the normal sample cohort. Under this integrative analysis framework, 20 cancer types exhibited significant differential expression of OSR2, namely ACC, BLCA, BRCA, CESC, COAD, DLBC, GBM, HNSC, KICH, LIHC, OV, PAAD, PRAD, READ, SKCM, STAD, TGCT, THCA, UCEC, and UCS (Fig. 1D).Fig. 1Pan-cancer expression landscape of OSR2 in tumor and normal tissues: A Tissue-specific distribution of OSR2 expression across normal human tissues based on data from the GTEx database. B Evaluation of OSR2 expression profiles in normal tissues from the GTEx database and protein-level data from the Human Protein Atlas (HPA). C Comparison of OSR2 expression levels between tumor tissues and corresponding normal tissues derived from the TCGA cohort. D Integrated analysis of OSR2 expression in tumor and normal tissues using combined TCGA and GTEx datasets. E Expression of OSR2 across different pathological stages in pan-cancer cohorts. Statistical significance is indicated as follows: ns not significant (P > 0.05); *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. ACC Adrenocortical Carcinoma, BLCA Bladder Urothelial Carcinoma, BRCA Breast Cancer, CESC Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma, CHOL Cholangiocarcinoma, COAD Colorectal Adenocarcinoma, DLBC Diffuse Large B-cell Lymphoma, ESCA Esophageal Carcinoma, GBM Glioblastoma Multiforme, HNSC Head and Neck Squamous Cell Carcinoma, KICH Kidney Chromophobe, KIRC Kidney Renal Clear Cell Carcinoma, KIRP Kidney Renal Papillary Cell Carcinoma, LAML Acute Myeloid Leukemia, LGG Low Grade Glioma, LIHC Liver Hepatocellular Carcinoma, LUAD Lung Adenocarcinoma, LUSC Lung Squamous Cell Carcinoma, MESO Malignant Mesothelioma, OV Ovarian Cancer, PAAD Pancreatic Adenocarcinoma, PCPG Pheochromocytoma and Paraganglioma, PRAD Prostate Adenocarcinoma, READ Rectal Adenocarcinoma, SARC Sarcoma, SKCM Cutaneous Melanoma, STAD Stomach Adenocarcinoma, TGCT Testicular Germ Cell Tumor, THCA Thyroid Cancer, THYM Thymoma, UCS Uterine Carcinosarcoma, UCEC Uterine Corpus Endometrial Carcinoma, UVM Uveal Melanoma Pan-cancer expression landscape of OSR2 in tumor and normal tissues: A Tissue-specific distribution of OSR2 expression across normal human tissues based on data from the GTEx database. B Evaluation of OSR2 expression profiles in normal tissues from the GTEx database and protein-level data from the Human Protein Atlas (HPA). C Comparison of OSR2 expression levels between tumor tissues and corresponding normal tissues derived from the TCGA cohort. D Integrated analysis of OSR2 expression in tumor and normal tissues using combined TCGA and GTEx datasets. E Expression of OSR2 across different pathological stages in pan-cancer cohorts. Statistical significance is indicated as follows: ns not significant (P > 0.05); *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; ****P ≤ 0.0001. ACC Adrenocortical Carcinoma, BLCA Bladder Urothelial Carcinoma, BRCA Breast Cancer, CESC Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma, CHOL Cholangiocarcinoma, COAD Colorectal Adenocarcinoma, DLBC Diffuse Large B-cell Lymphoma, ESCA Esophageal Carcinoma, GBM Glioblastoma Multiforme, HNSC Head and Neck Squamous Cell Carcinoma, KICH Kidney Chromophobe, KIRC Kidney Renal Clear Cell Carcinoma, KIRP Kidney Renal Papillary Cell Carcinoma, LAML Acute Myeloid Leukemia, LGG Low Grade Glioma, LIHC Liver Hepatocellular Carcinoma, LUAD Lung Adenocarcinoma, LUSC Lung Squamous Cell Carcinoma, MESO Malignant Mesothelioma, OV Ovarian Cancer, PAAD Pancreatic Adenocarcinoma, PCPG Pheochromocytoma and Paraganglioma, PRAD Prostate Adenocarcinoma, READ Rectal Adenocarcinoma, SARC Sarcoma, SKCM Cutaneous Melanoma, STAD Stomach Adenocarcinoma, TGCT Testicular Germ Cell Tumor, THCA Thyroid Cancer, THYM Thymoma, UCS Uterine Carcinosarcoma, UCEC Uterine Corpus Endometrial Carcinoma, UVM Uveal Melanoma Pan-cancer analysis of OSR2 expression across clinical stages and its correlation with HDAC3 Using the TCGAbiolinks package (Version 2.36.0), we retrieved clinical annotation data, including pathological stage, for patients enrolled in TCGA projects and conducted a systematic assessment of OSR2 expression patterns across distinct clinical stages. The results demonstrated that OSR2 expression differed significantly among clinical stages in six tumor types, namely BLCA, BRCA, KICH, KIRC, KIRP, and THCA (Fig. 1E), whereas no comparable stage-associated differences were detected in the remaining cancer types. These findings suggest that OSR2 expression is associated with pathological stage in selected malignancies, possibly reflecting tumor-type-specific biological or clinical heterogeneity. Given previous reports linking OSR2 to HDAC3-dependent T-cell exhaustion, we evaluated the correlation between OSR2 and HDAC3 expression across 33 TCGA cancer types. With P < 0.05 and Pearson R > 0.3 as thresholds, OSR2 expression was significantly and positively correlated with HDAC3 in LGG (R = 0.42), KICH (R = 0.44), ACC (R = 0.33), and UVM (R = 0.40) (Fig. 2). The moderate co-expression of OSR2 and HDAC3 in these malignancies suggests a possible relationship between the two genes in a tumor-context-dependent manner; however, whether this reflects functional cooperation requires further mechanistic investigation.Fig. 2OSR2 expression is correlated with HDAC3 expression across 33 TCGA cancer types OSR2 expression is correlated with HDAC3 expression across 33 TCGA cancer types Associations of OSR2 expression with copy number variation and DNA methylation Based on the GSCA platform, we systematically evaluated the associations of OSR2 expression levels with copy number variation and DNA methylation across multiple cancer types, and statistically significant results were further extracted for analysis. Spearman correlation analysis demonstrated a moderate positive correlation between OSR2 mRNA expression and CNV in several tumor types, with the strongest correlations observed in UVM (R = 0.52) and ACC (R = 0.52) (Fig. 3A). Consistent positive associations were also detected in ESCA (R = 0.35) and READ (R = 0.33) (Fig. 3A). These findings suggest that OSR2 expression is associated with CNV alterations in these cancers and may be consistent with a possible gene-dosage-related relationship. In parallel, we further assessed epigenetic alterations by analyzing methylation differences between tumor and normal tissues. All examined cancer types showed statistically significant differential methylation (Fig. 3B). Notably, COAD, HNSC, LIHC, and PRAD exhibited an overall decreasing trend in methylation levels in tumor tissues relative to normal controls (Fig. 3B). In contrast, BLCA, BRCA, ESCA, KIRC, KIRP, LUAD, PAAD, THCA, and UCEC displayed increased methylation levels in tumors, indicating a consistent hypermethylation pattern across these cancer types (Fig. 3B). Collectively, these results suggest that aberrant methylation is a widespread feature in tumorigenesis, with distinct cancer type–specific directions of methylation change. To determine whether these methylation changes were associated with OSR2 transcriptional activity, we further examined the correlation between DNA methylation and OSR2 mRNA expression. Methylation showed a moderate negative correlation with mRNA expression across multiple tumor types, including UVM (R = − 0.49), SARC (R = − 0.41), LUAD (R = − 0.40), LGG (R = − 0.39), UCEC (R = − 0.36), and LIHC (R = − 0.31) (Fig. 3C), which is consistent with methylation-associated transcriptional repression. In contrast, MESO exhibited a moderate positive correlation (R = 0.37) (Fig. 3C), highlighting a distinct, tumor-type-specific association between methylation and gene expression.Fig. 3Copy number variation and DNA methylation analyses of OSR2 across pan-cancer: A Association between OSR2 copy number variation (CNV) and mRNA expression across multiple cancer types based on the GSCA database. Bubble plots depict the correlation between OSR2 transcript levels and CNV status. Blue and red bubbles indicate negative and positive correlations, respectively, with color intensity reflecting correlation strength. Bubble size corresponds to the false discovery rate (FDR) significance, and black outlines denote statistically significant correlations (FDR ≤ 0.05). B Differential DNA methylation of OSR2 between tumor and corresponding normal tissues across various cancer types. Bubble plots illustrate methylation differences, with color gradients from blue to red representing relative decreases or increases in methylation levels in tumor samples. Bubble size reflects FDR-adjusted significance, and only comparisons meeting the significance threshold (FDR ≤ 0.05) are shown. C Correlation analysis between OSR2 DNA methylation levels and mRNA expression across pan-cancer cohorts using the GSCA platform. Bubble plots represent the direction and magnitude of correlations, where blue and red bubbles indicate negative and positive associations, respectively. Color intensity denotes correlation strength, bubble size reflects FDR significance, and black outlines highlight statistically significant correlations (FDR ≤ 0.05) Copy number variation and DNA methylation analyses of OSR2 across pan-cancer: A Association between OSR2 copy number variation (CNV) and mRNA expression across multiple cancer types based on the GSCA database. Bubble plots depict the correlation between OSR2 transcript levels and CNV status. Blue and red bubbles indicate negative and positive correlations, respectively, with color intensity reflecting correlation strength. Bubble size corresponds to the false discovery rate (FDR) significance, and black outlines denote statistically significant correlations (FDR ≤ 0.05). B Differential DNA methylation of OSR2 between tumor and corresponding normal tissues across various cancer types. Bubble plots illustrate methylation differences, with color gradients from blue to red representing relative decreases or increases in methylation levels in tumor samples. Bubble size reflects FDR-adjusted significance, and only comparisons meeting the significance threshold (FDR ≤ 0.05) are shown. C Correlation analysis between OSR2 DNA methylation levels and mRNA expression across pan-cancer cohorts using the GSCA platform. Bubble plots represent the direction and magnitude of correlations, where blue and red bubbles indicate negative and positive associations, respectively. Color intensity denotes correlation strength, bubble size reflects FDR significance, and black outlines highlight statistically significant correlations (FDR ≤ 0.05) Prognostic significance of OSR2 expression in pan-cancer Univariate Cox regression analyses were performed to evaluate the associations between OSR2 mRNA expression and survival outcomes, including OS, DSS, DFI, and PFI across pan-cancer cohorts. The results demonstrated that high OSR2 expression was significantly correlated with poorer OS in several cancer types, including BLCA (HR = 1.169, P = 0.004), KIRC (HR = 1.239, P = 0.012), KIRP (HR = 1.433, P = 0.008), LGG (HR = 1.344, P < 0.001), THCA (HR = 1.359, P = 0.030), and UVM (HR = 3.429, P = 0.001) (Fig. 4A). In contrast, elevated OSR2 expression acted as a protective factor for OS in CESC (HR = 0.745, P = 0.011), HNSC (HR = 0.795, P = 0.002), and LUSC (HR = 0.873, P = 0.035) (Fig. 4A). These findings indicate that the prognostic association of OSR2 is cancer-type dependent. Consistent with these findings, OSR2 acted as a risk factor for DFI in KICH and KIRP (Fig. S1), suggesting that elevated OSR2 expression was associated with an increased likelihood of recurrence. Regarding DSS, high OSR2 expression was significantly correlated with poorer outcomes in BLCA, KIRC, KIRP, LGG, and UVM (Fig. S2), whereas it served as a protective factor in DLBC and CESC. Similarly, for PFI, increased OSR2 expression predicted unfavorable prognosis in BLCA, KICH, KIRP, LGG, PAAD, and UVM, while a protective association was observed in CESC (Fig. S3). Furthermore, Kaplan–Meier survival analysis revealed a context-dependent prognostic role of OSR2 expression: elevated OSR2 levels were associated with significantly prolonged survival in CESC, HNSC, and LUSC, whereas high OSR2 expression predicted poorer outcomes in BLCA, LGG, MESO, and UVM (Fig. 4B–H). ROC analysis demonstrated that OSR2 exhibited strong discriminatory performance across five cancer types (AUC > 0.7), including BLCA (AUC = 0.705, 95% CI 0.597–0.792), KICH (AUC = 0.880, 95% CI 0.794–0.938), PRAD (AUC = 0.750, 95% CI 0.669–0.824), THCA (AUC = 0.799, 95% CI 0.744–0.852), and UCEC (AUC = 0.972, 95% CI 0.919–0.993) (Fig. 4I–M). Since these prognostic analyses were based on univariate Cox regression and Kaplan–Meier survival analysis, the observed associations should be interpreted cautiously and not as evidence of independent prognostic value.Fig. 4Prognostic value of OSR2 across pan-cancer cohorts. A Univariate Cox proportional hazards regression analysis assessing the association between OSR2 expression and overall survival (OS) across 33 TCGA cancer types. Cancer types with P < 0.05 and hazard ratio (HR) > 1 are highlighted in red. B–H Kaplan–Meier survival analyses evaluating the impact of OSR2 expression on overall survival in BLCA, CESC, HNSC, LGG, LUSC, MESO, and UVM. I–M Receiver operating characteristic (ROC) curve analyses assessing the predictive performance of OSR2 in BLCA, KICH, PRAD, THCA, and UCEC Prognostic value of OSR2 across pan-cancer cohorts. A Univariate Cox proportional hazards regression analysis assessing the association between OSR2 expression and overall survival (OS) across 33 TCGA cancer types. Cancer types with P < 0.05 and hazard ratio (HR) > 1 are highlighted in red. B–H Kaplan–Meier survival analyses evaluating the impact of OSR2 expression on overall survival in BLCA, CESC, HNSC, LGG, LUSC, MESO, and UVM. I–M Receiver operating characteristic (ROC) curve analyses assessing the predictive performance of OSR2 in BLCA, KICH, PRAD, THCA, and UCEC Associations between OSR2 expression and immune cell infiltration and immune-related genes across pan-cancer Based on the CIBERSORT analysis, correlations with coefficients > 0.3 and P < 0.05 suggested that OSR2 expression was associated with immune cell infiltration across multiple cancer types. Overall, OSR2 expression showed a significant positive correlation with CD8⁺ T-cell infiltration in KICH and UVM (Fig. 5A), whereas negative associations were observed with specific CD4⁺ T-cell subsets, including naive and memory T-cell subsets, in certain cancers, suggesting that OSR2 expression may be associated with T-cell-related immune states or transcriptomic features. In addition, OSR2 was associated with macrophage polarization, exhibiting predominantly positive correlations with M1 macrophages and bidirectional associations with M2 macrophages across different tumor types (Fig. 5A). Notably, in UVM, OSR2 was concurrently correlated with multiple immune cell populations, including CD8⁺ T cells, CD4⁺ memory T cells, and follicular helper T cells, implying that OSR2 may be associated with a distinct immune microenvironmental profile in this cancer type (Fig. 5A). Using the MCP-counter algorithm, we found that the only negative association with a correlation coefficient greater than 0.3 and P < 0.05 was the correlation between OSR2 and CD8⁺ T cells in THYM (Fig. 5B). No other negative correlations exhibited a correlation strength greater than 0.3 with a P < 0.05. In THCA, fibroblasts demonstrated the strongest positive correlation with OSR2 expression. Moreover, in KICH, PAAD, UVM, CHOL, and PCPG (Fig. 5B), the correlations between OSR2 and CD8⁺ T cells were all greater than 0.3, suggesting that OSR2 expression is preferentially associated with T-cell-enriched immune contexts in selected tumor types. Consistent with the ssGSEA algorithm, all CD8⁺ T cells subsets exhibited positive correlations across multiple cancer types (Fig. 5C). Collectively, the three algorithms showed a reproducible association between OSR2 and T-cell-related infiltration signals, particularly CD8⁺ T-cell signatures, whereas associations with other immune populations were more method-dependent. Given the central involvement of immune checkpoint–related genes in immune infiltration and immunotherapeutic responses, we systematically examined the expression correlations between OSR2 and immune checkpoint–related as well as other immune-associated genes across multiple human cancers. Notably, OSR2 expression showed significant correlations with the majority of immune checkpoint–related genes in BLCA, CHOL, COAD, KICH, LAML, LGG, PAAD, PCPG, PRAD, TGCT, THCA, THYM, and UVM (P < 0.001) (Fig. 5D). The ESTIMATE algorithm was applied to systematically assess the correlations between OSR2 expression and stromal score, immune score, as well as the overall ESTIMATE score, thereby providing insights into the relationship between OSR2 and the tumor microenvironment (Fig. 5E). Notably, the immune associations of OSR2 were not uniformly distributed across all leukocyte populations but were enriched in T-cell-related immune contexts in several tumor types. When interpreted together with the reported role of OSR2 as a biomechanical checkpoint that promotes CD8⁺ T-cell exhaustion, these findings raise the possibility that OSR2-associated T-cell infiltration may reflect the presence of dysfunctional or exhausted T cells, although this cannot be determined from the current data. Likewise, the broad correlations between OSR2 and immune checkpoint-related genes suggest that OSR2 expression is correlated with transcriptional features associated with immunosuppressive states. Although these correlative observations do not establish direct causality, they are consistent with the possibility that OSR2 expression is linked to T-cell dysfunction, checkpoint activity, and context-dependent changes in immune cell composition within the tumor microenvironment.Fig. 5Association between OSR2 expression and the immune landscape across pan-cancer. A Correlation analysis between OSR2 mRNA expression and immune cell infiltration across multiple cancer types estimated using the CIBERSORT algorithm. Statistical significance is indicated as *P < 0.05 and **P < 0.01. B, C Evaluation of the relationship between OSR2 expression and the abundance of infiltrating immune cell populations across various cancers using the MCP-counter (B) and ssGSEA (C) methods. D Spearman correlation analysis assessing the associations between OSR2 expression and immune-related genes, including genes involved in major histocompatibility complex (MHC), mismatch repair (MMR), immune checkpoint regulation, chemokines, and chemokine receptors. Statistical significance is indicated as *P < 0.05 and **P < 0.01. E Correlation analysis between OSR2 expression and stromal score, immune score, and ESTIMATE score calculated using the ESTIMATE algorithm. F Radar plots illustrating the correlations between OSR2 expression and tumor mutational burden (TMB) as well as microsatellite instability (MSI) across pan-cancer cohorts Association between OSR2 expression and the immune landscape across pan-cancer. A Correlation analysis between OSR2 mRNA expression and immune cell infiltration across multiple cancer types estimated using the CIBERSORT algorithm. Statistical significance is indicated as *P < 0.05 and **P < 0.01. B, C Evaluation of the relationship between OSR2 expression and the abundance of infiltrating immune cell populations across various cancers using the MCP-counter (B) and ssGSEA (C) methods. D Spearman correlation analysis assessing the associations between OSR2 expression and immune-related genes, including genes involved in major histocompatibility complex (MHC), mismatch repair (MMR), immune checkpoint regulation, chemokines, and chemokine receptors. Statistical significance is indicated as *P < 0.05 and **P < 0.01. E Correlation analysis between OSR2 expression and stromal score, immune score, and ESTIMATE score calculated using the ESTIMATE algorithm. F Radar plots illustrating the correlations between OSR2 expression and tumor mutational burden (TMB) as well as microsatellite instability (MSI) across pan-cancer cohorts Correlation of OSR2 expression with tumor mutation burden and microsatellite instability in pan-cancer Based on pan-cancer correlation analyses with a statistical significance threshold of P < 0.05, we systematically investigated the associations between OSR2 expression and two key indicators of genomic instability, namely tumor mutation burden (TMB) and microsatellite instability (MSI). The results revealed marked heterogeneity in the relationship between OSR2 expression and TMB across different cancer types. Specifically, OSR2 expression was significantly negatively correlated with TMB in PRAD, SKCM, THCA, UCEC, and UVM (Fig. 5F), suggesting that elevated OSR2 expression may be associated with a reduced mutational load in these tumors. In contrast, significant positive correlations between OSR2 expression and TMB were observed in BRCA, COAD, LGG, LIHC, LUAD, LUSC, MESO, STAD, and THYM (Fig. 5F). Among these, LGG exhibited the strongest association (R = 0.45), suggesting a possible association between OSR2 expression and genomic instability-related features in a cancer-type-dependent manner. Further evaluation of the relationship between OSR2 expression and MSI demonstrated that OSR2 was positively correlated with MSI in COAD, GBM, KIRP, LGG, LUAD, LUSC, and UVM, with the highest correlation observed in GBM (R = 0.29) (Fig. 5F). These findings suggest that increased OSR2 expression may be associated with MSI-related molecular features in selected tumor types. Conversely, significant negative correlations between OSR2 expression and MSI were identified in BLCA, BRCA, LAML, PCPG, PRAD, and UCEC (Fig. 5F), indicating that OSR2 expression may reflect distinct genomic contexts in different tumor types. Overall, OSR2 expression showed cancer-specific associations with both TMB and MSI, supporting further investigation into its relevance to genomic instability-related features. Associations between OSR2 expression and drug sensitivity Drug sensitivity analyses were conducted using the GSCA platform. Based on analyses of the CTRP database, most drugs displayed a positive correlation with OSR2 expression, whereas only a small subset of EGFR-related inhibitors showed a significant negative correlation with OSR2 expression, including PD-153035, Erlotinib, and Afatinib (Fig. 6A). These findings suggest that OSR2 expression may be associated with differential responses to specific EGFR-targeted therapies, implying a potential role in mechanisms underlying targeted therapy resistance. Based on the GDSC database, we analyzed the correlation between OSR2 expression levels and sensitivity to anticancer drugs, and ranked the top 30 drugs according to their correlation coefficients. The results demonstrated that OSR2 expression was significantly associated with the sensitivity to multiple targeted agents. Specifically, OSR2 expression showed a strong positive correlation with the sensitivity to Dabrafenib, PLX4720, CHIR-99021, CEP-701, UNC0638, and FK866 (Fig. 6B), suggesting that high OSR2 expression may be linked to enhanced responsiveness to these drugs. Notably, several drugs exhibited a negative correlation with OSR2 expression, including CGP-082996, AKT inhibitor VIII, GSK1904529A, Afatinib, CCT007093, and Gefitinib (Fig. 6B), indicating that elevated OSR2 expression may be associated with resistance to these agents. Overall, OSR2 expression was associated with differential sensitivity to multiple anticancer agents, warranting further investigation into its potential relevance to therapeutic response.Fig. 6Association between OSR2 expression and drug sensitivity across pan-cancer. A Correlation analysis between OSR2 expression and sensitivity to anticancer agents based on data from the Cancer Therapeutics Response Portal (CTRP) across pan-cancer cohorts. B Correlation analysis between OSR2 expression and sensitivity to anticancer agents derived from the Genomics of Drug Sensitivity in Cancer (GDSC) database across pan-cancer cohorts Association between OSR2 expression and drug sensitivity across pan-cancer. A Correlation analysis between OSR2 expression and sensitivity to anticancer agents based on data from the Cancer Therapeutics Response Portal (CTRP) across pan-cancer cohorts. B Correlation analysis between OSR2 expression and sensitivity to anticancer agents derived from the Genomics of Drug Sensitivity in Cancer (GDSC) database across pan-cancer cohorts Discussion OSR2 is a protein-coding gene and a paralog of OSR1 [35]. OSR2 functions as a sequence-specific DNA-binding transcription factor, interacting with regulatory regions associated with RNA polymerase II–mediated transcription. Dysregulation of OSR2 has been implicated in the pathogenesis of several developmental and reproductive disorders, including deficient palate growth [36]. Notably, recent studies have identified OSR2 as a mechanobiological checkpoint that promotes CD8⁺ T-cell exhaustion through the recruitment of HDAC3 [20]. This recently described function suggests that OSR2 may be relevant to cancer immunotherapy. In this study, OSR2 was systematically analyzed using integrated bioinformatics approaches based on multiple publicly available databases and analytical platforms. Our analyses revealed statistically significant differences in OSR2 expression between tumor and normal tissues across more than half of the cancer types examined. In addition, OSR2 expression was significantly correlated with clinicopathological stage in six cancers: BLCA, BRCA, KICH, KIRC, KIRP, and THCA. Additionally, OSR2 expression exhibited a significant positive correlation with HDAC3 expression in LGG, KICH, ACC, and UVM. Based on the GSCA platform, we evaluated the associations between OSR2 expression, CNV, and DNA methylation across multiple cancers. OSR2 expression was moderately and positively correlated with CNV in several tumor types, notably UVM and ACC, suggesting a gene dosage-dependent regulatory effect. In contrast, DNA methylation exhibited widespread, cancer type-specific alterations and was generally inversely correlated with OSR2 expression, consistent with epigenetic repression, with MESO representing an exception. In this study, we systematically evaluated the associations between OSR2 expression and multiple survival endpoints across pan-cancer cohorts using univariate Cox regression and Kaplan–Meier survival analyses. Overall, the prognostic impact of OSR2 exhibited marked cancer-type dependency. Specifically, elevated OSR2 expression was consistently associated with unfavorable outcomes in terms of OS, DSS, and PFI in several cancer types, including BLCA, KIRP, LGG, and UVM, suggesting that OSR2 expression may be associated with clinicopathological or biological features in these tumors. In contrast, higher OSR2 expression showed a protective association in CESC, potentially reflecting differences in tumor differentiation status or the tumor immune microenvironment. Furthermore, Kaplan–Meier survival analysis highlighted the context-dependent prognostic role of OSR2 expression. High OSR2 levels were associated with significantly prolonged survival in CESC, HNSC, and LUSC, whereas elevated OSR2 expression predicted poorer survival outcomes in BLCA, LGG, MESO, and UVM. In addition, receiver operating characteristic analysis demonstrated that OSR2 exhibited relatively strong discriminatory performance in several cancer types, with particularly high AUC values in UCEC and KICH, suggesting potential discriminative value in specific tumor contexts, which requires further validation. Nevertheless, the current findings were derived from univariate models and did not account for potential clinical confounders such as tumor stage, therapeutic regimens, or molecular subtypes. Across pan-cancer analyses, OSR2 expression was broadly associated with the tumor immune microenvironment, particularly with T-cell-related immune features. The concordant signals observed across CIBERSORT, MCP-counter, and ssGSEA suggest that the relationship between OSR2 and immune infiltration is unlikely to be attributable to a single deconvolution method. CIBERSORT, MCP-counter, and ssGSEA consistently indicated that OSR2 expression was associated with T-cell-related infiltration signals (particularly CD8⁺ T cells in KICH and UVM), CD4⁺ T-cell subset shifts, and macrophage polarization (generally positive with M1 and context-dependent with M2), with UVM showing multi-immune-cell co-associations. Importantly, the preferential associations with CD8⁺ T cells and selected CD4⁺ T-cell subsets indicate that OSR2 may be associated not only with the abundance of infiltrating lymphocytes but also with transcriptomic features related to their differentiation states and functional polarization within the tumor microenvironment. OSR2 also exhibited widespread correlations with immune checkpoint and immune-related genes in several tumor types and was linked to stromal/immune components by ESTIMATE. Moreover, OSR2 showed cancer-specific relationships with genomic instability, displaying both positive and negative correlations with TMB and MSI, suggesting context-dependent roles in mutational burden and mismatch repair deficiency. Finally, drug-response analyses (GDSC/GSCA) revealed that OSR2 expression was significantly associated with responses to multiple targeted agents, including several EGFR inhibitors, suggesting a potential role for OSR2 in therapy response and resistance mechanisms. The context-dependent prognostic role of OSR2 across cancer types likely reflects substantial biological heterogeneity in both tumor-intrinsic programs and immune ecosystem composition. In some tumors, elevated OSR2 may coincide not only with immunosuppressive remodeling, checkpoint activation, and dysfunctional CD8⁺ T-cell states, but also with invasion-associated matrix remodeling and epithelial–mesenchymal plasticity, thereby associating with poor prognosis and greater metastatic aggressiveness. In contrast, in other tumor types, OSR2 expression may be linked to lineage-specific differentiation programs, distinct stromal contexts, or compensatory immune activation, resulting in an apparently protective association. These findings suggest that the biological meaning of OSR2 expression cannot be interpreted in a uniform manner across cancers and should instead be understood within the tissue-specific and microenvironment-specific context. A plausible mechanistic framework may explain the link between OSR2 and immune cell infiltration across cancers. Prior evidence suggests that OSR2 promotes CD8⁺ T-cell exhaustion through HDAC3-dependent epigenetic remodeling, thereby suppressing cytotoxic programs and driving T cells toward dysfunctional states. Accordingly, tumors with high OSR2 expression may exhibit preserved or increased T-cell infiltration; however, whether these infiltrating cells are functionally exhausted cannot be determined from the current correlative analyses alone. This interpretation is supported indirectly by the observed associations with immune checkpoint and immune-related signatures, but requires further experimental validation. In addition, given the reported link between OSR2 and biomechanical signaling, one possible explanation is that OSR2 upregulation may be associated with stromal remodeling and mechanically stressed tumor niches, which in turn could influence cytokine gradients, immune cell trafficking, and checkpoint activity. Consistently, the significant associations observed here between OSR2 and chemokines, chemokine receptors, immune checkpoint genes, and ESTIMATE-derived stromal and immune scores support a role for OSR2 in the coordinated regulation of both immune recruitment and immune suppression. In addition, OSR2 may also be linked to metastasis-associated structural remodeling of the tumor microenvironment. Invadopodia are actin-rich invasive structures that promote extracellular matrix degradation and local invasion, whereas pre-metastatic niche formation depends on stromal–immune crosstalk, vascular remodeling, and extracellular matrix reprogramming [37, 38]. Therefore, given the reported biomechanical role of OSR2 and its association with stromal and immune features in our analysis, OSR2 upregulation may be associated with features of a metastasis-permissive microenvironment, although this hypothesis requires further experimental validation. Together, these non-mutually exclusive factors may help explain why OSR2 shows preferential associations with T-cell-related immune features across multiple cancer types. The discovery and clinical success of immunotherapy have had a profound impact on oncology and human health [39, 40]. Effector T cells are widely recognized as central mediators of antitumor immunity, and the remarkable therapeutic efficacy of immune checkpoint inhibitors targeting cytotoxic T-lymphocyte–associated antigen 4 (CTLA4) or the programmed cell death 1 (PD-1) pathway has enabled an increasing number of patients with diverse malignancies to benefit from these treatments [41–43]. Therefore, investigating the relationships between OSR2 and key immune checkpoint molecules as well as broader immune-related genes may provide valuable insights into its potential involvement in immunotherapy and its utility as a predictive biomarker for immune-based treatments. Mismatch repair (MMR) system–related genes (such as MLH1, MSH2, MSH6, PMS2, and EPCAM) play a pivotal role in maintaining genomic stability [44]. Defects in this pathway can lead to microsatellite instability and increased tumor mutation burden, thereby promoting neoantigen generation and enhancing tumor immunogenicity; consequently, MMR deficiency is widely regarded as an important predictive factor for response to immune checkpoint blockade [45]. Immune checkpoint–related genes (including PDCD1, CD274, CTLA4, LAG3, and TIGIT) regulate the intensity of T-cell activation and help maintain immune homeostasis [46, 47]. Tumors can exploit these pathways to induce T-cell exhaustion and achieve immune evasion, and thus their expression levels are commonly used to characterize immunosuppressive tumor microenvironments and predict potential sensitivity to immunotherapy [48]. Notably, OSR2 expression shows significant correlations with multiple immune checkpoint genes in several tumor types, including UVM, suggesting that OSR2 may serve as an important risk-associated factor in these malignancies. In addition, chemokines and their receptors constitute a core regulatory network governing immune cell recruitment and localization, thereby determining the extent and spatial distribution of immune cell infiltration (e.g., CD8⁺ T cells) within tumors and shaping “hot” versus “cold” immune phenotypes [47]. Taken together, the close associations between OSR2 and gene sets involved in tumor immunogenicity, antigen presentation efficiency, immune suppression, and immune cell infiltration provide supportive evidence for its potential role as an immunological biomarker and a candidate indicator for immunotherapy-related stratification. From a translational perspective, OSR2 may have potential clinical utility in patient stratification and may warrant further evaluation as a therapeutic target. The observed associations between OSR2, immune checkpoint genes, CD8⁺ T-cell infiltration, and genomic instability metrics suggest that OSR2 expression may be associated with tumors exhibiting features of an exhausted immune phenotype. OSR2 may have potential utility in patient stratification when integrated with TMB, MSI, and other tumor-specific immune features, although further validation is required. In addition, if its mechanistic role in promoting T-cell dysfunction is further validated, targeting the OSR2-associated regulatory axis may represent a potential strategy, subject to further experimental validation. These possibilities warrant prospective clinical validation and tumor-specific mechanistic studies. Recent experimental evidence has indicated that loss of OSR2 alleviates exhaustion in tumor-reactive CD8⁺ T cells, whereas enforced OSR2 expression aggravates dysfunctional T-cell states in solid tumor models [20]. Therefore, a causal role of OSR2 in regulating CD8⁺ T-cell exhaustion has been suggested in experimental settings. In contrast, the present study does not directly test causality, but extends these observations at the pan-cancer level by showing that OSR2 expression is systematically associated with T-cell infiltration patterns, immune checkpoint–related genes, and multiple immune microenvironment features across human malignancies. Taken together, our findings support the hypothesis that OSR2 may be involved in the regulation of tumor immune dysfunction, while highlighting the need for tumor-specific mechanistic validation using in vitro and in vivo models. Therefore, further validation using multi-source datasets and experimental approaches is required to substantiate the proposed biological functions and molecular mechanisms. In summary, our pan-cancer analyses nominate OSR2 as a candidate prognostic and immune-associated biomarker with potential relevance to cancer immunotherapy. These findings provide a rationale for future mechanistic and translational studies to determine whether OSR2 can be clinically exploited as a biomarker or therapeutic target in selected tumor contexts. Limitations Several limitations of this study merit consideration. Due to its retrospective design and reliance on publicly available datasets, the study is potentially susceptible to selection bias and variability in data quality. Moreover, because the analyses were entirely computational and predominantly correlative, without experimental or prospective clinical validation, causal relationships cannot be conclusively established. The exclusive use of bulk transcriptomic data, in the absence of single-cell or spatial transcriptomic profiling, further constrains the resolution of the cellular origin of OSR2 expression and the characterization of functional heterogeneity within the tumor microenvironment. In addition, disparities in sample size and cohort composition across cancer types may limit the robustness and comparability of the pan-cancer analyses. Importantly, the present study does not provide mechanistic evidence linking OSR2 to CD8⁺ T-cell exhaustion. Accordingly, future investigations incorporating single-cell, spatial, and functional perturbation approaches, along with rigorous prospective validation, will be essential to clarify the biological role and translational potential of OSR2. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4. Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4.