| Objective: Osteosarcoma（OS）is the most common malignant bone tumor in children and adolescents.Osteosarcoma typically presents in distal femoral metaphysis,followed by the upper extremities,pelvis,and the like.Surgical resection and chemotherapy significantly prolong survival duration of patients with local osteosarcoma.Non-metastatic patients reach the 60–70% of the 5-years survival rate following diagnosis.However,osteosarcoma is characterized by a high risk of metastasis and recurrence.For metastatic or recurrent subjects,the 5-year survival rate was only 20%.Therefore,there is an urgent need to explore biomarkers to help clinicians predict survival outcomes and provide the basis for personalized medicine.Growing evidence suggests that transcription factors（TFs）play an important role in osteosarcoma.However,the molecular characterization of transcription factors in osteosarcoma remains poorly defined.Here,we develop a TFs-related prognostic signature in osteosarcoma.Methods: We collected m RNA expression data and corresponding clinical dataof OS samples from the Therapeutically Applicable Research to Generate Effective Treatments（TARGET）database and the GEO database（GSE39055 dataset）,respectively.Differential analysis between metastatic and non-metastatic samples was performed on TARGET database to identify differentially expressed TFs（DETFs）.Univariate and multivariate Cox regression analyses were applied to identify DETFs associated with overall survival,and we established a TFs-related prognostic signature（TRPS）.We externally validated in an independent cohort（GSE39055 dataset）.Kaplan-Meier analysis was used to compare overall survival between high-and low-risk groups.Receiver operating characteristic（ROC）curves and subgroup analyses were used to assess the predictive power of TRPS.Univariate and multivariate Cox analyses were applied to identify independent predictors of OS.Then,a prognostic nomogram was constructed by integrating risk score,age,sex,tumor site,and metastasis.The ESTIMATE algorithm was used to assess immune/stromal cell scores in the tumor microenvironment.Immune cell infiltration scores and immune-related pathway activity between highand low-risk groups were calculated using single-sample gene set enrichment analysis（ss GSEA）.Results: In the TARGET dataset,compared to non-metastatic primary samples（n=65）,we screened 24 DETFs in metastatic osteosarcoma samples（n=22）.Among them,23 TFs were down-regulated in metastatic sample tissues,and 1 TF was up-regulated in metastatic sample tissues.We established a prognostic risk signature based on two genes,ZNF597 and MESP1.Then,the OS samples were divided into high and low risk groups according to the best cut-off value（2.24）of the risk score calculated by the X-Tile software.The Kaplan-Meier survival curve showed that the prognosis of OS patients in the low-risk group was significantly better than that in the high-risk group.The ROC curve results showed that the AUC of 1,2,and 3 years were 0.870,0.868,and 0.777,respectively,indicating that our TRPS has high predictive performance.In the TARGET cohort,TRPS 1-,2-,and 3-year OS predicted AUC higher than age,sex,and site.We found that the risk scores were significantly higher in the metastatic OS sample than in the non-metastatic OS sample,suggesting that this signature was strongly associated with metastasis.Multivariate analysis demonstrated that this TRPS was an independent prognostic predictor of OS and was further confirmed in the GSE39055 dataset and multiple clinical subtypes.Using the ESTIMATE algorithm,we found that the stromal score was significantly negatively correlated with the risk score.And ss GSEA showed that the relative abundance of NK cells in the low-risk prognosis group was significantly higher than that in the high-risk prognosis group.Conclusions:1.A total of 24 differentially expressed transcription factors（DETFs）were screened in the TARGET dataset of metastatic samples（n=22）and non-metastatic samples（n=65）.2.We established a TRPS（MESP1 and ZNF597）in osteosarcoma,and based on the model,each osteosarcoma sample in the TARGET cohort was risk-scored and divided into high-and low-risk prognostic groups.This model had high diagnostic and prognostic efficacy in OS patients and was further validated in another independent cohort（GSE39055）.This feature may optimize the prognostic management of patients with osteosarcoma and help to personalize treatment.3.The risk score of metastatic OS samples was significantly higher than that of non-metastatic OS samples.4.TRPS is an independent prognostic predictor of OS.5.In the TARGET database,the overall survival rate of osteosarcoma patients in the low stromal score group was lower than that in the high stromal score group,and the stromal score of osteosarcoma samples was significantly negatively correlated with the risk score.6.The ss GSEA showed that the relative abundance of NK cells in the low-risk prognosis group was significantly higher than that in the high-risk prognosis group. |
