STSs is a rare group of malignancies with 50 histological subtypes and performs differing in behavior, biology, and sensitivity to treatment . However, therapies for each subtype remains similar in situ STSs that surgical resection is the main method and supplemented with radiotherapy . The application of appropriate biomarkers is critical in tumor biology for prediction or risk stratification. For example, genotyping methods based on genomics have been widely used to classify tumors, thus conducting the clinical trials. Recent studies have confirmed that DNA methylation provide insights into various tumor early diagnosis, molecular classification, and precise treatment [34, 14, 30]. Meanwhile, aberrant DNA methylation has been regarded as one of the hallmarks of cancer tissues [35, 36], alterations in DNA methylation also play a virtual role in the progression and development of STSs [20, 21]. In addition, S. Peter Wu et all had confirmed that methylation-based classifier could be used to provide diagnostic assistance in bone sarcoma . Therefore, we carried out this discovery to indicate the potential application of DNA methylation in STSs epigenomes classification. The TCGA database is a publicly available resource that contains more than 30 large cohorts of human tumors with a comprehensive multidimensional analysis , these large sample sizes are absolutely the basis for us to provide an in-depth understanding of the etiology of STSs. In this study, the whole genome DNA methylation sites corresponding to 269 STSs samples were also obtained from TCGA database and methylation sites in the gene promoter regions were first applied to select the prognosis associated CpG sites. Four specific prognosis subgroups, classified by 1,445 intersected independent prognostic CpG sites, were developed to present a molecular stratification for individual tumors, which has the significance of making therapeutic decisions and exploring the biological mechanisms involved in the progression of RCC.
STSs is not only a heterogeneous tumor, but also has been found presenting in almost every part of the human body, and patients with different histological types are significantly distinct in clinical outcomes. Thus, analysis combined with the comprehensive clinical characteristics, such as age, gender, tissue or organ of origin, and histological, might effectively improve the accuracy of prognosis in STSs patients . In this study, the distribution of the disease-specific OS in these 4 distinct prognostic subtypes of STSs was seemed to be predicted, as well as original sites, histological classification, age, and gender distribution. As the results suggested, the survival curves of these 4 specific subgroups were distinct from each other. Meanwhile, our classification scheme also provided an accurate diagnosis for individual tumors. For example, while patients were assigned to Cluster 1, they were found to have high chances to be diagnosed as leiomyosarcoma primary from retroperitoneum and connected with better prognosis, and these results might prompt clinicians to re-evaluate the treatment for STSs patients. Conclusion, our classifications of 4 subtypes based on the DNA methylation sites can classify STSs more accurately and guide clinicians in terms of clinical diagnosing, treating strategies and prognostic judgment of different STSs patients.
Previous researches have reported that CpG island methylation was shown to promote carcinogenesis by disrupting the function of tumor suppressor genes or oncogenes . Furthermore, it is also verified that the hypomethylation of hub genes was significantly correlated to tumor proliferation and metastasis . In terms of tumor progression, methylation has also been identified significantly associated with multiple biological processes and signaling pathways in cancers, including tumor stem cell growth , self-renewal , ultraviolet-induced DNA damage response , focal adhesion pathway , and so on. So, for comprehensively evaluating the mechanism of DNA methylation sites in STSs progression, the enrichment analysis of CpG sites corresponded genes was also applied in this study. Based on the gene expression profiles and CpG sites on 4 subgroups of STSs, we found that the levels of gene expression and DNA methylation were consistent, which revealed that methylation sites might affected the pathogenesis of STSs through modulating the expression of corresponded genes. Furthermore, the functional analysis discovered that these annotated genes were enriched in several biological processes and signaling pathways, like Cell adhesion process, apoptosis, SMAD pathway, phosphatidylinositol 3 − kinase (PI3K) pathway, intermediate filament cytoskeleton formation, and others. All of these biological processes and signaling pathway has been declared significantly correlated with tumorigenesis and progression [44–46]. These explores could provide clues to emphasize the relationship between these specific methylation sites and STSs biological processes and signaling pathways.
Moreover, in this study, we identified whether these specific methylation sites could be used at the prognostic level and followed by the construction of a STSs prognostic prediction model. Eventually, we focused on the differential CpG sites among 4 clusters and developed a novel thirteen methylation sites for prognostication. The constructed risk model was determined with a robust prognostic value and demonstrated to be an independent prognostic factor for STSs. In addition, the risk score was also confirmed with a major advantage of its biological implications for predicting STSs intrinsic histological subtypes and its original sites. A similar scenario was also observed in the nomogram analysis that risk signature played a virtual role in predicting the OS of STSs, which may be caused by the intensive correlation between the risk signature and STSs pathogenesis.