Sarcomas are malignant soft tissue and bone tumours affecting adults, adolescents and children. They represent a morphologically heterogeneous class of tumours and some entities lack defining histopathological features. Therefore, the diagnosis of sarcomas is burdened with a high inter-observer variability and misclassification rate. Here, we demonstrate classification of soft tissue and bone tumours using a machine learning classifier algorithm based on array-generated DNA methylation data. This sarcoma classifier is trained using a dataset of 1077 methylation profiles from comprehensively pre-characterized cases comprising 62 tumour methylation classes constituting a broad range of soft tissue and bone sarcoma subtypes across the entire age spectrum. The performance is validated in a cohort of 428 sarcomatous tumours, of which 322 cases were classified by the sarcoma classifier. Our results demonstrate the potential of the DNA methylation-based sarcoma classification for research and future diagnostic applications.

Tumor microenvironment is considered nowadays as one of the main players in cancer development and progression. Tumor microenvironment is highly complex and consist of non-tumor cells (i.e. cancer-associated fibroblast, endothelial cells or infiltrating leucocytes) and a large list of extracellular matrix proteins and soluble factors. The way in that microenvironment components interact among them and with the tumor cells themselves is very complex and only partially understood. However, it is now clear that these interactions govern and modulate many of the cancer hallmarks such as cell proliferation, the resistance to death, the differentiation state of tumor cells, their ability to migrate and metastasize and the immune response against tumor cells. One of the microenvironment components that has emerged in the last years with strength is a heterogeneous group of multifaceted proteins grouped under the name of matricellular proteins. Matricellular proteins are a family of non-structural matrix proteins that regulate a variety of biological processes in normal and pathological situations. Many components of this family such as periostin (POSTN), osteopontin (SPP1) or the CNN family of proteins have been shown to regulate key aspect of tumor biology, including proliferation, invasion, matrix remodeling and dissemination to pre-metastatic niches in other organs. Matricellular proteins can be produced by tumor cells itself or by other tumor-associated cells, and their synthesis can be affected by intrinsic and/or extrinsic tumor cell factors. In this review, we will focus on the role of periostin in the development and progression of cancer. We will describe their functions in normal tissues and the mechanisms involved in their regulation. We will analyze the tumors in which their expression is altered and their usefulness as a biomarker of tumor progression. Finally, we will speculate about future directions for research and therapeutic approaches targeting periostin.

Background: Retinoblastoma, a prototype of hereditary cancer, is the most common intraocular tumour in children and potential cause of blindness from therapeutic eye ablation, second tumours in germ line carrier's survivors, and even death when left untreated. The molecular scanning of RB1 in search of germ line mutations lead to the publication of more than 900 mutations whose knowledge is important for genetic counselling and the characterization of phenotypic-genotypic relationships. Results: A searchable database (RBGMdb) has been constructed with 932 published RB1 mutations. The spectrum of these mutations has been analyzed with the following results: 1) the retinoblastoma protein is frequently inactivated by deletions and nonsense mutations while missense mutations are the main inactivating event in most genetic diseases. 2) Near 40% of RB1 gene mutations are recurrent and gather in sixteen hot points, including twelve nonsense, two missense and three splicing mutations. The remainder mutations are scattered along RB1, being most frequent in exons 9, 10, 14, 17, 18, 20, and 23. 3) The analysis of RB1 mutations by country of origin of the patients identifies two groups in which the incidence of nonsense and splicing mutations show differences extremely significant, and suggest the involvement of predisposing ethnic backgrounds. 4) A significant association between late age at diagnosis and splicing mutations in bilateral retinoblastoma patients suggests the occurrence of a delayed-onset genotype. 5) Most of the reported mutations in low-penetrance families fall in three groups: a) Mutations in regulatory sequences at the promoter resulting in low expression of a normal Rb; b) Missense and in-frame deletions affecting non-essential sequence motifs which result in a partial inactivation of Rb functions; c) Splicing mutations leading to the reduction of normal mRNA splicing or to alternative splicing involving either true oncogenic or defective (weak) alleles. Conclusion: The analysis of RB1 gene mutations logged in the RBGMdb has shown relevant phenotype-genotype relationships and provided working hypothesis to ascertain mechanisms linking certain mutations to ethnicity, delayed onset of the disease and low-penetrance. Gene profiling of tumors will help to clarify the genetic background linked to ethnicity and variable expressivity or delayed onset phenotypes.