Tumor collagens predict genetic features and patient outcomes

Abstract

The extracellular matrix (ECM) is a critical determinant of tumor fate that reflects the output from myriad cell types in the tumor. Collagens constitute the principal components of the tumor ECM. The changing collagen composition in tumors along with their impact on patient outcomes and possible biomarkers remains largely unknown. The RNA expression of the 43 collagen genes from solid tumors in The Cancer Genome Atlas (TCGA) was clustered to classify tumors. PanCancer analysis revealed how collagens by themselves can identify the tissue of origin. Clustering by collagens in each cancer type demonstrated strong associations with survival, specific immunoenvironments, somatic gene mutations, copy number variations, and aneuploidy. We developed a machine learning classifier that predicts aneuploidy, and chromosome arm copy number alteration (CNA) status based on collagen expression alone with high accuracy in many cancer types with somatic mutations, suggesting a strong relationship between the collagen ECM context and specific molecular alterations. These findings have broad implications in defining the relationship between cancer-related genetic defects and the tumor microenvironment to improve prognosis and therapeutic targeting for patient care, opening new avenues of investigation to define tumor ecosystems.