SCISSORS · Single cell profiling of breast tumors and the residing immune cells leveraged by integration with multi-dimensional molecular data from thousands of tumors

Single cell profiling of breast tumors and the residing immune cells leveraged by integration with multi-dimensional molecular data from thousands of tumors

Breast cancer is the most common type of cancer in Europe, responsible for the highest women cancer mortalities each year. Characterisation of breast tumours teased them apart to distinct subtypes and facilitated targeted treatments that improved survival rates significantly, yet some aggressive subtypes remain difficult to treat. New therapies that harness the patient own immune system to fight the tumour show remarkable success in several cancers but limited one in breast cancer. Single cell genomics methods that profile the tumour and its surrounding cells hold the potential to better understand the dynamics of tumour development and its resistance to therapy and improve current cancer treatment. The SCISSORS project aimed to profile normal and cancerous breast tissue samples on the single cell level to elucidate the interplay between genomic clones, their functional profile, and the response of the stromal environment as they all evolve. The objectives were 1) analyse a comprehensive single cell expression atlas of healthy human breast tissue, 2) develop a computational framework to infer somatic clones from single cell RNA-seq data, 3) profile the complete environment of normal and malignant breast tissues at single cell and clonal levels, and 4) extend results by integration with existing data on thousands of multi-dimensional molecularly profiled tumour samples. The technical developments made in the project facilitated the analysis of a large cohort of breast tumours from various types. Using a normal breast cell atlas as reference enabled the classification of single cells to phenotypes. We observed single cells mixing several normal epithelial breast phenotypes, a phenomenon confirmed to be unique to tumours. However, a minority of the tumours evaded this phenotypical plasticity and remained extremely homogenous. We were able to infer the genomic clones by combining several published and in-house tools. Integration of these genomic clones with the phenotypic classifications shed light on the dynamics of clonal evolution and the corresponding transcriptional changes.

Текст от CORDIS, на английски · Данни: CORDIS, © Европейски съюз