Prognostic and Predictive Somatic Mutations in Colorectal Cancer

Sammanfattning: Colorectal cancer (CRC) is the third most incident and the second deadliest cancer worldwide. Even though CRC incidence is strongly correlated with age, it has been increasing in developing countries and younger individuals. Patients diagnosed with early stage local disease have 90% survival chance, but those detected with late stage metastatic disease have their survival odds reduced to 12-14%. Besides clinical and pathological tumour stages, molecular markers provide valuable information on patient outcome and help guide decision for different therapies. The aim of this thesis was therefore to identify and study somatic mutations and other genetic alterations in CRC samples to be used as prognostic and predictive biomarkers.In Paper I, we explored the mutational prevalence of known cancer genes in an unselected population cohort of metastatic CRC (mCRC) and compared this to what is reported in clinic-trial populations. This study demonstrated that BRAF-V600E, microsatellite instability-high (MSI-H) and right-sided tumour location were more common in populations than in previous clinical-based studies. Half of the patients had a tumour that was treatable with an FDA-approved targeted drug for mCRC, underlining the importance of evaluating targeted therapies upfront.In Paper II, we investigated what clinical and genetic factors lead to complete response (CR) after radiotherapy, alone or with chemotherapy, in a large population-based rectal cancer cohort, as well which factors impact overall survival and time to recurrence. Tumour stage, size, treatment, and mutation in SMAD4 or SYNE1 were predictors of CR while mutation in KRAS was a predictor of non-CR. BRAF V600E mutation increased the risk of recurrence.In Paper III, we studied the impact of somatic alterations in CRC through whole-genome and transcriptome sequencing. Successful sequencing was possible for 1,063 CRC primary tumours. Sequencing data analyses defined the somatic genomic and expression landscape and identified prognostic somatic coding and non-coding mutations, mutational signatures, structural variants and new expression subgroups. We could also link tumour hypoxia to different genetic alterations and subdivide the MSI samples in two distinct classes.In Paper IV, we integrated pan-cancer and pan-Ephrin mutational data to identify recurrent EPHB1 somatic mutations for further functional evaluation. Selected mutants and wild-type EPHB1 were transduced in DLD1 CRC cells and studied in compartmentalisation and phosphorylation assays. From the selected mutants, 7 lacked impact, 2 enhanced, and 6 reduced or strongly compromised cell compartmentalisation. This is, to date, the first integrative study of pan-cancer EPH receptor mutations followed by in vitro validation.