CAUSES OF ANEUPLOIDY IN TUMOURS AND THE CONSEQUENCES ON GENE EXPRESSION AND PROTEIN COMPLEX STOICHIOMETRY

Protein complexes are dynamic assemblies in which proteins bind each other in different physiological cell conditions and stoichiometries to perform cellular functions. Failures in maintaining complex stoichiometries cause proteotoxic stress, and are associated with proliferative and survival disadvantages in normal cells. Tumours are characterised by massive dysregulation of genes leading to imbalances in protein dosage and thus in protein complex stoichiometries. Aneuploidies, arm- or chromosome-level copy number aberrations, are one of the main reasons for transcriptional dysregulation, yet paradoxically they frequently occur in cancer genomes. We use aneuploid tumours, harbouring chromosome-level amplifications and deletions, as a model to understand how tumour cells compensate for aneuploidy induced transcriptional dysregulation. We observe that regulation of co-complex members in trans acts as a compensatory mechanism to deal with abundance changes on the aneuploid chromosome itself. We show that this compensation is stronger for aggregation-prone proteins of aneuploid chromosomes and those involved in a smaller number of complexes suggesting the role of protein complex organisation in modulating those compensatory mechanisms. Further, we provide evidence that this compensation in aneuploid tumours is established through post-translational regulation, and that higher degree of success in this compensation is associated with better tumour fitness, and failure results in activation of protein degradation programs. However, it is still unclear why aneuploidies and focal copy number alterations are occurring repeatedly in cancer genomes if they cause those compensation problems through dysregulation. To address this, we ask if we can model the observed frequencies of genomic amplifications and deletions as a function of avoiding transcriptional dysregulation of co-complex members or affecting large number of genes, as an example of negative selection, probability of occurrence (e.g. distance to telomere/centromere), and amplifying oncogenes or deleting tumour suppressor genes, representing positive selection, by using machine learning models. We find a balance among these factors in explaining to a certain degree the observed genomic alteration patterns in cancer genomes. Taken together, our findings describe the need for compensation mechanisms to deal with the imbalances in protein complex stoichiometry induced by aneuploidy, and highlight the importance of protein complex components as potential vulnerabilities for the identification of drug targets for clinical use, in addition to providing insights into understanding tumour genome evolution and factors driving frequently observed genomic alterations.