Mathematical modelling of stress signalling and cell fate during genotoxic stress

Apoptosis is an intracellular signalling pathway that initiates cell suicide upon diverse stress signals, such as genotoxic stress caused by DNA damaging agents. Its deregulation is often associated with carcinogenesis and failure of chemotherapy. In human cells, mitochondrial outer membrane permeabilisation (MOMP) is a key event in linking intrinsic stress signals to the induction of apoptosis, and hence a major decision point that determines the cell’s fate. MOMP is caused and regulated by pro- and anti-apoptotic proteins of the BCL-2 family. Besides their various compositions and concentrations among different tissues, these proteins exhibit a high diversity in their molecular interactions. This diversity and the different binding affinities between proteins of this group aggravate reliable predictions of cell fates or patient responses to chemotherapy, based solely on protein expression levels.

In this thesis, we designed a computational model that includes the topology of interaction of key members of the pro- and anti-apoptotic BCL-2 family and considered the distinct binding affinities of these proteins. Using the model, we were able to support the hypothesis of mode I and mode II inhibition of MOMP whereby the anti-apoptotic BCL-2 proteins are more effective in inhibiting the pro-apoptotic proteins BAK and BAX than the pro-apoptotic BH3-only proteins. Based on quantification of the BCL-2 proteins, we studied the concept of direct and indirect activation and found that the direct activation model successfully described cancer cell responses to the DNA damaging agents 5-fluorouracil and oxaliplatin. Applied to colorectal cancer patients, the designed model predicted patients’ clinical outcome to adjuvant and neo-adjuvant therapy based on patient-specific protein profiles. We investigated the model’s ability to determine whether, and to which extent, cancer cells can be re-sensitised to novel therapeutic agents that inhibit anti-apoptotic BCL-2 proteins. In conclusion, we provided a predictive and prognostic tool that may help to optimise chemotherapeutic treatment of colorectal cancer patients.