Nils Blüthgen: "Using mathematical modelling to dissect complex feedback regulation in EGFR signalling: Towards tailored interference strategies"

Nils Blüthgen
Institute of Pathology / Theoretical Biology, Charite Universitätsmedizin Berlin

Using mathematical modelling to dissect complex feedback regulation in EGFR signalling: Towards tailored interference strategies

The EGFR receptor signal transduction network is controlled by a multitude of feedbacks. Most of these feedbacks act negatively on the network's activity, providing robustness, adaptation and homeostasis in the healthy cell. Depending on the feedback strength and structure, different consequences arise if one wants to interfere with the network, such as in cancer treatment. Many small-molecule inhibitors targeting different proteins in the network are available or in (pre-)clinical tests. Thus, the question arises whether one can predict the efficiency and consequences of applying these inhibitors from the feedback structure. If this is possible, one can device a treatment option tailored to the mutational patterns observed in a cancer, giving rise to personalised medicine.

Mathematical modelling predicts that strong feedbacks provide robustness against interference within the pathway. For example, we recently characterised a very strong feedback from ERK to RAF, which renders inhibition of MEK, one of the prime drug targets, inefficient. Upon mutation in RAF (BRAF V600E), this strong feedback is broken, and thus MEK inhibion becomes efficient. Similarly, we find that the response to the RAF-inhibitor Sorafinib is very inefficient unless as long as the feedback is intact.

In an attempt to rationalise and systematize the search and quantitative characterisation of feedbacks and their effects on drug treatment, we set up combined experimental and theoretical screening pipeline. In particular, we use combinatorial perturbation experiments coupled with medium-scale proteomics that allow us to generate coarse-grain semi-quantitative models for feedback regulation. This in turn generates hypothesis that are subsequently followed up by more targeted experiments. For example, this approach predicted a relatively weak feedback from Erk to the EGFR. While weak feedbacks do not hamper drug efficiency, our model predicts other consequence: If one inhibits MAPK siganalling, EGFR is activated and it starts to signal into the AKT pathway. Follow-up experiments confirmed this prediction. Therefore, inhibition of the pro-proliferation MAPK pathway can have the dramatic and undesired consequences as it activates the anti-apoptotic AKT pathway. This suggests that combinatorial treatment is required to successfully commit cancer cells to apoptosis.