Activation of JNK signalling is required for therapeutically mediated apoptosis in a variety of tumour types, however the mechanism that regulates rapid JNK activation under these conditions remains poorly characterised. We previously hypothesised that an unidentified positive feedback loop embedded within the kinases upstream of JNK was responsible for this ultrasensitive JNK activation (Fey, Croucher et al, Frontiers in Physiology, 2012). Our experimental analysis now confirms that JNK phosphorylates its own upstream kinase, MKK7, and relieves auto-inhibition mediated by the N-terminal region of MKK7. Furthermore, we have shown that JNK activation is constrained by inhibitory crosstalk from Akt to both MKK4 and MKK7, and regulated through scaffolding with the MAP3K ZAK.
Utilising this experimentally resolved network structure, we generated a mathematical model capable of simulating JNK activation kinetics and predicting dynamic behaviour within this kinase network. We further adapted the model to generate patient-specific simulations of JNK activation, based upon gene expression data for 369 neuroblastoma patients. Survival analysis performed using these patient-specific simulations revealed that impaired JNK activation in silico is an independent prognostic indicator of poor overall survival for Neuroblastoma patients. This novel approach demonstrates that modelling of signalling pathways can not only decipher the behaviour of complex networks on the cellular level, but can also relate pathway behaviour to clinical outcomes on the level of individual patients.