COUPLED OSCILLATORS: DETECTING THE FUNCTIONAL CONSEQUENCES OF SIGNALLING PATHWAY INTERACTIONS

Many of the activities of biological cells are regulated by proteins which carry signals that modify the expression of different genes at a given time, but how these signals do so is not known. Typical signals in general, such as those detected by a radio, may be encoded in terms of their amplitude (amplitude-modulated - AM) or frequency (FM), and until recently it was assumed that it was changes in the concentration (amplitude) of these signalling proteins that were important. Recently we have shown by studies in single cells that the signals in one of these pathways, the so-called NF-kappaB pathway, are oscillatory and that it is the frequency of these oscillations that seems to determine the downstream response. However, there are many other signalling pathways in a cell, and what is also unknown is the extent to which different signalling pathways are coupled to each other. Many years ago, anomalies in the orbit of Neptune led to the recognition that they must be caused by the presence of an unknown i.e. unobserved planet (Pluto), and calculations allowed astronomers both to predict the orbit of the unknown planet and thus to discover it observationally. This recognition that interactions between two dynamical systems A and B allows one to infer the presence of functional interactions solely by looking for the effects of B on A provides a novel and powerful tool that we wish to exploit here. This is because interactions between nonlinear systems oscillating respectively at A Hz and B Hz causes the production in pathway A of 'beat' frequencies of (A plus/minus nB) Hz. Since these frequencies can be measured with high precision, we seek to develop and exploit this idea for the determination of functional interactions between signalling pathways whose 'natural' frequencies of oscillation differ. This will be done both computationally (by studying a computer model of the NF-kappaB pathway, which we shall extend) and experimentally. A specific focus will be on interactions with a related and important pathway called the p53 pathway, which is modified in a large percentage of human cancers. The result will be a novel set of tools with which we can determine the interactions between signalling pathways by looking at their frequency components.

We have applied single cell imaging and computational modelling to show that the NF-kappaB signalling system functions as a non-linear oscillator that controls target gene expression. We demonstrated that the excellent temporal resolution of our single cell imaging data allows detailed analysis of the characteristics of the (~100 min) oscillations of NF-kappaB proteins between the cytoplasm and nucleus. We applied a computational model, based on 64 parameters and 24 variables, to simulate the core NF-kappaB:IkappaBalpha negative feedback loop that regulates the timing and amplitude of oscillations. Despite the clear utility of this model, it lacks important components of the NF-kappaB pathway and fails to predict important aspects of our experimental data (particularly from multi-pulse re-stimulation experiments). There is also a need to consider possible stochastic as well as deterministic characteristics of the system. We will therefore improve and expand the NF-kappaB model while identifying critical parameters that are important in controlling output dynamics and function of the system. Experimentally, we have detected oscillations in other components of the NF-kappaB pathway such as in the RelB protein which can form a second oscillatory negative feedback loop with p100/p52. The RelB and RelA modules represent coupled oscillators. A major aim of this proposal is to understand from a theoretical point of view how the interaction between coupled oscillatory systems will affect output dynamics and function. We will use Fourier analysis to analyse the imaging time-course data and we will study the theoretical sensitivity of model parameters in the frequency domain. We will simulate how modulation of a hypothetical coupled oscillatory pathway would affect the NF-kappaB oscillations and relate this to the coupling of the RelA-IkappaBalpha and RelB-p100/p52 modules. Finally, we will study the interaction of NF-kappaB with the important p53-MDM2 system. We have already been able to experimentally measure oscillations in the p53-MDM2 system in single cells which occur with a 400 min period. We will now establish a computational model for the p53 system. The NF-kappaB and p53 systems are probably the two most important stress response systems in the mammalian cell that together regulate cell fate. A great deal of evidence has shown that these systems interact with each other although different mechanisms have been proposed for this cross-talk. We will study, on a theoretical and experimental basis, how these two systems with different frequencies interact with one another. These studies will act as a paradigm for the coupling between oscillatory signalling pathways and will also provide key mechanistic information on two systems that are important drug targets for inflammatory disease and cancer.