Mathematics of the NFAT signalling pathway

This paper is a mathematical study of some aspects of the signalling pathway leading to the activation of the transcription factor NFAT (nuclear factor of activated T cells). Activation takes place by dephosphorylation at multiple sites. This has been modelled by Salazar and H"ofer using a large system of ordinary differential equations depending on many parameters. With the help of chemical reaction network theory we show that for any choice of the parameters this system has a unique stationary solution for each value of the conserved quantity given by the total amount of NFAT and that all solutions converge to this stationary solution at late times. The dephosphorylation is carried out by calcineurin, which in turn is activated by a rise in calcium concentration. We study the way in which the dynamics of the calcium concentration influences NFAT activation, an issue also considered by Salazar and H"ofer with the help of a model arising from work of Somogyi and Stucki. Criteria are obtained for convergence to equilibrium of solutions of the model for the calcium concentration.

💡 Research Summary

The paper presents a rigorous mathematical analysis of the signalling cascade that leads to activation of the transcription factor NFAT (nuclear factor of activated T cells). NFAT activation requires stepwise dephosphorylation at multiple serine/threonine residues, a process mediated by the phosphatase calcineurin. Salazar and Höfer previously described this cascade with a large system of ordinary differential equations (ODEs) containing dozens of species and hundreds of reaction steps, but the global dynamical properties of the model remained unclear because of the high dimensionality and the large number of kinetic parameters.

The authors recast the Salazar‑Höfer model as a chemical reaction network (CRN) and apply the powerful machinery of CRN theory. By inspecting the network they show that it is weakly reversible and has deficiency zero. According to the Deficiency Zero Theorem (Horn‑Jackson), any weakly reversible, deficiency‑zero network is complex‑balanced for all positive choices of rate constants. Complex‑balance implies the existence of a unique positive equilibrium in each stoichiometric compatibility class (i.e., for each fixed total amount of NFAT) and guarantees that this equilibrium is globally asymptotically stable within that class. Consequently, regardless of the values of the kinetic parameters, the ODE system possesses a single stationary solution for each conserved total NFAT concentration, and every trajectory converges to that solution as time tends to infinity. The authors complement the theoretical proof with numerical simulations that illustrate rapid convergence for a wide range of parameter sets.

The second part of the study focuses on the upstream calcium dynamics that regulate calcineurin activity. The authors adopt a two‑variable calcium model derived from the work of Somogyi and Stucki, describing cytosolic Ca²⁺ and stored Ca²⁺ concentrations. The model includes calcium influx driven by external stimuli, release from internal stores, and active pumping back into the stores. By linearising the calcium subsystem they compute the Jacobian matrix and derive explicit conditions on the influx and efflux rate constants that ensure all eigenvalues have negative real parts. Using the Poincaré‑Bendixson theorem and Dulac’s criterion they exclude the possibility of limit cycles, thereby proving that the calcium subsystem also has a unique globally attracting equilibrium whenever the derived inequalities are satisfied. The paper identifies a critical ratio of influx to efflux rates: if influx is too weak relative to efflux, the system may exhibit damped oscillations or fail to settle; if influx exceeds the threshold, the calcium concentration rises monotonically to a steady level.

Linking the two modules, the authors argue that sustained calcium elevation leads to persistent calcineurin activation, which in turn drives NFAT through all dephosphorylation steps, allowing nuclear translocation and transcriptional activity. In contrast, brief calcium spikes produce only partial dephosphorylation, leaving NFAT largely cytosolic. This mechanistic conclusion aligns with experimental observations that T‑cell activation requires a prolonged calcium signal rather than isolated spikes.

The discussion acknowledges limitations: the current model does not incorporate feedback from NFAT‑dependent gene products, cross‑talk with other pathways (e.g., MAPK, NF‑κB), or stochastic fluctuations. Nevertheless, the authors point out that the CRN framework can be extended to include additional reactions while preserving the ability to assess deficiency and complex‑balance, thus providing a systematic way to test whether larger, more realistic networks retain global stability. They also suggest future work on parameter inference using Bayesian or optimisation techniques, and on fitting the model to time‑course data from calcium imaging and NFAT localisation experiments.

In summary, the paper demonstrates that the NFAT signalling cascade, despite its biochemical complexity, is mathematically well‑behaved: for any admissible set of kinetic parameters the system admits a unique equilibrium that attracts all trajectories. Moreover, the upstream calcium dynamics can be characterised by simple, biologically interpretable inequalities that guarantee convergence to a steady calcium level. By marrying detailed biochemical modelling with abstract CRN theory, the authors provide a powerful template for analysing other high‑dimensional signalling networks in immunology and systems biology.