Decoding cytokine dynamics with biochemical networks

T cells defend their hosts by scanning other cells for ligands from potentially pathogenic sources.Upon antigen detection, they produce and consume extracellular proteins called cytokines to communicate with each other.Those messenger proteins play a fundamental role in scaling the nonlinear, out-of-equilibrium immune response.Yet, no quantitative theory exists to systematically account for the dynamics of multiple cytokines at once.However, state-of-the-art experimental data from our collaborators (Altan-Bonnet lab, NIH) reveal that cytokine concentrations encode information about the strength, or quality, of the ligands detected by the T cells.In this thesis, we combine cytokine biology, information theory, and nonlinear dynamics to establish theoretical models for cytokines.Previous modelling attempts in this direction suffer from excessive complexity or describe only one cytokine, so we adopt a different angle.We formulate an inverse problem: how do T cells decode cytokine time courses to infer ligand quality?We give a proper definition of information processing in this context and look for biochemical networks that perform this function.Extending the evolutionary algorithm -evo, we find a network relying on a combination of cytokines IL-2, IL-6, and TNF-.We are able to simplify and solve it analytically.The simplified network predicts ligand quality with an accuracy of 80% in various experimental repeats collected over two years by our collaborators.We obtain a lower bound of (1.05 0.08) bits for the information content of cytokine time courses about ligand quality.The network's interpretability yields important biophysical insights: for instance, it provides a simple mechanism for time-averaging cytokine signals in T cells.It also opens up new questions grounded in physics about collective computations and information processing in biology.vi I first want to thank my advisor, Prof. Paul Franois, for his flawless support over the past two years.His inexhaustible capacity to find new ideas and his optimism proved salutary for my own motivation.He is both a driving and humane mentor.In addition, I am very lucky to have collaborated with Grgoire Altan-Bonnet's laboratory.They displayed outstanding hospitality in welcoming me for two separate

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.