Macrophage phenotype transitions in a stochastic gene-regulatory network model

Abstract

Polarization is the process by which a macrophage cell commits to a phenotype based on external signal stimulation. To know how this process is affected by random fluctuations and events within a cell is of utmost importance to better understand the underlying dynamics and predict possible phenotype transitions. For this purpose, we develop a stochastic modeling approach for the macrophage polarization process. We classify phenotype states using the Robust Perron Cluster Analysis and quantify transition pathways and probabilities by applying Transition Path Theory. Depending on the model parameters, we identify four bistable and one tristable phenotype configuration. We find that bistable transitions are fast but their states less robust. In contrast, phenotype transitions in the tristable situation have a comparatively long time duration, which reflects the robustness of the states. The results indicate parallels in the overall transition behavior of macrophage cells with other heterogeneous and plastic cell types, such as cancer cells. Our approach allows for a probabilistic interpretation of macrophage phenotype transitions and biological inference on phenotype robustness. In general, the methodology can easily be adapted to other systems where random state switches are known to occur.

Efficient construction of Markov state models for stochastic gene regulatory networks by domain decomposition

“…The macrophage polarization model presented in [24] consists of four elementary reactions corresponding to synthesis and degradation of STAT1 ( x 1 ) and STAT6 ( x 2 ), respectively, as well as four stochastic propensity functions presented in Fig. 5.…”

“…5. We consider a parameter set from [24] for which the system has three metastable regions corresponding to STAT1/STAT6 being low/high, low/low, or high/low, as visualized by the potential energy landscape U ( x ) = − log( π ( x )) in Fig. 10a.…”