Comparative mapping of crawling-cell morphodynamics in deep learning-based feature space

Abstract

Navigation of fast migrating cells such as amoeba Dictyostelium and immune cells are tightly associated with their morphologies that range from steady polarized forms that support high directionality to those more complex and variable when making frequent turns. Model simulations are essential for quantitative understanding of these features and their origins, however systematic comparisons with real data are underdeveloped. Here, by employing deep-learning-based feature extraction combined with phase-field modeling framework, we show that a low dimensional feature space for 2D migrating cell morphologies obtained from the shape stereotype of keratocytes, Dictyostelium and neutrophils can be fully mapped by an interlinked signaling network of cell-polarization and protrusion dynamics. Our analysis links the data-driven shape analysis to the underlying causalities by identifying key parameters critical for migratory morphologies both normal and aberrant under genetic and pharmacological perturbations. The results underscore the importance of deciphering self-organizing states and their interplay when characterizing morphological phenotypes.

Mass-Conservation Increases Robustness in Stochastic Reaction-Diffusion Models of Cell Crawling

“…Therefore, some other new models consider only a single variable for the membrane together with a global feedback to account for the mass-conservation condition for a single concentration [17,32,38]. When the dynamics of the concentration at the membrane is combined with other membrane concentrations more complex models including also excitable signalling networks are developed [39,47,48,55]. In contrast with such models, in this study we chose an existing complex model where the condition of bistability is finely tuned [23] and add the global feedback condition to account for the conservation of the protein.…”

“…9, however we could have employed a similar condition for parameter k 18 or any other parameter which exhibits a hysteresis behavior on the concentration P. Actually, note that other global feedback were already considered in the original version of the model [23]; see the dynamics of the global inhibitors, defined by G R and G P in Eqs 2-5 respectively, similarly to other models of the motion of Dictyostelium discoideum cells based on global cytosolic quantities [12,48]. The addition of global control quantities which can determine the available concentration inside the cell has previously been employed in other models [32,38,47].…”

“…Therefore, the parameter k 14 is dynamically controlled depending on the total amount of the protein of the inducer P at the cell. A new control term related to k 14 is incorporated, following previous similar approaches in simple models [32,38,47]:…”

“…4. Following similar previous approaches [34,47], we included this conservation as a global feedback condition in the reactiondiffusion equations of the model, see Section 2.3.…”

“…Therefore, we propose a constraint on the conservation of a component of the signal pathway, controlling the autocatalytic mechanism, in order to increase the robustness of the model in representing changes in parameter values [46]. Similar types of conservation have previously been employed in other models of motion of Dictyostelium discoideum cells [34,38,47,48], and we show here that a constraint of this kind is an useful and reasonable condition to systematically increase the robustness of the crawling mechanism, increasing the window of parameter values allowing cell migration.…”