Theoretical study of the impact of adaptation on cell-fate heterogeneity and fractional killing

Abstract

Fractional killing illustrates the cell propensity to display a heterogeneous fate response over a wide range of stimuli. The interplay between the nonlinear and stochastic dynamics of biochemical networks plays a fundamental role in shaping this probabilistic response and in reconciling requirements for heterogeneity and controllability of cell-fate decisions. The stress-induced fate choice between life and death depends on an early adaptation response which may contribute to fractional killing by amplifying small differences between cells. To test this hypothesis, we consider a stochastic modeling framework suited for comprehensive sensitivity analysis of dose response curve through the computation of a fractionality index. Combining bifurcation analysis and Langevin simulation, we show that adaptation dynamics enhances noise-induced cell-fate heterogeneity by shifting from a saddle-node to a saddle-collision transition scenario. The generality of this result is further assessed by a computational analysis of a detailed regulatory network model of apoptosis initiation and by a theoretical analysis of stochastic bifurcation mechanisms. Overall, the present study identifies a cooperative interplay between stochastic, adaptation and decision intracellular processes that could promote cell-fate heterogeneity in many contexts.

Protein supersaturation powers innate immune signaling

“…We observed slow and heterogeneous cell death via FADD signaling, a phenomenon also reported by other labs (75,82,83), and which has thwarted attempts to exploit extrinsic apoptosis in treating cancer (84). We suggest that it results from the stochasticity of nucleation and/or heterogeneity of effective FADD concentrations around its phase boundary.…”

Two-level modeling approach to identify the regulatory dynamics capturing drug response heterogeneity in single-cells

“…A longer delay capacity is restricted to the trajectories which are closer to the stable manifold of the saddle node. Importantly, these observations do not imply the existence of a well defined separatrix dividing the space into initial conditions leading to cell death or survival, as suggested earlier in 32 or more recently in 35 . Although a saddle node introduces a division between two locally stable steady states in our model, we found that both resistant and sensitive phenotypes correspond to the same high C8 steady state, while the low steady state corresponds to non responding cells.…”

“…To do so, the authors had added a genetic layer to a comprehensive model of apoptosis 15 , by introducing an ON/OFF promoter activity model, to study the effect of stochastic protein synthesis in generating a large variety of cell responses. More recently, the idea of a separatrix was further developed to suggest that a saddle node can introduce the mechanism for cell decision between death or survival, with each phenotype corresponding to a different steady state, while perturbations in initial conditions impact fate probabilities 35 .…”

Measuring adaptation dynamics to hydrogen peroxide in single human cells using fluorescent reporters

“…Such heterogeneity in a clonal cell line can arise from variation in proteome [51, 68, 69]. It was shown using mathematical modeling of both stress response network and death decision pathway that cell-to-cell variation and noise in signaling networks can lead to fractional killing [70, 71, 72]. It is tempting to ask whether and how the cell-to-cell variability that we observed here affects cell death or survival.…”