Stochastic modeling of aging cells reveals how damage accumulation, repair, and cell-division asymmetry affect clonal senescence and population fitness

Abstract

Background: Asymmetry during cellular division, both in the uneven partitioning of damaged cellular components and of cell volume, is a cell biological phenomenon experienced by many unicellular organisms. Previous work based on a deterministic model claimed that such asymmetry in the partitioning of cell volume and of aging-associated damage confers a fitness benefit in avoiding clonal senescence, primarily by diversifying the cellular population. However, clonal populations of unicellular organisms are already naturally diversified due to the inherent stochasticity of biological processes.

Results: Applying a model of aging cells that accounts for natural cell-to-cell variations across a broad range of parameter values, here we show that the parameters directly controlling the accumulation and repair of damage are the most important factors affecting fitness and clonal senescence, while the effects of both segregation of damaged components and division asymmetry are frequently minimal and generally context-dependent.

Conclusions: We conclude that damage segregation and division asymmetry, perhaps counterintuitively, are not necessarily beneficial from an evolutionary perspective.

Fig. 1

a Illustration of the model. The cell grows until it reaches a critical volume and divides. It accumulates damage over time, which slows down volume growth. The accumulated damage can also be repaired. There is no separate mechanism for killing a cell due to damage, because high level of accumulated damage will prevent a cell from dividing and cause it to be rapidly overtaken by faster-dividing cells. b The cell volume module of the model. The volume grows exponentially until it reaches a generation-dependent critical volume and the cell divides (blue dashed line). c Two mechanisms of particular interest in this study: segregation of damaged proteins in mother cells, and division asymmetry of cell volume. Yellow dots indicate normal proteins, while green dots indicate damaged proteins. d Illustration of the exponential growth of simulated cell population. An initial population of 2000 cells were simulated for 6000 min with periodic resampling (blue dashes) every time the population size exceeds 2 million cells. The red dashed line indicates the expected population size without sampling

Fig. 2

a Sensitivity analysis for the clonal senescence outcome. The parameter combinations tested were partitioned into groups such that all combinations in each group differs only in the value of one parameter. The fraction of groups with divergent senescence outcome is plotted for each parameter. b Sensitivity analysis for the doubling time outcome. The parameter combinations tested were partitioned into groups such that all combinations in each group differs only in the value of one parameter. The fraction of groups where the minimum doubling time is at least 5% below the maximum is plotted for each parameter. Level of transparency indicates the size of the effect. In each panel, the color indicates the value of the parameter corresponding to the most-fit combination in the group: red indicates that the smallest parameter value is the most fit; blue indicates that the largest parameter value is the most fit; and green indicates that an intermediate parameter value is the most fit

Fig. 3

a-h Relative representation of the other parameters in the cases where changing the nonlinearity of the effect of damage on volume growth rate caused a significant (> 5%) fitness difference. In each panel, the color indicates the level of nonlinearity resulting in maximum fitness: red indicates that no nonlinearity (α = 1) is the most fit; blue indicates that maximum nonlinearity (α = 4) is the most fit; and green indicates that an intermediate level of nonlinearity is the most fit

Fig. 4

a-h Relative representation of the other parameters in the cases where changing the level of damage segregation caused a significant (> 5%) fitness difference. In each panel, the color indicates the level of damage segregation resulting in maximum fitness: red indicates that no segregation is the most fit; blue indicates that full segregation is the most fit; and green indicates that an intermediate level of segregation is the most fit

Fig. 5

a-h Relative representation of the other parameters in the cases where changing the level of division asymmetry caused a significant (> 5%) fitness difference. In each panel, the color indicates the level of division asymmetry resulting in maximum fitness: red indicates that no division asymmetry is the most fit; blue indicates that maximum asymmetry (1:4) is the most fit; and green indicates that an intermediate level of asymmetry is the most fit

Fig. 6

a-h Relative representation of the other parameters in the cases where changing the level of damage-related noise caused a significant (> 5%) fitness difference. In each panel, the color indicates the level of damage-related noise resulting in maximum fitness: red indicates that no damage-related noise is the most fit; blue indicates that maximum damage-related noise (10%) is the most fit; and green indicates that an intermediate level of damage-related noise (5%) is the most fit

Fig. 7

a-h Relative representation of the other parameters in the cases where changing the level of volume module noise caused a significant (> 5%) fitness difference. In each panel, the color indicates the level of volume module noise resulting in maximum fitness: red indicates that no volume module noise is the most fit; blue indicates that maximum volume module noise (10%) is the most fit; and green indicates that an intermediate level of volume module noise (5%) is the most fit