An ecological and stochastic perspective on persisters resuscitation

Abstract

Resistance, tolerance, and persistence to antibiotics have mainly been studied at the level of a single microbial isolate. However, in recent years it has become evident that microbial interactions play a role in determining the success of antibiotic treatments, in particular by influencing the occurrence of persistence and tolerance within a population. Additionally, the challenge of resuscitation (the capability of a population to revive after antibiotic exposure) and pathogen clearance are strongly linked to the small size of the surviving population and to the presence of fluctuations in cell counts. Indeed, while large population dynamics can be considered deterministic, small populations are influenced by stochastic processes, making their behaviour less predictable. Our study argues that microbe-microbe interactions within a community affect the mode, tempo, and success of persister resuscitation and that these are further influenced by noise. To this aim, we developed a theoretical model of a three-member microbial community and analysed the role of cell-to-cell interactions on pathogen clearance, using both deterministic and stochastic simulations. Our findings highlight the importance of ecological interactions and population size fluctuations (and hence the underlying cellular mechanisms) in determining the resilience of microbial populations following antibiotic treatment.

Keywords: 37N25; 46N60; Microbial communities; Microbial interactions; Persisters.

Graphical abstract

Fig. 1

Schematic representation of the three-member community modelled herein. (A) Network describing the dynamics of the species A (in green), B (in purple) and C (in orange), and the possibility of the persister P formation (in yellow). (B) Random interaction matrix that produces a stable non-zero steady state. (C) Sample simulation with a final co-existence of all the three species. (D) Dynamics of the community with different initial abundances for each species lead to the same steady state. (E) Dynamics of the community under antibiotic (ABX) stress in the interval 12-18 h. (F) Changes in the final composition of abundances resulting from varying the strength of the interactions among the three community members (i.e. changing the interaction matrix; line opacity indicates interaction strength). (G) Percentage of communities in which species C disappears when the interaction strength increases.

Fig. 2

Changes in the main parameters of the model. A) Changes in antibiotic exposure time t_end, antibiotic killing rate ξ and persister generation rate γ have no impact on the end-point values of species C. B-D) Impact on the post-antibiotic dynamic of species C of antibiotic exposure time (B), its strength (C) and persister generation rate (γ) (D). E) The effect of different antibiotic killing rate on the optimal resuscitation time. F) The effect of antibiotic exposure time on the optimal resuscitation time.

Fig. 3

The effect of fluctuations in a three-member community. A-B) A sample of deterministic (straight line) and stochastic (step line) simulations. It is shown the overlap between the two models, both in a non-antibiotic (A) and antibiotic (B) scenario. However, stochastic effects may become relevant when the size of species C population reaches very low values. C) Deterministic simulation of nine communities using different interaction matrices that resulted in the extinction of species C. D) Using the same interaction matrices, species C reappears when modelling the community dynamics stochastically while A and B disappear from the community. Not all interaction matrices led to persister resuscitation, and thus to species C reappearance. Here we show in E) the average α values that lead to it (n=11,288 for each α), and in F) the average α values that lead species C to extinction even under the fluctuation system (n=16,391 for each α). In both matrices the standard deviation is reported. Furthermore, on G) we show the number of interaction matrices that led to the survival of species C when modelling the dynamics stochastically, and the number of simulations that used that specific matrix. In other words, there is a set of interaction coefficients that are more common than others (Species A: green, B: purple, C: red, P: yellow).