Spatial Structure, Transmission Modes and the Evolution of Viral Exploitation Strategies

Berngruber et al. extend prior studies on the effect of spatial structure on virulence evolution to consider the role that vertical transmission (in addition to horizontal transmission) can play in these dynamics. They address this scenario first with a compartmental model and then with a very cool microbial experiment, which show qualitatively similar results. Both of these methods address the competition between a “wild-type” avirulent strain of a pathogen and more virulent and more horizontally transmissible strain. Their simple compartmental model allows for both horizontal and vertical transmission of both pathogen strains. They place this simple compartmental model into a spatially explicit landscape and vary dispersal distance. Generally infection prevalence increases with time and increases more quickly as dispersal increases. They also study the relative frequencies of the two mutants through time using pair approximation methods demonstrating that at the beginning of the outbreak, when susceptibles are common, the mutant strain temporarily increases. The duration of this transient increase depends on the level of mixing with low mixing (local dispersal) resulting in very short term increases in the mutant strain. This phenomenon is due to a quick depletion of local susceptibles limiting horizontal transmission (the better transmission route for the virulent strain) under local dispersal. Local dispersal does, however, increase the effectiveness of vertical transmission for the virulent strain because hosts that are killed locally by the virulent strain are likely to be filled locally by the infected offspring of infected hosts. With global dispersal these vacancies could be filled by the offspring of any individual, including those infected with the avirulent/wild-type strain.

The authors observe quite similar phenomena and mechanisms in their experimental microbial system. Infected cells were inoculated onto a biofilm of susceptible cells in a 1:1 ratio of infection with the mutant or the wild-type strain of phage lambda. These biofilms were then either left undisturbed (no dispersal) disturbed for 30s, 24 hours, or 24 hours in the presence of liquid (global dispersal). Increased mixing changes the epidemic structure of the system but generally increased mixing increases overall prevalence. The effect of mixing on the fitness of the virulent strain also matches theoretical predictions. In the no-mixing treatment mutant fitness rapidly drops off after the first day while in the 24 hour wet (global mixing) treatment the mutant fitness shows a slight increase. Intermediate levels of mixing display intermediate responses (less extreme declines in fitness). They were able to parse the effect of horizontal and vertical transmission on mutant fitness by dyeing susceptible cells and tracking whether new infection events were in dyed cells (horizontal transmission) or undyed cells (vertical transmission). In agreement with theoretical predictions, increased mixing promotes an increased role for horizontal transmission. This study represents a great pairing of simple compartmental models with a tractable experimental system.

Berngruber, Thomas W., Sébastien Lion, and Sylvain Gandon. “Spatial structure, transmission modes and the evolution of viral exploitation strategies.” PLoS Pathog 11.4 (2015): e1004810.