Pathogen evolution during vaccination campaigns

Abstract

Following the initiation of the unprecedented global vaccination campaign against Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), attention has now turned to the potential impact of this large-scale intervention on the evolution of the virus. In this Essay, we summarize what is currently known about pathogen evolution in the context of immune priming (including vaccination) from research on other pathogen species, with an eye towards the future evolution of SARS-CoV-2.

Fig 1. The fate of a variant (i) is determined by 3 key components of fitness, each of which can be affected by multiple within-host mechanisms of adaptation.

All else being equal, variants with increased infectivity, increased transmissibility, or a long and early infectious period (i.e., long infections and a short generation interval) will have an increased fitness (rate of spread in a population). As indicated in Eq (1), fitness depends on both the degree of adaptation to naïve and primed hosts. Within-host processes affect the 3 components of fitness in each of the host types. Some within-host mechanisms of adaptation can be measured directly using in vitro assays. Some components of pathogen fitness can be inferred from evolutionary epidemiological studies.

Fig 2. Four types of immunity-adapted variants.

Solid lines depict the growth rate of the population of infected individuals for the wild type (blue) and for a variant (red) as a function of the fraction of the population that has been primed against infection by vaccination, previous infection, or both. Priming decreases the growth rate of the wild type (r_N>r_P). Quantities Δr_N and Δr_P are the differences in growth rate between the variant and the wild type in naïve and primed hosts, respectively. Colored shading indicates which type prevails evolutionarily: the wild type (light blue shading) or the variant (light red shading). Panels (a) and (b) show generalists; the variant is also better adapted to naive hosts (Δr_N>0). Generalist variants will outcompete the wild type even in the absence of priming. Panels (c) and (d) show specialists; the variant is maladapted to naïve hosts (Δr_N<0). Specialist variants will outcompete the wild type only above a critical threshold. Panels (a) and (c) show immunity-inhibited variants; the growth rate of the variant decreases with increasing fractions of primed hosts. As a result, the growth rate of infections after adaptation (i.e., after fixation of the fittest type) in a fully primed population (black dot) is always lower than that in a fully naïve population (white dot and dashed line). Panels (b) and (d) are immunity-facilitated variants; the growth rate of the variant increases with increasing fractions of primed hosts. As a result, the growth rate of infections after adaptation in a fully primed population (black dot) is always higher than that in a fully naïve population (white dot) for generalist variants (panel (c)) but it can go either way for specialists (panel (d); only the case where it is lower is shown). Panel (e) show a plot of the growth rate of variants in a fully naïve (r_i,N) and a fully primed (r_i,P) population. Blue dot indicates location of the wild type. Uncolored region corresponds to variants whose growth rate in primed hosts is less than that of the wild type and so are immunity-maladapted (and so ignored in our discussion). Different colored regions correspond to the 4 types of variants from panels (a–d). Finer distinctions within these types are presented in S1 Fig. See S2 Appendix for a discussion of alternative ways to visualize variants.

Fig 3. Selection and genetic variation.

Plots of the growth rate of all viable variants (i) in a fully naïve and a fully primed population (black dots). Large blue dot denotes the current wild type. Red dots are those variants that are most accessible from the wild type. Note that the location of all variants along the r_i,P axis is specific to an immune response and may be different for natural immunity and different vaccines. All variants in the white region are selectively advantageous but variants in the direction of the selection arrow are most strongly favored (dashed lines indicate contours of overall growth rate). Variants in the gray region are disfavored by selection. The direction of selection arrow is upwards in a fully naïve population (p = 0) (panel (a)) and shifts towards the right as the fraction of primed individuals increases (panels (b and c)).

Fig 4. Pathogen adaptation as the fraction of primed individuals increases.

Plots of the growth rate of all viable variants in a fully naïve and a fully primed population (dots). Large blue dot denotes the phenotype of the current wild type and black arrow indicates direction of selection (i.e., the variants that are most advantageous). Variants in the gray region are disadvantageous. Note that the location of all variants along the r_i,P axis is specific to immune response and may be different for natural immunity and different vaccines. Colored regions indicate the 4 different kinds of variants. (a) Early in a novel host–pathogen association when a small fraction of hosts are primed. Many potential new variants will be better adapted to both host types (i.e., they will be generalists). (b) Later in the association, when the pathogen is better adapted to its novel host (and a larger fraction of hosts are primed). The evolutionary trajectory of successive fixation events leading to the new wild type variant is indicated by the succession of blue dots. Note how the change in the location of the blue dot can affect the typology of some variants (i.e., a variant that was identified as a generalist in the early stage of adaption could later become a specialist relative to the more recent form of the pathogen). Once the level of adaptation is high (panel (b)), most advantageous variants that appear will tend to be specialists. Even though generalists are still more strongly favored by selection there are fewer of them that can arise.