Table 1.
Consequences of evolution for traditional live attenuated and recombinant vector vaccines.
Fig 1.
Impact of pre-host evolution on within-host evolution.
The blue curve depicts the time course of revertant frequency in competition with vaccine, where the revertant has a 10% fitness advantage and starts at a frequency of 10−6. The curve shows the well-known population genetic principle that, while the favored type (revertant) is rare, its absolute frequency changes very little. But the frequency eventually reaches a level at which evolution is rapid. The yellow box represents a possible period of pre-host evolution, the green box representing the period of within-host evolution. The periods of within-host evolution are drawn to be the same length in right and left panels, as if the vaccine has the same within-host duration in both cases. The arrow represents a possible point at which manufacture would end and an inoculum be created, thus defining the boundary between pre-host and within-host evolution. The left panel depicts a short period of vaccine manufacture, the right a longer period of vaccine manufacture and one in which more pre-host evolution has occurred. It is thus easy to see the potential importance of pre-host vaccine evolution on within-host evolution, for even when the revertant is not a large component of the inoculum, it can be poised for rapid evolution within the host (right panel). The curve obeys , in which pt represents the revertant frequency in generation t and w the fitness of revertant relative to vaccine. The curve is drawn for a common evolutionary process across pre-host and within-host evolution, but evolution in the within-host phase will typically experience different parameters than evolution in the pre-host phase.
Fig 2.
Independent growth of vaccine (blue) and revertant (green).
The revertant virus has the superior growth rate, but in the absence of interference between the two, vaccine growth is unimpeded and immunity is triggered.
Fig 3.
Diagram of model processes and interactions.
This figure gives all the processes in the full model that includes resource limitation with innate and adaptive immunity. Solid lines represent variables (V, W, R, Z, X, and Y) and dashed lines represent influences. Note that only the top-most box in gray, the specific immune response to the vaccine antigen, acts differentially on the vaccine vs revertant virus. Not all of these components are included in each iteration of the model. Furthermore, this figure omits pre-host processes that occur during vaccine manufacturing that affect inoculum composition.
Fig 4.
Representative dynamics contrasting vaccine evolution with no evolution.
The combination of intrinsic fitness effects and revertant abundance in the inoculum has the most profound effect in depressing immunity, but depressions are observed even when any evolution is allowed. The vertical axes use a log scale, thus diminish the visual appearance of changes. (Top left): Absence of revertant (i.e. no pre-host or within-host evolution). (Top right): The revertant is included at half the inoculum (representing pre-host evolution); it has no intrinsic fitness advantage over vaccine. Immunity (to vaccine) is reduced to just over a third of the level with no evolution. (Lower left): The revertant is a small fraction of the inoculum (0.01) but it has a 20% fitness advantage over vaccine. The level of immunity is 71% that with no evolution. (Lower right): The revertant is half the inoculum and has a 20% fitness advantage over vaccine. The level of immunity is now less than 0.3% that with no evolution—a depression of almost 3 orders of magnitude. The trials are parameterized so that virus is controlled by innate immunity with final clearance due to adaptive immunity; the mutation rate is 0 in all cases. Equations, initial conditions and parameter values not shown here are given in S1 Appendix; R code is included in S1 File.
Fig 5.
Viral load and the level of immunity to the vaccine antigen depend on evolution and vaccine composition (pre-host evolution).
The final vaccine load and immunity against the vaccine antigen depends heavily on two parameters, the inoculum composition (plotted on the x-axis as initial abundance of the revertant virus, W(0)) and the growth advantage of the revertant within the host (c, plotted on the y-axis). The heat maps show how, as the composition shifts toward revertant or as vector superiority increases (as we move to the right or up), there is a reduction in the viral load of the vaccine (defined as ∫ V dt, left panel) and in the magnitude of immunity to the vaccine antigen (X, right panel). The initial amount of vaccine virus is always V(0) = 1 (i.e. logV(0) = 0). Note that the graphs span high frequencies of revertant in the inoculum that should be easily avoided (log W(0) = 1, i.e. W(0) = 10 V(0))—if the researcher is alert to the possibility. We include such extremes merely to show that the outcome is relatively insensitive to small changes in vaccine composition. Equations, initial conditions and parameter values not shown here are given in S1 Appendix; R code is included in S1 File.
Fig 6.
Effect of evolution on the suppression of immunity by impairment parameters.
The final level of immunity to the vaccine antigen (X) depends heavily on the inhibitory parameters kX and kY—which respectively describe how strongly immunity to the vaccine (X) and revertant (Y) suppress the viral populations. The left plot considers the absence of revertant, hence no evolution. The right panel introduces revertant at 3/4 the inoculum, with the same total inoculum size as in the left panel. The revertant reduces immunity X, but the effect of increasing kX is not made worse by the revertant. Intrinsic fitness differences are absent; mutation rate of vaccine to revertant is set to 0 and fitness benefit of revertant (c) is 0. Equations, initial conditions and parameter values not shown here are given in S1 Appendix; R code is included in S1 File.
Fig 7.
Effects of manipulating the inoculum on immunity to the vaccine.
Small inocula that contain vaccine plus revertant are more prone to reduced immunity levels than are large inocula with little revertant. Composition of the vaccine has the larger effect for the inoculm sizes and initial revertant fractions shown, as indicated by the contours being more horizontal than vertical. An intrinsic fitness cost of c = 0.1 was set for these trials. Smaller c values would lead to higher vaccine and immunity levels across the graphs. Equations, initial conditions and parameter values not shown here are given in S1 Appendix; R code is included in S1 File.