The role of evolution in the emergence of infectious diseases

Abstract

It is unclear when, where and how novel pathogens such as human immunodeficiency virus (HIV), monkeypox and severe acute respiratory syndrome (SARS) will cross the barriers that separate their natural reservoirs from human populations and ignite the epidemic spread of novel infectious diseases. New pathogens are believed to emerge from animal reservoirs when ecological changes increase the pathogen's opportunities to enter the human population¹ and to generate subsequent human-to-human transmission². Effective human-to-human transmission requires that the pathogen's basic reproductive number, R₀, should exceed one, where R₀ is the average number of secondary infections arising from one infected individual in a completely susceptible population³. However, an increase in R₀, even when insufficient to generate an epidemic, nonetheless increases the number of subsequently infected individuals. Here we show that, as a consequence of this, the probability of pathogen evolution to R₀ > 1 and subsequent disease emergence can increase markedly.

Main

The emergence of a disease combines two elements: the introduction of the pathogen into the human population and its subsequent spread and maintenance within the population. Ecological factors such as human behaviour can influence both of these elements, and consequently ecology has been recognized to have an important role in the emergence of disease^1,2,4. In contrast, evolutionary factors including the adaptation of the pathogen to growth within humans and the subsequent transmission of the pathogen between humans are mostly considered in terms of changes in the virulence of the pathogen, and are often thought to have a lesser role in the initial emergence of pathogens⁴. One exception⁵ suggests that immunocompromised individuals might provide “stepping stones” for the evolution of pathogens.

The successful emergence of a pathogen requires R₀ to exceed one in the new host. Only then can an introduction trigger emergence². (Here we use R₀ to refer to spread in human populations, not in the natural reservoir.) If R₀ for a potential pathogen exceeds one, this scenario represents an epidemic waiting to happen. By contrast, when R₀ is initially less than one, infections will inevitably die out and there will be no epidemic unless genetic or ecological changes drive R₀ above one.

There are a number of ways in which R₀ can increase. Ecological changes such as changes in host density or behaviour can increase R₀, as can genetic changes in the pathogen population or in the population of its new host. Genetic changes in the pathogen can arise either through ‘coincidental’ processes such as neutral drift or coevolution of the pathogen and its reservoir host, or through adaptive evolution of the pathogen during chains of transmission in humans. Genetic changes of the new host might be more likely for domesticated or endangered species than for humans.

Here we show that factors, such as ecological changes, that increase the R₀ value of the potential pathogen to a level not sufficient to cause an epidemic (that is, R₀ remains less than one) can greatly increase the length of the stochastic chains of disease transmission. These long transmission chains provide an opportunity for the pathogen to adapt to human hosts, and thus for the disease to emerge.

Our model is illustrated in Fig. 1. Introductions occur stochastically from the natural reservoir of the pathogen, and each primary case is followed by stochastic transmission that generates a variable number of subsequent infections in the human population. We assume that the number of secondary cases follows a Poisson distribution with a mean equal to R₀. Each introduction thus forms branched chains of transmission, which stutter to extinction if R₀ < 1, and the pathogen cannot evolve (µ = 0). The probability that the pathogen evolves and a secondary infection is caused by the mutant is equal to µ for each of the secondary infections (we note that µ incorporates not only the pathogen's mutation rate, but also its dynamics within the host and its transmissibility). We use multi-type branching processes^6,7,8,9 to describe the initial spread of the infection, incorporating the evolution of the pathogen.

Figure 1: Schematic for the emergence of an infectious disease.

Introductions from the reservoir are followed by chains of transmission in the human population. Infections with the introduced strain (open circles) have a basic reproductive number R₀ < 1. Pathogen evolution generates an evolved strain (filled circles) with R₀ > 1. The infections caused by the evolved strain can go on to cause an epidemic. Daggers indicate no further transmission.

In the simplest case (see Fig. 2a) only one mutation is required for the R₀ of the evolved pathogen to exceed one. The probability that a single introduction evolves, causing one (or more) infections with the evolved pathogen (a filled circle in Fig. 1) before it goes extinct, depends very strongly on the R₀ of the introduced pathogen, particularly for low µ values as R₀ approaches 1. This probability is approximately linearly dependent on the rate of evolution µ.

Figure 2: One-step evolution. A single change is required for the pathogen to evolve to R₀ > 1.

a, The probability that an introduction leads to an infection with an evolved strain of the pathogen (filled circle in Fig. 1) is highly sensitive to R₀, and is approximately linearly dependent on the mutation rate µ. Lines correspond to numerical solutions to the branching process model (see Supplementary Information) and symbols correspond to Monte-Carlo simulations following 10⁵ introductions. b, The probability of emergence per introduction depends on the R₀ value of the introduced pathogen and of the evolved pathogen. The solid, dashed and dotted lines correspond to the evolved pathogen having an R₀ of 1,000, 1.5 and 1.2 respectively.

The probability that the introduction leads to an epidemic (the ‘probability of emergence’) depends on the probability of evolution and the probability that the evolved infections do not go extinct due to stochastic effects. In Fig. 2b we plot the probability of emergence for three different R₀ values of the evolved strain. The probability of emergence approaches the probability of evolution when R₀ of the evolved strain is large, and is lower when the R₀ of the evolved strain is close to 1. We find that the probability of emergence depends most strongly on the R₀ of the introduced pathogen, increases approximately linearly with the mutation rate µ, and depends only modestly on the R₀ of the evolved pathogen.

We extend the simple one-step mutation model to consider the situation in which multiple evolutionary changes are required for the pathogen to attain R₀ > 1 in the human population. We begin with a simple scenario, which we call the jackpot model, where the R₀ of the pathogen with the intermediate mutations is the same as that of the introduced pathogen, and where only the addition of the final mutation results in an increase in R₀ to greater than one. As seen in Fig. 3, increasing the number of required evolutionary steps greatly reduces the probability of emergence and increases its sensitivity to changes in R₀. The probability of emergence is approximately proportional to the mutation rate to the power of the number of evolutionary steps required (see Supplementary Information).

Figure 3: Multiple-step evolution.

Here multiple evolutionary changes are required for evolution of the pathogen to have an R₀ > 1. a, Jackpot model with µ = 0.1 and n intermediate changes each with R₀ equal to that of the introduced pathogen: increasing the number of steps (n) greatly decreases the probability of evolution, and makes it more sensitive to the R₀ of the introduced pathogen. b, Alternative multi-step models for the one-intermediate (n = 1) case. The jackpot model (solid line), additive model (dashed line) and fitness valley model (dotted line) are shown (see text for details).

The fitness landscape on which evolution occurs is important in determining the outcome¹⁰. Figure 3b illustrates this for the case of a single intermediate type. As should be expected, changing the jackpot model to an additive model, where the fitness of the intermediate is the average of the fitness of the introduced strain and fully evolved strain, increases the probability that the pathogen evolves to R₀ > 1, whereas changing it to a fitness valley model, where the fitness of the intermediate is lower than the fitness of the introduced strain, decreases the probability of emergence.

The key characteristic distinguishing our model from the conventional view is the R₀ of the introduced pathogen. In the conventional view the R₀ of these infections must be greater than one, whereas in the mechanism described here it is less than one, and evolution during the stochastic chains of transmission allows R₀ to increase above one. In the case of human infections such as HIV^1,11,12, SARS^{13,14,15,16,17} and (potentially) monkeypox^18,19,20,21, seroprevalence studies among groups at high risk of infection from the reservoir would allow evaluation of whether crossover events are usually dead ends, as we expect in our model, or whether they are associated with a large number of secondary cases. In the case of diseases emerging into non-human populations, such as the Nipah virus, which moved from bats to pigs^22,23,24, it may be possible to conduct additional tests involving controlled experimental infections to estimate the R₀ (in this case of the bat Nipah virus in pigs) and to determine whether it evolves during the course of chains of transmission. Such studies may additionally help to identify pathogens that have an ability to evolve rapidly and thus have a high potential for emergence.

The framework presented here has special relevance for pathogens that have been driven to extinction by vaccination. In the case of smallpox there are probably reservoirs of related zoonoses (such as, but not restricted to, monkeypox) from which smallpox may have originated. Although the R₀ of monkeypox in the human population is clearly less than one, there are occasional chains of transmission in the human population^18,19,20,21. As the level of herd immunity to smallpox wanes in the absence of continued vaccination, we expect an increase in R₀ of infections with monkeypox albeit to a level still less than one (the smallpox vaccine provides about 85% cross immunity against monkeypox). Our results suggest that this increase in the effective R₀ of monkeypox in the human population could markedly increase the probability of evolution of monkeypox, allowing it to emerge into a successful human pathogen (which, depending on the evolutionary trajectory followed, may be similar to or differ from smallpox).

The present study could be extended in a number of directions. These include explicitly incorporating the details of ecological interactions such as heterogeneity in the transmission in different areas and subpopulations^2,25,26,27 and incorporating genetic diversity of the pathogen in its reservoir. Finally, we note that this framework can be applied to the more general problem of biological invasions^28,29.

Methods

We describe the dynamics and evolution of emerging diseases as a multi-type branching process with the following probability-generating functions^6,7,8,9:

We calculated the extinction probabilities of the above process numerically. For details and definitions see Supplementary Information.