Males can evolve lower resistance to sexually transmitted infections to infect their mates and thereby increase their own fitness

Abstract

Sexually transmitted infections (STIs) often lower their host’s future reproductive success by inducing sterility. Females can minimise the reproductive cost of infection by plastically increasing their current reproductive effort (i.e. terminal investment) before they become sterile. In polyandrous systems, long-term female survival or fecundity is often irrelevant to male fitness. Mating with an infected, terminally investing female potentially yields greater fitness gains for males than mating with an uninfected female. Males might consequently benefit from infecting females with an STI. We construct mathematical models of the evolutionary consequences of a sterilising STI. We show that females should terminally invest in response to an STI when immune investment is relatively ineffective at delaying STI-induced sterility. Cost-effective immune responses may conversely select for reduced reproductive effort after infection (‘terminal divestment’). Crucially, we then show that female terminal investment can select for lower STI resistance in males. This selection is driven by fitness gains to males that acquire the STI and subsequently infect their mates, which offset any costs of infection (e.g. male sterility). This type of adaptive mate harm generates sexual conflict over the optimal level of resistance to STIs. It could partly explain why immune reactions to new infections are weaker in males than females of many species.

Appendix

Calculating fitness for the epidemiological model

Here we calculate the fitness of mutant individuals for the epidemiological model. We assume all other individuals in the population play the same sex-specific strategies. Demographic change is assumed to happen much faster than evolutionary change, so that we can calculate fitness against the background of a stable population as described in the main text. We mark variables that depend on the mutant’s strategies with a star.

Consider first a mutant female. As in our initial simple model, we can partition her lifetime fitness $w$ as the sum of fitness gained while uninfected and while infected [cf. Eq. (3)]. Each term equals her expected number of matings ($n_{U}$ or $n_{I}$) times the expected fitness gain per mating ($v_{U}$ or $v_{I}$):

$$w\left( {x_{U}^{*} ,y_{U}^{*} ,x_{I}^{*} ,y_{I}^{*} } \right) = n_{U} v_{U} + n_{I} v_{I}$$

(20)

As before, the expected fitness gain from mating is $v_{U} = \frac{{x_{U}^{*} }}{{1 + \mu_{U}^{*} }}$ for an uninfected female and $v_{I} = \frac{{x_{I}^{*} }}{{1 + \mu_{I}^{*} }}$ for an infected female. However, the numbers of matings must be adjusted for two reasons. First, they now depend on both reproductive and immune effort. Second, matings with sterile males do not yield viable offspring, so it is more convenient to only tally matings with fertile mates (i.e. the expected numbers of matings $n_{U}$ and $n_{I}$ are defined to only include matings with fertile males). Note that matings with sterile males are still costly, as they consume the same time and reproductive resources as matings with fertile males.

The probability that a female becomes infected during any given mating is $p^{*} = \alpha^{*} \left( {r_{I} + r_{S} } \right)$. The probability that her mate is fertile, given that she remains uninfected after mating, is $\frac{{r_{U} + \left( {1 - \alpha^{*} } \right)r_{I} }}{{r_{U} + \left( {1 - \alpha^{*} } \right)\left( {r_{I} + r_{S} } \right)}}$. The expected number of matings while uninfected is then [cf. Eq. (1)]:

$$n_{U} = \left( {1 - p^{*} } \right)\left( {\frac{{\mu_{U}^{*} + 1}}{{\mu_{U}^{*} + p^{*} }}} \right)\left( {\frac{{r_{U} + \left( {1 - \alpha^{*} } \right)r_{I} }}{{r_{U} + \left( {1 - \alpha^{*} } \right)\left( {r_{I} + r_{S} } \right)}}} \right)$$

(21)

Similarly, when a female is newly infected by her mate, the probability that her mate is fertile is $\frac{{r_{I} }}{{r_{I} + r_{S} }}$. For all subsequent matings while infected, the mate is fertile with probability $r_{U} + r_{I}$. The expected number of matings while infected is consequently [cf. Eq. (2)]:

$$n_{I} = p^{*} \left( {\frac{{\mu_{U}^{*} + 1}}{{\mu_{U}^{*} + p^{*} }}} \right)\left( {\frac{{r_{I} }}{{r_{I} + r_{S} }} + \frac{{r_{U} + r_{I} }}{{\mu_{U}^{*} + \beta^{*} }}} \right)$$

(22)

Male lifetime fitness can be partitioned similarly as:

$$\tilde{w}\left( {\tilde{x}_{U}^{*} ,\tilde{y}_{U}^{*} ,\tilde{x}_{I}^{*} ,\tilde{y}_{}^{*} } \right) = \tilde{n}_{U} \tilde{v}_{U} + \tilde{n}_{I} \tilde{v}_{I}$$

(23)

Unlike the case for females, we count the mating in which a male becomes infected under $\tilde{n}_{U}$ rather than $\tilde{n}_{I}$, because his expected fitness gain from that mating does not change due to his becoming infected (infection only influences his future mating rate). A mutant male can leave the uninfected state in two ways: by dying, which occurs at a rate of $\tilde{\mu }_{U}^{*}$, or by becoming infected, at a rate of $\tilde{p}^{*} = \tilde{\alpha }^{*} \left( {\tilde{r}_{U,I}^{*} + \tilde{r}_{U,S}^{*} } \right)$. While uninfected, he mates with fertile females at a total rate of $\tilde{r}_{U,R}^{*} + \tilde{r}_{U,U}^{*} + \tilde{r}_{U,I}^{*}$. The expected number of matings while uninfected is then:

$$\tilde{n}_{U} = \frac{{\tilde{r}_{U,R}^{*} + \tilde{r}_{U,U}^{*} + \tilde{r}_{U,I}^{*} }}{{\tilde{p}^{*} + \tilde{\mu }_{U}^{*} }}$$

(24)

The probability that the male becomes infected at some point in his lifetime is $\frac{{\tilde{p}^{*} }}{{\tilde{p}^{*} + \tilde{\mu }_{U}^{*} }}$. Once infected, he mates with fertile females at a rate of $\tilde{r}_{I,R}^{*} + \tilde{r}_{I,U}^{*} + \tilde{r}_{I,I}^{*}$. His reproductive life ends when he becomes sterile, at a rate of $\tilde{\beta }^{*}$, or dies, at a rate of $\tilde{\mu }_{I}^{*}$. His expected number of matings while infected is then:

$$\tilde{n}_{I} = \left( {\frac{{\tilde{p}^{*} }}{{\tilde{p}^{*} + \tilde{\mu }_{U}^{*} }}} \right)\left( {\frac{{\tilde{r}_{I,R}^{*} + \tilde{r}_{I,U}^{*} + \tilde{r}_{I,I}^{*} }}{{\tilde{\beta }^{*} + \tilde{\mu }_{I}^{*} }}} \right)$$

(25)

Matings with uninfected females have fitness value $\frac{{x_{U} }}{{1 + \mu_{U} }}$, whereas those with infected females have fitness value $\frac{{x_{I} }}{{1 + \mu_{I} }}$. The average fitness value of mating while uninfected is given by weighting each of these by the proportion of matings with uninfected females ($R$ or $U$) or infected females ($I$) out of all matings with fertile females:

$$\tilde{v}_{U} = \left( {\frac{R + U}{R + U + I}} \right)\left( {\frac{{x_{U} }}{{1 + \mu_{U} }}} \right) + \left( {\frac{I}{R + U + I}} \right)\left( {\frac{{x_{I} }}{{1 + \mu_{I} }}} \right)$$

(26)

If an infected male mates with an uninfected female, he will infect her with probability $\alpha$, after which his expected fitness gain is $\frac{{x_{I} }}{{1 + \mu_{I} }}$. If he does not infect her, or if she was already infected, then the situation is unchanged from Eq. (26). The average fitness value of mating while infected is consequently:

$$\tilde{v}_{I} = \left( {\frac{{\left( {1 - \alpha } \right)\left( {R + U} \right)}}{R + U + I}} \right)\left( {\frac{{x_{U} }}{{1 + \mu_{U} }}} \right) + \left( {\frac{{\alpha \left( {R + U} \right) + I}}{R + U + I}} \right)\left( {\frac{{x_{I} }}{{1 + \mu_{I} }}} \right)$$

(27)