Double Trouble: Multiple infections and the coevolution of virulence and resistance in nested host-parasite populations

Using a nested co-evolutionary model, this study demonstrates that parasite coinfection drives the joint evolution of lower virulence and reduced host resistance through kin selection effects, with the magnitude of these outcomes contingent on the spatial coupling of host populations.

The Gist

The Big Picture: A Game of Musical Chairs with a Twist

Imagine a host (like a human or an animal) is a house. The parasites (bacteria, viruses, or worms) are squatters trying to move in.

Usually, we think of a house having one squatter. But in reality, houses often get multiple squatters at the same time. This paper asks a big question: How does having multiple squatters fighting each other inside the house change the game for both the squatters and the homeowner?

The authors built a computer model to simulate this "nested" world:

1. Inside the house (Within-host): The squatters fight for space and food.
2. Between the houses (Between-host): The squatters try to jump to new houses to start new families.

They wanted to see how this internal fighting changes two things:

• Virulence: How mean the squatters are (how much they damage the house).
• Resistance: How hard the homeowner fights back to kick them out.

---

The Three Ways Squatters Fight

The paper compares three different "rules of engagement" for when two squatters meet in the same house:

1. The "First Come, First Served" Rule (Preemption):

   • Analogy: The first squatter moves in and locks the door. If a second one tries to enter, they are blocked.
   • Result: The squatter just needs to be fast enough to get in first. They don't need to be super aggressive; they just need to be there.

2. The "King of the Hill" Rule (Superinfection):

   • Analogy: The door is unlocked. If a stronger, meaner squatter arrives, they immediately kick the current squatter out and take over the whole house.
   • Result: This creates a "winner-takes-all" scenario. To win, you must be the most aggressive and reproduce the fastest.

3. The "Roommate" Rule (Coinfection):

   • Analogy: Two squatters move in and have to share the living room. They fight, but neither can fully kick the other out immediately. They coexist for a while.
   • Result: This is the paper's main focus. It's a messy middle ground where they have to share resources.

---

The Main Findings: What Happens When They Share?

The researchers found that the "Roommate" rule (Coinfection) changes the outcome dramatically compared to the "King of the Hill" rule.

1. The "Family Dinner" Effect (Kin Selection)

When two squatters are in the same house, they are often closely related (like brothers or cousins from the same "parent" parasite).

• The Logic: If you are a squatter and you see your brother squatter, you don't want to destroy the house too much, because if the house collapses, your brother dies too.
• The Result: In the "Roommate" scenario, the squatters become less aggressive. They realize that being too mean hurts their own family members. This leads to lower virulence (less damage to the host).

2. The Homeowner's Reaction

• In the "King of the Hill" world: The squatters are super aggressive. The homeowner panics and builds a massive, expensive fortress (high resistance) to fight back.
• In the "Roommate" world: Because the squatters are being a bit more polite (less virulent), the homeowner doesn't need to spend as much energy building a fortress. They invest less in resistance.

Summary: When parasites share a host, they act more like a cooperative family and less like a gang. This makes them less dangerous, which in turn makes the host less defensive.

---

The "Travel Cost" Factor (Spatial Coupling)

The paper also looked at how often parasites jump from one house to another (dispersal).

• High Travel (Well-connected world): If parasites jump houses easily, the "King of the Hill" rules dominate. The most aggressive strain wins everywhere.
• Low Travel (Isolated villages): If parasites stay in their local neighborhood, the "Family Dinner" effect is stronger. They are more likely to meet their relatives, so they stay gentle.

The authors found that the more connected the population is, the more aggressive the parasites become. The more isolated they are, the more they cooperate to keep the host alive.

---

Why This Matters in the Real World

This isn't just about squatters; it's about real diseases like malaria, flu, or antibiotic-resistant bacteria.

1. Multiple Infections are Common: In the real world, people often get infected by multiple strains of a virus or bacteria at once.
2. Evolutionary Trap: If we assume parasites are always "King of the Hill" (super aggressive), we might overestimate how dangerous they will be. If they are actually "Roommates," they might evolve to be less deadly.
3. Treatment Implications: Understanding that parasites cooperate when they share a host could help us design better treatments. For example, if we can force them to share a host, maybe they will naturally become less virulent.

The Takeaway in One Sentence

When parasites are forced to share a host, they often "play nice" to avoid killing their own relatives, which makes them less deadly and causes the host to relax its defenses—a delicate balance that depends on how often they travel to new hosts.

Technical Summary

1. Problem Statement

The evolution of host-parasite interactions is driven by complex eco-evolutionary feedback loops. A critical gap in current theoretical epidemiology is the integration of multiple infections (coinfections) into coevolutionary models.

• The Challenge: Most existing models rely on phenomenological assumptions that either ignore multiple infections (single infection/preemption), assume instantaneous competitive exclusion (superinfection), or fail to link within-host dynamics to between-host epidemiological traits.
• The Conflict: Parasite competition within a host creates selective pressures that often oppose between-host selection. Within-host competition typically favors higher replication (and thus higher virulence) to outcompete other strains, while between-host selection often favors intermediate virulence to maximize transmission duration.
• The Goal: To develop a nested co-evolutionary model that explicitly derives epidemiological traits (transmission, virulence, recovery) from within-host interactions between parasite strains, and to analyze how different competition regimes (preemption vs. dominance vs. coinfection) shape the joint evolution of parasite virulence and host resistance.

2. Methodology

The authors employ a multi-scale modeling approach combining metapopulation ecology and adaptive dynamics.

A. Model Structure

1. Between-Host Scale (Metapopulation):

   • The host population is modeled as a Susceptible-Infected-Susceptible (SIS) system reframed as a metapopulation of discrete patches (hosts).
   • Dynamics track the fraction of empty patches, susceptible hosts, and infected hosts.
   • Host demography includes birth (dependent on patch occupancy) and death (background + virulence).

2. Within-Host Scale:

   • A dynamic model tracks parasite density ($X$) based on replication ($r_0\theta$), density-dependent regulation ($r_1X$), and host-induced suppression ($\gamma\sigma$).
   • Traits:
     • Parasite: Investment in host exploitation ($\theta$).
     • Host: Investment in resistance ($\sigma$).
   • The model assumes a fast time-scale separation: within-host dynamics reach equilibrium ($X^*$) much faster than host infection/recovery events.

3. Bridging the Scales:

   • Epidemiological parameters are derived from the equilibrium parasite load ($X^*$):
     • Transmission ($\beta$): Linear function of $X^*$.
     • Virulence ($\alpha$) & Recovery ($\tau$): Non-linear trade-off functions of $X^*$.
   • Costs: Resistance is an inducible cost (only infected hosts pay the cost via reduced fecundity), not constitutive.

B. Competition Regimes

The model simulates three distinct competition scenarios:

1. Preemption (Single Infection): A resident strain prevents reinfection; competition occurs only for new susceptible hosts.
2. Superinfection (Dominance): A secondary strain instantly displaces the resident; competition is resolved by strict hierarchy.
3. Coinfection (Transient Coexistence): Strains coexist within the host for a period. The model tracks the frequency of mutant vs. resident strains and allows for "preemptive effects" where the first colonizer gains an advantage before exclusion occurs.

C. Evolutionary Analysis

• Adaptive Dynamics: The authors calculate invasion fitness ($W$) for rare mutants in a resident population.
• Parasite Fitness: Defined as the average lifetime population success ($R_m$), incorporating both transmission to new hosts and within-host competitive ability.
• Key Metric: The fitness gradient is decomposed to identify the balance between between-host selection (maximizing $R_0$) and within-host selection (maximizing competitive dominance).
• Relatedness ($R$): The model explicitly calculates relatedness between co-infecting strains, introducing a kin selection component.

3. Key Contributions

1. Mechanistic Derivation: Unlike previous phenomenological models, this work derives transmission, virulence, and recovery rates directly from mechanistic within-host growth equations and host resistance traits.
2. Unified Framework: It bridges metapopulation ecology and epidemiology, treating hosts as patches and parasites as dispersing propagules, allowing for the calculation of relatedness and kin selection effects in a standard SIS context.
3. Competition-Colonization Trade-off: The model formalizes a trade-off between maximizing competitive ability within a host (favored by high replication/virulence) and maximizing colonization of new hosts (favored by longer infectious periods/lower virulence).
4. Inducible Resistance Costs: By assuming resistance costs are paid only by infected individuals, the model captures a demographic feedback loop where high infection prevalence reduces the population-wide cost of resistance, altering evolutionary outcomes compared to models with constitutive costs.

4. Key Results

A. Competition Regime Effects

• Superinfection (Dominance): Selects for higher virulence and higher host resistance. Without preemptive effects, the "winner-takes-all" dynamic drives parasites to maximize replication rates, increasing host damage.
• Coinfection (Coexistence): Selects for lower virulence and lower host resistance compared to superinfection.
  • Mechanism: The local coexistence of strains creates a kin selection effect. When related strains co-infect, selection favors "prudent" exploitation to avoid killing the shared host patch too quickly. This dampens the drive for maximum virulence.

B. Coevolutionary Dynamics

• Coevolutionarily Stable Strategies (Co-CSS): The system converges to a single stable strategy for both host and parasite.
• Dispersal and Virulence:
  • In superinfection, high parasite dispersal leads to high virulence and high resistance.
  • In coinfection, the relationship is more complex. High dispersal increases the force of infection, but the kin selection effect and preemptive competition dampen the selection for extreme virulence.
• Recovery Rates: Increased baseline recovery (shorter infectious periods) selects for higher parasite investment and higher host resistance, particularly in superinfection models.

C. Deviation from $R_0$ Optimization

• In simple single-infection models, parasites evolve to maximize the basic reproduction number ($R_0$).
• In this model, multiple infections cause a deviation from $R_0$ optimization.
  • When infection is rare (near viability boundaries), selection reverts to $R_0$ maximization.
  • When infection is common (intermediate dispersal), the competition-colonization trade-off dominates. Parasites are selected to maximize within-host competitive ability (often at the cost of $R_0$), leading to higher virulence than predicted by $R_0$-only models.

D. Host Resistance Patterns

• Contrary to some classical models where resistance peaks at intermediate virulence, this model shows that resistance decreases when the disease is widespread.
• Reasoning: Because resistance costs are induced (paid only by infected hosts), a high prevalence of infection means a large portion of the population is already paying the cost of infection. In this scenario, the marginal benefit of additional resistance is outweighed by the demographic depression caused by reduced fecundity in the infected class.

5. Significance

• Theoretical Advancement: The paper provides a rigorous mathematical framework for understanding how within-host ecology (competition, kin selection) scales up to drive between-host epidemiology.
• Kin Selection in Pathogens: It demonstrates that the genetic relatedness of co-infecting strains is a critical determinant of virulence evolution, supporting the idea that "altruistic" restraint in virulence can evolve if strains are related.
• Ecological Context Dependence: The study highlights that evolutionary outcomes are not static but depend heavily on ecological parameters like dispersal rates, recovery rates, and host demography.
• Implications for Disease Control: The findings suggest that interventions altering transmission rates (dispersal) or recovery rates can have non-linear and counter-intuitive effects on the evolution of virulence and resistance, depending on the underlying competition regime (e.g., whether coinfections are common).

In summary, "Double Trouble" argues that ignoring the nuances of multiple infections and the resulting kin selection effects leads to inaccurate predictions about the evolution of virulence and resistance. The interplay between dominance competition (driving virulence up) and kin selection/preemption (driving virulence down) is the central driver of host-parasite coevolution in this framework.