Differing effects of parasite-parasite interaction types on the spatial epidemiology of co-circulating parasites

By combining spatially-explicit modeling with laboratory experiments on *Paramecium* and bacterial parasites, this study demonstrates that within-host interactions altering host susceptibility have the most significant impact on parasite spatial epidemiology, particularly by amplifying spatial priority effects during sequential invasions.

The Gist

Imagine a bustling city (the host population) where two different types of "invaders" (parasites) are trying to take over. Let's call them Invader A and Invader B.

For a long time, scientists have known that if these two invaders meet inside a single house (a single host), they might fight, ignore each other, or even help one another. But this paper asks a bigger question: How does their relationship change the way they spread across the entire city?

The researchers used a mix of computer simulations and a tiny, real-life laboratory "city" made of test tubes to figure this out. Here is the story of what they found, broken down simply.

1. The Three Ways Invaders Can Interact

The study looked at three main ways Invader B could mess with Invader A's plan to spread:

• The "Door Lock" Effect (Within-Host Interaction): Imagine Invader B moves into a house first and changes the locks. Now, Invader A can't get in, or has a much harder time getting in. This is about changing the host's susceptibility.
• The "Empty House" Effect (Demographic Interaction): Imagine Invader B is so destructive that it kills the homeowners. Now, there are fewer houses left for Invader A to occupy. This is about changing the host population size.
• The "Traffic Jam" Effect (Dispersal Interaction): Imagine Invader B makes the homeowners too sick to travel. Since the invaders travel with the homeowners, Invader A gets stuck in the same neighborhood and can't move to new towns. This is about host movement.

2. The Big Discovery: The "Door Lock" Wins

The researchers found that the "Door Lock" effect was by far the most powerful.

If Invader B makes it hard for Invader A to infect a host (by changing the locks), Invader A's spread across the city slows down dramatically. The city-wide infection rate drops, and the infection becomes patchy—some neighborhoods are full of Invader A, while others are completely empty.

The other two effects (killing homeowners or stopping them from traveling) had much smaller impacts. It turns out, preventing the invader from getting inside the house in the first place is a much bigger deal than just making the house smaller or stopping the owner from moving.

3. The "First-Come, First-Served" Rule (Priority Effects)

This is the most fascinating part. The timing of the invasion matters immensely.

• Scenario A: The Race (Simultaneous Arrival)
  If Invader A and Invader B arrive in the city at the exact same time, they fight it out. Invader A might still win some battles, but it's a messy, chaotic fight.
• Scenario B: The Takeover (Sequential Arrival)
  If Invader B arrives first and settles into the houses, it effectively "claims" the territory. When Invader A arrives later, it finds the doors locked. The study found that if Invader B gets there first, Invader A often fails to spread at all.

This is called a Spatial Priority Effect. It's like a game of musical chairs where the first person to sit down gets to keep the seat, and the latecomers are left standing. If the "bad" parasite gets there first, it can block the "new" parasite from ever taking hold in that area.

4. The Real-Life Experiment

To prove this wasn't just a computer fantasy, the scientists built a mini-city in the lab.

• The City: A line of 5 test tubes connected by rubber tubing.
• The Hosts: Tiny single-celled organisms called Paramecium (like microscopic swimmers).
• The Invaders: Two types of bacteria (Holospora undulata and Holospora obtusa).

They set up the tubes and let the bacteria swim from one tube to the next.

• Result: When the "blocking" bacteria (H. obtusa) was present, the "target" bacteria (H. undulata) moved much slower. It often got stuck in the first few tubes and never made it to the end of the line. When the blocking bacteria wasn't there, the target bacteria zoomed through the whole chain.

The Takeaway

This paper teaches us that who arrives first matters just as much as who is stronger.

In the real world, this helps us understand diseases. If we want to stop a new disease from spreading, knowing that existing parasites can act as a "shield" (by locking the doors) or a "barrier" (by arriving first) is crucial. It suggests that the history of infections in a population—what's already there and when it got there—determines how fast and how far a new disease can travel.

In short: Parasites don't just fight inside a body; they fight over the map. And the one who gets there first often gets to draw the borders.

Technical Summary

1. Problem Statement

While it is well-established that co-circulating parasite species interact within individual hosts (altering susceptibility, virulence, or transmission), the spatial consequences of these interactions at the landscape level remain poorly understood. Existing literature has largely focused on:

• Within-host dynamics: How co-infection affects individual health.
• Static spatial overlap: Observational studies identifying hotspots of co-occurrence.
• Limited theoretical scope: Previous models often focused on specific interaction types (e.g., cross-immunity) without systematically comparing how different mechanisms (demographic vs. dispersal vs. susceptibility) scale up to influence disease spread across a landscape.

The authors aim to bridge this gap by determining how distinct types of parasite-parasite interactions influence the spatiotemporal dynamics (prevalence, heterogeneity, and rate of spread) of co-circulating parasites.

2. Methodology

The study employs a dual approach combining stochastic spatially-explicit modeling with controlled laboratory experiments.

A. Epidemiological Modeling

• Framework: A stochastic metapopulation model was developed for a linear chain of 20 connected patches.
• Host Dynamics: Follows a Susceptible-Infected-Susceptible (SIS) framework with an additional class for co-infected hosts ($I_C$). Host births are density-dependent; mortality is increased by infection.
• Parasite Species: Two species were modeled: a "background" parasite ($B$) and a "focal" parasite ($F$).
• Interaction Scenarios: The model tested three distinct interaction mechanisms where the background parasite affects the focal parasite:
  1. Within-host interaction: Altering host susceptibility to the focal parasite (parameter $\beta'_{BF}$).
  2. Demographic interaction: Increasing host mortality due to the background parasite (parameter $\mu'_B$).
  3. Dispersal interaction: Reducing host movement probability due to the background parasite (parameter $p'_{move,B}$).
• Epidemiological Scenarios: Simulations were run under two invasion timings:
  1. Simultaneous: Both parasites introduced at $t=0$.
  2. Sequential (Delayed): The focal parasite introduced after the background parasite reached a steady state ($t=200$).
• Metrics: The study measured landscape-wide prevalence, spatial heterogeneity (Coefficient of Variation), and the time to reach the final patch (rate of spread).

B. Experimental Validation

• System: A microcosm system using the protist host Paramecium caudatum and two bacterial parasites: Holospora undulata (focal) and H. obtusa (background).
• Setup: Linear landscapes of 5 interconnected microcosms (tubes).
• Treatments:
  1. Single-parasite: Only H. undulata introduced.
  2. Mixed-parasite: Both H. undulata and H. obtusa introduced simultaneously.
• Procedure: Hosts were allowed to disperse between patches for 3 hours, 3 times a week, over 4 weeks. Prevalence and density were tracked.
• Analysis: Linear mixed models (LMM) and Generalized Linear Mixed Models (GLMM) were used to analyze spread rates and final prevalence.

3. Key Contributions

• Mechanistic Differentiation: The study is the first to systematically compare how different types of parasite interactions (susceptibility vs. mortality vs. dispersal) scale up to affect spatial epidemiology.
• Identification of Dominant Mechanisms: It demonstrates that within-host interactions (specifically those altering host susceptibility) have a disproportionately large impact on spatial dynamics compared to demographic or dispersal-mediated interactions.
• Spatial Priority Effects: The research highlights that the timing of invasion is critical. Sequential invasion (where one parasite establishes before the other) amplifies interaction effects, leading to strong "spatial priority effects" where the resident parasite blocks the invader.
• Theory-Experiment Integration: It provides robust empirical validation for theoretical predictions regarding parasite co-circulation in spatially structured environments.

4. Key Results

Modeling Results

• Within-Host Interactions ($\beta'_{BF}$): This had the strongest effect.
  • Prevalence: When the background parasite reduced susceptibility ($\beta'_{BF} < 1$), focal parasite prevalence dropped drastically (from ~65% to ~2% in extreme antagonism).
  • Spread Rate: Strong antagonism significantly slowed the spread of the focal parasite.
  • Priority Effects: In sequential scenarios, if the background parasite established first, the focal parasite often failed to reach the end of the network entirely (0% success in sequential vs. 20% in simultaneous at extreme antagonism).
• Demographic & Dispersal Interactions:
  • These had much weaker effects on overall prevalence and spatial heterogeneity compared to within-host interactions.
  • They did slightly slow the rate of spread, but only in sequential invasion scenarios. In simultaneous invasions, these interactions had negligible impact on spread rates.
• Heterogeneity: Strong within-host antagonism increased spatial heterogeneity (patch-to-patch variation) in prevalence.

Experimental Results

• Validation of Model: The experiments confirmed the model's predictions.
  • Spread Rate: The presence of H. obtusa significantly slowed the spatial spread of H. undulata. In mixed treatments, the focal parasite failed to colonize the most distant patches in 50% of replicates, whereas it succeeded in nearly all single-parasite replicates.
  • Prevalence: Landscape-wide prevalence of the focal parasite was reduced by nearly half (0.24 vs. 0.43) in the presence of the background parasite.
  • Mechanism Confirmation: Preliminary data confirmed that H. obtusa reduces the infection success of H. undulata (supporting the $\beta'_{BF} < 1$ mechanism).
• Stochasticity: The study highlighted that stochastic effects (random dispersal events) play a major role. Small population sizes of the focal parasite, caused by suppression from the background parasite, increased the probability of local extinction and failure to colonize downstream patches.

5. Significance

• Disease Management: The findings suggest that managing the timing of parasite introductions (or the order of arrival in natural systems) is as critical as the nature of the interaction itself. Early establishment of a dominant parasite can effectively block the spread of a secondary pathogen via spatial priority effects.
• Predictive Modeling: Future models of disease spread must prioritize within-host susceptibility interactions over demographic or dispersal interactions when predicting landscape-scale outcomes, as the former drives the most significant changes in prevalence and spread.
• Conservation and Public Health: Understanding these spatial dynamics is crucial for managing co-circulating pathogens in wildlife conservation (e.g., preventing invasive parasites) and public health (e.g., managing vector-borne diseases where cross-immunity or competition occurs).
• Stochasticity in Metapopulations: The study underscores that in spatially structured populations, stochastic extinction events driven by reduced population sizes (due to interspecific competition) can halt disease spread even if the basic reproduction number ($R_0$) theoretically supports it.

In conclusion, the paper establishes that parasite-parasite interactions are not just biological phenomena but spatial drivers, with within-host susceptibility changes acting as the primary lever for altering the geography of disease spread, particularly when invasion timing creates priority effects.