Decoupling of resistance and tolerance against one of two related parasites (Eimeria) in mice

Resistance (the host’s capacity to reduce parasite burden) and tolerance (the host’s capacity to reduce impact on host health of a given parasite burden) manifest two different lines of immune defenses. In some host-parasite systems these two defenses are balanced against each other, while in others they are uncoupled. In hybrid hosts, resistance has sometimes been interpreted as having an effect on fitness without considering the modulatory effect of tolerance. Here, we used two closely related parasite species of genus Eimeria and measured proxies for resistance and tolerance in four wild-derived strains of inbred mice from two subspecies during controlled infection. We found a negative correlation between resistance and tolerance against E.  falciformis, while the two are uncoupled against E.  ferrisi. This might be explained by trade-offs, as resistance limits infection load and thereby the scope of possible tolerance, and both resistance and tolerance can be costly in terms of resource allocation. Resistance can be assumed to be limited by immunopathogenicity, tolerance by carrying capacity of the host or energy drained by the parasite. Findings of resistance in natural populations of hybrid mice have to be interpreted carefully in this context. Resistance and tolerance have to be studied in conjunction.


Introduction
Host defense mechanisms evolving in response to feedback between hosts and parasites can be categorised into two components: resistance and tolerance (Little, Shuker, Colegrave, Day, & Graham, 2010). Resistance (the ability of a host to reduce its parasite burden) results from defense against parasite infection or proliferation early after infection (Råberg, Graham, & Read, 2009). Resistance can be energetically costly and therefore limited by resource allocation as measured by a decrease of other fitness components (e.g. delayed maturity, lower fecundity) in the absence of infection (Langand, Jourdane, Coustau, Delay, & Morand, 1998;Sheldon & Verhulst, 1996;Vijendravarma, Kraaijeveld, & Godfray, 2009). Additionally, too strong immune response against pathogens can lead to a negative impact on health or immunopathology (Graham, Allen, & Read, 2005). Tolerance balances damage caused by parasites themselves and immunopathology (Medzhitov, Schneider, & Soares, 2012) through control mechanisms like stress response, damage repair and cellular regeneration (Soares, Teixeira, & Moita, 2017). This is why, just like resistance, tolerance can involve energetic costs (Simms & Triplett, 1994). In natural populations, costs of the two lines of defense against parasites predict that resistance and tolerance are negatively correlated (Råberg, 2014;Råberg, Sim, & Read, 2007). They can also be found uncoupled if they are at intermediate levels (Athanasiadou, Tolossa, Debela, Tolera, & Houdijk., 2015). As resistance alone is not an estimator of parasite impact on health, understanding how resistance and tolerance are coupled is necessary to conclude on health effects of parasitism.
The house mouse subspecies Mus musculus musculus and M. m. domesticus (hereafter Mmm and Mmd, respectively), whose genomes diverged some 0.5 million years ago, hybridize in a secondary contact zone running through Europe (Boursot, Auffray, Britton-Davidian, & Bonhomme, 1993;Duvaux, Belkhir, Boulesteix, & Boursot,2011). Hybrids show elevated resistance to parasites compared to both parental subspecies Balard et al., 2019). Newly generated diversity in the immune system can result in novel interplay in immunological response; interpretations of these results in terms of health or even fitness effects, however, have been attempted (Sage, Heyneman, Lim, & Wilson, 1986) and criticised .
Given differences in pathogenicity and prevalence between the two Eimeria species we suspected that coupling between resistance and tolerance might differ. We assessed this experimentally in controlled infections of Mmm and Mmd. We employed four wild-derived inbred strains representing the two mouse subspecies and assessed the symptoms both at the level of host subspecies and inbred strains.

Parasite strains
The three parasite isolates used in this study were isolated from feces of mice captured in a house mouse hybrid zone (HMHZ) running through Brandenburg, Germany ), in 2016(capture permit No. 2347/35/2014. They belong to both the most prevalent Eimeria species in the wild (Jarquín-Díaz, Balard, Mácová, et al., 2019), namely E. ferrisi (isolates Brandenburg64 and Brandenburg139) and E. falciformis (isolate Brandenburg88).
Hybrid index (HI) of each individual wild-caught mouse was calculated to account for the admixture of mouse genomes across the HMHZ as a proportion of Mmm alleles in a set of 14 diagnostic markers . Isolate Brandenburg64 was isolated in a 92% Mmd individual (HI = 0.08), isolate Brandenburg139 in a 85% Mmm (HI = 0.85) and isolate Brandenburg88 in a 80% Mmd (HI = 0.2). Pre-patency and the peak day of parasite shedding for these isolates were estimated during infection in NMRI laboratory mice (Al-khlifeh et al.,

Mouse strains
We used four wild-derived fully-inbred mouse strains: two representing Mmd: SCHUNT  Parasites of the Eimeria genus are known to induce host immune protection against reinfection (Rose, Hesketh, & Wakelin, 1992;Smith & Hayday, 2000). To ensure that our mice were Eimeria-naive, mice fecal samples were tested before infection for the presence of Eimeria spp. oocysts, by flotation in saturated NaCl solution followed by washing and observation under light microscope. All individuals were negative for Eimeria at the beginning of our experiment.

Experimental infection
Mice were kept in individual cages during infection. Water and food (SNIFF, Rat/Mouse maintenance feed 10 mm) were provided ad libitum supplemented with 1 g of sunflower and barley seeds per day. Mice were orally infected with 150 sporulated oocysts of one Eimeria isolate suspended in 100 µl phosphate-buffer saline (PBS) and monitored daily until their sacrifice by cervical dislocation at 11 days after infection (dpi) (experiment license Reg. 0431/17). Individuals presenting severe health deficiency and/or a weight loss approaching 18% relative to their starting weight were sacrificed earlier. Weight was recorded and feces collected on a daily basis. Fecal pellets were collected every day from each individual cage and suspended in 2% potassium dichromate. Parasite oocysts were recovered using NaCl flotation (see above). In total, 108 mice were infected. Mice were randomly allocated to experimental groups ensuring homogeneous distribution of ages and sexes between groups.
Our experiments were conducted in four consecutive batches for easy handling. The first two groups were infected by the two E. ferrisi isolates (Brandenburg64 and Brandenburg139), the two second by one E. ferrisi isolate (Brandenburg64) and one E. falciformis isolate (Brandenburg88). Summarised experiment design is shown in Table 1 We observed Eimeria oocysts in the feces of 9 mice belonging to the last experimental batch at the day of infection, likely due to cross-contamination between batches. Moreover, before arrival to the infection facility, nematode eggs were observed in flotated feces of mice belonging to all genotypes. Nematode infection is common in breeding facilities (Baker, 1998).

Choice of measurements for resistance and tolerance
Resistance is the capacity of a host to reduce its parasite burden, therefore it is usually estimated by the inverse of infection intensity (Råberg et al., 2009). As a proxy we used the number of oocysts per gram of feces (OPG) at the day of maximal shedding. This measure is tightly correlated with the sum of oocysts shed throughout the experiment (Pearson Tolerance is usually defined as the slope of the regression of host fitness, approximated by health condition, on infection intensity (Råberg, 2014). The major measurable symptom in murine Eimeria infections is weight loss. Therefore, the impact of parasites on host health was measured as the maximum relative weight loss compared to day 0 (body weight taken at the start of the experimental infection).
We defined a tolerance index for each individual, describing how its health varied with infection intensity, between day 0 of infection (weight = 100%, parasite intensity = 0 oocyst per mouse gram) and highest impact (weight = maximum weight loss relative to day 0, parasite intensity = maximum parasite number per gram of feces). This index was then standardised by log10 transformation, after addition of 1e-8 to the ratio to avoid infinite values. The obtained log10 transformed ratio that ranged between -8 (high tolerant) and -5.6 (low tolerant) was divided by the negative constant -8 to obtained a final index positively correlated with tolerance: Tolerance index = (log10(maximum relative weight loss / maximum number of oocysts per gram of feces + 1e-8) / -8

Statistical analyses
Appropriate distribution for maximum number of oocysts per gram of feces, maximum weight loss relative to day 0, and tolerance index were selected based on log likelihood and AIC criteria and by comparing goodness-of-fits plots (density, CDF, Q-Q, P-P plots) between usual distributions (R packages MASS (Venables & Ripley, 2002)  We then compared the coupling between proxies of resistance and tolerance between mouse subspecies. Using the resistance index and tolerance index defined above, we fitted a linear model to explain the variation of tolerance with resistance, Eimeria species and their interaction.
To All analyses were performed using the R software version 3.5.2 (R Development Core Team, 2018). Graphics were produced using the R package ggplot2 (Wickham, 2016) and compiled using the free software inkscape (https://inkscape.org).
All codes and data used for this article can be found at: https://github.com/alicebalard/Article_RelatedParasitesResTol

General parasitology
The life cycle of all isolates was successfully completed in all mouse strains (Figure 2). For E. ferrisi (both isolates), the pre-patent period was 5 dpi and the median day of maximal oocyst shedding was 6 dpi (standard deviation sd=0.73 and 0.61, respectively). The median day of maximum weight loss was 5 dpi for both isolates (sd=2.1 and 1.9 respectively). For E. falciformis (isolate Brandenburg88) pre-patency was 7 dpi, median day of maximal shedding was 8 dpi (sd=1.2) and median day of maximal weight loss 9 dpi (sd=1.5). All tested Eimeria isolates infected all individuals of the tested mouse strains.
A considerable number of Mmm mice (8/14; 5 of BUSNA and 3 of PWD) infected with E. falciformis (isolate Brandenburg88) died (or had to be sacrificed at humane end points specified in animal experimental procedures) before the peak of oocyst shedding. Moreover, one Mmd mouse (strain SCHUNT) infected by E. ferrisi isolate Brandenburg139 had liquid diarrhea in the peak shedding day, making its feces not collectable. These mice were assessed as missing data for both resistance and following tolerance measurements.

Figure 2. Parasite density (A) and relative weight loss (B) during Eimeria infection. Parasite density is
calculated as number of oocysts detected (x10e6) per gram of feces, relative weight loss is calculated compared to day 0. Mean and 95% CI are plotted for each parasite isolate. All hosts strains are pooled together.

M. m. domesticus is less resistant to E. falciformis than to E. ferrisi
To establish differences between the two house mouse subspecies and between the parasite species, we analysed the maximum number of oocysts per gram of feces (OPG) as a measure of resistance after infection with both Eimeria species on the 99 mice alive by the time of median peak shedding. We found statistically significant effects of parasite species (LRT: G = 18.9, df = 2, P < 0.001), mouse subspecies (LRT: G = 16.2, df = 2, P < 0.001) as well as an interaction between parasite species and mouse subspecies (LRT: G = 10.2, df = 1, P < 0.01). Post-hoc multiple comparison tests showed than the subspecies Mmd shed more OPG at the peak of shedding when infected with E. falciformis than with E. ferrisi.
Moreover, the Mmm subspecies shed more OPG at the peak of shedding than the Mmd subspecies when infected with E. ferrisi (Table 2; Figure 3A). Figure 3A.  Table 2); (B) Impact on host health measured as the maximum weight loss during patent period relative to starting weight (%) (see Table 4); (C) Tolerance index measured as (log10(maximum relative weight loss / maximum number of oocysts per gram of feces + 1e-8) / -8 (see Table 6) We then tested the influence of mouse strain and parasite isolate on maximum number of OPG. We found statistically significant effects of parasite isolate (LRT: G = 35.5, df = 8, P < 0.001), mouse strain (LRT: G = 36.3, df = 9, P < 0.001) as well as an interaction between parasite isolate and mouse strain (LRT: G = 21.8, df = 6, P < 0.01). We found no significant difference between the strains of the same subspecies within a given parasite species infection, nor between the isolates of the same species within a given mouse strain (Table 3; Figure 4A).

Table 3. Post-hoc statistical test for maximum oocyts per gram of feces (Tukey Multiple Comparisons of Means) between each mouse strain
and parasite isolate. See Figure 4A. (A) Maximum oocysts per gram of feces used as a proxy for (inverse of) resistance (see Table 3); (B) Impact on host health measured as the maximum weight loss during patent period relative to starting weight (%) (see Table   5); (C) Tolerance index measured as (log10(maximum relative weight loss / maximum number of oocysts per gram of feces + 1e-8) / -8 (see Table 7

E. ferrisi
Analysing the weight loss upon infection as a proxy for impact on host health of the full dataset (N = 108), we found statistically significant differences both between the mouse subspecies (LRT: G = 10, df = 2, P < 0.01) and between the parasite species (LRT: G = 18.6, df = 2, P < 0.001). Post-hoc multiple comparison tests showed that Mmd lost less weight than Mmm when infected by E. falciformis (9.3% vs 18.7%), and Mmm lost more weight when infected by E. falciformis than by E. ferrisi (Table 4, Figure 3B).

Then we modelled maximum weight loss separating Eimeria by isolates (instead of species)
and mice by strains (instead of subspecies). We found differences between parasite isolates (LRT: G = 30.7, df = 8, P < 0.001) and mouse strains (LRT: G = 23, df = 9, P < 0.01). Notably, PWD (Mmm) mice infected with Brandenburg64 (E. ferrisi) lost significantly more weight than STRA mice (Mmd) infected with the same isolate, following the pattern described at the mouse subspecies-parasite species level (Mmd losing less weight than Mmm when infected by E. ferrisi). Overall, we did not find any significant difference between mouse strains of the same subspecies within a given parasite species infection or between parasite isolates of the same species within a given mouse strain (Table 5; Figure 4B).

M. m. musculus is less tolerant to E. falciformis than to E. ferrisi
Combining resistance and the impact on health, we assessed tolerance of the mouse subspecies to both Eimeria species on the 99 mice for which this index could be calculated.
Post-hoc multiple comparison tests showed that tolerance index upon infection with E. falciformis was higher for Mmd than for Mmm. Within Mmm subspecies, animals had a lower tolerance index to E. falciformis than to E. ferrisi (Table 6; Figure 3C).Testing at the level of Eimeria isolates and mouse strains, we found between-parasite isolates differences (LRT: G = 24.1, df = 8, P < 0.01), between-mouse strains differences (LRT: G = 35.1, df = 9, P < 0.001), and interaction between the two factors (LRT: G = 19.9, df = 6, P < 0.01). Post-hoc multiple comparison tests showed statistically significant differences between PWD infected with Brandenburg88 against both Mmd strains infected with the same parasite isolate , and against animals of the same strain infected with Brandenburg64 Overall, we did not find any significant difference in tolerance between mouse strains of the same subspecies within a given parasite species infection, or between parasite isolates of the same species within a given mouse strain (Table 7; Figure 4C). These results indicate that the lower tolerance of Mmm to E. falciformis compared to Mmd is consistent at the strain level.

Coupling of resistance and tolerance differs between Eimeria species
To test coupling between resistance and tolerance of the mouse subspecies within each of the parasite species, we plotted the mean maximum relative weight loss on maximum oocysts per gram of feces (Figure 5). High tolerance to a given parasite species means that the weight is lowly affected even in case of high parasite load, which corresponds to the lower right corner of the plot, and inversely the upper left corner represents low tolerance. We see that for E. falciformis, there is a trade-off between resistance and tolerance, with high tolerance-low resistance for Mmd, and high resistance-low tolerance for Mmm. In the case of E. ferrisi, resistance varies between both mouse subspecies, but tolerance does not vary consequently, showing a lack of coupling between resistance and tolerance for this parasite.

Discussion
In this study, we used a controlled infection experiment to test whether two closely related parasites differ in their impact on their hosts -house mice. We found tolerance to be decoupled from resistance against E. ferrisi. The two types of responses against E. falciformis were negatively correlated, suggesting a trade-off between resistance and tolerance for this parasite. While resistance decreases parasite fitness and prevalence in natural populations, tolerance generally has no impact on parasite fitness and either increases or does not affect prevalence (Miller, White, & Boots, 2005;Roy & Kirchner, 2000). This allows speculation on host-parasite co-evolution for both Eimeria species.
E. ferrisi commits to sexual reproduction after a relatively short time with few cycles of asexual expansion (Al-khlifeh et al., 2019;Ankrom et al., 1975). As E. ferrisi infections do not reach extremely high intensities with this infection strategy, high tolerance might be the optimal strategy for both house mouse subspecies. Resistance could then evolve relatively freely without any major impact of the parasite on the hosts' health. Enhanced virulence (reduction of host fitness upon infection e.g. due to prolonged asexual replication before commitment to sexual replication and transmission) might not evolve because the low resistance of the host already allows an optimal transmission rate, especially considering the fast production of transmission stages (Anderson & May, 1982). A global optimum of high tolerance might also be the reason why no subspecies-specific adaptation of Mmd or Mmm infecting strains, i.e. increased tolerance of matching host-parasite pairs, could be detected in this parasite species.
E. falciformis has a relatively long life cycle (Al-khlifeh et al., 2019;Haberkorn, 1970). This means that parasites multiply asexually for a relative long time leading to potentially higher tissue loads and -once it starts to reproduce sexually -extremely high reproductive output in strongly impacted hosts. Therefore, tolerance of this parasite might, on the one hand, lead to prohibitively high intensities if the parasite is allowed to expand asexually and damage the tissue (Ehret et al., 2017) without enough resistance. On the other hand, immunopathology has been observed in advanced E. falciformis infections. For example, proinflammatory T cell mediators have been shown to decrease parasite load but increase body weight loss upon infection (Stange et al., 2012). This might lead to multiple different optima for resistance and tolerance (Råberg et al., 2007). In this context, two alternative response strategies against E. falciformis might have evolved and stabilised in the house mouse subspecies: while Mmm rather resists E. falciformis, Mmd tends to tolerate it. Instead of such more or less stable optima in the two mouse subspecies we could speculate two related alternative explanations. Firstly, E. falciformis could originally be a Mmd parasite dissipated into Mmm territory by a spillover through the hybrid zone. As an argument against this explanation, no significant difference in E. falciformis prevalence at each side of the hybrid zone have be observed (unpublished data). Secondly, the E. falciformis isolate Brandenburg88 employed here was taken close to the hybrid zone center but from a predominantly Mmd mouse (hybrid index 0.2). The isolate could hence be adapted to Mmd.
Experiments with an additional E. falciformis isolate from Mmm are needed to answer the question whether host subspecies adaptation can lead to tolerance in matching pairs of E. falciformis and mouse subspecies.
Resistance and tolerance to parasites are highly relevant to the house mouse hybrid zone. As a so-called tension zone, this zone is maintained by a balance between dispersal and endogenous selection against hybrids (Barton & Hewitt, 1985;Macholán et al., 2007;Payseur, Krenz, & Nachman, 2004;Raufaste et al., 2005). It has been shown that hybrid mice are more resistant not only to Eimeria but also to other parasites including pinworms Balard et al., 2019). Impact on tolerance could not be measured under natural conditions . The effect of parasites on hosts' fitness in particular and the role they can play in the evolution of species barriers is thus still rather ambiguous. We here show that it is indispensable to measure both resistance and tolerance in Eimeria infections of house mice. Such measurements can be made in future laboratory experiments involving hybrid mice.
The contrast between two different Eimeria spp. invites future research on the relationship between infection intensity, parasite reproductive output, host health and immune response.
This might allow us to better understand both the process and mechanisms of the evolution of tolerance and resistance in the context of hybrid hosts and beyond.

Funding
This work was funded by the German Research Foundation (DFG) Grant [HE 7320/1-1] to EH. VHJ is an associated student of GRK 2046 funded by the DFG. The maintenance of wildderived strains was supported by the ROSE program from Czech Academy of Sciences and the Czech Science Foundation (project 16-23773S) to JP.