Evolution of host resistance and damage-limiting mechanisms to an emerging bacterial pathogen

Understanding how hosts minimise the cost of emerging infections has fundamental implications for epidemiological dynamics and the evolution of pathogen virulence. Despite this, few experimental studies conducted in natural populations have explicitly tested whether hosts evolve resistance, which prevents infections or reduces pathogen load through immune activation, or tolerance, which limits somatic damages without decreasing pathogen load. In addition, none have done so controlling for the virulence of the pathogen isolate used, despite critical effects on host responses to infection. Here, we conducted an experimental inoculation study to test whether eastern North American house finches (Haemorrhous mexicanus) have evolved resistance or tolerance to their emerging bacterial pathogen, Mycoplasma gallisepticum, using 55 distinct isolates of varying virulence. First, we show that peak pathogen loads, which occurred around 8 days post-inoculation, did not differ between experimentally inoculated finches from disease-exposed (eastern) versus unexposed (western) population. However, pathogen loads subsequently decreased faster and to a greater extent in finches from exposed populations, indicating that they were able to clear the infection through adaptive immune processes. Second, we found no between-population difference in the regression of clinical symptom severity on pathogen load; if tolerance had evolved then the slope of this regression is predicted to be shallower (less negative) in the exposed population. However, finches from exposed populations displayed lower symptom severity for a given pathogen load, suggesting that damage-limitation mechanisms have accompanied the evolution of immune clearance. These observations show that resistance and damage-limitation mechanisms - including, but not limited to the standard conceptualisation of tolerance - should not be seen as mutually exclusive. Nevertheless, we propose that host resistance is especially likely to evolve in response to pathogens such as M. gallisepticum that require virulence for successful infection and transmission.


Introduction
Hosts can alleviate the costs of infection by evolving one of two distinct -though not 47 necessarily mutually exclusive -strategies [1,2]. They can evolve resistance, which serves to 48 reduce the establishment of infectious pathogens and/or to clear pathogens following 49 establishment [3,4]. Alternatively, hosts can evolve tolerance, which serves to mitigate 50 collateral somatic damage caused by the infection without reducing pathogen load [5][6][7]. 51 Whether and when hosts evolve resistance or tolerance in response to emerging pathogens 52 has far-reaching consequences for predicting virulence evolution and epidemiological  Hypotheses based on the evolution of resistance versus tolerance make two contrasting 57 predictions. First, if resistance (but not tolerance) has evolved, hosts will display reduced 58 pathogen load during an infection relative to non-evolved hosts [8,11,12]. Second, if 59 tolerance (but not resistance) has evolved, hosts will have a shallower regression of clinical 60 symptom severity on pathogen load, since tolerant hosts are better able to mitigate the impact 61 of an increasing pathogen load [12]. However, given these contrasting predictions, it is 62 important to note that damage-limitation mechanisms could evolve in conjunction with 63 resistance (e.g. by limiting immunity or initiating repair) [13,14]. Thus, resistance and 64 tolerance-mechanisms need not be mutually exclusive, and evidence for the evolution of one 65 is not necessarily evidence against evolution of the other. 66 There have been few experimental tests of the predictions for the evolution of 67 resistance and tolerance in response to naturally emerging pathogens, and the handful of 68 studies to date have also yielded rather mixed conclusions. For example, strong evidence for 69 the evolution of resistance comes from observations of the epidemic of myxoma virus in 70 European rabbits (Oryctolagus cuniculus) in Australia [15,16]. Following initially dramatic 71 population declines, at 7 years post-outbreak rabbits from disease-exposed populations 72 displayed mortality rates of only ~25% in response to experimental infection. In contrast 73 mortality rates of over 88% were found in unexposed wild and domestic rabbits [17,18] here we use a greater number of host individuals and of pathogen isolates of varying levels of 98 virulence to reduce the possibility of type 2 error. 99 We conducted a large-scale infection experiment using 108 naïve house finches from 100 disease-exposed (N=51) and unexposed (N=57) populations, and 55 bacterial isolates  that house finches from exposed populations displayed less severe symptoms than those from 108 unexposed populations [31]. In this study, we test the key contrasting predictions set out 109 above to determine whether this host evolutionary response is principally attributable to 110 changes in resistance or tolerance.

111
First, if finches from exposed populations have evolved resistance, we would expect 112 them to display lower pathogen loads during infection than birds from unexposed populations 113 (i.e. populations that have not evolved resistance). Specifically, because resistance to M. 114 gallisepticum is thought to be mediated through the ability to mount a cell-mediated immune 115 response [32], given evolved resistance, finches from exposed populations are expected to 116 show reduced pathogen load from approximately 2 weeks post-infection (i.e. the time 117 required to mount a pathogen-specific immune response). By contrast, if tolerance alone has 118 evolved, we predict no differences in pathogen load over the course of the month-long infection experiment, but predict that the relationship between symptom severity and 120 pathogen load will be shallower in birds from exposed populations [12].

123
Over the course of the experiment, the median peak bacterial load observed across the 108  Figure 1). Overall, however, there was 131 no directional pattern and average peak load was similar in birds from exposed and 132 unexposed populations (mixed GLM; population effect (unexposed relative to exposed)  se 133 = -0.10  0.29, χ 2 = 0.11, df = 1, p =0.74).

134
The absence of a population difference in peak load is not necessarily incompatible 135 with evolved resistance if pathogen loads are peaking prior to the time when genetically 136 resistant birds are able to mount an effective immune response. Here bacterial loads were 137 highest (on average) in birds from both populations at 8 days post-inoculation (dpi) and 138 thereafter declined significantly (mixed GLM; main dpi effect: estimate  se = -0.07  0.009, 139 χ 2 = 54.7, df = 1, p < 0.0001). However, there was a significant population by dpi interaction 140 (estimate  se = 0.07  0.02, χ 2 = 13.9, df = 1, p < 0.0002), with birds from exposed 141 populations clearing the pathogen approximately 3 times faster than those from unexposed 142 populations ( Figure 2). Differential clearing rates are such that birds from exposed 143 populations have a 4-fold lower bacterial load than birds from unexposed populations by 28 144 dpi. This finding is consistent with the hypothesis that genetic resistance through acquired 145 immunity has evolved in exposed populations.

146
The key symptom of M. gallisepticum infection in house finches is conjunctivitis, 147 which, when severe, causes blindness and death in the wild through starvation or predation 24.4, df = 1, p < 0.0001), but the slope of this regression did not differ between birds from 155 exposed vs. unexposed populations (population × integral of pathogen load interaction effect: 156 estimate  se = -0.03  0.08, χ 2 = 0.4, df = 1, p = 0.73), as predicted under the hypothesis that 157 tolerance alone had evolved ( Figure 3). Nonetheless, birds from exposed populations have 158 10% lower clinical symptoms for a given pathogen load (population main effect: estimate  159 se = 0.21  0.04, χ 2 = 6.3, df = 1, p = 0.012). Thus, while there is no evidence that (slope) 160 tolerance differs between populations, our results do suggest that some mechanism(s) to limit 161 immune damage have evolved in tandem with resistance. Note that results were qualitatively 162 equivalent when using peak pathogen load rather than integral of pathogen load (see  To test whether house finch hosts from disease-exposed populations have evolved 167 resistance or tolerance to infection to the emerging bacterial pathogen M. gallisepticum, we 168 conducted an inoculation experiment of house finches from disease-unexposed and exposed 169 populations using 55 isolates collected over a 20-year period from epidemic outbreak. We 170 found that birds from exposed and unexposed populations had comparable peak pathogen 171 loads, which were maximal at 8 days post-inoculation in both cases. However, thereafter 172 birds from previously exposed populations cleared the pathogen more rapidly and to a greater 173 extent during our experiment. We interpret these patterns as clear evidence that, in the 174 exposed populations, hosts have evolved resistance in response to the emerging pathogen M. 175 gallisepticum. 176 In contrast to the evidence supporting the hypothesis of evolved resistance, evidence for 177 the evolution of tolerance in the exposed finch population was equivocal. Notably, the 178 gradient of the regression of symptom severity on pathogen load was comparable in birds 179 from exposed and unexposed populations. Because we did not observe the predicted 180 differences in the response slopes of exposed vs. unexposed host populations, our results are 181 not consistent with the hypothesis that tolerance to M. gallisepticum has evolved.

182
Nonetheless, birds from exposed populations did exhibit lower clinical symptoms across the 183 range of pathogen loads experienced (i.e. the intercept of the regression line is lower than in 184 unexposed populations). This pattern suggests that some mechanism(s) to limit damage (i.e. 185 symptom severity) has evolved, although whether this should be interpreted as tolerance is 186 perhaps a matter of perspective. The ambiguity arises because, while we have adopted a  In this study, we found a difference in reaction norm intercept, but not slope, between 192 unexposed and exposed finch populations. Strictly, this provides evidence for the evolution of were equivalent between finches from disease-exposed and unexposed populations, but those 257 from the latter displayed higher symptoms overall.

258
In conclusion, we provide evidence that house finches have evolved resistance 259 following the infectious outbreak of the bacterial pathogen, M. gallisepticum, with finches 260 from disease-exposed populations likely reducing pathogen load through acquired immune 261 processes [25]. Further, however, we also found evidence to suggest that the ability to limit

286
During the same time period, we also caught hatch-year house finches from populations 287 known to have been exposed to M. gallisepticum since the disease outbreak (exposed 288 populations). These were captured from urban areas and suburban parks in Alabama (see 289 [31]). Birds were similarly banded weighed, and sampled for blood and choanal swabs. They  304 We inoculated each of the birds with one of fifty-five M. galliseptum isolates sampled 305 over the course of the epidemic (N total inoculated =108, consisting of 51 birds from exposed 306 populations inoculated with 51 isolates and 57 birds from unexposed populations inoculated 307 with 55 isolates; 2 isolates were each inoculated into 2 birds from unexposed populations).

308
Isolates were administered via 20 μL of culture containing 1 × 10 4 to 1 × 10 6 colour changing   , and figures were made using ggplot2 [64]. Response variables (pathogen load, conjunctival swelling data) were natural log-transformed to better meet the assumptions 344 of residual normality. To determine whether peak pathogen load differed between disease-345 exposed and unexposed host populations, we ran a mixed effects model with ln(pathogen 346 load) as the response term, host population (exposed vs unexposed), dpi (as a continuous 347 covariate) and their interaction as explanatory terms. Bacterial isolate identity and bird 348 identity were both included as random effects. Finally, we tested for population differences in   Declaration of Interests

376
The authors have declared that no competing interests exist.

379
All data will be made available on Dryad Digital Repository (https://datadryad.org). over ythe course of the experiment(log-transformed) for birds from exposed (blue) and 598 unexposed (orange) populations. Raw values are shown as triangles (exposed) or circles 599 (unexposed populations); lines are predicted from the model (solid = exposed; dashed = 600 unexposed), with confidence intervals represented by ribbons.