Diet composition modulates animals’ ability to resist parasites and recover from stress. Broader diet breadths enable omnivores to mount dynamic responses to parasite attack, but little is known about how plant/prey mixing might influence responses to infection.

Using omnivorous deer mice (Peromyscus maniculatus) as a model, we examine how varying plant and prey concentrations in blended diets influence resistance and body condition following infestation by Rocky Mountain wood ticks (Dermacentor andersoni).

In two repeated experiments, deer mice fed for four weeks on controlled diets that varied in proportions of seeds and insects, were then challenged with 50 tick larvae in two sequential infestations.

The numbers of ticks successfully feeding on mice declined by 25% and 66% after the first infestation (in the first and second experiments, respectively), reflecting a pattern of acquired resistance, and resistance was strongest when plant/prey ratios were more equally balanced in mouse diets, relative to seed-dominated diets.

Diet also dramatically impacted the capacity of mice to cope with tick infestations. Mice fed insect-rich diets lost 15% of their body weight when parasitized by ticks, while mice fed seed-rich diets lost no weight at all.

While mounting/maintaining an immune response may be energetically demanding, mice may compensate for parasitism with fat and carbohydrate-rich diets.

Altogether, these results suggest that a diverse nutritional landscape may be key in enabling omnivores’ resistance and resilience to infection and immune stressors in their environments.

Mice

Prior to this experiment, deer mice (Peromyscus maniculatus subspecies sonoriensis) were originally obtained from the Peromyscus Genetic Stock Center (University of South Carolina, USA) to begin a laboratory colony. All mice were tick naïve for approximately 18 generations and kept in a climate-controlled room at ca. 21ºC, with a relative humidity of ca. 30% and in a long day (16h), short night (8h) cycle. Mice were housed with same sex individuals with 2 – 6 mice per cage until used in experiments when mice were held individually in cages with a 9.83 cm long, 5.08 cm diameter cardboard tube (Bio-Serv, USA) for environmental enrichment. All mice were fed ad libitum using a breeder-type diet of laboratory rodent chow obtained from Harlan Laboratories (USA) until the start of each experiment, at which time mice were gradually transitioned to a diet consisting of a mixture of seeds and freeze-dried insects. Adult mice (2 – 16 months old) were in an approximate 1:1 sex ratio. While mouse starting weights after transitioning to the experimental seed-and insect rich diets were not significantly different from each other, mice on the insect-rich diet lost 8% of their pre-transition weight. Mouse genetic relatedness and genetic diversity were not determined for the colony, but no siblings were used for breeding pairs. For each experiment described below, mice were age-matched and distributed evenly by sex among the diet groups. For experiment 1, all mice were sexually mature, ca. four months old (12 females, 12 males). Mice in experiment 2 were all sexually mature, ca. three months old (six females, five males). Each mouse was weighed daily until the start of the tick infestations, at which time they were weighed before ticks were applied and after ticks dropped off (7-d interval = 2 d tick-contact period + 5 d blood feeding and detachment). Peromyscus maniculatus reaches sexual maturity at approximately 48 days (Clark 1938), at which point growth rate declines to near zero (Dice & Bradley 1942), so we expected only marginal growth of the sexually mature mice in our experiments. All animal use was approved by the Institutional Animal Care and Use Committees (IACUC) of Washington State University, USA and the University of Idaho, USA.

Ticks

Tick larvae used in these experiments were obtained from a disease-free colony maintained by the USDA-ARS at the University of Idaho, USA. To start the colony, adult ticks were originally collected from Reynolds Creek, Idaho, USA, and to assure genetic heterogeneity within the tick colony, field collected ticks were added to the colony every three – four years. Colony ticks were tested for parasites/pathogens using PCR. The adult ticks were fed on Holstein cattle to produce larvae (Scoles et al., 2005). Egg masses from several engorged female ticks were mixed, aliquoted into glass vials and maintained for hatching in humidity chambers at 25°C using a saturated salt solution of potassium nitrate to provide a constant relative humidity of ca. 95.5% (Winston and Bates, 1960). After hatching, humidity chambers were kept in an incubator at 15ºC in a light cycle of 12 h light: 12 h dark until the larvae were used.

Infestation Procedure

Fifty RMWT larvae per mouse were brushed into a 9.83 cm long, 5.08 cm diameter cardboard tubes (Bio-Serv, USA). This challenge approximated a median tick burden observed among hosts of larval D. andersoni in the wild (Sonenshine et al., 1976). Each cardboard tube containing larvae was then placed into a solid, plastic mouse cage (29.21cm L x 21.27cm W x 15.88cm H) with a small handful of bedding material (TEK-Fresh Laboratory Animal Bedding, Harlan Laboratories, USA). A mouse contacted the tick larvae when taking refuge in the cardboard tube over the initial 48 h period. After the initial contact period, each mouse was transferred to a wire-bottom cage (29.85cm L x 20.96cm W x 15.24cm H) along with the cardboard tube for environmental enrichment. Both the cardboard tube and the plastic cage were checked thoroughly for any tick larvae that did not attach to a mouse during the 48 h period. No unattached tick larvae were ever found in the cardboard tube or the plastic cage after transfer of mice to the wire-bottom cages. Our cage set-up made it possible to recover all ticks that did not attach to mice, and we closely inspected each mouse for attached ticks after the completion of each infestation. We assumed that any ticks not recovered were eaten by mice. The wire-bottom cage was suspended over a pan of water. Over a 5 d period, replete (blood-fed) larvae dropped from the mouse and into the water pan. Replete ticks were recovered from the water pan. All recovered ticks were blood-fed and we never observed ticks remaining attached to the mice after the 7-day infestation period. Feeding success was determined by recovering and counting the number of replete ticks that dropped into the water pan. The recovered ticks represented the number of larvae that (i) were able to successfully encounter the mouse, (ii) subsequently attach to the mouse, and (iii) fully engorge.

Measures of Resistance and Body Mass

Resistance was measured as the number of RMWT larvae able to obtain a blood meal out of the 50 unfed ticks initially placed with the mouse. Following parasitism, daily measurements of body mass were taken, an important fitness correlate in small mammals (Balčiauskas et al., 2022; Myers and Master, 1983), that likely predicts their ability to tolerate parasite burdens (Budischak and Cressler, 2018; Kutzer and Armitage, 2016) and is reflective of body condition and endogenous capital from which they are able to fuel reproductive processes (Clutton-Brock 1984; Hamel et al., 2008). Body mass commonly serves as a proxy for tolerance metrics (Bordes et al., 2012; Sorci, 2013).

Diet Composition and Feeding Regime 

Experiment 1: Two-level variation in diet composition

Diets were prepared to match the composition and volume of food items mice could encounter and consume in the wild (Jameson, 1952; Sealander 1952; Whitaker Jr, 1966; Williams, 1959). Each diet contained a mixture of seeds (pine nuts: Pinaceae; sunflower: Asteraceae) and insects (mealworms: Coleoptera:Tenebrionidae; crickets: Orthoptera: Gryllidae) totaling 18 kilocalories mouse⁻¹ day⁻¹ (Table 1). The number of kcals day⁻¹ was determined from preliminary measures that revealed fewer kcals mouse⁻¹ day⁻¹ could cause weight loss (>25% of body weight), emaciation and mortality. This approach differs from laboratory studies where mice are fed chow ad-libitum for growth and reproduction, as it was restricted to a maintenance diet with a precise caloric value to reflect natural limits to food availability they would experience in the wild. Diet selection and rates of consumption of plant/prey foods were examined in experiment 3, described below. All diet items were analyzed for gross energy (bomb calorimetry), crude protein (combustion method to measure nitrogen) and crude fat (ether extract) by the Washington State University Wildlife Habitat and Nutrition Laboratory (http://cahnrs.wsu.edu/soe/facilities/wildlifehabitat/). Mice were fed a diet of 18 kcals day^-1 that was either seed-rich (~90% of kcals from seeds) (n = 12 mice; 6 female, 6 male) or insect-rich (~60% of kcals from insects) (n = 12 mice; 6 female, 6 male) over a period of nine weeks (Table 1). The proportions of diet items were determined from pilot studies that revealed the conditions of mice would deteriorate (decreased body weight or vigor) on diets that were >90% of kcals from seeds, or >60% of kcals from insects. Based on the gross energy estimate (kcals g⁻¹) of each diet item, the amount of seeds and insects in each diet combination were determined by wet weight. For example, the diet of 90% seeds required that 16.2 kcals come from the two seed types. We weighed the amount of sunflower seeds and pine nuts that provided roughly 8.1 kcals for each seed type (Table 1). The amount of mealworm and cricket material was similarly measured for the insect fraction of the diet.

Table 1 Variation in isocaloric diet composition: Energy and macronutrient breakdown.

^aValues determined by analysis of food stuffs at the Wildlife Habitat and Nutrition Laboratory at Washington State University.

^bValues determined from the literature and packaging nutrition facts [63].

Four weeks prior to the first infestation with RMWT larvae, all mice were transitioned from rodent chow to their respective diets. This acclimation period allowed adequate time for the differential nutritional contents of each diet to assimilate sufficiently to affect diet-driven immune function. Three to six weeks is a common acclimation period for diets prior to immune challenges in mice, (Kostovcikova et al., 2019, Sweeny et al., 2021). During this transition period, all mice were weighed daily during 13:00 h and 14:00 h using a digital bench-top scale (±0.01 g) to be sure mouse body weights did not decline by greater than 25% (Lobo and Millar, 2011). Mice were fed at approximately the same time each day (14:00 h), and any remaining food items from the previous day were removed from the cage and replaced with a new batch of the balanced diet items. Amounts of remaining diet items were not measured, so we did not calculate consumption by each mouse.

We determined qualitative differences in diet item intake by comparing the stable isotope signatures of δ ¹⁵N in mice from the high- and low-seed groups of experiment 1. Nitrogen isotope ratio (¹⁵N/¹⁴N) is used to determine relative trophic levels, and can distinguish mammals that consume high levels of animal matter from those that consume primarily plant matter (Miller et al., 2008). A blood sample was collected from each mouse prior to the first experimental infestation, and the samples were sent to the Stable Isotope Research facility at the University of Utah.

Experiment 2: Four-level variation in diet composition

Based on results of Experiment 1, diets were designed to establish a gradient from the maximum to the minimum amount of seeds that would sustain deer mice. Diet compositions were mixed from pine nuts, sunflower seeds, mealworms and crickets to provide 3 g of food mouse⁻¹ day⁻¹ at 4 levels of seed percentage: 90% seeds (n = 3 mice; 2 female, 1 male), 75% seeds (n = 3 mice; 2 female, 1 male), 50% seeds (n = 2 mice; 1 female, 1 male) and 40% seeds (n = 3 mice; 1 female, 2 male) (Table 2). The percentage of seeds and insects were based on the total amount of food provided by wet weight (e.g. 90% seed diet consists of 3 g x 0.9/2 seed types = 1.35 g of each seed type). The two ends of the diet gradient (90% seeds and 40% seeds), matched the seed-rich and insect-rich diets examined in experiment 1. Only two mice were in the 50:50 diet group due to unexpected mortality of one mouse during the experiment (cause of death remained unknown following a veterinary necropsy). Each mouse was fed only one diet type for four weeks prior to the first infestation. Mice were fed at ca. the same time each day (17:00 h), and any remaining food items from the previous day were removed from the cage and replaced with a new batch of food items. As with experiment 1, mice typically ate all diet items, but amounts of any remaining food were not measured.

Table 2 Diet composition and kilocalories in four mouse diets that varied in seed:insect ratios.

^aValues determined by analysis of food stuffs at the Wildlife Habitat and Nutrition Laboratory at Washington State University.

^bValues determined from the literature and packaging nutrition facts [63].

Infestation timing

Five and eight weeks after transition to experimental diets all mice in experiments 1 & 2 were infested with 50 RMWT larvae each as described previously. The exposure of 50 larvae mouse⁻¹ represents a moderate burden of ticks, as field estimates of tick burdens on Peromyscus spp. mice range from 1 to >100 ticks per individual (Hersh et al., 2014). Diet treatments were maintained during both infestations. Mice were only weighed before and after each 7 d infestation period, so as not to disrupt attachment and blood feeding of the larvae.

Experiment 3: Diet Item Consumption Experiment 

Diet item consumption was not measured in Experiments 1 & 2, so a third experiment was conducted to measure types and amounts of diet items eaten by deer mice provided with mixed diets. At 45 days old, weaned, age-matched mice were transitioned from laboratory mouse chow to a mixture of seed (sunflower and pine nuts) and insect (mealworm) food items. Mice in a seed-rich group (S) were each provided ca. 0.71 g sunflower seeds, 1.12 g pine nuts, and 0.21 g mealworms daily. Mice in an insect-rich group (I) were each provided ca. 0.61 g sunflower seeds, 0.42 g pine nuts, and 1.28 g mealworms daily. In the seed-rich group there were 8 females and 10 males, while in the insect-rich group there were 9 female and 11 male mice. All mice were housed individually. Daily, any uneaten food items from the previous day were collected from the cage and stored, and fresh diet items were supplied to each mouse according to the treatment group. Mice were individually weighed using a digital scale every 1-3 days for a period of 5 months. At the end of the experiment, uneaten diet items collected from each mouse were weighed to determine the proportion of the total diet item provided that each mouse had consumed (= (total mass provided - uneaten mass)/total mass provided).

Statistical analyses

All analyses were completed in R version 4.0.2. Pre-infestation body weights of mice in each diet group were compared using ANOVA, and we report type II sums of squares for unequal sample sizes. To evaluate how mouse diets influence their resistance to tick infestation, we analyzed counts of ticks that fed successfully with generalized linear mixed models using the glmer() function in the lme4 package (Bates et al., 2014), assuming a Poisson distribution. We tested interactions between diet composition and infestation sequence (first or second), with mouse sex as a main effect, and mouse ID as a random effect accounting for non-independence between repeated infestations. Mass lost by mice following each tick infestation was analyzed as a proportion of the baseline weight prior to infestation. We used linear mixed effects models assuming a normal distribution with the lme() function in the nlme package (Pinheiro et al., 2017), with mouse ID as a random effect. To determine mouse preferences for plant- and prey-based food and evaluate how consumption rates of insects and seeds varied across diet treatments, we used a linear mixed effects model with logit-transformed proportions of seeds and prey consumed as the dependent variable (Warton and Hui, 2011), and tested an interaction between food item (sunflower seeds, pine nuts, mealworms) and diet (seed-rich and insect-rich), with mouse ID as a random effect. We also used a simple t-test to determine how diets influenced mouse body mass over an extended period (5 months). Test statistics for main effects and interaction terms were extracted from ANOVA tables using the Anova() function in the car package (Fox et al., 2012). Means were separated between diet treatment groups using Tukey’s HSD posthoc tests when significant effects were revealed using the glht() function in the multcomp package (Hothorn et al., 2016). We checked model assumptions visually using residual plots and histograms (where normality was assumed).