Supplemented nutrition decreases helminth burden and increases drug efficacy in a natural host–helminth system

Gastrointestinal (GI) helminths are common parasites of humans, wildlife, and livestock, causing chronic infections. In humans and wildlife, poor nutrition or limited resources can compromise an individual's immune response, predisposing them to higher helminth burdens. This relationship has been tested in laboratory models by investigating infection outcomes following reductions of specific nutrients. However, much less is known about how diet supplementation can impact susceptibility to infection, acquisition of immunity, and drug efficacy in natural host–helminth systems. We experimentally supplemented the diet of wood mice (Apodemus sylvaticus) with high-quality nutrition and measured resistance to the common GI nematode Heligmosomoides polygyrus. To test whether diet can enhance immunity to reinfection, we also administered anthelmintic treatment in both natural and captive populations. Supplemented wood mice were more resistant to H. polygyrus infection, cleared worms more efficiently after treatment, avoided a post-treatment infection rebound, produced stronger general and parasite-specific antibody responses, and maintained better body condition. In addition, when applied in conjunction with anthelmintic treatment, supplemented nutrition significantly reduced H. polygyrus transmission potential. These results show the rapid and extensive benefits of a well-balanced diet and have important implications for both disease control and wildlife health under changing environmental conditions.


Methods
Here we provide additional information on methods used in the field and in the laboratory experiment, as well as rationale for approach to age and body condition analysis [1].

Experimental details
In 2015, we live-trapped three grids (1 supplemented grid, 2 control grids, 49 trapping stations per grid with 2 traps/station, 10m between each trap, for a total area of 3600m 2 ), while in 2016, we trapped four grids (2 supplementation and 2 control grids; with each grid set up as a 6x5 array of 30 trapping stations with 2 traps/station, 10m between each trap, for a total area of 2000m 2 ). All grids in both years were spaced a minimum of 50m from each other to minimise mouse movement between grids, and grids were randomly assigned to nutrition regimes prior to the start of the experiment.
Experimental grids were supplemented TransBreed TM -a high-nutrient, standard veterinary feed which is formulated for optimum breeding performance in laboratory mice and offers whole-diet nutrition to the wild mice in this study (20% protein, 10% fat, 38% starch, high content of micronutrients (Table   S1). Anthelmintic treatment regime consisted of a combination of Ivermectin and Pyrantel pamoatebroad-spectrum anthelmintics which target adult and larval stages (Ivermerctin) and adult stages (Pyrantel) of H. polygyrus in both laboratory and wild mice [7][8][9]. Previous work in wild A. sylvaticus found that the combination of Ivermectin and Pyrantel at 9.4mg/kg and 100 mg/kg, respectively, efficiently cleared H. polygyrus infection for 12-16 days [10].During trapping, each trap was set with cotton wool bedding, and was baited with seeds, carrot, mealworms, and TransBreed TM pellets (on supplemented grids only), set in the early evening (16.00-18.00) and then checked early the following morning.
bones) as detailed in [2]. After field data collection, body condition index (BCI) was calculated by obtaining the residuals of an ordinary least squares (OLS) regression of mass against length [3] for inclusion as a fixed effect in models of infection and immunity. We did not calculate body condition index for the colony mice as those measures would be obscured by the fact that they are, on average, much heavier compared to their wild counterparts.
Sex and reproductive status were assigned by visual examination of the genitals as male wood mice have a greater urogenital distance than females. Males were classed into the following reproductive categories: abdominal (testes non-visible); descended, or scrotal. Females were classed into the formalin for dissection of eye lenses for higher resolution estimates of animal age. Eye lens mass has been shown to strongly correlate with age in many species (rodents and others), and has successfully been used to distinguish age classes for both laboratory and wild mice [4,5]. Eyes collected from sacrificed animals were removed from their container and left at room temperature for 5-10 minutes to allow the formalin to evaporate. Eye lenses were then extracted and dried at 70°C overnight. They were then weighed to the nearest mg using a precision balance. The combined weight of both eye lenses (log-transformed) for each individual were used as a proxy for age. We calculated the relationship between age and eye lens weight using wood mice of known ages from our colony to be ( Figure S2): Although anthelmintic drugs were administered to half of the experimental group before secondary challenge, there was no difference in worm clearance (as indicated by EPG) between drug-treated and control mice ( Figure S8B) and thus they were combined within diet groups for these analyses. Over the course of this experiment, 5 mice exhibited weight loss over the threshold for our experimental protocol (not related to the diet supplementation or H. polygyrus infection) and were culled and removed from further analysis.

Diet Information
Wood mice in the laboratory colony were fed two different formulations of standard laboratory rodent chow. Control mice were fed with standard maintenance chow (RM1 TM ), whereas supplemented mice were fed a specially-formulated chow (TransBreed TM ), which has been designed to include higher fat, protein, and micronutrient contents (Table S1). We selected those diets to approximate control and supplemented nutrition in our field experiment as closely as possible. Both diets were fed ad libitum and provided adequate maintenance nutrition. Compared to field animals, lab wood mice had higher body mass (Lab mean weight = 23.88g; Wild mean weight = 20.32g; T-test, t = -2.99 p = 0.005) and better body condition than wild wood mice (Lab mean total fat score = 9.08/10; Wild mean total fat score = 5.7/10, Wilcoxon Rank-Sum test, W = 127, p < 0.001).

Parasite inoculation
Third stage larvae (L3) of H. polygyrus were originally derived from wild wood mice from Callendar Park that were infected with H. polygyrus [6]. Since then they have been passaged approximately ten times through colony-housed wood mice at the University of Edinburgh. In order to extract H.
polygyrus eggs from faecal samples, the pellets were broken up and mixed with inactivated charcoal to mimic soil. The charcoal-soil mix was spread thinly on moist filter paper maintained in petri dishes at 17C. Larvae started to hatch after approximately 12 days and were collected into sterilised water and kept at 4C until use. Prior to infections, larvae concentrations were adjusted to a final concentration of 200 L3/ 150µL. Infective doses were administered to mice via oral gavage.

Faecal Sampling
Faecal samples in the colony were collected by changing the cage bedding ~12 hr prior to each collection and collecting faecal pellets from the freshly-used bedding and then preserving the samples in 10% buffered formalin. A small sample of 2-3 pellets was also collected to measure faecal IgA.

Antibody assays
ELISAs were performed to measure (1) total faecal IgA concentration and (2)

Raw Data Summary, Full Model Output & Model Variation Output
This section includes raw data summaries and full model output for further interpretation of data presented in the main text. Additionally, we describe the models using a 3-level factor for supplementation in the wild to more accurately classify mice who were captured on both supplemented and control grids. Only 18% (n = 16) of mice were captured on both grid types, but to test the possibility that effects of supplemented nutrition could be dependent on the time spent on these grids, we specified diet as 'control', 'mix', or 'supplemented', where 'mix' represented mice that were found on both control and supplement grids across the experiment. These models were very similar to main models, with the exception of a diet-by-treatment interaction as group sizes were not sufficient to fit this interaction. When compared to models with a 2-level diet group, this modification did not significantly improve the fit for models of EPG at first capture, post-treatment, or at end point (ΔAIC = 1.71, ΔAIC = -1.57, ΔAIC = 1.28, respectively; Fig S1) nor did the inclusion of this factor level change any of the main effects ( Figures S1 and S4).     Figure S2. Correlation of wood mouse age predicted from eye lens weight and known age (in weeks) from colony wood mice. Pearson's r with 95% credibility intervals and significance of the correlation is included for both 2015 and 2016 data. Points have been jittered by 10% of raw values to aid in visualisation.