Effect of Housing Condition and Diet on the Gut Microbiota of Weanling Immunocompromised Mice

Gastrointestinal microbiota are affected by a wide variety of extrinsic and intrinsic factors. In the husbandry of laboratory mice and design of experiments, controlling these factors where possible provides more reproducible results. However, the microbiome is dynamic, particularly in the weeks immediately after weaning. In this study, we characterized the baseline gastrointestinal microbiota of immunocompromised mice housed under standard conditions for our facility for 6 weeks after weaning, with housing either in an isolator or in individually ventilated cages and a common antibiotic diet (trimethoprim sulfamethoxazole). We compared these conditions to a group fed a standard diet and a group that was weaned to a standard diet then switched to antibiotic diet after 2 weeks. We found no clear effect of diet on richness and α diversity of the gastrointestinal microbiota. However, diet did affect which taxa were enriched at the end of the experiment. The change to antibiotic diet during the experiment did not convert the gastrointestinal microbiome to a state similar to mice consistently fed antibiotic diet, which may highlight the importance of the initial post-weaning period in the establishment of the gastrointestinal microbiome. We also observed a strong effect of housing type (isolator compared with individually ventilated cage) on the richness, α diversity, β diversity, and taxa enriched over the course of the experiment. Investigating whether the diet or microbiome affects a certain strain’s phenotype is warranted in some cases. However, our findings do not suggest that maintaining immunocompromised mice on antibiotic feed has a clinical benefit when potential pathogens are operationally excluded, nor does it result in a more consistent or controlled microbiome in the post-weaning period.

Gastrointestinal microbiota (GM) have significant effects on animal health and are affected by many factors including diet, antibiotic administration, age, sex, strain, and cage environment.^{4,5,7,8,11,16} When using mice to study human disease, particularly when studying immune response to disease, recognizing which portions of a study are or are not controlled is an important element.⁹ In the case of the GM, characterizing the populations of bacteria under defined housing conditions for commonly used strains is an important step in improving reproducibility and translatability of these models.⁹Feeding diets compounded with antibiotics is common practice in irradiated animals to prevent opportunistic infections in the postprocedural period, and that practice has been extended to other populations that are known to be immunocompromised, such as SCID or nude mice.^3,19 In our facility, standard practice is to maintain populations of known and presumed immunodeficient mice as well as immunologically-humanized mice, on trimethoprim-sulfamethoxazole (TMS) antibiotic chow after weaning. However, mice that consume TMS from ad libitum commercial diets or water do not achieve therapeutic plasma concentrations of antibiotic against most pathogens.^14,15 Non-judicious use of antibiotics also contributes to the growing problem of antimicrobial resistance.¹⁹ Because immunodeficient mice are used to study human disease, immunity, and autoimmunity, the use of an antibiotic that may affect immunity and be clinically irrelevant for preventing disease creates a significant confounding variable.^1,9,10,20,21 In addition, medicated chow is more expensive than standard rodent chow, so this policy represents a significant operational cost.This study had 2 primary goals. The first was to characterize the microbiome of our most common immunodeficient mouse model after weaning under 2 standard housing (individually ventilated cage (IVC) or open-top cage within a semirigid isolator (isolator)) and 3 diet (TMS diet, standard rodent chow, or a switch from standard to TMS diet) conditions. This information may be important for future model characterization and may assist in rescue of model phenotypes that depend on an inhouse microbiome that may drift over time. We hypothesized that housing would have a greater effect on the GM than would the type of chow because of the low concentration of TMS in the medicated chow. We also hypothesized that mice fed only a TMS diet would show lower GM species richness and α diversity as compared with groups fed standard rodent chow due to antibiotic dampening of certain bacterial populations.Concern about maintaining disease phenotypes that depend on diet is a major roadblock to changing our current practice of providing TMS diet. Therefore, our second goal was to determine the stability of the GM after a diet change in the post-weaning period. To assess this, we changed the diet of a subset of cages from standard diet back to TMS diet 2 weeks after weaning. We hypothesized that a change to the TMS diet after a 2-week period on standard chow would²⁴ negate any TMS-related effects on the microbiome. This outcome would be expressed as an alignment in richness, diversity, and relative populations with that found in the group fed TMS diet over the entire experiment under a given housing condition.