Metaphylactic Antibiotic Treatment to Prevent the Transmission of Corynebacterium bovis to Immunocompromised Mouse Offspring

Abstract

Current methods for eradicating Corynebacterium bovis, such as depopulation, embryo transfer, and cesarean rederivation followed by cross fostering, are expensive, complex, and time-consuming. We investigated a novel method to produce immunocompromised offspring free of C. bovis from infected NOD.Cg-Prkdc^scid Il2rg^tm1Wgl/SzJ (NSG) breeding pairs. Adult NSG mice were infected with C. bovis, paired, and randomly assigned to either a no-antibiotic control group (NAB, n = 8) or a group that received amoxicillin–clavulanic acid (0.375 mg/mL) in their drinking water for a mean duration of 7 wk (AB group, n = 7), spanning the time from pairing of breeders to weaning of litters. The AB group also underwent weekly cage changes for 3 wk after pairing to decrease intracage C. bovis contamination, whereas the NAB mice received bi-weekly cage changes. Antibiotics were withdrawn at the time of weaning. All litters (n = 7) in the AB group were culture- and qPCR-negative for C. bovis and remained negative for the duration of the study, whereas all litters in the NAB group (n = 6) remained C. bovis positive. A single adult from each breeding pair was sampled at weaning and at 5 and 10 wk after weaning to confirm the maintenance of (NAB) or to diagnose the reemergence (AB) of C. bovis infection. By the end of the study, C. bovis infection had returned in 3 of the 7 (43%) tested AB adults. Our data suggest that metaphylactic antibiotic use can decrease viable C. bovis organisms from adult breeder mice and protect offspring from infection. However, using antibiotics with frequent cage changing negatively affected breeding performance. Nevertheless, this technique can be used to produce C. bovis-free NSG offspring from infected adults and may be an option for salvaging infected immunocompromised strains of mice that are not easily replaced.

Corynebacterium bovis is a common bacterial pathogen among immunocompromised mouse colonies. Clinical signs associated with C. bovis infection in athymic nude mice (Foxn1, nu/nu) include an asymptomatic persistent carrier state and 2 to 7 d of dermal hyperkeratosis, dehydration, lethargy, and decreased body condition.^5,9,11,24 Haired immunocompromised mouse strains such as NOD.Cg-Prkdc^scid Il2rg^tm1Wgl/SzJ (NSG) mice also can develop clinical C. bovis infection and present with rough hair coat, decreased body condition, scaly skin, alopecia, conjunctivitis, and erythematous pinnae.^5,9,25

C. bovis has a negative effect on cancer research due to the induced changes to the immune response and some institutions choose to exclude this pathogen from their facilities.^20,29 Depopulation, environmental and equipment sanitation, and repopulation are often elected, given that the majority of susceptible strains can easily be purchased from C. bovis-free vendors. However, when unique noncommercial immunocompromised strains are infected, depopulation may not be an option. Rederivation methods that are likely to be effective in the elimination of C. bovis include embryo transfer and cesarean rederivation followed by cross fostering. These measures can be expensive and require a high degree of technical skill and considerable time. Conversely, antibiotic administration through feed or drinking water requires minimal labor, treats multiple animals concurrently, provides no additional stress to the animals, and requires no specialized surgical skills.¹⁷

In one study, antibiotic administration prevented the detection of C. bovis from infected mouse skin tested by culture during treatment.⁵ However, after discontinuing long-term antibiotic administration, C. bovis can again be cultured, with definitive clearance of only 13% of infected adult mice.⁵ We recently demonstrated that prophylactic antibiotic therapy can prevent the infection of immunodeficient mice after acute exposure.¹⁶ As opposed to prophylaxis, metaphylaxis is the treatment of an animal population that has been infected with a microbial agent but is not currently experiencing any level of clinical disease.³ Metaphylaxis differs from prophylaxis and therapeutic antimicrobial use because the expectation is to control an infection that is already present, rather than to prevent or cure infection. Using antibiotics in a metaphylactic manner requires a population-management mindset.

The goal of our study was to investigate the metaphylactic use of amoxicillin–clavulanic acid for C. bovis-infected NSG breeding pairs in order to prevent post-parturition transmission of C. bovis prior to weaning. We hypothesized that the metaphylactic antibiotics would either eliminate all viable C. bovis or decrease skin burden to undetectable levels on adult breeders. This situation would facilitate a C. bovis-free window during which transmission to neonates and weanlings would be prevented, thus providing a novel and noninvasive method to eliminate C. bovis from unique immunocompromised mouse strains that are not easily replaced.

Materials and Methods

Animals and facility.

Thirty male and 26 female, 7 wk old, NOD.Cg-Prkdc^scid Il2rg^tm1Wgl/SzJ mice (NSG) were purchased from Jackson Laboratories, Bar Harbor, ME. The mice were documented to be free of ectromelia virus, Theiler mouse encephalomyelitis virus, Hantaan virus, K virus, lactic dehydrogenase elevating virus, lymphocytic choriomeningitis, mouse adenovirus, mouse cytomegalovirus, mouse hepatitis virus, mouse minute virus, mouse parvovirus, mouse thymic virus, pneumonia virus of mice, polyoma virus, reovirus 3, rotavirus, Sendai virus, Bordetella spp., CAR bacillus, Citrobacter rodentium, Clostridium piliforme, Corynebacterium kutscheri, Corynebacterium bovis, Mycoplasma pulmonis, Salmonella spp., Streptobacillus monoliformis, Encephalitozoon cuniculi, fur mites, lice, follicle mites, pinworms, roundworms, and tapeworms. On arrival, mice were tested by skin swab and confirmed negative for C. bovis by qPCR analysis. All animal manipulations were approved by the IACUC of University of Colorado–Denver Anschutz Medical Campus, an AAALAC-accredited institution. Personal protective equipment required to enter the facility include a hair bonnet and disposable gown worn over personal clothing; additional equipment was required in the quarantine animal housing room and included an additional disposable gown and shoe covers.¹⁵

Animal housing and manipulations.

On arrival, all mice were acclimated for 1 wk and housed 4 per cage according to sex. Mice were moved to a quarantine room, where they were housed in autoclaved JAG 75 cages (Allentown Caging, Allentown, NJ) on a 70-cage, single-sided, IVC rack (MicroVent, Allentown Caging) providing 40 air changes per hour. The intracage environment included aspen chip bedding, a compressed cotton square, sterile cage furniture or cardboard roll, ad libitum irradiated rodent diet (2920X, Teklad Extruded Diet, Envigo, Indianapolis, IN), and autoclaved reverse-osmosis–purified, hyperchlorinated (3 to 5 ppm) water in 375-mL water bottles (Allentown Caging). The macroenvironment was maintained at 22.2 ± 1 °C (72 °F) and 30% to 40% humidity with at least 12 complete air changes per hour and a controlled 14:10-h light:dark cycle. All animal manipulations were conducted in an animal transfer station (ATS2, Allentown Cages) with work surfaces moistened with disinfectant (1:18:1 Clidox S, Pharmacal, Naugatuck, CT). Cages were changed on a 2-wk cycle for the duration of the experiment, with the exception that the AB group (but not the NAB group) received weekly changes during the first 3 wk after breeder pairing. A clean glove-changing practice was used to prevent contamination of C. bovis infection between cages. Because this bacterium is environmentally stable, the animal housing room and all contents were considered to be C. bovis-positive. A C. bovis-free glove box, containing standard nitrile exam gloves, was stored inside a zip-lock plastic bag within the housing room. Once the cage lid was removed, these gloves were donned to manipulate all aspects of the intracage environment without touching anything else in the room.

Bacterial inoculation.

C. bovis CUAMC1, isolated in 2014 from an infected nude mouse on the Anschutz Medical Campus, was removed from frozen stock, and cultured on tripticase soy agar with 5% sheep blood (catalog no. 221261, Becton Dickenson, Franklin Lakes, NJ) at 37 °C for 72 h. At 24 hours prior to mouse inoculation, CUAMC1 was propagated in heart infusion broth (catalog no. 238400, Becton Dickenson) containing 1% Tween 80 for 24 h. Bacterial concentration was determined by comparing the culture's absorbance at OD₆₀₀ with a standard curve generated by serial dilutions. Mice were anesthetized with isoflurane, and 50 µL of culture broth containing 2 × 10⁷ cfu of C. bovis was applied to the dorsal neck of each mouse. A sterile swab (BactiSwab Dry, Remel, Lenexa, KS) was used to distribute the inoculum on the skin, outer and inner pinnae, muzzle, lip commissures, and eyelids bilaterally. At 2 wk after experimental inoculation, mice in each cage were swabbed, to confirm infection by qPCR analysis. Cages that failed to generate an established infection were supplemented with 50 mL of soiled bedding from an experimentally positive cage. At 2 wk after exposure to C. bovis-soiled bedding, all remaining cages were confirmed positive by qPCR analysis. Animals were positive via qPCR analysis for a minimum of 3 wk (21.6 ± 1.4 d) prior to the start of the experiment.

Experimental design.

After qPCR confirmation of C. bovis infection, male and female NSG mice were randomly grouped into breeding pairs, which then were assigned to receive either treatment with antibiotic-containing drinking water (AB, n = 8) or standard facility water (NAB, n = 8). Because a low birth rate occurred in AB, 10 additional breeding pairs were added to reach a statistically significant sample size. On the day after pairing, the AB group began receiving drinking water containing amoxicillin–clavulanic acid (amoxicillin trihydrate–clavulanate potassium, 0.375 mg/mL, Zoetis, Parsippany-Troy Hills, NJ). Medicated water was prepared by mixing 0.37 g of amoxicillin trihydrate–clavulanate potassium, previously determined to be equivalent to 1/4 teaspoon, into 375 mL of facility water for an anticipated dose of 75 mg/kg based on an estimated 5 mL of water consumption daily.¹⁷ Water bottles were refreshed weekly,^19,23,28 and were provided to the AB group throughout gestation and until the date of weaning. Antibiotic administration averaged approximately 7 wk (49.8 ± 2.1 d). Breeding pairs that failed to produce a litter or demonstrated infanticide did not progress further in the study.

Viable offspring from breeding pairs (AB group, 7 litters; NAB group, 6 litters) were weaned at 28 d of age. Litters were weaned into sex-segregated cages, with a maximum of 1 cage per sex with 5 mice per cage. All additional weanlings were euthanized by using CO₂ asphyxiation followed by cervical dislocation. Litters born before the 28-d weaning of the first litter were euthanized by decapitation after anesthesia (approximately 5 to 10 min of hypothermia). At the time of weaning, offspring in the AB group were transferred to an adjacent housing room to minimize the risk of contamination from the positive control (NAB) group. For 10 wk after weaning, weanlings and one parent were monitored by culture and qPCR analysis for reemergence of C. bovis (Figure 1). At the conclusion of the monitoring period, all animals were euthanized as described for weanling mice. For the AB group, the experiment used 2 cohorts of mice; the first cohort was bred in December, the second in April.

Figure 1.

Illustration of the experimental design from C. bovis inoculation to the end of the study. Male mice, blue; female mice, pink; either sex, white. After inoculation, mice were housed in sex-segregated cages for a minimum of 3 wk prior to randomization into breeding pairs and treatment groups. Treatment provided to the AB group involved weekly cage changing from pairing to birth (*) and antibiotics in the drinking water from pairing to weaning. C. bovis was monitored by qPCR analysis and culture at each time point from pairing to 10 wk after weaning.

Sample collection.

Pelt swabs were collected for C. bovis qPCR analysis and bacterial culture by passing a swab (BD BBL CultureSwab EZ, Becton Dickinson, Sparks, MD) over the animal's right lateral thorax, inner and outer pinnae, periocular region, muzzle, and oral cavity in a consistent pattern. When both parents were present, the female mouse was used to obtain qPCR samples. Swabs were submitted to the University of Colorado Denver Quantitative PCR Core for DNA isolation and analysis as previously described.¹⁵ Time points of sample collection for C. bovis qPCR analysis included at arrival, after C. bovis inoculation, 24 h after pairing for breeding, 3 d after birth, 28 d after birth (weaning), and 5 and 10 wk after weaning. qPCR was used as the primary assay for C. bovis status determination prior to and after the withdrawal of antibiotics. We anticipated that during antibiotic administration, mice would be culture negative, yet positive by qPCR due to residual C. bovis DNA in their fur. An increase in copy number after antibiotic withdrawal would suggest an active C. bovis infection.

Pelt swabs for bacterial culture of C. bovis were obtained by using the same swabs, sampling procedure, and interval as for qPCR analysis. To reduce bacterial load on the swab, we did not sample the oral cavity. When both parents were available, culture samples were taken from the male. The swabs were shipped to IDEXX BioAnalytics for C. bovis isolation and identification. Culture swab tips were vortexed in sterile PBS, and a 100-μL aliquot was plated on BBL Trypticase Soy Agar with 5% sheep blood (TSA II, Becton Dickinson). Plates were incubated at 35 °C with 7% CO₂ and monitored for colony growth for 72 h. Individual bacterial colonies displaying morphology consistent with Corynebacterium spp. were analyzed by MALDI-TOF mass spectrometry for species-level identification, as described previously.²⁵ C. bovis culture was the primary test used to determine C. bovis status during the administration of antibiotics.

Data analysis.

Both qPCR and culture results were used to evaluate the presence of C. bovis. Any qPCR value greater than 0 was considered to be C. bovis-positive. Statistical significance was assessed using the Fisher exact test for differences between the AB and NAB groups in the proportion of mice with C. bovis at each time period. The numbers of viable and nonviable litters were compared between AB and NAB groups by using the Fisher exact test. Treatment group characteristics were summarized according to bacterial counts and percentages to describe and visually compare each categorical value. The McNemar test was used to assess changes in C. bovis-positive status of mice between the withdrawal of antibiotics and the end of the study, for both qPCR and culture results. To compare agreement between qPCR and culture results, we applied intraclass correlation with a 95% confidence interval based on agreement by using a 2-way mixed-effects model. To aid in visualization due to the extreme range in qPCR values, copy numbers were transformed by using log₁₀(qPCR + 1) and graphed in correlation with culture results by using R version 3.5.0 (The R Foundation, www.r-project.org). Significance was defined as a P value of 0.05 or lower and based on a 2-sided alternative. All analyses were performed by using R version 3.5.0 (The R Foundation, www.r-project.org) and SAS version 9.4 (SAS Institute, Cary, NC).

Results

C. bovis Infection.

Mice were qPCR-positive for C. bovis for a minimum of 3 wk prior to the start of the experiment. During the study, 2 C. bovis-positive adult breeding pairs and 1 set of weaned offspring, all from the NAB group, developed characteristic clinical symptoms of C. bovis infection. All other mice, independent of C. bovis status, remained asymptomatic for the duration of study.

qPCR and culture diagnostics.

Adult breeding pairs in the AB group showed a decrease in qPCR copy number from pairing to weaning, which was statistically significant (P < 0.05, Figure 2) at the time of weaning. When the C. bovis culture status was assessed for the AB group, all breeding pairs were positive at pairing. At 3 wk after antibiotic administration, corresponding to the time of pup birth, 2 of the 7 (29%) breeding pairs in the AB group remained culture-positive (Table 1). All of the 7 AB breeding pairs were culture-negative at weaning. At 5 and 10 wk after antibiotics were discontinued, 3 of the 7 (43%) adults returned to culture-positive status. In comparing litter culture status at 10 wk after weaning, the AB group differed significantly (P < 0.05) from the NAB group: the AB group had a lower proportion of C. bovis-positive litters. The weaned offspring from the AB group were qPCR- and culture-negative at both 5 and 10 wk after weaning (Table 1). One pair of weaned males in the AB group was retained beyond the end of the study, monitored at 31 wk after weaning, and shown to be negative for C. bovis via qPCR and culture analysis (data not shown).

Figure 2.

Longitudinal overview of qPCR results by mouse breeding pair. Left panel, breeding pairs in the nonantibiotic-treated (NAB) group; right panel, breeding pairs in the antibiotic-treated (AB) group. Lines indicate trajectories for individual breeding pair, and points corresponding to a culture result at a given measurement time are color-coded. qPCR results are shown on the log₁₀(qPCR+1) scale.

Table 1.

Numbers of breeders and offspring culture-positive for C. bovis

	Treatment group
	AB (n = 7)	NAB (n = 6)
Adults
At breeder pairinga	7e	2
At birth of pups	2e	6
At weaning of pupsb	0e	6
5 wk after weaningc	3	6
10 wk after weaningc	3	6
Offspring
5 wk after weaning	0e	6
10 wk after weaningd	0e	6

^aqPCR data used at breeder pairing to designate positive C. bovis status

^bAdults and pups cultured immediately prior to weaning by using the same sample

^cOne adult of the breeding pair was retained for sampling at 5 and 10 wk after weaning of pups

^dBoth male and female offspring were cultured by using the same sample

^eP ≤ 0.05 (Fisher exact test) between treatment groups

As anticipated, adult breeding pairs in the NAB group showed an increase in mean qPCR copy number from pairing to weaning (Figure 2). Culture results for breeders in the NAB group at the time of pairing did not agree with qPCR results, with only 2 of 6 breeder pairs positive according to culture and all positive by PCR analysis. Nevertheless, NAB group showed both qPCR- and culture-positive status among all pup cages and adult cages from birth through 10 wk after weaning (Table 1).

The qPCR and culture test results did not agree across the entire experiment. However, they displayed intraclass correlation (95% CI) of 0.74 (0.64–0.82), which indicates moderate agreement between the 2 methods.

Breeding performance.

Breeding success was judged by visually assessing viable litters at 4 wk after parturition. In the AB group, 16 of the 18 (89%) pairs gave birth to live pups, compared with all 8 (100%) pairs in the NAB group (Table 2). Of the 16 litters produced in AB group, 9 litters (56%) were cannibalized. For NAB, 8 litters were born, and 2 (25%) were cannibalized. The pairing-to-parturition interval did not differ significantly between groups and ranged from 19 to 31 d. Litter size was similar between groups (pups per litter: AB 6.2 +/- 3.7, NAB 6.8+/-1.5). The number of viable litters surviving to weaning age was 7 of 18 (39%) in the AB group and 6 of 8 (75%) in the NAB group (Table 2).

Table 2.

Breeding performance

	Treatment group
	AB			NAB
Cohort no.	1	2	Totala	1
Number of breeding pairs	8	10	18	8
Parturition rate	6/8 (75%)	10/10 (100%)	16/18 (89%)	8/8 (100%)
Litter mortality	5/8 (63%)	6/10 (60%)	11/18 (61%)	2/8 (25%)
Mean interval (d) between pairing and parturition	21.67	23	23.3	20.3
Viable litters	3/8 (38%)	4/10 (40%)	7/18 (39%)	6/8 (75%)
No. of C. bovis-positive litters	0	0	0	6
No. of C. bovis-negative litters	3	4	7	0

^aAverage provided where appropriate

Discussion

Techniques to produce SPF offspring are common in laboratory rodent colonies. Viruses and bacteria have been eradicated successfully during gestation through weaning in mice by embryo transfer, cesarean section, cross fostering, and antimicrobial administration.^6,7,26,27 In the current study, we demonstrate the use of metaphylactic antibiotic treatment to control C. bovis infections in adult breeding pairs of NSG mice to produce C. bovis-free offspring. The C. bovis-negative status of offspring was confirmed at 10 wk and, in a subset of offspring, at 31 wk after weaning. The success of this method depends on antibiotic-induced elimination of viable C. bovis from the skin of infected mice after appropriate antibiotic treatment.⁵ We hypothesized that if the skin of breeding pairs are culture-negative for C. bovis, the risk of pups acquiring an infectious dose would be diminished or eliminated. In addition, eliminating bacterial shedding from infected mice would eliminate intracage contamination. Here we demonstrated that metaphylactic administration of amoxicillin–clavulanic acid decreased C. bovis skin copy numbers based on qPCR analysis, and that mice became culture-negative while on antibiotics (Figure 2 and Table 1). When combined with weekly cage changes during gestation and changing gloves prior to handling anything within each cage, transmission of infection to neonates and weanlings was prevented. In combination, metaphylactic antibiotic treatment and specific management practices can be used as a tool to facilitate the propagation of C. bovis-free NSG mice from breeders infected with C. bovis.

C. bovis is a facultatively anaerobic, gram-positive bacterium that is susceptible to a wide range of antibiotics.³⁰ For this study, we chose amoxicillin–clavulanic acid (amoxicillin trihydrate–clavulanate potassium) due to its broad spectrum of activity, ease of administration through drinking water, efficacy against C. bovis,³⁰ ability to reach the minimal inhibitory concentration of amoxicillin for C. bovis at a reasonable oral dose,¹⁷ and routine use at our facility with a history of reliable consumption by research mice. Our treatment is regarded as metaphylactic because we do not expect the antibiotic treatment to cure the adult mice. As previously demonstrated, infected mice placed on antibiotics for more than 8 wk became culture-positive 0 to 55 d after treatment was discontinued.⁵ However, experimental infection with C. bovis can be prevented by prophylactic administration of oral amoxicillin–clavulanic acid.¹⁶ To show that metaphylactic antibiotic treatment reliably creates a C. bovis-free ‘window’ only while mice are on treatment, we discontinued antibiotics for breeders after weaning the offspring and monitored for the reemergence of C. bovis. Concurrently, to decrease the potential for cross-contamination to offspring, weanlings in the AB group were moved into a separate room, which was entered before entry into the room containing both the AB and NAB breeder adults and NAB offspring. At 10 wk after the removal of antibiotic treatment from the breeders in the AB group, 57% (4 of 7) were PCR-positive and 34% (3 of 7) were culture-positive (Figure 2 and Table 1). These results continue to support that antibiotic therapy alone does not eliminate established C. bovis infections from immunodeficient mice. In contrast, by monitoring the pups in the AB group for 10 wk after weaning, we show that metaphylaxis successfully protected offspring in AB group from infection.

In our study, we could not determine when neonates began to ingest antibiotics. Although we did not quantify the antibiotic dose obtained by pups while nursing, some literature suggests that antibiotics can be transported across the mammary gland and reach the nursing pups.¹⁸ Antibiotic metabolites are similar to blood plasma components and, depending on factors including concentration gradients, protein affinity, lipid solubility, and state of ionization, they can successfully penetrate breast milk.¹⁸ When considering the safety of amoxicillin–clavulanic acid in regard to a human fetus, the American Academy of Pediatrics Committee on Drugs considers amoxicillin–clavulanic acid and penicillin-like drugs safe for pregnant mothers to use.¹ No harmful effects have been reported in infants who consumed breast milk that potentially contained amoxicillin or clavulanic acid.² In addition, the milk:plasma ratio reportedly is very low when oral penicillin-like drugs and clavulanic acid are prescribed orally to nursing human mothers.² However, once mouse pups transition from nursing to drinking the provided antibiotic-containing water, we anticipate that they would consume a prophylactic dose prior to weaning and separation from the dam. Despite clarity regarding whether antibiotics consumed by the pups contributed to preventing infection transmission, the offspring of the AB group were diagnostically negative for bacteria.

Antimicrobial stewardship has become an area of focus across many professions.^10,13,21,22 As the incidence of bacterial resistance grows and alternative therapeutics are scant, prudent use is mandated to keep the population safe. The metaphylactic manner in which this study deploys antibiotics adheres to antimicrobial stewardship principles. By placing infected adults on antibiotics prior to birth, offspring are able to mature in a C. bovis-free environment, preventing infection. C. bovis can affect research and cause mice to become clinically ill as well as to propagate the bacteria as chronic carriers. Bacterial shedding results in environmental contamination, ultimately perpetuating the infection in other susceptible mice. Giving these mice antibiotics in conjunction with facility-wide measures to clear the infection will ultimately reduce overall antibiotic use.

At our institution, nude mice are most likely to be from commercial stock and, when naïve, can easily be replaced. In contrast, NSG mice or mice with similar genetic deficits have been crossed with other strains to produce new strains that are not commercially available. These mice are extremely valuable to researchers, eliminating depopulation as an option for eradicating infectious agents. Therefore, we chose NSG mice for this study because several valuable NSG crosses required C. bovis eradication at our facility. Additional studies must be conducted to determine whether this technique can be applied to other strains of immunocompromised mice. Our data support the hypothesis that when therapeutic levels of antibiotics are administered, viable C. bovis is eliminated from the skin of infected adult breeding pairs, suggesting that similar outcomes can be expected in other mouse strains.

In the current study, pup mortality occurred within the first 3 d of birth. Perinatal pup mortality in 0% to 50% of litters has been reported in previous studies.^4,14,31,32 For AB cohorts 1 and 2, pup mortality from breeding pairs occurred in 3 of 6 (50%) and 6 of 10 (60%) litters, respectively. In contrast, only 2 of 6 (25%) litters in the NAB group experienced pup mortality. Little research exists to explain preweaning or perinatal pup mortality. Some studies suggest that litter survivability increases when nest building behaviors are displayed and when the dam spends increased time within the nest.³³ Whereas parity does not affect litter loss, strain appears to have some predictive value for litter survivability.³² Documented strategies to improve litter survivability include supplying the cage with the appropriate type and quantity of nesting materials,¹² reducing the frequency of cage changing, and providing environmental enrichment like polycarbonate houses and tissue paper.¹⁴ As a result, not only the increased cage-change frequency after pairing but also the complete removal and replacement of nesting material might have contributed to poor litter survivability of the AB group as compared with the NAB group. Nevertheless, our proposed method appears to affect pup viability, which will require confirmation through future studies. However, in practice, once the breeding animals have received initial antibiotic treatment and cage changing, subsequent litters can be produced by maintaining breeder animals on antibiotics alone, thus potentially increasing litter survivability for subsequent litters. Although we did not test this method on subsequent litters from antibiotic-treated breeding pairs, we have no reason to suspect that antibiotics would become ineffective for subsequent litters from the same breeders.

We found discrepancies among the results from culture and qPCR diagnostics throughout the study. When results were analyzed at pairing of breeding animals, 4 of 13 (32%) breeding pairs were qPCR-positive but culture-negative. For facility-wide monitoring measures, qPCR analysis is the test of choice.⁸ However, qPCR testing has little reliability in definitively determining C. bovis status (positive or negative) during antibiotic administration. Statistical evaluation revealed that these 2 diagnostic tests have moderate to good reliability, according to intraclass correlation assessment.

Potential limitations of this study include the experimental nature of the infection tested in mice. For nonexperimental applications of this method, animals are likely to be infected chronically, whereas our mice were infected for a minimum of 3 wk. Even with this relatively short duration of infection, our data mirror natural C. bovis infection, because the trend for C. bovis to return after the withdrawal of antibiotics was preserved.⁵ Although mice were culture- and qPCR-positive for C. bovis, the majority did not display clinical signs of infection and appeared healthy. This situation may have had a positive influence on breeding success, given that clinically affected NSG mice exhibit rough hair coat, decreased body condition, dehydration, scaly skin, alopecia, conjunctivitis, and erythematous pinnae.^5,9,25 These signs contribute to the animal's overall dull demeanor, which may lead to low fecundity. The AB group required 2 cohorts of breeding to reach statistically significant sample size. Even though the cohorts did not differ significantly, seasonality could be a factor. Another variable between the AB and NAB groups was the initial 3 wk of increased cage changing after pairing the breeders in the AB groups. We hypothesize that the period of increased cage-changing decreased litter viability in the AB group; a sham cage manipulation was not applied to the NAB group. Despite these limitations, the ability to produce C. bovis-free offspring with management practices and antibiotics is valuable when more invasive eradication methods are not feasible.

C. bovis affects many studies involving immunocompromised mice.²⁹ Eradication is a difficult and expensive process, but may ultimately result in the use of fewer animals in research. Our method aligns antimicrobial stewardship and reduction of animal numbers to achieve immunocompromised mouse offspring free of C. bovis. Our study demonstrates a novel way to eradicate C. bovis from immunocompromised mouse strains by using metaphylactic antibiotics in infected immunocompromised breeding pairs. The ability to maintain unique noncommercial immunocompromised strains during facility wide eradication of C. bovis is essential to research groups that would otherwise be unable or unwilling to use expensive, time-consuming, and invasive eradication methods.